mechanistic studies of the methylthiolation reaction

The Pennsylvania State University

The Graduate School

Department of Chemistry

MECHANISTIC STUDIES OF THE METHYLTHIOLATION REACTION

CATALYZED BY THE RADICAL SAM ENZYME RIMO

A Dissertation in

Chemistry

by

Bradley J. Landgraf

2016 Bradley J. Landgraf

Submitted in Partial Fulfillment

of the Requirements

for the Degree of

Doctor of Philosophy

December 2016

The dissertation of Bradley J. Landgraf was reviewed and approved* by the following:

Squire J. Booker

Professor of Chemistry

Professor of Biochemistry and Molecular Biology

Dissertation Advisor

Chair of Committee

Carsten Krebs


Professor of Biochemistry and Molecular Biology

Amie K. Boal

Assistant Professor of Chemistry

Assistant Professor of Biochemistry and Molecular Biology

William O. Hancock

Professor of Bioengineering

Kenneth Feldman


Graduate Program Chair

*Signatures are on file in the Graduate School

iii

ABSTRACT

The S12 protein, a component of the bacterial 30S subunit of the ribosome, contains a

universally conserved aspartic acid at position 89 (D89) in Escherichia coli (Ec). D89 is the target

of a unique post-translational modification (PTM), methylthiolation (–SCH3), at its C3 position to

form 3-methylthioaspartyl 89. This reaction is chemically challenging, requiring the activation

of an unactivated sp3-hybridized carbon center for insertion of the -SCH3 group. The enzyme

responsible for catalyzing this PTM is RimO (ribosomal modification O), a member of the

superfamily of enzymes named radical SAM (RS). RS enzymes reductively cleave S-adenosyl-L-

methonine (SAM) upon its binding to a reduced [4Fe-4S] cluster to generate methionine and a 5'-

deoxyadenosyl 5'-radical (5'-dA•), a potent oxidant. This radical is used to abstract one of the

prochiral hydrogen atoms from C3 of D89, thereby activating it for sulfur- or methylthio-

insertion. The protein ligates an additional [4Fe-4S] cluster, known as an auxiliary cluster, in its

N-terminal region. This auxiliary cluster is thought to be a sacrificial source of sulfide for the

methylthiolation reaction in vitro. Radical recombination between the substrate and a µ3-sulfido

ion of the auxiliary cluster would result in a thiolated intermediate of D89. In addition to its use

of SAM as a precursor to a 5'-dA•, RimO also catalyzes the transfer of a methyl group from a

second molecule of SAM, presumably to the sulfur atom of the thiolated intermediate of D89, or

to an acceptor site on the RimO polypeptide that acts as an intermediary that then transfers the

methyl group to the inserted sulfur atom and completes the reaction.

Biochemical experiments described in chapter 2 determined that RimO catalyzes methyl

transfer from SAM to an acceptor site on itself in the absence of a chemical reductant or an S12

peptide substrate, indicating that methyl transfer precedes radical chemistry. Radiotracing studies

in which RimO was incubated with [14

C-methyl]-SAM and the mixture subsequently separated by

size-exclusion chromatography demonstrated that radioactivity was associated with the protein.

iv

HPLC analysis of the protein fraction using an acidic mobile phase resulted in the complete loss

of radioactivity from SAM-derived breakdown products, which suggested that the radioactivity

was liberated upon treatment with acid. GC-MS analysis of headspace injections taken from

sealed vials containing RimO incubated with SAM or [methyl-d3]-SAM showed production of

methanethiol or d3-methanethiol, respectively, indicating that the acid-labile auxiliary [4Fe-4S]

cluster was the methyl acceptor site. This methylated cluster intermediate was shown to be

chemically and kinetically competent, and the presence of methanethiol in reaction mixtures of

RimO resulted in its incorporation in the S12 peptide substrate and also enabled the enzyme to

catalyze ~ 3 turnovers.

RimO from the gut bacterium Bacteroides thetaiotaomicron (Bt) was characterized and

shown to be similar to RimO from T. maritima in chapter 3. One of the key differences of Bt

RimO was the fact that the flavodoxin/flavodoxin reductase/NADPH (Fld/Fdx/NADPH) reducing

system from Ec was a competent source of electrons required for catalysis, thereby obviating the

use of the chemical reductant sodium dithionite. Use of the Fld/Fdx/NADPH reducing system

decreased the amount of 5'-deoxyadenosine (5'-dAH) formed abortively—meaning uncoupled

from methylthiolated product formation—but did not eliminate it. The flavodoxin semiquinone

was used as a spectroscopic handle to estimate that Bt RimO uses ~ 1 electron for each

methylthiolated product formed. It was determined that Bt RimO does not harbor any additional

sulfide or persulfide species through the use and quantification of the fluorescent sulfur-labeling

reagent, I-AEDANS.

In chapter 4, chemoenzymatic synthesis of 3-pro-R and 3-pro-S deuterium-labeled

aspartic acid was achieved by exploiting the enzymatic reaction catalyzed by aspartate ammonia-

lyase. The resulting product identities and retention of the deuterium label were confirmed by 1H

NMR after orthogonal protecting groups appropriate for solid phase peptide synthesis (SPSS)

were added. The labeled and protected aspartic compounds were incorporated into synthetic

v

peptides corresponding to residues 83-95 of the S12 protein by SPSS, and confirmation of the

correct peptide and retention of the deuterium label was obtained by MALDI-TOF MS. When

RimO from Bt and Tm were incubated under turnover conditions with the pro-R or pro-S labeled

S12 peptide substrates, observation of deuterium incorporation into 5'-deoxyadenosine occurred

only with the peptide substrate containing deuterium at the pro-S position, indicating that the

enzyme stereoselectively abstracts the pro-S hydrogen atom from its target substrate. This finding

also established the stereochemical course of methylthiolation to occur with inversion of

configuration, and the apparent primary kinetic isotope effect for H-atom versus D-atom

abstraction of ~ 1.9 indicates that this step is at least partially rate-limiting. Additionally, a large

apparent secondary isotope effect of ~1.4 was observed with the pro-R labeled substrate.

In chapter 5, the Tm S12 protein was overproduced in Ec, purified from inclusion bodies

under denaturing conditions, and slowly refolded to yield homogenous protein after size-

exclusion chromatography. S12 was shown to be a competent substrate when incubated with Tm

RimO under turnover conditions, with incorporation of an -SCH3 group into S12 observed by

MALDI-TOF MS. Variant proteins of Tm RimO in which one conserved amino acid found in the

protein active site was substituted by site-directed mutagenesis were characterized; the specific

Tm RimO variant were as follows: K12A, K12Q, Y227A, Y227F, and Q192A. Substitution of

K12 with alanine or glutamine abolished 5'-dAH and methylthiolated product formation and

decreased both the amount and rate of SAH formation, but did not affect the ability of the K12A

variant to bind SAM. These results suggested that the lysine residue may play a minor role in

methyl transfer, but is required in some unknown capacity for generation of the 5'-dA•. Both

Y227A and Y227F Tm RimO variants catalyzed methyl transfer and formation of 5'-dAH, but in

neither reaction was the methylthiolated product observed. Reactions containing either the

Y227A or Y227F variant in ~60% D2O resulted in no deuterium enrichment into 5'-

deoxyadenosine, suggesting that the 5'-dA• abstracts a hydrogen atom from a site on the S12

vi

peptide or the RimO protein that is not solvent exchangeable. Substitution of F for Y at position

227 had little effect on the determined dissociation constant for SAM binding compared to the

wild-type enzyme as determined by ITC. The Q192A variant was capable of catalyzing the full

methylthiolation reaction, albeit it to lower extents and at slower rates compared to the wild-type

enzyme, making the role for this conserved residue nebulous.

vii

TABLE OF CONTENTS

List of Figures .......................................................................................................................... x

List of Tables ........................................................................................................................... xiv

Acknowledgements .................................................................................................................. xv

Chapter 1 Methylthiolation: A post-translational modification of a bacterial ribosomal

protein catalyzed by an enzymatic radical-mediated reaction.......................................... 1

The ribosome: an elaborate, ornately decorated machine ........................................ 1 Identification of a novel post-translational modification of the S12 protein in

the bacterial 30S ribosomal subunit ................................................................. 2 Methylthiolation: a rare modification in biology ..................................................... 6 Discovery of the radical SAM superfamily of enzymes .......................................... 7 Radical SAM enzymes use a [4Fe-4S]

1+ cluster to reductively cleave SAM ........... 8

SAM: the universal methyl donor in the cell ........................................................... 11 Discovery and characterization of the first gene product that catalyzes

methylthiolation ................................................................................................ 12 Sulfur-inserting RS enzymes ligate a second iron-sulfur cluster ............................. 13 Identification of the gene product responsible for methylthiolation of D89 ............ 15 The methylthiotransferase subfamily of radical SAM enzymes .............................. 17 Characterization of RimO from E. coli .................................................................... 18 Characterization of RimO from T. maritima ............................................................ 22 A methylated cluster intermediate in RimO ............................................................. 28 The stereochemical course of the RimO reaction..................................................... 30 A proposed mechanism for the methylthiolation of D89 of the ribosomal S12

protein ............................................................................................................... 30 Conclusions ...................................................................................................................... 33 References ........................................................................................................................ 34

Chapter 2 Identification of an Intermediate Methyl Carrier in the Radical S-

adenosylmethionine Methylthiotransferase RimO ........................................................... 37

Introduction ...................................................................................................................... 38 Materials and Methods ..................................................................................................... 42 Materials........................................................................................................................... 42 Methods ............................................................................................................................ 42

Preparation of Substrates for Tm RimO Reactions .................................................. 42 Cloning and Overexpression of the Tm rimO gene ................................................. 44 Purification of Tm RimO ......................................................................................... 44 Protein, Iron, and Sulfide Quantification ................................................................. 45 Chemical Reconstitution of Tm RimO ..................................................................... 46 Tm RimO Activity Assays ....................................................................................... 46

viii

Tm RimO Radioactivity Assays ............................................................................... 47 Determination of Tm RimO-Dependent Production of Methanethiol ..................... 48 Tm RimO Differential Labeling Assays .................................................................. 48

Results .............................................................................................................................. 49 Turnover by Tm RimO ............................................................................................. 51 Radiotracing methyl transfer from ([methyl-

14C])SAM to Tm RimO ...................... 52

Tm RimO-Catalyzed Formation of Methanethiol .................................................... 56 Turnover in the Presence of Exogenously Supplied Methanethiol .......................... 58 Chemical and Kinetic Competence of a Potential Intermediate ............................... 61

Discussion ........................................................................................................................ 64 References ........................................................................................................................ 70

Chapter 3 Characterization of RimO from the mesophilic gut bacterium Bacteroides

thetaiotaomicron .............................................................................................................. 72

Introduction ...................................................................................................................... 72 Materials and Methods ..................................................................................................... 74 Materials........................................................................................................................... 74 Methods ............................................................................................................................ 75

Cloning and overexpression of the Bt rimO gene .................................................... 75 Purification of Bt RimO ........................................................................................... 76 Construction, overexpression, and purification of the Y225F variant of Bt

RimO ................................................................................................................ 77 Protein, Iron, and Sulfide Quantification ................................................................. 77 Chemical Reconstitution of Bt RimO ....................................................................... 78 Determination of the oligomeric state of Bt RimO .................................................. 78 EPR characterization of the Fe/S clusters of Bt RimO ............................................. 79 Bt RimO Activity Assays ......................................................................................... 80 Determination of Persulfide Content of Bt RimO by Fluorescent Labeling ............ 81 Quantification of flavodoxin semiquinone with Bt RimO under turnover

conditions ......................................................................................................... 82 Results .............................................................................................................................. 83

Cloning and overexpression of the Bt rimO gene .................................................... 83 Analysis of Fe/S cluster content by quantitative Fe and S analyses and EPR

spectroscopy ..................................................................................................... 86 Determination of the oligomeric state of Bt RimO .................................................. 89 Determination of Bt RimO WT and Y225F activity with dithionite or the Ec

flavodoxin reducing system .............................................................................. 92 Determination of persulfide content of Bt RimO by fluorescent labeling ................ 95 Quantification of flavodoxin semiquinone consumption by Bt RimO under

turnover conditions ........................................................................................... 100 Discussion ........................................................................................................................ 105 References ........................................................................................................................ 110

Chapter 4 The Stereochemical Course of the Reaction Catalyzed by the Radical SAM

Methylthiotransferase RimO ............................................................................................ 112

Introduction ...................................................................................................................... 112 Materials and Methods ............................................................................................. 114

ix

Materials ................................................................................................................... 114 Methods .................................................................................................................... 115 Cloning and Overexpression of the Ec aspA gene ................................................... 115 Purification of Ec AspA ........................................................................................... 116 Chemoenzymatic syntheses of (2S,3R)-3-[

2H1] aspartic acid (pro-R) and

(2S,3S)-[2,3-2H2] aspartic acid (pro-S) and their incorporation into

synthetic S12 13-mer peptide substrates .......................................................... 117 Determination of the stereospecificity of hydrogen atom abstraction by Bt

RimO ................................................................................................................ 120 Results .............................................................................................................................. 122 Discussion ........................................................................................................................ 128 References ........................................................................................................................ 131

Chapter 5 Assessment of Tm RimO activity with the Tm S12 protein as a substrate and

biochemical and biophysical characterization of Tm RimO active site variants .............. 133

Introduction ...................................................................................................................... 133 Materials and Methods ..................................................................................................... 136

Materials ................................................................................................................... 136 Methods .................................................................................................................... 137 Cloning and overexpression of the Tm rpsL (S12) gene .......................................... 137 Purification of Tm S12 ............................................................................................. 138 Activity assays with Tm S12 .................................................................................... 139 MALDI-TOF analysis of the Tm RimO reaction with Tm S12 ............................... 140 Site-directed mutagenesis, overexpression, and purification of Tm RimO

variants ............................................................................................................. 140 Quantitative iron and sulfide analyses and concentration determination of Tm

RimO variants ................................................................................................... 141 Activity and methyl transfer assays with Tm RimO variants................................... 141 Determination of dissociation constants for SAM or SAM analogues with Tm

RimO wild type and active site variants by isothermal titration calorimetry ... 142 Results .............................................................................................................................. 142

Cloning and overexpression of the Tm S12 gene ..................................................... 142 Purification of Tm S12 ............................................................................................. 143 Tm RimO activity assays with Tm S12 .................................................................... 144 Identification of conserved active site residues from sequence alignments and

the Tm RimO crystal structure .......................................................................... 147 Overexpression, purification, and characterization of Tm RimO variants ............... 149 Assessment of methyl transfer activity of Tm RimO variants .................................. 152 Assessment of methylthiolation activity of Tm RimO variants ................................ 155

Discussion ........................................................................................................................ 163 References ........................................................................................................................ 171

x

LIST OF FIGURES

Figure 1-1. D89 and the chemical derivitizations conducted to determine its post-

translational modification................................................................................................. 5

Figure 1-2. Post-transcriptionally modified adenosine 37 of tRNAs with a methylthio

group. ............................................................................................................................... 7

Figure 1-3. SAM bound via its α-carboxy- and α-amino groups to the [4Fe–4S] cluster of

PFL activase. .................................................................................................................... 9

Figure 1-4. X-ray crystallographic structures of the RS enzymes BioB and MoaA . ............ 10

Figure 1-5. Reductive cleavage of SAM to generate the 5'-deoxyadenosyl radical.. ............. 10

Figure 1-6. Methyl transfer from SAM to a nucleophilic acceptor via a polar SN2

mechanism.. ..................................................................................................................... 11

Figure 1-7. The reactions catalyzed by the sulfur-inserting RS enzymes BioB and LipA . ... 14

Figure 1-8. X-ray crystal structure of the Thermus thermophilus S12 protein in complex

with 16S rRNA, a 4-U mRNA codon mimic, and a 17-nucleotide anticodon stem

loop mimic of tRNA. ...................................................................................................... 16

Figure 1-9. The working mechanistic model for RimO proposed by Lee et al. ..................... 21

Figure 1-10. X-ray crystal structure of apo-Tm RimO.. .......................................................... 23

Figure 1-11. X-ray crystal structure of holo-Tm RimO.. ........................................................ 27

Figure 1-12. Electrostatic protein contact potential map determined from the X-ray

crystal structure of holo-Tm RimO. . .............................................................................. 28

Figure 1-13. The proposed mechanism for the methylthiolation of D89 of S12.. ................... 32

Figure 2-1. Reactions of the three major classes MTTases: MiaB; MtaB; and RimO. ........... 41

Figure 2-2. UV/vis spectra of AI and RCN Tm RimO. ............................................................ 50

Figure 2-3. Tm RimO-catalyzed reactions at 37 °C. ................................................................ 52

Figure 2-4. Elution profiles of Tm RimO incubated with [methyl-14

C]SAM or [adenosyl-14

C]SAM. . ...................................................................................................................... 54

Figure 2-5. HPLC elution profiles monitored at 260 nm of AGFC protein fraction from

Tm RimO incubated with [adenosyl-14

C]SAM or [methyl-14

C]SAM. ............................. 55

Figure 2-6. GC-MS total ion chromatogram of methanol at various concentrations using

single-ion monitoring at m/z = 31. . ................................................................................. 57

xi

Figure 2-7. Time-dependent formation of SAH and methanethiol by Tm RimO. .................. 58

Figure 2-8. Time-dependent formation of MS-1 in the presence of 1 mM methanethiol

and 2 mM SAM or d3-SAM. ............................................................................................ 59

Figure 2-9. Time-dependent formation of MS-1 and d3-MS-1 in the presence of 2 mM

methanethiol and 2 mM d3-SAM. .................................................................................... 61

Figure 2-10. Time courses for the formation of 5'-dA, SAH, MS-1, d3-MS-1, and

consumption of 1 by Tm RimO incubated with d3-SAM for 3 min after previous

incubation with unlabeled SAM for -15 h followed by AGFC. ....................................... 63

Figure 2-11. Time courses for the formation of MS-1 and d3-MS-1 by Tm RimO

incubated with d3-SAM for 1 h or 3 h after previous incubation with unlabeled SAM

for 15 h followed by AGFC. . ......................................................................................... 63

Figure 2-12. Working hypothesis for the reaction catalyzed by Tm RimO.. ........................... 69

Figure 3-1. SDS-PAGE analysis of Bt RimO overexpression.. ............................................... 83

Figure 3-2. SDS-PAGE of Bt RimO purification.. .................................................................. 84

Figure 3-3. HR 26/60 Sephacryl S200 elution profile of Bt RimO. ........................................ 85

Figure 3-4. SDS-PAGE analysis of reconstituted and S200-purified Bt RimO.. ..................... 86

Figure 3-5. EPR spectra of 400 µM Bt RimO RCN ................................................................ 88

Figure 3-6. Molecular-sieve chromatographic analysis of Bt RimO RCN.. ........................... 90

Figure 3-7. LC-MS analysis of the reaction of 100 µM Bt RimO RCN with 1 mM SAM,

1 mM 13 mer peptide substrate, and either the Fld/FldR/NADPH reducing system

reducing system or dithionite. .......................................................................................... 93

Figure 3-8. Active site of Tm RimO. ....................................................................................... 96

Figure 3-9. Labeling of protein-bound persulfide by the fluorescent dye 1,5-I-AEDANS.. ... 97

Figure 3-10. Standard curves of 1,5-I-AEDANS. .................................................................... 98

Figure 3-11. The electronic forms of the flavin mononucleotide cofactor. ............................. 101

Figure 3-12. Time-dependent formation of 5'-dAH, SAH, unlabeled MS-1 product, d3-

labeled MS-1 product, and time-dependent consumption of Fld SQ by Bt RimO

RCN. ................................................................................................................................ 105

Figure 4-1. 1H NMR spectrum of (2S, 3R)-3-[

2H1] Fmoc-N-aspartic acid β-tert-butyl ester

(pro-R 3-[2H1]-aspartate). ................................................................................................. 118

xii


2H1] Fmoc-N-aspartic acid β-tert-butyl ester

(pro-S 2,3-[2H1]-aspartate) ............................................................................................... 119

Figure 4-3. Bt RimO-catalyzed time-dependent formation of SAH, 5'-dAH, and

methylthiolated product (3-MS-1) with the Ec Fld/FldR/NADPH reducing system. ..... 123

Figure 4-4. Quantification of 5'-dAD generated in the reaction of Bt RimO conducted in

90% D2O with unlabeled peptide. .................................................................................... 124

Figure 4-5. Synthetic routes for (2S,3R)-3-[2H1] Fmoc-N-aspartic acid β-tert-butyl ester

(pro-R) and (2S,3S)-[2,3-2H2] Fmoc-N-aspartic acid β-tert-butyl ester (pro-S). .............. 125

Figure 4-6. 1H NMR spectra from 2.7 to 4.7 ppm of unlabeled , pro-R labeled, and pro-S

labeled aspartate . ............................................................................................................. 126

Figure 4-7. Bt RimO catalyzed reactions at 37 ˚C in the presence of Ec

Fld/FldR/NADPH, SAM and peptides 1, 2, or 3.. ........................................................... 127

Figure 4-8. Quantification of methionine generated in the Bt RimO reaction.. ....................... 128

Figure 5-1. SDS-PAGE of the overexpression of the S12 gene from Thermotoga

maritima in E. coli BL21(DE3) cells. .............................................................................. 143

Figure 5-2. SDS-PAGE of the purification of Tm S12 under denaturing conditions............... 144

Figure 5-3. MALDI-TOF mass spectra of Tm S12. ................................................................. 146

Figure 5-4. Time-dependent formation of SAH, and 5'-dAH, by 200 µM Tm RimO in the

presence of 2 mM SAM, 2 mM sodium dithionite, and 150 µM Tm S12 protein. .......... 147

Figure 5-5. Sequence alignment of RimO proteins from 11 different bacterial species.. ........ 148

Figure 5-6. Active site from the crystal structure of Tm RimO.. ............................................. 149

Figure 5-7. SDS-PAGE of Tm RimO variants following purification by IMAC and size-

exclusion chromatography.. ............................................................................................. 150

Figure 5-8. UV-Visible spectra of Tm RimO WT and K12A, K12Q, Y227A, Y227F, and

Q192A variants normalized to the maximum absorbance at 280 nm. ............................. 151

Figure 5-9. Time-dependent formation of SAH by 100 µM Tm RimO wild type and

K12A, K12Q, Y227F, Y227A, and Q192A variants in the presence of 1 mM SAM

over 3 h ............................................................................................................................ 153

Figure 5-10. Time-dependent formation of SAH (A & B), 5'-dAH (C & D), and MS-1

product (E & F) by 100 µM Tm RimO wild type and K12A, K12Q, Y227F, Y227A,

and Q192A variants. ........................................................................................................ 158

Figure 5-11. Isothermal titration calorimetry in which 650 µM SAM or TeSAM was

titrated into 150 µM Tm RimO WT. .............................................................................. 162

xiii

Figure 5-12. Isothermal titration calorimetry in which 800 µM SAM or 1.5 mM SAM

was titrated into 150 µM Tm RimO K12A or 125 µM Tm RimO Y227F. . ................... 163

Figure 5-13. Working hypothesis for the reaction catalyzed by Tm RimO.. ........................... 170

xiv

LIST OF TABLES

Table 3-1. Fit parameters of Bt RimO reactions containing SAM, a synthetic peptide

substrate, and the flavodoxin reducing system or dithionite as the reductant. ................. 94

Table 3-2. Summary of results of 1,5-I-AEDANS labeling of persulfides present on Bt

RimO RCN. ...................................................................................................................... 100

Table 3-3. Fit parameters of pre-methylated Bt RimO RCN reactions containing [methyl-

d3]SAM, a synthetic peptide substrate, and flavodoxin semiquinone. ............................. 105

Table 5-1. The forward and reverse primers used to make Tm RimO variants. ..................... 141

Table 5-3. Fit parameters of methyl transfer reactions containing 100 µM of the indicated

RimO protein and 1 mM SAM......................................................................................... 153

Table 5-4. Fit parameters of turnover reactions containing 100 µM of the indicated RimO

protein .............................................................................................................................. 159

xv

ACKNOWLEDGEMENTS

I am very proud of this document and the work detailed herein. It's been a long journey

up to this point, but the end isn't what this whole process was about. Sure, this will (hopefully!)

culminate in a doctoral degree, but ultimately, that's a certification. It is the knowledge and

experience that I gained along the way that made this pursuit and journey worth making.

Essentially what I'm trying to say is "It's about the journey, not the destination." Truer words have

never been spoken when it comes to describing earning a Ph.D.

Of course no one goes about these journeys alone, and I'm quite fortunate to have a select

group of incredible individuals in my life who have always supported me no matter what the

pursuit. My Dad and Stepmom are my rocks, my constants in life. They have always pushed me

to be the best I can possibly be, and they instilled in me the belief that anything is possible with a

lot of work, sacrifice, and some blood, sweat, and tears. I'm forever grateful for their advice, their

always open ears, their unwavering support, and their eternal optimism. I'm also grateful for my

Mom, who always puts things into perspective for me and also supports me and my decisions no

matter what. Being the oldest of 4 kids, there was a lot of pressure on me to set a good example.

For my siblings, I hope I've set the bar high, because only good things can come of it. There were

days when I didn't want to keep going, but I kept going because of all of you.

I've made some terrific friends through grad school. First and foremost is Nick, the older

brother I never had. Our conversations, both scientific and very unscientific, were always

enriching, or at the very least entertaining. The general disdain we held for most of those around

us—we really were the grumpy old men of the lab—and our commiserating helped me survive

the bad days. Liz also became an incredible friend over the years, and we talked as much about

science as we did about food I think. Some of the meme e-mails we had running still make me

laugh. And to both of you, thank you for being there for me and letting me crash at your place on

xvi

more than one occasion for more than one day at a time. You both truly have become like a

brother and sister to me.

Maria, where to begin? I'm not sure how your gigantic heart fits in such a small body.

Thank you for collaborating with me, for our chats, for that time you invited me to Greece, the

list goes on. I will always admire your intelligence, compassion, and work ethic. You truly are

the energizer bunny, and your success speaks for itself.

The two Matts: Matt Radle and Matt Bauerle. You guys could always make me laugh no

matter what the circumstances were and both of you were always dependable and there for me.

You're both incredible people in your own ways, and I'm very grateful to be friends with both of

you. The summer we went whitewater rafting was awesome, and the picture I have of all of us (so

young and so happy!) makes me smile every time I look at it. We made some good memories and

will make more in the future.

My mentor Kyung, you molecular biology wizard. Your presence in lab has been sorely

missed! Also, to all of the original Booker lab members I haven't mentioned yet, Tyler, Allison,

Doug, and Lauren, thanks for your patience with me as I learned to work with bacteria and for

being incredibly helpful both inside and outside of lab. Also, to all of my labmates, thank you for

putting up with my caffeine-induced "get everything done as fast as possible and don't let anyone

get in my way" bouts. Not everyone is as understanding as you all were.

Last but not least, I need to thank my committee members for agreeing to oversee this

Ph.D. pursuit, most of all Dr. Booker. Thank you for taking a chance on me. I wasn't one of the

all-stars with a background in biochemistry, but you agreed to take me on, and for that I will

always be grateful. I didn't realize just how very high the bar could be set until I joined your

group. You've taught me to be incredibly critical in both my thinking and data analysis and

writing manuscripts with you was like have my own personal writing tutor. You've got a way

with words that I'll always admire. I was and continue to be impressed by your breadth of

xvii

knowledge, not just of science, but of life, and your happy-go-lucky attitude and enthusiasm for

science is contagious.

Chapter 1

Methylthiolation: A post-translational modification of a bacterial ribosomal

protein catalyzed by an enzymatic radical-mediated reaction

The ribosome: an elaborate, ornately decorated machine

The ribosome is a paragon of Nature's ability to craft a finely tuned and well-oiled

macromolecular machine to synthesize the proteins and enzymes needed for "life." Indeed, the

ribosome is "probably the most sophisticated machine ever made" (1). Protein synthesis is high

stakes work requiring both speed and accuracy for survival. Polypeptide formation needs to be

fast to enable organisms to respond to sudden environmental changes, and fast it is: a 250 amino

acid protein is synthesized in approximately 15 seconds. The formation of peptide bonds

between amino acids also needs to be accurate: misincorporation of a single amino acid can

disrupt both proper protein folding and the correct placement of amino acids, which are crucial to

the functional integrity of a protein (2). The overall miscorporation rate of the bacterial ribosome

is estimated to be between 1 in 1000 and 1 in 10,000 (3).

The bacterial ribosome is composed of a smaller 30S subunit and a larger 50S subunit,

both of which are comprised of ribosomal RNA (rRNA) and proteins. Specifically, the 30S

subunit contains 16S rRNA and 23 proteins, S1-S23, and the 50S subunit is constituted of 23S

and 5S rRNA and 33 proteins, L1-L36 (4). While these two subunits associate to form the fully

assembled ribosome, they each play distinct roles during translation. The 30S subunit is charged

with maintaining translational fidelity by mediating interactions between anticodons of tRNAs

and codons of the mRNA being translated to determine the correct sequence of amino acids in the

synthesized protein (5). The 50S subunit harbors the peptidyl-transferase center where the

2

formation of peptide bonds of nascent proteins is catalyzed (6). Interestingly, the ribosome is

technically a ribozyme, given that it is the RNA that carries out the catalytic functions of protein

synthesis, whereas the proteins in both subunits mostly play structural roles (7).

The rRNA and some proteins of the ribosome are decorated with post-transcriptional and

post-translational modifications that further expand the chemical reactivity beyond that of the 4

nucleotides and 20 amino acids. The 33 post-transcriptional modifications of rRNA in E. coli

(Ec) consist of four 2'-O-methylations and 19 instances of methylation of other base sites, in

addition to 10 pseudouridylations and one further modified pseudouridine. The extent and

complexity of modifications varies between and within phylogenetic kingdoms, with archaeal and

eukaryotic rRNA typically more ornately adorned than that found in prokaryotes (4). Similarly, a

number of ribosomal proteins contain post-translational modifications in Ec. Specifically, 6

proteins are methylated (S11, L3, L11, L7/L12, L16, and L33), 3 are acetylated (S5, S18, and

L7), additional glutamic acid residues are appended to S6, some C-terminal amino acids are

removed from L31, the N-terminal methionine is removed from 37 of the 57 proteins, and last but

not least, the S12 protein is methylthiolated (-SCH3) (Figure 1-1A) (8-11). Like the RNA

modifications, these protein alterations vary in their extent and complexity, with eukaryotic

ribosomal proteins containing more of these decorations than prokaryotes.

Identification of a novel post-translational modification of the S12 protein in the bacterial

30S ribosomal subunit

Of the post-translational modifications (hereafter PTM) of ribosomal proteins in Ec, the

exact modification of the S12 protein was among the last to be determined, spanning some 30

years of research. In 1977, the first indication that Ec S12 contained a PTM was the inability to

3

assign an identity to the amino acid at position 89 (X89) of the protein following peptide mapping

and subsequent Edman degradation (12). The preceding residue was identified as lysine, and it

was observed that the amide bond between K88 and X89 was not cleaved by trypsin. The authors

cited their unpublished results in which they determined aspartic acid as the residue at position 89

in several Ec mutant strains with streptomycin resistance; however, they never established nor

published whether the mutation occurred at position 89, leaving the identity at this position in the

S12 protein ambiguous (12). Three years later, the nucleotide sequence of the rpsL gene that

encodes the S12 protein was determined, and the codon for residue 89 corresponded to aspartic

acid (13). Another 16 years passed until matrix-assisted laser desorption/ionization-time of flight

mass spectrometric (MALDI-TOF MS) analysis of the S12 protein isolated from purified 30S

subunits was conducted, which resulted in a spectrum in which two peaks were observed: a minor

peak corresponding to the predicted mass of S12 with position 89 occupied by aspartic acid

(13,605.2 Da), and a major peak +46 Da heavier than the D89 S12 protein (13652.1 Da). The

same mass increase of +46 Da was observed in MALDI-TOF MS analysis of tryptic digests of

S12, and the ratio of the minor and major peaks, corresponding to the unmodified and modified

peptide, was approximately 1:3 (11). Mass selection for the modified peptide and careful analysis

of the product ions produced by post-source decay both verified the peptide sequence and

identified unequivocally that D89 was the residue containing the modification (11).

The identity of the modification on S12 as methythiolation was established by chemical

derivatization and subsequent MS analysis. Performic oxidation of the unmodified and modified

peptides resulted in incorporation of two oxygen atoms in the latter (+32 Da) and no mass

changes observed for the former (Figure 1-1B). Given the overall mass change with the

modification present, the incorporation of two oxygen atoms was consistent with modified D89

containing a sulfur atom in a thioether (-SCH3) linkage that was oxidized to form a sulfone group

(-SO2CH3) (Figure 1-1B) rather than a methylthio (-CH2SH) linkage, which would be oxidized to

4

the corresponding sulfonic acid (-CH2SO3) with incorporation of three oxygen atoms (Figure 1-

1C) (11). Esterification of the modified and unmodified D89 containing peptides in methanolic

HCl resulted in mass increases of +28 Da in both peptides, and post-source decay analysis

identified the esterification sites to be the carboxylic acids of D89 and the C-terminus of the

peptide (Figure 1-1F). Interestingly, performic oxidation of the esterified peptide containing the

modification resulted in β-elimination (-80 Da) corresponding to loss of -SO2CH3 and -H from

the β- and α-carbons, respectively (Figure 1-1G). Last but not least, treatment of the modified

D89-containing peptide with Raney Ni, a catalyst that reductively removes sulfur from organic

substrates, resulted in a mass shift of -46 Da, indicating removal of the thiomethyl group and

corroborating the linkage as -SCH3 and not CH2SH (11) (Figure 1-1D, E). Collectively, these

results demonstrated unequivocally that the amino acid at position 89 of S12 is aspartic acid, that

this residue is the site of the PTM, and that the PTM is methylthiolation. The authors proposed

that the β-carbon, or C3, of D89 is the site of methylthiolation (Figure 1-1A) based on the fact

that there are no known PTMs occurring at the α-carbon of any amino acid, whereas β-carboxy

and β-hydroxy aspartic acid residues have been observed (11, 14, 15).

5

Figure 1-1. D89 and the chemical derivitizations conducted to determine its post-translational

modification (PTM). Removal of a hydrogen atom from C3 of D89 and subsequent insertion of a

thiomethyl- group gives rise to the methylthiolation PTM (A). Performic acid oxidations to

establish that the PTM linkage is via a thioether (-SCH3), resulting in an m/z shift of +32 Da

corresponding to incorporation of two oxygen atoms to form a sulfone moiety (B) and not a

methylthio (-CH2SH) linkage, which would result in formation of a sulfonic acid with an m/z

shift of +47 Da, which was not observed (C). Raney Nickel treatment of modified D89 peptide

resulted in desulfurization giving rise to an m/z shift of -46 Da and not m/z -32 Da, confirming

the linkage is via a thioether (D) and not a methylthio (E). Esterification of both the C-terminal

and β-carboxylic acid moieties resulted in an m/z shift of +28 Da (F). Performic acid oxidation

of the esterified D89 modified peptide resulted in β-elimination of the α-proton and the

methylsulfonate (G).

6

Methylthiolation: a rare modification in biology

The thiomethyl modification of S12 is interesting for a number of reasons, one of them

stemming from its rarity in biology. In fact, S12 is the only known protein to contain this

modification. Whereas most of the PTMs—methylation, acetylation, selective proteolysis—are

commonplace, the only other known instances of thiomethylation in bacteria are found in the

post-transcriptional modifications of adenosine 37 (A37) of certain transfer RNAs (tRNAs).

Specifically, N6-isopentenyladenosine 37 (i

6A37) of tRNAs that read codons starting with uridine

(except tRNASer I,V

) is methylthiolated at C2 to form 2-methylthio-N6-isopentenyladenosine

(ms2i6A) in Ec (Figure 1-2A) (16). In Salmonella typhimurium (St), Pseudomonas aeruginosa,

and other gram negative bacteria, the modified base N6-4-hydroxyisopentenyladenosine 37

(io6A), which is not found in Ec, is also methylthiolated at C2 to form 2-methylthio-N

6-(4-

hydroxy)isopentenyladenosine 37 (ms2io

6A) (Figure 1-2B) (17). Analogously, the same position

of N6-threonylcarbamoyladenosine 37 (t

6A) of tRNAs that read codons starting with adenosine is

methylthiolated to form 2-methythio-N6-threonylcarbamoyladenosine (ms

2t6A) (Figure 1-2C)

(18), and N6-4-hydroxynorvalyladenosine is methylthioated at C2 in thermophilic and some

mesophilic bacteria and archaea to form 2-methylthio-N6-4-hydroxynorvalyladenosine (Figure 1-

2D) (19). These hypermodified adenosines are all adjacent to the 3' end of the anticodon, and

their modifications are thought to stabilize codon-anticodon interactions and aid in the prevention

of frameshifting during translation (20).

7

Figure 1-2. Post-transcriptionally modified adenosine 37 of tRNAs with a methylthio group.

Discovery of the radical SAM superfamily of enzymes

In 2001, a seminal bioinformatics study by Sofia et al. discovered a superfamily of

enzymes that at that time consisted of 600 related enzymes all containing a Cx3Cx2C motif (21).

These enzymes were proposed to utilize S-adenosyl-L-methionine (SAM) in radical-mediated

reactions, and the superfamily and the enzymes within were appropriately named "Radical SAM"

(RS) (21). Several well characterized RS enzymes at that time were lysine 2,3-aminomutase

(LAM), biotin synthase (BioB), lipoyl synthase (LipA), and the activases of pyruvate formate-

lyase (PFL-AE) and anaerobic ribonucleotide reductase (RNR-AE). In the case of LAM and PFL-

AE, the SAM-derived byproducts of its reductive cleavage to methionine and 5'-deoxyadenosine

had been observed in in vitro activity assays (22, 23). Further experiments with LAM in the

presence of tritium labeled SAM showed the transfer of tritium between the 5'-carbon of the

deoxyadenosyl moiety of SAM and lysine, which suggested that LAM catalyzed formation of a

8

5'-deoxyadenosyl radical, akin to that observed in coenzyme B12 (adenosylcobalamin)-dependent

isomerization reactions (24-26). It is worth noting that since its discovery 15 years ago, the RS

superfamily has grown tremendously to include over 113,000 members that span the entire

phylogenetic kingdom (27).

Radical SAM enzymes use a [4Fe-4S]1+

cluster to reductively cleave SAM

RS enzymes have been shown to contain at least one [4Fe-4S]2+

cluster in their resting

states, requiring reduction by one electron to form the catalytically active [4Fe-4S]1+

cluster. This

"radical SAM" cluster, of which three of its four Fe ions are ligated by the cysteines comprising

the Cx3Cx2C motif, contains a free Fe site to which SAM binds via its α-amino and carboxylate

moieties (Figure 1-3) (28). Structural determinations of RS enzymes have shown them all to

contain a full (α8β8) or partial (αnβn) triose phosphate isomerase (TIM) barrel comprising a radical

SAM fold (Figure 1-4) (29-37), and of those structures in which SAM is bound to the RS cluster,

the sulfonium moiety of SAM was observed to be ~ 4 Å from both the Fe and S ions in the cluster

(29-32). This proximity is thought to facilitate electron transfer from the RS cluster into the C5'-

S bond of SAM to effect its homolytic cleavage to form methionine and a 5'-deoxyadenosyl 5'-

radical (5'-dA•) (Figure 1-5).

