elongation factor p-dependent translation and …...elongation factor p-dependent translation and...
TRANSCRIPT
1
i
Elongation Factor P-Dependent Translation and
Metabolic Phenotypes of Salmonella
by
Steven Jeremy Hersch
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy
Department of Molecular Genetics
University of Toronto
© Copyright by Steven J. Hersch (2016)
ii
Elongation Factor P-Dependent Translation and
Metabolic Phenotypes of Salmonella
Steven Jeremy Hersch
Doctor of Philosophy
Department of Molecular Genetics
University of Toronto
2016
Abstract
The facultative intracellular pathogen Salmonella has acquired adaptations contributing
to virulence, including pathogenicity islands and regulatory alterations. Elongation factor P (EF-
P) is a universally conserved translation factor, and deletion of efp in Salmonella results in
pleiotropic phenotypes, including hyperactive metabolism, reduced stress resistance, and
avirulence. Crystal structures demonstrated that EF-P adopts the shape of a tRNA and binds to
the ribosome to stimulate peptide bond synthesis. Preliminary analyses suggested that EF-P is
not a general translation factor but rather influences a specific subset of proteins. In this thesis I
investigate the mechanism underlying EF-P’s impact on translation and how its deletion leads to
the observed phenotypes. I present the Salmonella EF-P-dependent proteome, and particular
proteins offer explanations for efp mutant phenotypes. Meanwhile, other groups discovered that
EF-P rescues ribosomes stalled at polyproline motifs. I expand upon their work and show that
polyproline motifs are not strictly necessary or solely sufficient to render a protein dependent on
EF-P. I demonstrate that residues upstream of the polyproline motif significantly influence stall
strength. Furthermore, translation initiation rate greatly impacts EF-P-dependence. I present a
model wherein elongation stalls only reduce protein synthesis if they are more rate-limiting than
translation initiation.
iii
While examining efp mutants, I discovered a novel phenotype of wild-type Salmonella.
When dicarboxylates are the sole carbon source, Salmonella represses its growth using the
general stress response sigma factor RpoS, the RpoS-stabilizer IraP, and repression of the
dicarboxylate transporter DctA. Moreover, E. coli does not exhibit this phenotype and the E. coli
dctA promoter is sufficient to induce Salmonella growth. This divergence suggests a potential
role in virulence, and I examine the accumulation of succinate and itaconate in activated
macrophage. I demonstrate that the dicarboxylate phenotype does not influence intracellular
survival, and I propose novel hypotheses of why Salmonella has evolved this trait.
iv
Acknowledgments
Many people have influenced my graduate school experience and deserve thanks. All the
people who have donated strains, plasmids, ideas and time; the people who ask questions at
seminars because sometimes even those dreaded questions that you can’t answer and make you
think end up leading to a new figure in a paper; and the people who don’t necessarily ask
questions but still say “nice talk” afterward – you’re noticed and greatly appreciated!
Specifically I’d like to thank a few people from the Molecular Genetics department including
Ryan Gaudet for donating a significant amount of his time to teaching me how to culture and
infect macrophage. Joe Bondy-Denomy and Vanessa Tran for always being willing to discuss
cool science and get as excited as I do about microbiology conferences! I would also like to
thank my supervisory committee members Dr. Leah Cowen and Dr. Jack Greenblatt who have
always offered their time, advice and knowledge to support me over the course of my degree.
Their mentorship and motivation have been invaluable!
I also want to say thank you to all the members of the Navarre lab that I’ve interacted
with over the years: Former members Hazel Soto-Montoya, Emily Beckett, Grace Tong, and
Jeremy Soo for their support and ideas; Sabrina Ali and Andrea Leung for their guidance and for
creating such a positive working environment in the lab; and current members Vivian Cheung,
Kamna Singh, Erin Wong and Sinwan Basharat for listening to my ramblings and bouncing ideas
back and forth with me. Primarily I would like to thank Betty Zou, who not only established the
foundation for my project and was always willing to discuss EF-P, but moreover hugely
contributed to training, supporting and guiding me through much of my graduate career. All of
these people have not only been great colleagues but also great friends.
My family has always supported me in so many ways. My brothers Mike and Rob are my
closest friends and are always interested not just in hearing about my work, but in really learning
what it is and means despite it being in a completely different field than their own very exciting
careers. My parents, Ann and Edwin, whose high regard for education, discussion and doing
what I find interesting allowed me to pursue this career! They’ve contributed so much support in
everything from advice to finance to just being cool people that I like to hang out with.
Predominantly I want to thank them for raising me to be the person that I am today and for their
guidance and love, always. Furthermore I want to thank my fiancée Annie Walton who is always
there for me and has kept me sane over the years. She’s my favourite person in the world.
v
Finally I would like to thank my supervisor Dr. William Navarre. Will has been a
fantastic mentor and without him graduate school would have been nowhere near as exciting,
enlightening and successful. He gave me the freedom to pursue my crazy ideas, but was always
willing and able to figure out problems and guide me through them. He has also been an
extraordinary source of ideas and leadership. I will always remember the ‘best lab meeting ever’
when he saw a bunch of my odd data and on the spot compiled it into the primary model of my
thesis that made everything make sense! Beyond that he’s been a great person to work and get
along with as not just a mentor and supervisor, but as a friend.
Thank you everyone once again.
…And now on to science!!
vi
Table of Contents
Abstract .......................................................................................................................................... ii
Acknowledgments ........................................................................................................................ iv
Table of Contents ......................................................................................................................... vi
List of Tables ................................................................................................................................. x
List of Figures ............................................................................................................................... xi
List of Appendices ...................................................................................................................... xiii
1 Introduction ........................................................................................................................... 1
1.1 A brief overview of bacterial metabolism and response to stress ................................... 1
1.2 Salmonella....................................................................................................................... 4
1.2.1 Overview ..................................................................................................................................4
1.2.2 Virulence traits and divergence from E. coli ...........................................................................4
1.3 Bacterial translation ........................................................................................................ 6
1.3.1 Overview ..................................................................................................................................6
1.3.2 The role of ribosome stalling and translation rates during elongation ...................................9
1.4 PoxA, YjeK and EF-P, the ‘PYE pathway’ .................................................................... 9
1.4.1 Identification and overview .....................................................................................................9
1.4.2 Post-translational modification of EF-P ...............................................................................12
1.4.3 PYE mutants display pleiotropic phenotypes.........................................................................12
1.4.4 EF-P does not appear to be a general translation factor ......................................................12
1.5 Thesis rationale ............................................................................................................. 13
2 Methods ................................................................................................................................ 14
2.1 Bacterial strains ............................................................................................................. 14
2.1.1 Salmonella strain background and efp mutants ....................................................................14
2.1.2 Chromosomal replacement of the dctA promoter ..................................................................14
2.1.3 SGSC and ECOR collections .................................................................................................15
2.2 Plasmids ........................................................................................................................ 15
2.2.1 GFP translational fusion plasmids ........................................................................................15
2.2.2 DctA overexpression plasmid ................................................................................................16
2.2.3 STM3120 transcriptional fusion plasmid ..............................................................................17
2.3 Experimental Methods .................................................................................................. 17
2.3.1 N-Phenyl-1-Naphtylamine accumulation assay .....................................................................17
vii
2.3.2 Ethidium Bromide (EtBr) Accumulation Assay .....................................................................18
2.3.3 Stable isotope labelling of amino acids in cell culture (SILAC) ............................................18
2.3.4 Mass spectrometry and proteomic data analysis ...................................................................19
2.3.5 DAVID analysis .....................................................................................................................20
2.3.6 Translational fusion assay .....................................................................................................20
2.3.7 Reverse transcriptase quantitative PCR (RT-qPCR) .............................................................21
2.3.8 Immunoblotting ......................................................................................................................21
2.3.9 Curve Fitting ..........................................................................................................................22
2.3.10 Growth using dicarboxylates as a sole carbon source ......................................................22
2.3.11 Catalase assay ...................................................................................................................23
2.3.12 Acidified media survival assay ..........................................................................................23
2.3.13 Promoter induction fluorescence assay.............................................................................24
2.3.14 Macrophage survival assay ...............................................................................................25
2.3.15 Fluorescence microscopy ..................................................................................................25
2.3.16 Identification of STM3121-STM3117 orthologs in Salmonella serovars ..........................26
3 Analysis of the Salmonella efp mutant proteome ............................................................. 27
3.1 Overview ....................................................................................................................... 28
3.2 Results ........................................................................................................................... 29
3.2.1 Salmonella efp mutants are selectively more permeable to NPN ..........................................29
3.2.2 Deletion of yfcM has no effect on NPN permeability ............................................................31
3.2.3 Deletion of kdgM partially complements the permeability defect .........................................31
3.2.4 Identification of EF-P regulated proteins by SILAC .............................................................34
3.2.5 EF-P rescues ribosomes stalled at polyproline motifs ..........................................................36
3.2.6 Polyproline motifs are not always necessary or sufficient to confer EF-P-dependence .......36
3.2.7 Functional annotation analysis of EF-P-dependent proteins ................................................39
3.3 Discussion ..................................................................................................................... 40
4 Factors that influence EF-P-dependence .......................................................................... 42
4.1 Overview ....................................................................................................................... 43
4.2 Results ........................................................................................................................... 43
4.2.1 PoxB requires EF-P activity for its efficient translation .......................................................43
4.2.2 PoxB is dependent on EF-P due to a novel six amino acid motif, GSCGPG ........................47
4.2.3 AtpD and AtpA as model proteins to study polyproline motifs ..............................................50
4.2.4 Upstream residues influence EF-P-dependence in AtpD and AtpA ......................................54
4.2.5 The 5’ UTR plays a significant role in EF-P dependence .....................................................56
viii
4.2.6 The 5’ UTR regions that include the SD sequence and a potential translation enhancer play
a significant role in EF-P dependence ...............................................................................................58
4.2.7 Mutations that affect translation initiation influence EF-P dependence ...............................61
4.2.8 Initiation rate and stall strength correlate with EF-P dependence .......................................64
4.2.9 Modeling the interplay between translation initiation rate and elongation stalls .................67
4.2.10 Fate of stalled peptides in vivo ..........................................................................................68
4.3 Discussion ..................................................................................................................... 70
5 Salmonella evolved to suppress its uptake of dicarboxylic acids during stress ............. 75
5.1 Overview ....................................................................................................................... 76
5.2 Results ........................................................................................................................... 76
5.2.1 Wild-type but not efp or rpoS mutant Salmonella displays delayed growth using
dicarboxylic acids as a sole carbon source ........................................................................................76
5.2.2 Many Salmonella but few E. coli strains delay growth using succinate ................................79
5.2.3 The RpoS stabilizer IraP contributes to growth shutdown in succinate media .....................82
5.2.4 Supplementation with proline or citrate induces diauxic growth using succinate and
subsequent repression requires the stringent response ......................................................................85
5.2.5 Growth lag appears to be due to repression of succinate import ..........................................88
5.2.6 The E. coli dctA promoter is sufficient to induce Salmonella growth using succinate ..........88
5.3 Discussion ..................................................................................................................... 91
6 Salmonella versus macrophage-produced dicarboxylic acids......................................... 96
6.1 Overview ....................................................................................................................... 97
6.2 Results ........................................................................................................................... 97
6.2.1 Hypothesis and rationale that Salmonella may repress uptake of dicarboxylic acids to
improve survival in activated macrophage .........................................................................................97
6.2.2 Constitutive expression of dctA from a plasmid reduces Salmonella survival in acidified
succinate and in macrophage ...........................................................................................................100
6.2.3 Chromosomal E. coli dctA promoter does not reduce Salmonella survival ........................102
6.2.4 Salmonella contains genes that can degrade itaconate .......................................................104
6.2.5 The itaconate degradation operon is induced by STM3121 in the presence of itaconate ...106
6.2.6 Deletion of the itaconate degradation operon does not reduce Salmonella survival ..........108
6.2.7 The itaconate degradation operon is induced in J774 but not THP-1 macrophage............110
6.3 Discussion ................................................................................................................... 112
7 Discussion and Future Directions .................................................................................... 117
ix
7.1 Summary and Discussion ............................................................................................ 117
7.2 Future Directions ........................................................................................................ 121
7.2.1 Regulation of EF-P and its potential regulatory activity ....................................................121
7.2.2 Characterize the repression of the Salmonella dctA promoter ............................................122
7.2.3 Elucidate the role of Salmonella dctA repression in macrophage, .....................................123
7.2.4 Investigate alternate itaconate defence mechanisms ...........................................................124
7.2.5 Scrutinize the induction of itaconate synthesis in human and mouse macrophage .............125
7.3 Conclusions ................................................................................................................. 126
References .................................................................................................................................. 127
x
List of Tables
Table 1: Constructs used in Figure 22 ......................................................................................... 66
xi
List of Figures
Figure 1: Regulation of rpoS at transcriptional and post-transcriptional levels ............................. 3
Figure 2: Translation in bacteria ..................................................................................................... 8
Figure 3: Structure and modification of EF-P............................................................................... 11
Figure 4: Salmonella efp mutants are more permeable to NPN but not EtBr. .............................. 30
Figure 5: Deletion of yfcM has no effect on permeability to NPN ............................................... 32
Figure 6: Deletion of kdgM partially complements the efp mutant permeability defect .............. 33
Figure 7: A subset of proteins show significantly altered abundance in Δefp Salmonella ........... 35
Figure 8: Comparison of proteins identified in SILAC and those with EF-P-dependent
polyproline motifs. ........................................................................................................................ 38
Figure 9: PoxB-sfGFP translation is impaired in PYE mutants .................................................... 45
Figure 10: PoxB mRNA levels are similar in wild-type and ∆efp strains .................................... 46
Figure 11: The GSCGPG motif of PoxB renders it dependent on EF-P ...................................... 48
Figure 12: PoxB dependence on EF-P is determined by the amino acid sequence ...................... 49
Figure 13: ATP Synthase components show altered abundance in the efp mutant. ..................... 51
Figure 14: AtpD but not AtpA requires EF-P for its synthesis ..................................................... 52
Figure 15: AtpD but not AtpA shows decreased fluorescence throughout the growth curve of efp
mutant Salmonella. ....................................................................................................................... 53
Figure 16: Upstream residues influence EF-P-dependence .......................................................... 55
Figure 17: The 5’ UTR plays a role in EF-P dependence ............................................................. 57
Figure 18: The 5’ UTR regions affecting EF-P dependence include the ribosome binding site and
correlate with expression levels .................................................................................................... 59
Figure 19: GFP levels determined by immunoblotting follow similar trends as levels determined
by fluorescence. ............................................................................................................................ 60
Figure 20: The effect of the 5’ UTR on EF-P dependence requires an intact stall motif. ............ 62
Figure 21: Single base mutations in the SD sequence or start codon alter expression and EF-P
dependence .................................................................................................................................... 63
Figure 22: Translation initiation and elongation stall strength influence protein level ................ 65
Figure 23: Peptide abundance ratios do not change significantly before and after APP, PPG or
PPP motifs ..................................................................................................................................... 69
Figure 24: Model of the interplay between translation initiation rate and elongation stalls ......... 74
xii
Figure 25: Wild-type Salmonella displays an extended lag phase using dicarboxylic acids as a
sole carbon source ......................................................................................................................... 78
Figure 26: E. coli grows earlier than Salmonella using succinate ................................................ 81
Figure 27: Deletion of the RpoS stabilizer IraP results in early growth on succinate .................. 84
Figure 28: Proline or citrate induce growth using succinate and subsequent repression involves
the stringent response .................................................................................................................... 87
Figure 29: Expression of dctA induces growth using succinate ................................................... 89
Figure 30: Working model of Salmonella’s response to succinate ............................................... 94
Figure 31: Alignment of Salmonella and E. coli dctA promoters ................................................. 95
Figure 32: Dicarboxylic acids can act as proton shuttles in acidified conditions ......................... 99
Figure 33: Overexpression of DctA from a plasmid decreases survival in macrophage due to
toxicity ........................................................................................................................................ 101
Figure 34: Replacement of the Salmonella dctA promoter does not influence survival ............. 103
Figure 35: The itaconate degradation operon ............................................................................. 105
Figure 36: The STM3120 promoter is inducible by itaconate and STM3121 is necessary and
sufficient for this induction ......................................................................................................... 107
Figure 37: Deletion of the Salmonella itaconate degradation operon or its regulator does not
influence survival in acidified itaconate or macrophage ............................................................ 109
Figure 38: The STM3120 promoter is induced in mouse but not human macrophage .............. 111
Figure 39: Salmonella serovars Typhi, Paratyphi A and Agona uniquely lack the itaconate
degradation operon...................................................................................................................... 116
xiii
List of Appendices
Hersch_Steven_J_201611_PhD_datatable1.xlsx
Hersch_Steven_J_201611_PhD_datatable2.xlsx
1
1 Introduction
1.1 A brief overview of bacterial metabolism and response to stress
Bacteria must constantly adapt to changing environmental conditions by sensing their
surroundings and integrating signals to initiate rapid growth in nutrient rich situations or instigate
defence mechanisms and metabolic hibernation in response to stress1. Bacteria have evolved
many different means of sensing and responding to environmental cues, but many of the best
studied are those of E. coli. For instance, E. coli can respond to the presence or absence of the
rich carbon source glucose by means of a process called catabolite repression2-7. In brief, the
absence of glucose leads to the phosphorylation of the CyaA protein, which synthesizes the
second messenger molecular cyclic AMP (cAMP). cAMP is bound by the cAMP receptor
protein (CRP) allowing it to bind to DNA and regulate gene expression to allow for import,
degradation and use of alternate carbon sources. In addition to CRP-cAMP, many other response
systems also modulate bacterial metabolism. For instance the catabolite repressor activator, Cra
(also known as FruR), is able to sense metabolic flux to alter gene expression pertaining to
central metabolic pathways1,8-10. These include glycolysis, wherein glucose is degraded into
pyruvate, and the tricarboxylic acid (TCA) cycle, which further converts pyruvate to carbon
dioxide to produce the reduced intermediates NADH and FADH2 (Ecocyc, described in
reference 11, is a very useful database and tool for examining metabolic pathways of E. coli).
These intermediates can then be employed by the electron transport chain to produce energy via
ATP Synthase12.
Metabolic modulation is important during conditions of stress as exemplified by the
bacterial stringent response, wherein the RelA protein senses amino acid starvation and produces
the second messenger molecule, guanosine 5’-disphosphate 3’-diphosphate (ppGpp)13-16. This
second messenger is also produced by the SpoT protein in response to other cellular stress cues
to influence transcription of amino acid biosynthesis genes, transfer RNA (tRNA), and the
translation apparatus, including ribosomal and translation factor genes (amongst others). Bacteria
can also alter gene expression in response to stress using alternative sigma factors that associate
with the RNA polymerase holoenzyme to guide transcription. A particularly well studied
instance is the general stress response sigma factor RpoS (σS). RpoS can be activated in response
to a variety of different conditions including starvation, hyper-osmolarity and oxidative stress,
2
and can be regulated at all levels of synthesis from transcription to protein degradation17 (Figure
1). Moreover, RpoS has been linked to virulence in a number of pathogenic bacteria, likely by
contributing to survival within an infected host18. Indeed, pathogenic bacteria have further
adapted their mechanisms of sensing and reacting to their environment such that they are
especially equipped to survive and replicate within their particular host niche. A useful approach
to study differences in pathogenic and commensal bacteria is comparing the genomes and
phenotypes of two well characterized enterobacteria, E. coli and Salmonella. These bacteria are
closely related, yet Salmonella has acquired a number of adaptations in regulatory pathways,
metabolism and its response to stress that accommodate its virulent lifestyle19.