9

Figure 1-3. SAM bound via its α-carboxy- and α-amino groups to the [4Fe–4S] cluster of PFL

activase (3CB8). The carbon atoms of SAM and the cysteine ligands to the cluster are shown in

green and cyan respectively. Sulfur, iron, oxygen and nitrogen atoms are shown in yellow,

orange, red and blue, respectively. The 5′ carbon bearing the radical upon homolytic cleavage of

SAM is shown as a green sphere. The structures were prepared using the PyMOL Molecular

Graphics program (http://www.pymol.org).

10

Figure 1-4. X-ray crystallographic structures of the RS enzymes BioB (left) and MoaA (right).

The full (α8β8) triose phosphate isomerase (TIM) barrel of BioB is shown in pale green, while the

partial (α/β)6 TIM barrel of MoaA is shown in lavender. The iron and sulfur ions comprising the

clusters of each enzyme are depicted as orange and yellow spheres, respectively. The structures

were prepared using the PyMOL Molecular Graphics program (http://www.pymol.org). The PDB

accession codes for BioB and LipA are 1R30 and 1OLT, respectively.

Figure 1-5. Reductive cleavage of SAM to generate the 5'-deoxyadenosyl radical. The [4Fe-4S]2+

cluster is reduced by one electron to the catalytically active [4Fe-4S]1+

state. The source of

reducing equivalents in vivo is the flavodoxin:flavodoxin oxidoreductase:NADPH reducing

system, while sodium dithionite and illuminated deazaflavin are commonly used reductants in in

vitro studies. The binding of SAM via its α-amino and α-carboxy moieties to the free iron site of

the iron-sulfur cluster brings the sulfur atom of SAM within close proximity of the [4Fe-4S]1+

cluster to allow for electron transfer into this bond into an antibonding orbital to effect homolytic

cleavage of the C5’-S bond, thereby producing a 5'-deoxyadenosyl radical and L-methionine.

11

SAM: the universal methyl donor in the cell

Although some 113,000 individual sequences of RS enzymes have been identified by

bioinformatics (27), the use of SAM to generate the 5'-dA• is not its most common mode of

reactivity in biology. In fact, SAM is typically thought of as the universal methyl donor in the

cell. Over 300 SAM-dependent methyltransferases have been characterized, donating methyl

groups to substrates such as proteins, nucleic acids, phospholipids, and secondary metabolites

(38, 39). SAM is used by these methyltransferases as a methyl donor to impact myriad key

cellular processes, including transcription, translation, gene regulation, signal transduction, and

the biosynthesis of essential metabolites (38-40). These reactions exploit the electrophilicity of

the methyl moiety bound to the sulfonium of SAM, affording transfer of the methyl group to

nucleophilic acceptors by an SN2 mechanism (Figure 1-6).

Figure 1-6. Methyl transfer from SAM to a nucleophilic acceptor via a polar SN2 mechanism.

The carbon atom of the methyl group (red) is bound to the sulfonium of SAM (green), making the

carbon sufficiently electrophilic for attack by a nucleophilic moiety on an acceptor molecule,

12

resulting in transfer of the methyl group from SAM to the acceptor and formation of the methyl

transfer byproduct, S-adenosylhomocysteine (SAH).

Discovery and characterization of the first gene product that catalyzes methylthiolation

Studies of the biosynthesis of ms2i6A paved the way towards our initial understanding of

the biochemistry behind methylthiolation modifications. The gene product responsible for

insertion of the methylthio group in i6A37, named miaB in Ec, was first identified by genetic

studies in Ec and St in 1999 (41). Sequence alignments of MiaB protein homologues revealed a

conserved triad of cysteine residues residing in a Cx3Cx2C motif, and the authors noted that this

cysteine motif was also found in a set of enzymes known as "radical activating" enzymes that

presumably bind iron (42). Additionally, previous biochemical and genetics studies had shown

that methylthiolation of tRNAs was iron-dependent (43-45). Shortly thereafter, Ec MiaB was

overproduced and characterized spectroscopically and was shown to ligate an iron-sulfur cluster

of unknown configuration. The enzyme also required the presence of the iron-sulfur cluster to

methylthiolate i6A37 in vivo in a miaB

- Ec strain (46).

A similar study on MiaB isolated from the thermophilic bacterium Thermotoga maritima

(Tm) identified the Cx3Cx2C motif to be present and demonstrated that this MiaB could

complement the formation of ms2i6A in a miaB

- Ec strain at elevated temperatures (47). This

study also demonstrated that a second molecule of SAM was the source of the methyl group in

the methylthio- modification by tracing the transfer of a tritium radiolabel in [3H3-methyl]-SAM

to the methylthiolated tRNA substrate. Additionally, the cluster configuration was assigned to be

a [4Fe-4S]2+

cluster after analysis by UV-visible, Mössbauer, resonance Raman, and variable

temperature magnetic circular dichroism spectroscopic techniques, leading the authors to place

MiaB on the growing list of radical SAM enzymes (47). MiaB was also noted to be a bifunctional

RS enzyme, due to its use of SAM as both a precursor to a 5'-dA• and as a methyl group donor,

13

making the enzyme unique in its ability to exploit two different reactivities of SAM within the

same polypeptide (48). A spectroscopic and analytical study three years later on Tm MiaB

revealed that the enzyme housed two [4Fe-4S]2+

clusters, one of them being the RS cluster, and

the other an auxiliary cluster coordinated by three conserved cysteine residues in the N-terminal

region of the protein (49).

Sulfur-inserting RS enzymes ligate a second iron-sulfur cluster

The presence of the second cluster in MiaB was reminiscent of the RS enzymes biotin

synthase (BioB) and lipoyl synthase (LipA), which ligate a [2Fe-2S]2+

cluster and a [4Fe-4S]2+

,

respectively, in addition to their RS clusters. These two enzymes catalyze the insertion of sulfur

atoms into unactivated C–H bonds of their respective substrates, dethiobiotin and N-

octanoyllysine (Figure 1-7). The second cluster in each of these enzymes was proposed to be

sacrificed as a source of sulfur for the reaction, given that they both exhibit ~ one turnover in

vitro. Indeed, isotopic labeling of the clusters in these enzymes with [34

S]- or [35

S]-sulfide showed

transfer of the sulfur isotope into their respective products (50-52). Further spectroscopic

evidence supporting a sacrificial role for the second cluster was obtained for each enzyme. In the

case of BioB, hyperfine sublevel correlation spectroscopy (HYSCORE) spectra of the reduced

enzyme in the presence of its reaction intermediate, 13

C9-9-mercaptodethiobiotin, exhibited strong

cross peaks indicative of strong isotropic cross-coupling between the nuclear spin of 13

C9-9-

mercaptodethiobiotin and the reduced [2Fe-2S]1+

remnant cluster; modeling of a covalent linkage

between the reaction intermediate and the [2Fe-2S]1+

produced hyperfine coupling constants in

agreement with those experimentally observed, thereby capturing a spectroscopic snapshot of the

sulfur transfer in BioB (53). LipA was shown both biochemically and spectroscopically to form a

14

cross-linked intermediate through its auxiliary [4Fe-4S] cluster (54). Specifically, LipA and a

protein substrate, N-octanoyllysine lipoyl carrier protein (LCP), when incubated under turnover

conditions with coeluted during anion-exchange chromatography, but were separated when

fractions were analyzed by SDS-PAGE, which suggested the LCP cross-link to the iron-sulfur

cluster of LipA was destroyed upon aerobic degradation of the cluster, since a covalent linkage

would survive SDS-PAGE. Mössbauer spectoscopic analysis of the cross-linked intermediate

determined that one of the clusters, presumably the auxiliary cluster to which the cross-link was

bound, was no longer a [4Fe-4S] cluster, but rather most resembled a [3Fe-4S]0 cluster. The

partially disassembed cluster implied that one of its sulfur atoms was shared with the cross-linked

intermediate (54). Altogether, the similarities between the MiaB enzyme and the reaction it

catalyzes with those of BioB and LipA led to the conclusion that MiaB likely uses a similar

mechanism to form ms2i6a (49).

Figure 1-7. The reactions catalyzed by the sulfur-inserting RS enzymes BioB (A) and LipA (B).

BioB catalyzes the insertion of sulfur at C9 and then C6 of dethiobiotin to form biotin. This

reaction proceeds through the 9-mercaptodethiobiotin intermediate that is cross-linked to a sulfur

atom of the [2Fe-2S] cluster. LipA catalyzes sulfur insertion at C6 and C8 of N-octanoyllysine

LCP to form lipoyl-LCP. Analogous to the BioB reaction, this reaction proceeds through a cross-

link between C6 of the N-octanoyllysine chain a sulfur atom of the partially degraded [3Fe-4S]

cluster.

15

Identification of the gene product responsible for methylthiolation of D89

Due to the similarity between the ms2i6A modification catalyzed by MiaB and the

methylthiol modification of the S12 protein, a bioinformatics study was conducted to find

sequences in Ec that were similar to MiaB to identify possible protein candidates responsible for

the S12 modification. Only one gene product of unknown function, yliG, exhibited strong

similarity to the MiaB sequence, and because putative yliG homologues were identified in T.

thermophilus and R. palustris, the only other organisms in which methylthiolation of S12 had

been observed, further genetics and bioinformatics studies on yliG were conducted (55).

Two mutant Ec strains in which the S12 modification was absent were identified by

analyzing the S12 protein by MALDI-TOF MS. The modification was restored by introduction

of a plasmid harboring constitutively expressed yliG, thereby confirming that the loss of the

modification in the mutant strains was due to inactive yliG and also confirming that yliG was

indeed responsible for methylthiolation of S12 (55). The gene product was appropriately renamed

RimO (ribosomal modification O). Studies in a ΔrimO Ec strain identified only a slightly slower

growing phenotype, indicating that rimO confers a slight growth advantage in Ec, in addition to

these studies demonstrating that the modification is not essential in Ec (55). Interestingly,

however, is the fact that D89 is universally conserved among S12 homologues, and all attempts at

generating D89 variants have failed, suggesting this amino acid plays some important role (56).

In the X-ray crystal structure of the 30S ribosomal subunit from T. thermophilus, the loop on

which D89 resides interacts with 16S rRNA and is in close contact in the A-site where

aminoacylated-tRNAs bind (Figure 1-8) (5, 57). Furthermore, several amino acids in proximity

of D89 in sequence or when associated with the ribosome are among the most conserved in S12

homologues, and variants of these amino acids (K43, K88, L90, P91, G92, R94) confer resistance

to or dependence on the antibiotic streptomycin, resulting in a hyperaccurate phenotype that is

16

thought to result from these mutations disrupting stabilizing S12-16S rRNA interactions (58, 59).

Of note, P90R and P90W S12 variants lacked the methylthiol modification on D89, whereas

K88R and K88E variants did not (59). While it is clear that D89 is important, the reason for its

evolutionary conservation and/or the conditions under which D89 plays a role have yet to be

determined.

Figure 1-8. X-ray crystal structure of the Thermus thermophilus S12 protein in complex with 16S

rRNA, a 4-U mRNA codon mimic, and a 17-nucleotide (nt) anticodon stem loop (ACSL) mimic

of tRNA. The S12 protein is shown in cyan sheets and gray loops, the 16S rRNA segment in

orange, the 4-U mRNA codon mimic in pale green, and the 17-nt ACSL in pale pink. Conserved

amino acids in S12 are depicted as lines, with K42 in pink, K43 in maroon, K88 in yellow, D89

(as sticks) in bright green, L90 in black, P91 in blue, G92 in bright pink, and R94 in purple.

Nitrogen and oxygen atoms of D89 are shown in blue and red, respectively. K42, K43, and K88

interact with the phosphate backbone of the 16S rRNA. The structures were prepared using the

PyMOL Molecular Graphics program (http://www.pymol.org). The PDB accession code for the

30S subunit crystal structure is 1IBM.

17

The methylthiotransferase subfamily of radical SAM enzymes

Phylogenetic analysis of sequences similar to those of MiaB and RimO showed that all of

them belonged to one of four clades: three were comprised exclusively of bacteria; the remaining

clade contained exclusively archeal and eukaryotic members (55). The bacterial "RimO" clade

contained Ec RimO as well as RimOs from T. thermophilus, and R. palustris, while the "MiaB"

clade contained MiaBs from Ec and T. maritima. The third bacterial clade contained no

characterized members but was named YqeV, based on YqeV being the putative enzyme

responsible for the biosynthesis of ms2t6A in Bacillus subtilis. Later studies confirmed that YqeV

was the enzyme responsbile for formation of ms2t6A, and this clade was renamed MtaB after

YqeV was renamed to methylthio-threonylcarbamoyl-adenosine transferase B (60, 61). The

fourth clade—the archaeal and eukaryotic containing clade—was predicted to catalyze formation

of ms2hn

6A and was named Mj0867 based on a member found in Methanocaldococcus

jannaschii. Since the proteins comprising these clades were all believed to catalyze

methylthiolation reactions, they were deemed "methylthiotransferases" or MTTases. A sequence

analysis of the MTTases showed them all to contain three domains: an N-terminal domain named

UPF0004 (unknown protein function 0004), a central radical SAM domain, and a C-terminal

TRAM domain involved in RNA binding. The presence of the TRAM domain in RimO is

somewhat perplexing given that its substrate is a protein, and may suggest that the enzyme

recognizes S12 when assembled in the ribosome, where the TRAM domain could bind to rRNA

(55).

Of particular interest are two human MTTases, CDK5RAP1 (cyclin-dependent kinase 5

regulatory subunit-associated protein 1), a MiaB ortholog, and CDKAL1 (cyclin-dependent

kinase 5 regulatory subunit-associated protein 1-like), an MtaB ortholog. Studies of mice in

which the CDK5RAP1 gene was deleted (ΔCDK5RAP1) resulted in a mitochondrial dysfunction

18

phenotype, in addition to increased frameshifting during translation due to loss of ms2i6A,

indicating that this mammalian MTTase may play a role in the mitochondrial stress response (62,

63). Mutations in the CDKAL1 gene were identified by several genomic studies as one of the

leading risk factors in the development of type 2 diabetes across all ethnic groups (64-67). It was

determined that the loss of ms2t6A due to mutations within CDKAL1 led to misreading of lysine

codons. Proinsulin contains two lysine residues, one of which is located at the cleavage site of

proinsulin in its maturation to insulin. Pancreatic β-cells from ΔCDKAL1 mice showed lower

levels of 14

C-lysine incorporation into proinsulin compared to control wild-type cells, indicating

that CDKAL1 deficiency likely results in poor translational fidelity for lysine codons.

Misincorporation of amino acids other than lysine at the two sites in proinsulin likely disrupts its

downstream processing into insulin, leading to the development of type 2 diabetes (68).

Characterization of RimO from E. coli

The first biochemical and spectroscopic characterization of RimO from Ec was

conducted by Lee et al. (69). The rimO gene product was coexpressed with a plasmid containing

the isc operon from Azotobacter vinelandii‒which encodes for proteins involved in iron-sulfur

cluster biosynthesis (70)‒in M9 minimal media supplemented with 57

Fe, and the protein was

subsequently purified to homogenity under anaerobic conditions. UV-visible spectrophotometric

analysis of RimO as-isolated (RimOai) resulted in a spectrum exhibiting a broad feature at ~410

nm, which in conjunction with the brown color of the isolated protein, suggested the presence of

Fe-S cluster(s). Quantification of the iron and sulfide content of RimO determined that 4.4 + 0.2

equivalents of iron and 3.9 + 0.4 equivalents of sulfide were present, indicating that RimOai was

isolated with less than a full complement of two 4Fe-4S clusters. Further analysis of RimOai by

EPR spectroscopy in the presence of the low-potential reductant, sodium dithionite, resulted in

19

the formation of an axial signal accounting for 0.22 equivalents of spin with g values of 1.94 and

2.06. These parameters coupled with the fact this signal exhibited temperature-dependent

relaxation properties were consistent with those of reduced [4Fe-4S]1+

clusters with an S = 1/2

ground state (69).

To complement and corroborate the findings obtained by X-band EPR spectroscopy, the

configuration(s) and relative amounts of Fe associated in Fe-S clusters were determined by

Mössbauer spectroscopy. The Mössbauer spectrum of RimOai recorded at 4.2 K and 53 mT

exhibited a quadrupole doublet with an isomer shift (δ) parameter of 0.43 mm/s and a quadrupole

splitting (ΔEq) parameter of 1.07 mm/s, which are typical of [4Fe-4S]2+

clusters. A less intense

line was observed at ~0.5 mm/s and was assigned to the high-energy line observed in spectra of

[2Fe-2S]2+

clusters. A second spectrum recorded in a 6 T external magnetic field and simulated

with the δ and ΔEq parameters obtained from the 53 mT spectrum determined the electronic

ground state to be diamagnetic (S = 0), further corrorborating the presence of [4Fe-4S]2+

and

[2Fe-2S]2+

, given that these cluster species have diamagnetic ground states. In conjunction with

the quantitative iron and sulfide analyses of RimOai, the Mössbauer spectroscopic analysis

indicated the enzyme to harbor 4.0 Fe ions as a [4Fe-4S] cluster. Perturbations in the Mössbauer

spectrum of RimOai in the presence of SAM suggested that the majority of the clusters present are

ligated by the RS Cx3Cx2C motif (69).

Reconstitution of RimOai with 57

Fe and sodium sulfide and subsequent desalting resulted

in an intensely brown protein, in which quantitative iron and sulfide analyses determined that

11.4 + 0.2 equivalents of iron and 11.1 + 0.7 equivalents of sulfide were present. The Mossbauer

spectrum of RimOrcn obtained at 4.2 K and 53 mT exhibited a quadrupole doublet with

parameters similar to those obtained for RimOai, indicating the presence of [4Fe-4S]2+

clusters.

Quantitative iron analysis coupled with the Mössbauer spectroscopic analysis determined 7.1 +

0.7 iron ions to be present in RimOrcn in the form of [4Fe-4S]2+

clusters, which corresponded to

20

1.8 + 0.2 [4Fe-4S] clusters per RimOrcn polypeptide, thereby supporting the hypothesis that RimO

ligates two [4Fe-4S] clusters . The Mössbauer spectrum of RimOrcn in the presence of a 13 amino

acid peptide (P1) corresponding to residues 83-95 of the S12 protein exhibited no apparent

perturbations (69).

Formation of S-adenosylhomocysteine (SAH), a SAM-derived product resulting from

methyl transfer, and 5'-deoxyadenosine (5'-dAH), by RimOrcn in the presence of SAM, the

chemical reducant sodium dithionite, and P1 substrate was used to assess turnover. A peptide

substrate was used in place of the S12 protein, which was found to be insoluble. Over the course

of 3 hours, 0.33 equivalents of 5'-dAH and 0.04 equivalents of SAH were formed per RimOrcn.

Addition of a 50 nucleotide RNA mimicking the 530 stem loop of 16S rRNA found near S12 in

the X-ray crystal structure of the ribosome did not enhance formation of either 5'-dAH or SAH.

Similarly, the addition of a peptide in which the target aspartic acid was substituted with alanine

resulted in minute amounts of 5'-dAH formation. In all instances, even in the absence of

substrate, the same amount of SAH was formed in an enzyme-dependent manner, which

suggested that SAH was not on the catalytic pathway, that the methyl group of the methylthio

moiety is not SAM-derived, or that the protein catalyzes methyl transfer from SAM to an

unknown acceptor site on the polypeptide (69).

Electrospray ionization tandem mass spectrometry (ESI+-MS/MS) of a turnover reaction

mixture containing SAM, P1, and dithionite showed the presence of a peptide with Δm/z +47

greater than the expected mass of P1, which, within instrumental error, closely matched the Δm/z

+46 expected for methylthiolation. Although only ~4% of the peptide was methylthiolated, it was

definitively shown by analysis of the resultant b and y ions of the methylthiolated peptide that the

modification did indeed take place on the target aspartic acid (69). The authors proposed a

working mechanistic hypothesis, based on other RS sulfur-insertion enzymes, in which the

methylthio- group is inserted in a stepwise manner in which SAM is first used for radical

21

chemistry and sulfur-insertion to produce a thiolated intermediate species, followed by a second

molecule of SAM donating a methyl group to the thiolated intermediate to form the

methylthiolated product and SAH (Figure 1-9). This model proposes that the auxiliary [4Fe-4S]

cluster is the sacrificial source of sulfide for the reaction, thereby limiting RimO to one turnover

in vitro.

Figure 1-9. The working mechanistic model for RimO proposed by Lee et al. SAM is reductively

cleaved to generate L-methionine and a 5'-dA•, which abstracts a hydrogen atom from C3 of D89

to form 5'-dAH and a radical at C3. The substrate radical combines with a bridging µ3-sulfido ion

of the auxiliary 4Fe-4S cluster in two possible scenarios. In the first, the cluster is degraded upon

sulfur insertion to form a thiolated intermediate. An SN2 attack by the C3 thiol on the

electrophilic methyl group of SAM results in the formation of SAH and capping of the inserted

sulfur atom to form methylthiolated D89. The second scenario depicts the thiolated intermediate

remaining cross-linked to Fe-S cluster until the sulfur atom attacks the methyl group of SAM,

resulting in degradation of the [4Fe-4S] cluster and formation of methylthiolated D89. The

degradation of the cluster in both pathways limits RimO to one turnover in vitro, whereas in vivo,

the cluster is likely reassembled via Fe-S cluster assembly proteins in a process that is not well

understood.

22

Characterization of RimO from T. maritima

Shortly after the characterization of Ec RimO, a similar study was conducted by Arragain

et al. in which RimO from T. maritima (Tm) was biochemically and spectroscopically

characterized. As in Ec RimO, Tm RimOai was found to harbor substoichiometric amounts of iron

and sulfide, but after reconstitution contained nearly a full complement of iron and sulfide in the

form of two [4Fe-4S] clusters as determined by quantitative iron and sulfide analyses in

conjunction with Mössbauer spectroscopic analyses. The quaternary structure of Tm RimO was

shown to be primarily monomeric by analytical gel filtration chromatography. Diffraction quality

crystals of apo-Tm RimO were obtained in the presence of small amounts of the protease

subtilisin and resulted in the determination of an incomplete the X-ray crystal structure depicting

the central RS and C-terminal TRAM domains to 2.0 Å resolution (Figure 1-10) (71).

Unfortunately, the loop containing the Cx3Cx2C motif was not present, likely due to this loop

being disordered in the absence of the RS cluster and its cleavage by subtilisin. Nonetheless, this

first RimO structure showed the central SAM domain to be composed of an incomplete α6β7 TIM

barrel, and the most closely related domains of known structure were identified in RS enzymes

oxygen-independent coproporphyrinongen III oxidase (HemN) and molybdenum cofactor

biosynthesis A (MoaA). Similarly, the TRAM domain of RimO was most closely related to that

found in the 23S rRNA methyltransferase RumA. Although TRAM domains bind negatively

charged RNA molecules (72, 73), the TRAM domain in RimO is highly acidic and is found at the

distal edge of the concave surface of the RS domain, which supported the hypothesis that this

domain in RimO and MiaB facilitates the binding of their respective macromolecular substrates

(71)

23

Figure 1-10. X-ray crystal structure of apo-Tm RimO. The α-helices and β-sheets of the α6β7 TIM

barrel comprising the radical SAM domain are shown in teal and bright green, respectively. The 5

β-sheets and the loops comprising the TRAM domain are shown in dark red. The N-terminus

portion of the protein is missing in this structure, and the first N-terminal residue, 135, is

indicated by the black arrow. The C-terminus is also indicated by a black arrow. The structure

was prepared using the PyMOL Molecular Graphics program (http://www.pymol.org). The PDB

accession code for apo-Tm RimO is 2QGQ.

Methylthiolation activity of Tm RimO was assessed in assays containing SAM,

dithionite, and a 20 amino acid peptide substrate corresponding to residues flanking D89 of the

S12 protein. Incorporation of -SCH3 on the target aspartate residue, in addition to appendage of a

second -SCH3 group at an unidentified position of the peptide was observed by ESI+ MS/MS.

Interestingly, ~ 0.3 nmol of holo-RimO inserted 0.8 nmol of sulfur into the substrate, resulting in

24

the formation of 0.6 nmol of methylthiolated and bismethylthiolated products. This result

indicated that additional sulfur atoms tightly associated with the protein after its reconstitution

and subsequent purification by anion-exchange chromatography were used in the

methylthiolation reaction. Quantification of 5'-dAH and SAH formed in the reaction determined

the former was generated in 4-fold excess and the latter in 1.5-fold excess of methylthiolated

product. Formation of SAH was also observed in the absence of the peptide substrate, whereas

formation of 5'-dAH was detected only in its presence (71).

A follow-up study by Forouhar et al. on RimO and MiaB from Tm shed light on the

methylthiolation reactions they catalyze. Reconstitution of RimO and quantification of the

amount of iron and sulfide retained resulted in the incorporation of excess sulfide (11.6 + 0.8

sulfide ions) versus Fe (8.5 + 0.2 iron ions) per protein. It was determined that the additional

sulfide (2 + 1 per protein) present was in the S(0) oxidation state, indicating sulfur was bound to

the protein. Similar results were obtained for MiaB. When incubated under turnover conditions,

RimOrcn catalyzed formation of ~ 3 enzyme equivalents of MS-D89, indicating that the retained

sulfur atoms afforded multiple turnovers, and it was observed that there was a strong correlation

between the amounts of additional sulfur atoms retained and MS-D89 formed. The amount of 5'-

dAH and SAH formed in these reactions was not stated for RimO, but MiaB produced 1.2

equivalents of 5'-dAH and 2.0-3.4 equivalents of SAH per ms2i6A formed. The excess SAH

produced appeared to be an uncoupled side reaction, as its formation was observed in the

presence of dithionite and absence of substrate, and analysis of the protein by ESI-MS revealed

that the protein itself was not the methyl acceptor (37).

When exogenous sulfide was added to reaction mixtures of RimO or MiaB, both enzymes

catalyzed additional turnovers beyond that observed in its absence‒a total of 5 turnovers by RimO

and 12-21 turnovers by MiaB‒indicating these enzymes to be catalytic in the presence of sulfide.

Similar results were obtained when RimO or MiaB were incubated with methylsulfide (CH3S-) or

25

methylselenide (CH3Se-): the former catalyzed 5 turnovers with both cosubstrates and the latter

catalyzed 6 turnovers with methylsulfide and 10 turnovers with methylselenide. In the

methylselenide-containing reactions, the MSe-D89 and mse2i6a products were almost exclusively

formed, revealing the retained sulfur atoms were not readily used for product formation by either

enzyme under these conditions. In reactions containing methylselenide and [14

C-methyl]-SAM,

radioactivity was observed only in the minor HPLC peak corresponding to ms2i6A, thereby

demonstrating that CH3S- and CH3Se

- are functional cosubstrates that are incorporated into the

RimO and MiaB substrates.

To determine whether the auxiliary cluster acted as a binding site for sulfide,

methylsulfide, and methylselenide, variant MiaB and RimO proteins in which their RS clusters

were absent were studied by HYSCORE spectroscopy in the presence of CH377

Se-. For both

enzymes, addition of CH377

Se- resulted in the formation of a new feature in their spectra, which

was centered on the nuclear frequency of 77

Se and which exhibited the shape characteristic of

weak 77

Se hyperfine coupling (3.8 + 0.5 MHz) to the reduced Fe-S cluster. In silico modeling and

DFT calculations were consistent with CH377

Se- binding to the unique Fe site of the auxiliary

cluster rather than replacing one of the cluster's constituent sulfide ions (37). The stability of the

auxiliary cluster in MiaB in the presence of 2000-fold excess CH3S- both before and after

reduction with dithionite was assessed by quantifying the amount of sulfur bound to the enzyme.

Under these conditions, no sulfur was liberated from the protein, thereby indicating the cluster to

be stable under strongly reducing conditions (37).

The same study also reported an X-ray crystal structure of holo-Tm RimO at 3.3 Å

resolution, the first such structure of an MTTase with both 4Fe-4S clusters present. The reported

distance between the two 4Fe-4S clusters was ~ 8 Å, which is substantially less than those

reported for other RS enzymes containing two clusters (BioB, 12 Å; MoaA, 16 Å; LipA 13 Å;

BtrN, 16 Å; anSME, 17 Å; HydG, 24 Å) (30, 31, 74-78). Intriguingly, a chain of electron density

26

linking the two unique iron sites of the RS and auxiliary clusters was observed. The electron

density was best refined when modeled as a pentasulfide chain, which, given that excess sulfide

was present in the condition under which the crystals were grown, made this model reasonable

(Figure 1-11). Modeling of the SAM molecule bound to the RS cluster in the structure of MoaA

into this Tm RimO structure resulted in no steric clash with protein side chains, but it did clash

with the final three sulfur atoms of the pentasulfide chain, indicating that the active site can only

accomodate SAM and up to two sulfur atoms bound to the auxiliary cluster. An electrostatic map

of the protein structure revealed a negatively charged funnel ~ 40 Å deep in which the S12

protein is thought to bind (Figure 1-12). Docking of S12 in this funnel resulted in little steric

clash and essentially sealed the active site, implying that SAM and any cosubstrates bind before

the substrate (37).

27

Figure 1-11. X-ray crystal structure of holo-Tm RimO. The UPF0004, radical SAM, and TRAM

domains are shown in indigo, gray, and light teal, respectively. The iron and sulfide ions of the

two 4Fe-4S clusters are depicted as orange and yellow spheres, respectively. The pentasulfide

chain bridging the two clusters is shown as contiguous yellow sticks. Cysteine residues

coordinating the auxiliary and RS clusters are shown as cyan and red sticks, respectively. The N-

and C-termini are indicated by black arrows. The structure was prepared using the PyMOL

Molecular Graphics program (http://www.pymol.org). The PDB accession code for holo-Tm

RimO is 4JC0.

28

Figure 1-12. Electrostatic protein contact potential map determined from the X-ray crystal

structure of holo-Tm RimO. Red and blue indicates areas of negative and positive charge,

respectively. The funnel leading to the active site is lined with negatively charged residues, which

are thought to aid in binding of the positively charged S12 protein. The patch of blue above the

funnel corresponds to the TRAM domain, which binds the phosphate backbone of RNA in other

proteins and is appropriately positively charged. The structure was prepared using the PyMOL

Molecular Graphics program (http://www.pymol.org). The PDB accession code for holo-Tm

RimO is 4JC0.

A methylated cluster intermediate in RimO

An in-depth biochemical study of RimO from Tm was conducted by Landgraf et al (79)

and detailed in Chapter 2. Incubation of the enzyme with SAM, but in the absence of a chemical

reductant and a 13 amino acid peptide substrate (P1), resulted in the formation of ~ 1 enzyme

29

equivalent of SAH formed, which stood in contrast to the proposed model wherein radical

chemistry preceeded methyl transfer. In radiotracer studies, in which RimO and MiaB were

incubated with [14

C-methyl]-SAM, radioactivity was found to be associated with the protein

fraction after the reaction mixture was applied to a size-exclusion column. The protein fraction

that eluted from the size-exclusion column was acid-denatured, and the resulting supernatant

containing any SAM-derived molecules previously bound to the protein was separated by HPLC.

Subsequent scintillation counting of the eluted molecules showed no radioactivity to be present in

any SAM-derived compounds, which suggested that the radioactivity was liberated from an acid-

labile acceptor of the enzyme upon acid denaturation. GC-MS headspace analysis of sealed vials

containing RimO and SAM that were incubated and subsequently acid denatured detected

methanethiol (m/z +47), the product of methyl transfer to a sulfide ion, in a 1:1 stoichiometric

ratio with the amount of SAH formed during the incubation. Furthermore, the kinetics of SAH

and methanethiol formation were essentially identical, thereby demonstrating kinetic competence

of the methylated cluster intermediate. The chemical competence of this intermediate was

demonstrated by differential labeling studies in which the enzyme was incubated with [methyl-

d3]-SAM, subsequently applied to a size-exclusion column to remove unreacted [methyl-d3]-

SAM, then incubated under turnover conditions with natural abundance SAM. This treatment

resulted in a burst of d3-methylthiolated product formed, followed by slower formation of

unlabeled product. Essentially identical results were obtained from the same experiments

conducted with MiaB. Finally, RimO and MiaB were shown to catalyze multiple turnovers in the

presence of excess sulfide and methylsulfide, as was previously reported (37), but to lesser

extents (79).

30

The stereochemical course of the RimO reaction

In a recent X-ray crystal structure of the ribosome from Ec at 2.4 Å resolution, electron

density corresponding to the methythio- group appended to D89 of S12 was observed. The

introduction of the -SCH3 group to C3 of D89 by RimO makes it a chiral center. The resolution of

the crystal structure was sufficiently high that the absolute configuration at C3 was assigned as

3R (80), leaving only the stereoselectivity of hydrogen atom abstraction from C3 by RimO

unanswered. Landgraf and Booker chemoenzymatically synthesized aspartic acid labeled

stereospecifically with deuterium at either the pro-R or pro-S position of C3, and the deuterated

aspartic acids were incorporated into S12 peptide substrate mimics. Incubation of both of the

isotopologue-containing peptides with RimO under turnover conditions afforded deuterium

enrichment only in the presence of the pro-S deuterated aspartic acid containing peptide,

indicating that the RimO active site finely controls the 5'-dA• such that the pro-S hydrogen atom

is stereoselectively abstracted. This finding completed the overall stereochemical course of the

methylthiolation of C3 of D89, wherein abstraction of the pro-S hydrogen atom and insertion of

the -SCH3 group occurs with an overall inversion of configuration (79).

A proposed mechanism for the methylthiolation of D89 of the ribosomal S12 protein

The culmination of decades of work described herein allows for a mechanism to be

proposed for the insertion of a methylthio- group to form 3-methylthioaspartyl D89 on the

ribosomal S12 protein. (Figure 1-13). In the first step, SAM binds to the RS [4Fe-4S]2+

cluster,

anchoring and positioning it for nucleophilic attack by a sulfide ion bound to the unique iron site

(Fea) of the N-terminal auxiliary cluster, resulting in the formation of a methylated auxiliary

cluster intermediate. Release of SAH from the active site opens up the Fea of the RS cluster, to

31

which a second molecule of SAM binds via its α-amino and α-carboxylate moieties. Reduction of

the RS [4Fe-4S]2+

by one electron to the [4Fe-4S]1+

state affords electron transfer into the anti-

bonding orbital of the sulfur atom of SAM to effect its homolytic cleavage to L-methionine and a

5'-dA•. The highly potent 5'-dA• abstracts the pro-S hydrogen atom from C3 of D89 to generate

5'-dAH and a C3-centered radical. The substrate radical combines with the pendant -SCH3 group

on the Fea of the auxiliary cluster with concomitant transfer of one electron from the scission of

the Fea-S bond of the auxiliary cluster to form C3-methylthioaspartyl 89 S12 and an auxiliary

[4Fe-4S]1+

cluster, which may transfer an electron to the RS cluster for another reductive

cleavage of SAM upon substrate binding or to an external electron acceptor. Although the fate of

the auxiliary cluster has yet to be definitively determined, if it is indeed used as a sacrificial

source of sulfur, mechanisms in vivo likely reassemble the cluster to afford additional turnovers;

however, the nature of such processes remains unknown.

32

Figure 1-13. The proposed mechanism for the methylthiolation of D89 of S12. In step 1, SAM

bound to the RS cluster is positioned for nucleophilic attack on its methyl group (blue) by a

pendant sulfide ion (green) on the unique iron of the auxiliary cluster, resulting in the formation

of a methylated cluster intermediate and SAH, wherein the latter vacates the active site. In step 2,

a second molecule of SAM binds to the RS cluster, the RS cluster is reduced by one electron, and

S12 binds to trigger formation of the 5'-dA• (red). In step 3, the 5'-dA• abstracts the pro-S

hydrogen atom from C3 of D89 of S12 (orange). In step 4, 5'-dAH is released from the active site

and the C3-centered radical (red) combines with an electron from the bond between the unique

iron of the auxiliary cluster and the sulfur atom of the pendant SCH3 group to form (3R)-3-

methylthioaspartyl 89 S12 (3-MS-D89 S12) shown in step 5. In step 6, 3-MS-D89 S12 vacates

the active site and one electron from the auxiliary cluster is either transferred to the RS cluster or

to an external acceptor. Sulfide and SAM bind to the vacant active site for another round of

catalysis.

33

Conclusions

The reaction catalyzed by RimO is significant for a number of reasons. The enzyme

performs the feat of cleaving a C-H bond from relatively inert carbon center, C3 of D89 S12.

RimO harnesses the oxidative power of a 5'-dA•, generated by the reductive cleavage of SAM

bound to a [4Fe-4S]1+

cluster to effect such a homolytic cleavage. This enzyme performs another

feat by not only utilizing SAM as the precursor to a 5'-dA• but as a methyl donor to synthesize a

methylthio- group, making it one of the few enzymes able to exploit two different modes of

reactivity from the same coenzyme. The unique iron site of the auxiliary cluster plays key roles in

the chemistry that takes place in the RimO reaction. It serves as a binding site for sulfide and is

capable of stabilizing an unpaired electron, which is required in the homolytic cleavage of the Fe-

S bond during methylthio- insertion. While there are some questions concerning the RimO

reaction that remain to be answered‒is the auxiliary cluster a binding site or a sacrificial source of

sulfide? If the cluster is sacrificed, what repair processes are at play? What minimum players are

required in the repair process to make MTTases, and other sulfur-insertion RS enzymes,

catalytic? Do RimO and MiaB utilize the same mechanisms to cleave an sp3-hybridized and an

sp2-hybridized C-H bond, respectively?‒much work has elucidated key details of this interesting

and challenging enzymatic transformation.

34

References

1. Garrett R. Nature. 1999: 400, 811-812

2. Frank J. Chemistry & Biology. 2000: 7, R133-R141

3. Kaczanowska M, Rydén-Aulin M. Microbiology and Molecular Biology Reviews. 2007:

71, 477-494

4. Decatur WA, Fournier MJ. Trends in Biochemical Sciences. 2002: 27, 344-351

5. Carter AP, Clemons WM, Brodersen DE, et al. Nature. 2000: 407, 340-348

6. Green R, and, Noller HF. Annual Review of Biochemistry. 1997: 66, 679-716

7. Nissen P, Hansen J, Ban N, et al. Science. 2000: 289, 920-930

8. Arnold RJ, Reilly JP. Analytical Biochemistry. 1999: 269, 105-112

9. Ben-Bassat A, Bauer K, Chang SY, et al. Journal of Bacteriology. 1987: 169, 751-757

10. Chang FN, Budzilowicz C. Journal of Bacteriology. 1977: 131, 105-110

11. Kowalak JA, Walsh KA. Protein Science. 1996: 5, 1625-1632

12. Funatsu G, Yaguchi M, Wittmann-Liebold B. FEBS Letters. 1977: 73, 12-17

13. Post LE, Nomura M. Journal of Biological Chemistry. 1980: 255, 4660-4666

14. Christy MR, Barkley RM, Koch TH, et al. Journal of the American Chemical Society.

1981: 103, 3935-3937

15. McMullen BA, Fujikawa K, Kisiel W, et al. Biochemistry. 1983: 22, 2875-2884

16. Díaz I, Ehrenberg M. Journal of Molecular Biology. 1991: 222, 1161-1171

17. Buck M, McCloskey JA, Basile B, Ames BN. Nucleic Acids Research. 1982: 10, 5649-

5662

18. Qian Q, Curran JF, Björk GR. Journal of Bacteriology. 1998: 180, 1808-1813

19. Reddy DM, Crain PF, Edmonds CG, et al. Nucleic Acids Research. 1992: 20, 5607-5615

20. Kirchner S, Ignatova Z. Nat Rev Genet. 2015: 16, 98-112

21. Sofia HJ, Chen G, Hetzler BG, et al. Nucleic Acids Research. 2001: 29, 1097-1106

22. Knappe J, Neugebauer FA, Blaschkowski HP, Gänzler M. Proceedings of the National

Academy of Sciences of the United States of America. 1984: 81, 1332-1335

23. Wu W, Booker S, Lieder KW, et al. Biochemistry. 2000: 39, 9561-9570

24. Moss M, Frey PA. Journal of Biological Chemistry. 1987: 262, 14859-14862

25. Frey PA, Abeles RH. Journal of Biological Chemistry. 1966: 241, 2732-2733

26. Abeles RH, Dolphin D. Accounts of Chemical Research. 1976: 9, 114-120

27. Akiva E, Brown S, Almonacid DE, et al. Nucleic Acids Res. 2014: 42, D521-D530

28. Walsby CJ, Ortillo D, Broderick WE, et al. Journal of the American Chemical Society.

2002: 124, 11270-11271

29. Layer G, Moser J, Heinz DW, et al. The EMBO Journal. 2003: 22, 6214-6224

30. Berkovitch F, Nicolet Y, Wan JT, et al. Science. 2004: 303, 76-79

31. Hänzelmann P, Schindelin H. Proceedings of the National Academy of Sciences of the

United States of America. 2004: 101, 12870-12875

32. Lepore BW, Ruzicka FJ, Frey PA, Ringe D. Proceedings of the National Academy of

Sciences of the United States of America. 2005: 102, 13819-13824

33. Goto-Ito S, Ishii R, Ito T, et al. Acta Crystallographica Section D. 2007: 63, 1059-1068

34. Vey JL, Yang J, Li M, et al. Proceedings of the National Academy of Sciences. 2008:

105, 16137-16141

35

35. Boal AK, Grove TL, McLaughlin MI, et al. Science (Washington, DC, U. S.). 2011: 332,

1089-1092

36. Vey JL, Drennan CL. Chemical Reviews. 2011: 111, 2487-2506

37. Forouhar F, Arragain S, Atta M, et al. Nature Chemical Biology. 2013: 9, 333-338

38. Mato J, Alvarez L, Ortiz P, Pajares MA. Pharmacology & Therapeutics. 1997: 73, 265-

280

39. Lieber CS, Packer L. The American Journal of Clinical Nutrition. 2002: 76, 1148S-

1150S

40. Roje S. Phytochemistry. 2006: 67, 1686-1698

41. Esberg B, Leung H-CE, Tsui H-CT, et al. Journal of Bacteriology. 1999: 181, 7256-7265

42. Sun X, Eliasson R, Pontis E, et al. Journal of Biological Chemistry. 1995: 270, 2443-

2446

43. Buck M, Griffiths E. Nucleic Acids Research. 1982: 10, 2609-2624

44. Wettstein FO, Stent GS. Journal of Molecular Biology. 1968: 38, 25-40

45. Rosenberg AH, Gefter ML. Journal of Molecular Biology. 1969: 46, 581-584

46. Pierrel F, Björk GR, Fontecave M, Atta M. Journal of Biological Chemistry. 2002: 277,

13367-13370

47. Pierrel F, Hernandez HL, Johnson MK, et al. Journal of Biological Chemistry. 2003: 278,

29515-29524

48. Pierrel F, Douki T, Fontecave M, Atta M. J. Biol. Chem. 2004: 279, 47555-47563

49. Hernandez HL, Pierrel F, Elleingand E, et al. Biochemistry. 2007: 46, 5140-5147

50. Tse Sum Bui B, Florentin D, Fournier F, et al. FEBS Letters. 1998: 440, 226-230

51. Gibson KJ, Pelletier DA, Turner IM. Biochemical and Biophysical Research

Communications. 1999: 254, 632-635

52. Cicchillo RM, Booker SJ. J. Am. Chem. Soc. 2005: 127, 2860-2861

53. Fugate CJ, Stich TA, Kim EG, et al. J. Am. Chem. Soc. 2012: 134, 9042-9045

54. Lanz ND, Pandelia M-E, Kakar ES, et al. Biochemistry. 2014: 53, 4557-4572

55. Anton BP, Saleh L, Benner JS, et al. Proceedings of the National Academy of Sciences

USA. 2008: 105, 1826-1831

56. Carr JF, Hamburg D-M, Gregory ST, et al. Journal of Bacteriology. 2006: 188, 2020-

2023

57. Ogle JM, Brodersen DE, Clemons WM, et al. Science. 2001: 292, 897-902

58. Finken M, Kirschner P, Meier A, et al. Molecular Microbiology. 1993: 9, 1239-1246

59. Carr JF, Hamburg D-M, Gregory ST, et al. J. Bacteriol. 2006: 188, 2020-2023

60. Anton BP, Russell SP, Vertrees J, et al. Nucleic Acids Research. 2010: 38, 6195-6205

61. Arragain S, Handelman SK, Forouhar F, et al. Journal of Biological Chemistry. 2010:

285, 28425-28433

62. Wei F-Y, Zhou B, Suzuki T, et al. Cell Metabolism. 2015: 21, 428-442

63. Landgraf BJ, McCarthy EL, Booker SJ. Annual Review of Biochemistry. 2016: 85, 485-

514

64. Saxena R, Voight BF, Lyssenko V, et al. Science. 2007: 316, 1331-1336

65. Scott LJ, Mohlke KL, Bonnycastle LL, et al. Science. 2007: 316, 1341-1345

66. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. Nat Genet. 2007: 39, 770-775

67. Zeggini E, Weedon MN, Lindgren CM, et al. Science. 2007: 316, 1336-1341

68. Wei F-Y, Suzuki T, Watanabe S, et al. The Journal of Clinical Investigation. 2011: 121,

3598-3608

69. Lee K-H, Saleh L, Anton BP, et al. Biochemistry. 2009: 48, 10162-10174

70. Johnson DC, Unciuleac M-C, Dean DR. Journal of Bacteriology. 2006: 188, 7551-7561

36

71. Arragain S, Garcia-Serres R, Blondin G, et al. Journal of Biological Chemistry. 2010:

285, 5792-5801

72. Lee TT, Agarwalla S, Stroud RM. Structure. 2004: 12, 397-407

73. Anantharaman V, Koonin EV, Aravind L. FEMS Microbiology Letters. 2001: 197, 215-

221

74. Harmer Jenny E, Hiscox Martyn J, Dinis Pedro C, et al. Biochemical Journal. 2014: 464,

123-133

75. Goldman PJ, Grove TL, Booker SJ, Drennan CL. Proceedings of the National Academy

of Sciences. 2013: 110, 15949-15954

76. Goldman PJ, Grove TL, Sites LA, et al. Proceedings of the National Academy of

Sciences. 2013: 110, 8519-8524

77. Dinis P, Suess DLM, Fox SJ, et al. Proceedings of the National Academy of Sciences.

2015: 112, 1362-1367

78. Lanz ND, Booker SJ. Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics.

2012: 1824, 1196-1212

79. Landgraf BJ, Arcinas AJ, Lee K-H, Booker SJ. J. Am. Chem. Soc. 2013: 135, 15404-

15416

80. Noeske J, Wasserman MR, Terry DS, et al. Nature Structural and Molecular Biology.

2015: 22, 336-341

Chapter 2

Identification of an Intermediate Methyl Carrier in the Radical S-

adenosylmethionine Methylthiotransferase RimO

RimO is a radical S-adenosylmethionine (SAM) enzyme that catalyzes the attachment of

methylthio (–SCH3) group at C3 of aspartate 89 of protein S12, a component of the 30S subunit

of the bacterial ribosome. This enzyme is a prototypical member of a subclass of radical SAM

(RS) enzymes called methylthiotransferases (MTTases). Like all RS enzymes, RimO contains a

[4Fe–4S] cluster to which SAM associates (RS cluster), and which participates in the reductive

cleavage of SAM to methionine and a 5’-deoxyadenosyl 5’-radical. Unlike most RS enzymes,

RimO also contains an additional [4Fe–4S] cluster (auxiliary cluster) that is believed to be the

source of the sulfur atom of the inserted –SCH3 group. It had been assumed that the sequence of

MTTase reactions involves initial sulfur insertion into the organic substrate followed by capping

of the inserted sulfur atom with a SAM-derived methyl group. In this work, however, we show

that RimO from Thermotoga maritima (Tm) catalyzes methyl transfer from SAM to an acid/base

labile acceptor on the protein in the absence of its macromolecular substrate as well as the

requisite reductant that triggers radical chemistry. Consistent with the assignment of the acceptor

as an iron–sulfur (Fe/S) cluster (presumably the auxiliary [4Fe–4S] cluster), denaturation of the

SAM-treated protein with acid results in production of methanethiol. When RimO is first

incubated with SAM in the absence of substrate and reductant, and then incubated with excess S-

adenosyl-L-[methyl-d3]methionine ([methyl-d3]-SAM) in the presence of substrate and reductant,

production of the unlabeled product precedes production of the deuterated product, showing that

the methylated intermediate is chemically and kinetically competent. Last, introduction of

methanethiol in Tm RimO reactions conducted with [methyl-d3]-SAM affords both unlabeled and

38

deuterated products in equal ratios, showing that methanethiol itself can serve as a

methylthiolating agent.

Introduction

The radical S-adenosylmethionine (SAM) methylthiotransferases (MTTases) catalyze the

attachment of methylthio (–SCH3) groups at specific locations on tRNAs or ribosomal proteins,

resulting in thioether bonds (1). Three major classes of MTTases are currently recognized, which

are represented by the enzymes MiaB, MtaB, and RimO (1-3). MiaB catalyzes the final step in

the biosynthesis of the hypermodified tRNA nucleoside 2-methylthio-N6-(isopentenyl)adenosine

(ms2i6A), which is the methylthiolation of C2 of N

6-(isopentenyl)adenosine (i

6A) — found at

position 37 of certain tRNAs — while MtaB catalyzes the methylthiolation of the same carbon

center of N6-(threonylcarbamoyl)adenosine (t

6A) to afford 2-methylthio-N

6-

(threonylcarbamoyl)adenosine (ms2t6A) (Figure 2-1, panels A and B, respectively). By contrast,

RimO acts on a protein substrate, catalyzing the methylthiolation of the -carbon of aspartate 89

(Ec numbering) of ribosomal protein S12 (Figure 2-1, panel C). These proteins, along with

biotin synthase (BS) and lipoyl synthase (LS), constitute a special subfamily of radical SAM (RS)

enzymes that catalyze sulfur insertion (1, 4-6). All RS enzymes that catalyze sulfur insertion

contain two distinct iron–sulfur (Fe/S) clusters: a [4Fe–4S] cluster ligated by cysteines in a

Cx3Cx2C motif (RS cluster), and either a [2Fe–2S] cluster (BS) (7-9) or an additional [4Fe–4S]

cluster (LS, and MTTases) (auxiliary cluster) (3, 10-12). The RS cluster binds in contact with

SAM and, in its reduced state ([4Fe–4S]+), participates in the reductive fragmentation of SAM to

a 5’-deoxyadenosyl 5’-radical (5’-dA•), a common intermediate among RS reactions (6, 13, 14).

The mechanistic details associated with sulfur insertion are not completely understood; however,

it is believed that substrate radicals, generated by abstraction of hydrogen atoms (H•) from target

39

carbon centers by the 5’-dA•, attack the bridging µ-sulfido ions of the auxiliary clusters (4,15).

Because their auxiliary clusters are thought to be sacrificed during catalysis, RS enzymes that

catalyze sulfur insertion typically catalyze no more than one turnover in vitro, although instances

of higher product ratios have been reported (3, 16).

All known MTTases that act on tRNA modify adenosine 37 (A37), which resides

immediately adjacent to the third nucleotide (position 36) of the anticodon. Before

methylthiolation takes place, A37 must be modified at N6 with either an isopentenyl (or 4-

hydroxyisopentenyl) group (MiaB family) or a threonylcarbamoyl group (MtaB family). MiaA, a

dimethylallyl pyrophosphate:tRNA dimethylallyltransferase, catalyzes the first committed step in

formation of ms2i6A, which is the transfer of an isopentenyl (dimethylallyl) group from

dimethylallyl pyrophosphate (DMAPP) to N6 of A37 (17-19). By contrast, four proteins (YgjD,

YrdC, YjeE, and YeaZ) are required to generate the threonylcarbamoyl group at A37 of tRNAs

that are modified by MtaB and its related proteins (20, 21). Hypermodifications of A37 are

typically found on tRNAs that contain adenosine or uracil at position 36. Although nonessential

(22-24), they are believed to induce slight structural perturbations in the tRNA that permit

increased exposure of the Watson–Crick faces of the anticodon to the RNA codon. This improved

base pairing increases recognition of cognate tRNAs in the A-site over near-cognate tRNAs,

thereby reducing ribosomal A-site pausing (25). Moreover, improving the relatively weak

adenosine-uridine pairing at the first base of the codon prevents ribosomal P-site slippage. The

improvements in A-site and P-site recognition result in enhanced reading frame maintenance and

therefore translational fidelity (22-25). Recently, a member of the eukaryotic MtaB class of

MTTases was shown to be one of the most reproducible genetic risk factors in the etiology of

type 2 diabetes across multiple ethnic groups (26-30).

RimO catalyzes methylthiolation of the -carbon of aspartate 89 of protein S12 of the

small subunit of the bacterial ribosome in Escherichia coli (Ec) and a number of other bacteria,

40

including Thermotoga maritima (Tm). The purpose of this modification is not yet known; it is

neither universal nor essential for ribosome function. However, the inability to generate variants

of D89 suggests that this residue — which projects toward the acceptor site of the ribosome — is

essential, and may play a role in some aspect of ribosome function (31, 32).

The MTTases represent a growing subclass of RS enzymes that use SAM both as a

radical generator and a methyl donor. The best characterized members of this class are the RS

methyltransferases/methylsynthases, which are represented by RlmN and Cfr (33). These two

proteins catalyze the synthesis of a methyl group onto C2 and C8, respectively, of adenosine 2503

in 23S rRNA (34, 35), employing a ping-pong-like mechanism of catalysis (34). In the first half-

reaction, SAM binds to the unique Fe ion of the sole [4Fe–4S] cluster in each protein and donates

a methyl group to a conserved Cys residue, releasing S-adenosylhomocysteine (SAH) as the

byproduct of the reaction (34, 36, 37). In the second half-reaction a second molecule of SAM

binds to the same site, but is reductively cleaved to a 5’-dA•, which initiates turnover by

abstracting a hydrogen atom (H•) from the methylCys residue. After radical addition to C2 or C8

of the adenine ring and loss of an electron to an undetermined acceptor, a methylene-bridged

protein-substrate crosslink is resolved by disulfide-bond formation with concomitant release of an

enamine, which tautomerizes to the methyladenosine product upon acquiring a proton from a

general acid in the active site (34, 38).

In this work, we show that Tm RimO and Tm MiaB also show characteristics of a ping-

pong-like reaction. Each protein catalyzes formation of ~1 equiv of SAH in the absence of

substrate and reductant, and an equal amount of methanethiol upon acid-denaturation of the

protein. Moreover, introduction of methanethiol in assays conducted with S-adenosyl-L-[methyl-

d3]methionine ([methyl-d3]-SAM) results in formation of both unlabeled and deuterated products,

showing that exogenous methanethiol can intercept the natural methylthiolating agent. Last,

treatment of each protein with SAM in the absence of substrate or low-potential reductant (i.e.

41

dithionite) followed by treatment with [methyl-d3]-SAM in the presence of substrate and

dithionite results in a burst of unlabeled product followed by slower formation of the labeled

product, suggesting that the radical-dependent transfer of a methylthio group to the substrate is

fast relative to SAM-dependent methylation of the protein.

Figure 2-1. Reactions of the three major classes MTTases: (A) MiaB; (B) MtaB; (C) RimO. In

each reaction two molecules of SAM are cleaved to give one molecule of 5'-dAH and one

molecule of SAH

42

Materials and Methods

Materials

All DNA-modifying enzymes and reagents were from New England Biolabs (Ipswich,

MA). Sodium sulfide (nonahydrate), L-tryptophan, 2-mercaptoethanol, L-(+)-arabinose, ferric

chloride, sodium methanethiolate, 5’-deoxyadenosine (5’-dA), and S-adenosyl-L-homocysteine

(SAH) were purchased from Sigma Corp (St. Louis, MO). N-(2-hydroxyethyl)piperizine-N'-(2-

ethanesulfonic acid) (HEPES) was purchased from Fisher Scientific (Pittsburgh, PA), and

imidazole was purchased from J. T. Baker Chemical Co. (Phillipsburg, NJ). Potassium chloride,

glycerol, and expression vectors pET-28a and pET-26b were purchased from EMD Chemicals

(Gibbstown, NJ), while dithiothreitol (DTT) and nickel nitrilotriacetic acid (Ni-NTA) resin were

purchased from Gold Biotechnology (St. Louis, MO). Coomassie blue dye-binding reagent for

protein concentration determination was purchased from Pierce (Rockford, IL), as was the bovine

serum albumin standard (2 mg/mL). Nick, NAP-10, and PD-10 pre-poured gel-filtration columns,

as well as Sephadex G-25 resin were purchased from GE Biosciences (Piscataway, NJ). All other

buffers and chemicals were of the highest grade available.

Methods

Preparation of Substrates for Tm RimO Reactions

SAM, S-adenosyl-L-[methyl-d3]methionine (d3-SAM), S-adenosyl-L-[methyl-

14C]methionine ([methyl-

14C]SAM), and S-[8-

14C]adenosyl-L-methionine ([adenosyl-

14C]SAM)

were synthesized and purified as described previously (39). Oligonucleotide sequencing was

43

conducted at the Penn State Huck Nucleic Acid Facility. The S12 peptide substrate (1) for RimO

(NH2-RGGRVKDLPGVRY-COOH) and a synthetic peptide substrate (2), used as an external

standard (NH2-PMSAPARSM-COOH), was synthesized by the Peptide Synthesis Facility at New

England Biolabs (Ipswich, MA) as described previously (12) or by the Penn State Hershey

College of Medicine Macro Core Facility. The sequence of the peptide corresponds to residues

83-95 of the Tm S12 protein, and the Asp residue (D) in bold type corresponds to D89, the site of

methylthiolation.

UV/vis spectra were recorded on a Cary 50 spectrometer from Varian (Walnut Creek,

CA) using the WinUV software package for spectral manipulation and to control the instrument.

Oxygen-sensitive samples were prepared in an anaerobic chamber and aliquoted into cuvettes that

were sealed before being removed from the chamber. High performance liquid chromatography

(HPLC) was conducted on an Agilent Technologies (Santa Clara, CA) 1100 system that

contained a variable wavelength detector and an autosampler for sample injection. The instrument

was operated via the ChemStation software package, which was also used for data analysis.

Liquid chromatography/mass spectrometry (LC/MS) was conducted on an Agilent Technologies

1200 system coupled to an Agilent Technologies 6410 QQQ mass spectrometer with

simultaneous UV/vis analysis using an Agilent diode-array detector. The system was operated

with the associated MassHunter software package, which was also used for data collection and

analysis. Sonic disruption of Ec cell suspensions was carried out as described previously (12), and

liquid scintillation counting was conducted on a Beckman LS 6500 scintillation counter using 5

mL of Ecoscint scintillation cocktail per mL of aqueous sample.

44

Cloning and Overexpression of the Tm rimO gene

The Tm rimO gene was amplified from Tm genomic DNA using the following forward

(5’- CGC GGC GTC CAT ATG AGG GTT GGT ATA AAG GTT CTA GGA TGT CC -3’) and

reverse (5’- CGC GGC GTC GAA TTC TCA TAT CAC TGA CCC CCA CAT GTC GTA CTC

G-3’) primers. The forward primer included an NdeI restriction site (underlined) flanked by a

nine-base GC clamp and the first 29 bases of the rimO gene. The reverse primer contained an

EcoRI restriction site (underlined) flanked by a nine-base GC clamp and the last 31 bases of the

rimO gene, including the stop codon. After amplification, the product was digested with NdeI and

EcoRI and ligated into similarly digest pET-28a by standard procedures. The correct construct

was verified by DNA sequencing and designated pTmRimO.

Expression vector pTmRimO was transformed into Ec BL21(DE3) along with plasmid

pDB1282 as previously described (41, 42). Bacterial growth and gene expression was carried out

at 37 °C in 16 L of M9 minimal media distributed evenly among 4 Erlenmeyer flasks with

moderate shaking (180 rpm). At an optical density (OD) at 600 nm of 0.3, solid L-(+)-arabinose

was added to each flask to a final concentration of 0.2 % (w/v), while cysteine and ferric chloride

were added to final concentrations of 300 µM and 50 µM, respectively. At an OD600 of 0.6, a

sterile solution of IPTG was added to each flask to a final concentration of 200 µM. Expression

was allowed to take place for 16 h at 18 °C before the cells were harvested by centrifugation at

10,000 g for 10 min at ambient temperature.

Purification of Tm RimO

Purification of Tm RimO was carried out by immobilized metal affinity chromatography

(IMAC) using Ni-NTA resin. All purification steps were performed in a Coy (Grass Lakes, MI)

45

anaerobic chamber (unless specifically stated otherwise), which was kept under an atmosphere of

N2 and H2 (95%/5%). The O2 concentration was maintained below 1 ppm by using palladium

catalysts. Buffers used during the purification of Tm RimO were as follows: lysis buffer (50 mM

HEPES, pH 7.5, 300 mM KCl, 10 mM 2-mercaptoethanol, 20 mM imidazole, and 1 mg/mL

lysozyme); wash buffer (50 mM HEPES, pH 7.5, 300 mM KCl, 10 mM 2-mercaptoethanol, 10%

(v/v) glycerol, and 40 mM imidazole); elution buffer (wash buffer containing 250 mM

imidazole). After lysing the cells by sonication (41), the cell suspension was transferred into

sterile centrifuge tubes, which were subsequently sealed and heated at 70 °C for 1 h outside of the

anaerobic chamber. After subjecting the heat-treated solution to centrifugation at 50,000 g and

4°C for 1 h, the supernatant was loaded onto a Ni-NTA column, which was subsequently washed

with 200 mL of wash buffer. After addition of elution buffer to the column, fractions containing

RimO, distinguished by their dark brown color, were pooled and concentrated using an Amicon

stirred ultrafiltration apparatus (Millipore, Billerica, MA) fitted with a YM-30 membrane (30,000

molecular weight cutoff). The protein was exchanged into gel-filtration buffer (GFB) (50 mM

HEPES, pH 7.5, 300 mM KCl, 20% glycerol, and 1 mM DTT) using a Sephadex G-25 column

(2.5 13 cm), reconcentrated, and stored in aliquots in a liquid N2 dewar until ready for use.

Protein, Iron, and Sulfide Quantification

The concentrations of Tm RimO were determined by the procedure of Bradford (43)

using bovine serum albumin (Fraction V) as a standard. Quantitative amino acid analysis,

conducted as described previously (44), indicates that the procedure of Bradford overestimates

the concentration of Tm RimO by a factor of 1.47. Iron and sulfide analyses were performed

according to the procedures of Beinert (45-47).

46

Chemical Reconstitution of Tm RimO

Tm RimO was treated with 10 mM DTT before being incubated for 10 min with an 8-fold

molar excess of FeCl3. An 8-fold molar excess of sodium sulfide was added over the course of 3

to 4 h, upon which the solution was subjected to centrifugation at 18,000 g. The supernatant

was exchanged into storage buffer by gel-filtration (G-25) chromatography and concentrated by

ultrafiltration using an Amicon stirred ultrafiltration apparatus fitted with a YM-10 membrane.

Following chemical reconstitution, Tm RimO was further purified by fast protein liquid

chromatography (FPLC) on an S-200 column using an ÄKTA liquid chromatography system (GE

Biosciences) housed in an anaerobic chamber. The column was equilibrated in buffer consisting

of 50 mM HEPES, pH 7.5, 300 mM KCl, 5 mM DTT, and 10% glycerol. Fractions were pooled

based on absorbencies at 280 and 400 nm, and concentrated and stored as described above.

Tm RimO Activity Assays

Tm RimO reactions contained the following in a final volume of 180 µL: 67 µM Tm

RimO, 700 µM SAM, 300 µM S12 peptide substrate (1), 50 mM Na-HEPES, pH 7.5, 2 mM

dithionite, and 1 mM tryptophan (IS). All components except SAM were incubated at 37 °C for 3

min before initiating the reaction with the omitted component. Aliquots (20 µL) of the reaction

mixture were withdrawn at various times from 0-180 min and added to 20 µL of 0.1 M H2SO4

containing 20 µM peptide 2 (ES) to quench the reaction. Precipitated protein was removed by

centrifugation at 18,000 g for 15 min, and a 20 µL aliquot of the resulting supernatant was

subjected to analysis by ESI+ LC/MS with single-ion monitoring (SIM). Solvent A consisted of

ammonium acetate (40 mM) and methanol (5% v/v) titrated to pH 6.2 with acetic acid, while

solvent B was 100 % acetonitrile. The column was equilibrated in 100% solvent A at a flow rate

47

of 0.5 mL min-1

. After sample injection (2 µL), a gradient was applied from 0% solvent B to 2%

solvent B over 0.5 min, 2% to 28% over 4.5 min, and then 28% to 0% over 3 min. The monitored

ions (m/z) and retention times (min), respectively, were 385.1 and 0.8 (SAH), 188.0 and 1.2

(tryptophan), 252.1 and 2.1 (5'-dA), 474.4 and 3.2 (peptide 2, ES), 498.1 and 4.4 (peptide 1), and

507.1 and 4.4 (MS-1). Calibration curves were generated with known concentrations of each

analyte and run under identical conditions to determine the concentration of products generated in

assays. Data were analyzed using the Agilent Technologies MassHunter qualitative and

quantitative analysis software.

Tm RimO Radioactivity Assays

Tm RimO was incubated for 1 to 2 h at 37 °C with [methyl-14

C]SAM (specific

radioactivity: 910 cpm/nmol) or [8-14

C]SAM (specific radioactivity: 1110 cpm/nmol) in reactions

containing the following in a total volume of 100 µL: 50 mM HEPES, pH 7.5, 273 µM Tm RimO,

and 1 mM radiolabeled SAM. After incubation, the reactions were applied to pre-poured gel-

filtration columns equilibrated in (i) 10 mM HEPES, pH 7.5, 150 mM KCl, 10% glycerol, and 5

mM DTT; (ii) 10 mM HEPES, pH 7.5, 150 mM KCl, 10% glycerol, 5 mM DTT, and 200 mM

NaOH; or (iii) 10 mM HEPES, pH 7.5, 150 mM KCl, 10% glycerol, 5 mM DTT, and 8 M urea. A

200 µL aliquot of the protein-containing fraction (400 µL total volume) was analyzed directly by

scintillation counting. A 50 µL aliquot of the protein-containing fraction was added to 10 µL of a

carrier solution containing 100 µM each of SAM, SAH, 5’-dA, adenine, and MTA. The resulting

solution was acidified by addition of 60 µL of 100 mM H2SO4 before a 100 µL aliquot was

withdrawn for HPLC analysis as described above for detection of 5’-dA and SAH. Fractions were

collected throughout the entire chromatographic procedure, and fractions with retention times

corresponding to those of each carrier component were pooled and subjected to scintillation

48

counting. Control samples containing either of the two radiolabeled forms of SAM, but lacking

Tm RimO, were prepared and treated as described above for the complete assays.

Determination of Tm RimO-Dependent Production of Methanethiol

Assays contained the following components in a final volume of 700 µL: 50 mM HEPES,

pH 7.5, 1 mM SAM, 1 mM tryptophan, 67 µM Tm RimO or 100 µM Tm MiaB, 1 mM SAM, and

when appropriate, 300 µM S12 peptide or 200 µM ACSL RNA. Reactions were performed in

triplicate at ambient temperature in septum-sealed vials, and were initiated by addition of SAM.

At designated times, 20 µL aliquots were removed and added to an equal volume of 0.1 M H2SO4

for quantification of SAH by LC/MS. An equal volume of 1 M HCl was injected into the

remaining 80 µL, and the reaction was incubated further at 42 °C for 30 min to allow

equilibration of methanethiol between the liquid and gas phases. An aliquot (500 µL) of the

headspace was removed using a gas-tight syringe and analyzed by gas chromatography/mass

spectrometry (GC/MS) using a Shimadzu GC-17A gas chromatograph connected to a Shimadzu

GCMS-QP500 mass spectrometer and a Restek Rxi-1ms 30 m column (ID: 0.32 narrow bore;

film: 4.0 µm) (Restek; Bellefonte, PA). The inlet and oven temperatures were both maintained at

30 °C, while the detector was set to 300 °C. Total ion chromatograms were generated under SIM

conditions (m/z of 47).

Tm RimO Differential Labeling Assays

A reaction mixture containing 533 µM Tm RimO, 50 mM Na-HEPES, pH 7.5, and 1 mM

SAM in a total volume of 100 µL was incubated for 18 h at 37 °C and then subjected to AGFC to

remove SAH and unreacted SAM. After determining the concentration of Tm RimO following

49

AGFC (132 µM), the protein (66 µM) was incubated with 0.7 mM [methyl-d3]SAM for 3 min, 1

h, or 3 h, before initiating turnover by addition of dithionite. At appropriate times, aliquots of the

reaction mixture were removed and quenched with an equal volume of 0.1 M H2SO4 containing

peptide 2 (ES) and then analyzed for SAH, 5’-dA, and labeled and unlabeled MS-1 by LC-MS.

Results

In our previous studies of Ec RimO, we reported that the enzyme catalyzed formation of

SAH in the absence of dithionite and substrate (12). This behavior was also observed in our

studies of the methylsynthases, Cfr and RlmN, which represent the best characterized of the RS

enzymes that use SAM both as a precursor to a 5’-dA• to initiate radical-dependent chemistry and

the source of an appended methyl group (37). This observation suggested the possibility that,

similar to RlmN and Cfr, the MTTases might also operate via a ping-pong mechanism, wherein

the methyl group is first appended to an amino acid residue or enzyme prosthetic group before

being transferred to the product. Although the amount of turnover by Ec RimO was exceedingly

low, SAH was generated in amounts similar to those of the methylthiolated product (12). Similar

studies by Arragain and coworkers on Tm RimO showed that this enzyme is better suited for

mechanistic interrogation; it supported production of ~2 nmol of product per nmol of RimO

polypeptide (3). Interestingly, the authors reported that Tm RimO catalyzed production of the

intended mono-methylthiolated product as well as a bis-methylthiolated product when a 20 aa

peptide containing the sequence surrounding D89 of protein S12 was used as a substrate. The

former product was shown to contain the methylthiol modification at the intended location (D89),

but the location of the second methylthiol group was not determined. The authors also reported

that SAH was produced in a reaction mixture lacking the peptide substrate but requiring

dithionite, but did not provide supporting data (3, 16).

50

To determine whether methyl transfer in the absence of substrate and dithionite is a

general characteristic of MTTases, we cloned the gene that encodes RimO from Thermotoga

maritima (Tm) to study the behavior of its encoded protein. The Tm rimO gene was co-expressed

with genes on plasmid pDB1282, which encodes proteins involved in Fe/S cluster biosynthesis

and insertion in Azotobacter vinelandii (42, 51). Tm RimO was produced with an N-terminal

hexahistidine tag and was routinely reconstituted (RCN) with additional iron and sulfide using

previously described methods (12, 42). Displayed in Figure 2-2 is a typical UV-vis traces of AI

and RCN Tm RimO (solid and dashed lines respectively) which is similar to those reported

previously (3, 10, 12).

Figure 2-2. UV/vis spectra of AI (solid lines) and RCN (dashed lines) Tm RimO. Protein

concentrations were: 19.5 µM (2.4 ± 0.08 Fe and 4.7 ± 0.26 S2-

) and 22.3 µM (4.6 ± 0.08 Fe and

8.5 ± 0.24 S2-

) for AI and RCN Tm RimO.

51

Turnover by Tm RimO

Turnover by Tm RimO was measured using a 13 aa synthetic peptide composed of the

sequence surrounding D89 (bold type) of protein S12 (NH2-RGGRVKDLPGVRY–COOH),

which is perfectly conserved between Ec and Tm. Quantification of SAM, SAH, 5’-dA, and the

S12 peptide (heretofore termed 1) was conducted by LC/MS using standard curves that were

constructed with authentic compounds. The methylthiolated peptide (MS-1) was quantified using

1, with the assumption that it ionized with similar efficiency. Consistent with this assumption, the

time-dependent concentrations of MS-1 formation and 1 decay were similar, and no other

peptide-related species were observed during analysis (Figure 2-3, panel A). The S12 peptide

exhibits m/z = 491.8 (+3 charge state), and MS-1 exhibits m/z = 507.1 (+3 charge state). No

evidence for m/z = 522.4 (+3 charge state) was observed, which would correspond to a bis-

methylthiolated species. The bis-methylthiolated species was also not observed when the +1 or

+2 charge states were monitored. Figure 2-3, panel A depicts the time-dependent formation of

MS-1 (black line), 5’-dA (red line), and SAH (blue line) under turnover conditions in the

presence of 67 µM Tm RimO, as well as the time-dependent loss of 1 (green line). In contrast to

our previous studies on Ec RimO, wherein the amount of SAH generated was meager (<10% the

concentration of enzyme), Tm RimO catalyzed formation of ~3 equiv of 5’-dA and 4 equiv of

SAH per equiv of enzyme in ~80 min. Importantly, the concentration of MS-1 formed (~114 µM)

is nearly twice the Tm RimO concentration in the assay (66 µM) as well as the amount of peptide

consumed (~131 µM), and the initial rate for MS-1 formation (5.8 + 0.3 µM min-1

)) is similar to

the initial rate of consumption of 1 (9.7 + 0.3 µM min-1

). Therefore, it appears that Tm RimO

catalyzes more than one turnover, as shown previously (3, 16). Figure 2-3, panel B depicts a Tm

RimO reaction containing SAM and dithionite but lacking 1. SAH is still produced at a similar

concentration after 80 min of reaction; however, neither time-dependent formation of 5’-dA (red

52

line) or MS-1 is observed, implying that the peptide substrate triggers radical generation but not

SAH formation. Interestingly, SAH is generated relatively rapidly (ν = 13.7 + 0.2 µM min-1

), but

in a 3-fold lower concentration when both dithionite and 1 are omitted (Figure 2-3, panel C).

Moreover, the concentration of SAH generated (73 µM) is almost equivalent to the concentration

of enzyme in the assay (67 µM). This behavior implies that Tm RimO does not require that

radical chemistry take place before methyl transfer, and presents the possibility that sulfur

insertion may not precede methyl transfer as had been suggested previously (1).

Figure 2-3. Tm RimO catalyzed reactions at 37 °C (A) under turnover conditions with SAM, 1,

and dithionite, (B) in the presence of SAM and dithionite, but absence of 1, and (C) in the

presence of SAM, but absence of dithionite and 1. Blue squares, SAH formation; red circles, 5'-

dA formation; black triangles, MS-1 formation; and green diamonds, consumption of 1. The

reactions were conducted as described in Materials and Methods, and contained, where

appropriate, 67 µM Tm RimO, 300 µM 1, 1 mM SAM, and 2 mM dithionite. The lines are fits to

a first-order single-exponential equation, resulting in the following kinetic parameters: (A) SAH

formation: A = 239 + 3 µM, ν = 11.9 + 0.7 µM min-1

; 5'-dA formation: A = 183 + 17 µM, ν =

47.6 + 4.4 µM min-1

; MS-1 formation: A = 115 + 6 µM, ν = 5.8 + 0.3 µM min-1

; consumption of

1: A = 139 + 4 µM, ν = 9.7 + 0.3 µM min-1

; (B) SAH formation: A = 359 + 65 µM, ν = 3.6 + 0.7

µM min-1

; (C) SAH formation: A = 72 + 1 µM, ν = 13.7 + 0.2 µM min-1

, k = 0.19 + 0.02 min-1

.

Radiotracing methyl transfer from ([methyl-14

C])SAM to Tm RimO

To determine the nature of the species to which SAM donates its methyl moiety in the

absence of substrate and/or reductant, studies were conducted with S-adenosyl-L-[methyl-

14C]methionine ([methyl-

14C]SAM) or S-[8-

14C]adenosyl-L-methionine ([adenosyl-

14C]SAM)

53

(Figure 2-4). When Tm RimO (43.1 nmol) was incubated with [methyl-14

C]SAM to allow for

methyl transfer, and then the reaction subjected to anaerobic gel-filtration chromatography

(AGFC), two peaks of radioactivity were observed, an early peak containing the protein fraction

and a later peak containing only small molecules (Figure 2-4, panel B). From the specific

radioactivity of SAM, it was calculated that 33.4 nmol (~0.75 equiv) of radioactivity was attached

to the protein (Figure 2-4, panel B). A control experiment, in which an equal concentration and

amount of [methyl-14

C]SAM was applied to the gel-filtration column in the absence of Tm RimO,

showed the presence of only 1 peak of radioactivity, which elutes with small molecules (Figure

2-4, panel A). As shown in Figure 2-4, panel C, when Tm RimO was incubated with excess

[methyl-14

C]SAM and then applied to a gel-filtration column equilibrated in gel-filtration buffer

(GFB) containing 0.2 N NaOH, no radioactivity eluted with the protein fraction, consistent with

attachment of the radioactive moiety to a species that is unstable under very basic conditions. In a

similar experiment, in which the reaction mixture was applied to a gel-filtration column

equilibrated in GFB containing 8 M urea, only 10 nmol of radioactivity eluted with the protein

fraction (Figure 2-4, panel E), suggesting that the stability of the methyl acceptor is influenced

by the integrity of the overall fold of the protein.

54

Figure 2-4. Elution profiles of Tm RimO incubated with [methyl-14

C]SAM or [adenosyl-

14C]SAM and analyzed subsequent to anaerobic gel-filtration chromatography (AGFC) under

various conditions. (A) [methyl-14

C]SAM in gel-filtration buffer (GFB); (B) Tm RimO +

[methyl-14

C]SAM in GFB; (C) Tm RimO + [methyl-14

C]SAM in GFB + 200 mM NaOH; (D) Tm

RimO + [adenosyl-14

C]SAM; (E) Tm RimO + [methyl-14

C]SAM in GFB containing 8 M urea.