3
Figure 1: Regulation of rpoS at transcriptional and post-transcriptional levels. Figure is
shown as it appears in reference 17 and the publisher states that they allow for its use in this thesis
without special permission. Regulation of rpoS transcription (A), translation (B), activity (C),
and stability (D). The stringent response second messenger ppGpp and the antiadaptor IraP,
which binds to RssB to stabilize RpoS are particularly relevant in the later chapters of this thesis.
4
1.2 Salmonella
1.2.1 Overview
Salmonella are gram negative pathogenic bacteria, and the genus contain two species, S.
enterica and S. bongori. The S. enterica species contains thousands of serovars (serological
variants) that are grouped into six subspecies (numbered I to VI with group VII representing S.
bongori)20. Subspecies I, also known as subsp. enterica, contains the majority of serovars that
cause disease in humans, including the most studied Salmonella enterica subsp. enterica serovar
Typhimurium (non-typhoidal; causes gastroenteritis in humans) and the most lethal Salmonella
enterica subsp. enterica serovar Typhi (typhoidal; causes typhoid fever in humans)21. Salmonella
can invade the gut epithelium and non-typhoidal strains induce an inflammatory response
resulting in gastroenteritis. In contrast, Typhoid strains do not induce a prominent inflammation
response in the gastrointestinal tract but instead cause a systemic disease called typhoid fever
characterized by fever and abdominal pain22. Most Salmonella serovars can infect animal hosts
as well as (or instead of) humans, but serovars Typhi and Paratyphi A are human specific23,24.
Globally per annum, Salmonella is estimated to cause almost 100 million cases of gastroenteritis
resulting in about 150 000 deaths, and an estimated 22 million cases of typhoid fever resulting in
about 200 000 deaths25,26. Thus Salmonella does not only represent a useful model organism that
can be genetically manipulated in a lab setting, but also has a significant global disease burden,
and research on its virulence could have a significant impact on human health.
1.2.2 Virulence traits and divergence from E. coli
The study of how Salmonella evolved to be a pathogen has been greatly facilitated by
comparing it to the closely related bacteria Escherichia, including E. coli. Both genera are
members of the family Enterobacteriaceae; they diverged from their last common ancestor
approximately 100-160 million years ago, and many of the adaptations acquired by Salmonella
since then pertain to its virulent lifestyle19. The Salmonellae have acquired a number of genomic
islands by horizontal gene transfer called Salmonella pathogenicity islands (SPI). For instance,
all Salmonella including S. bongori have SPI-1, a cluster of more than 40 genes encoding a type
three secretion system and associated virulence factors that allow the bacteria to inject effector
5
proteins into host epithelial cells for the purpose of invasion and intracellular growth27. This
differs from most strains of E. coli (with the exception of enteroinvasive E. coli and Shigella
spp., which are now classified as variants of E. coli) which do not invade the epithelium and
remain in the gut lumen28. Over 20 SPIs have been identified in various serovars that each play
specific roles in Salmonella virulence21. For instance SPI-2, which is found in all S. enterica
isolates but not S. bongori, encodes a second type three secretion system and is involved in
intracellular survival and replication within host cells such as macrophage29. This is another
instance where E. coli and Salmonella differ – survival within host macrophage is important for
Salmonella virulence30,31. The induction of SPI-2 and other survival mechanisms within host
cells is therefore critical for Salmonella’s pathogenic lifestyle.
Salmonella has adapted its regulatory networks to recognize and react to environments
that it encounters during infection, such as the phagosome of macrophage. For example, E. coli
and Salmonella both encode a ‘two-component’ signalling system known as PhoPQ. The PhoPQ
system of Salmonella, however, has evolved to respond to conditions within a macrophage
phagosome, including low pH, low magnesium concentration, and the presence of antimicrobial
peptides32-35. Activation of the PhoPQ system contributes to induction of virulence genes,
reorganization of the Salmonella outer membrane to defend against the innate immune system,
and stimulation of the stress response via RpoS36-38.
In addition to regulatory adaptations like that of PhoPQ, Salmonella has also acquired
subtle but important metabolic differences from E. coli. An interesting example is Salmonella’s
distinctive ability to employ tetrathionate as a terminal electron acceptor using the ttrRSBCA
gene cluster in SPI-239. Though tetrathionate is not normally present in the gut, the inflammation
induced by Salmonella invasion of host cells leads to an oxidative burst and the oxidation of
readily available thiosulfate to tetrathionate40. This provides Salmonella in the anaerobic gut
lumen with a terminal electron acceptor allowing it to respire under conditions where E. coli
cannot. Salmonella can thereby use alternative carbon sources such as ethanolamine that require
respiration to provide energy, and ultimately out-compete other microbes in the gut41. Thus, over
the course of a non-typhoidal infection, Salmonella “wants” to induce inflammation in order to
acquire tetrathionate and provide it with a fitness edge over the normal microbiome, allowing it
to replicate and disseminate efficiently.
6
The identification and characterization of the factors and adaptations that enable
Salmonella to be a pathogen is ongoing, and novel discoveries continue to be made. During his
post-doctoral work, my supervisor, Dr. William Navarre, identified mutations in two poorly
characterized genes, poxA and yjeK, that render Salmonella avirulent. As discussed below, these
were later identified to play a role in translation and so the next section will provide an overview
of protein synthesis in bacteria.
1.3 Bacterial translation
1.3.1 Overview
Cellular protein levels are the sum of transcription, translation and mRNA and protein
degradation rates, and all of these incorporate their own regulatory mechanisms. During
translation, one protein is produced for each ribosome that successfully recognizes and
assembles on an mRNA transcript (initiation), constructs a polypeptide (elongation) and releases
the full-length amino acid chain (termination). The process of decoding an mRNA transcript to
synthesize a polypeptide that can fold into a protein is complex, and bacteria have evolved a
number of translation factors to catalyze the process. The ribosome itself is a complex enzyme
consisting of large (50S) and small (30S) subunits totalling 3 ribosomal RNAs (rRNA) and about
50 proteins that combine to form the complete 2.5 megadalton (70S) ribosome that catalyzes
peptide bond formation (Figure 2A)42.
The ribosome first assembles on the 5’ untranslated region (UTR) of an mRNA transcript
with the help of three initiation factors (IF1-3) and usually an initiator tRNA charged with a
methionine amino acid corresponding to the ATG start codon (although other codons, including
GTG, can act as start codons for certain proteins). Elongation of the nascent peptide then
progresses in cycles as the elongation factor EF-Tu recruits amino acyl-charged tRNAs to the
‘acceptor’ (A) site of the ribosome. ATP hydrolysis by EF-Tu ensures that only tRNAs that
properly match the mRNA codon remain bound. Following this decoding event, EF-Tu releases
the tRNA, which rocks forward toward the tRNA in the ‘peptidyl’ (P) site of the ribosome. At
this point the ribosome catalyzes the formation of a peptide bond between the amino acid residue
bound to the P site tRNA and the new residue in the A site. Thus the amino acid and any
7
polypeptide chain attached to it are transferred to the A site tRNA. Another elongation factor,
EF-G then contributes to translocation wherein the A site tRNA (with the growing polypeptide
chain) shifts to the P site, and the former P site tRNA (now uncharged) shifts to the ‘exit’ (E) site
for release from the ribosome. This process repeats until the mRNA encodes a stop codon, which
is recognized by release factor (RF) 1 or 2 to release the new protein and (with the aid of RF-3)
dissociate the subunits of the ribosome for reuse on other transcripts.
Decreased efficiency of any of these phases (initiation, elongation or termination) reduces
the number of ribosomes able to complete translation and, hence, lowers total protein production.
To maximize the efficient use of resources, the rate-limiting step usually occurs at the point of
initiation and extensive work has made it clear that the strength of ribosome binding greatly
affects the amount of protein produced from an mRNA transcript43. By altering elements of the
ribosome binding site (RBS) translation rates can be controlled in order to fine tune gene
expression at the post-transcriptional level44-46. One such element is the Shine-Dalgarno (SD)
sequence of canonical eubacterial mRNA, which is complementary to a region of the 16S
ribosomal RNA that recruits the 30S ribosome to the mRNA transcript. Thus the strength of
binding between the SD sequence and the 16S rRNA is directly related to translation initiation
rate. Other elements of the 5’ UTR can also play a role in translation initiation including mRNA
secondary structure, small RNA (sRNA) binding sites, the start codon, and translation enhancers
– AT-rich regions in proximity to the SD sequence that improve small subunit binding47-49.
8
A)
B)
Figure 2: Translation in bacteria. A) Detailed overview of translation in bacteria. Figure is
shown with permission as it appears in reference 42. B) Simplified cartoon depicting key aspects
of bacterial translation that are particularly relevant for this thesis.
Start codon Stop codon
Open reading frame (ORF)
5’ UTR
mRNA
Initiation Elongation Termination(Rate limiting step)
70S5’ 3’
3’ UTR
50S
30SRibosome
9
1.3.2 The role of ribosome stalling and translation rates during elongation
Though initiation is typically the primary regulatory step in translation, slower
progression through elongation can also influence protein synthesis. This has been studied
extensively for rare codons that slow the elongation rate due to lower cellular tRNA
concentration, and starvation conditions that restrict the amount of aminoacylated tRNA50-54. In
these instances ribosomes must pause until a properly charged tRNA accurately decodes at the
A-site, and the impact of this delay on protein levels can be influenced by rare codon clustering,
repetition, or position within an open reading frame (ORF)55-59.
Elongation stalls have been shown to regulate mRNA structures affecting ribosome
binding, influence transcription of downstream genes via impeding the transcriptional terminator
Rho, potentially fascilitate proper protein folding, or directly mediate protein production60-65. For
instance, during translation of secM, ribosomes pause at a particular amino acid sequence until
the N-terminal peptide is recognized by the SecA secretion apparatus – thus directing secM
expression66,67. Interaction between the nascent polypeptide and the ribosome exit tunnel has
been implicated to mediate this and other stalls in elongation or termination68-71. Though these
stall motifs bear little sequence similarity, many involve prolines, presumably because of their
unique characteristics and relatively poor peptide bonding capability72,73.
1.4 PoxA, YjeK and EF-P, the ‘PYE pathway’
1.4.1 Identification and overview
During his post-doctoral work, Dr. Navarre identified mutations in two genes, poxA and
yjeK, that decreased Salmonella growth, increased their resistance to the nitric oxide (NO)-
donating compound S-Nitrosoglutathione (GSNO), and rendered them avirulent in a mouse
model of infection74. Very little was known about either gene other than that PoxA was required
for the efficient expression of the pyruvate oxidase (poxB) gene75,76. Bioinformatic analysis of
genomic location, synteny, and species distribution of both genes revealed that poxA and yjeK
are genetically linked to a third gene, efp, encoding elongation factor P (EF-P). This was a poorly
understood translation factor that had only been demonstrated to modestly stimulate peptide
bond formation in vitro77,78. Crystal structures of EF-P, both alone and in complex with the
10
ribosome, revealed that EF-P adopts the size and shape of a tRNA and that it binds to the
ribosome between the P and E sites and can align itself adjacent to the P site tRNA (Figure 3A
and B)79-81. Moreover, EF-P is universally conserved in all bacteria, and eukarya and archaea
also encode homologs called initiation factor 5A (e/aIF-5A), strongly suggesting that it plays a
critical role in translation82.
11
Figure 3: Structure and modification of EF-P. A) EF-P is similar in size and shape to a tRNA.
Arrows indicate sites of amino acylation and β-lysylation for tRNA and EF-P respectively.
Figure is shown as it appears in reference 80 and the publisher states that they allow for its use in
this thesis without special permission. B) EF-P occupies a unique position between the peptidyl
(P) and exit (E) sites. Figure is reproduced with permission from reference 79. PTC, peptidyl
transferase center. I, II and III indicate domains of EF-P. C) Modification of EF-P by YjeK and
PoxA. Figure is shown as it appears in reference 83 and the publisher states that they allow for its
use in this thesis without special permission.
A.
B.
C.
12
1.4.2 Post-translational modification of EF-P
In collaboration with Dr. Michael Ibba at the Ohio State University, Dr. Navarre and a
former graduate student in the lab, Dr. Betty Zou, determined that YjeK and PoxA coordinate to
post-translationally modify EF-P74,83-85. Specifically, YjeK synthesizes (R)-β-lysine, which PoxA
attaches to lysine 34 of EF-P to form a unique lysyl-(R)-β-lysine moiety (Figure 3C).
Importantly, this modification extends towards the peptidyl transferase centre of the ribosome to
modulate peptide bond formation. This unique β-lysyl modification is critical for the function of
EF-P in E. coli and Salmonella. Recently it was discovered that the YfcM protein further
hydroxylates the β-lysyl moiety of modified EF-P, however the role of this hydroxylation is
unclear, as it appears to be dispensable for EF-P activity under all conditions tested thus far86,87.
1.4.3 PYE mutants display pleiotropic phenotypes
The Navarre lab has collectively termed PoxA, YjeK and EF-P the “PYE pathway”. As
might be expected, individual Salmonella PYE mutants phenocopy one another and display a
wide variety of pleiotropic phenotypes74,84. Of particular note are the results of Biolog Phenotype
MicroarraysTM 88, which show that Salmonella strains containing defects in the PYE pathway
display misregulated metabolism and increased respiration when growing on certain nutrient
sources, and yet these mutants are hypersusceptible to a range of cellular stressors, including
hypo-osmolarity, detergents, and many pharmacologically unrelated antibiotics. PYE mutants
also display reduced motility and virulence. Analysis of membrane proteins revealed that efp
mutant Salmonella greatly overexpresses kdgM encoding an outer membrane porin84. This may
influence the permeability of PYE mutants and proposes the hypothesis that they are
hypersusceptible to stress due to increased permeability of their membranes.
1.4.4 EF-P does not appear to be a general translation factor
The question of what role EF-P plays in translation remained unresolved until very
recently, and most of the progress studying this factor occurred in our lab and others during the
course of my time as a student. Evidence including the analysis of membrane proteins mentioned
13
above suggested that EF-P is not a general translation factor but rather plays a role in the
translation of specific proteins. The notion that EF-P affects the production of a limited set of
proteins was further supported by two-dimensional difference in gel electrophoresis (2D-DIGE)
analysis of a Salmonella poxA mutant74. In this method, total cellular proteins from wild type and
poxA mutant Salmonella are labeled with fluorescent dyes, mixed in equal amounts, and then
separated by 2D-polyacrylamide gel electrophoresis (PAGE) prior to analysis by mass
spectrometry. Using this technique, Betty Zou was able to determine that only a relatively small
subset of proteins are affected by perturbations in the PYE pathway – a finding in agreement
with earlier work on the efp mutant of Agrobacterium89. However, she could only
unambiguously identify a small number of these proteins due to crowding on the 2D gel. This
warranted further investigation into the EF-P-dependent proteome by more sensitive, quantitative
and comprehensive methods.
1.5 Thesis rationale
When I began my doctoral research, EF-P was known to be a universally conserved
translation factor that resulted in a range of pleiotropic phenotypes when deleted from a number
of bacteria, including Salmonella. However, the mechanisms underlying these phenotypes were
uncharacterized. Data demonstrating the overexpression of outer membrane porins implicated a
potential role for increased permeability in the hyper-susceptibility of PYE mutants to stress. As
well, the increased respiration when growing on specific nutrient sources suggested that PYE
mutants may also misregulate their metabolism, potentially limiting their ability to adapt to
stress. Furthermore, preliminary proteomic analyses suggested that EF-P is not a general
translation factor but rather impacts the synthesis of a specific subset of proteins. Yet the
mechanism of how it influences translation and why only specific proteins require it remained
unknown. In light of these observations, I began my graduate work in January of 2011 with the
goal of characterizing the EF-P-dependent proteome of Salmonella and determining the
mechanism by which it influences the translation of a specific subset of proteins. Moreover I
aimed to ascertain the basis of PYE mutant phenotypes and the role of EF-P in Salmonella
metabolism, stress resistance, and virulence.
14
2 Methods
2.1 Bacterial strains
2.1.1 Salmonella strain background and efp mutants
(Used in all chapters)
With the exception of the SILAC experiment and the SGSC collection, all Salmonella
strains are derivatives of Salmonella enterica subsp. enterica serovar Typhimurium (S.
Typhimurium) strain 14028s. Lambda red recombination was used to generate all of the gene
knockout mutants in this background84,90. Two efp mutant strains were generated by Dr. Betty
Zou: the original mutant, WN934, is a deletion for the entire efp open reading frame including
part of the adjacent yjeK promoter, effectively creating a double mutant. In contrast, WN1405 is
a deletion of the majority of the efp open reading frame (basepairs 145 to 424) that leaves the
yjeK promoter intact. In this thesis, WN934 is the efp mutant in the original permeability assays
(Figure 4). For all other data (with the exception of SILAC) the Salmonella efp mutant is
WN1405.
For SILAC I employed strain WN1269, an arginine and lysine auxotroph (ΔargH ΔlysA)
derivative of S. Typhimurium strain SL1344 that was utilized in an earlier study and was a
generous gift from Dr. Brian Coombes and Dr. Colin Cooper91. I confirmed the auxotrophy by
testing growth in minimal media in the absence of lysine or arginine. I transferred the efp
mutation from WN934 into SL1344 by transduction using the HT105/1 int-201 derivative of
phage P22 to generate strain WN130892.
2.1.2 Chromosomal replacement of the dctA promoter
(Used in Chapters 5 and 6)
I generated chromosomal swaps to replace the dctA promoter in the Salmonella
chromosome with the E. coli dctA promoter. I first inserted the 500bp upstream of the E. coli
dctA start codon into the pXG10sf plasmid upstream of the full length Salmonella dctA ORF and
downstream of the chloramphenicol resistance cassette used to maintain the plasmid93. I then
used PCR to amplify from the end of the dctA ORF to the end of the chloramphenicol resistance
15
cassette and including overhangs homologous to the Salmonella genome. I then inserted the
fragment into the Salmonella chromosome by lambda red recombination and subsequent
transduction using p22HT phage.
2.1.3 SGSC and ECOR collections
(Used in Chapter 5)
The Salmonella genetic stock centre (SGSC) collection contains a range of Salmonella
strains isolated from a variety of sources and maintained by the University of Calgary SGSC. I
employed three subgroups: The SARA collection contains 72 strains from 5 serovars within
subspecies I (including Salmonella Typhimurium) allowing for a more thorough analysis with
many strains from each serovar94. The SARB collection contains 37 strains including a more
broad range of serovars from subspecies I95. Finally the SARC collection contains two isolates
from each Salmonella enterica subspecies and also of Salmonella bongori96. The E. coli
reference (ECOR) collection is a collection of 72 E. coli strains from various sources that are
commonly used as a sample spanning the genetic diversity of E. coli97. E. coli K12 is a common
lab strain and specifically BW25113 was the strain used as the genetic background for the
generation of the Keio collection98.