In a subsequent experiment, Tm RimO was incubated for 2 h at 37 °C with [methyl-

14C]SAM in the absence of substrate and dithionite and then subjected to AGFC. The protein

fraction (6.8 nmol) was treated with 50 mM H2SO4 (final concentration), and a fraction of the

resulting supernatant obtained after centrifugation was analyzed by HPLC with radiometric

detection. As shown in Figure 2-5, panel B, very little radioactivity in SAM (0.2 nmol), 5’-dA

(0.003 nmol), SAH (0.003 nmol), adenine (0.29 nmol), or MTA (0.002 nmol) eluted with the

protein after AGFC, and no other significant peaks of radioactivity were found in any other

region of the chromatograph. Experiments conducted with [adenosyl-14

C]SAM corroborated the

observations obtained using [methyl-14

C]SAM. When Tm RimO (43.1 nmol) was incubated with

55

excess [adenosyl-14

C]SAM and then subjected to AGFC, ~24.6 nmol of radioactivity eluted with

the protein fraction (Figure 2-4, panel D). Upon analysis of a portion of the protein-containing

fraction by HPLC with radiometric detection, the vast majority of the radioactivity (2.8 nmol)

eluted with the SAH standard (Figure 2-5, panel A), while only 0.23 nmol eluted with SAM.

These observations suggest that upon SAM binding to Tm RimO, transfer of a methyl group from

SAM to an acid- and base-labile acceptor takes place. Moreover, the labile acceptor appears to be

volatile under acidic conditions, given that the radiolabeled methyl group is not observed during

HPLC with radiometric detection. The instability of the methylated species in the presence of

urea argues that the methyl group is not transferred to an amino acid (e.g. Glu or Asp) to afford

an ester or some other acid- or base-labile organic species, but rather to an acceptor whose

presence depends on the integrity of the overall protein fold.

Figure 2-5. HPLC elution profiles monitored at 260 nm of AGFC protein fraction from Tm RimO

incubated with (A) [adenosyl-14

C]SAM and (B) [methyl-14

C]SAM. Relative amount (nmol) of

radioactivity for each compound is indicated in red. The elution times are as follows: SAM, 3

min; adenine, 4.3 min; SAH, 5.7 min; 5’dA, 6.4 min; methylthioadenosine (MTA), 11 min.

56

Tm RimO-Catalyzed Formation of Methanethiol

SAH is not a typical degradation product of SAM; its formation at significant rates

requires enzymatic assistance (39). The results described above suggest that Tm RimO (i)

catalyzes the adventitious attack of a water molecule onto the activated methyl group of SAM, or

(ii) catalyzes transfer of the methyl group from SAM onto perhaps one of the bridging µ-sulfido

ions (or an externally ligated sulfide ion) of, most probably, the N-terminal [4Fe–4S] cluster. In

the former case, methanol would be produced, and would need to be tightly bound to the enzyme

to survive gel-filtration. In the latter case, methanethiol (CH3SH) would be produced, but only

after treatment of the methylated enzyme with acid or base. Under acidic conditions, CH3SH is

volatile, which would explain our inability to detect it radiometrically in our HPLC

chromatograms of acid-quenched samples. Figure 2-6 shows chromatograms of varying

concentrations of methanol analyzed by GC-MS. As can be observed, the limit of detection of

methanol is significantly less than 8 µM. When Tm RimO (67 µM) was incubated with SAM for

2 h in the absence of substrate and dithionite to allow for methyl transfer, methanol was not

detected in either the liquid or gas phases upon GC/MS analysis of the reaction after quenching in

either acid or base.

57

Figure 2-6. GC-MS total ion chromatogram of methanol at various concentrations using single-

ion monitoring at m/z = 31. Methanol standards eluted at 2.1 min from an Rxi-1MS column. GC

was performed using the following parameters: 100 °C injection temperature; 300 °C interface

temperature; oven temperature gradient from 50 °C to 75 °C over 5 min. 1 µL of methanol

standards were injected, and were prepared and treated as described for methanethiol samples.

The lower limit of detection is less than 8 µM.

GC-MS was also used to analyze for time-dependent CH3SH formation in assays

containing Tm RimO and SAM, but in the absence of substrate and dithionite. Assays were

conducted in septum-sealed vials and quenched in acid at appropriate times. The quenched

samples were incubated further at 42 °C to allow equilibration of CH3SH between the liquid

phase and the headspace of the vial before an aliquot of the headspace was removed and

analyzed. A standard curve was generated with commercially available sodium methanethiolate

(NaSCH3), which was added to reaction mixtures containing all components except Tm RimO.

The samples comprising the standard curve were quenched and treated as described above for the

experimental samples. In Figure 2-7, the time-dependent formation of SAH (blue squares) and

CH3SH (red circles) is displayed for reactions containing SAM and 67 µM Tm RimO. The lines

in each of the graphs is a fit of the data to a first-order, single-exponential kinetic equation,

58

affording the following amplitudes (A) and initial rates (ν) for SAH and CH3SH formation,

respectively: A = 40 ± 0.8 µM, ν = 4.0 + 0.1 µM min-1

and 37 ± 2 µM, ν = 3.7 + 0.2 µM min-1

. As

can be observed, CH3SH formation closely parallels SAH formation in amplitude and initial rate.

Figure 2-7. Time-dependent formation of SAH and methanethiol by Tm RimO. Each reaction

was performed in triplicate. The dashed lines are fits to a first-order single-exponential kinetic

equation: (A) SAH formation: A = 40 + 0.8 µM, ν = 4.0 + 0.1 µM min-1

, (blue squares);

methanethiol formation: A = 37 + 2 µM, ν = 3.7 + 0.2 µM min-1

, (red circles). The reactions were

conducted as described in Materials and Methods. Reaction mixtures contained 67 µM Tm RimO

or 100 µM MiaB, 50 mM Na-HEPES (pH 7.5), 1 mM tryptophan, and 1 mM SAM.

Turnover in the Presence of Exogenously Supplied Methanethiol

To assess whether the methylated sulfur ion is in exchange with free CH3SH in catalysis

by Tm RimO (67 µM), reactions were conducted with unlabeled SAM or S-adenosyl-[methyl-

d3]methionine (d3-SAM) in the absence or presence of sodium methanethiolate (NaSCH3). In

Figure 2-8, panel A, the Tm RimO-catalyzed time-dependent production of MS-1 in the presence

of SAM (2 mM), 1 (300 µM), and NaSCH3 (1 mM) is displayed (black line). As can be observed,

in the presence of NaSCH3, Tm RimO catalyzes multiple turnovers (~4.5 per polypeptide).

Moreover, in contrast to reactions that lack NaSCH3, the amount of SAH produced (blue line) is

less than the amount of product produced, presumably because NaSCH3 incorporation onto or

into the acceptor site blocks methyl transfer from SAM. The initial rate for MS-1 formation (25.8

59

+ 0.9 µM min-1

) is similar to the initial rate for 5’-dA formation (40.4 + 0.9 µM min-1

) and

disappearance of 1 (33.7 + 1.2 µM min-1

); however, the initial rate for SAH formation is

somewhat slower (10.5 + 0.2 µM min-1

). It should be noted that the initial rate of SAH formation

recapitulates the initial rate for product formation in Figure 2-3, panel A, suggesting that in the

absence of methanethiol, methyl transfer from SAM to the acceptor limits the rate of the reaction.

The faster initial rate of product formation in assays containing methanethiol suggests that the

small molecule is efficiently incorporated into a binding site on the enzyme, and that at the

earliest times, product containing a methylthio group from methanethiol predominates over

product containing a methyl group from SAM. Although it appears that the reaction subsides only

after all of the substrate has been consumed, reactions conducted with much smaller

concentrations of enzyme (~13 µM) also show a leveling off after ~4 turnovers.

Figure 2-8. Time-dependent formation of MS-1 in the presence of 1 mM methanethiol and 2 mM

SAM (A) and time-dependent formation of MS-1 and d3-MS-1 in the presence of 2 mM

methanethiol and 2 mM d3-SAM (B). SAH formation (blue squares), 5'-dA formation (red

circles), MS-1 formation (black triangles), d3-MS-1 formation (yellow right triangles), MS-1 +

d3-MS-1 formation (gray crosses), and consumption of 1 (green diamonds). The reactions were

conducted as described in Materials and Methods. Both reaction mixtures contained 67 µM Tm

RimO, 50 mM Na-HEPES (pH 7.5), 1 mM tryptophan, 2 mM dithionite, and 350 µM 1. The lines

are fits to a first-order single-exponential equation, with the following obtained parameters: (A)

SAH: A = 228 + 4 µM, ν = 10.5 + 0.2 µM min-1

,; 5'-dA: A = 367 + 8 µM, ν = 40.4 + 0.9 µM min-

1, k; MS-1: A = 304 + 11 µM, ν = 25.8 + 0.9 µM min

-1,; 1: A = 306 + 11 µM, ν = 33.7 + 1.2 µM

min-1

,. (B) SAH: A = 257 + 13 µM, ν = 10.8 + 0.5 µM min-1

; 5'-dA: A = 347 + 7 µM, ν = 41.6

+ 0.8 µM min-1

; d3-MS-1: A = 172 + 6 µM, ν = 10.3 + 0.4 µM min

-1, ; MS-1: A = 164 + 16 µM,

60

ν = 29.5 + 2.9 µM min-1

; MS-1 + d3-MS-1: A = 325 + 21 µM, ν = 35.8 + 2.3 µM min-1

; 1: A =

287 + 4 µM, ν = 27.3 + 0.4 µM min-1

.

To show convincingly that a methylthio group from NaSCH3 is incorporated into the

product during turnover, Tm RimO assays were conducted with d3-SAM (2 mM) in the presence

(2 mM) and absence of NaSCH3. Product arising from the endogenous, natural, pathway should

contain a deuterated methyl group deriving from d3-SAM, while product arising from the

exogenous pathway should contain an unlabeled methyl group deriving from NaSCH3. In Figure

2-8, panel B, the Tm RimO (67 µM)-catalyzed time-dependent production of unlabeled (–SCH3)

MS-1 (black trace), labeled (–SCD3) MS-1 (yellow trace), SAH (blue trace), and 5’-dA (red

trace) are displayed, as well as the time-dependent loss of 1 (green trace). As can be seen, initial

production of unlabeled MS-1 is faster (29.5 + 2.9 µM min-1

) than that of labeled MS-1 (10.3 +

0.4 µM min-1

); however, both labeled and unlabeled species are produced in approximately

equimolar concentrations (A = 164 ± 16 µM and 172 ± 6 µM, respectively). It appears that for

production of labeled MS-1, the rate-limiting step is methyl transfer, given that SAH is produced

with a similar initial rate (ν = 10.8 + 0.5 µM min-1

). The gray trace in Figure 2-8, panel B, is the

sum of the black and yellow traces; it closely mirrors 5’-dA production (red trace) in both

amplitude (325 ± 21 µM vs. 347 ± 7 µM) and initial rate (41.6 + 0.8 µM min-1

vs. 35.8 + 2.3 µM

min-1

), consistent with both unlabeled and labeled products being generated via a radical-

dependent process, and tighter coupling of radical generation and product formation in the

presence of methanethiol. The black and yellow traces, only, are shown in Figure 2-9, allowing

for better visualization of production of the two differentially labeled products at early time

points.

61

Figure 2-9. Time-dependent formation of MS-1 (black triangles) and d3-MS-1 (yellow right

triangles) in the presence of 2 mM methanethiol and 2 mM d3-SAM. The lines are fits to a first-

order single-exponential equation, with the following obtained parameters: MS-1: A = 164 + 16

µM, ν = 29.5 + 2.9 µM min-1

; d3-MS-1: A = 172 + 6 µM, ν = 10.3 + 0.4 µM min

-1.

Chemical and Kinetic Competence of a Potential Intermediate

If Tm RimO follows a ping-pong mechanism, it should be possible to isolate the

intermediate form of the protein after incubating the protein with the first substrate in the reaction

(ping step), and then re-introduce the intermediate form into a reaction containing only the second

substrate (pong step). One caveat of this common method to show chemical competence is that in

the MTTases, the same cosubstrate (SAM) is required in both steps of the reaction. However, the

finding that SAM is used for distinctly different types of reactivities in each step, one of which

requires the presence of a low-potential reductant (dithionite), allows differentiation of the two

steps by omitting the low-potential reductant required to initiate radical chemistry. Therefore, the

first step, methylation of an acceptor, was conducted with unlabeled SAM in the absence of

dithionite, while the second step, radical-dependent introduction of a methylthio group into the

organic substrate, was conducted with d3-SAM in the presence of dithionite. Figure 2-10 displays

the results of these differential labeling experiments with Tm RimO, wherein the protein was

62

treated with excess SAM for 15 h and then subjected to AGFC before it was incubated with d3-

SAM, dithionite, and 1 (turnover conditions). In Figure 2-10, panel A, the time-dependent

formation of MS-1 (black triangles), d3-MS-1 (yellow triangles), SAH (blue squares), and 5’-dA

(red circles) is displayed, as well as the time-dependent loss of 1 (green squares), for a sample

that was incubated with d3-SAM for 3 min before addition of 1 and dithionite (in that order) to

initiate the reaction. Formation of unlabeled MS-1 occurs relatively rapidly (ν = 21.2 + 1.4 µM

min-1

); however, the concentration of unlabeled MS-1 plateaus at ~50 µM (0.75 equiv of

enzyme). Formation of d3-MS-1 occurs with a lag phase, implying a slow step that precedes d3-

MS-1 formation, which may involve methyl transfer, dissociation of SAH, and rebinding of

another molecule of SAM needed for radical generation. In Figure 2-10, panel B, only the black

and yellow curves are displayed, better revealing the pronounced lag associated with d3-MS-1

formation. In addition, this labeled product is produced in a two-fold greater ratio (~100 µM) than

the unlabeled product after 30 min of reaction time. Figure 2-11 displays repeats of the

experiment described in Figure 2-10, in which the intermediate form of Tm RimO was incubated

with d3-SAM for 1 h (A) and 3 h (B) before initiating the second phase of the reaction by

introduction of 1 and dithionite. As can be observed, these extended incubation times have no

significant effect on the distribution of the labeled and unlabeled MS-1 products, indicating that

exchange between the methylated acceptor and the methyl group of SAM does not take place, and

that the methyl donor in the second step of the reaction is not a bound molecule of SAM that

survived AGFC.

63

Figure 2-10. Time courses for the formation of 5'-dA (red circles), SAH (blue squares), MS-1

(black triangles), d3-MS-1 (yellow right triangles), and consumption of 1 (green diamonds) by Tm

RimO incubated with d3-SAM for 3 min after previous incubation with unlabeled SAM for -15 h

followed by AGFC (A); (B) panel A, but displaying only the formation of MS-1 (black triangles)

and d3-MS-1 (yellow right triangles). The lines are fits to a first-order single-exponential

equation, with the following obtained kinetic parameters for formation of MS-1 and d3-MS-1,

respectively: A = 47 + 3.0 µM, ν = 21.2 + 1.4 µM min-1

; A = 174 + 31 µM, ν = 5.7 + 1.0 µM min-

1. Reactions were conducted as described in Materials and Methods, and contained 67 µM RimO,

50 mM Na-HEPES (pH 7.5), 1 mM tryptophan, 300 µM 1, and 1 mM d3-SAM.

Figure 2-11. Time courses for the formation of MS-1 (black triangles) and d3-MS-1 (yellow right

triangles) by Tm RimO incubated with d3-SAM for (A) 1 h, and (B) 3 h after previous incubation

with unlabeled SAM for 15 h followed by AGFC. The lines are fits to a first-order single-

exponential equation, with the following obtained kinetic parameters for formation of MS-1 and

d3-MS-1, respectively: (A) A = 47 + 3 µM, ν = 21.2 + 1.4 µM min

-1,; A = 159 + 26 µM, ν = 6.0 +

64

1.0 µM min-1

; (B) A = 56 + 3 µM, ν = 23.0 + 1.2 µM min-1

,; A = 142 + 14 µM, ν = 9.8 + 1.0 µM

min-1

, . Reactions were conducted as described in Materials and Methods, and contained 67 µM

RimO, 50 mM Na-HEPES (pH 7.5), 1 mM tryptophan, 300 µM 1, and 1 mM d3-SAM.

Cumulatively, these results are consistent with a mechanism wherein a methyl group

from SAM is transferred to a sulfur ion—presumably located on one of the [4Fe–4S] clusters—by

an SN2 mechanism, which is followed by a radical-dependent transfer of an intact methylthio

group from the protein to the substrate. Based on the amount of SAH formed in the initial

methylation of the protein, it would appear that two sites on the protein are equally available for

methylation, but only one site is actually used to donate the methylthio group. Upon donation of

this methylthio group, this one site becomes available for one or two more rounds of methyl

transfer and subsequent methylthiolation.

Discussion

Previous in vivo studies on Ec MiaB led to the suggestion that the sequence of

methylthiolation involves initial sulfhydrylation of the substrate followed by capping of the sulfur

atom with a SAM-derived methyl group (54). Starvation of an Ec (rel met cys) mutant for

methionine (a precursor to SAM), but not cysteine, resulted in the trapping of a cytokinin-active

species suspected to be 2-thio-N6-(

2-isopentenyl)adenosine (s

2i6A), given that its treatment with

[methyl-14

C]SAM and a crude MiaB preparation resulted in incorporation of radioactivity into the

species. The observation that the species was cytokinin-active suggested that it contained a

dimethylallyl group, and the observation that radioactivity from [methyl-14

C]SAM was not

incorporated into tRNA isolated from Ec mutants starved for sulfur (cysteine or sulfate) lent

credibility to its assignment as s2i6A (54).

65

The MTTases represent one of a few classes of enzymes wherein a single polypeptide

directs two distinct chemical outcomes for SAM—reductive cleavage to a 5’-dA• and SN2-based

methyl transfer to an acceptor, affording SAH as a co-product. In the best studied class,

represented by the RS methylases RlmN and Cfr, which catalyze the synthesis of methyl groups

at C2 and C8, respectively, of adenosine 2503 of 23S rRNA, catalysis takes place via a ping-pong

like mechanism, involving an initial SN2-based transfer of a methyl group to a target Cys residue

before it is transferred to the nucleotide substrate via radical-dependent chemistry (34, 38).

Studies detailed herein provide strong evidence for a similar ping-pong-like mechanism for

MTTases. As we observed for Ec RimO (12), Tm RimO catalyzeS formation of SAH from SAM

in the absence of substrate and/or dithionite, a reductant with a suitably low redox potential to

initiate radical-dependent chemistry. In the absence both of substrate and dithionite, the formation

of SAH follows hyperbolic kinetics, with the maximum concentration generated approaching the

concentration of enzyme in the reaction. Our results are consistent with the transfer of a methyl

group from SAM to an acceptor on the protein that is labile in the presence of acid and base, and

moderately labile in the presence of chaotropic agents such as urea. The lability of the acceptor in

the presence of agents that denature the overall fold of the protein suggests that the acceptor is

most likely not an amino acid residue whose methylated side chain can be hydrolyzed in the

presence of acid or base (e.g. methylglutamate or methylaspartate), and our inability to detect

methanol after denaturing the protein under acidic or basic conditions indicates that the acceptor

is not a tightly bound water molecule. Indeed, subsequent to treatment of Tm RimO with SAM in

the absence of dithionite and denaturing the protein in acid, methanethiol is produced in amounts

that are stoichiometric with SAH. These results are consistent with a polar (SN2) transfer of a

methyl group from SAM to a sulfide ion.

When Tm RimO is incubated with SAM in the presence of dithionite, no 5’-dA is formed

unless substrate is present, indicating that radical-dependent chemistry is strongly coupled to

66

substrate binding. However, the presence of dithionite strongly affects the extent to which SAH is

formed, with the maximum concentration produced significantly exceeding the concentration of

enzyme. Although we do not know the exact basis for enhanced SAH production in the presence

of dithionite, it may derive from reduction of a reservoir of sulfane sulfur that is methylated by

SAM. This sulfane sulfur was recently observed in the holo crystal structure of Tm RimO,

wherein a pentasulfide bridge was observed to connect the unique iron ions of each of the [4Fe–

4S] clusters (16). It should be mentioned that dithionite is not a physiological reductant, and that

its unspecific reactivity can short-circuit natural catalytic sequences, as has been observed in BS,

which also contains two redox-active Fe/S clusters (8). The Ec flavodoxin/flavodoxin

reductase/NADPH reducing system appears capable of supplying the requisite electron for SAM

cleavage in most RS enzymes from Ec and some other organisms; however, it is relatively

ineffective in our Tm RimO reactions. Previous studies on Tm MiaB have shown that the

auxiliary cluster has a relatively high redox potential; it is fully reduced upon treatment with

dithionite, and the triple variant lacking the cysteines that coordinate the RS cluster is partially

reduced simply after isolation and reconstitution (10). Whether the oxidized or reduced form of

the auxiliary cluster functions in the initial stages of the physiological reaction mechanism is

currently unknown.

Our results further suggest that the methylated species is a chemically and kinetically

competent intermediate in the reaction. Not only is methanethiol produced after incubating Tm

RimO with SAM and then denaturing in acid, methanethiol introduced exogenously in reaction

mixtures serves as a perfectly good methylthiolating agent in the presence of SAM and dithionite,

as has recently been demonstrated by others (16). In fact, in reactions containing NaSCH3 and

[methyl-d3]SAM, production of unlabeled product is initially favored over the d3-containing

product. This observation suggests that exogenous methanethiol is efficiently activated toward

radical-dependent incorporation into organic substrates. Moreover, when Tm RimO is first treated

67

with unlabeled SAM in the absence of dithionite to allow for methyl transfer to the target

acceptor, and then treated with [methyl-d3]SAM under turnover conditions, production of the

unlabeled product precedes production of the labeled product. This behavior is dramatic in the Tm

RimO reaction, wherein a clear lag phase associated with d3-MS-1 production is observed during

the burst phase of MS-1 production. Not only is this initial methyl-containing species chemically

competent for methylthiolation, but the initial rate of product formation from the methyl-

containing species generated in the first half-reaction (k ~ 0.45 min-1

) indicates that it is also

kinetically competent. In fact, these differential labeling studies, as well as the studies detailed

above using exogenous methanethiol, suggest that methyl transfer from SAM is the rate-limiting

step in these reactions.

At present, we cannot readily explain the stoichiometry of unlabeled product to labeled

product that we see in our differential labeling experiments. Before initiating these experiments,

the mechanistic prediction was that we would observe a maximum of one equiv of

methylthiolated product per equiv of MTTase, and that the product would bear exclusively an

unlabeled methyl group. Surprisingly, in the Tm RimO reaction, we observed 0.7 equiv of the

unlabeled product, while an additional ~1.4 equiv of the labeled product was formed in a slower

process, whereas in the Tm MiaB reaction, we observed ~0.5 equiv of the unlabeled product,

while another 0.5 equiv was formed in a slower process. The recent crystal structure of holo Tm

RimO provides possible insight into these findings. Based on our observations, we suggest that

~70% of our Tm RimO reacts productively, and that our RCN Tm RimO contains a trisulfide

substituent coordinated to the unique iron ion of the auxiliary cluster (Figure 2-12). This idea is

consistent with the observation that the occupancy of the sulfur atoms in the pentasulfide bridge

is quite low. In other words, not all sulfur atoms are present at all times. In the Tm RimO reaction

we propose that methyl transfer from SAM to the external sulfide ion of the polysulfide

substituent takes place via polar SN2-based chemistry, most likely from SAM bound to the RS

68

[4Fe–4S] cluster. Upon reductive cleavage of SAM and abstraction of a H• from substrate by the

resulting 5’-dA•, the substrate radical attacks the terminal sulfur atom of the polysulfide chain

attached to the auxiliary cluster in its reduced state, resulting in transfer of the methylthiol group

to afford the product. This reaction produces a polysulfide chain that is shorter by one sulfur

atom, but which bears a nucleophilic terminal persulfide for another round of the exact same

chemistry (Figure 2-12). This proposed reaction mechanism, wherein SAM bound to the RS

[4Fe–4S] cluster is activated toward two distinct types of chemistry, is also consistent with the

relatively short distance (~8 Å) between the two Fe/S clusters as compared to that in MoaA (17

Å) (55) and the recently solved structure of the anaerobic sulfatase maturating enzyme from

Clostridium perfringens (12.9 Å) (56).

Recent studies suggest that a similar ping-pong-like mechanism may be operative in the

reaction catalyzed by the RS enzyme NifB. This enzyme plays a key role in the maturation of the

M cluster of Mo-nitrogenase, the metalloenzyme responsible for reduction of N2 to ammonia.

Mo-nitrogenase contains a complex metallocluster of the core composition 1Mo:7Fe:9S:1C. At

the center of this metallocluster is a carbide atom coordinated to six iron ions, which emanates

from the activated methyl group of SAM (57-60). Treatment of a NifEN-B fusion protein—in

which NifB is fused to the scaffold proteins NifEN—with SAM under turnover conditions results

in the production both of 5’-dA and SAH. Further labeling experiments with d3-SAM show

deuterium enrichment in 5’-dA, as was observed in the reactions catalyzed by RlmN and Cfr (34,

35). The authors propose a mechanism involving initial transfer of a methyl group from SAM to

some atom on a precursor to the M-cluster, followed by abstraction of at least one H• by a 5’-dA•

generated via reductive cleavage of another molecule of SAM (60). It appears that ping-pong

mechanisms for RS methylation reactions may be relatively common.

69

Figure 2-12. Working hypothesis for the reaction catalyzed by Tm RimO. Step 1: transfer of a

methyl group from SAM bound to the RS [4Fe–4S] cluster to the external sulfur ion of a

polysulfide group attached to the unique iron ion of the auxiliary [4Fe–4S] cluster. Step 2:

Reductive fragmentation of a second molecule of SAM bound to the RS [4Fe–4S] cluster to a 5’-

dA• and abstraction of a H• from bound substrate. Step 3: Attack of a substrate radical onto the

methylated sulfur atom of the polysulfide chain to afford the methylthiolated product and a [4Fe–

4S]2+ cluster with a terminal persulfide.

70

References

1. Atta M, Mulliez E, Arragain S, et al. Curr. Opin. Struct. Biol. 2010: 20, 1–9

2. Anton BP, Russell SP, Vertrees J, et al. Nucleic Acids Res. 2010: 38, 6195–6205

3. Arragain S, García-Serres R, Blondin G, et al. J. Biol. Chem. 2010: 285, 5792–5801

4. Booker SJ, Cicchillo RM, Grove TL. Curr. Opin. Chem. Biol. 2007: 11, 543-552

5. Challand MR, Driesener RC, Roach PL. Nat. Prod. Rep. 2011: 28, 1696–1721

6. Frey PA, Hegeman AD, Ruzicka FJ. Crit. Rev. Biochem. Mol. Biol. 2008: 43, 63–88

7. Cosper MM, Jameson GNL, Hernández HL, et al. Biochemistry. 2004: 43, 2007-2021

8. Ugulava NB, Gibney BR, Jarrett JT. Biochemistry. 2001: 40, 8343-8351

9. Ugulava NB, Surerus KK, Jarrett JT. J. Am. Chem. Soc. 2002: 124, 9050-9051

10. Hernández HL, Pierrel F, Elleingand E, et al. Biochemistry. 2007: 46, 5140-5147

11. Cicchillo RM, Lee K-H, Baleanu-Gogonea C, et al. Biochemistry. 2004: 43, 11770-11781

12. Lee K-H, Saleh L, Anton BP, et al. Biochemistry. 2009: 48, 10162–10174

13. Walsby CJ, Ortillo D, Yang J, et al. Inorg. Chem. 2005: 44, 727-741

14. Vey JL, Drennan CL. Chem. Rev. 2011: 111, 2487–2506

15. Fugate CJ, Stich TA, Kim EG, et al. J. Am. Chem. Soc. 2012: 134, 9042–9045


17. Bartz JK, Kline LK, Soll D. Biochem. Biophys. Res. Commun. 1970: 40, 1481-1487

18. Caillet J, Droogmans L. J. Bacteriol. 1988: 170, 4147-4152

19. Rosenbaum N, Gefter ML. J. Biol. Chem. 1972: 247, 5675-5680

20. Elkins BN, Keller EB. Biochemistry. 1974: 13, 4622-4628

21. Deutsch C, El Yacoubi B, de Crecy-Lagard V, Iwata-Reuyl D. J. Biol. Chem. 2012: 287,

13666–13673

22. Connolly DM, Winkler ME. J. Bacteriol. 1989: 171, 3233-3246

23. Connolly DM, Winkler ME. J. Bacteriol. 1991: 173, 1711-1721

24. Esberg B, Björk GR. J. Bacteriol. 1995: 177, 1967-1975

25. Urbonavicius J, Qian Q, Durand JMB, et al. EMBO J. 2001: 20, 4863–4873

26. Dehwah MA, Wang M, Huang QY. Genet. Mol. Res. 2010: 9, 1109-1120

27. Saxena R, Voight BF, Lyssenko V, et al. Science. 2007: 316, 1331-1336

28. Scott LJ, Mohlke KL, Bonnycastle LL, et al. Science. 2007: 316, 1341-1345

29. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. Nat. Genet. 2007: 39, 770-775

30. Zeggini E, Weedon MN, Lindgren CM, et al. Science. 2007: 316, 1336-1341

31. Anton BP, Saleh L, Benner JS, et al. Proc. Natl. Acad. Sci. USA. 2008: 105, 1826–1831

32. Carr JF, Hamburg DM, Gregory ST, et al. J. Bacteriol. 2006: 188, 2020-2023

33. Yan F, LaMarre JM, Röhrich R, et al. J. Am. Chem. Soc. 2010: 132, 3953-3964

34. Grove TL, Benner JS, Radle MI, et al. Science. 2011: 332, 604–607

35. Yan F, Fujimori DG. Proc. Natl. Acad. Sci. U S A. 2011: 108, 3930–3934

36. Boal AK, Grove TL, McLaughlin MI, et al. Science. 2011: 332, 1089–1092

37. Grove TL, Radle MI, Krebs C, Booker SJ. J. Am. Chem. Soc. 2011: 133, 19586–19589

38. McCusker KP, Medzihradszky KF, Shiver AL, et al. J. Am. Chem. Soc. 2012: 134,

18074-18081

39. Iwig DF, Booker SJ. Biochemistry. 2004: 43, 13496-13509

40. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular Cloning: A Laboratory Manual.

Plainview, New York: Cold Spring Harbor Laboratory Press

41. Cicchillo RM, Iwig DF, Jones AD, et al. Biochemistry. 2004: 43, 6378-6386

42. Lanz ND, Grove TL, Gogonea CB, et al. Methods Enzymol. 2012: 516, 125-152

43. Bradford M. Anal. Biochem. 1976: 72, 248-254

71

44. Grove TL, Lee KH, St Clair J, et al. Biochemistry. 2008: 47, 7523–7538

45. Beinert H. Methods Enzymol. 1978: 54, 435-445

46. Beinert H. Anal. Biochem. 1983: 131, 373-378

47. Kennedy MC, Kent TA, Emptage M, et al. J Biol Chem. 1984: 259, 14463-14471

48. Soderberg T, Poulter CD. Biochemistry. 2000: 39, 6546–6553

49. Crain PF. Methods Enzymol. 1990: 193, 782–790

50. Ghehrke CW, Kuo KC, McCune RA, et al. J. Chromatogr. 1982: 230, 297–308

51. Johnson DC, Unciuleac MC, Dean DR. J. Bacteriol. 2006: 188, 7551–7561

52. Pierrel F, Douki T, Fontecave M, Atta M. J. Biol. Chem. 2004: 279, 47555-47653

53. Soderberg T, Poulter CD. Biochemistry. 2001: 40, 1734–1740

54. Agris PF, Armstrong DJ, Schafer KP, Soll D. Nucleic Acids Res. 1975: 2, 691-698

55. Hänzelmann P, Schindelin H. Proc. Natl. Acad. Sci. USA. 2004: 101, 12870-12875

56. Goldman PJ, Grove TL, Sites LA, et al. Proceedings of the National Academy of

Sciences. 2013: 110, 8519-8524

57. Spatzal T, Aksoyoglu M, Zhang L, et al. Science. 2011: 334, 940

58. Einsle O, Tezcan FA, Andrade SLA, et al. Science. 2002: 297, 1696–1700

59. Lancaster KM, Roemelt M, Ettenhuber P, et al. Science. 2011: 334, 974–977

60. Wiig JA, Hu Y, Lee CC, Ribbe MW. Science. 2012: 337, 1672–1675

Chapter 3

Characterization of RimO from the mesophilic gut bacterium Bacteroides

thetaiotaomicron

Introduction

It is estimated that the human body contains ~ 1014

cells, of which only 10% are its own,

with the remainder belonging to symbiotic prokaryotes (1). The vast majority of these bacteria

reside in the human intestinal tract (gut microbiota) (2), and en masse can be thought of as an

active organ carrying out complex biochemical reactions to aid in digestion of foods that would

otherwise be indigestible (3), especially the plant-derived carbohydrates amylose, amylopectin,

pullulan, and maltooligosaccharides (4). These organisms also aid in absorption of important fatty

acids, vitamins, and minerals from food (5). In addition to their nutritional benefits, gut

microbiota play a role in regulating immune homeostasis not just in the gastrointestinal tract (GI),

but systemically as well (6). Symbiotic bacteria play crucial roles in resistance to colonization by

their opportunistic, pathogenic counterparts by successfully evading the host immune response

and forming a barrier in the gut mucosal lining (7). Disruption of their homeostasis, through the

administration of antibiotics, vaccinations, hygiene, and diet alterations, can lead to colonization

by opportunistic bacteria, such as Helicobacter pylori (Crohn's disease), Clostridium difficile

(colitis), Salmonella enterica (salmonellosis), Vibrio cholerae (cholera), certain strains of

Escherichia coli (dysentery), Corynebacterium diphtheriae (diphtheria), and Listeria

monocytogenes (listeriosis and meningitis) to name a few (6, 8). Additionally, homeostasis

disruption can lead to "leaky gut syndrome" wherein symbionts can be translocated to other sites

of the body causing peritonitis, septicemia, and meningitis (8). The two most common species

73

that cause infections resulting from bacterial translocation from the mucosal lining are

Bacteroides fragilis (Bf) and Bacteroides thetaiotaomicron (Bt), likely due to their abundance

(109

to 1010

per gram of dry feces) (9). In fact, the predominant genus found in the human GI is

Bacteroides, accounting for ~ 30% of all fecal isolates (9).

The first genome from the genus Bacteroides to be sequenced was from Bt, and, as a

result, this species has become one of the most widely studied (10). The Bt proteome contains the

highest number of glycosylhydrolases‒enzymes used to break down the plant-derived

polysaccharides mentioned above‒of all sequenced enteric bacteria to date (10), making this

species, and others from this genus found in the GI, major contributors in supplying 10-15% of

our total daily calories through their fermentation of ingested dietary plant matter (11).

Additionally, Bt contains several types of mobile genetic elements, allowing it to contribute to

horizontal gene transfer within its genus and to others present in the gut (10). Indeed, over the

past several decades, Bacteroides clinical isolates have shown resistance to the antibiotics

tetracycline and clindamycin (12), and this genus has developed the highest resistance rates to

antimicrobial agents of all anaerobic bacteria (13). This resistance is worrisome due to the

predominance of this genus in the GI, where conditions are conducive for horizontal gene transfer

events that could lead to more bacteria developing antibiotic resistance (13).

Given the recent interest in studying the links between human health and gut microbiota,

along with our desire to study radical SAM enzymes from some of the mesophilic anaerobes

found in the GI, we chose to study the methylthiotransferase enzyme, RimO, from Bacteroides

thetaiotaomicron. RimO catalyzes the synthesis and subsequent transfer of a methylthio- group to

C3 of asparate 89 (D89) found on the ribosomal protein S12 (14-18). While D89 is absolutely

conserved among its homologues, the post-translational modification, which occurs only in

bacteria, is not, and in fact, is non-essential (17). Methylthiolation of S12 D89 is believed to

maintain translational fidelity, given that this residue resides on a loop that projects into the

74

acceptor site where tRNAs bind to the ribosome (19, 20). RimOs from E. coli and T. maritima

have been characterized (16, 18), and mechanistic studies have unraveled some of the details

concerning the methylthiolation reaction (14, 15); however, some issues have precluded more in-

depth study of this reaction. Herein, we describe the characterization of RimO from Bacteroides

thetaiotaomicron, demonstrate its activity with the flavodoxin/flavodoxin

oxidoreductase/NADPH reducing system from E. coli and compare rates of formation with this

reducing system with those observed in the presence of the chemical reductant sodium dithionite.

We also show that Bt RimO does not harbor additional sulfide above that required to form its 2

[4Fe-4S] clusters, is limited to one turnover without exogenous sources of sulfide, and likely uses

one electron per equivalent of methylthiolated formed. A variant of Bt RimO, in which a

conserved tyrosine was substituted with phenylalanine (Y225F), was isolated to compare its

activity to that of the cognate variant (Y227F) of Tm RimO, which was previously shown to

catalyze formation of 1 equivalent of 5'-dAH but no methylthiolated product (see Chapter 5).


Materials


MA). Sodium sulfide (nonahydrate), L-tryptophan, 2-mercaptoethanol, L-(+)-arabinose, ferric

chloride, sodium methanethiolate, 5’-deoxyadenosine (5’-dA), and S-adenosyl-L-homocysteine

(SAH) were purchased from Sigma Corp (St. Louis, MO). N-(2-hydroxyethyl)piperizine-N'-(2-

ethanesulfonic acid) (HEPES) was purchased from Fisher Scientific (Pittsburgh, PA), and

imidazole was purchased from J. T. Baker Chemical Co. (Phillipsburg, NJ). Potassium chloride,

glycerol, and expression vector pET-28a were purchased from EMD Chemicals (Gibbstown, NJ),

75

while dithiothreitol (DTT) and nickel nitrilotriacetic acid (Ni-NTA) resin were purchased from

Gold Biotechnology (St. Louis, MO). Coomassie blue dye-binding reagent for protein

concentration determination was purchased from Pierce (Rockford, IL), as was the bovine serum

albumin standard (2 mg/mL). 5-((2-[(iodoacetyl)amino]ethyl)amino)naphthalene-1-sulfonic acid

(I-AEDANS) was obtained from Life Technologies (Carlsbad, CA). Nick, NAP-10, and PD-10

pre-poured gel-filtration columns, as well as Sephadex G-25 resin were purchased from GE

Biosciences (Piscataway, NJ). All other buffers and chemicals were of the highest grade

available.

Methods

Cloning and overexpression of the Bt rimO gene

A plasmid (pSGC-His) encoding an N-terminally hexahistidine-tagged form of Bt RimO

containing a 15 amino acid linker (SSGVDLGTENLYFQS) was a generous gift from Dr. Steven

Almo at the Albert Einstein College of Medicine. The plasmid DNA was extracted and purified

from E. coli DH5α cells using a NucleoSpin Plasmid kit (Macherey-Nagel, Düren, Germany), and

its sequence was verified before its transformation into Ec BL21 (DE3) cells containing the

pDB1282 plasmid as previously described (15, 21, 22)

Bacterial growth and gene expression was carried out at 37 °C in 16 L of M9 minimal

media distributed evenly among 4 Erlenmeyer flasks with moderate shaking (180 rpm). At an

optical density (OD) at 600 nm of 0.3, solid L-(+)-arabinose was added to each flask to a final

concentration of 0.2 % (w/v), while cysteine and ferric chloride were added to final

concentrations of 150 µM and 25 µM, respectively. At an OD600 of 0.6, the flasks were placed on

ice with intermittent shaking for 1 h, and then a sterile solution of IPTG was added to each flask

76

to a final concentration of 200 µM. Cysteine and ferric chloride were added again to final

concentrations of 300 µM and 50 µM, respectively, and expression was allowed to take place for

16 h at 18 °C before the cells were harvested by centrifugation at 10,000 g for 10 min at 4°C.

Purification of Bt RimO

Purification of Bt RimO was carried out by immobilized metal affinity chromatography

(IMAC) using Ni-NTA resin. All purification steps were performed in a Coy (Grass Lakes, MI)

anaerobic chamber (unless specifically stated otherwise), which was kept under an atmosphere of

N2 and H2 (95%/5%). The O2 concentration was maintained below 1 ppm by using palladium

catalysts. Buffers used during the purification of Bt RimO were as follows: lysis buffer (50 mM

HEPES, pH 7.5, 300 mM KCl, 10 mM 2-mercaptoethanol, 10 mM imidazole, and 1 mg/mL

lysozyme); wash buffer (50 mM HEPES, pH 7.5, 300 mM KCl, 10 mM 2-mercaptoethanol, 10%

(v/v) glycerol, and 20 mM imidazole); elution buffer (wash buffer containing 250 mM

imidazole). After lysing the cells by sonication (23), the cell suspension was transferred into

sterile centrifuge tubes and centrifuged at 50,000 g and 4 °C for 1 h. The supernatant was

loaded onto a Ni-NTA column, which was subsequently washed with 200 mL of wash buffer.

After addition of elution buffer to the column, fractions containing RimO, distinguished by their

dark brown color, were pooled and concentrated using an Amicon stirred ultrafiltration apparatus

(Millipore, Billerica, MA) fitted with a YM-30 membrane (30,000 molecular weight cutoff). The

protein was exchanged into gel-filtration buffer (GFB) (50 mM HEPES, pH 7.5, 300 mM KCl,

20% glycerol, and 1 mM DTT) using a Sephadex G-25 column (2.5 13 cm), reconcentrated,

and stored in aliquots in a liquid N2 dewar until ready for use.

77

Construction, overexpression, and purification of the Y225F variant of Bt RimO

The gene for the Bt RimO Y225F variant was constructed using the Stratagene

QuikChange II site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) according

to the manufacturer’s specifications, and as described previously (24). The following were used

as forward and reverse primers, respectively, with base changes underlined: 5′-

GAGTGGATTCGTCTGCATTTTGCGTATCCGGCAC-3′ and 5'-

GTGCCGGATACGCAAAATGCAGACGAATCCACTC-3'. These primers were added to a

typical QuikChange II reaction mixture to a final concentration of 20 µM with 100 ng of pSC-His

Bt RimO template DNA. 15 cycles of the following program were initiated: 95 °C for 1 min, 55

°C for 1 min, and 68 °C for 10 min. Upon completion of the cycling program, the reaction

mixture was incubated for 15 min at 68 °C before being cooled to 4 °C. Subsequent to this step,

the procedure followed the manufacturer’s specifications. The correct mutation was verified by

DNA sequencing, and the resulting plasmid was designated pBtRimOY225F. Transformation of

pBtRimOY225F into Ec BL21(DE3) cells and overexpression and purification of the Bt RimO

Y225F gene product was conducted as described above for Bt RimO wild type.

Protein, Iron, and Sulfide Quantification

The concentrations of Bt RimO were determined by the procedure of Bradford (25) using

bovine serum albumin (Fraction V) as a standard. Quantitative amino acid analysis, conducted as

described previously (26), indicates that the procedure of Bradford overestimates the

concentration of Bt RimO by a factor of 1.38. Iron and sulfide analyses were performed

according to the procedures of Beinert (27-29).

78

Chemical Reconstitution of Bt RimO

Quantitative iron analysis on as-isolated (AI) Bt RimO established the number of iron

ions present in the enzyme. The protein was diluted to 100 µM in GFB on ice and treated with 1

mM DTT before being incubated for 10 min with an amount of FeCl3 equal to the difference

between 8 equiv and the number of equiv of Fe ions present in AI Bt RimO, plus two (e.g.

quantitative iron analysis determines that AI Bt RimO contains 5 equiv, so 5 equiv of FeCl3 are

added). The same number of equiv of sodium sulfide as FeCl3 was added in 8 increments over 2

h, upon which the reconstitution mixture was incubated overnight on ice. The next day, the

mixture was concentrated by ultrafiltration using an Amicon stirred ultrafiltration apparatus fitted

with a YM-10 membrane. Following concentration, the mixture was loaded onto a pre-poured

PD-10 column equilibrated in GFB to remove excess Fe and sulfide, further concentrated, and

centrifuged at 18,000 g for 1 min prior to purifying it further by molecular sieve

chromatography using an ÄKTA FPLC system with a HiPrep 26/60 Sephacryl S-200 size-

exclusion column (GE Healthcare Piscataway, NJ) housed in an anaerobic chamber. The column

was equilibrated in buffer consisting of 50 mM Na-HEPES, pH 7.5, 300 mM KCl, 5 mM DTT,

and 10% glycerol. Fractions were pooled based on absorbances at 280 and 400 nm and

concentrated and stored as described above.

Determination of the oligomeric state of Bt RimO

Purified and reconstituted Bt RimO was diluted to 130 µM in a final volume of 500 µL

100 mM HEPES, pH 7.5, 10 % glycerol, and 500 mM KCl (high salt buffer) or 50 mM KCl (low

salt buffer). Alternatively, a sample was prepared in low salt buffer containing 500 µM of

racemic SAM (R and S diastereomers of the chiral sulfur atom, hereafter referred to as RS-SAM),

79

which was a generous gift from Nicholas Lanz (30). Molecular sieve chromatography was

performed using an ÄKTA FPLC system with a HiPrep 16/60 Sephacryl S-200 size-exclusion

column (GE Healthcare Piscataway, NJ) equilibrated in the same buffer with which the samples

were prepared. The flow rate was maintained at 0.5 mL·min-1

, and the absorbance was monitored

at 280 and 400 nm. A standard curve was generated from a molecular weight markers kit (Sigma,

St. Louis, MO) consisting of proteins of known molecular masses: Cytochrome c from horse

heart (12.4 kDa), carbonic anhydrase from bovine erythrocytes (29 kDa), bovine serum albumin

(66 kDa), alcohol dehydrogenase from yeast (150 kDa), and β-amylase from sweet potato (200

kDa). Blue dextran (2,000 kDa) was used for the determination of the column void volume (V0).

The elution volumes (Ve) of the standards were obtained, and the ratios of Ve/V0 were plotted as a

function of the log of their respective molecular masses. The standard curve was then used to

calculate the apparent molecular mass of Bt RimO in high salt, low salt, and low salt + RS-SAM

conditions from their corresponding elution volumes.

EPR characterization of the Fe/S clusters of Bt RimO

Electron paramagnetic resonance (EPR) samples of Bt RimO were prepared anaerobically

in a total volume of 250 µL and contained, where appropriate, 400 µM Bt RimO RCN, 50 mM

Na-HEPES pH 7.5, 2 mM sodium dithionite, 2 mM 13-mer S12 peptide, and 2 mM SAM.

Samples were transferred to 4 mM O.D. thin quartz EPR tubes (Wilmad Labglass, Vineland, NJ)

and then flash frozen in semi-frozen isopentane and removed from the anaerobic glovebox. Each

sample tube was wiped clean of isopentane and placed in liquid N2 until analysis. EPR spectra

were obtained at 12 K on a Bruker (Billerica, MA) ESP 300 spectrometer as previously described

(24).

80

Bt RimO Activity Assays

Bt RimO reactions contained the following in a final volume of 220 µL unless otherwise

specified: 100 µM Bt RimO, 1 mM SAM, 1 mM S12 peptide substrate (1), 50 mM Na-HEPES,

pH 7.5, and 2 mM dithionite or the Ec flavodoxin:NADPH-flavodoxin oxidoreductase reducing

system (50 µM flavodoxin (Fld), 25 uM flavodoxin oxidoreductase (Fdr), 1 mM NADPH). Fld

and Fdr were overexpressed and purified as previously described (26). All components except

SAM were incubated at 37 °C for 3 min before initiating the reaction with the omitted

component. Aliquots (15 µL) of the reaction mixture were withdrawn at various times from 0-180

min and added to 20 µL of 0.1 M H2SO4 containing 50 µM peptide 2 (ES) to quench the reaction.

The quenched samples were neutralized with 15 µL of 0.5 M ammonium acetate, pH 6.0.

Precipitated protein was removed by centrifugation at 18,000 g for 15 min, and a 35 µL aliquot

of the resulting supernatant was analyzed by ESI+ LC/MS with single-ion monitoring (SIM).

Solvent A consisted of ammonium acetate (40 mM) and methanol (5% v/v) titrated to pH 6.0

with acetic acid, while solvent B was 100 % acetonitrile. The column was equilibrated in 100%

solvent A at a flow rate of 0.5 mL min-1

. After sample injection (2 µL), a gradient was applied

from 0% solvent B to 50% solvent B over 5 min and then 50% to 0% over 2 min. The monitored

ions (m/z) and retention times (min), respectively, were 385.1 and 3.8 (SAH), 188.0 and 4.0

(tryptophan), 252.1 and 4.1 (5'-dA), 474.4 and 4.2 (2, ES), 498.1 and 4.4 (peptide 1, 3+ charge

state), and 507.1 and 4.4 (MS-1, 3+ charge state). Calibration curves were generated with known

concentrations of each analyte and run under identical conditions to determine the concentration

of products generated in assays. Data were analyzed using the Agilent Technologies MassHunter

qualitative and quantitative analysis software.

81

Determination of Persulfide Content of Bt RimO by Fluorescent Labeling

For detection of any persulfide groups on Bt RimO, a previously described method was

used with some modifications (31). Monomeric and dimeric fractions of Bt RimO RCN isolated

by size-exclusion chromatography were exchanged into buffer A (50 mM Na-HEPES, 20 mM

MgCl2, pH 8.0) to remove DTT using a pre-poured PD-10 column. The eluate was concentrated

using Microcon YM-10 centrifugal filters (Millipore, Billerica, MA) and the concentration of the

retentate was determined by the method of Bradford (25). Samples were made in triplicate,

starting with incubation of 10 µM Bt RimO in buffer A with 1 mM Na2S in a total volume of 100

µL at 37 °C for 30 min to allow for formation of persulfides. Control samples made with

monomeric Bt RimO RCN were treated in exactly the same manner with the omission of Na2S.

After incubation, each sample was concentrated using centrifugal filters and washed four times

with 350 µL of buffer A to remove excess Na2S. The samples were concentrated to a final

volume of ~ 90 µL. 1,5-I-AEDANS dissolved in buffer A was added to each sample to a final

concentration of ~ 1 mM and allowed to react at 37 °C for 30 min to derivatize thiol groups in the

protein (32). Where noted, guanidinium hydrochloride was added to a final concentration of 1 or

4 M. The samples were loaded into centrifugal filter units, washed four times with buffer A to

remove any unreacted 1,5-I-AEDANS, and concentrated to ~ 50 µL total volume. 50 µL of 10

mM DTT was then added to each sample, and the samples were incubated at 37 °C for 30 min to

release any persulfide-bound dye. Each sample was loaded into centrifugal filters and centrifuged

to dryness. To each sample filtrate, 900 µL of buffer A was added, and the fluorescence of the

diluted filtrate was quantified using a Cary Eclipse Fluorescence Spectrophotometer (Varian,

Walnut Creek, CA) with an excitation wavelength of 337 nm and an emission wavelength of 498

nm. Standard curves were generated with 0.2 to 50 µM of 1,5-I-AEDANS in 100 µL of buffer A

containing 5 mM DTT and diluted with 900 µL of buffer A.

82

Quantification of flavodoxin semiquinone with Bt RimO under turnover conditions

The one electron reduced Ec flavodoxin semiquinone (Fld SQ) was generated in a Coy

(Grass Lakes, MI) anaerobic chamber as described above by incubating Ec flavodoxin (Fld) with

0.55 equivalents of sodium dithionite for 2 h at 37 °C in buffer consisting of 50 mM Na-HEPES,

pH 7.5, and 200 mM KCl. Formation of Fld SQ was apparent by the observed change in color

from yellow-orange to gray-blue. Dithionite was removed from the Fld SQ reaction mixture by

gel-filtration with a pre-poured NAP-10 column equilibrated in the same buffer and the purple-

blue eluate was concentrated using Microcon YM-10 centrifugal filter units. Fld SQ was placed

in an eppendorf tube, sealed in a 50 mL conical tube, and subsequently transferred to an MBraun

(Peabody, MA) anaerobic glove box maintained at < 0.1 ppm with palladium catalysts. The

concentration of Fld SQ was determined spectrophometrically (ε = 4570 M-1

cm-1

at 579 nm)(33)

using an Agilent 8453 UV-Visible spectrophotometer (Agilent Technologies, Santa Clara, CA)

housed in the anaerobic glovebox. Bt RimO RCN was pre-methylated by incubating with SAM

for 2 h and gel-filtered to remove any unreacted SAM and weakly bound SAH. An 850 µL

reaction mixture consisting of 72 µM pre-methylated Bt RimO RCN, 100 mM Na-HEPES, pH

7.5, 200 mM KCl, 400 µM [methyl-d3]SAM, 800 µM 13 mer Bt S12 peptide, and 1 mM

tryptophan (internal standard) was used to blank the spectrophotometer at 579 nm prior to

initiating the reaction by adding 50 uL of Fld SQ to a final concentration of 50 µM (dilution of all

components by addition of Fld SQ was accounted for). The absorbance at 579 nm was monitored

and recorded at specific times. An identical reaction in 150 µL total volume was run in tandem

from which 15 µL aliquots at specific times were removed and quenched with a mixture of 0.1 M

H2SO4 and an appropriate external standard to quantify SAH, 5'-dAH, and the methylthiolated

product by LC/MS as described above.

83

Results

Cloning and overexpression of the Bt rimO gene

A plasmid (pSGC-His) encoding an N-terminal hexahistidine-tagged form of Bt RimO

(pSC-His-BtRimO) was used to transform Ec BL21 (DE3) cells containing the pDB1282

plasmid, which encodes an arabinose inducible isc operon from Azotobacter Vinelandii for

expression of iron-sulfur cluster machinery proteins (34, 35). Induction of expression of pSC-

His-BtRimO with 200 µM IPTG in M9 minimal media at 18 °C subsequent to a 1 h cold shock

on ice resulted in the greatest protein yield (Figure 3-1).

Figure 3-1. SDS-PAGE analysis of Bt RimO overexpression. Lanes 1 and 16: molecular weight

markers (in kDa); lanes 2-5: cell culture samples prior to induction of the pDB1282 plasmid with

arabinose; lanes 7-10: cell culture samples after induction with arabinose; lanes 12-15: cell

culture samples after induction of the pSC-His-Bt RimO plasmid for 16 h at 18°C.

Purification of the Bt RimO gene product was conducted anaerobically by IMAC using

Ni-NTA resin with lower concentrations of imidazole than is typically used to ensure protein

binding to the resin. Protein yields ranged from 15 - 20 mg / L of cell culture. The two 4Fe-4S

clusters of Bt RimO were then reconstituted with FeCl3 and Na2S in the presence of the reductant

DTT, and the protein was subsequently purified by size-exclusion chromatography (Figure 3-2)

84

to remove protein aggregates resulting from reconstitution and to increase the homogeneity of the

enzyme as evidenced by SDS-PAGE. (Figure 3-3). Typical protein yields following

reconstitution and size-exclusion chromatography were ~ 10 mg/L of cell culture.

Figure 3-2. SDS-PAGE of Bt RimO purification. Lane 1: molecular weight markers (in kDa);

lane 2: insoluble fraction of cell lysate; lane 3: soluble fraction of cell lysate; lane 4: flow through

from Ni-NTA resin; lane 5: wash through of Ni-NTA resin; lanes 6-8: 2, 5, and 10 µL samples of

the eluate from the Ni-NTA resin.

85

Figure 3-3. HR 26/60 Sephacryl S200 elution profile of Bt RimO. Wavelengths monitored were

280 (blue) and 400 nm (red), corresponding to wavelengths at which protein and Fe-S clusters,

respectively, absorb maximally. Fractions 6-18, corresponding to the second peak centered at 130

mL, were pooled and concentrated, as were fractions 19-29 from the third peak centered at 160

mL.

86

Figure 3-4. SDS-PAGE analysis of reconstituted and S200-purified Bt RimO. Lane 1: molecular

weight markers (in kDa); lanes 2-4: pooled fractions 19-29 from the S200 column with the

volume of sample loaded indicated; lanes 6-8: pooled fractions 6-18.

Analysis of Fe/S cluster content by quantitative Fe and S analyses and EPR spectroscopy

To determine the extent to which both AI and RCN Bt RimO ligated iron and sulfide,

which is indicative of the number of Fe/S clusters present, quantitative iron and sulfide analyses

of Bt RimO were performed. Analyses for the iron and sulfide content of AI enzyme yielded 3.7

+ 0.7 of the former and 4.9 + 1.2 of the latter (average and standard deviation of three

independent determinations). Determination of the iron and sulfide content of RCN Bt RimO

resulted in 7.3 + 0.9 of the former and 8.6 + 1.3 of latter (average and standard deviation of seven

independent determinations). Like RimOs from Tm and Ec, which both harbor two 4Fe-4S

clusters (16, 18) and share 35% and 39% sequence identity with Bt RimO, respectively, it is

highly likely that Bt RimO also contains two 4Fe-4S clusters. The results of quantitative iron and

sulfide analyses indicated that the AI enzyme is purified with less than two full 4Fe-4S clusters;

however, reconstitution with FeCl3 and Na2S under reducing conditions and subsequent

purification by anaerobic size-exclusion chromatography resulted in Bt RimO containing ~ 8 iron

87

ions and ~ 9 sulfide ions, which is consistent with the number of iron and sulfide ions expected if

two 4Fe-4S clusters are present.

To definitively show that Bt RimO ligates two 4Fe-4S clusters, EPR analysis of Bt RimO

RCN was conducted. The spectrum of a dithionite reduced sample of Bt RimO RCN (Figure 3-

5A) recorded at 12 K exhibits features at g = 2.06 and 1.93, which are typical of a [4Fe-4S]+

cluster with an S = 1/2 ground state and consistent with g values observed with RimOs from Ec

and Tm (16, 18). The additional feature at g = 2.03 was not previously observed; however, this

feature can be attributed to the presence of co-purified S-adenosylhomocysteine (SAH) bound to

or near, presumably, the radical SAM [4Fe-4S]+ cluster, since acid-denaturation of the protein

and subsequent LC/MS analysis identified the presence of SAH. The addition of SAM to Bt

RimO for 1 hr (Figure 3-5B) prior to reduction of the enzyme with dithionite alters the spectrum

slightly with broadening of the feature at g = 2.03 such that the signal at g = 2.06 is obscured. The

feature at g = 1.91 is also broadened in the presence of SAM. The perturbations of the EPR

spectra of dithionite-reduced Bt RimO in the presence of SAM are consistent with its binding to

or near a [4Fe-4S]+ cluster, which has been observed for Ec and Tm RimOs (16, 18). Notably,

signal intensities of samples containing SAM were ~ 50% that of samples in which these

components were omitted, and spin quantification of these samples confirmed that the sample

containing SAM had 0.19 equiv of spin versus 0.33 for those lacking SAM, which suggests that

SAM binding to the radical SAM cluster may decrease its redox potential. The EPR spectrum of

Bt RimO in the presence of the 13 mer peptide substrate and dithionite (Figure 3-5C) exhibits an

EPR spectrum that is essentially identical to that of dithionite reduced Bt RimO, indicating that

the substrate does not perturb the electronic properties of either cluster. Collectively, these results

suggest that Bt RimO ligates two 4Fe-4S clusters with EPR signal parameters closely matching

those of others characterized RimOs (16, 18).

88

Figure 3-5. EPR spectra of 400 µM Bt RimO RCN reduced with 2 mM dithionite (A) in the

presence of 2 mM SAM for 1 h then reduced (B); in the presence of 13 mer peptide substrate for

1 h then reduced (C). Spectra were collected at 12 K with a microwave power of 0.1 mW, a

microwave frequency of 9.48 GHz, and a modulation amplitude of 10 G.

89

Determination of the oligomeric state of Bt RimO

The apparent molecular mass of Bt RimO RCN was determined under several conditions

by analytical molecular-sieve chromatography to gain insight on the oligomeric state of the

enzyme. Blue dextran (2,000 kDa) was used to determine the void volume (V0) of the S200

column, and a mixture of proteins with known molecular masses was used as standards (Figure

3-6A). A standard curve was generated by plotting the ratio of the elution volume (Ve) of each

protein standard to the void volume (i.e. Ve/V0) versus the log of the protein molecular weight

(Figure 3-6B). Alcohol dehydrogenase was omitted from the standard curve plot due to its

ambiguous elution volume. The elution volumes obtained for Bt RimO RCN under various

conditions (Figure 3-6C) were used in the ratio Ve/V0 to calculate the apparent molecular mass of

the enzyme under each condition.

90

Figure 3-6. Molecular-sieve chromatographic analysis of Bt RimO RCN. Protein standards of

known molecular weight were loaded on an S200 column to determine their elution volumes (A)

in order to generate a standard curve (B). Bt RimO RCN was analyzed by molecular-sieve

chromatography under the following conditions: high salt (blue trace), low salt (black trace), low

salt + 500 µM RS-SAM (red trace), and the determined elution volumes of the enzyme in each

condition was used to calculate its apparent molecular mass. Absorbances were measured at 280

nm and 400 nm (only 280 nm shown for clarity). The peak at 17 mL results from an unknown

contaminant eluting from the column.

91

Bt RimO RCN samples were analyzed in buffer containing high salt (500 mM KCl), low

salt (50 mM KCl), or low salt + 500 µM RS-SAM. In each condition, two protein fractions eluted

from the S200 column, indicating that two oligomeric states of the enzyme were likely present

(Figure 3-6C). Under conditions of low salt in the absence or presence of RS-SAM, the first

fraction of Bt RimO eluted at 44.1 mL (Figure 3-6C, black and red traces), while the first fraction

of Bt RimO in high salt buffer eluted slightly later at 45.5 mL (Figure 3-6C, blue trace). The

calculated apparent molecular mass of this first protein fraction in low salt with or without SAM

was 154.5 kDa, while that in high salt was 134.8 kDa. These apparent masses are consistent with

this fraction of Bt RimO existing has a homotrimer (theoretical mass of 160.2 kDa) or homodimer

(theoretical mass of 106.8 kDa), given the molecular weight of Bt RimO with its hexahistidine tag

and 15 amino acid linker is 53.4 kDa.

The second protein fraction in low salt and the absence or presence of SAM eluted at

volumes of 53.1 and 54.3 mL, which resulted in apparent molecular masses of 66.7 and 60.6 kDa,

respectively. A slight shift in elution volume (56.3 mL) was observed for this fraction in high

salt, yielding a molecular mass of 51.4 kDa. The masses calculated for this second protein

fraction are consistent with the theoretical mass of Bt RimO and likely correspond to its

monomeric form. Curiously, while the exact same amount of Bt RimO in each buffer condition

was loaded on the S200 column, the absorbance of the protein fractions varied; however, the ratio

of the maximum absorbance at 280 nm of the first and second protein fractions were consistent in

each condition—the absorbance of the first fraction was 1.5 to 1.7-fold greater than the second

fraction—indicating that the conditions tested did not significantly perturb the population of

either oligomeric state of Bt RimO RCN.

92

Determination of Bt RimO WT and Y225F activity with dithionite or the Ec flavodoxin

reducing system

The activity of Bt RimO WT and Bt RimO Y225F was determined with a 13 amino acid

peptide substrate (1)—corresponding to residues 83-95 of the Bt S12 protein—with the chemical

reductant sodium dithionite or the in vivo flavodoxin/flavodoxin oxidoreductase/NADPH

(Fld/FldR/NADPH) reducing system from E. coli. The Y225F variant was tested to compare its

activity to the analagous variant (Y227F) in RimO from Tm, which formed 1 equivalent of 5'-

dAH but no methylthiolated product (see Chapter 5). ESI+ LC/MS analysis of reactions of Bt

RimO in the presence of SAM, 1, and dithionite showed time-dependent formation of 5'-dAH ,

SAH, and methylthiolated peptide (MS-1) with m/z values of 252.1, 385.1, and 507.1,

respectively (Figure 3-7B). Gratifyingly, reactions containing Bt RimO, SAM, 1, and the Ec

Fld/FldR/NADPH reducing system resulted in robust formation of the same products observed in

reactions when dithionite was used as the requisite source of electrons (Figure 3-7A). The time-

dependent formation of 5'-dAH, SAH, and MS-1 in reactions containing the in vitro or in vivo

reducing system was each fitted to a first-order exponential equation shown in equation 1 with

their respective kinetic parameters reported in Table 3-1, wherein A is the amplitude, k is the rate

constant, and t is time. In addition, the initial rates of formation of these reaction products were

obtained from the slopes resulting from fitting the linear portion of the data (Figure 3-7 C & D

and Table 3-1).

(1)

93

Figure 3-7. LC-MS analysis of the reaction of 100 µM Bt RimO RCN with 1 mM SAM, 1 mM 13

mer peptide substrate, and either the Fld/FldR/NADPH reducing system reducing system,

comprised of 50 µM Fld, 25 µM FldR and 2 mM NADPH (A), or 2 mM sodium dithionite (B) as

the requisite source of electrons. The formation of 5'-dAH (red circles), SAH (blue squares), and

MS-1 product (black triangles) were best fit to a first-order exponential equation. The inset boxes

correspond to the linear portion of the data for the Fld/FldR/NADPH reducing system shown in

(C) and dithionite (D).

94

Table 3-1. Fit parameters of Bt RimO reactions containing SAM, a synthetic peptide substrate,

and the flavodoxin reducing system or dithionite as the reductant.

Flavodoxin Dithionite

A (µM) k (min-1

) ν

(µM·min-1

) A (µM) k (min

-1)

ν

(µM·min-

1)

5'-dAH 372

± 29

0.015

± 0.002

4.96

+ 0.16

787

± 38

0.011

± 0.001

7.57 +

0.18

SAH 138

± 10

0.023

± 0.004

2.54

± 0.18

198

± 41

0.007

± 0.002

1.18

± 0.12

Methylthiolated

Peptide (MS-1)

85

± 4

0.014

± 0.001

1.04

± 0.03

97

± 27

0.005

± 0.002

0.423

± 0.044

In the presence of the Fld/FldR/NADPH reducing system, SAM, and 1, Bt RimO

exhibited appreciably higher MS-1 formation (69 vs 42 µM) with a nearly 3-fold greater rate

constant and more than 2-fold greater initial rate than in the presence of dithionite after 2 h

(Table 3-1). The discrepancies between the actual amount of MS-1 formed and the amplitudes

from fits of the data reported in Table 3-1 can be attributed to premature termination of the

reactions at 2 h; allowing the reaction to proceed for additional time would provide amplitudes

generally in line with the total concentration of products formed at the end of the time course. In

addition to more robust activity exhibited by Bt RimO with the in vivo reducing system, use of

this source of electrons affords less uncoupling of 5'-dAH from MS-1 product formation. This

uncoupling, or abortive cleavage of SAM to form 5'-dAH in a non-productive manner, is

commonly observed in radical SAM enzymes, especially in the presence of dithionite

(references). As a result, the use of dithionite in the Bt RimO reaction caused 5'-dAH to be

formed in greater abundance (579 vs 311 µM) and with a higher rate constant and initial rate

(Table 3-1) than the corresponding parameters determined with the Fld/FldR/NADPH reducing

system. Likewise, formation of SAH in the presence of dithionite was five-fold greater than the

amount of MS-1 formed, indicating that aberrant formation of SAH occurred. This uncoupling of

SAH and MS-1 formation has been observed before in studies of Tm RimO and is not well

95

understood (14, 15, 18). Collectively, these results demonstrate that not only is the Ec

Fld/FldR/NADPH reducing system capable of transferring electrons to Bt RimO for use in the

methylthiolation reaction, but it is a better overall electron source with less uncoupling of 5'-dAH

and SAH formation from MS-1 production, and affords formation of MS-1 with a rate constant

that is ~ 3-fold greater and an initial rate that is ~ 2-fold greater than those observed with

dithionite. The Y225F variant of Bt RimO did not form any 5'-dAH or methylthiolated product in

detectable quantities, in contrast to the cognate variant in Tm RimO, which catalyzed formation of

one equivalent of 5'-dAH but no detectable methylthiolated product. However, the Y225F Bt

RimO variant was capable of SAH formation, leaving the exact role for this conserved tyrosine

residue in both Bt and Tm RimOs to be determined.

Determination of persulfide content of Bt RimO by fluorescent labeling

The crystal structure of Tm RimO RCN revealed a chain of electron density linking the

unique iron sites of the two 4Fe-4S clusters, which, when modeled and refined, fit well to a

covalently bonded pentasulfide chain (14) (Figure 3-8). Additionally, biochemical evidence

showed that Tm RimO was capable of multiple turnovers in the presence of sulfide, selenide, and

methanethiol, which suggested that the enzyme may harbor sulfide or a sulfide species for use in

the reaction, potentially by binding such a species to the unique iron ion of a 4Fe-4S cluster (14,

15). To determine whether a persulfide species was present or could be formed on Bt RimO RCN,

the monomeric form and the di- or trimeric form of the protein was treated with the fluorescent

dye 5-((2-[(iodoacetyl)amino]ethyl)amino)naphthalene-1-sulfonic acid (I-AEDANS). This

reagent reacts with the terminal sulfur of protein-bound persulfides to form a dye-protein adduct,

which is resolved upon incubation with dithiothreitol (DTT), thereby releasing the thiolated dye

96

into solution (Figure 3-9). Separation and collection of the dye from the protein allows for its

detection and quantification by fluorescence spectroscopy (31).

Figure 3-8. Active site of Tm RimO, depicting the iron ions (orange spheres) and sulfide ions

(yellow spheres) of the two 4Fe-4S clusters. The sulfur atoms of the modeled pentasulfide bridge

are shown as yellow sticks, with the cysteine residues ligating the auxiliary cluster and radical

SAM cluster shown in purple and black, respectively.

97

Figure 3-9. Labeling of protein-bound persulfide by the fluorescent dye 1,5-I-AEDANS. The dye

undergoes an SN2 reaction with the terminal sulfur of the persulfide, resulting in loss of iodide

and the formation of a protein-dye crosslink. Addition of DTT resolves the crosslink, resulting in

release of the thiolated dye into solution and loss of one sulfur atom from the former protein-

bound persulfide.

Bt RimO was incubated in the presence and absence of Na2S, spin-filtered to remove

adventitious Na2S, then incubated with 1,5-I-AEDANS to label any protein-bound persulfides

that were present after reconstitution or were formed during incubation with Na2S. In some

instances, labeling with 1,5-I-AEDANS was conducted in the presence of 1 or 4 M guanidinium

chloride in an attempt to disrupt some protein folding interactions, thereby affording greater

access to the active site for the dye to react with any persulfide species. A standard curve was

generated by serial dilution of a known amount of 1,5-I-AEDANS in buffer containing 5 mM

DTT, with subsequent analysis by fluorescence spectroscopy (Figure 3-10).

98

Figure 3-10. Standard curves of 1,5-I-AEDANS ranging from 0.78 to 50 µM (A) or 0.2 to 25 µM

(B) used to quantify the amount of 1,5-I-AEDANS that had reacted with any persulfides present

on Bt RimO RCN.

Quantification of the amount of 1,5-I-AEDANS released from the analyzed protein samples

showed that very few persulfides were present on either oligomeric form of Bt RimO RCN under

the conditions tested (Table 3-2). Specifically, the amount of protein-bound persulfide that had

reacted with the fluorescent dye under all conditions ranged from 0.27 to 0.85µM. In each of the

conditions, the concentration of Bt RimO RCN was 10 µM, indicating that only 2.7 to 8.5% of the

enzyme harbored a persulfide that reacted with the dye. Pre-incubation of the enzyme with Na2S

99

had no effect on the amount of persulfide detected versus the control experiments in which Na2S

was omitted (0.27 + 0.14 µM for the monomer and 0.45 + 0.08 µM for the di-/trimer in the

presence of Na2S, and 0.42 + 0.24 µM for the control experiments with the monomer lacking

Na2S incubation). Likewise, the presence of guanidinium chloride at concentrations of 1 and 4 M

had no effect on the amount of persulfide detected. Taken together, these results suggest that

reconstituted Bt RimO, regardless of its oligomeric state or the conditions under which the

persulfide labeling experiments were conducted, contained insufficient amounts of persulfide to

explain the observations of multiple turnovers in RimO from Thermotoga maritima. These data

also suggest that the proposed pentasulfide bridge linking the two 4Fe-4S clusters in the crystal

structure of Tm RimO is likely not physiologically relevant and, in all likelihood, is an artifact of

the excess iron and sulfide present in the conditions in which the crystals formed (14)

100

Table 3-2. Summary of results of 1,5-I-AEDANS labeling of persulfides present on Bt RimO

RCN. Underlined samples were omitted from calculating the average persulfide concentration

and its corresponding standard deviation within a set of samples.

Quantification of flavodoxin semiquinone consumption by Bt RimO under turnover

conditions

The methylthiolation reaction catalyzed by RimO requires at least one electron for

reductive cleavage of SAM to form the 5'-dA• that is used to activate the substrate for

101

methylthio- insertion; however, the exact number of electrons required for one turnover is

unknown. To determine the number of requisite electrons needed in the RimO reaction, we aimed

to quantify the amount of flavodoxin semiquinone (Fld SQ) consumed in reactions with Bt RimO

and correlate its consumption with product formation. The flavodoxin protein from E. coli binds

one molecule of flavin mononucleotide, which can be reduced by one or two electrons to its

semiquinone or hydroquinone forms, respectively (Figure 3-11).

Figure 3-11. The electronic forms of the flavin mononucleotide cofactor.

The semiquinone form of FMN is a blue-purple color and exhibits a distinct UV-Visible

spectrum with an absorbance maximum at 579 nm, which allows it to be monitored and

quantified. Accordingly, Fld SQ was formed by incubating the oxidized protein under an inert

atmosphere with 0.55 equivalents of the chemical reductant sodium dithionite. This reducing

agent forms two one electron reducing equivalents in the form of the SO2- radical anion due to its

sulfur-sulfur bond readily dissociating in solution (36). Following its isolation from dithionite by

gel-filtration and its quantification using a previously established extinction coefficient (ε579 =

4570 M-1

·cm-1

) (37), a known concentration of the Fld SQ was added to reaction mixtures

containing pre-methylated and gel-filtered Bt RimO RCN, [methyl-d3]SAM, and 1 to initiate the

102

reaction; the concentration of Fld SQ was monitored as a function of time. In a separate, but

identical, reaction mixture of 150 µL total volume, aliquots of the reactions were removed and

quenched in acid for quantification of Bt RimO reaction products. Shown in Figure 3-12A is the

time-dependent formation of 5'-dAH, SAH, MS-1 product resulting from the methyl group of

SAM used in the pre-incubation step, d3-MS-1 product resulting from [methyl-d3]SAM present in

the reaction mixture, the sum of unlabeled and d3-labeled MS-1, as well as the consumption of the

Fld SQ by Bt RimO. The formation or decay of all components in the reaction mixture was fitted

to a first-order single exponential equation as shown above in equation 1 except for that of 5'-

dAH and the sum of products, which were fitted to a double exponential equation (Equation 2).

The parameters obtained from fits of the data are summarized in Table 3-3.

(2)

Pre-methylation of 925 µM Bt RimO with 1.2 mM unlabeled SAM resulted in formation

of 505 µM SAH, corresponding to 0.55 equiv, or 55%, of the enzyme harboring an unlabeled

methyl group. Initiation of the reaction containing 72 µM of the pre-methylated enzyme with 85

µM Fld SQ in the presence of [methyl-d3]SAM and 1 resulted in rapid formation of unlabeled

MS-1 product (10.8 µM) in the first 10 min of the reaction followed by slower formation of d3-

MS-1 product (21.8 µM) , indicating that the semiquinone was indeed capable of reducing the

[4Fe-4S]2+

cluster to the catalytically active [4Fe-4S]1+

state (Figure 3-12C). It should be noted

here that 85 µM of Fld SQ was added to the reaction; however, immediately after its addition and

mixing of the reaction, the absorbance at 579 nm was recorded and the calculated concentration

of Fld SQ was 53.9 µM, resulting in a difference in concentration of 31.1 µM. This discrepancy

can be explained by rapid reduction of the fraction of enzyme capable of catalyzing methyl

transfer, or, in other words, the fraction of active enzyme. Of the 72 µM of Bt RimO in the

103

reaction, 55%, corresponding to 39.6 µM of the enzyme, was methylated, which is in close

agreement with the observed difference in the determined Fld SQ concentrations and the amount

of methylthiolated product formed. An analysis of the initial rates of formation determined by

fitting the linear portion of the data shows that of 5'-dAH is 6-fold greater than the initial rate of

decay of Fld SQ (Figure 3-13 and Table 3-3). Since formation of 5'-dAH requires reduction of

the [4Fe-4S]2+

cluster to which SAM binds, the fact that the initial rate of formation of 5'-dAH is

much greater than rate of decay of Fld SQ supports the hypothesis that the discrepancy in the

concentration of Fld SQ was due to rapid electron transfer to the enzyme to support reductive

cleavage of SAM.

The concentration of Fld SQ consumed during the course of the reaction, assuming rapid

reduction of the enzyme did indeed occur, was 69.4 µM, which is approximately two-fold greater

than the total concentration of methylthiolated product formed (32.6 µM). Interestingly, the rate

constant and initial rate of decay of Fld SQ are nearly identical to the corresponding parameters

for formation of the d3-MS-1 product, which may indicate that the 31.1 µM of Fld SQ that rapidly

reduced the enzyme was used in the first phase of the reaction in which the unlabeled methyl

group was transferred to the substrate to form 10.8 µM of MS-1, and the remaining 38.3 µM of

Fld SQ was used in the second phase of the reaction to form 21.8 µM of d3-MS-1 for a total of

two equivalents of Fld SQ used to form 1 equivalent of methylthiolated product. This fails to

explain why the amount of Fld SQ consumed in each phase of the reaction—31.1 µM in the first

rapid phase to form 10.8 µM of methylthiolated product and 38.3 µM in the second, slower phase

to form d3-methylthiolated product—is greater than the amount of product formed. Alternatively,

a more likely scenario is that the 31.1 µM of Fld SQ that rapidly reduced the enzyme was used in

one round of catalysis to form methyl- and d3-methylthiolated product, and release of these

products gated a second reduction of Bt RimO corresponding to slower Fld SQ decay.