2.2 Plasmids
2.2.1 GFP translational fusion plasmids
(Used in Chapter 4)
I generated plasmids used for the PoxB translational fusion assay by PCR amplification
of the 5’ UTR (28 base pairs upstream of the start codon) and full-length (excluding the stop
codon) or C-terminally truncated poxB gene using S. Typhimurium strain 14028s genomic DNA
as a template. I inserted these amplicons into the NsiI and NheI sites of pXG10sf to generate
translational fusions to “super-folder” green fluorescent protein (sfGFP)93,99. The plasmid
employs the constitutively active PLtet0-1 promoter and the tightly controlled pSC101 low copy-
number origin of replication to ensure minimal variation in transcript levels. I generated point
16
mutations by site directed mutagenesis of the full-length PoxB construct. Plasmids are named
according to the truncation site in poxB (i.e. the PoxB 68 construct is truncated after codon 68 of
poxB) or the location of a specific point mutation in full-length poxB (e.g. PoxB P76L contains a
proline to leucine mutation at codon 76). LacZ186, included as a control, consists of the first 186
codons of the E. coli lacZ gene inserted into the NsiI and NheI sites of the pXG10sf plasmid,
similar to the PoxB fusions.
I constructed the plasmids used for the AtpD and AtpA translational fusion assay in a
similar manner as the PoxB fusion plasmids with one main exception: Since AtpD and AtpA are
translated form a single mRNA encoding all eight ATP Synthase subunits, I did not include the
entirety of the upstream mRNA in the plasmid. Instead I included only the 74 or 75 base-pairs
(for AtpD and AtpA respectively) upstream of the ATG start codon as a 5’ UTR. I generated
mutations and swaps using Gibson isothermal assembly cloning with overlapping primers100.
Plasmids are named according to the regions that are swapped from AtpA into AtpD-sfGFP or
vice versa (i.e. the ‘AtpD - AtpA UTR’ construct encodes full-length AtpD-sfGFP with the 5’
UTR from atpA).
2.2.2 DctA overexpression plasmid
(Used in Chapter 5 and 6)
I made the constitutive dctA expression plasmid by inserting the full length dctA 5’ UTR
and ORF into pXG10sf under the control of the constitutively active PLtet0-1 promoter. I
generated the IraP complementation plasmid by inserting the iraP ORF and the upstream 300bp
into pXG10sf. Since the plasmid and the original dctA and iraP mutant strains employ
chloramphenicol resistance, the chloramphenicol cassette was first flipped out of the dctA and
iraP mutants using the pCP20 plasmid encoding FLP recombinase101. The heat-unstable pCP20
plasmid was eliminated from the strain by passaging overnight at 42°C.
17
2.2.3 STM3120 transcriptional fusion plasmid
(Used in Chapter 6)
To test the induction of the itaconate degradation operon (STM3120 to STM 3117) I
generated a plasmid construct with the STM3120 promoter (P3120) driving expression of sfGFP.
Specifically I inserted P3120 (the 333bp upstream of the STM3120 start codon and thereby
including the 5’ UTR and promoter to 25bp upstream of the predicted -35 box) into pXG10sf to
replace the PLtet0-1 promoter and drive expression of LacZ186-sfGFP93. For the STM3121-P3120
construct, I extended the same promoter to 1570bp upstream of the STM3120 start codon to
include the entire STM3121 ORF and a predicted transcriptional terminator following it.
For fluorescence microscopy I modified the pXG10sf plasmid by inserting constitutively
expressed (PLtet0-1 promoter) mCherry into a transcriptionally independent region of the
plasmid. I designated this plasmid ‘independent constitutive mCherry’ or pICM. To generate
pICM-P3120, I inserted the constitutive mCherry into the pXG10sf-P3120 plasmid in the same way.
The same was true for the pICM-STM3121-P3120 plasmid. Thus for the pICM plasmids, mCherry
is constitutively expressed, whereas sfGFP levels are driven by the promoter in question.
2.3 Experimental Methods
2.3.1 N-Phenyl-1-Naphtylamine accumulation assay
(Used in Chapter 3)
This protocol was adapted from others previously established to measure the intrinsic
permeability of bacterial membranes102-104. LB broth containing appropriate antibiotics was
inoculated from overnight cultures of the test strains. The cells were grown to early stationary
phase (10 hours), centrifuged (13,000 rpm for 1 min) and washed twice in the assay buffer (5
mM HEPES pH 7.2, 137 mM NaCl). The cells were resuspended in buffer and the OD600 was
adjusted to 1.0. Fractions from these suspensions were diluted and plated in triplicate to acquire
CFU counts. 100 μl were also loaded into wells of a black 96-well fluorescence microplate
(Greiner Bio-One). Controls containing only buffer were included and three technical replicates
were conducted. A 20 μM solution of NPN (Sigma-Aldrich) was prepared in the assay buffer just
prior to experimentation and 100 μl was added to appropriate wells of the microplate, yielding a
18
final bacterial OD600 of 0.5 and a final NPN concentration of 10 μM. Controls were included in
which buffer was added instead of dye. Immediately after the addition of the dye, the plate was
inserted into a Tecan Infinite M200 microplate reader (a delay of approximately 15s) and was
read at excitation and emission wavelengths of 355 nm and 402 nm, respectively. Readings were
taken every 25 seconds for 10 minutes. Analysis of the fluorescence values was conducted using
Microsoft Excel. Background fluorescence (NPN in buffer only) was subtracted from the raw
values and the result was divided by the number of CFU in the corresponding sample. Finally,
the fluorescence value of wild-type Salmonella at time 0 was defined to be 100% and all other
values were normalized accordingly.
2.3.2 Ethidium Bromide (EtBr) Accumulation Assay
(Used in Chapter 3)
EtBr uptake was measured very similarly to NPN uptake with several differences: The
buffer used was 50 mM KH2PO4 pH 7.0, 137 mM NaCl. As well, cells were diluted to an OD600
of 0.4 (final OD600 of 0.2 after addition of EtBr). The concentration of the stock EtBr solution
was 12 μM (final concentration of 6 μM after adding to cells). EtBr fluorescence was read at
excitation and emission wavelengths of 545 nm and 600 nm, respectively. Finally, since wild-
type fluorescence values were at background levels, the value for wild-type at time 0 was zero
for some trials. The values could therefore not be represented as a percentage of wild-type at
time 0. The data is instead presented relative to the acrAB mutant at time 0.
2.3.3 Stable isotope labelling of amino acids in cell culture (SILAC)
(Used in Chapter 3)
For SILAC, I supplemented MOPS minimal media with amino acids at the
concentrations previously described91 and used 0.2% (w/v) glycerol as a carbon source. For
heavy isotope labeled samples arginine and lysine were replaced with 13C6-Arg and 2H4-Lys
isotopes at the equivalent molar concentration. WN1269 (wild-type) was grown in heavy
19
arginine and lysine isotopes and WN1308 (Δefp) was grown in light isotopes in this media for
16h to ensure complete labeling of all proteins. Strains were subsequently subcultured 1:200 into
the same media and grown to an optical density (600nm) of 0.5 (mid-log phase) at which point
the cells were harvested by centrifugation. The pellets were subjected to lysis by heating to 99°C
for 5 minutes in fresh lysis buffer (1% deoxycholate [DOC] in 50mM ammonium bicarbonate
[NH4HCO3] at pH 8). Cell debris was removed by centrifugation at 13,000 x g for 15 minutes
and the supernatant was frozen at -80°C until used.
2.3.4 Mass spectrometry and proteomic data analysis
(Used in Chapter 3)
The mass spectrometry and data analysis described here were conducted by Kyung-Mee
Moon and Dr. Leonard Foster. I have included the method in this thesis as background to clarify
how the SILAC analysis data was produced. For analysis of the isotope-labeled lysates, 30μg of
protein from each of WN1269 (heavy) and WN1308 (light) were combined, fractionated into 12
pieces by gel electrophoresis and in-gel trypsin digested following a previously outlined
procedure105. The resulting peptides were subjected to liquid chromatography coupled to tandem
mass spectrometry using an Orbitrap XL (Thermo Scientific) as described previously106. Data
analysis was conducted using the MaxQuant software107 to generate an average normalized
heavy / light ratio over three biological replicates and Significance B values were calculated
using Perseus108. To determine significance, we used the cutoff of a Significance B score of less
than 0.01 in at least one trial. This statistic measures significance within a single trial even for
proteins that were only identified in one the three biological replicates. When we assess
significance using the average of Significance B scores across all three trials (excluding scores
where the protein of interest was not identified in a given trial), we obtain similar percentages of
proteins containing EF-P target motifs that show significantly altered abundance. For example,
using an average Significance B score of less than 0.05, we identified 107 proteins with
significantly altered abundance. 61 of these are greater than two-fold less abundant in the efp
mutant and 25 of these 61 contain a PPP, PPG or APP motif.
20
2.3.5 DAVID analysis
(Used in Chapter 3)
Lists of proteins affected by EF-P were generated based on the SILAC data and/or on the
presence of putative EF-P target motifs in Salmonella proteins. The proteins identified in SILAC
had gene identification numbers corresponding to the SL1344 strain used. I identified each
corresponding homolog gene identification number in the S. Typhimurium strain LT2 so that I
could upload the gene lists to the Database for Annotation, Visualization and Integrated
Discovery (DAVID) online analysis software109. The program compares a list of genes with
functional annotation databases including GO terms, KEGG pathways and SP-PIR keywords
amongst others. I employed the Functional Annotation Tools to determine which functional
groups were overrepresented amongst my query genes. DAVID employs a one-tailed Fisher
Exact Probability Test to calculate p-values of individual annotation groups and cluster p-values
are generated as the geometric mean of the p-values of all constituent groups. Functional
annotation clusters with p-values less than 0.05 are shown in figures.
2.3.6 Translational fusion assay
(Used in Chapter 4)
Wild-type (S. Typhimurium str. 14028s) and the isogenic Δefp mutant carrying plasmids
encoding poxB, atpD or atpA translational fusions to sfGFP were grown in MOPS minimal
media supplemented with 0.2% (w/v) glucose and 20μg/mL chloramphenicol. Growth was
conducted for 16 hours at 37°C in a TECAN Infinite M200 microplate reader with constant
aeration. Optical density (OD600nm) and GFP fluorescence readings (excitation and emission
wavelengths of 475nm and 511nm respectively) were taken every 15 minutes. For both OD600
and GFP, background values taken from no-cell controls were subtracted from all readings. For
clarity the wild-type / Δefp fluorescence ratios are displayed as GFP fluorescence per OD600 unit
at 10 hours of growth. Results obtained when cultures were measured at 6, 8, 12, or 16 hours of
growth or in mid-log phase were similar to those at 10 hours110.
21
2.3.7 Reverse transcriptase quantitative PCR (RT-qPCR)
(Used in Chapter 4)
Wild-type and efp mutant Salmonella containing the translational fusion plasmids were
grown to an OD600 of 0.5 (OD600 ≈ 0.2 as read in TECAN Infinite M200) in MOPS minimal with
0.2% glucose as a carbon source and supplemented with chloramphenicol (20μg/mL). 1.5 ml of
each sample was pelleted in RNA Protect Bacterial Reagent (Qiagen) according to
manufacturer’s instructions and stored at -80°C. Subsequent RNA preparations were performed
using the Aurum Total RNA Mini kit (BioRad). Reverse transcription was performed using the
iScript cDNA synthesis kit (BioRad) using random hexamer primers. The cDNA generated was
used for quantitative PCR analysis using iQ SYBR Green Mix (BioRad) according to the
manufacturer’s protocol. To measure poxB-sfGFP mRNA levels at two points on the transcript I
used primers specific to either poxB or sfGFP. The transcript of 16S rRNA was used as an
internal standard for normalization. Since the strains also encode poxB in their genomes, the
mRNA levels obtained for the lacZ186 construct with the poxB-specific primers was subtracted
as background from all other strains. Under the conditions tested this background poxB mRNA
level was negligible (greater than 300 fold lower) compared to any of the strains expressing
PoxB-GFP from a plasmid.
2.3.8 Immunoblotting
(Used in Chapter 4)
Salmonella strains were grown in MOPS minimal media with 0.2% glucose as the carbon
source to mid log phase (OD600 ~ 0.5), washed twice in wash buffer (1 mM Tris pH 8.0, 5 mM
magnesium acetate) and lysed by sonication in lysis buffer (9.32 M urea, 2.67 M thiourea, 40
mM Tris, 86.78 mM CHAPS, pH 8.5). 10μg of total cell lysate was mixed with 2x SDS loading
buffer and boiled for 10 min at 95ºC. Proteins were separated by SDS-PAGE and transferred
(semidry) to a nitrocellulose membrane. Following 1 h blocking at room temperature in 5% milk
in TBST (1x Tris-buffered saline, 0.05% Tween 20), immunoblotting was conducted overnight
in TBST +5% milk at 4ºC using a mouse anti-DnaK antibody (1:50,000; Enzo Life Sciences) and
a mouse monoclonal antibody specific for the beta subunit of E. coli ATP Synthase (1:1000;
22
MitoSciences) or a mouse monoclonal antibody specific for GFP (1:1000; Santa Cruz
Biotechnology). Blots were washed and subsequently incubated for 1 h at room temperature with
HRP-fused goat anti-mouse antibody (1:10,000 in TBST +5% milk) for ECL imaging (Thermo
Scientific). Quantification of AtpD protein levels relative to DnaK was done using Image Lab
software (Bio-Rad Laboratories).
2.3.9 Curve Fitting
(Used in Chapter 4)
Fluorescence data was fit to a ‘1-exp’ curve of the form 𝑦 = 𝐴(1 − 𝑒−𝐵𝑥) by the
minimum sum of chi-squares method: For each observed wild-type fluorescence value (x-
coordinate), I calculated chi-square values by comparing the corresponding observed ∆efp
fluorescence to the y-value calculated by the equation. The chi-squares for all constructs with the
same ORF were summed and the values of A and B yielding the minimal sum of chi-squares was
solved using Microsoft Excel’s ‘Solver’ function. Coefficient of determination (R2) comparing
observed ∆efp data and y-values predicted by the solved 1-exp equation was calculated using
Excel’s ‘RSQ’ function.
2.3.10 Growth using dicarboxylates as a sole carbon source
(Used in Chapter 5)
Strains were grown overnight in LB media inoculated from single colonies. The next day
the cultures were centrifuged, the LB was removed and the cells were resuspended in MOPS
minimal media with no carbon source to an optical density (OD600) of approximately 1.75. This
suspension was used to inoculate (1/200 dilution) MOPS minimal media containing 0.2% (unless
otherwise indicated) succinate as the sole carbon source. For some experiments, carbon sources
other than succinate were used and these are indicated. Growth was conducted in a TECAN
Infinite M200 plate reader at 37ºC with shaking and OD600 was read every 15 minutes. Of note,
for salts and hydrates of carbon sources the final concentration reflects the percent of the carbon
23
source itself; for example 0.2% succinate was made as 0.47% sodium succinate dibasic
hexahydrate.
The SGSC and ECOR collections screen was conducted by testing 47 strains per run in a
96-well plate such that each strain had two technical replicates and wild-type and rpoS mutant
Salmonella could be included on every plate as quality controls. Each strain was tested on at
least three separate days. To identify the first timepoint for each technical replicate where the
OD600 surpassed 0.1, I employed Excel’s Match and Index functions and values from at least two
technical replicates were averaged for each biological replicate.
2.3.11 Catalase assay
(Used in Chapter 5)
For each replicate of the SGSC and ECOR collections screen, I tested each strain for
catalase activity as an analog for RpoS function111. Following inoculation of the LB overnight
cultures used as inoculum, I spotted 10μl of each culture onto an LB plate. The next day these
spots were tested for catalase activity by the addition of 10μl of hydrogen peroxide. Bubbling
was scored compared to wild-type (catalase positive) and rpoS mutant (catalase negative)
Salmonella.
2.3.12 Acidified media survival assay
(Used in Chapter 6)
LPM media was made as described previously91 and succinate or itaconate were added to
either 0.2% or 0.4% as indicated in figures. The pH of the media was then adjusted to 4.4. For
the assay, 0.1 OD units (the equivalent of 100μl of culture at an OD600 of 1) of an LB overnight
culture was centrifuged and the supernatant was removed. The pellet was resuspended in 1ml of
acidified LPM media by vortexing for 15 seconds in a 1.5ml tube. The tube was then incubated
in a 37ºC water bath. At time points the tube was vortexed for 10 seconds and 5 spots of 10μl
were plated for colony forming units (CFU). An additional 10μl was also serially diluted and
24
plated for CFU (5 spots of 10μl at each dilution). To allow for multiple test strains while
maintaining exact timepoints, tubes were staggered by 1 minute to allow for the sampling of one
strain prior to the next sample reaching the same timepoint. For strains including a plasmid,
antibiotics were included in the overnight culture used as an inoculum but were not included in
the challenge media. Since I found that residual chloramphenicol from the overnight culture
increased survival in this assay, I generated ampicillin resistant versions of the plasmids as
described in Bacterial Strains and Plasmids. All plasmid-containing strains used in this assay
contained the ampicillin resistant version of the plasmid.
2.3.13 Promoter induction fluorescence assay
(Used in Chapter 6)
I tested induction of the Salmonella STM3120 promoter using a transcriptional fusion of
P3120 to sfGFP in either the pXG10sf or pICM plasmid. Data from the two plasmids were
combined as the inducible region is identical and the plasmids only differ in the constitutively
active mCherry expressed independently from pICM. Overnight LB cultures were used to
inoculate (1/200 dilution) either LB, LPM, or MOPS minimal media containing 0.2% of the
indicated carbon source. Supplements for testing fluorescence induction were added to a
concentration of 0.2%. Of note, for salts and hydrates of carbon sources the final concentration
reflects the percent of the carbon source itself; for example 0.2% succinate was made as 0.47%
sodium succinate (dibasic) hexahydrate. Growth was conducted in a TECAN Infinite M200 plate
reader at 37ºC with shaking and OD600 and GFP fluorescence (475nm and 511nm excitation and
emission wavelengths respectively) were read every 15 minutes. For clarity, bar graphs show
fluorescence at 16h post inoculation. Chloramphenicol was included in all media at a
concentration of 20μg/ml to maintain the plasmids.