104

Curiously, the concentration of 5'-dAH formed in the reaction was 91 µM, which was

1.5-fold greater than the concentration of Fld SQ used in the reaction. This difference may be

explained by the fact that abortive cleavage of SAM to form the 5'-dA• results in its quenching

with an accessible H• in a homolytic reaction to form 5'-dAH and another radical. This new

radical could transfer the unpaired electron back to the RS cluster to be used again in another

round of 5'-dA• generation, thereby obviating stoichiometric consumption of Fld SQ for 5'-dA•

formation.

105

Figure 3-12. Time-dependent formation of 5'-dAH, SAH, unlabeled MS-1 product, d3-labeled

MS-1 product, and time-dependent consumption of Fld SQ by Bt RimO RCN (A). Formation of

unlabeled MS-1 product and d3-labeled MS-1 product with other products formed omitted for

clarity (B). Formation of the sum of unlabeled MS-1 product and d3-labeled MS-1 product, and

time-dependent consumption of the Fld SQ (C). Linear portions of the data that were fitted to

determine initial rates of formation of reaction products (SAH not shown for clarity) and the

initial rate of decay for Fld SQ.

Table 3-3. Fit parameters of pre-methylated Bt RimO RCN reactions containing [methyl-

d3]SAM, a synthetic peptide substrate, and flavodoxin semiquinone.

A1 (µM) k1 (min-1) A2 (µM) k2 (min-1) ν (µM·min-1)

5'-dAH 36.4 0.116 98.3 0.009 2.87

± 5.2 ± 0.025 ± 71 ± 0.003 ± 0.45

SAH 29.7 0.027

NA NA 1.21

± 2.1 ± 0.006 ± 0.26

MS-1 product 9.7 0.217

NA NA 0.69

± 0.8 ± 0.049 ± 0.26

d3-MS-1 product

20.4 0.019 NA NA

0.45

± 0.6 ± 0.002 ± 0.01

Sum of MS-1 products

8.3 0.310 21.2 0.022 1.14

± 2.8 ± 0.196 ± 2.5 ± 0.005 ± 0.25

Fld SQ 39.9 0.011

NA NA 0.45

± 2.2 ± 0.002 ± 0.04

Discussion

Previous studies of RimOs from Ec and Tm used the chemical reductant dithionite as the

requisite source of electrons (14-16, 18). RimO from Ec exhibited meager activity with dithionite

(16), and we found use of the Ec flavodoxin reducing system with Ec RimO did not support any

appreciable methylthiolation activity. Experiments with Tm RimO and the chemical reductant

106

resulted in robust turnover; however, much abortive cleavage of SAM to form 5'-dAH and

aberrant formation of SAH was observed, which precluded the determination of the number of

equivalents of 5'-dA• required for each methylthiolated product (14, 15, 18). Thermotoga

maritima lacks the genes encoding the flavodoxin reducing system, instead relying on five

ferredoxins as its electron mediators, one of which may be the in vivo source of electrons for the

RimO reaction. Given the recent explosion in gut microbiome research, we have been interested

in studying RS enzymes from mesophilic gut bacteria, and, additionally, in finding a RimO that

uses the Fld/FldR/NADPH reducing system to obviate use of dithionite. We chose to overexpress

and purify RimO from Bacteroides thetaiotaomicron (Bt), a major bacteria species found in adult

intestine that encodes a flavodoxin with 37% sequence identity to that in E. coli. The

overexpression of the pSC-His-BtRimO gene and subsequent purification and reconstitution of its

gene product by IMAC and size-exclusion chromatography yielded nearly homogenous protein

containing ~ 8 iron ions and ~ 9 sulfide ions with which we conducted the studies described

herein.

The activity of reconstituted Bt RimO was measured in the presence of the chemical

reductant dithionite or the Ec Fld/FldR/NADPH reducing system. These two sources of the

requisite electron required to reduce the [4Fe-4S]2+

RS cluster to its active [4Fe-4S]+ state for 5'-

dA• generation both supported radical generation and formation of the methylthiolated product,

as detected and quantified by LC/MS. While it was expected that Bt RimO would exhibit

methylthiolation activity in the presence of dithionite, since previous studies in which it was used

with RimOs from Ec(16) and, to a greater extent, from Tm(14, 15, 18) showed the proteins to be

active, the finding that the Ec Fld/FldR/NADPH reducing system supported turnover was

somewhat surprising, given that this system does not support RimO activity in its native

organism. Bt RimO exhibited greater methylthiolation activity overall, with a determined rate

constant ~ 3-fold greater and an initial rate of formation ~2-fold greater than those determined for

107

the reaction with dithionite. Importantly, this reducing system decreased the amount of 5'-dAH

formed abortively, but unfortunately did not eliminate it; however, these experiments should be

replicated to confirm that these findings are reproducible. The observation of significant SAM

abortive cleavage with the in vivo reducing system from Ec may be attributed to the use of a

truncated peptide substrate rather than the full S12 protein, which could be missing residues

necessary for proper docking of the substrate to react productively with 5'-dA•. Aberrant

formation of SAH in the presence of dithionite observed with Tm RimO (14, 15, 18), and in this

study of Bt RimO, is not well understood; however, use of the Fld/FldR/NADPH reducing system

minimized its off-pathway production. The amounts of SAH formed were consistently ~2-fold

that of both methylthiolated product formed and the presumed concentration of active enzyme in

the reaction, and can be rationalized by a second methyl transfer from SAM to the acceptor site of

the active enzyme following product release. Since the Ec Fld/FldR/NADPH reducing system

supported activity with Bt RimO and seemed to be a tractable system to study, we decided to

further characterize Bt RimO in an attempt to gain further insight on the mechanistic details of the

methylthiolation reaction.

As expected, quantification of iron and sulfide content along with EPR characterization

of Bt RimO strongly suggested the presence of two [4Fe-4S]+ clusters. The addition of SAM

allowed for differentiation of the two clusters, since its addition resulted in shifting of the g-value

of the RS cluster slightly lower. The observed g-values observed agreed with those of Ec and Tm

RimO, as was expected (16, 18).

Analytical molecular sieve chromatography of reconstituted Bt RimO showed that the

enzyme adopts two oligomeric states in low salt, high salt, or low salt buffer containing SAM: a

dimeric or trimeric state that was the major species present, and a monomeric form. Of the RS

enzymes with determined oligomeric states, the majority are monomeric(38) (22) (30, 39-41) or

dimeric (42), with only one example of formation of a homotrimer (43). Taking this into account

108

and given the results of analytical molecular sieve chromatgraphy, we conclude that Bt RimO

adopts monomeric and dimeric quarternary structures under the conditions we tested; it remains

to be determined which of these structures the enzyme adopts during catalysis.

The quantification of persulfides present in Bt RimO RCN with the fluorescent dye 1,5-I-

AEDANS determined that only 2.7 to 8.5% of the total protein harbored persulfide under the

examined conditions. This result stands in contrast to the observations that reconstituted Tm

RimO has been shown to catalyze up to 3 turnovers, which requires the enzyme to harbor at least

3 sulfide ions in some form (14, 15). It is possible that adventitiously bound sulfide was present

in these studies and was used by the enzyme to support multiple turnovers, but the rates of

product formation do not show any indications of being biphasic, which would be expected if the

first sulfide ion in the proper location for the reaction was used for one turnover relatively rapidly

in the first phase followed by slow, adventitious occupation of this binding site by another sulfide

ion for an additional round of turnover. This scenario is similar to spontaneous Fe-S cluster

reassembly observed in the RS enzyme biotin synthase, in which insertion of 2 sulfur ions

harbored as constituents of a 2Fe-2S cluster were used relatively quickly in one round of catalysis

to form solely 32

S-containing biotin product, followed by slow reconstitution of the 2Fe-2S

cluster with 34

S present in the reaction mixture to support much slower formation of 34

S-

containing product (44). We have yet to observe with Bt RimO significant concentrations of

methylthiolated product above that of the enzyme in reactions lacking exogenously supplied

sulfide or methanethiol. The low levels of persulfide detected with this enzyme, in addition to its

quantified sulfide never exceeding 9 sulfide ions per polypeptide, suggest it does not harbor

additional sulfide ions for the methylthiolation reaction, and, instead, may rely on (an)other

protein(s) in vivo to supply sulfide for multiple rounds of catalysis.

The exact number of electrons required for formation of methylthiolated product by

RimO is unknown. At least one electron is necessary for reductive cleavage of SAM to form 5'-

109

dA•. With the finding that the Ec Fld/FldR/NADPH reducing system was a competent source of

electrons for Bt RimO, we now had a spectroscopic handle with which we could quantify the

number of electrons used in the reaction in the form of the Fld SQ. Based on our results, two

equivalents of Fld SQ were consumed per equivalent of methylthiolated product formed. We

hypothesize that one of the Fld SQ equivalents was used for formation of product, and the second

is used to reduce Bt RimO again to be primed for another round of catalysis, similar to our

observations of SAH production wherein one equivalent is formed for methylthiolation of the

substrate, and a second equivalent results from another methyl transfer to the enzyme for a second

reaction. Rapid reduction of Bt RimO pre-methylated with SAM resulted in a burst of unlabeled

methylthiolated product, followed by slower decay of the Fld SQ and formation of d3-

methylthiolated product. The sum of products formed (32.6 µM) is consistent with the

concentration of Fld SQ consumed in the rapid reduction of Bt RimO (31.1 µM); the slower

decay of Fld SQ (k = 0.011 + 0.002 min-1

and v = 0.45 + 0.04 µM · min-1

) is consistent with

relatively slow d3-product formation (k = 0.019 + 0.002 min-1

and v = 0.45 + 0.01 µM · min-1

)

and release that gates the second reduction of the enzyme.

110

References

1. Savage DC. Annual Review of Microbiology. 1977: 31, 107-133

2. Luckey TD. The American Journal of Clinical Nutrition. 1972: 25, 1292-1294

3. Savage DC. Annual Review of Nutrition. 1986: 6, 155-178

4. Shipman JA, Cho KH, Siegel HA, Salyers AA. Journal of Bacteriology. 1999: 181,

7206-7211

5. Hooper LV, Midtvedt T, Gordon JI. Annual Review of Nutrition. 2002: 22, 283-307

6. Wu H-J, Wu E. Gut Microbes. 2012: 3, 4-14

7. Guarner F, Malagelada J-R. The Lancet. 2003: 361, 512-519

8. Sekirov I, Russell SL, Antunes LCM, Finlay BB. Physiological Reviews. 2010: 90, 859-

904

9. Salyers AA. Annual Review of Microbiology. 1984: 38, 293-313

10. Xu J, Bjursell MK, Himrod J, et al. Science. 2003: 299, 2074-2076

11. Bergman EN. Physiological Reviews. 1990: 70, 567-590

12. Shoemaker NB, Vlamakis H, Hayes K, Salyers AA. Applied and Environmental

Microbiology. 2001: 67, 561-568

13. Ulger N, Celik C, Cakici O, Soyletir G. Anaerobe. 2004: 10, 255-259



15416



USA. 2008: 105, 1826-1831


285, 5792-5801

19. Brodersen DE, Clemons Jr WM, Carter AP, et al. Journal of Molecular Biology. 2002:

316, 725-768

20. Ogle JM, Brodersen DE, Clemons WM, et al. Science. 2001: 292, 897-902

21. Lanz ND, Grove TL, Gogonea CB, et al. 2012. In Methods in Enzymology, ed. AH

David, pp. 125-152: Academic Press

22. Yu L, Blaser M, Andrei PI, et al. Biochemistry. 2006: 45, 9584-9592

23. Cicchillo RM, Iwig DF, Jones AD, et al. Biochemistry. 2004: 43, 6378-6386

24. Cicchillo RM, Lee K-H, Baleanu-Gogonea C, et al. Biochemistry. 2004: 43, 11770-11781

25. Bradford MM. Analytical Biochemistry. 1976: 72, 248-254

26. Lanz ND, Grove TL, Gogonea CB, et al. Methods in Enzymology. 2012: 516, 125-152

27. Beinert H. Analytical Biochemistry. 1983: 131, 373-378


29. Kennedy MC, Kent TA, Emptage M, et al. Journal of Biological Chemistry. 1984: 259,

14463-14471

30. Lanz ND, Lee K-H, Horstmann AK, et al. Biochemistry. 2016: 55, 1372-1383

31. Liu Y, Dos Santos PC, Zhu X, et al. Journal of Biological Chemistry. 2012: 287, 5426-

5433

32. Hudson EN, Weber G. Biochemistry. 1973: 12, 4154-4161

33. Jenkins CM, Waterman MR. Biochemistry. 1998: 37, 6106-6113

34. Zheng L, Cash VL, Flint DH, Dean DR. Journal of Biological Chemistry. 1998: 273,

13264-13272

35. Johnson DC, Unciuleac M-C, Dean DR. Journal of Bacteriology. 2006: 188, 7551-7561

111

36. Fox JL. FEBS Letters. 1974: 39, 53-55

37. Mayhew SG, Foust GP, Massey V. Journal of Biological Chemistry. 1969: 244, 803-810

38. Challand MR, Martins FT, Roach PL. Journal of Biological Chemistry. 2010: 285, 5240-

5248

39. Broderick JB, Duderstadt RE, Fernandez DC, et al. Journal of the American Chemical

Society. 1997: 119, 7396-7397

40. Quitterer F, List A, Eisenreich W, et al. Angewandte Chemie International Edition. 2012:

51, 1339-1342

41. Zhang Y, Zhu X, Torelli AT, et al. Nature. 2010: 465, 891-896

42. Sanyal I, Gibson KJ, Flint DH. Archives of Biochemistry and Biophysics. 1996: 326, 48-

56

43. Lehtiö L, Goldman A. Protein Engineering Design and Selection. 2004: 17, 545-552

44. Farrar CE, Siu KKW, Howell PL, Jarrett JT. Biochemistry. 2010: 49, 9985-9996

112

Chapter 4

The Stereochemical Course of the Reaction Catalyzed by the Radical SAM

Methylthiotransferase RimO

RimO is a member of the growing radical S-adenosylmethionine (SAM) superfamily of

enzymes, which use a reduced [4Fe−4S] cluster to effect reductive cleavage of the 5′ C−S bond of

SAM to form a 5′-deoxyadenosyl 5′-radical (5′-dA•) intermediate. RimO uses this potent oxidant

to catalyze the attachment of a methylthio group (−SCH3) to C3 of aspartate 89 of protein S12,

one of 21 proteins that compose the 30S subunit of the bacterial ribosome. However, the exact

mechanism by which this transformation takes place has remained elusive. Herein, we describe

the stereochemical course of the RimO reaction. Using peptide mimics of the S12 protein bearing

deuterium at the 3 pro-R or 3 pro-S positions of the target aspartyl residue, we show that RimO

from Bacteroides thetaiotaomicron (Bt) catalyzes abstraction of the pro-S hydrogen atom, as

evidenced by the transfer of deuterium into 5′-deoxyadenosine (5′- dAH). The observed kinetic

isotope effect on H atom versus D atom abstraction is ∼1.9, suggesting that this step is at least

partially rate determining. We also demonstrate that Bt RimO can utilize the

flavodoxin/flavodoxin oxidoreductase/NADPH reducing system from Escherichia coli as a

source of requisite electrons. Use of this in vivo reducing system decreases, but does not

eliminate, formation of 5′-dAH in excess of methylthiolated product.

Introduction

RimO (ribosomal modification O) catalyzes the posttranslational modification of

aspartate 89 (D89) of protein S12 to give 3-methylthioaspartate (3-MS-D89) (1). Protein S12 is a

component of the 30S subunit of the bacterial ribosome, and the loop on which D89 resides

projects into the acceptor site where aminoacyl tRNAs bind (2). The modification itself is not

113

essential, but Escherichia coli (Ec) that are capable of catalyzing this methylthiolation reaction

have a slight growth advantage over those that are not. This growth advantage is believed to be

related to enhanced translational fidelity (1). Two recent X-ray structures of the Thermotoga

maritima (2.3−2.5 Å) and Ec (2.4 Å) ribosome showed electron density for 3-MS-D89, and in the

Ec structure the absolute stereochemistry at C3 was observed to be R (3,4). The methylthio group

points toward the 6-oxo group of N7-methylguanosine 527, a modified nucleobase in 16S rRNA;

however, the purpose of the interaction between the modified protein residue and the nucleobase

is unknown.

As a member of the radical SAM (RS) superfamily of enzymes, RimO uses a [4Fe−4S]+

cluster to promote the reductive cleavage of the 5′ C−S bond of SAM to form a 5′-deoxyadenosyl

5′-radical (5′-dA• ) (5-7). This potent oxidant has been suggested to abstract a hydrogen atom (H•

) from C3 of D89 to activate it for methylthiolation, although this particular step of the reaction

has never been demonstrated (1,6,7). In addition to the [4Fe−4S] cluster that participates in the

reductive cleavage of SAM (RS cluster), RimO harbors an additional [4Fe−4S] cluster (auxiliary

cluster) in its N-terminal region. The auxiliary cluster was previously believed to be sacrificed

during the first phase of the reaction to provide the inserted sulfur atom (6) in a mechanism

analogous to those proposed for the RS sulfurtransferases, lipoyl synthase (LipA) (8−12) and

biotin synthase (BioB) (13−17). In the second phase of RimO catalysis, the inserted sulfur atom

was proposed to undergo methylation by a canonical SAM-dependent SN2 mechanism (6). Recent

studies suggest, however, that RimO and the related enzyme, MiaB, catalyze the initial synthesis

of a methylthio group that is most likely attached externally to the auxiliary Fe/S cluster and that

the entire methylthio group is subsequently transferred intact to C3 of the aspartyl residue via

radical chemistry (18,19). These observations indicate that RimO and MiaB are members of an

emerging subclass of RS enzymes that, within a single active site, activate SAM both for

methyltransfer and for generation of a 5′-dA• (20,21).

114

Although RimOs both from Ec (6) and from Thermotoga maritima (Tm) (7) have been

characterized, Tm RimO is better behaved and exhibits significantly greater turnover. However,

the need to use the artificial reductant, sodium dithionite, in activity determinations of Tm RimO

induces production of SAH and 5′-dAH that was believed to be uncoupled from product

formation. In this study, we show that RimO from the gut bacterium, Bacteroides

thetaiotaomicron, can utilize the Ec flavodoxin/flavodoxin oxidoreductase/NADPH

(Fld/FldR/NADPH) reducing system as a source of electrons for catalysis. We also determine the

stereochemistry of H• abstraction from D89 of the S12 protein through the use of

chemoenzymatically synthesized deuterated isotopomers at C3 of an aspartate residue

incorporated into a 13-amino acid peptide mimic (13-mer) of the S12 protein. We observe

transfer of deuterium into 5′-deoxyadenosine only from the 13-mer containing [(2S,3S)-2,3-2H2]-

aspartate and not [(3R)-3-2

H1]-aspartate, indicating that H• abstraction is indeed stereoselective,

and that insertion of the −SCH3 group occurs with inversion of configuration at C3. The apparent

kinetic isotope effect associated with H versus D atom abstraction by the 5′-dA• is ∼1.9,

indicating that H atom abstraction is at least partially rate limiting.


Materials


MA). L-tryptophan, 2-mercaptoethanol, L-(+)-arabinose, ferric chloride, sodium methanethiolate,

5’-deoxyadenosine (5’-dA), 2-methylpropene (isobutylene), triphenylphosphine, p-

toluenesulfonic acid monohydrate, dimethyl acetylenedicarboxylate, copper (II) sulfate

pentahydrate, NADPH, H2S, and S-adenosyl-L-homocysteine (SAH) were purchased from Sigma

115

Corp (St. Louis, MO). N-(2-hydroxyethyl)piperizine-N'-(2-ethanesulfonic acid) (HEPES) was

purchased from Fisher Scientific (Pittsburgh, PA), and imidazole was purchased from J. T. Baker

Chemical Co. (Phillipsburg, NJ). Potassium chloride, glycerol, and expression vectors pET-28a

and pET-26b were purchased from EMD Chemicals (Gibbstown, NJ), while dithiothreitol (DTT)

and isopropyl-β-D-1-thiogalactopyranoside (IPTG) were purchased from Gold Biotechnology (St.

Louis, MO). Coomassie blue dye-binding reagent for protein concentration determination was

purchased from Pierce (Rockford, IL), as was the bovine serum albumin standard (2 mg/mL).

PD-10 pre-poured gel-filtration columns, as well as Sephadex G-25 resin were purchased from

GE Biosciences (Piscataway, NJ). All other buffers and chemicals were of the highest grade

available.

Methods

Cloning and Overexpression of the Ec aspA gene

The Ec aspA gene was amplified from Ec genomic DNA using polymerase chain reaction

(PCR) technology. The forward amplification primer (5’- CGC GGC GTC CAT ATG TCA AAC

AAC ATT CGT ATC GAA GAA GAT CTG TTG G -3’) included an NdeI restriction site

(underlined) flanked by a nine-base GC clamp and the first 34 bases of the aspA gene. The

reverse primer (5'- CGC GGC GTC CTC GAG TTA CTG TTC GCT TTC ATC AGT ATA GCG

TTT TGC-3') contained a XhoI restriction site (underlined) flanked by a nine-base GC clamp and

the last 33 bases of the aspA gene, including the stop codon. The PCR was performed with a

Stratagene Robocycler thermocycler (La Jolla, CA) as described previously (22). The product

was isolated and digested with NdeI and XhoI and ligated into similarly digested pET-28a by

standard procedures . The correct construct, encoding a 10 amino acid linker between the gene

116

product and an N-terminal hexahistidine tag, was verified by DNA sequencing and designated

pEcAspA.

Expression vector pEcAspA was transformed into Ec BL21(DE3) for gene expression.

Bacterial growth and gene expression was carried out at 37 °C in 8 L of Luria-Bertani media

distributed evenly among 4 Erlenmeyer flasks with moderate shaking (180 rpm). At an optical

density at 600 nm of 0.6, a sterile solution of IPTG was added to each flask to a final

concentration of 200 µM. Expression was allowed to take place for 5 h at 37 °C before the cells

were harvested by centrifugation at 10,000 g for 10 min at ambient temperature.

Purification of Ec AspA

Purification of Ec AspA was carried out by immobilized metal affinity chromatography

(IMAC) using Ni-NTA resin at 4˚C (Qiagen, Valencia, CA). Buffers used during the purification

of Ec AspA were as follows: lysis buffer (50 mM HEPES, pH 7.5, 300 mM KCl, 10 mM 2-

mercaptoethanol, 20 mM imidazole, and 1 mg/mL lysozyme); wash buffer (50 mM HEPES, pH

7.5, 300 mM KCl, 10 mM 2-mercaptoethanol, 10% (v/v) glycerol, and 40 mM imidazole); elution

buffer (wash buffer containing 250 mM imidazole). After lysing the cells by sonication, the cell

suspension was transferred into sterile centrifuge tubes, which were subsequently subjected to

centrifugation at 50,000 g at 4˚C for 1 h. The supernatant was loaded onto Ni-NTA resin

equilibrated in lysis buffer and was subsequently washed with 120 mL of wash buffer. After

addition of elution buffer to the column, protein-containing fractions, as determined by UV-Vis

spectrophotometric analysis at 280 nm on a Cary 50 spectrophotometer (Varian, Walnut Creek,

CA), were pooled and concentrated using an Amicon stirred ultrafiltration apparatus (Millipore,

Billerica, MA) fitted with a YM-30 membrane (30,000 molecular weight cutoff). The protein

117

was exchanged into gel-filtration buffer (50 mM potassium phosphate, pH 8, 10% glycerol, using

a Sephadex G-25 column (2.5 13 cm), concentrated again and stored in aliquots at -80˚C until

ready for use.

Chemoenzymatic syntheses of (2S,3R)-3-[2H1] aspartic acid (pro-R) and (2S,3S)-[2,3-

2H2]

aspartic acid (pro-S) and their incorporation into synthetic S12 13-mer peptide substrates

(2S,3R)-3-[2H1] aspartic acid (pro-R) and (2S,3S)-[2,3-

2H2] aspartic acid (pro-S) were

prepared as previously described by Young and colleagues (23) and Richards and colleagues (24).

After crystallization of the labeled aspartic acids, each was chemically activated for solid phase

peptide synthesis using standard Fmoc and t-butyl protection strategies for the amino and β-

carboxylic acid moieties of aspartic acid, respectively. Each labeled aspartic acid was added to a

thick-walled glass bottle containing dioxane in which 2-fold molar excess para-toluenesulfonic

acid had been dissolved. The bottle was placed in an ice bath and stirred. 2-methylpropene

(isobutylene) was condensed in a flask placed in a dry ice/acetone bath and was quickly added to

the glass bottle, tightly capped, and allowed to react with vigorous stirring for 72 h at room

temperature. The cap was carefully removed, and the reaction was vigorously stirred and gently

heated to evolve excess isobutylene gas. Once isobutylene bubbles were no longer observed (~ 3

h), the solution was titrated to pH 10 with 10% sodium carbonate and placed in an ice bath.

Fmoc N-hydroxysuccinimide ester (Fmoc-OSu) (AnaSpec, Inc. Fremont, CA), in 3-fold molar

excess of the labeled aspartic acid, was dissolved in dioxane and added dropwise to the mixture,

stirred for 1 h on ice, and then allowed to react overnight while stirring at room temperature. The

next day the reaction was poured into an equal volume of ice water and then extracted with

diethyl ether to remove unreacted Fmoc-OSu. The aqueous fraction was slowly acidified to pH 2

with 6 M HCl, vigorously stirred to promote evolution of CO2, and subsequently extracted 3 times

118

with ethyl acetate. The ethyl acetate extractions were combined and dried over MgSO4, filtered,

and concentrated in vacuo to yield a thick, yellow syrup, which contained a mixture of the labeled

Fmoc-N-aspartic acid and the desired labeled Fmoc-N-aspartic acid β-tert-butyl ester. This

mixture was dissolved in 3:1 0.1% TFA, 60% methanol:acetonitrile, placed on ice, and allowed to

crystallize to yield the desired (2S,3R)-3-[2H1] Fmoc-N-aspartic acid β-tert-butyl ester or (2S,3S)-

[2,3-2H2] Fmoc-N-aspartic acid β-tert-butyl ester. NMR analysis to confirm the identity of each

compound and retention of the deuterium label was conducted on a Bruker Avance III HD 500.20

MHz spectrometer equipped with a 5 mm Prodigy BBO z-gradient probe (Figures 4-1 and 4-2).

NMR spectral acquisition parameters were as follows: 32 scans, 1 second relaxation delay, 3.3 s

acquisition time, and 64k points. NMR spectra were processed with MNova software (MestreLab

Research, S.L., Santiago, Spain) with the following functions: Exponential: 1.6 Hz; Gaussian:

1.10 Hz; TRAF: 0.2 Hz.


2H1] Fmoc-N-aspartic acid β-tert-butyl ester (pro-R

3-[2H1]-aspartate).

119

(2S, 3R)-3-[2H1] Fmoc-N-aspartic acid β-tert-butyl ester:

1H NMR (500 MHz, chloroform-d) δ

7.76 (d, J = 7.5 Hz, 2H), 7.59 (t, J = 5.6 Hz, 2H), 7.35 (dt, J = 44.7, 7.5 Hz, 4H), 5.83 (d, J = 8.5

Hz, 1H), 4.65 (dd, J = 8.6, 4.5 Hz, 1H), 4.49 – 4.30 (m, 2H), 4.24 (t, J = 7.1 Hz, 1H), 2.98 (d, J =

4.4 Hz, 1H), 1.46 (s, 9H).


2H1] Fmoc-N-aspartic acid β-tert-butyl ester (pro-S

2,3-[2H1]-aspartate)

(2S, 3S)-[2,3-2H2] Fmoc-N-aspartic acid β-tert-butyl ester:

1H NMR (500 MHz, chloroform-d) δ

7.76 (d, J = 7.5 Hz, 2H), 7.59 (t, J = 5.7 Hz, 2H), 7.35 (dt, J = 44.4, 7.4 Hz, 4H), 5.84 (s, 1H),

4.48 – 4.31 (m, 2H), 4.24 (t, J = 7.1 Hz, 1H), 2.77 (s, 1H), 1.46 (s, 9H).

Following NMR confirmation of the desired deuterium-labeled aspartates, each

compound was incorporated into a synthetic peptide by the Penn State Hershey College of

Medicine Macro Core Facility. The peptides consisted of 13 amino acids (NH2-

120

RGGRVKDLPGVRY-COOH) corresponding to residues 83-95 of the Bacteroides

thetaiotaomicron S12 ribosomal protein, in which D denotes the pro-R- or pro-S-labeled aspartic

acid. Hereafter these peptides are referred to as 2 and 3, respectively. MALDI-TOF analysis of

the crude peptides was used to verify the target peptide mass and retention of their respective

deuterium labeling. 2 and 3 were then purified on an Agilent 1100 series high-performance liquid

chromatography (HPLC) system (Agilent Technologies, Santa Clara, CA) using a ZORBAX SB-

C18 9.4 mm 25 cm semi-prep column (Agilent, Santa Clara, CA) with 0.1% trifluoroacetic

acid (solvent A) and acetonitrile (solvent B) flowing at 4 mL/min with UV-vis detection at 275

nm. The column was equilibrated in 100% solvent A. After sample injection, the following

gradient was applied: 25% solvent B (0-10 min), 30% solvent B (10-15 min), 100% solvent B

(15-20 min), 0% solvent B (20-25 min). The target peptides eluted at 12.6 min and were

collected, analyzed by ESI+ LC/MS to confirm their identities, lyophilized, resuspended in

anaerobic 18 MΩ water, titrated to pH 7 with NaOH, and flash frozen in liquid N2 until use.

Determination of the stereospecificity of hydrogen atom abstraction by Bt RimO

The pro-R and pro-S labeled peptide substrates were used in reactions with Bt RimO to

determine which compound afforded deuterium enrichment into 5'-deoxyadenosine (5'-dAH),

thereby indicating which hydrogen atom is removed by the 5'-deoxyadenosyl radical. Each Bt

RimO reaction contained the following in a final volume of 220 µL: 100 µM Bt RimO, 3 mM

SAM, 3 mM 1, 2, or 3 as substrate, 50 mM Na-HEPES, pH 7.5, and, where appropriate, 200 µM

Ec flavodoxin, 50 µM Ec flavodoxin reductase, and 3 mM NADPH. In some instances 2 mM

dithionite was used to replace the flavodoxin reducing system. All components except SAM

were incubated at 37 °C for 15 min before initiating the reaction with the omitted component.

Aliquots (15 µL) of the reaction mixture were withdrawn at various times from 0-180 min and

121

added to 15 µL of 0.5 M H2SO4 containing 100 µM AtsA peptide (NH2-PMSAPARSM-COOH,

4) and 100 µM tryptophan as external standards to quench the reaction. Precipitated protein was

removed by centrifugation at 18,000 g for 15 min, and a 40 µL aliquot of the resulting

supernatant was subjected to analysis by ESI+ LC/MS with single-ion monitoring (SIM) as

previously described (19). Solvent A consisted of ammonium acetate (40 mM) and methanol (5%

v/v) titrated to pH 6.0 with acetic acid, while solvent B was 100% acetonitrile. The column was

equilibrated in 100% solvent A at a flow rate of 0.5 mL min-1

. After sample injection (5 µL), a

gradient was applied from 0% solvent B to 100% solvent B over 10 min and then from 100% to

0% over 3 min. The monitored ions (m/z) and retention times (min), respectively, were 385.1 and

2.7 (SAH), 188.0 and 3.0 (tryptophan), 252.1 and 3.2 (5'-dAH), 253.1 and 3.2 (5'-dAD), 474.4

and 3.8 (4), 737.1 and 4.0 (1), 737.6 and 4.0 (2), 738.1 and 4.0 (3), 760.1 and 4.1 (methylthiolated

peptide, 3-MS-1), 760.6 and 4.1 (monodeuterated methylthiolated peptide, 3-MS-2), 761.6 and

4.1 (dideuterated methylthiolated peptide, 3-MS-3). Calibration curves were generated with

known concentrations of each analyte and were run under identical conditions to determine the

concentration of products generated in assays. Data were analyzed using the Agilent

Technologies MassHunter qualitative and quantitative analysis software. For methionine

quantification, ESI+ LC/MS with multiple reaction monitoring (MRM) was used. Solvent A

consisted of 0.2% formic acid, while solvent B was 100% methanol. The column was equilibrated

in 97% solvent A, 3% solvent B at a flow rate of 0.5 mL min-1

. After sample injection (10 µL),

isocratic conditions were maintained from 0 to 4.5 min, a gradient of 3% solvent B to 90%

solvent B was applied from 4.5 to 6.5 min, and then from 90% solvent B to 3% solvent B over 3

min. The monitored transition (m/z) and retention time (min) for methionine were 150.1/104.1

and 1.7, respectively.

122

Results

In previous studies of Ec and Tm RimOs, dithionite was used as the source of requisite

electrons for catalysis. Tm RimO catalyzed methylthiolation in the presence of dithionite, but not

in the presence of the Ec Fld/FldR/NADPH reducing system, while Ec RimO did not exhibit

appreciable activity with either reductant. The use of the chemical reductant in the Tm RimO

reaction, however, was believed to induce abortive cleavage of SAM, as evidenced by production

of 5′-dAH and SAH that were in excess over the methylthiolated product (7,18,19). The Tm

genome does not appear to encode flavodoxins but does encode five ferredoxins, one or more of

which might function to deliver electrons to RS enzymes. However, because of our interest in RS

enzymes from gut microbiota and the desire to work with enzymes that function optimally closer

to ambient temperature, we chose to study RimO from the major human gut bacterium,

Bacteroides thetaiotaomicron (Bt), which does contain an annotated flavodoxin gene. Bt RimO is

35% identical to Ec RimO and 39% identical to Tm RimO, while Bt flavodoxin is 37% identical

to Ec flavodoxin. The gene encoding Bt RimO was coexpressed with plasmid pDB1282, and its

protein product was isolated as previously described for Tm and RimOs (6,19) When Bt RimO

was incubated under turnover conditions in the presence of the Fld/FldR/NADPH reducing

system and a 13-aa peptide corresponding to residues 83−95 of the Bt S12 protein, formation of

SAH (m/z = 385.0), 5′-dAH (m/z = 252.1), and methylthiolated peptide, 3-MS-1 (m/z = 760.1),

was observed by LC/MS (Figure 4-5). Interestingly, however, the final concentrations of SAH

and 5′-dAH were 1.1- to 2.5-fold higher than that of the methylthiolated product, which is similar

to that observed when assays were conducted using dithionite as the requisite source of electrons.

123

Figure 4-3. Bt RimO-catalyzed time-dependent formation of SAH, 5'-dAH, and methylthiolated

product (3-MS-1) with the Ec Fld/FldR/NADPH reducing system. The reaction was conducted as

described in the methods and contained 100 μM Bt RimO, 3 mM SAM, 3 mM peptide (1), 200

μM Fld, 50 μM FldR, 3 mM NADPH, and 50 mM Na-HEPES pH 7.5.

To determine initial rates of formation for SAH, 5′-dAH, and 3-MS-1, the amplitudes of

each of the corresponding curves were multiplied by the first-order rate constants obtained from

fits of the data to a single exponential equation, resulting in initial rates of 19.7 ± 0.68 μM min−1

(SAH) 6.03 ± 0.01 μM min−1

(5′-dAH), and 3.95 ± 0.01 μM min−1

(3-MS-1). These results

suggest that methylation of RimO by SAM takes place relatively rapidly as compared to

formation of 5′-dAH and product and that the Ec Fld/FldR/NADPH reducing system is indeed

capable of delivering electrons to Bt RimO for catalysis. Interestingly, while formation of 5′-

dAH, even in the presence of the in vivo reducing system, is in excess of methylthiolated product,

reactions conducted in ∼90% D2O result in no 5′-dAD above natural abundance. This observation

suggests that any 5′-dA• that does not lead to the correct product abstracts an enzyme- or

substrate-derived H atom that is not from a solvent exchangeable site (Figure 4-6).

124

Figure 4-4. Quantification of 5'-dAD generated in the reaction of Bt RimO conducted in 90% D2O

with unlabeled peptide (1). The reaction was conducted as described above with the following

modifications: the peptide (1), NADPH, and Na-HEPES were dissolved in D2O, lyophilized, and

resuspended in D2O; the protein components (Bt RimO, Ec Fld, and Ec FldR) were diluted in

D2O for two hours on ice prior to the reaction. The reaction mixture contained 100 µM Bt RimO,

2 mM SAM, 3 mM peptide (1), 150 µM Ec Fld, 37.5 µM Ec FldR, 2 mM NADPH, 50 mM Na-

HEPES pH 7.5 in D2O. The lines are fits to a first-order exponential equation with the following

kinetic parameters: SAH formation: A = 119.9 + 10.5 µM, v = 4.3 + 0.4 µM min-1

; 5'-dAH

formation: A = 286.3 + 8.9 µM, v = 5.1 + 0.2 µM min-1

; product formation: A = 121.2 + 4.8 µM, v

= 1.9 + 0.1 µM min-1

.

To determine the stereoselectivity of H atom abstraction, deuterium-containing

isotopologs (pro-R or pro-S) at C3 of aspartate were incorporated into peptide substrates, which

were subsequently used in the RimO reaction. The synthesis of the 3-pro-R and 3-pro-S

deuterated substrates followed the synthetic strategies (Figure 4-7) described by Young et al.

(23) and Richards et al., (24) which exploit the ability of aspartate ammonia-lyase (AAL) to

catalyze the stereoselective incorporation of deuterium from D2O into the C3 pro-R position of

aspartate when incubated with fumarate and excess ammonium chloride to afford (2S,3R)-3- [2H1]

aspartate. Similarly, AAL catalyzes the stereoselective incorporation of a proton in the C3 pro-R

125

position of aspartate and, when incubated with [2,3-2H2]-fumarate and ammonium chloride in

H2O, affords (2S,3S)-[2,3-2H2] aspartate (23).

Figure 4-5. Synthetic routes for (2S,3R)-3-[2H1] Fmoc-N-aspartic acid β-tert-butyl ester (pro-R)

and (2S,3S)-[2,3-2H2] Fmoc-N-aspartic acid β-tert-butyl ester (pro-S).