25
2.3.14 Macrophage survival assay
(Used in Chapter 6)
The THP-1 human monocyte cell line and the J774 mouse macrophage cell line were
maintained in RPMI Medium 1640 (with L-glutamine) supplemented with 10% FBS and 1%
Glutamax, and grown at 37ºC and 5% CO2. For infection assays, THP-1 cells were seeded in 96-
well plates at 50,000 per well with 50nM PMA (phorbol 12-myristate 13-acetate) added to the
media to induced them to become adherent macrophage. After 48h, the media was replaced with
normal growth media (no PMA) overnight. For infections with J774 macrophage the cells were
seeded in 96-well plates at 50,000 per well overnight in normal growth media. Salmonella were
added onto seeded cells in RPMI growth media at a multiplicity of infection (MOI) of
approximately 20 bacteria to 1 macrophage and centrifuged for 10 minutes at 1000rpm for
maximum cell contact. After centrifuging the plate was placed at 37ºC (5% CO2) and this was
called ‘time 0’. After 30 minutes, non-adherent Salmonella were washed off by three washes
with PBS followed by replacement with fresh media containing 100 μg/ml gentamicin to kill
extracellular Salmonella. At 2 hours the media was replaced with media containing gentamicin at
10 μg/ml. At timepoints, media was replaced with PBS containing 1% Triton X-100 and
incubated at room temperature for 5 minutes. Following resuspension by vigorous pipetting,
samples were serially diluted. For each dilution, five 10μl spots were plated for CFU counting.
Each sample included three separate wells as technical replicates and CFU for that biological
replicate was the average from each well (a total of 15 x 10μl spots counted per sample).
2.3.15 Fluorescence microscopy
(Used in Chapter 6)
Fluorescence microscopy was conducted similarly to the macrophage survival assay with
some exceptions: Cells were seeded in 24-well plates containing glass coverslips at 125,000 per
well. Bacteria were infected at an MOI of approximately 100 to maximize the instance of
macrophage that contain bacteria. At timepoints the media was removed and cells were washed
three times with PBS. They were then fixed for 10 minutes at room temperature in PBS
containing 4% paraformaldehyde (PFA). Following three more PBS washes the cells were
26
permeabilized for 10 minutes in PBS + 0.2% Triton X-100 + 1% BSA. Coverslips were washed
again with PBS, dipped in water, and edges dabbed to remove excess liquid. They were then
mounted on slides using 3μl mounting media containing DAPI and allowed to dry overnight in
the dark. Coverslips were then sealed to the slide with nail polish and allowed to dry for 1h in the
dark. Slides where viewed using a Zeiss Observer.z1 microscope using a 100x oil immersion
objective and the Zeiss Zen microscopy software. Images were taken with a Zeiss Axiocam 506
mono camera mounted on the microscope. For viewing, a positive control constitutively
expressing GFP and mCherry was used to establish exposure times for all fluorophores and the
same exposure time was employed for all samples. For the pICM-PSTM3120 experiment presented
here I employed a 2 second exposure for mCherry and 1 second exposure for sfGFP. For
quantification, ImageJ software was employed to calculate fluorescence intensities in the red and
green channels and a ratio was generated for greater than 125 bacteria per test condition across
three biological replicates.
2.3.16 Identification of STM3121-STM3117 orthologs in Salmonella serovars
(Used in Chapter 6)
To explore the presence of the itaconate degradation operon (STM3120-STM3117) and
its regulator (STM3121) across many serovars of Salmonella, I employed the Prokaryotic
Genome Analysis Tool (PGAT)112. Specifically I used the Ortholog Search Tool to identify the
presence or absence of the operon in the 69 Salmonella strains in the database. The Synteny
Mapper was used to align the operon region of select genomes.
27
3 Analysis of the Salmonella efp mutant proteome
Primary references
Much of the data presented in this chapter is published work reproduced with permission from:
Zou SB, Hersch SJ, Roy H, Wiggers B, Leung AS, Buranyi S, Xie JL, Dare K, Ibba M,
Navarre WW. Loss of elongation factor P disrupts bacterial outer membrane integrity. J
Bacteriol. 2012;194(2):413-425. 84
Bullwinkle TJ, Zou SB, Rajkovic A, Hersch SJ, Elgamal S, Robinson N, Smil D,
Bolshan Y, Navarre WW, Ibba M. (R)-β-lysine-modified elongation factor P functions in
translation elongation. J Biol Chem. 2013;288(6):4416-4423. 87
Hersch SJ, Wang M, Zou SB, Moon K-M, Foster LJ, Ibba M, Navarre WW. Divergent
protein motifs direct elongation factor P-mediated translational regulation in Salmonella enterica
and Escherichia coli. mBio. 2013;4(2). 110
Acknowledgements
Dr. Betty Zou generated the efp mutant strains used throughout this work as well as the
yfcM mutant. She also identified that kdgM was overexpressed in the efp mutant, generated the
Δefp ΔkdgM double knockout strain and conducted experiments assessing the role of KdgM in
stress susceptibility. I conducted the experiments assessing the permeability of these strains.
I conducted the SILAC experiment in collaboration with Dr. Leonard Foster and Kyung-
Mee Moon who performed the LC-MS/MS and initial analysis to generate protein abundance
ratios from the mass spectrometry output data. I prepared the labelled Salmonella protein lysates
and later conducted further data analysis.
Mengchi Wang in the lab of our collaborator Dr. Michael Ibba conducted the analysis of
the SILAC dataset that identified APP as an additional tripeptide polyproline motif showing EF-
P-dependence in vivo.
28
3.1 Overview
PYE mutants display similar pleiotropic phenotypes, including reduced growth in rich
media, reduced motility, altered metabolism on certain nutrient sources and hypersusceptibility
to an array of cellular stressors and antimicrobial compounds74,84. Previous proteomic analyses of
poxA and efp mutants suggested that EF-P acts in a targeted manner rather than as a global
translation factor, implying that EF-P is required for the efficient translation of specific
transcripts and that failure to efficiently produce these proteins in an efp mutant could underlie
the phenotypes observed74,84,89,113. In this chapter I present data showing that efp mutant
Salmonella demonstrates increased permeability that is partially due to kdgM overexpression.
However, concomitant work conducted by Dr. Betty Zou indicated that deletion of kdgM does
not affect other Δefp phenotypes such as increased stress susceptibility – implicating a role in
these phenotypes for additional unidentified EF-P-dependent proteins.
I conducted an unbiased in vivo analysis of the Salmonella efp mutant using stable-
isotope labeling of amino acids in cell culture (SILAC). This revealed a subset of proteins with
significantly altered abundance in the efp mutant, and I further conducted a functional annotation
analysis highlighting ontological groups that were overrepresented amongst these proteins. Some
of the proteins and functional groups showing altered steady state levels in the efp mutant can
provide a parsimonious explanation for the observed PYE mutant phenotypes. Shortly after the
completion of this SILAC dataset, two publications from other labs employed in vitro techniques
to demonstrate that EF-P rescues the translation of ribosomes that are stalled at particular
tripeptide motifs, specifically Proline-Proline-Proline (PPP) or Proline-Proline-Glycine
(PPG)114,115. These findings established that EF-P functions to rescue ribosomes stalled at
specific motifs and enabled me to compare the presence of these proline containing motifs with
my in vivo SILAC dataset of EF-P-dependent proteins. Interestingly, my analysis reveals that the
presence of a polyproline motif is not always necessary or sufficient to render a protein
dependent on EF-P for its efficient translation in vivo.
29
3.2 Results
3.2.1 Salmonella efp mutants are selectively more permeable to NPN
The hypersensitivity of PYE mutants to a broad range of cellular stressors suggests that
they may have an altered cell envelope resulting in increased membrane permeability. I tested
the permeability of the S. Typhimurium (Str. 14028s) efp mutant (WN934) by measuring the
uptake of the fluorescent probes ethidium bromide (EtBr) and N-phenyl-1-naphthylamine
(NPN), two dyes that fluoresce upon entering the cell but for different reasons103,104,116. EtBr will
not fluoresce unless it permeates both the inner and outer membranes to intercalate into nucleic
acid. In contrast, NPN is a nonpolar dye that will fluoresce strongly in a hydrophobic
environment such as the lipid bilayer of the inner membrane or the inner leaflet of the outer
membrane. The outer leaflet of the outer membrane of gram negative bacteria is composed of
lipolysaccharide (LPS), which forms a tightly packed semicrystalline structure that effectively
excludes both hydrophilic and hydrophobic compounds, including NPN103,117.
I found that uptake of NPN was significantly higher in the efp mutant than in wild-type
Salmonella or a complemented strain expressing exogenous efp from a plasmid (Figure 4A). This
uptake happened rapidly and equilibrium was almost achieved within the 10 to 15 seconds
between adding the dye and measuring fluorescence. In contrast, I observed no significant
difference in permeability to EtBr between wild-type and efp mutant Salmonella (Figure 4B).
This was not a systematic error of the method, as an isogenic strain lacking the AcrAB multidrug
efflux pump served as a positive control. While the strain lacking AcrAB is expected to exhibit
the same rate of dye influx as wild-type Salmonella, a decrease in the cell’s efflux capability
results in the accumulation of the dye and subsequent fluorescence. As expected, this acrAB
mutant showed increased uptake of both NPN and EtBr, but notably achieved equilibrium at a
slower rate than the efp mutant. These kinetics suggest that, although the efp strain contains a
functional acrAB efflux system, increased membrane permeability enhances its rate of dye influx
such that it exceeds the capacity of the efflux system.
30
Figure 4: Salmonella efp mutants are more permeable to NPN but not EtBr. Fluorescent dye
uptake for NPN (A) or EtBr (B) by the S. Typhimurium 14028s wild-type, efp mutant (Δefp
WN934), and complemented (Δefp pEF-P pYjeK) strains. The isogenic ΔacrAB strain was
included as a positive control. Dyes were added at time 0 with a 10 to 15 second delay while the
plate was inserted into the plate reader. Data shows fluorescence per CFU normalized to the
wild-type strain at time 0 (for NPN) or the ΔacrAB strain at time 0 (for EtBr). Values are the
average of four biological replicates. p-values indicated in the legend were calculated using a
two-tailed t-test (assuming unequal variance) comparing strains to the wild-type at 10 minutes (*
p < 0.05, ** p < 0.01, *** p < 0.001). Figure is as it appears in Zou, Hersch et al. (2012), J
Bacteriol 84.
31
3.2.2 Deletion of yfcM has no effect on NPN permeability
A third EF-P-modifying protein called YfcM was recently identified that acts after PoxA
and YjeK to hydroxylate modified EF-P in E. coli86. However despite its role in post-
translationally modifying EF-P, and in contrast to the PYE mutants, no phenotypes have yet been
observed for yfcM deletion mutants in either Salmonella or E. coli87. Indeed I find that in contrast
to efp, deletion of the Salmonella yfcM gene has no effect on NPN permeability (Figure 5). These
data suggest that the hydroxylation modification performed by YfcM is not necessary for EF-P
function.
3.2.3 Deletion of kdgM partially complements the permeability defect
By examining altered protein levels in membrane fractions Dr. Betty Zou identified
specific proteins showing altered abundance in an efp mutant. In particular, an outer membrane
porin called KdgM was much more highly expressed in the efp mutant than in wild-type
Salmonella. To address the possibility that the increased permeability that I observed for efp
mutants was due to the overexpression of kdgM, I tested the NPN permeability of the kdgM
mutant and efp kdgM double mutant. I observed a partial complementation of the efp mutant
permeability defect when the kdgM gene was also deleted (Figure 6). This finding supports the
hypothesis that the increased permeability of the efp mutant is at least partially due to the
overexpression of kdgM. In parallel, Betty tested the efp kdgM double mutant under stress
conditions, such as sensitivity to the aminoglycoside antibiotic gentamicin and the zwitterionic
detergent lauryl-sulfobetaine. Despite the role that I found for kdgM in permeability, Betty
observed no change in the double mutant compared to the efp single mutant for resistance to
either compound or for any of the other phenotypes tested. Cumulatively these data suggest that
overproduction of KdgM partially accounts for the increased permeability of the efp mutant but
the mechanisms underlying the other PYE mutant phenotypes cannot be attributed to it and likely
involve other EF-P-dependent proteins.
32
Figure 5: Deletion of yfcM has no effect on permeability to NPN. As in Figure 4A,
fluorescent dye uptake for NPN by the S. Typhimurium 14028s wild-type, efp mutant (Δefp),
yfcM mutant (ΔyfcM) and efp yfcM double knockout (Δefp ΔyfcM) strains. The isogenic ΔacrAB
strain was included as a positive control. Dyes were added at time 0 with a 10 to 15 second delay
while the plate was inserted into the plate reader. Data shows NPN fluorescence in arbitrary units
(AU) normalized to one million colony forming units (CFU). Values are the average of four
biological replicates and error bars showing one standard deviation are shown at 0, 2.5, 5 and 10
minutes. Figure is as it appears in Bullwinkle, Zou, Rajkovic, Hersch et al. (2013) J Biol
Chem87.
33
Figure 6: Deletion of kdgM partially complements the efp mutant permeability defect. As in
Figure 4A, fluorescent dye uptake for NPN by the S. Typhimurium 14028s wild-type, efp mutant
(Δefp), kdgM mutant (ΔkdgM) and efp kdgM double knockout (Δefp ΔkdgM) strains. The
isogenic ΔacrAB strain was included as a positive control. Dyes were added at time 0 with a 10
to 15 second delay while the plate was inserted into the plate reader. Data shows fluorescence
per CFU normalized to the wild-type strain at time 0. Values are the average of four biological
replicates. p-values indicated in the legend were calculated using a two-tailed t-test (assuming
unequal variance) comparing strains to the efp mutant at 10 minutes (* p < 0.05, ** p < 0.01, ***
p < 0.001).
0%
100%
200%
300%
400%
500%
600%
700%
800%
900%
0 100 200 300 400 500 600
% F
luo
resc
en
ce
Time (seconds)
NPN
WT
efp-
1131-
efp- 1131-
acrAB-
WT **
Δefp
ΔkdgM ***
Δefp ΔkdgM *
ΔacrAB *
34
3.2.4 Identification of EF-P regulated proteins by SILAC
Previous proteomic analysis of a Salmonella poxA mutant using 2D-DIGE suggested that
only a relatively small subset of proteins were affected by perturbations in the PYE pathway – a
finding in agreement with earlier work on the efp mutant of Agrobacterium74,89. However, only a
limited number of proteins could be unambiguously identified due to crowding on the 2D gels.
To gain a more comprehensive view of the effect of EF-P on protein levels, I collaborated with
the lab of Dr. Leonard Foster to conduct stable isotope labeling of amino acids in cell culture
(SILAC) in conjunction with quantitative mass spectrometry-based proteomics to examine the
proteome of an efp mutant strain of Salmonella enterica Sv. Typhimurium strain SL1344 (strain
WN1308). I generated heavy isotope-labelled samples and we were able to detect, quantify and
identify a total of 1517 proteins, or approximately 34% of the 4514 proteins predicted to be
encoded in the S. Typhimurium strain SL1344 genome (Supplementary DataTable1). Using
Microsoft Excel, I identified proteins showing greater than two-fold altered abundance in the efp
mutant and further screened for a significant difference using a Significance B cutoff of 0.01
(Figure 7). By this criterion, 87 proteins showed changes of two-fold or greater and 28 displayed
a change of greater than ten-fold. Of the 87 proteins with significantly altered abundance, 49
showed a decrease in steady-state levels in the efp mutant strain and are more likely to be
directly dependent on EF-P owing to its characterized stimulatory effect on translation.
35
Figure 7: A subset of proteins show significantly altered abundance in Δefp Salmonella.
Histogram outlining the distribution of protein expression ratios identified in SILAC comparing
the efp mutant (WN1308) with the isogenic wild-type background strain (WN1269) such that
proteins showing increased abundance in the efp mutant are further left and those with decreased
abundance in the efp mutant are further right. Columns indicate the number of proteins with
average expression ratio between two neighbouring x-axis values. Note that underlined values in
the x-axis indicate a change in scale. The inset table shows the number of SILAC hits
demonstrating a greater than two-fold difference in protein level between the efp+ and ∆efp
strains. The second column further indicates the proteins with a Significance B value of less than
0.01 in at least one trial. Expression ratios shown are the average normalized heavy/light ratios
of three biological replicates. Figure is shown as it appears in Hersch et al. (2013), mBio 110.
36
3.2.5 EF-P rescues ribosomes stalled at polyproline motifs
Shortly after the completion of my SILAC proteomic analysis of efp mutant Salmonella,
two papers were published in Science demonstrating that EF-P rescues ribosomes stalled at
sequences of consecutive prolines114,115. Specifically, these articles identified two tripeptide
motifs, three consecutive prolines (PPP) or two prolines followed by a glycine (PPG), that
require EF-P for their efficient translation in vitro. They went on to demonstrate that the
ribosome stalls at these motifs and stalling is dependent on the amino acid combinations rather
than on specific codons. Furthermore they showed that in the absence of EF-P the ribosome fails
to efficiently make a peptide bond between the second and third amino acids of the motif,
resulting in the stall.
Their findings were supported by a computer script written by Mengchi to search my
SILAC data for over-represented tripeptide motifs within the ten percent of proteins showing the
greatest reduction of abundance in the efp mutant relative to wild-type. He identified PPP and
PPG motifs as the two most prominent tripeptide motifs amongst this group110. Moreover he also
identified Alanine-Proline-Proline (APP) as the third most prominent motif and the only other
polyproline-containing tripeptide in his top ten hits. This suggests that APP is another
polyproline motif that is dependent on EF-P for its efficient translation.
3.2.6 Polyproline motifs are not always necessary or sufficient to confer EF-P-dependence
As described earlier, 1517 proteins were conclusively identified by the SILAC analysis
and 49 showed significantly (Significance B score < 0.01) reduced abundance in the efp mutant
relative to wild-type (Figure 7). With the subsequent publications showing that EF-P is required
for the efficient translation of polyproline motifs, I began a comparison of the presence of these
motifs with EF-P-dependent expression in vivo. I generated a Venn diagram overlapping the 49
proteins that showed significantly reduced abundance in the efp mutant with the 422 Salmonella
proteins that contain at least one of the polyproline motifs PPP, PPG or APP as annotated in the
Salmonella Typhimurium str. SL1344 genome (Figure 8A) (GenBank ID: FQ312003.1). I found
that of the 49 proteins that showed reduced expression in the efp mutant in vivo, 20 of them
contained a polyproline motif – as might be expected based on the two Science papers114,115.
37
However, there remained 29 proteins that were conclusively identified in SILAC and showed
reduced expression in the efp mutant yet did not contain a characterized EF-P-dependent motif.
While it is possible that many of these are indirect effects, some of these proteins may contain a
novel EF-P-dependent motif, suggesting that polyproline motifs may not always be necessary to
render a transcript dependent on EF-P for its efficient translation.
My analysis also led to the discovery that polyproline motifs may not in-and-of
themselves be sufficient to confer EF-P-rescuable ribosome stalling in vivo. Specifically, 100
proteins containing a characterized polyproline motif (PPP, PPG or APP) were conclusively
identified in my SILAC data but only 20 of those showed significantly reduced abundance in the
efp mutant. Examples include ZipA, SseA and YtfM, which were identified to have average
wild-type/efp mutant expression ratios of 1.07, 0.74 and 0.84 respectively with Significance B
scores that were far from significant in all three replicates. Though they did not show
significantly altered abundance in the efp mutant, ZipA has two distinct APP motifs as well as an
APPP motif, SseA has an APPG motif, and YtfM contains a PPP motif. Even a more
conservative analysis of proteins with SILAC ratios of less than two without regard for the
significance score found that there were 45 proteins identified in SILAC that did not show
reduced abundance in the efp mutant yet have a PPP, PPG or APP motif (12, 22 and 16 proteins
respectively contain each motif and 5 proteins have two of the three motifs). This demonstrates
that a large percentage of proteins that contain a putative EF-P-dependent motif are not produced
in lower abundance in the efp deletion strain, suggesting that additional factors beyond short
polyproline motifs dictate whether a transcript will induce an EF-P-rescuable ribosome
stall114,115. I have made significant progress towards elucidating the factors governing whether a
polyproline motif will or will not induce an EF-P-rescuable ribosome stall and I will discuss this
work in detail in Chapter 4.
38
Figure 8: Comparison of proteins identified in SILAC and those with EF-P-dependent
polyproline motifs. A) Venn diagram outlining the overlap of the 1517 proteins conclusively
identified in SILAC (red), the 49 proteins with significantly reduced abundance in the efp mutant
(green), and the 422 Salmonella proteins containing a PPP, PPG or APP motif (blue). B) DAVID
analysis of the 49 SILAC hits that showed significantly reduced abundance in the efp mutant.
Functional annotation clusters showing significant overrepresentation (p-value < 0.05) are
shown. C) DAVID analysis showing the most significantly overrepresented clusters amongst the
422 proteins that contain a known EF-P-dependent polyproline motif. For clarity, only groups
with p < 0.001 are shown. D) DAVID analysis showing the only significantly overrepresented
cluster (p-value < 0.05) amongst the 20 proteins that belong to all three categories. Figure was
adapted from Hersch et al. (2013), mBio 110.
39
3.2.7 Functional annotation analysis of EF-P-dependent proteins
I subjected my SILAC data to a functional annotation analysis using the Database for
Annotation, Visualization and Integrated Discovery (DAVID) software package109,118,119. Upon
analysis of the 49 proteins with significantly lower abundance in the efp mutant in SILAC, I find
that four clusters demonstrate overrepresentation with a cluster p-value of less than 0.05 (Figure
8B). Most prominent amongst these groups are two-component regulatory systems, with
particular emphasis on proteins involved in chemotaxis and motility. Furthermore, metabolic
proteins also show reduced abundance in the efp mutant as annotated by functions in nucleotide
binding, oxidative phosphorylation, or proteolysis.
I subsequently examined the 422 ORFs in the S. Typhimurium SL1344 genome encoding
a PPP, PPG or APP motif (112, 195, and 185 ORFs, respectively, with 70 proteins containing
more than one motif). An obvious caveat of this analysis is that the presence of a polyproline
motif is not always sufficient to confer dependence on EF-P in vivo, and so many of these
proteins will not show altered levels in an efp mutant despite containing a polyproline motif. My
DAVID cluster analysis of these genes found that the overrepresented functional groups were
similar to those identified amongst proteins that were significantly depleted in the efp mutant in
SILAC including those involved in motility, two-component systems and nucleotide binding
proteins (Figure 8C).
I identified twenty proteins that both contain a polyproline motif and showed
significantly reduced abundance in the efp mutant in SILAC. Upon analysis using DAVID, seven
of these were identified as nucleotide binding proteins – the only overrepresented category in this
subset (Figure 8D).
40
3.3 Discussion
In this chapter I demonstrated that an efp mutant is more permeable than wild-type
Salmonella to the fluorescent dye NPN but not to ethidium bromide. Since NPN can fluoresce
upon bypassing the outer membrane whereas EtBr must also bypass the inner membrane, these
data suggested that efp mutants have a more permeable outer membrane that may underlie their
increased susceptibility to cellular stressors such as antibiotics74,84. In support of this I showed
that deletion of the kdgM gene partially complemented the NPN permeability defect. However,
Dr. Betty Zou demonstrated that deletion of kdgM in the efp mutant did not rescue any of the
other phenotypes, suggesting that the permeability defect (at least the portion of it conferred by
kdgM overexpression) was not the sole mechanism underlying the hyper-susceptibility of the efp
mutant to cellular stressors. Many of the proteins identified in SILAC that contain a
PPP/PPG/APP motif are predicted membrane proteins, suggesting a potential role in
permeability. Proteins involved in DNA repair were also overrepresented suggesting a potential
role for EF-P in translating proteins involved in resistance to DNA damaging agents.
Some proteins identified as EF-P-dependent by SILAC or by the presence of a
polyproline motif can provide a parsimonious explanation for previously described phenotypes
of Salmonella strains lacking PoxA or EF-P. For example, gamma-glutamyl transferase (Ggt) is
present at a level approximately 16-fold lower in the efp mutant compared to wild-type
Salmonella (Supplementary DataTable1). Ggt contains a PPP motif at residues 291-293 and its
strongly reduced synthesis in the efp mutant likely explains why Salmonella strains lacking EF-
P, PoxA, or YjeK are simultaneously unable to utilize -glutamyl-glycine as a nitrogen source
and are resistant to the compound GSNO (S-nitrosoglutathione)74,84. Finally, the impaired
synthesis of proteins involved in motility and chemotaxis, such as CheA, which is depleted
approximately four-fold in the efp mutant, likely contributes to the observed motility defect in
PYE mutant strains84.
Some of the EF-P-dependent proteins are involved in central metabolism and may
contribute to efp mutant phenotypes, including impaired growth: AtpD is the catalytic subunit of
the FOF1 ATPase and was previously identified as downregulated in the proteome of poxA
mutants74. PfkB is a phosphofructokinase that functions in glycolysis. Three of the other EF-P-
dependent nucleotide binding proteins (HflB/FtsH, HslU, and ClpB) play a role in protein
41
stability and turnover. The function of the polyproline motifs in these proteins is unlikely to be
universally conserved. For example, in AtpD120, ClpB121, and PfkB (PDB ID: 3UMP) the
putative EF-P dependent motifs are not proximal to the region of the protein that interacts with
ATP, whereas in HflB/FtsH the putative EF-P dependent motif (GPPG) makes contact with
AMP122.
Following the discovery that EF-P rescues ribosomes stalled at particular tripeptide
motifs, I probed the proteomic data for the presence of such sequences in proteins showing
reduced levels in the efp mutant. I discovered that these tripeptide polyproline motifs may not
always be necessary or solely sufficient to confer EF-P-dependence in vivo. This is reinforced by
the fact that the tripeptide motifs can be relatively common in the proteome and yet the overall
number of proteins affected strongly by the loss of EF-P is considerably smaller. Indeed, of the
100 polyproline motif-containing proteins identified in SILAC, only twenty showed significantly
reduced abundance in the efp mutant. Thus, EF-P-dependence in vivo may depend on additional
factors as well.
It is clear that not all ribosomal stalls are created equal and that many subtle factors will
influence when translation will pause. For example, many peptides known to stall during
elongation will stall despite the presence of EF-P. Multiple groups have reported on proteins that
instigate a translational stall via not only a particular sequence in the vicinity of the PTC but also
upstream residues in the nascent polypeptide chain that interact with the ribosomal exit tunnel.
These stalls may be dependent on external factors such as the antibiotic erythromycin in the case
of ermAL1 and tryptophan for TnaC, or may be self-mediated as for SecM and MifM61-64,123,124.
The laboratory of Allen Buskirk identified a number of stall sequences containing polyproline
tracts near the PTC that also require upstream residues for efficient stalling70,125. In Chapter 4 I
will present evidence elucidating additional factors influencing EF-P-dependence including a
role for the amino acids upstream of a proline-rich motif and a significant impact of the rate of
translation initiation.
42
4 Factors that influence EF-P-dependence
Primary references
Much of the data presented in this chapter is published work reproduced with permission from:
Hersch SJ, Wang M, Zou SB, Moon K-M, Foster LJ, Ibba M, Navarre WW. Divergent
protein motifs direct elongation factor P-mediated translational regulation in Salmonella enterica
and Escherichia coli. mBio. 2013;4(2). 110
Hersch SJ, Elgamal S, Katz A, Ibba M, Navarre WW. Translation initiation rate
determines the impact of ribosome stalling on bacterial protein synthesis. J Biol Chem.
2014;289(41):28160-28171. 126
Elgamal S, Katz A, Hersch SJ, Newsom D, White P, Navarre WW, Ibba M. EF-P
dependent pauses integrate proximal and distal signals during translation. PLoS Genet.
2014;10(8):e1004553. 127
Acknowledgements
Dr. Michael Ibba at the Ohio State University and members of his lab have been close
collaborators on this project. Ribosome profiling and computational modelling were conducted
by Sara Elgamal and Dr. Assaf Katz and some of their findings are mentioned in this chapter.
Further details about their data and methods can be found in Elgamal, Katz, Hersch et al. (2014)
PLoS Genet127 (for ribosome profiling) and in Hersch et al. (2014) J Biol Chem126 (for
computation modelling of translation).
The pXG10sf plasmid used for generating translational fusions was a generous gift from
Dr. Jörg Vogel93.
43
4.1 Overview
Recent work demonstrated that three amino acid proline-containing motifs require EF-P
for their efficient translation in vitro114,115. Comparing my SILAC proteomic data to the presence
of polyproline motifs revealed that a three amino acid polyproline motif is not always necessary
or sufficient to induce EF-P-dependent protein levels in cells lacking EF-P. In this chapter I
identify a novel six amino acid EF-P-dependent motif that contains only a single proline in
pyruvate oxidase (PoxB), which was the first protein in the literature to demonstrate PYE
pathway-dependent expression75,76. I also employ the model proteins AtpA and AtpD, which
each contain a PPG motif and are both encoded on the same mRNA transcript, yet AtpD
consistently exhibits lower levels in efp mutants whereas AtpA abundance is unaffected by the
loss of EF-P128,129. Using these model proteins I demonstrate that residues up to three positions
upstream of the PPG motif can influence the degree of stalling in efp mutants. Furthermore, I
show that the component most critical for the difference in EF-P dependence is within the 5’
untranslated region (UTR). Further mutagenesis and modeling studies demonstrate a relationship
between the translation initiation rate of a transcript and the impact of an EF-P-rescuable stall on
protein production. Taken together these data indicate that not all polyproline motifs induce
stalling to equal degrees and that EF-P dependence is only observed when a given polyproline
sequence impacts the rate of protein synthesis more than translation initiation.
4.2 Results
4.2.1 PoxB requires EF-P activity for its efficient translation
Levels of pyruvate oxidase (PoxB) are reduced five- to eight-fold in E. coli poxA
mutants, implicating it as a protein dependent on EF-P for its synthesis75,76. The PoxB protein,
however, lacks any of the proline-rich sequences (e.g. PPG) implicated to be EF-P-dependent
motifs, which leaves open the question as to whether the PoxB protein is indeed a target of EF-P
and, if so, what particular sequence renders it dependent on EF-P for its synthesis. I was unable
to detect PoxB in the SILAC assay presented in Chapter 3, likely due to the particular growth
conditions employed. To circumvent this issue, I constructed a reporter vector wherein the PoxB
5’ UTR and open reading frame was fused to a “super-folder” variant of green fluorescent
44
protein (sfGFP)99. A constitutive PLtet0-1 promoter drives transciption of the chimeric gene,
enabling me to measure translation independently of transcription. In a Salmonella background
the full-length PoxB protein fused to sfGFP show a marked and reproducible dependence on EF-
P for synthesis (Figure 9). I observed this decrease in fluorescence not only in the efp mutant but
also in poxA and yjeK mutants. In contrast, I found no significant difference in synthesis of the
PoxB-sfGFP fusion construct in a strain lacking the YfcM protein. These results are consistent
with previous data showing that YfcM-mediated hydroxylation of EF-P is not critical for its
function87.
To ensure that this difference was due to an effect on translation, I measured mRNA
levels using reverse transcriptase quantitative PCR. Results indicated that transcription of the
chimeric PoxB-sfGFP fusion construct was similar between the mutant and wild-type strains
(Figure 10). Interestingly, though similar poxB transcript levels indicate that transcription is not
altered, gfp mRNA levels were decreased in strains expressing EF-P-dependent constructs. Since
previous in vitro data measured the effect of EF-P in the absence of nucleases114,115,125, this
degradation is likely not the cause of decreased GFP fluorescence. Instead it is likely that
ribosome stalling at EF-P-dependent motifs leads to reduced ribosome-mediated nuclease
protection of the 3’ end of the poxB-gfp transcript. This supports the concept that the altered
expression of poxB-sfGFP in PYE mutants reflects PoxB requiring EF-P during translation rather
than an artefact of mRNA stability.
45
Figure 9: PoxB-sfGFP translation is impaired in PYE mutants but not in the yfcM mutant.
Fluorescence of sfGFP fusion constructs in mutant strains lacking an EF-P-modifying protein.
Mutant strains are indicated in the top row of the x-axis, the sfGFP fusion constructs are
indicated in the bottom row of the x-axis. LacZ indicates the first 186 amino acids of LacZ,
PoxB FL indicates full-length PoxB (572 codons), and PoxB 21 indicates the first 21 residues of
PoxB are fused to sfGFP. Values were taken at 10 hours post-inoculation and are shown in
arbitrary fluorescence units (AU) normalized to optical density at 600nm (OD600). The ‘no
plasmid’ controls do not contain a sfGFP fusion plasmid and were grown in the absence of
chloramphenicol. All values are the average of at least 3 biological replicates. Error bars show
one standard deviation.
0
1
2
3
4
5
6
7
8
9
WT Δefp WT Δefp ΔyfcM ΔpoxA ΔyjeK WT Δefp ΔyfcM ΔpoxA ΔyjeK WT Δefp ΔyfcM ΔpoxA ΔyjeK
no plasmid LacZ PoxB FL PoxB 21
GFP
Flu
ore
sce
nce
(1
00
00
AU
)
Construct in pXG10sf and Salmonella strain
46
Figure 10: PoxB mRNA levels are similar in wild-type and ∆efp strains for both EF-P
dependent and independent constructs. GFP fluorescence normalized to optical density (black),
poxB mRNA transcript levels (light grey) and sfGFP mRNA transcript levels (dark grey) were
assayed at mid-log phase. Transcript levels were assessed by RT-qPCR using gene-specific
primers. Since the strains used encode poxB in their genomes, the mRNA levels obtained for
lacZ186 with the poxB-specific primers was subtracted as background from all other values. This
background poxB mRNA level was negligible (greater than 300 fold lower) compared to any of
the strains expressing PoxB-GFP from a plasmid. All values are shown as the ratio of the efp
mutant relative to the wild-type strain expressing the same construct. All values are the average
of at least 3 biological replicates with the exception of those for the LacZ186 construct, which
are the average of 2 biological replicates. For the replicate where LacZ186 was not included,
background poxB mRNA levels were obtained from another strain expressing a non-poxB
construct from the same plasmid. Error bars show one standard deviation. Figure is shown as it
appears in Hersch et al. (2013), mBio 110.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
LacZ186 PoxB FL PoxB 68 PoxB 82 P76L
Δef
p/
WT
rati
o
GFP fluorescence
poxB mRNA
sfGFP mRNA
47
4.2.2 PoxB is dependent on EF-P due to a novel six amino acid motif, GSCGPG
To identify the region of PoxB that rendered it dependent on EF-P I systematically
generated C-terminal truncations of the poxB sequence fused to sfGFP such that each construct
encoded the poxB 5’ UTR followed by varying lengths of the poxB ORF (Figure 11A). By
comparing fluorescence of these constructs in wild-type and efp mutant Salmonella, I was able to
narrow down the critical region to a 14 amino acid stretch between residues 69 and 82 with the
sequence: VCAGSCGPGNLHLI (Figure 11B and 8C). Furthermore I confirmed that it is the
amino acid (as opposed to the nucleotide) sequence in this region that is critical for EF-P
dependence by generating a construct with this region shifted to the +1 reading frame and
observing an abolition of the requirement for EF-P (Figure 12). Mutating the proline codon to
any of the other three proline codons also had no effect (Figure 12).
Residues 69 to 82 of PoxB contain a GPG motif that I hypothesized could, like PPG,
mediate the EF-P dependence of the PoxB protein. To test this, I performed site directed
mutagenesis on the GPG motif in the full-length PoxB-sfGFP construct as well as on additional
upstream and downstream residues, converting each residue to a leucine. I chose leucine because
of its absence from the top scoring motifs predicted by Mengchi Wang’s tripeptide motif analysis
(See Chapter 3). I found that mutation of any residue in the GPG motif or the three amino acids
upstream (sequence: GSC) restored GFP fluorescence in the efp mutant (Figure 11D). In
contrast, mutating residues further downstream or upstream of the GSCGPG motif had negligible
effect on the synthesis of the reporter construct. This novel EF-P dependent motif of PoxB is
considerably larger than the previously described tripeptide motifs and can cause a protein to be
dependent on EF-P for proper synthesis in vivo. Notably, in Salmonella the GSCGPG sequence
is only found in PoxB.
48
Figure 11: The GSCGPG motif of PoxB renders it dependent on EF-P. A) Outline of
translational fusion constructs expressing full-length (FL) PoxB or C-terminal truncations fused
to sfGFP. Construct designations are shown in bold at left as a figure key and indicate the length
in codons of the truncated poxB gene. Arrow indicates the transcriptional start site under the
control of the PLtet0-1 constitutively active promoter and fMet indicates the ATG start codon of
PoxB. Numbers indicate the number of codons from the poxB start codon to the C-terminal
truncation and fusion to sfGFP. For clarity only a selection of constructs are illustrated. Amino
acids 69-82 are highlighted. B) Sequence of amino acids 69-82 of PoxB. The GSCGPG motif is
emboldened and underlined. C) Relative GFP fluorescence of PoxB truncations. Values were
taken at 10 hours post-inoculation, were normalized to optical density at 600nm (OD600) and are
shown as a ratio of the efp mutant relative to the wild-type strain expressing the same construct.
The same plasmid encoding the first 186 codons of lacZ instead of poxB is included as a control.
D) As in (C) but showing single residue mutations to leucine in the full-length PoxB construct.
All values are the average of at least 3 biological replicates. Error bars show one standard
deviation. Figure was adapted from Hersch et al. (2013), mBio 110.
C)
0.00.20.40.60.81.01.21.41.61.8
GFP
Flu
ore
sce
nce
(∆
efp
/ W
T)
Construct in pXG10sf
D)
0.00.20.40.60.81.01.21.41.61.8
GFP
Flu
ore
sce
nce
(∆
efp
/ W
T)
Construct in pXG10sf
A)
B)
49
Figure 12: PoxB dependence on EF-P is determined by the amino acid sequence as opposed
to nucleotides. A) Nucleotide (upper line) and amino acid (lower line) sequences of PoxB FL and
PoxB Shifted constructs are shown from residues 66-85. The PoxB Shifted construct was
generated by shifting codons 69-82 (emboldened) of full-length PoxB to the +1 reading frame by
introducing an upstream single base-pair insertion (underlined adenine) and downstream deletion
(struck-out cytosine). B) Relative GFP fluorescence values taken at 10 hours post-inoculation,
normalized to optical density at 600nm (OD600) and shown as a ratio comparing expression in the
efp mutant relative to wild-type Salmonella expressing the same construct. P76::CCX constructs
are mutated at proline 76 to an alternate proline codon. LacZ186, PoxB FL, PoxB 82 and PoxB
68 from Figure 11 are shown for comparison. All values are the average of at least 3 biological
replicates. Error bars show one standard deviation. Figure was adapted from Hersch et al.