We confirmed the selective deuterium incorporation in both labeled aspartates by 1H

NMR after converting their side chain carboxylic acids to tert-butyl esters and protecting their

amine groups with Fmoc. Displayed in panels A, B, and C of Figure 4-8 are the expanded

regions (2.7 to 4.7 ppm) of the 1H NMR spectra of the unlabeled, pro-R labeled, and pro-S

labeled protected aspartates, respectively, where the C2 and C3 proton signals are observed. The

C3 hydrogens of unlabeled Fmoc-Asp-β-OtBu in panel A exhibit both geminal coupling and

vicinal coupling to the hydrogen on C2, resulting in two sets of doublets of doublets. The doublet

of doublets corresponding to the pro-R hydrogen is centered at 2.77 ppm, while that of the pro-S

hydrogen is observed at 2.98 ppm. In panel B, the replacement of the pro-R hydrogen with

deuterium results in the disappearance of the doublet of doublets at 2.77 ppm. Additionally, one

of the doublets at 2.98 ppm that was present due to geminal coupling is now absent, confirming

that the pro-R deuterium was indeed retained. Similarly, in panel C, the replacement of the pro-S

hydrogen at C3 and the hydrogen at C2 with deuterium eliminated geminal and vicinal proton

coupling and resulted in the collapse of the doublet to a single peak at 2.77 ppm, corresponding to

126

the C3 pro-R hydrogen. The absence both of the proton signal at 2.98 ppm and of proton coupling

confirmed the presence of deuterium at C2 and in the pro-S position at C3.

Figure 4-6. 1

H NMR spectra from 2.7 to 4.7 ppm of unlabeled (A), pro-R labeled (B) and pro-S

labeled (C) aspartate with its amino and β-carboxylic acid moieties protected with Fmoc and tert-

butyl ester groups, respectively.

Catalysis by RimO is believed to involve H atom abstraction from C3 of aspartate by a

5′-dA• generated from the reductive cleavage of SAM. To determine the overall stereochemical

course of the RimO reaction, the stereoselectively labeled aspartates were appropriately protected

for solid phase peptide synthesis and incorporated at the target position of the 13-mer peptide

derived from the Bt S12 protein. The peptide containing the C3 pro-R aspartate (2) and the

peptide containing the C3 pro-S aspartate (3) were used as substrates in reactions with Bt RimO

127

to determine which of the two peptides supports the formation of 5′-deoxyadenosine enriched

with deuterium (5′-dAD), thereby indicating which of the H atoms attached to C3 is abstracted.

Panels A, B, and C of Figure 3 show the time-dependent formation of 5′-dAH, 5′-dAD, and

methylthiol-containing product (3-MS) obtained using peptides 1, 2, or 3, respectively. Formation

of 5′-dAD is only observed with peptide 3 containing the C3 pro-S-labeled aspartate, indicating

that the 5′-dA• abstracts the pro-S H atom from the substrate and that the overall RimO reaction

proceeds with inversion of configuration. A primary kinetic isotope effect of ∼1.9 was calculated

from the ratios of the rates of 5′-dAH formation with peptide 1 (5.4 ± 0.1 μM min−1

) and 5′-dAD

+ 5′-dAH formation with peptide 3 (2.9 ± 0.1 μM min−1

), indicating that H atom abstraction is at

least partially rate-limiting. Furthermore, there appears to be a substantial secondary isotope

effect (∼1.4) in the reaction using the 3-pro-R-labeled substrate (2), given the rates of 5′-dAH

formation with peptide 1 (5.4 ± 0.1 μM min−1

) versus peptide 2 (3.7 ± 0.1 μM min−1

).

Figure 4-7. Bt RimO catalyzed reactions at 37 ˚C in the presence of Ec Fld/FldR/NADPH, SAM

and peptides 1 (A), 2 (B), or 3 (C). The reactions were conducted as described in the supporting

information and contained 100 µM Bt RimO, 3 mM SAM, 3 mM peptide (1, 2, or 3 as indicated),

200 µM Fld, 50 µM FldR, 3 mM NADPH, and 50 mM Na-HEPES pH 7.5.

The source of sulfur that is inserted in the RimO methylthiolation reaction has yet to be

definitively determined; however, it has been proposed that the auxiliary N-terminal [4Fe−4S]

cluster serves this role (6) or serves as a binding site for a sulfide species that is methylated and

subsequently inserted into the substrate in a radical-dependent process (18,19). To determine

128

whether methionine, a byproduct formed during the reductive cleavage of SAM, is used by RimO

as a source of sulfur in the methylthio group, the concentrations of methionine and 5′-dAH

formed in a reaction of Bt RimO incubated under turnover conditions with 1 and the Ec

Fld/FldR/NADPH reducing system were quantified and compared (Figure 4-9). Methionine and

5′-dAH were formed in a 1:1 ratio, thereby ruling out methionine as a source of

sulfide/methylthio group for the RimO reaction.

Figure 4-8. Quantification of methionine generated in the Bt RimO reaction. The reaction was

conducted as described above and contained 150 µM Bt RimO, 1.5 mM SAM, 1.5 mM peptide

(1), 150 µM Ec Fld, 150 µM Ec FldR, 1 mM NADPH, and 50 mM Na-HEPES pH 7.5. The lines

are fits to a first-order exponential equation for 5'-dAH, product, and methionine, and the line for

SAH was fit to a second-order exponential equation with the following kinetic parameters: 5'-

dAH formation: A = 58.8 + 2.6 µM, v = 1.1 + 0.1 µM min-1

; product formation: A = 56.4 + 2.6

µM, v = 1.2 + 0.1 µM min-1

; methionine formation: A = 53.6 + 6.5 µM, v = 1.1 + 0.1 µM min-1

;

SAH formation: A1 = 19.5 + 1.7 µM, v1 = 23.6 + 2.0 µM min-1

; A2 = 22.9 + 1.3 µM, v2 = 0.8 + 0.1

µM min-1

.

Discussion

Previous studies of Tm RimO conducted by two different laboratories shed light on the

mechanism by which this enzyme catalyzes methylthiolation of D89 (6,7,18,19) Contrary to the

129

initial proposed mechanism, based on the mechanisms of the RS sulfurtransferases BioB (15, 26-

31) and LipA, (11, 32) in which the auxiliary clusters of these enzymes were shown to be the

sacrificial source of sulfide, RimO has been shown to synthesize an −SCH3 group that is

presumably bound to the unique Fe ion of its auxiliary cluster (18,19). Subsequent generation of

the 5′-dA• for H atom abstraction, now known from this study to be the pro-S H atom, generates

a substrate-based radical with which the synthesized −SCH3 group presumably combines to form

the methylthiolated product. Although the details of the attachment of the methylthio group onto

the substrate remain elusive, the determination of both the stereospecificity of H atom abstraction

and the absolute configuration at C3 of 3-MS-D89 allows us to conclude that −SCH3 insertion

occurs with inversion of configuration. Together, these results make RimO the third RS enzyme

for which the stereochemical outcomes of H atom abstraction and sulfur/methylthio group

insertion is known (12,17).

The finding that the Ec Fld/FldR/NADPH reducing system acts as a competent source of

reducing equivalents for the Bt RimO reaction led us to believe that it would allow demonstration

of the expected product ratios of 1:1:1 (methylthiolated product/SAH/5′-dAH). Surprisingly, 5′-

dAH and SAH formation in excess of product was still observed, mirroring the results that we,

and others, have reported for Tm RimO (7,18,19). In previous studies of the Tm RimO reaction in

which dithionite was used, formation of both SAH and 5′-dAH in 2-fold or greater excess of

product was observed, which was attributed to abortive cleavage of SAM, a known side reaction

of RS enzymes (7,18,19). The use of the Ec Fld/FldR/NADPH reducing system did decrease the

amount of SAH and 5′-dAH formed per methylthiolated product; however, abortive cleavage still

resulted.

Reactions conducted in D2O demonstrated that solvent and solvent exchangeable H

atoms do not quench any 5′- dA• formed productively or abortively, suggesting that abstraction

of an H atom derived from the RimO polypeptide or the peptide substrate occurred. Future studies

130

will address the questions concerning the overall stoichiometry of reactants and products and the

source of sulfide in the methylthiolation reaction.

131

References

1. Anton, B. P.; Saleh, L.; Benner, J. S.; Raleigh, E. A.; Kasif, S.; Roberts, R. J. Proc. Natl.

Acad. Sci. USA 2008, 105, 1826.

2. Brodersen, D. E.; Clemons Jr, W. M.; Carter, A. P.; Wimberly, B. T.; Ramakrishnan, V.

J. Mol. Biol. 2002, 316, 725.

3. Noeske, J.; Wasserman, M. R.; Terry, D. S.; Altman, R. B.; Blanchard, S. C.; Cate, J. H.

D. Nat. Struct. Mol. Biol. 2015, 22, 336.

4. Polikanov, Y. S.; Melnikov, S. V.; Söll, D.; Steitz, T. A. Nat. Struct. Mol. Biol. 2015, 22,

342.

5. Akiva, E.; Brown, S.; Almonacid, D. E.; Barber, A. E.; Custer, A. F.; Hicks, M. A.;

Huang, C. C.; Lauck, F.; Mashiyama, S. T.; Meng, E. C.; Mischel, D.; Morris, J. H.;

Ojha, S.; Schnoes, A. M.; Stryke, D.; Yunes, J. M.; Ferrin, T. E.; Holliday, G. L.; Babbitt,

P. C. Nucleic Acids Res. 2014, 42, D521.

6. Lee, K.-H.; Saleh, L.; Anton, B. P.; Madinger, C. L.; Benner, J. S.; Iwig, D. F.; Roberts,

R. J.; Krebs, C.; Booker, S. J. Biochemistry 2009, 48, 10162.

7. Arragain, S.; Garcia-Serres, R.; Blondin, G.; Douki, T.; Clemancey, M.; Latour, J.-M.;

Forouhar, F.; Neely, H.; Montelione, G. T.; Hunt, J. F.; Mulliez, E.; Fontecave, M.; Atta,

M. J. Biol. Chem. 2010, 285, 5792.

8. Cicchillo, R. M.; Iwig, D. F.; Jones, A. D.; Nesbitt, N. M.; Baleanu-Gogonea, C.; Souder,

M. G.; Tu, L.; Booker, S. J. Biochemistry 2004, 43, 6378.

9. Cicchillo, R. M.; Lee, K.-H.; Baleanu-Gogonea, C.; Nesbitt, N. M.; Krebs, C.; Booker, S.

J. Biochemistry 2004, 43, 11770.

10. Douglas, P.; Kriek, M.; Bryant, P.; Roach, P. L. Angew. Chem. Int. Ed. 2006, 45, 5197.

11. Lanz, N. D.; Pandelia, M.-E.; Kakar, E. S.; Lee, K.-H.; Krebs, C.; Booker, S. J.

Biochemistry 2014, 53, 4557.

12. Parry, R. J.; Trainor, D. A. J. Am. Chem. Soc. 1978, 100, 5243.

13. Cosper, M. M.; Jameson, G. N. L.; Hernández, H. L.; Krebs, C.; Huynh, B. H.; Johnson,

M. K. Biochemistry 2004, 43, 2007.

14. Escalettes, F.; Florentin, D.; Tse Sum Bui, B.; Lesage, D.; Marquet, A. J. Am. Chem. Soc.

1999, 121, 3571.

15. Fugate, C. J.; Stich, T. A.; Kim, E. G.; Myers, W. K.; Britt, R. D.; Jarrett, J. T. J. Am.

Chem. Soc. 2012, 134, 9042.

16. Taylor, A. M.; Stoll, S.; Britt, R. D.; Jarrett, J. T. Biochemistry 2011, 50, 7953.

17. Trainor, D. A.; Parry, R. J.; Gitterman, A. J. Am. Chem. Soc. 1980, 102, 1467.

18. Forouhar, F.; Arragain, S.; Atta, M.; Gambarelli, S.; Mouesca, J.-M.; Hussain, M.; Xiao,

R.; Kieffer-Jaquinod, S.; Seetharaman, J.; Acton, T. B.; Montelione, G. T.; Mulliez, E.;

Hunt, J. F.; Fontecave, M. Nat. Chem. Biol. 2013, 9, 333.

19. Landgraf, B. J.; Arcinas, A. J.; Lee, K.-H.; Booker, S. J. J. Am. Chem. Soc. 2013, 135,

15404.

20. Grove, T. L.; Benner, J. S.; Radle, M. I.; Ahlum, J. H.; Landgraf, B. J.; Krebs, C.;

Booker, S. J. Science 2011, 332, 604.

21. Bauerle, M.; Schwalm, E.; Booker, S. J. J. Biol. Chem. 2014, 290, 3995.

22. Cicchillo, R. M.; Lee, K.-H.; Baleanu-Gogonea, C.; Nesbitt, N. M.; Krebs, C.; Booker, S.

J., Biochemistry 2004, 43, 11770.

23. Gani, D.; Young, D. W. J. Chem. Soc., Perkin Trans. 1 1983, 2393.

24. Richards, E. M.; Tebby, J. C.; Ward, R. S.; Williams, D. H. J. Chem. Soc. C. 1969, 1542.

132

25. Lanz, N. D.; Grove, T. L.; Gogonea, C. B.; Lee, K.-H.; Krebs, C.; Booker, S. J. Methods

Enzymol. 2012, 516, 125.

26. Berkovitch, F.; Nicolet, Y.; Wan, J. T.; Jarrett, J. T.; Drennan, C. L. Science 2004, 303,

76.

27. Tse Sum Bui, B.; Benda, R.; Schünemann, V.; Florentin, D.; Trautwein, A. X.; Marquet,

A. Biochemistry 2003, 42, 8791.

28. Tse Sum Bui, B.; Escalettes, F.; Chottard, G.; Florentin, D.; Marquet, A. Euro. J.

Biochem. 2000, 267, 2688.

29. Tse Sum Bui, B.; Lotierzo, M.; Escalettes, F.; Florentin, D.; Marquet, A. Biochemistry

2004, 43, 16432.

30. Tse Sum Bui, B.; Mattioli, T. A.; Florentin, D.; Bolbach, G.; Marquet, A. Biochemistry

2006, 45, 3824.

31. Ugulava, N. B.; Surerus, K. K.; Jarrett, J. T. J. Am. Chem. Soc. 2002, 124, 9050.

32. Cicchillo, R. M.; Booker, S. J. J. Am. Chem. Soc. 2005, 127, 2860.

Chapter 5

Assessment of Tm RimO activity with the Tm S12 protein as a substrate and

biochemical and biophysical characterization of Tm RimO active site variants

Introduction

Ribosomes are macromolecular complexes composed of ribosomal RNA (rRNA) and

proteins that bind mRNA and tRNA, among other molecules and proteins, to synthesize the

proteins necessary for life. The bacterial ribosome sediments during ultracentrifugation as a 70S

particle, which is composed of a small subunit (30S) and a large subunit (50S). The 30S subunit

is comprised of 16S rRNA and 21 small proteins named S1-S21, while the 50S subunit is

comprised of 23S and 5S rRNA and 33 proteins, named L1-L36 (1). These rRNAs and proteins

are embellished with post-transcriptional and post-translational modifications that expand the

chemical repertoire beyond that of the 4 nucleotides and 20 amino acids, with methylation being

the most common modification (2-5).

The S12 protein of the 30S subunit contains a unique post-translational modification:

methylthiolation of C3 of an aspartic acid residue (D89 in E. coli) (6, 7). Genetic, bioinformatic,

and biochemical studies identified the E. coli (Ec) gene product yliG as the enzyme responsible

for attachment of the methylthiol group onto C3 of D89 and was subsequently renamed RimO

(ribosomal modification O) (8). While D89 is absolutely conserved across all S12 homologues,

the methylthio- modification is nonessential, and Ec ΔrimO mutants are completely viable and

exhibit only a minor slow growth phenotype (8). Structural determinations of the ribosome

observed that the loop on which D89 resides projects into the acceptor site where tRNAs bind and

that the methythio group on D89 makes contact with m7G527 of 16S rRNA. While the exact

134

function of the modification is unknown, it is thought to play a role in maintaining translational

fidelity or in serving a function under stress or variable growth conditions (9-12).

RimO is a member of the radical SAM (RS) superfamily of enzymes, which all use an

electron donated from a [4Fe-4S]1+

cluster to reductively cleave the 5' C-S bond of S-adenosyl-L-

methionine (SAM) to generate methionine and a highly reactive 5'-deoxyadenosyl radical (5'-

dA•). Target hydrogen atoms (H•) on protein, nucleic acid, or small molecule substrates are

abstracted by this potent radical to initiate catalysis (13-17). RimO belongs to a subclass of RS

enzymes that catalyze sulfur insertion at unactivated C-H bonds, of which characterized members

include BioB, LipA, MiaB, and MtaB. BioB and LipA catalyze sulfur insertion into small

molecule precursors to synthesize biotin and the lipoyl cofactor, respectively, whereas RimO,

MiaB, and MtaB are believed to catalyze the methylation of a sulfur atom to synthesize a

methylthio group, which is subsequently inserted by radical-mediated reactions to generate their

respective methylthiolated products, making these enzymes methylthiotransferases (18-23). All

of the members of the sulfur-insertion subclass of RS enzymes harbor two Fe-S clusters: the RS

4Fe-4S cluster, ligated by a triad of cysteines residing in a signature Cx3Cx2C motif, and an

auxiliary cluster, which in the case of BioB is a 2Fe-2S cluster, and in the other enzyme members

is a second 4Fe-4S cluster (18, 24). The auxiliary clusters in BioB and LipA are the sacrificial

sources of sulfur atoms that are inserted into their substrates, while those found in RimO, MiaB,

and MtaB, are either thought to provide sulfur in a sacrificial capacity or to utilize the unique Fe

site for sulfide binding, thereby obviating cluster degradation (18-25). The exact role of the

auxiliary clusters in the methylthiotransferases (MTTases) remains to be definitively determined.

The MTTases contain a TRAM domain (tRNA methyltransferase 2 and MiaB) that is

unusually acidic and that has been shown to bind RNA in a diverse set of enzymes (26, 27).

MiaB and MtaB both modify adenosine 37 of different tRNA substrates, so the TRAM domain

likely aids in substrate binding in these enzymes. The presence of the TRAM domain is

135

intriguing, since this enzyme does not modify RNA, but rather a residue on a small, highly basic

S12 protein. While the nature of the substrates differ, it is likely that the acidic TRAM domain

also plays a role in binding the basic substrate in RimO. Whether the substrate is the lone S12

protein, the 30S ribosome, or the fully assembled 70S ribosome remains to be determined.

Studies using the S12 protein as a substrate have been precluded by its insolubility when

overproduced in E. coli (19, 25).

Some of the mechanistic details of the methylthiolation reactions catalyzed by the RS

MTTases have been elucidated. Both RimO and MiaB have been demonstrated to catalyze

methyl transfer from SAM to an acceptor site on the proteins through radiolabel tracing

experiments. The acceptor site is thought to be a sulfide ion bound to the unique iron site of the

auxiliary cluster, or a sulfide ion within the cluster itself to form a methylated cluster

intermediate, which was shown to be both kinetically and chemically competent (21). In the

presence of exogenous sulfide, methanethiol, or methaneselenol, both RimO and MiaB are

capable of catalyzing more than one turnover, suggesting that these small molecules are activated

for their incorporation into the final products of both enzymes (20, 21). A recent crystal structure

of Tm RimO at 3.3 Å resolution showed the two Fe-S clusters to be ~ 8 Å apart, and, intriguingly,

the clusters were linked by a chain of electron density that was modeled well as a pentasulfide

bridge (20). While the physiological relevance of the pentasulfide bridge is questionable, it is

clear that the active site of RimO can accomodate and bind sulfide species that could be used in

the methylthiolation reaction. It remains to be established whether MTTases bind sulfide for use

in the reactions they catalyze or whether the auxiliary cluster is sacrificed, as in BioB and LipA,

as the source of sulfide.

Herein, we describe the purification of the S12 protein from Thermotoga maritima (Tm)

overproduced in E. coli and its use as a substrate in the methylthiolation reaction catalyzed by Tm

RimO. We find that the Tm S12 protein is a competent substrate that triggers formation of 5'-

136

dAH and is indeed found to have an appended methylthio group upon reaction with RimO as

evidenced by MALDI-TOF analysis. Using the previously determined structure of Tm RimO and

aligning a wide array of bacterial RimO sequences, we identified conserved amino acids in the

enzyme active site—K12, Q192, and Y227—and subsequently constructed protein variants and

characterized their abilities to catalyze methyl transfer and formation of 5'-dAH and

methylthiolated product. We find the K12A and K12Q variants to exhibit decreased methyl

transfer ability and to be unable to form any detectable amounts of 5'-dAH or methylthiolated

product. Y227A and Y227F variants catalyze formation of 5'-dAH, but do not catalyze

methylthiolation. The Q192A variant catalyzed methyl transfer and formation of 5'-dAH and

methylthiolated product, albeit it to lower extents and with slower rates than those observed with

the wild type enzyme. Lastly, we determined the dissociation constants of SAM binding to Tm

RimO wild type, K12A, and Y227F by isothermal titration calorimetry (ITC) and found all three

proteins to bind SAM with low micromolar affinities.


Materials


MA). L-tryptophan, 2-mercaptoethanol, L-(+)-arabinose, ferric chloride, sodium methanethiolate,

5’-deoxyadenosine (5’-dA), sodium sulfide nonahydrate, phenylmethylsulfonyl fluoride and S-

adenosyl-L-homocysteine (SAH) were purchased from Sigma Corp (St. Louis, MO). N-(2-

hydroxyethyl)piperizine-N'-(2-ethanesulfonic acid) (HEPES) was purchased from Fisher

Scientific (Pittsburgh, PA), and imidazole and urea were purchased from J. T. Baker Chemical

Co. (Phillipsburg, NJ). Potassium chloride, glycerol, and expression vector pET-28a were

137

purchased from EMD Chemicals (Gibbstown, NJ), while dithiothreitol (DTT) and isopropyl-β-D-

1-thiogalactopyranoside (IPTG) were purchased from Gold Biotechnology (St. Louis, MO).

Coomassie blue dye-binding reagent for protein concentration determination was purchased from

Pierce (Rockford, IL), as was the bovine serum albumin standard (2 mg/mL). PD-10 pre-poured

gel-filtration columns as well as Sephadex G-25 resin were purchased from GE Biosciences

(Piscataway, NJ). Egg white lysozyme was purchased from Alfa Aesar (Ward Hill, MA). All

other buffers and chemicals were of the highest grade available.

Methods

Cloning and overexpression of the Tm rpsL (S12) gene

The Tm S12 gene was amplified from Tm genomic DNA using polymerase chain reaction

(PCR) technology. The forward amplification primer (5’- CGC GGC GTC CAT ATG CCA ACG

ATC AAT CAA TTG ATC AGG TAC G -3’) included an NdeI restriction site (underlined)

flanked by a nine-base GC clamp and the first 28 bases of the S12 gene. The reverse primer (5'-

CGC GGC GTC GAA TTC TCA CTT CTT TTG ATC CTT GGG TCT TTT CGC- 3') contained

an EcoRI restriction site (underlined) flanked by a nine-base GC clamp and the last 30 bases of

the S12 gene, including the stop codon. The PCR was performed with a Stratagene Robocycler

thermocycler (La Jolla, CA) as described previously (28). The product was isolated and digested

with NdeI and EcoRI and ligated into similarly digested pET-28a by standard procedures (29).

The correct construct, encoding a 10 amino acid linker between the gene product and an N-

terminal hexahistidine tag, was verified by DNA sequencing and designated pTmS12.

138

Expression vector pET28a-TmS12 was transformed into Ec BL21(DE3) for gene

expression. Bacterial growth and gene expression was carried out at 37 °C in 16 L of Luria-

Bertani media distributed evenly among 4 Erlenmeyer flasks with moderate shaking (180 rpm).

At an optical density at 600 nm of 0.6, a sterile solution of IPTG was added to each flask to a

final concentration of 1 mM. Expression was allowed to take place for 5 h at 37 °C before the

cells were harvested by centrifugation at 10,000 g for 10 min at ambient temperature.

Purification of Tm S12

Purification of Tm S12 was carried out by immobilized metal affinity chromatography

(IMAC) using Ni-NTA resin at 4˚C (Qiagen, Valencia, CA). Because the majority of the

overproduced protein was found in inclusion bodies, as well as the fact that the protein exhibits a

relatively simple teriatry structure and contains no cofactors, a strategy to purify it under

denaturing conditions and then refold it was formulated. Buffers used during the purification of

Tm S12 were as follows: lysis buffer (50 mM HEPES, pH 7.5, 500 mM KCl, 10 mM 2-

mercaptoethanol, 20 mM imidazole, 1 mM PMSF and 1 mg/mL lysozyme); solubilization buffer

(lysis buffer with 10% (v/v) glycerol and 6 M urea and lysozyme omitted), wash buffer

(solubilization buffer with 40 mM imidazole); refolding buffer (50 mM HEPES, pH 7.5, 500 mM

KCl, 10 mM 2-mercaptoethanol, 10% glycerol); elution buffer (refolding buffer containing 20%

(v/v) glycerol and 500 mM imidazole). After lysing the cells by sonication, the cell suspension

was transferred into sterile centrifuge tubes, which were subsequently centrifuged at 50,000 g at

4 °C for 30 min. The supernatant was transferred to a sterile bottle and stored at 4 °C, the

pelleted cell debris and inclusion bodies were removed, placed in solubilization buffer, and stirred

over night at 4° C. The next day, the resuspended cell pellet was placed in sterile centrifuge tubes

and centrifuged at 50,000 g at 4 °C for 30 min; the denatured supernatant was combined with

139

the first supernatant and stirred with Ni-NTA resin for 4 h prior to loading the resin-supernatant

mixture into a column. The resin was then washed with 250 mL of wash buffer. The protein was

refolded on the resin by slow removal of urea, which was achieved by flowing a linear gradient of

750 mL of wash buffer and 750 mL of refolding buffer over the resin bed; an additional 250 mL

of refolding buffer was flowed over the resin to ensure complete removal of urea. After addition

of elution buffer to the column, protein-containing fractions, as determined by UV-Vis

spectrophotometric analysis at 280 nm on a Cary 50 spectrophotometer (Varian, Walnut Creek,

CA), were pooled and concentrated using an Amicon stirred ultrafiltration apparatus (Millipore,

Billerica, MA) fitted with a YM-3 membrane (3,000 molecular weight cutoff). The protein was

exchanged into gel-filtration buffer (50 mM HEPES, pH 7.5, 500 mM KCl, 10% glycerol, using a

Sephadex G-25 column (2.5 13 cm), concentrated again and further purified by fast protein

liquid chromatography (FPLC) on a 16/60 Hi-Prep S-200 column using an ÄKTA liquid

chromatography system (GE Biosciences) housed in an anaerobic chamber. The column was

equilibrated in gel-filtration buffer. Fractions were pooled based on their absorbance at 280 nm

and concentrated. The protein concentration was determined by UV-Visible spectrophotometric

analysis at 275 nm using a calculated extinction coefficient (ε280 = 5960 M-1

cm-1

) based on the

presence of four tyrosyl residues.

Activity assays with Tm S12

Assessment of activity of Tm RimO with Tm S12 as a substrate was conducted as

previously described (21), except that the concentrations of Tm RimO and Tm S12 were 200 and

150 µM, respectively, and the final concentration of SAM was 2 mM.

140

MALDI-TOF analysis of the Tm RimO reaction with Tm S12

A reaction containing 200 µM Tm RimO, 150 µM Tm S12, and 2 mM dithionite in 50

mM Na-HEPES, pH 7.5, was initiated with the addition of 2 mM SAM. Control reactions were

conducted simultaneously in which SAM and dithionite, just SAM, or just dithionite were

omitted. The reactions were lyophilized, diluted in water, and subsequently exchanged into water

by multiple rounds of dilution and concentration using a Microcon YM-3 ultrafiltration device (3

kDa molecular weight cutoff). MALDI-TOF mass spectra were acquired on a Bruker

Ultraflextreme mass spectrometer operated in linear positive-ion mode. Samples were prepared

by mixing 1 µL of protein (~1 mg/mL) with 1 µL of 1% trifluoroacetic acid and 2 µL of a matrix

solution, which was prepared by dissolving 10 mg 4-chloro-α-cyanocinnamic acid (Sigma, St.

Louis, MO) in 30% aqueous acetonitrile containing 1% trifluoroacetic acid. 1 µL of this mixture

was applied to a brushed steel target and allowed to dry. The instrument was calibrated using

Bruker Clinprot standard mixture, and the data were acquired using the factory-configured default

parameters for 5-20 kDa range.

Site-directed mutagenesis, overexpression, and purification of Tm RimO variants

The genes for the Tm RimO K12A, K12Q, Y227F, Y227A, and Q192A variant proteins

were constructed using the Stratagene QuikChange II site-directed mutagenesis kit (Agilent

Technologies, Santa Clara, CA) according to the manufacturer’s specifications and as described

previously (28) . The forward and reverse primers used to construct each variant gene are listed

in Table 5-1, with base changes underlined. These primers were added to a typical QuikChange II

reaction mixture to a final concentration of 20 µM with 100 ng of pET28A Tm RimO template

DNA. 15 cycles of the following program were initiated: 95 °C for 1 min, 55 °C for 1 min, and 68

141

°C for 10 min. Upon completion, the reaction mixture was incubated for 15 min at 68 °C before

being cooled to 4 °C. Subsequent to this step, the procedure followed the manufacturer’s

specifications. The correct mutations were verified by DNA sequencing, and Ec BL21(DE3) cells

were transformed with the resulting plasmids. Overexpression and purification of the Tm RimO

variant gene products was conducted as described previously for Tm RimO wild type (21).

Table 5-1. The forward and reverse primers used to make the indicated amino acid substitutions

in Tm RimO by site-directed mutagenesis with the changed nucleotides highlighted in red.

Quantitative iron and sulfide analyses and concentration determination of Tm RimO

variants

Iron and sulfide analyses were performed according to the procedures of Beinert (30-32)

as described previously (21, 33). Protein concentrations were determined by the method of

Bradford (34) using the correction factor of 1.46 that was previously determined by quantitative

amino acid analysis (21).

Activity and methyl transfer assays with Tm RimO variants

Assessment of methylthiolation and methyl transfer activity exhibited by Tm RimO

variants was conducted as previously described (21) unless noted otherwise. Assays conducted in

142

D2O were at least ~60% D2O by volume, and all components except SAM were incubated in D2O

for at least one hour at room temperature to allow for hydrogen-deuterium exchange.

Determination of dissociation constants for SAM or SAM analogues with Tm RimO wild

type and active site variants by isothermal titration calorimetry

Isothermal titration calorimetry (ITC) experiments were conducted with a MicroCal VP-

ITC (Malvern Instruments., Malvern, Worcestershire, UK) housed in an MBraun anaerobic

glovebox (Stratham, NH) kept under an atmosphere of N2 with the concentration of O2

maintained below 1 ppm. The ITC cell contained 1.4 mL of the indicated protein in ITC buffer

(50 mM Na-HEPES, pH 7.5, 300 mM KCl). The analyzed proteins were exchanged extensively

into ITC buffer by gel-filtration using pre-poured PD-10 columns. After applying the protein to

the column,the eluate was isolated and concentrated using centrifugal spin filters and then

reapplied to a newly equilibrated column. This process was then repeated two additional times.

The syringe contained SAM or SAM analogue diluted in ITC buffer to the indicated

concentration. After a 2 µL initial injection and an initial delay of 60 s, 5 µL of titrant was

injected at 480 s intervals. The cell was stirred at 242 RPM and maintained at 25 °C with a

reference power of 10 µCal/s. The instrument was operated in automatic fast equilibration and

high feedback modes.

Results

Cloning and overexpression of the Tm S12 gene

The gene encoding Tm S12 was cloned into the NdeI and EcoRI restriction sites of

expression vector pET28a, which yields a protein containing an N-terminal hexahistidine tag and

143

a 10 amino acid linker preceding its natural start codon. The resulting plasmid, pET28a-TmS12,

was transformed into BL21 (DE3) cells. Induction of expression at 37 °C in Luria-Bertani media

for 5 h resulted in the greatest yield of protein as monitored by SDS-PAGE (Figure 5-1).

Figure 5-1. SDS-PAGE of the overexpression of the S12 gene from Thermotoga maritima in E.

coli BL21(DE3) cells. Lane 1: molecular weight markers (in kDa); lanes 2-5: cell culture samples

prior to induction of pET28a-TmS12 plasmid with IPTG; lanes 6-9: cell culture samples after

induction with IPTG. The molecular weight of the pET28a-TmS12 gene product is 16.2 kDa. The

protein runs higher than its molecular weight due to its high pI of 11.5.

Purification of Tm S12

All previous attempts to purify Tm S12 overproduced in E. coli cells failed to yield any

appreciable amounts of soluble protein; it was found exclusively in inclusion bodies when

overexpressed in cells cultured in Luria-Bertani broth or M9 minimal media at 37 °C and 18 °C

with varied concentrations of IPTG. To purify the protein from the inclusion bodies, a

denaturation step was employed that has been used previously to purify insoluble proteins (35).

The cell pellets were resuspended in 8 M urea and any soluble protein was allowed to bind to Ni-

NTA resin. After washing and slow removal of the urea to allow the protein to refold, soluble Tm

S12 was eluted with 500 mM imidazole (Figure 5-2A). Subsequent purification by size-exclusion

chromatography resulted in nearly homogenous protein (Figure 5-2B). The concentration of

144

isolated Tm S12 was determined by UV-visible spectrophometric analysis at 280 nm. The

approximate yields of protein ranged from 2 to 3.5 mg / L of cell culture.

Figure 5-2. SDS-PAGE of the purification of Tm S12 under denaturing conditions (A). Lane 1:

molecular weight markers (in kDa); lane 2: cell pellet; lane 3: blank; lane 4: the combined

supernatant from cell lysis and the supernatant of solubilized inclusion bodies, with Tm S12

indicated by the black arrow; lane 5: flow through from the Ni-NTA resin; lane 6: Ni-NTA resin

wash; lane 7: half way through the refolding step; lane 8: the end of the refolding step; lane 9:

combined eluate fractions; lane 10: concentrated eluate fractions. SDS-PAGE of Tm S12 after

size-exclusion chromatrography (B). Lane 1: molecular weight markers (in kDa); lane 2:

combined eluate fractions from the size-exclusion column.

Tm RimO activity assays with Tm S12

In previous studies the activity of Tm RimO was examined with 13- or 20-amino acid

peptide substrates—corresponding to amino acids surrounding the aspartate residue that RimO

modifies—due to failed attempts at isolating soluble Tm S12 protein (20, 21, 25). To assess the

activity of RimO with its full-length protein substrate, Tm S12, isolated under denaturing

conditions and subsequently refolded, was incubated with Tm RimO in the presence or absence of

SAM and sodium dithionite; the time-dependent formation of SAH and 5'-dAH was determined

145

by LC/MS, while the modification of the full length Tm S12 substrate was qualitatively assessed

by MALDI-TOF mass spectrometry. Shown in Figure 5-3 are the mass spectra of the Tm S12

protein under these assay conditions. As can be seen in panels A-C, the mass of Tm S12 was

16,091 Da at both t = 0 and t = 60, which was expected since these assays lacked components

required for the methylthiolation reaction. In panel D, the red spectrum features a peak at 16,136

Da that is not present at t = 0 and which corresponds to +45 Da greater than the unmodified Tm

S12 mass. The attachment of a methylthio- group results in an overall mass increase of +46 Da,

which is 1Da greater than the observed mass but within the mass error associated with the

calibration standards used. LC/MS analysis of the complete reaction allowed for quantification of

the time-dependent formation of SAH and 5'-dAH as shown in Figure 5-4. The formation of

SAH in the presence of the S12 protein substrate was best fitted by a first-order exponential

function (equation 1), wherein A is the amplitude, k is the rate constant, and t is time, with the

following parameters: A = 671 + 125 µM; k = 0.01 + 0.004 min-1

. The initial rate of formation (v

= 6.7 + 0.5 µM min-1

) was determined by fitting the linear portion of the data. The formation of

5'-dAH under the same conditions was best fitted to a second-order exponential function

(equation 2) with the following parameters: A1 = 198 + 13 µM; k1 = 0.03 + 0.01 min-1

; A2 = 76 +

18 µM; k2 = 0.3 + 0.1 min-1

. The initial rate (ν) was determined to be 17.8 + 7.9 µM min-1

. A

comparison of the amplitudes and initial rates corresponding to SAH and 5'-dAH formation with

the S12 protein to those obtained with the S12 peptide (vide supra) showed that both SAH and 5'-

dAH initial rates are 2- to 3-fold lower in the presence of the S12 protein. Moreover, the amount

of SAH formed with the protein substrate is twice that with peptide substrate. Disappointingly,

the full-length protein substrate does not appear to reduce the aberrant formation of SAH or the

abortive cleavage of SAM to form 5'-dAH with dithionite as the reductant. Nevertheless, these

results provide the first in vitro evidence that the Tm S12 protein is a substrate for Tm RimO.

(1)

146

(2)

Figure 5-3. MALDI-TOF mass spectra of Tm S12 that was incubated with Tm RimO in the

absence of SAM and dithionite (A); in the absence of dithionite (B); in the absence of SAM (C);

in the presence of both SAM and dithionite (D). The black spectra are from samples at t = 0; the

red spectra at t = 60 minutes. Only when all components required for methylthiolation are present

was a mass shift of +45 Da observed, corresponding to Tm RimO catalyzed methylthiolation of

the S12 protein.

147

Figure 5-4. Time-dependent formation of SAH (blue) and 5'-dAH (red) by 200 µM Tm RimO in

the presence of 2 mM SAM, 2 mM sodium dithionite, and 150 µM Tm S12 protein. Data for 5'-

dAH formation were fitted to a second-order exponential equation and that of SAH were fit to a

first-order exponential equation.

Identification of conserved active site residues from sequence alignments and the Tm RimO

crystal structure

Sequence alignments of RimO proteins from a wide variety of bacteria revealed residues

that are strictly conserved and likely to play important roles in the structure and function of this

enzyme (Figure 5-5). Of the conserved residues identified, three were found to reside directly in

the purported active site in the crystal structure of Tm RimO: lysine 12 (K12), glutamine 192

(Q192), and tyrosine 227 (Y227) (Figure 5-6) (20). Based on the results from RimO sequence

alignments and the crystal structure of Tm RimO depicting these residues near its two 4Fe-4S

clusters, RimO proteins in which one of these amino acids was substituted with alanine,

glutamine, or phenylalanine were constructed in an attempt to determine the role(s) of these

conserved amino acids in the methylthiolation reaction. Specifically, Tm RimO variant proteins

148

K12A, K12Q, Y227A, Y227F, and Q192A were created through site-directed mutagenesis as

described above.

Figure 5-5. Sequence alignment of RimO proteins from 11 different bacterial species. Strictly

conserved residues are highlighted in color, with K12, Q192, and Y227 colored purple, pink, and

teal, respectively.

149

Figure 5-6. Active site from the crystal structure of Tm RimO. Conserved amino acids Lys 12,

Gln 192, and Tyr 227 are shown in purple, pink, and teal, respectively. A pentasulfide chain

linking the unique iron sites of each 4Fe-4S cluster is shown in yellow sticks, while the sulfide

and iron ions of the clusters are depicted as yellow and orange spheres, respectively. Cysteine

residues ligating the auxiliary cluster are shown as cyan sticks, while those ligating the RS cluster

are shown as red sticks.