(2013), mBio 110.
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1.400
1.600
1.800
2.000
GFP
Flu
ore
sce
nce
(∆
efp
/WT)
Construct in pXG10sf
PoxB FL… GAG CTG GCA GTG TGC GCC GGT TCA TGT GGA CCG GGC AAC CTG CAC CTG ATC AAT GGC CTG …… E L A V C A G S C G P G N L H L I N G L …
PoxBShifted
… GAG CTG GCA AGT GTG CGC CGG TTC ATG TGG ACC GGG CAA CCT GCA CCT GAT AAT GGC CTG …… E L A S V R R F M W T G Q P A P D N G L …
Positionɪ
70ɪ
75ɪ
80ɪ
85
A)
B)
50
4.2.3 AtpD and AtpA as model proteins to study polyproline motifs
Only 20 of the 100 proteins identified in SILAC that contain a putative EF-P-dependent
polyproline motif actually showed a decrease in protein abundance in the efp mutant, suggesting
that additional factors influence whether or not a polyproline motif will require EF-P for efficient
translation. Literature examining other kinds of ribosome stalls (that stall despite the presence of
EF-P) suggest that many stall sites involve upstream amino acids interacting with the ribosome
exit tunnel to influence stalling61,70,124,125. Combined with my PoxB data where I found that up to
six amino acids rather than just three were required for EF-P-dependence, these data led to the
hypothesis that upstream amino acids influence polyproline motifs to either induce or protect
against ribosome stalling. To examine this possibility I employed two model proteins, AtpA and
AtpD. Both are subunits of the FOF1 ATP Synthase and are encoded on the same mRNA
transcript (Figure 13). In addition, both contain a PPG motif; however, my proteomic SILAC
analysis showed that only AtpD levels were affected by the deletion of efp (20.6 fold reduced
abundance in the efp mutant, contrasting with 1.05 for AtpA). A similar trend was observed in a
later SILAC experiment in E. coli wherein the authors found a 5.18 fold difference in the
abundance of AtpD in the efp mutant compared to 1.88 for AtpA130.
The identical polyproline motif, yet divergent degree of EF-P-dependence in vivo,
suggested AtpD and AtpA as prime candidates to assess the role of upstream residues on
polyproline-mediated ribosome stalling. I first verified the SILAC results using western blotting
that confirmed that the level of AtpD protein is lower in the Salmonella efp mutant (Figure 14A
and quantified in 11B). I then employed the previously used pXG10sf plasmid to generate C-
terminal translational fusions of the atpD and atpA genes to sfGFP93,99,110. Consistent with the
proteomic analysis, the fluorescence of the AtpD construct was dependent on EF-P whereas that
of AtpA was not (Figure 14C) and this was true throughout the growth curve (Figure 15).
Notably, when I mutated the PPG motif of AtpD to PLG (P214L), the requirement for EF-P was
abolished. Unsurprisingly, mutation of the AtpA PPG motif to PLG (P281L) did not have a
significant effect, as AtpA translation was already independent of EF-P. Conversely, when I
extended the PPG motif of AtpA to PPPG (R279P), I found that I could induce strong EF-P-
dependence in AtpA.
51
A)
B)
Protein
SILAC ratio
(WT/Δefp) SigB < 0.01
Polyproline
motif
AtpF 04.43 - -
AtpH 02.32 - -
AtpA 01.05 - PPG
AtpG 07.47 < 0.01 -
AtpD 20.56 < 0.01 PPG
AtpC 03.55 - -
Figure 13: ATP Synthase components show altered abundance in the efp mutant. A) Outline of the ATP Synthase operon encoding the eight subunits of the F1F0 ATP Synthase. The
black arrow indicates the transcription start site. Blue arrows indicate ORFs, yellow arrows
indicate ORFs containing a PPG motif. B) Six out of the eight components of FOF1 ATP
Synthase were identified in my SILAC anlysis of efp mutant Salmonella. Shown are their
WT/Δefp abundance ratios, whether or not their Significance B score was under 0.01 (indicating
significantly altered protein level in the efp mutant) and also if they contain a polyproline motif.
I B E HF atpA G CatpD
ATP Synthase operon
52
Figure 14: AtpD but not AtpA requires EF-P for its synthesis. A) Representative western blot
showing AtpD levels in the indicated strains of Salmonella. DnaK was included as a loading
control. B) Densitometry quantification of western blots showing mean AtpD/DnaK ratio relative
to wild-type (WT) across three biological replicates. Error bars indicate one standard deviation.
C) Fluorescence comparing synthesis of plasmid-encoded AtpD- and AtpA-sfGFP translational
fusions. Fluorescence values were taken at 10 hours post-inoculation, normalized to optical
density at 600nm (OD600) and are shown as a ratio comparing expression in the efp mutant
relative to wild-type Salmonella expressing the same construct. LacZ is included as a control.
Specific point mutations affecting the PPG motifs are indicated: P214L and P281L mutate the
PPG motifs of AtpD and AtpA respectively to PLG. R279P in AtpA generates a PPPG motif
conferring stronger ribosome stalling. Values are the average of at least 3 biological replicates
and error bars show one standard deviation. Figure was adapted from Elgamal S, Katz A, Hersch
et al. (2014), PLoS Genet 127.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
LacZ AtpD AtpDP214L
AtpA AtpAP281L
AtpAR279P
GFP
Flu
ore
sce
nce
(∆
efp
/WT)
Construct in pXG10sf
0
0.2
0.4
0.6
0.8
1
1.2
1.4
WT Δefp Δefp pEF-P
ΔatpD
Atp
D L
evel
s (r
elat
ive
to W
T)
A) B)
C)
53
Figure 15: AtpD but not AtpA shows decreased fluorescence throughout the growth curve
of efp mutant Salmonella. 16 hour growth curves of wild-type (solid lines) and efp mutant
(dashed lines) Salmonella containing pXG10sf plasmids with AtpD (A & B) or AtpA (C & D)
inserted as a translational fusion to sfGFP. GFP fluorescence (black lines in A & C) and optical
density at 600nm (grey lines in A & C) were measured every 15 minutes. The ratio of GFP
fluorescence / OD600 is also shown (B & D). At least three independent replicates were
conducted and one representative replicate is shown. AFU, arbitrary fluorescence units. Figure is
shown as it appears in Hersch et al. (2014), J Biol Chem 126.
54
4.2.4 Upstream residues influence EF-P-dependence in AtpD and AtpA
I examined the regions upstream of the AtpD and AtpA PPG motifs to delineate if they
contribute to stalls requiring rescue by EF-P. I swapped up to 40 codons upstream of the AtpA
PPG motif into the corresponding region of the pXG10sf-AtpD-sfGFP construct while leaving
the PPG motif and the remainder of the ORF unaltered. I also performed the reverse swaps
wherein AtpD regions were swapped into the pXG10fs-AtpA-sfGFP construct. I found that
swapping as few as two upstream codons from AtpA into the AtpD-sfGFP construct led to a
significant increase in expression in the efp mutant relative to wild-type Salmonella (Figure 16).
The effect was similar when I switched six or more residues but interestingly increased when
four upstream residues were swapped. I observed a similar reversal of EF-P-dependence when
swapping upstream regions of AtpD into the AtpA-sfGFP construct: a four amino acid swap led
to a very small increase in EF-P-dependence, which increased only marginally when the
swapped region was increased to six or more codons. My observation that swapping greater than
six residues upstream had marginal or no additional effect suggests that at least in this instance
the important interactions with the ribosome exit tunnel are occurring close to the peptidyl
transferase center and prior to the exit tunnel constriction that has been implicated in other
extended translational stall motifs61,70,71.
Interestingly, I observed no effect for either AtpA or AtpD when I swapped the amino
acid immediately upstream of the PPG motif (E212R for AtpD; R279E for AtpA). However,
swapping the residue two positions upstream of the PPG resulted in a drastic effect on EF-P-
dependence (N211R for AtpD; R278N for AtpA). The AtpD N211R construct partially
alleviated AtpD’s dependence on EF-P, though to a lesser degree than when both N211 and E212
were swapped to the corresponding arginines found in AtpA. In contrast, changing the arginine
positioned two amino acids upstream of the PPG motif of AtpA to the asparagine in that position
in AtpD (R279N) led to a dramatic increase in the dependence of this construct on EF-P,
surpassing the effect of all other AtpA constructs where more residues were altered. Thus it
appears that at least in the case of AtpA and AtpD the residue two positions upstream of the PPG
motif plays an important role in determining whether the ribosome will stall and require rescue
by EF-P. Furthermore, the data that I have presented here also demonstrates that other nearby
residues can dampen this effect.
55
Figure 16: Upstream residues influence EF-P-dependence in AtpD and AtpA. A) Sequences
of Salmonella AtpD and AtpA in proximity to their PPG motifs (bold). The relative position
when the glycine of PPG occupies the A site of the ribosome is shown above. The amino acid
position of the second proline of the PPG motif in each ORF is indicated below. B) Fluorescence
ratios comparing expression of plasmid borne AtpD-sfGFP translational fusions in efp mutant
relative to wild-type Salmonella. ‘Swap’ constructs indicate that the sequence from AtpA has
been swapped into AtpD-sfGFP for the specified number of amino acids upstream of the PPG
motif. LacZ, the unaltered AtpD-sfGFP construct (wt), P214L and R279P constructs are included
for comparison. Data shows the Δefp / WT GFP fluorescence ratio at 10 hours post-inoculation
normalized to optical density (600nm). Values are the average of at least 3 biological replicates
and error bars show one standard deviation. C) As in (B) but with the exception that sequences
from AtpD have been swapped into the AtpA-sfGFP construct. Figure was adapted from
Elgamal S, Katz A, Hersch et al. (2014), PLoS Genet 127.
V A Y R Q I S L L L R R P P G R E A
AtpD:
AtpA:
D K V S L V Y G Q M N E P P G N R L
- 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6 - 5 - 4 - 3 - 2 - 1 E P A + 1 + 2 + 3
214 (AtpD)281 (AtpA)
A
B
0
0.2
0.4
0.6
0.8
1
1.2
1.4
wt P214L 40aaswap
24aaswap
12aaswap
6aaswap
4aaswap
2aaswap
E212R N211R M210L
LacZ AtpD
GFP
Flu
ore
scen
ce (
∆ef
p/W
T)
LacZ
AtpD (with regions of AtpA swapped in)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
wt R279P 40aaswap
24aaswap
12aaswap
6aaswap
4aaswap
2aaswap
R279E R278N L277M
LacZ AtpA
GFP
Flu
ore
scen
ce (
∆ef
p/W
T)C
LacZ
AtpA (with regions of AtpD swapped in)
56
4.2.5 The 5’ UTR plays a significant role in EF-P dependence
The reversal of EF-P dependence upon swapping upstream amino acids was incomplete
when compared to a negative control (LacZ), to a PPG::PLG mutation in AtpD, or to an
RPPG::PPPG mutation in AtpA (Figure 16). Complete reversal was not achieved even when
swapping up to 40 codons upstream of the PPG motif, a distance that accounts for more amino
acids than can be accommodated in the ribosomal exit tunnel69,131. This suggested that factors
outside of the ribosome were negatively affecting AtpD translation in the efp mutant.
To investigate additional regions impacting EF-P dependence, I again employed the
pXG10sf plasmid, involving a constitutively active promoter and C-terminally fused sfGFP as a
reporter93,99,110,132. I systematically generated atpD and atpA hybrids by serially swapping regions
of increasing length from the 5’ end of the mRNA transcript to the PPG motif, or in the reverse
direction (Figure 17). Results from this analysis demonstrated that the coding region beyond 12
codons upstream of the PPG motif had no additional influence on EF-P dependence even when
the entire N-terminus from the ATG start codon to the polyproline motif was swapped.
Contrastingly, there was a stark change in EF-P dependence when I included the 5’ UTR in the
swapped region or when only the 5’ UTR was swapped. Specifically, the atpD 5’ UTR confers
strong EF-P dependence onto the AtpA protein whereas the 5’ UTR of atpA appears to
significantly reduce the amount of EF-P dependence displayed by AtpD. Combined replacement
of the 5’ UTR and the 12 codons upstream of the PPG motif have a cumulative effect, suggesting
that the two regions affect EF-P dependence by different mechanisms and that complete
switching requires both elements. Indeed, the atpD construct with both the atpA 5’ UTR and
upstream PPG flanking region demonstrates higher fluorescence in the efp mutant than in wild-
type Salmonella; this resembles the lacZ negative control and may be due to decreased levels of
protease in the efp mutant as found in my SILAC analysis in Chapter 3.
57
Figure 17: The 5’ UTR plays a role in EF-P dependence in addition to residues immediately
upstream of the PPG motif. A) Outline of pXG10sf-AtpD and -AtpA constructs. Swap constructs
are shown with regions of the atpA gene swapped into the pXG10sf-AtpD construct. Naming
scheme is indicated at left. B) Regions of the atpA gene were swapped into the pXG10sf-AtpD
construct (as shown in panel A). Data shows the ratio of GFP fluorescence in ∆efp/WT
Salmonella normalized to OD600. Unaltered (wt), a PPG::PLG mutation (P214L), the ‘12aa’
construct, and pXG10sf-LacZ are shown for comparison. C) As in panel B but showing
constructs with regions of the atpD gene swapped into pXG10sf-AtpA (inverse of depiction in
panel A). Values are the mean of at least three biological replicates and error bars show one
standard deviation. Figure is shown as it appears in Hersch et al. (2014), J Biol Chem 126.
58
4.2.6 The 5’ UTR regions that include the SD sequence and a potential translation enhancer
play a significant role in EF-P dependence
Though native ATP synthase proteins are encoded on a single mRNA transcript, both
atpD and atpA possess their own Shine Dalgarno (SD) sequence that is crucial for ribosome
recruitment during translation initiation. Furthermore, their start codons are spaced 12 and 26
bases (respectively) after the preceding gene’s stop codon, suggesting that the translation of each
is independent and not coupled to that of the preceding gene (NCBI accession: NC_016856.1).
For the plasmid-based translational fusion assay that I employed, the 5’ UTRs of the atpD and
atpA constructs consisted of the 74 or 75 bases (respectively) upstream of the start codon.
To further dissect the region within the 5’ UTR involved in EF-P dependence, I
systematically swapped 15 base-pair stretches between atpD and atpA and measured expression
of the resulting construct by GFP fluorescence. I found that EF-P dependence was reversed most
significantly when regions -15 to -1 or -30 to -16 (relative to the start codon) were swapped
(Figure 18). Notably, these regions are both predicted to be heavily involved in translation
initiation: The -15 to -1 region includes the SD sequence, and the -30 to -16 region of atpD is
particularly AT-rich (74%) compared to atpA (47%), suggesting that it may act as a translation
enhancer47,49. Moreover, for all of the serial swap constructs generated, there appeared to be a
correlation between EF-P dependence and the degree of overall expression (Figure 18).
Specifically, constructs that were poorly expressed exhibited little difference in fluorescence
levels between wild-type and efp mutant strains, whereas constructs with high levels of
expression in the wild-type background were the most reduced in the efp mutant.
To ensure that this apparent loss of EF-P dependence was not due to an inability to
accurately measure differences at low fluorescence levels, I conducted western blotting of select
AtpD and AtpA constructs. The patterns of EF-P dependence observed by western blotting were
similar to those found by fluorescence measurements (Figure 19). For a few constructs where
fluorescence levels are less than two-fold greater than no-plasmid controls, protein levels were
very low or non-detectable by western blot in both the wild-type and efp mutant strains. These
constructs, including the atpD construct containing the whole 5’ UTR of atpA, were omitted
from all downstream analyses.
59
Figure 18: The 5’ UTR regions affecting EF-P dependence include the ribosome binding
site and correlate with expression levels. A) Sequence of the 5’ UTRs included in pXG10sf-
AtpD and -AtpA constructs. Putative Shine-Dalgarno sequences are underlined. Base position
relative to the translation start site (ATG) are indicated below. B) Regions of the atpA 5’ UTR
were serially swapped into the corresponding position of pXG10sf-AtpD. Example: ‘-15-1’
indicates the 15bp from position -15 to -1 (relative to the ATG start codon) were swapped. Data
shows GFP fluorescence in arbitrary fluorescence units (AFU) normalized to OD600 in WT (dark
grey) and ∆efp (light grey) Salmonella. Numbers above the columns indicate the ∆efp/WT ratio.
C) As in panel B but with regions of the atpD 5’ UTR swapped into the pXG10sf-AtpA
construct. Values are the mean of at least three biological replicates and error bars show one
standard deviation. Figure was adapted from Hersch et al. (2014), J Biol Chem 126.
60
Figure 19: GFP levels determined by immunoblotting follow similar trends as levels
determined by fluorescence. Western blot probing for GFP expressed from pXG10sf constructs
in wild-type (WT) and ∆efp Salmonella grown to mid log phase under conditions similar to those
used for fluorescence measurements. GFP – control is wild-type cells expressing pXG10sf-AtpD
but with a FLAG tag replacing GFP. Unless otherwise indicated, 10µg protein lysate was loaded
for each sample. DnaK was included as a loading control. The experiment was conducted in
triplicate and one representative replicate is shown. For the ‘AtpD-AtpA UTR’ and ‘AtpD
AGAGG::AGACG’ constructs where fluorescence levels were less than two-fold greater than
no-plasmid controls, protein levels were very low or non-detectable in both the WT and efp
mutant strains. The ‘AtpA – AtpD UTR from -60-46’, ‘AtpA – AtpD from ATG – PPG’, ‘AtpD
– AtpA UTR & 12aa’, and ‘AtpD – AtpA UTR & P214L’ constructs also had similarly low
fluorescence levels (Figures 14 and 15). These constructs were omitted from all downstream
analyses. Figure is shown as it appears in Hersch et al. (2014), J Biol Chem 126.
61
4.2.7 Mutations that affect translation initiation influence EF-P dependence
To further assess whether the 5’ UTR influences EF-P dependence through its role in
translation initiation, I first confirmed that the UTR instigates its effect via the polyproline stall
rather than by a separate unrelated mechanism. Indeed, mutation of the PPG motif to PLG
resulted in restoration of GFP fluorescence in the efp mutant even at high expression levels
obtained by 5’ UTR swapping (Figure 20). Of note, wild-type cells express the PPG and PLG
constructs to a similar degree, indicating that the motif does not influence protein synthesis aside
from instigating an EF-P-rescuable stall. Thus the stall motif only acts on elongation, yet its
impact on protein output is strongly influenced by the 5’ UTR, which is far upstream of the stall,
is not translated, and would be expected to influence initiation but not elongation.