Overexpression, purification, and characterization of Tm RimO variants

Overexpression of the plasmids harboring the mutated Tm RimO genes (pTmRimOK12A,

pTmRimOK12Q, pTmRimOY227A, pTmRimOY227F, pTmRimOQ192A) was conducted as

described for Tm RimO WT, as was the purification of each gene product (21), resulting in

relatively pure proteins as assessed by SDS-PAGE (Figure 5-7). UV-Visible spectrophotometric

analysis of each variant protein showed that the amino acid substitutions had little to no effect on

the feature at ~ 400 nm corresponding to the two [4Fe-4S]2+

clusters (Figure 5-8). Similarly,

quantitative iron and sulfide analyses of each variant confirmed that the amino acid substitutions

150

had no major effects on the amount of iron and sulfide present, since iron content ranged from 6.2

- 7.1 Fe / polypeptide and acid-labile sulfide ranged from 4.3 to 8.0 S / polypeptide (Table 5-2).

Figure 5-7. SDS-PAGE of Tm RimO variants following purification by IMAC and size-exclusion

chromatography. Y227A and Q192A are shown without molecular weight markers present.

151

Figure 5-8. UV-Visible spectra of Tm RimO WT and K12A, K12Q, Y227A, Y227F, and Q192A

variants normalized to the maximum absorbance at 280 nm. Little to no perturbations in the

feature at 400 nm of each spectrum were observed compared to those of WT, indicating that the

amino acid substitutions did not have any major effects on the two [4Fe-4S]2+

clusters.

Table 5-2. Results of quantitative iron and sulfide analyses of Tm RimO active site variants.

Variant Fe/Protein S/Protein

K12A RCN S200 6.2 + 0.1 6.8 + 0.2

K12Q RCN S200 7.1 + 0.1 8.0 + 0.1

Y227A RCN S200 7.1 + 0.1 4.3 + 0.2

Y227F RCN S200 6.2 + 0.1 5.9 + 0.2

Q192A RCN S200 6.7 + 0.1 7.7 + 0.2

152

Assessment of methyl transfer activity of Tm RimO variants

To determine the effect(s) of amino acid substitution on the methyl transfer reaction in

active site variants of Tm RimO, several variant proteins were constructed by site-directed

mutagenesis: Tm RimO K12A, K12Q, Y227A, Y227F, and Q192A. Following the isolation and

reconstitution of the variant proteins, the methyl transfer activity of each enzyme was determined

in the presence of SAM. Time-dependent formation of SAH, with an m/z value of 385.1, was

assessed by ESI+ LC/MS and compared to that of Tm RimO wild type. The RimO variants

catalyzed methyl transfer, albeit to varying extents, as evidenced by the time-dependent formation

of SAH (Figure 5-9A). Fitting of the data obtained from assays of Tm RimO WT and K12A with

a second-order exponential equation (equation 3) resulted in a lower rate constant for SAH

formation by the K12A variant compared to that of WT (Table 5-3). Specifically, the K12A

protein catalyzed relatively rapid formation of SAH in early time points with a comparable rate

constant (k1 = 0.26 + 0.08 min-1

) to that of the wild type enzyme (0.23 + 0.03 min -1

); however,

the amplitude of SAH formation by K12A was more than two-fold lower than that of wild type.

This rapid phase was followed by a slower phase with a corresponding rate constant (k2 = 0.01 +

0.001 min-1

) that is 20-fold lower than the first. Along these lines, a comparison of the initial

rates of SAH formation by K12A and wild type RimO, obtained from fitting the linear portion of

each curve, determined that the initial rate of SAH formation by the K12A variant was two-folder

slower than that of wild type (Figure 5-9B).

(3)

(4)

153

Figure 5-9. Time-dependent formation of SAH by 100 µM Tm RimO wild type and K12A,

K12Q, Y227F, Y227A, and Q192A variants in the presence of 1 mM SAM over 3 h (A) and

during the linear phase of the reaction (B). Data for SAH formation by the WT and K12A

enzymes were fitted to a second-order exponential equation and that of K12Q, Y227F, Y227A,

and Q192A were fit to a first-order exponential equation (A). All of the variants tested exhibited

methyl transfer activity to varying extents and rates of formation.

Table 5-3. Fit parameters of methyl transfer reactions containing 100 µM of the indicated RimO

protein and 1 mM SAM in 50 mM Na-HEPES, pH 7.5. Wild type and K12A data were fitted to a

second-order exponential equation and K12Q, Y227F, Y227A, and Q192A data were fitted to a

first-order exponential equation. Amplitude, A1 and A2; rate constant k1 and k2; initial rate, ν;

N.A., not applicable. Activity is reported in enzyme equivalents.

154

Fitting of the data obtained from methyl transfer assays of Tm RimO K12Q, Y227A,

Y227F, and Q192A with a first-order exponential equation (equation 4) resulted in rate constants

at least 10-fold lower than those associated with the rapid phase of SAH formation exhibited by

the WT and K12A enzymes. Moreover, the rate constants for SAH formation of the former group

of RimO variants are virtually identical to those corresponding to the second phase exhibited by

WT and K12A (Table 5-2). Additionally, the initial rates of SAH formation by the K12Q,

Y227A, Y227F, and Q192A were all approximately 7-fold lower than that determined for the

wild type enzyme. Interestingly, only the K12A and K12Q variants possessed significantly

impaired methyl transfer activity, and, somewhat surprisingly, substitution of lysine 12 with

glutamine, which more closely resembles lysine in terms of size and polarity, resulted in more

marked decreases in its associated rate constant, initial rate, and extent of SAH formation than

those associated with the K12A variant (Table 5-2).

Substitution of tyrosine 227 in Tm RimO with phenylalanine or alanine showed similar

perturbations in both the kinetic profiles and the extent of SAH formation. Again, this behavior

was somewhat surprising, since we hypothesized the Y227F variant to exhibit methyl transfer

activity comparable to that of WT since phenylalanine is similar in size and is also aromatic.

However, the fact that the methyl transfer activity of the Y227A variant was nearly identical to

that of Y227F argues that the relative bulk and aromaticity of the tyrosyl residue in the wild type

enzyme are not key properties for methyl transfer. What seems to be more likely is that the 4'-OH

group of tyrosine, which phenylalanine lacks, may make an important hydrogen bond in the

active site that promotes SAM-binding and/or methyl transfer. Lastly, the formation of SAH by

the Q192A variant was slower than that of the wild type enzyme; however, over the course of the

3 h reaction, the former catalyzed formation of SAH in slight excess of the latter, indicating that

Q192 likely plays a minor role in SAM-binding and/or methyl transfer.

155

Assessment of methylthiolation activity of Tm RimO variants

To determine the effect(s) of amino acid substitutions on the methylthiolation activity of

the active site variants of Tm RimO, each of the variant proteins—K12A, K12Q, Y227A, Y227F,

Q192A— and wild type RimO were incubated in reactions containing 1 mM SAM, 1 mM 13-mer

S12 peptide substrate, and 2 mM sodium dithionite as the reductant. ESI+ LC/MS analysis of

these reactions was used to quantify any time-dependent formation of 5'-dAH , SAH, and

methylthiolated peptide (MS-1) with m/z values of 252.1, 385.1, and 507.1, respectively (Figure

5-10).

Each of the RimO variants exhibited methyl transfer activity under turnover conditions to

various extents (Figure 5-10A), with parameters obtained from fits of the data reported in Table

5-4. These results were expected, since previous studies had shown the wild type enzyme to

produce SAH in excess of the enzyme concentration when dithionite was present, which indicated

that some of the SAH formed under turnover conditions was aberrant and did not result from

methyl transfer required for formation of the methylthiolated product (20, 21). While aberrant

formation of SAH in the presence of dithionite with the wild type enzyme is not well understood,

it is likely that the variants exhibit similar off-pathway production of SAH. Indeed, of the

variants tested, Tm RimO Y227A produced SAH in two-fold excess of the WT enzyme with an

associated initial rate that is also two-fold that of WT. Of the Y227A and Y227F variants,

formation of SAH by the latter more closely resembled that of WT; however, since the chemistry

behind the aberrant formation of SAH by the WT and variant enzymes in the presence of

dithionite is unknown, ascribing any role(s) to the conserved tyrosine in SAH formation under

turnover conditions, or attempting to explain the data with the Y227A and Y227F variants would

be purely speculative.

156

The formation of SAH by the K12A, K12Q, and Q192A variants under turnover

conditions was best fitted with linear equations, in contrast to that of the WT, Y227A, and Y227F

variants which were best fitted with first-order single exponential equations, indicating these

amino acid substitutions had an effect on the kinetic profile of SAH formation (Figure 5-10A).

The final concentration of SAH formed by the K12A and K12Q variants (239 and 345 µM,

respectively) was similar to that of WT (311 µM), but the initial rates of SAH formation by these

two variants were ~ 6-fold slower. The Q192A variant exhibited paltry formation of SAH (64

µM) compared to WT (608 µM) and the other variant enzymes; additionally, its associated initial

rate was 20-fold lower than that of WT. These results are in contrast to those from methyl transfer

assays (vide supra) wherein Q192A-catalyzed methyl transfer most closely resembled that of

WT. The discrepancy between the results obtained for methyl transfer in the presence of

dithionite by the Q192A variant and WT may suggest that Q192 is, in some way, involved in the

off-pathway reaction forming SAH abortively. Clearly, the presence of the chemical reductant

sodium dithionite causes formation of SAH in an off-pathway process that confounds with SAH

formation resulting from methyl transfer involved in the methylthiolation reaction.

The amino acid substitutions in the RimO variants had a greater effect on 5'-dAH

formation compared to the WT enzyme, as shown in Figure 5-10C and D. While the Y227F,

Y227A, and Q192A variants exhibited 5'-dAH formation to various extents that were all less than

that of WT (Table 5-4), the K12A and K12Q variants catalyzed 100-fold less 5'-dAH compared

to the wild type enzyme, making it clear that lysine 12 plays some role in 5'-dA• formation. The

fact that the substitution of this lysyl residue affected two different reactivities of SAM—methyl

transfer and 5'-dA• generation—suggests the enzyme uses the same binding site or finely tunes it

to exploit these two modes of reactivity.

The observation that the Y227F variant supported more robust 5'-dAH formation

compared to that of the Y227A variant suggests the presence of a larger and/or aromatic amino

157

acid at position 227 is more favorable for 5'-dA• formation, possibly by excluding solvent or

interacting with SAM to promote its reductive cleavage. Of all the variants tested, only Q192A

exhibited 5'-dAH formation that was best fitted by a linear equation, indicating the alanine

substitution affected the kinetic profile of 5'-dAH formation, and, to a certain degree, it also

affected the extent of formation of 5'-dAH compared to WT (163 and 608 µM, respectively).

Interestingly, the Q192A variant was the only one capable of producing methylthiolated

product, which was meager at best (11 µM) with an initial rate 55-fold slower than WT. Since the

Q192A protein catalyzed methyl transfer, 5'-dAH, and methylthiolated product, albeit it at slower

rates and to lower extents than those of wild type, Q192 residue likely plays a minor or largely

redundant role in the overall reaction, although its conservation does imply it is important in some

capacity.

158

Figure 5-10. Time-dependent formation of SAH (A & B), 5'-dAH (C & D), and MS-1 product (E

& F) by 100 µM Tm RimO wild type and K12A, K12Q, Y227F, Y227A, and Q192A variants in

the presence of 1 mM SAM, 1 mM 13 mer S12 peptide, 2 mM dithionite, and 50 mM Na-HEPES

pH 7.5. Data in panels A, C and E were fit, where appropriate, with either a first-order

exponential equation or a linear equation to obtain the parameters summarized in Table 5-3, while

data in panels B, D, and F were fit to a linear equation to obtain initial rates of formation reported

in Table 5-3. All of the variants demonstrated significantly decreased abilities to form 5'-dAH,

and only the Q192A variant exhibited any appreciable methylthiolation activity.

159

Table 5-4. Fit parameters of turnover reactions containing 100 µM of the indicated RimO protein,

1 mM SAM, and 1 mM 13 mer S12 peptide, and 50 mM Na-HEPES pH 7.5. Data were fit, where

appropriate, to a first-order exponential equation or a linear equation to obtain the associated

amplitude (A), rate constant (k), and initial rate, ν parameters for each reaction. Activity is

reported in enzyme equivalents. N.A., not applicable, since these data were fit with linear

equations; N.O., not observed.

Determination of dissociation constants for SAM or TeSAM by Tm RimO wild type and active

site variants by isothermal titration calorimetry

Isothermal titration calorimetry (ITC) measures small changes in the heat released or

absorbed when a ligand of interest binds to a target macromolecule; these changes in heat are

used to calculate the dissociation constant, or binding affinity, of the ligand by the

macromolecule. Accordingly, to determine the dissociation constants associated with SAM and/or

160

SAM analogues binding to Tm RimO wild type and active site variants of this enzyme, we

analyzed these proteins with SAM and its analogues via ITC under strict anaerobic conditions.

Specifically, Tm RimO WT was analyzed with SAM and Te-adenosyl-L-methionine (TeSAM), in

which the sulfur atom of SAM is substituted with tellurium, while Tm RimO K12A and Tm RimO

Y227F variants were analyzed with SAM. Many experimental trials with the wild type enzyme

were conducted to obtain binding isotherms that yielded reasonably good data appropriate for

fitting and estimating the binding affinity of Tm RimO for SAM. Many experimental trials with

the wild type enzyme were conducted to obtain binding isotherms that yielded reasonably good

data appropriate for fitting and estimating the binding affinity of Tm RimO for SAM. The best

data were obtained when 650 µM SAM (5 µL injections) was titrated into 150 µM Tm RimO in

50 mM HEPES pH 7.5 and 150 mM KCl at 37 °C with a reference power of 10 µcal/sec. The

data were fitted to a single site binding model consisting of the following parameters:

stoichiometry/number of binding sites (N), dissociation constant (Kd), change in enthalpy (∆H),

and change in entropy (∆S) (Figure 5-11A). The calculated Kd of Tm RimO for SAM was ~ 3.3

µM, and the number of binding sites present was 0.45, despite the enzyme containing 8.3 and 9.4

Fe and S ions, respectively, per protein. The ∆H of SAM binding was -19.7 kcal/mol and the

corresponding ∆S value was -38.71 cal/mol/K. Importantly, C, a unitless parameter used to

describe the shape of the curve and assess how well the curve has been fitted (36), was calculated

by taking the product of the enzyme concentration, the determined stochiometry value (N), and

the determined association constant (Ka). Single binding site models are sigmoidal in shape and

provide the most reliable results when C-values fall between 10 and 100, with acceptable results

obtained from C-values between 5 and 500 (37). In this experiment, SAM binding to Tm RimO

resulted in a C-value of 20.6.

Corroboration for the results with Tm RimO and SAM was obtained by performing the

same titration experiment with 650 µM TeSAM (Figure 5-11B). In this case, the following

161

parameter values were determined: Kd, 676 nM; N, 0.5; ∆H, -13.7 kcal/mol; ∆S, -16.0

cal/mol/deg. The calculated C-value was 111, which is slightly outside the optimal range for C-

values, but still acceptable. From these results, it appears that Tm RimO binds SAM and TeSAM

with relatively high affinity, with Kd values ranging from ~ 0.7 to 3.3 µM, and that the number of

binding sites is equal to one-half of the enzyme concentration.

With the N and Kd values of Tm RimO wild type for SAM and TeSAM determined,

active site variants of this enzyme—K12A and Y227F—were tested to see whether the specific

amino acid substitutions had an effect on these parameters that could be rationalized by our in

vitro biochemical characterization of these enzymes. In the case of the K12A variant, the best

data were obtained using 800 µM SAM and 150 µM enzyme at 25°C (Figure 5-12A) with the

following parameters obtained from fitting the data to a single site binding model: Kd, 3.2 µM; N,

0.46; ∆H, -3.3 kcal/mol; ∆S, 13.9 cal/mol/deg; C, 21.5. In terms of SAM binding, the K12A

variant is remarkably similar to the wild type enzyme with nearly identical Kd and N values, yet

the methyl transfer activity of this variant is approximately half that of the wild type enzyme (vide

supra), which suggests this residue does not interact with SAM during its initial binding to RimO,

but may play a yet-to-be determined role in facilitating methyl transfer.

162

Figure 5-11. Isothermal titration calorimetry in which 650 µM SAM (A) or TeSAM (B) was

titrated into 150 µM Tm RimO WT. Both binding isotherms were fitted to a single site binding

model, yielding 0.45 and 0.5 binding sites (N), dissociation constants (KD) of 3.3 µM and 0.7 µM,

changes in enthalpy (∆H) of -19.7 and-13.7 kcal/mol, and changes in entropy (∆S) of -38.7 and -

16.0 cal/mol/K for SAM and TeSAM, respectively.

SAM (1.5 mM) was also titrated into 125 µM Tm RimO Y227F at 25°C with the

following parameters obtained from single site binding model: Kd, 23.4 µM; N, 0.94; ∆H, -6.6

kcal/mol; ∆S, -0.78 cal/mol/deg; C, 4.8 (Figure 5-12B). While the C-value falls just short of the

acceptable range due to the diminished slope of the sigmoidal curve, this data shows the Y227F

substitution has a small effect on the dissociation constant, increasing it by an order of magnitude,

but, more notably, nearly doubles the number of binding sites to 0.94. It is unclear whether the

discrepancy between the number of binding sites determined with the WT enzyme and the K12A

variant (~0.5), and the Y227F variant (~1) is due to the amino acid substitution or due to

163

differing concentrations of each enzyme capable of binding SAM. These results do, however,

narrow the range for the stoichiometry of SAM binding to Tm RimO to between 0.5 and 1.

Figure 5-12. Isothermal titration calorimetry in which 800 µM SAM (A) or 1.5 mM SAM (B)

was titrated into 150 µM Tm RimO K12A (A) or 125 µM Tm RimO Y227F (B) . Both binding

isotherms were fitted to a single site binding model, yielding 0.46 and 0.94 binding sites (N),

dissociation constants (KD) of 3.2 µM and 23.4 µM, changes in enthalpy (∆H) of -3.3 and -6.6

kcal/mol, and changes in entropy (∆S) of 13.9 and -0.78 cal/mol/K for the K12A and Y227F

variants, respectively.

Discussion

All studies to date of the RimO enzyme have used a 13 or 20 amino acid peptide

substrate in place of the full length S12 protein due to the propensity of the protein to be found in

inclusion bodies when overproduced in E. coli (19-21, 25, 38). To obviate this issue, a

purification strategy was employed in which the inclusion bodies and the proteins present within

164

were denatured to improve their solubility. From this denatured protein mixture, the S12 protein

was isolated by immobilized metal affinity chromatography and subsequently refolded while still

bound to the resin by the slow removal of urea. Following size exclusion chromatography,

homogenous S12 protein was obtained and used as a substrate in assays with Tm RimO. Unlike

some other RS enzymes (39-42), RimO does not form appreciable amounts of 5'-dAH in the

absence of a suitable substrate (19, 21, 25). LC/MS analysis of the reaction mixture confirmed

that the S12 protein induced Tm RimO to form 5'-dAH, indicating that the protein binds to RimO

sufficiently to trigger radical formation. MALDI-TOF MS analysis of the S12 protein that had

been incubated in the absence of dithionite, SAM, or both components showed no increases in the

mass of the protein; however, in the presence of all required reaction components, a mass

increase of +45 Da was observed, corresponding to the appendage of a methylthio group to the

S12 protein and confirming the protein is a competent substrate for RimO.

The use of the full length protein as a substrate was thought to reduce the amount of 5'-

dAH formed abortively versus that formed with the 13 amino acid peptide previously used (21,

25, 38). Quantification of the amount of 5'-dAH (266 µM) and SAH (435 µM) formed showed

the former was in slight excess of the concentration of RimO (200 µM), whereas the latter was

produced to a quantity ~ 2-fold greater than the enzyme concentration. Since the MALDI-TOF

MS analysis was qualitative, the amount of S12 that was methylthiolated was not determined;

however, the relative peak ratios of the unmodified and modified protein in the mass spectra

suggest that ~ 50% of the S12 protein was modified, corresponding to a concentration of ~ 100

µM Tm RimO that catalyzed methylthiolation. The concentration of 5'-dAH formed (266 µM)

was almost 3-fold greater than the estimated concentration of methylthiolated S12 protein,

indicating that the full length protein substrate did not decrease the formation of abortive 5'-dAH.

These results do, however, confirm that the S12 protein is a competent substrate. Whether the in

vivo substrate of RimO is the standalone S12 protein, the protein in complex with the 30S subunit

165

or the protein in complex with the completely assembled bacterial ribosome remains to be

determined, however, current evidence supports the standalone S12 protein as the RimO

substrate. Although RimO modifies a small ribosomal protein, it contains a C-terminal TRAM

domain found in tRNA-modifying enzymes that has been shown to directly bind RNA in the

RumA enzyme from E. coli (8, 26, 27, 43), which suggests the S12 protein may resemble RNA

or that RimO recognizes RNA associated with the growing ribonucleoprotein particle. The fact

that both the peptide and the full length protein support turnover, in addition to the previous

finding that RimO activity was not enhanced by the inclusion of a 50 base RNA oligomer

mimicking ribosomal RNA found near the S12 protein in the fully assembled ribosome (25),

suggests the post-translational modification likely occurs before association with the

ribonucleoprotein complex. Indeed, it has been conjectured that the methylthio modification may

play a role in aiding 30S subunit assembly by making an additional contact with the N7-methyl

group of modified nucleotide m7G527 of 16S rRNA (11), a scenario that likely requires the

modification to take place before the protein fully associates with the growing ribosome.

The second determined structure of Tm RimO provided the first snapshots of the active

site of the fully intact protein with both 4Fe-4S clusters present (20). Sequence alignments of

RimOs across several bacterial phyla identified a handful of strictly conserved residues that were

found to reside in the active site of the crystal structure, specifically K12, Q192, and Y227, which

suggested these residues were likely to play important roles in the overall structure or catalytic

function of RimO. We found that the amino acid substitutions in the following RimO variants—

K12A, K12Q, Q192A, Y227F, and Y227A—had little to no effect on the amount of Fe and S

harbored by each enzyme, indicating the amino acid substitutions had no major effects on the

proper folding of the protein or its ability to ligate the two 4Fe-4S clusters.

In terms of methyl transfer ability, all of the RimO variants catalyzed formation of SAH;

however, the K12A and K12Q proteins exhibited the lowest methyl transfer activity. ITC

166

experiments that analyzed the binding of SAM to both the wild type and K12A enzymes showed

essentially the same Kd values of 3.3 µM, indicating K12 is unlikely to be involved in the initial

binding of SAM, but, due to the decreased methyl transfer and 5'-dAH activity of both K12A and

K12Q variants, may be involved in facilitating methyl transfer and directing radical

initiation/generation. The substantial decrease in the ability of the K12Q enzyme to catalyze

methyl transfer may be due to the fact that glutamine can form two hydrogen bonds via its side

chain amide nitrogen and carbonyl oxygen moieties—compared to the inability of alanine to

hydrogen bond via its side chain—which may disrupt the native hydrogen bonding network such

that the active site is perturbed in a way that is far less amenable for methyl transfer activity. A

similar explanation may also extend to the K12A variant, wherein alanine in place of lysine may

eliminate an important hydrogen bond or polar contact that could decrease the enzyme's ability to

catalyze methyl transfer from SAM. While the exact reasons why the K12A variant exhibited a

similar kinetic profile but catalyzed formation of SAH to a lesser extent than the wild type

enzyme, and why the K12Q variant formed meager amounts of SAH slowly are unknown, it is

clear that substituting K12 with a small, nonpolar residue or a relatively large, polar residue

significantly decreases the methyl transfer ability of the enzyme.

The ability of the RimO variants to catalyze the complete methylthiolation reaction was

also assessed: both K12 variants were incapable of catalyzing formation of 5'-dAH and

methylthiolated product; both Y227 variants catalyzed 5'-dAH formation less than that of WT and

formed no detectable methylthiolated product; and Q192A was the only variant that catalyzed

formation of both 5'-dAH and methylthiolated product, albeit to much lower extents than that of

wild type. Since the overall activity of the enzyme was impaired—but not abolished—in the

Q192A variant, the glutamine residue, while conserved, does not appear to play an important role

in catalyzing methyl transfer or the complete methylthiolation reaction. It is possible this residue

facilitates SAH or product release from the active site, which would explain why the substitution

167

with alanine slows the reaction significantly, but additional experiments would need to be

conducted to test this idea.

The fact that both Y227 variants supported 5'-dA• formation but afforded no product is

intriguing, especially in light of the observation that formation of 5'-dAD is not observed when

the reaction is conducted in ~60% D2O, which would be expected if solvent was quenching any

5'-dA• that was uncoupled from methylthiolation. Similar results have been observed in the RS

enzyme biotin synthase from E. coli, in which asparagine and aspartate residues found in a highly

conserved "YNHNLD" motif were shown to form hydrogen bonds with the 2' and 3'-hydroxyl

groups of the ribose moiety of SAM (44, 45). Substitution of these residues with serine and

glutamate, respectively, resulted in variant enzymes that were both capable of binding SAM and

dethiobiotin substrate with relatively high affinity and catalyzed robust formation of 5'-dAH, but

neither the intermediate nor the final product of the reaction was detected; the same reactions

conducted in D2O resulted in no deuterium enrichment into 5'-dAH (46). Careful characterization

of the reaction products identified a new SAM-derived product with a mass corresponding to

incorporation of a sulfur atom into 5'-deoxyadenosine. The authors proposed the sulfur atom

likely comes from the auxiliary cluster in biotin synthase and reacts with 5'-dA• to form 5'-

mercapto-5'-deoxyadenosine (46). Along these lines, future studies with the Y227 variants of

RimO should carefully analyze for production of both SAM-derived and substrate-derived

products that may explain why radical formation does not lead to generation of the

methylthiolated product. While it is unknown what role the tyrosyl residue plays in forming the

methylthiolated product, the hydrogen bonding ability of its 4'-hydroxyl group is likely important

and may be involved in similar hydrogen bonding interactions as those of the conserved

asparagine and aspartate residues found in biotin synthase.

Few studies have determined the dissociation constants associated with SAM binding to

RS enzymes. The Kd values associated with SAM binding to RimO WT and the K12A and

168

Y227F variants tested were determined to be 3.3, 3.2, and 23.4 µM, respectively, which indicate

the enzyme's relatively high affinity for SAM. These values are in line with those determined for

SAM binding to the RS enzymes biotin synthase (1.0 + 0.5 µM) (46) and Cfr (~10 µM) (47), both

of which, like RimO, use two equivalents of SAM for each equivalent of product formed (21, 42,

48). The discrepancy between the number of SAM binding sites determined for RimO WT and

K12A (~0.5) and Y227F (~1) can most likely be attributed to the former enzymes being only ~

50% active and/or folded properly to bind SAM, and is less likely due to the possibility of RimO

exhibiting "half-of-the-sites" reactivity, akin to biotin synthase. In the case of BioB, the enzyme

is purified as a dimer, with each constituent monomer capable of binding one equivalent of SAM

(49). Although each monomer can bind SAM, the amount of biotin formed was consistently

equal to one-half the monomer concentration (i.e. the dimer concentration), which the authors

attributed to the biotin synthase dimer exhibiting "half-of-the-sites" reactivity (23). Our

characterization of RimO by molecular sieve chromatography has shown it exists predominantly

in its monomeric form, even in the presence of SAM (vide supra), and the amount of

methylthiolated product formed per monomer, while variable, more consistently approximates the

concentration of RimO monomer in the absence of exogenous sources of sulfide (21).

Altogether, these results support the assertion that RimO does not form a dimer under turnover

conditions, nor does it exhibit "half-of-the-sites" reactivity, and the stoichiometry of SAM

binding is very likely one-to-one; however further studies will need to be conducted to confirm or

disprove this conjecture.

The recent determination of the redox potentials of the two [4Fe-4S]2+

clusters of Tm

RimO by protein film voltammetry is among the first such studies to be conducted for RS

enzymes using this technique (50). The two clusters differ in their redox potentials by ~80 mV,

with the auxiliary cluster exhibiting a slightly higher potential at -370 mV compared to the RS

cluster at -450 mV in the absence of SAM. These determined redox potentials are in line with

169

those of other RS enzymes: MiaB (-495 + 10 mV, which is likely an approximation of the redox

potentials of the two [4Fe-4S]2+

clusters, but was attributed to the RS cluster) (51); BioB (-430

mV + 20 mV, attributed to indistinguishable potentials corresponding to its [2Fe-2S]2+

and [4Fe-

4S]2+

clusters) (52); anaerobic ribonucleotide reducatase activase (-550 mV; -620 mV in the

presence of SAM) (53); lysine 2,3-aminomutase (-479 + 5 mV, -430 + 2 mV in the presence of

SAM) (54); BtrN (-510 mV and -765 mV for its RS and auxiliary clusters, respectively) (55).

Like BtrN and lysine 2,3-aminomutase, the addition of SAM to Tm RimO caused the redox

potential of the RS cluster to shift slightly higher in potential by +50 mV (52, 55). Notably, the

redox potential of the auxiliary cluster (-370 mV) is slighly higher than that of the RS cluster in

the presence of SAM (-400 mV) but reasonably well matched. Since the auxiliary cluster is more

readily reduced, it could act as a conduit through which electrons flow from an in vivo reductase

to reduce the RS cluster, which is ~ 8Å away (20). Along these lines, the auxiliary cluster could

as an electron sink after formation of the methylthiolated product. In this scenario, the purported

C3-centered substrate radical would react with a methylated sulfur atom bound to the unique Fe

site of the auxiliary cluster, resulting in the homolytic cleavage of the S-Fe bond wherein one

electron combines with the substrate radical to form the methythiolated product, and the other

electron reduces the auxiliary [4Fe-4S]2+

cluster to [4Fe-4S]1+

. The reduced auxiliary cluster

could then transfer the electron back to RS [4Fe-4S]2+

cluster to form the catalytically active

[4Fe-4S]1+

for an additional round of catalysis, thereby obviating the need for the input of

exogenous electrons, or it could transfer the electron to an external reductase (Figure 5-13). One

of the most important unanswered questions concerning the RimO, MiaB, and MtaB

methylthiotransferases centers around the exact role(s) of their auxiliary clusters in the

methylthiolation reactions they catalyze.

170

Figure 5-13. Working hypothesis for the reaction catalyzed by Tm RimO. Step 1: transfer of a

methyl group from SAM bound to the RS [4Fe–4S] cluster to the external sulfur ion of a

sulfide/polysulfide (sulfide shown here for clarity) attached to the unique iron ion of the auxiliary

[4Fe–4S] cluster. Step 2: Reductive fragmentation of a second molecule of SAM bound to the RS

[4Fe–4S] cluster to a 5’-dA• and abstraction of a H• from bound substrate. Step 3: Attack of a

substrate radical onto the methylated sulfur atom of the polysulfide chain to afford the

methylthiolated product and a [4Fe–4S]1+

cluster. Subsequent electron transfer from the auxiliary

cluster to the RS cluster reforms the auxiliary [4Fe-4S]2+

cluster, and addition of sulfide primes

the enzyme for another round of catalysis.

171

References

1. Shajani Z, Sykes MT, Williamson JR. Annual Review of Biochemistry. 2011: 80, 501-526

2. Krzyzosiak W, Denman R, Nurse K, et al. Biochemistry. 1987: 26, 2353-2364

3. Kaczanowska M, Rydén-Aulin M. Microbiology and Molecular Biology Reviews. 2007:

71, 477-494

4. Chow CS, Lamichhane TN, Mahto SK. ACS Chemical Biology. 2007: 2, 610-619

5. Decatur WA, Fournier MJ. Trends in Biochemical Sciences. 2002: 27, 344-351

6. FEBS Letters. 1977: 73, 12-17

7. Kowalak JA, Walsh KA. Protein Science. 1996: 5, 1625-1632


USA. 2008: 105, 1826-1831

9. Powers T, Noller HF. Journal of Molecular Biology. 1994: 235, 156-172

10. Ogle JM, Murphy Iv FV, Tarry MJ, Ramakrishnan V. Cell. 2002: 111, 721-732

11. Polikanov YS, Melnikov SV, Söll D, Steitz TA. Nature Structural and Molecular

Biology. 2015: 22, 342-344

12. Noeske J, Wasserman MR, Terry DS, et al. Nature Structural and Molecular Biology.

2015: 22, 336-341

13. Sofia HJ, Chen G, Hetzler BG, et al. Nucleic Acids Research. 2001: 29, 1097-1106

14. Booker SJ, Cicchillo RM, Grove TL. Current Opinion in Chemical Biology. 2007: 11,

543-552

15. Frey PA. Annual Review of Biochemistry. 2001: 70, 121-148

16. Frey PA, Hegeman AD, Ruzicka FJ. Critical Reviews in Biochemistry and Molecular

Biology. 2008: 43, 63-88

17. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Chemical Reviews. 2014: 114,

4229-4317

18. Jarrett JT. Journal of Biological Chemistry. 2014:


285, 5792-5801



15416

22. Cicchillo RM, Booker SJ. J. Am. Chem. Soc. 2005: 127, 2860-2861

23. Farrar CE, Siu KKW, Howell PL, Jarrett JT. Biochemistry. 2010: 49, 9985-9996

24. Lanz ND, Booker SJ. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research.

2015: 1853, 1316-1334


26. Anantharaman V, Koonin EV, Aravind L. FEMS Microbiology Letters. 2001: 197, 215-

221

27. Lee TT, Agarwalla S, Stroud RM. Cell. 2005: 120, 599-611

28. Grove TL, Ahlum JH, Qin RM, et al. Biochemistry. 2013: 52, 2874-2887

29. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning : a laboratory manual.

Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory

30. Beinert H. Anal. Biochem. 1983: 131, 373-378


32. Kennedy MC, Kent TA, Emptage M, et al. Journal of Biological Chemistry. 1984: 259,

14463-14471

33. Lanz ND, Grove TL, Gogonea CB, et al. Methods in Enzymology. 2012: 516, 125-152

172

34. Bradford MM. Analytical Biochemistry. 1976: 72, 248-254

35. Middelberg APJ. Trends in Biotechnology. 2002: 20, 437-443

36. Wiseman T, Williston S, Brandts JF, Lin L-N. Analytical Biochemistry. 1989: 179, 131-

137

37. Myszka DG, Abdiche YN, Arisaka F, et al. Journal of Biomolecular Techniques : JBT.

2003: 14, 247-269

38. Landgraf BJ, Booker SJ. Journal of the American Chemical Society. 2016: 138, 2889-

2892

39. Moss ML, Frey PA. Journal of Biological Chemistry. 1990: 265, 18112-18115

40. Duschene KS, Broderick JB. FEBS Letters. 2010: 584, 1263-1267

41. McGlynn SE, Boyd ES, Shepard EM, et al. Journal of Bacteriology. 2010: 192, 595-598

42. Grove TL, Benner JS, Radle MI, et al. Science. 2011: 332, 604-607

43. Lee TT, Agarwalla S, Stroud RM. Structure. 2004: 12, 397-407

44. Berkovitch F, Nicolet Y, Wan JT, et al. Science. 2004: 303, 76-79

45. Nicolet Y, Drennan CL. Nucleic Acids Research. 2004: 32, 4015-4025

46. Farrar CE, Jarrett JT. Biochemistry. 2009: 48, 2448-2458

47. Challand MR, Salvadori E, Driesener RC, et al. PLoS ONE. 2013: 8, e67979

48. Taylor AM, Farrar CE, Jarrett JT. Biochemistry. 2008: 47, 9309-9317

49. Ugulava NB, Frederick KK, Jarrett JT. Biochemistry. 2003: 42, 2708-2719

50. Maiocco SJ, Arcinas AJ, Landgraf BJ, et al. Biochemistry. 2016: 55, 5531-5536

51. Pierrel F, Hernandez HL, Johnson MK, et al. Journal of Biological Chemistry. 2003: 278,

29515-29524

52. Ugulava NB, Gibney BR, Jarrett JT. Biochemistry. 2001: 40, 8343-8351

53. Mulliez E, Padovani D, Atta M, et al. Biochemistry. 2001: 40, 3730-3736

54. Hinckley GT, Frey PA. Biochemistry. 2006: 45, 3219-3225

55. Maiocco SJ, Grove TL, Booker SJ, Elliott SJ. Journal of the American Chemical Society.

2015: 137, 8664-8667

VITA

Bradley J. Landgraf

Education:

Allegheny College – Meadville, PA, B.S. Chemistry, 2005

The Pennsylvania State University, University Park, PA, Ph. D. (2009 to 2016)

Dissertation Title: “Mechanistic studies of the methylthiolation reaction catalyzed by the

radical SAM enymze RimO"

Publications:

Stephanie Maiocco, Arthur J. Arcinas, Bradley J. Landgraf, Squire J. Booker, and Sean J. Elliot.

Transformations of the FeS clusters of the methylthiotransferases MiaB and RimO, detected

by direct electrochemistry. Biochemistry. 2016: 55, 5531-5536

Eric Block, Squire J. Booker, Sonia Flores-Penalba, Graham George, Bradley J. Landgraf, Jun Liu,

Stephene N. Lodge, M. Jake Pushie, Sharon Rozovsky, Abith Vattekkatte, Rama Yaghi, and

Huawei Zeng. Trifluoroselenomethionine, a New Non-Natural Amino Acid with Enchanced

Methioninase-Induced Cytotoxicity toward Human Colon Cancer Cells, is Incorporated into

GB1 Proteins with Loss of the Trifluoromethyl Group. Chembiochem. 2016: 17, 1738-1751.

Bradley J. Landgraf, Erin L. McCarthy, and Squire J. Booker. Radical SAM enzymes in Human

Health and Disease. Annu. Rev. Biochem., 2016. 85, 485-514.

Bradley J. Landgraf and Squire J. Booker. Stereochemical Course of the Reaction Catalyzed by

RimO, a Radical SAM Methylthiotransferase. J. Am. Chem. Soc., 2016. 138 (9), 2889-2992.

Bradley J. Landgraf, Arthur J. Arcinas, Kyung-Hoon Lee, and Squire J. Booker. An Intermediate

Methyl Carrier in the Radical SAM Methylthiotransferases RimO and MiaB. J. Am. Chem.

Soc. 2013. 135 (41), 15404-15416

.

Bradley J. Landgraf and Squire J. Booker. Biochemistry: The ylide has landed. Nature., 2013 498,

45-47.

Tyler L. Grove, Jack S. Benner, Matthew I. Radle, Jessica H. Ahlum, Bradley J. Landgraf, Carsten

Krebs, and Squire J. Booker. A Radically Different Mechanism for S-Adenosylmethionine–

Dependent Methyltransferases. Science., 2011. 332 (6029), 604-607.

Yiqing Feng, Yuji Wang, Bradley Landgraf, Shi Liu, and Gong Chen. Facile Benzo-Ring

Construction via Palladium- Catalyzed Functionalization of Unactivated sp3 C-H Bonds

under Mild Reaction Conditions., Org. Lett., 2010. 12, 3414– 3417.

Meng-Dawn Cheng, Edwin Corporan, Matthew J. DeWitt, Bradley Landgraf. Emissions of Volatile

Particulate Components from Turboshaft Engines Operated with JP-8 and Fischer-Tropsch Fuels. Aerosol and Air Quality Research, 2009, Vol. 9, No. 2: 237-256.

mechanistic studies of the methylthiolation reaction

Documents