To explicitly examine the role of translation initiation in this phenomenon, I mutated
individual nucleotides within the SD sequence and the start codon of atpD and atpA. I improved
the atpA SD sequence (AGGGGA) by mutating it to the consensus sequence (AGGAGG). For
atpD, I weakened ribosome binding by mutating the ATG start codon or by altering the wild-
type SD sequence (AGAGG) with a G to C mutation (AGAGC). Furthermore, I also
strengthened the atpD SD sequence by a G insertion yielding the consensus sequence
(AGGAGG). Consistent with my previous observations, the constructs displaying the highest
levels of expression in wild-type cells were the most affected by the loss of EF-P (Figure 21).
Specifically, strengthening the SD sequence of either atpD or atpA yielded increased dependence
on EF-P, while weakening the SD or the start codon of atpD reduced its dependence. Since these
point mutations target regions critical for ribosome binding, these data suggest that the 5’ UTR
influences EF-P dependence via its role in translation initiation. This finding suggests that the
transcript’s initiation rate (governed by the 5’ UTR) and the EF-P-rescuable stall in elongation
may compete to be the rate-limiting step of translation.
62
Figure 20: The effect of the 5’ UTR on EF-P dependence requires an intact stall motif. A) The -45-31 region of the atpA 5’ UTR was swapped into the pXG10sf-AtpD construct in
combination with mutating the PPG motif to PLG. Data shows GFP fluorescence in arbitrary
fluorescence units (AFU) normalized to OD600 comparing expression in WT (dark grey) to ∆efp
(light grey) Salmonella. Numbers above the columns indicate the ∆efp/WT ratio. B) As in panel
A but with the 5’ UTR of atpD (-74-1) swapped into the pXG10sf-AtpA construct. Values are
the mean of at least three biological replicates and error bars show one standard deviation. Figure
is shown as it appears in Hersch et al. (2014), J Biol Chem 126.
63
Figure 21: Single base mutations in the SD sequence or start codon alter expression and
EF-P dependence. A) Mutations in the SD sequence or in the ATG start codon of pXG10sf-
AtpD. The wild-type sequence (AGAGG) is underlined in Figure 18A. Data shows GFP
fluorescence in arbitrary fluorescence units (AFU) normalized to OD600 comparing expression in
WT (dark grey) to ∆efp (light grey) Salmonella. Numbers above the columns indicate the
∆efp/WT ratio. Unaltered (wt) construct is shown for comparison. B) As in panel A but with
mutation in the SD sequence of the pXG10sf-AtpA construct. The wild-type sequence
(AGGGGA) is underlined in Figure 18A. Values are the mean of at least three biological
replicates and error bars show one standard deviation. Figure is shown as it appears in Hersch et
al. (2014), J Biol Chem 126.
64
4.2.8 Initiation rate and stall strength correlate with EF-P dependence
To gain insight into the relation between initiation rate and EF-P dependence, I plotted
fluorescence in wild-type versus in efp mutant Salmonella for the full set of 5’ UTR mutant
constructs (Figure 22). I grouped constructs together such that all members of each group have
the identical ORF and only differ in their 5’ UTR. Plotting in this manner allowed for
visualization of changes in expression in the efp mutant (subjected to stalls in elongation) as
initiation rate (and thereby wild-type expression) increases.
Strikingly, for both atpD and atpA, I find that as expression increases in wild-type,
expression in the efp mutant follows a distinct curve that eventually reaches a maximum. Indeed,
the observed data could be fit well to a ‘1-exp’ curve of the form: 𝑦 = 𝐴(1 − 𝑒−𝐵𝑥) , where x
and y represent fluorescence in the wild-type and efp mutant strains respectively, and A and B are
constants particular to the ORF being analyzed. Specifically, A is the asymptote of the curve and
can be interpreted as the theoretical maximum expression in the efp mutant for that particular
ORF, at which point progression through the stall motif has overtaken initiation as the rate-
limiting step for all transcripts. This supports my hypothesis that the initiation rate and
elongation stall clearance compete to be the rate-limiting step of translation. Similarly, B is
inversely related to the slope of the curve and how quickly it reaches the asymptote.
Interestingly, the curve for the AtpD ORF was steeper than that for AtpA and reached a lower
asymptote. This emphasizes that the stall in AtpD is more EF-P dependent than that of AtpA,
resulting in a lower maximum expression in the efp mutant and a significant hindrance on
translation even at lower initiation rates.
To examine the additive effects of the upstream region and translation initiation rate, I
applied my plotting analysis to constructs where the 12 amino acids upstream of the atpD or
atpA PPG motifs were swapped. Interestingly, the data for the 12aa swap constructs resembles
the protein from which the upstream region originated, suggesting that the upstream residues
govern stall strength and thereby the arc and maximum of the expression curve (Figure 22 grey
icons). Of note, this relation between translation initiation and EF-P dependence is not restricted
to AtpD and AtpA in Salmonella, but is consistent with findings on other proteins and in a
different species (E. coli), conducted by the lab of my collaborator, Dr. Mike Ibba126.
65
Figure 22: Translation initiation and elongation stall strength influence protein level. GFP
fluorescence data plotted to compare expression in WT (x-axis) and ∆efp (y-axis) Salmonella.
The y-axis is expanded from 0 to 1 to clarify differences at low fluorescence levels. Each point
represents fluorescence data for one pXG10sf construct. The specific construct for each data
point is indicated by numerical label referring to Table 1. Icon groups signify constructs that all
have the same ORF (indicated in key) and only differ from one another in the 5’ UTR. For the
AtpD and AtpA ORF groups, the data points were fit to a ‘1-exp’ curve shown in the inset at top
left. P214L and P281L groups were connected linearly and the dashed line indicates equal
fluorescence in the WT and ∆efp mutant. Data is in arbitrary fluorescence units (AFU)
normalized to OD600 and is the mean of at least three biological replicates. Error bars showing
one standard deviation are included for both x- and y-axes. ‘AtpD – AtpA 12aa’ indicates 12
codons upstream of the atpA PPG motif were swapped into the AtpD construct, ‘AtpA – AtpD
12aa’ is the reciprocal. Figure is shown as it appears in Hersch et al. (2014), J Biol Chem 126.
66
Table 1: Constructs used in Figure 22
ORF Mutation or Swap in 5'
UTR Label
LacZ wt 1
AtpD wt 2
AGAGG::AGAGC 3
AtpA UTR from -75-61 4
AtpA UTR from -60-46 5
AtpA UTR from -45-31 6
AtpA UTR from -30-16 7
AtpA UTR from -15-1 8
AGAGG::AGGAGG 9
ATG::GTG 10
ATG::TTG 11
AtpA UTR from -45-31
& AGAGG::AGGAGG 12
AtpA wt 13
AtpD UTR 14
AGGGGA::AGGAGG 15
AtpD UTR from -74-61 16
AtpD UTR from -45-31 17
AtpD UTR from -30-16 18
AtpD UTR from -15-1 19
AtpD UTR &
AGAGG::AGGAGG 20
AtpD UTR from -30-1 21
AtpD UTR from -30-1
& AGAGG::AGGAGG 22
AtpD
P214L
wt 23
AtpA UTR from -45-31 24
AtpA
P281L
wt 25
AtpD UTR 26
AtpA
R279P wt 27
AtpD -
AtpA 12aa
wt 28
AGAGG::AGGAGG 29
AtpA UTR from -45-31 30
AtpA -
AtpD 12aa
wt 31
AtpD UTR 32
AGGGGA::AGGAGG 33
67
4.2.9 Modeling the interplay between translation initiation rate and elongation stalls
My observation of a maximum expression in the efp mutant as initiation rate increases
prompted me to expand upon the relation between EF-P dependent stalls and initiation rate. I
employed high-throughput datasets from published works measuring EF-P dependence such as
my Salmonella efp mutant SILAC data110, another group’s published E. coli efp mutant SILAC
data130, or ribosome profiling data127. I attempted to correlate these measures of EF-P
dependence with analogs for translation initiation rate including ribosome profiling reads127,
protein abundance110,130, or protein per mRNA133 in wild-type cells. None of these analyses
yielded a significant correlation (data not shown). This may be due to significant variation in the
strength of different stalling sequences, which can be influenced by upstream amino acids in an
unpredictable manner. Thus EF-P dependence may be observed for some weakly expressed
genes if they have a strong stall motif, and similarly some highly expressed genes encoding a
polyproline motif may not show EF-P dependence if the motif triggers weak stalling. This
heterogeneity may mask the correlation between translation initiation rate and EF-P dependence
in these high-throughput datasets.
To further explore the interplay between initiation rate and the stall strength of not just
EF-P dependent but also other elongation pauses, Dr. Assaf Katz generated a computational
model simulating translation of a hypothetical transcript such that he could modulate the
initiation rate and elongation stall strength. A thorough description of this method and the results
obtained by it can be found in Hersch et al. (2014)126. In brief, by comparing the number of
ribosomes able to complete translation on transcripts with varying strengths of initiation rate and
rate of progression through an elongation stall, the computational modeling results follow a
similar curve as the data I observed for atpD and atpA wherein the impact of the elongation stall
increases with initiation rate. This modelling work strongly supports my hypothesis that
initiation rate and EF-P dependent stalls in elongation contend to be the rate-limiting step of
translation.
68
4.2.10 Fate of stalled peptides in vivo
A caveat of my rate-limiting step hypothesis and Assaf’s computational modelling is that
they assume that ribosomes do not undergo premature release from the transcript during an
extended stall at an EF-P dependent motif. To assess whether this is a frequent occurrence in a
Salmonella efp mutant, I reanalyzed peptide data from my previous SILAC investigation110. In
this experiment proteins were fragmented into tryptic peptides that were quantified by mass
spectrometry. I examined proteins identified in SILAC that contain an APP, PPG or PPP motif
and I binned individual peptides observed in mass spectrometry as either ‘pre-’ or ‘post-motif’. I
predicted that premature disengagement of the ribosome would release truncated peptides, which
would manifest as a relative increase in peptides observed upstream of the polyproline motif
compared to peptides after the motif in the efp mutant strain. Such a decrease should not be
observed in the wild-type strain where the entire protein should be synthesized to completion.
I constrained my analysis to proteins with at least two peptides conclusively identified
both before and after the motif, allowing for statistical comparison by t-test of pre- and post-
motif peptides. Out of the 40 proteins for which sufficient peptide data was available, only one
demonstrated a significant difference in relative expression pre- and post-motif (Figure 23).
Furthermore, SILAC ratios pre- and post-motif were linearly correlated with an R2 value of
0.851. Though it remains possible that truncated peptides are being degraded more rapidly than
they can be detected, this data suggests that ribosomes do not prematurely disengage from the
transcript at most polyproline motifs in the absence of EF-P, but rather eventually resume
synthesis and complete the full-length protein.
69
Figure 23: Peptide abundance ratios do not change significantly before and after APP, PPG
or PPP motifs in Salmonella. A) My SILAC data was analyzed comparing the mean peptide
ratios (WT/∆efp) before and after the first APP, PPG or PPP motif of the protein. Only proteins
with an APP, PPG or PPP motif and at least two peptides conclusively identified both before and
after the motif were analyzed (40 proteins). A t-test was used to calculate statistical significance
comparing the average peptide abundance ratio pre- versus post-motif. Gene names and p-values
are shown for proteins with a difference between pre- and post-motif of greater than two. AtpD
and AtpA are also indicated. The only protein with p < 0.05 is highlighted as a black square
(Lon). The linear regression is shown with coefficient of determination (R2) indicated at right.
Dashed line indicates a 1:1 regression. B) All Lon peptides detected in the my SILAC assay are
shown plotting peptide ratio (WT/∆efp) against their location within the full-length Lon protein.
The dashed line indicates the location of the PPG motif. The average of all peptide ratios before
or after the PPG motif are shown as a solid line and error bars indicate one standard deviation.
Figure is shown as it appears in Hersch et al. (2014), J Biol Chem 126.
70
4.3 Discussion
In this chapter I examined PoxB, which was the first protein to demonstrate PYE
pathway-dependent synthesis in the literature yet does not contain a polyproline motif75,76. I
identified the novel motif GSCGPG as responsible for the dependence of PoxB on EF-P. This
finding expands the known repertoire of EF-P target motifs and demonstrates that, at least in
some instances, they may require sequences longer than three amino acids. This suggests that
certain weaker stall motifs such as the GPG of PoxB may instigate ribosomal stalls if sufficiently
strengthened by upstream interactions.
This concept was reinforced by my work on AtpD and AtpA demonstrating that residues
up to six amino acids upstream of the first proline can impact EF-P-rescuable stalls at PPG
motifs. Specifically, the amino acid two positions prior to the PPG motif (-2 position) had a
drastic impact in the case of AtpD and AtpA. The arginine in that position in AtpA appeared to
have a protective effect on stalling in AtpD, particularly in combination with the adjacent (-1
position) arginine. In contrast the asparagine in the -2 position of AtpD drastically induced EF-P-
dependence in AtpA. Interestingly, this effect was negated by swapping surrounding residues as
well, suggesting that neighbouring residues can dampen the stall-inducing effects.
The six residue motif in PoxB and my data for AtpD and AtpA indicate that upstream
amino acids in the exit tunnel of the ribosome can modulate the induction of EF-P-rescuable
stalling. This is similar to what has been observed for the macrolide-sensitive stalling of
ErmAL164,124, the translocon-relieved stall of SecM61,63,66,134, and the tryptophan-responsive
stalling of TnaC123,135. However, the ability of individual amino acids to promote or protect
against EF-P-rescuable stalling is not straightforward. We are unable to find any simple or
general rule that dictates whether a particular sequence will stall – both in the instance of AtpA
and AtpD as well as in ribosome profiling analyses conducted on efp mutants of E. coli127,136.
Thus, while the data indicates that upstream amino acids influence ribosome stalling, this effect
cannot conclusively be attributed to particular amino acids in particular positions.
Interestingly, even when I swapped up to 40 amino acids upstream of the polyproline
motif (the longest estimates for the length of the exit tunnel69,131), I found that it was only enough
to partially reverse the EF-P-dependence of AtpD and AtpA. Moreover, ribosome profiling
71
conducted by the lab of our collaborator Dr. Michael Ibba indicated that (similar to AtpD)
ribosomes accumulated at the PPG site of AtpA in an efp mutant of E. coli despite this apparent
stalling having no influence on protein synthesis. These observations suggested that there exists a
crucial further factor(s) influencing whether or not a PPG motif-containing protein will be
dependent on EF-P for its efficient translation.
Upon expanding my swap analysis of AtpD and AtpA, I found that the 5’ UTR plays a
crucial role in governing EF-P dependence, and all data suggests that this is through its effect on
translation initiation rate. Simply put, the data suggest that many proteins containing polyproline
motifs are unaffected by the loss of EF-P because poor initiation limits the rate of synthesis more
than the stall itself. It follows that the impact of polyproline motifs on protein expression can be
superseded not only by the rate of translation initiation, but also by the rate of elongation through
alternate stalls, including slow codons or starvation-sensitive regions53,56,58,137,138. That is to say,
an EF-P dependent stall will only have an observable effect on protein level if it impacts the rate
of synthesis more than initiation and any other step in translation, including elongation or
termination. This is a parsimonious explanation for the many instances observed when a
polyproline containing protein does not demonstrate EF-P dependence in vivo74,84,110,130.
Experimentally it is possible to observe an influence of both initiation and the stall
sequence on translation. This can occur in situations where initiation and stalling impose similar
constraints on the rate of protein synthesis and is due to heterogeneity in the population of
transcripts. In a given population of mRNAs some transcripts will successfully initiate several
rounds of translation while other transcripts will fail to initiate at all. Furthermore, the EF-P
dependent stalls on some transcripts will stochastically resolve faster than others. When
transcripts have high overall initiation rates, most transcripts in the population will have engaged
with more ribosomes than can be cleared through the polyproline induced stall in the absence of
EF-P, resulting in a buildup of blocked ribosomes upstream of the stall (Figure 24). For
messages where initiation rates are low, ribosome binding is sufficiently rare that most stalled
ribosomes will resume elongation before the next ribosome reaches the stall site. In the latter
case, the polyproline motif will have no observable effect on protein levels. The net output of
protein depends on how the population of transcripts is distributed among these states.
72
The model in Figure 24 highlights the rate-limiting interplay between translation
initiation and clearance of ribosome stalls, and it is supported by my experimental results. For 5’
UTR mutants of atpD and atpA, as GFP fluorescence increased in wild-type cells, expression in
the efp mutant appeared to approach an asymptote where stall clearance becomes the
predominant rate-limiting step of translation. This effect was dependent on an intact PPG motif
and implies a calculable maximum expression in the efp mutant that is related to the rate of stall
clearance. This stall strength is particular to individual ORFs and allows for a comparison of EF-
P dependence across all initiation rates. Interestingly, the slope of the atpD curve declines much
more rapidly than that of atpA and reaches a significantly lower maximum expression,
supporting the idea that the EF-P alleviated stall of AtpD is more difficult to bypass than that of
AtpA. This is supported by my previous work, where I found that the region upstream of the
PPG motif plays a significant role in the degree of EF-P dependence. Indeed, when I swapped
twelve codons upstream of the PPG motifs, expression at multiple different initiation rates
resembled the protein from which the twelve codons originated. Taken together, my data
demonstrate that the region upstream of the PPG motif influences the strength of the stall and
thereby the steepness and maximum of the expression curve as a function of initiation rate. The
in silico translational model generated by our collaborators supports this concept.
My results also suggest that stalls at polyproline motifs in the absence of EF-P are
eventually resolved and that protein synthesis proceeds to completion in most cases. If the
ribosome disengaged from polyproline induced stalls, then I would expect that proteins
translated with low rates of initiation would show a degree of EF-P dependence similar to that of
proteins derived from transcripts with much higher initiation rates. This supposition is supported
by my SILAC data, where I do not observe a general decrease in peptide counts after the stall
sequence compared to sequences upstream of the stall.
It is unlikely that the balance between translation initiation rate and impaired elongation
is limited to EF-P dependent stalls. Indeed, Assaf’s computational modelling is not restricted to
EF-P dependent pauses and could be applied to any stall in translation elongation regardless of
mechanism. The correlation between codon usage and gene expression is well described, where
highly expressed genes have a strong bias toward using codons that correspond to the most
abundant tRNAs50,52,54,139,140. Chu et al. observed that the impact of slowly translated codons on
protein synthesis could be modulated by altering the 5’ UTRs of reporter plasmids in yeast57.
73
Other studies have described a translation ‘ramp’ wherein slow, non-preferred codons are more
abundant at the 5’ end of a coding region, and this is supported by ribosome profiling data where
increased ribosome occupancy was observed at the 5’ end of ORFs141,142. The length of this slow
ramp was significantly larger for transcripts with high ribosome densities, suggesting
evolutionary pressure on highly expressed genes to prevent downstream ribosome jamming57,141-
143. For various transcripts, such a ramp could also allow time for protein folding, binding of
cofactors to the N-terminal peptide, or enabling stalls to affect translation rate via blocking new
ribosomes from binding the mRNA56,144-147.
Cumulatively, these works support the interplay that I have presented here between
translation initiation rate and elongation stalls. This concept, demonstrated herein for EF-P
dependent pauses, can be readily modeled in silico and can be further applied to other forms of
ribosome stalls. Moreover, it suggests that including a weak ribosome binding site and thereby a
low rate of translation initiation can be protective against downstream ribosome stalls impacting
protein synthesis. For instance, if a peptide requires a translational stall for proper regulation of a
downstream gene, a low initiation rate could protect the transcript against ribosome jamming.
This establishes a mechanistic rationale for evolutionary selection of weak, inefficient ribosome
binding sites that lead to poor translation initiation rates on certain mRNA transcripts.
74
Figure 24: Model of the interplay between translation initiation rate and elongation stalls.
Depiction of translation of an mRNA transcript with low initiation rate (left) or high initiation
rate (right) and the subsequent effect of no elongation stall (top), a weak stall (middle) or a
strong stall (bottom). Black line, mRNA transcript. Purple box, start codon. Brown box, stop
codon. Blue ovals, translating ribosomes. Yellow stop sign, weak stall motif. Red stop sign,
strong stall motif. ++++, high degree of protein synthesis per transcript. ++, medium protein
synthesis per transcript. +, low level of protein synthesis per transcript. Figure is shown as it
appears in Hersch et al. (2014), J Biol Chem 126.
75
5 Salmonella evolved to suppress its uptake of dicarboxylic
acids during stress
Acknowledgements
Bushra Ilyas contributed greatly to testing growth using fumarate and malate in addition
to succinate during her undergraduate thesis research under my supervision. Figure 25C presents
data that was produced by her and is shown here to support my work. Bushra also generated the
original dctA mutant strain that I used in multiple experiments.
Bojana Radan was involved in the intial screen of compounds leading to the discovery
that proline can stimulate growth in succinate. As well, Bojana Radan was involved in testing
select SGSC and ECOR strains as a pilot study preceding my extensive screen of these
collections.
In this chapter I mention an unpublished microarray conducted on rpoS mutant
Salmonella. This was conducted by my supervisor, Dr. William Navarre, during his post-doctoral
work and in collaboration with Dr. Michael McClelland and Dr. Jonathan Frye.
76
5.1 Overview
Previous work employing Biolog Phenotype MicroarraysTM revealed that PYE mutants
display increased susceptibility to a wide variety of cellular stressors including
pharmacologically unrelated antibiotics74,84. The phenotype microarray data also demonstrated
that PYE mutants appear to display a hyper-active metabolism relative to wild-type Salmonella
when under specific nutrient limited conditions74,84. This indicates that perhaps efp mutants are
metabolically overactive and this could contribute to the hypersusceptibility phenotype due to a
failure to shut down metabolism in response to stress. In this chapter I set out to test metabolic
phenotypes of efp mutant Salmonella, yet I discovered a fascinating and counter-intuitive
adaptation of wild-type Salmonella. Specifically, when exposed to minimal media conditions
where the sole carbon source is a dicarboxylic acid such as succinate, wild-type Salmonella shuts
down its growth for an extended, yet surprisingly consistent, length of time (over 30 hours). In
contrast, PYE mutants, and also mutants in the stress response sigma factor rpoS, do not show
this prolonged lag phase and grow readily under these conditions. Moreover, this phenotype
appears to be a Salmonella specific phenomenon, as many strains of the closely related species
E. coli grow readily under the same conditions. I describe my progress toward identifying the
mechanisms underlying this Salmonella adaptation and how it represses growth in response to
dicarboxylates.
5.2 Results
5.2.1 Wild-type but not efp or rpoS mutant Salmonella displays delayed growth using
dicarboxylic acids as a sole carbon source
PYE mutants appear to display a hyper-active metabolism relative to wild-type
Salmonella when under specific nutrient limited conditions74,84,88. As well, my SILAC analysis
presented in Chapter 3 revealed that efp mutants have reduced levels of a number of metabolic
proteins, including members of the ATP Synthase complex such as the catalytic subunit AtpD,
which showed a 20-fold lower abundance in the Salmonella efp mutant. In light of these
findings, I attempted to examine metabolic phenotypes of efp mutants.
77
I first aimed to assess the ability of efp mutants to construct a functional ATP Synthase
complex by testing their growth in minimal media with succinate as the sole carbon source. This
assay was first described in 1971 with the rationale that when utilizing succinate as a carbon
source E. coli can only produce ATP via oxidative phosphorylation, and so strains lacking a
functional ATP Synthase cannot grow148,149. Of note, this rationale was later disproved by
demonstrating that the no-growth phenotype can be bypassed by overexpressing dctA, the
primary aerobic transporter of dicarboxylic acids – suggesting that ATP could be synthesized via
the TCA cycle yet loss of ATP Synthase activity somehow leads to repression of dctA150. Upon
testing the growth of Salmonella efp mutants in minimal media containing succinate as the sole
carbon source, I found that they were able to grow readily. This was similar to the Salmonella
rpoS mutant employed as a positive control (since E. coli rpoS mutants were previously
demonstrated to show improved growth on succinate151). Fascinatingly, and in stark contrast to
the efp and rpoS mutants, wild-type Salmonella appeared to repress its growth and exhibited an
extended lag phase for over 30 hours before initiating logarithmic growth (Figure 25A and B).
The fact that rpoS and efp mutants in an isogenic background could grow readily suggested that
the strain of Salmonella was capable of growing on succinate yet makes a regulatory decision not
to. The data suggest that this regulation involves RpoS and some unknown protein(s) that
requires efp for its efficient translation. As well, the dicarboxylic acid transporter, DctA, was
required for growth, and mutants in the dctA gene showed no sign of growth by 48 hours.
The extended lag of wild-type Salmonella also occurred during growth using two other
dicarboxylic acids, fumarate and malate (Figure 25C). Consistent with succinate, the efp mutant
grew significantly earlier than wild-type Salmonella using fumarate or malate as a sole carbon
source. This suggests that the growth repression instigated by wild-type Salmonella is not
specific to succinate but also occurs during growth using other dicarboxylic acids.
78
A)
B) C)
Figure 25: Wild-type Salmonella displays an extended lag phase using dicarboxylic acids as
a sole carbon source but efp and rpoS mutants do not. A) Representative growth curve of
Salmonella Typhimurium 14028s strains grown in MOPS minimal media with 0.2% succinate as
the sole carbon source. Data shows optical density at 600nm (OD600) for 48 hours from the time
of inoculation. The data shown is representative of greater than three biological replicates for all
strains. An OD600 of 0.1 is emphasized by a dashed line. B) Graphs showing the average time in
hours that wild-type (WT) or mutant Salmonella takes to reach an OD of 0.1 as an analog of the
length of lag phase using succinate as the sole carbon source. The data shows the average of at
least three biological replicates and error bars show one standard deviation. >48h indicates the
strain did not grow by 48 hours. C) As in (B) but comparing the use of three different
dicarboxylic acids as the sole carbon source. Data for panel C was produced by Bushra Ilyas
during her undergraduate thesis work under my supervision.
0
0.1
0.2
0.3
0.4
0 5 10 15 20 25 30 35 40 45 50
OD
(6
00
nm
)
Time (hours)
WT
∆efp
∆rpoS
∆dctA
0
10
20
30
40
Tim
e (
ho
urs
) to
rea
ch
OD
≥ 0
.1
∆rpoS ∆dctA∆efpWT
>48h0
10
20
30
40Ti
me
(h
ou
rs)
to r
each
O
D ≥
0.1 Succinate
Fumarate
Malate
∆efpWT
79
5.2.2 Many Salmonella but few E. coli strains delay growth using succinate
My finding that wild-type Salmonella significantly delays its growth using dicarboxylic
acids as a sole carbon source prompted the question of whether this was a common phenomenon
or specific to Salmonella. To address this I compared Salmonella to the closely related bacterium
E. coli and found that E. coli grew readily in minimal media using succinate as a sole carbon
source (Figure 26A). To examine if this difference extended to additional strains, I conducted a
screen testing growth using succinate for all 105 non-typhoidal strains in the Salmonella Genetic
Stock Centre (SGSC) collection, as well as all 72 strains of the E. coli Reference (ECOR)
collection. Indeed, I found that (though there are exceptions) the majority of E. coli strains grow
more readily in succinate media compared to most Salmonella strains, which display extended
lag phases (Figure 26B-E). Once logarithmic growth was initiated, Salmonella also appeared to
trend towards a slightly longer doubling time than the majority of E. coli strains (Figure 26C and
F). To ensure that the observed effects were not due to variations in RpoS, I also tested all SGSC
and ECOR collection strains for catalase activity as an analog for functional RpoS. I found that
regardless of catalase activity the same trend held true that E. coli strains generally showed
shorter lag phases than Salmonella when using succinate as the sole carbon source (Figure 26D).
80
0
0.2
0.4
0.6
0.8
0 10 20 30 40 50
OD
(60
0n
m)
Time (h)
E. coli K12 WT Salmonella WT Salmonella ∆rpoSE. coli K12Salmonella 14028s
Salmonella 14028s ∆rpoS
All SGSC SalmonellaSubsp. entericaAll ECOR E. coli
0
1
2
3
4
5
0 10 20 30 40 50 60
Do
ub
ling
tim
e (h
ou
rs)
fro
m O
D 0
.1 t
o O
D 0
.2
Time (hours) to reach OD ≥ 0.1
Salmonella
Salmonella with a >72h rep
E. coli
E. coli with a >72h rep
Salmonella
Salmonella with a >72h rep
E. coli
E. coli with a >72h rep
0
10
20
30
0 20 40 60 80
0
0.5
1
0 10 20 30 40 50 60
Cat
alas
e
Time (hours) to reach OD ≥ 0.1
Salmonella
Salmonella with a >72h rep
E. coli
E. coli with a >72h rep
+
−
+/−
Salmonella
Salmonella with a >72h rep
E. coli
E. coli with a >72h rep
All SGSC Salmonella
All ECOR E. coli
All SGSC Salmonella
All ECOR E. coli
A) B)
C)
D)
E) F)
81
Figure 26: E. coli grows earlier than Salmonella using succinate as the sole carbon source.
Growth in MOPS minimal media with 0.2% succinate as the sole carbon source. Growth was
conducted in a TECAN Infinite M200 plate reader and reads were taken every 15 minutes. A)
Representative growth curve. An OD600 of 0.1 is emphasized by a dashed line. B) Percentage of
SGSC Salmonella (green line) or ECOR E. coli (blue line) strains that (on average) had
surpassed an OD600 of 0.1 by indicated times post inoculation. Red line is for Salmonella but
excludes the 14 tested strains of the SARC collection that do not belong to subspecies enterica.
C) Overview of 105 Salmonella strains (SGSC collection) and 72 E. coli strains (ECOR
collection). Each strain is plotted by average time it takes to reach OD 0.1 (x-axis) compared to
doubling time during growth from OD 0.1 to OD 0.2. All Salmonella are coloured black and all
E. coli strains are coloured red. Strains where at least one replicate that did not grow by 72 hours
are shown as triangles; these values are the average of the remaining replicates. Inset at top right
shows zoomed out view to include outliers. D) Similar to (C) but comparing time to reach OD
0.1 with catalase result. + or − indicate that the strain showed positive or negative (respectively)
catalase result in all replicates. +/− indicates either delayed positive or inconsistent catalase
results across replicates. Data is the average of at least three biological replicates with the
exception of some strains that had at least one replicate that did not reach OD 0.1 by 72h (shown
as triangles that are the average of the remaining replicates). E) Average time in hours for strains
to reach an OD of 0.1. Data compares all 105 Salmonella strains to all 72 E. coli strains tested.
Line indicates the median, boxes show the 25th to 75th percentiles, and whiskers show the 10th
to 90th percentiles. An unpaired t-test with Welch’s correction indicated a p-value < 0.0001. F)
As in (E) but comparing doubling time in hours. An unpaired t-test with Welch’s correction
indicated a p-value < 0.05 with all data points included or p-value < 0.0001 when excluding slow
growing outliers (doubling time > 5h).
82
5.2.3 The RpoS stabilizer IraP contributes to growth shutdown in succinate media
The repression of wild-type Salmonella’s growth when using dicarboxylic acids as the
sole carbon source appears to involve the general stress response sigma factor RpoS, since
mutation of the rpoS gene resulted in early growth on succinate (Figure 25). One of the major
mechanisms by which RpoS activity is regulated – particularly under nutrient stress – is via its
degradation, and three anti-adaptor proteins are known (IraP, RssC and IraD) that inhibit RssB to
stabilize RpoS in response to specific kinds of stress (Figure 1)17,152. I investigated whether one
of these known RssB anti-adaptors is involved in sensing succinate media as a stress and
instigating growth shutdown via RpoS stabilization. I generated deletions of each of the anti-
adaptor genes and tested the mutant strains for growth in minimal media with succinate as the
sole carbon source. I found that deletion of iraP led to drastically earlier growth of Salmonella
under these conditions (similar to deletion of rpoS), and this phenotype could be partially
complemented by expressing IraP from its native promoter on a plasmid (Figure 27). In contrast,
deletion of the other anti-adaptors, RssC and IraD, had only minor or negligible effects. Deletion
of the rssB gene itself yielded inconsistent results, ranging from growth similar to wild-type
Salmonella to early growth similar to an rpoS mutant. This is likely due to suppressor mutations
in rpoS arising in this strain to compensate for the lack of rssB-mediated RpoS degradation. In
summary, though the mechanism by which iraP is induced remains unknown, my findings
demonstrate that IraP plays a role in the RpoS-mediated repression of Salmonella’s growth using
succinate as the sole carbon source.
83
A)
B)
C) D)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(6
00
nm
)
Time (h)
WT
ΔrpoS
ΔIraP
∆rssB
∆rssC
∆iraD
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(6
00
nm
)
Time (hours)
WT
ΔrpoS
ΔIraP
ΔIraP pIraP
0
10
20
30
40
50
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
0
10
20
30
40
50
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
WT
ΔrpoS
ΔiraP
ΔiraP pIraP
WT
ΔrpoS
ΔiraP
ΔrssB
ΔrssC
ΔiraD
84
Figure 27: Deletion of the RpoS stabilizer IraP results in early growth on succinate. Growth
of Salmonella in MOPS minimal media with 0.2% succinate as the sole carbon source. A) The
three known RssB antiadaptors were deleted from the Salmonella chromosome and growth is
shown along with an rssB mutant. Data is representative of at least three biological replicates for
all but the ΔrssB strain, which varied between replicates as described in the text. B) Exogenous
expression of the Salmonella iraP gene from its native promoter on a plasmid partially
complements the growth delay phenotype. Complementation data is representative of two
biological replicates. C and D) Graphs showing the average time in hours that wild-type (WT) or
mutant Salmonella takes to reach an OD of 0.1 as an analog of the length of lag phase using
succinate as the sole carbon source. The data shows the average of at least three biological
replicates for all strains in (C) except ‘∆iraD’ for which only two replicates were conducted.
Data in (D) is the average of two biological replicates. Error bars show one standard deviation.
85
5.2.4 Supplementation with proline or citrate induces diauxic growth using succinate and
subsequent repression requires the stringent response
To examine how IraP may be induced and whether Salmonella requires additional
nutrients, I supplemented succinate media with various compounds. The addition of minute
amounts of either proline or citrate induces growth in succinate media in a manner resembling a
diauxy wherein growth is repressed again following depletion of the proline or citrate (Figure
28A and C). This suggests that the addition of these compounds can provide Salmonella with a
metabolite that is either limiting with succinate as a sole carbon source or can act as a regulatory
signal stimulating growth. Interestingly, this critical metabolite does not appear to be proline
itself as growth induction by proline (but not citrate) required the enzyme PutA, which degrades
proline to glutamate.
In parallel, I also examined the role of a number of Salmonella mutants involved in
regulating starvation, including a relA spoT double mutant that cannot produce the stringent
response secondary messenger ppGpp. This ppGpp0 strain shows slightly earlier growth in
succinate than wild-type Salmonella but moreover does not repress growth following proline or
citrate stimulation (Figure 28B and D). This suggests that following induction by proline or
citrate ppGpp plays a significant role in restoring Salmonella’s repressed state and this signal
may contribute to growth shutdown even in the absence of these inducers.
86
A)
B)
C) D)
WT
ΔrpoS
WT + Pro
WT + Cit
ΔputA + Pro
ΔputA + Cit
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(6
00
nm
)
Time (hours)
WT
ΔrpoS
ppGpp0
ppGpp0 + Pro
ppGpp0 + Cit
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(6
00
nm
)
Time (hours)
0
10
20
30
40
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
0
10
20
30
40
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
WT
Pro Cit
∆putA
Pro Cit
ppGppO
Pro Cit
87
Figure 28: Proline or citrate induce growth using succinate and subsequent repression
involves the stringent response. A) Representative growth curve of Salmonella growth in
MOPS minimal media with 0.2% succinate as the primary carbon source and supplemented with
0.005% proline (+ Pro) or 0.005% citrate (+ Cit) where indicated. B) Representative growth
curve of Salmonella growth in MOPS minimal media with 0.2% succinate as the primary carbon
source. A relA spoT double mutant that cannot make the stringent response messenger ppGpp
(ppGpp0) exhibits an intermediate growth phenotype and does not shut down growth following
proline or citrate induction. C and D) Graphs showing the average time in hours that wild-type
(WT) or mutant Salmonella takes to reach an OD of 0.1 as an analog of the length of lag phase
using succinate as the sole carbon source. The data shows the average of at least three biological
replicates for all except ‘∆putA + Cit’ for which only two replicates were conducted. Error bars
show one standard deviation.
88
5.2.5 Growth lag appears to be due to repression of succinate import
I examined unpublished microarray data of a Salmonella rpoS mutant conducted during
Dr. Navarre’s post-doctoral research. I noticed that the gene encoding the primary dicarboxylate
transporter, dctA, demonstrated up to a 4.4 fold increase in expression in the rpoS mutant. This
suggests that RpoS may repress the expression of this transporter and thereby restrict Salmonella
from taking up dicarboxylates such as succinate for growth. To test if wild-type Salmonella does
not grow on succinate due to repression of dctA, I constitutively expressed dctA from a plasmid
and tested growth in minimal media with succinate as the sole carbon source. I found that
constitutive expression of dctA but not lacZ resulted in earlier growth in succinate media,
indicating that the limiting factor in Salmonella’s growth using succinate is synthesis of the
dicarboxylate importer DctA (Figure 29A and C). It therefore appears that a role of RpoS in
Salmonella is to repress dctA expression and restrict the uptake of dicarboxylic acids.
5.2.6 The E. coli dctA promoter is sufficient to induce Salmonella growth using succinate
In light of my findings that Salmonella’s suppressed growth phenotype is largely
mediated by repression of the dctA gene, I examined the role of the dctA promoter (PdctA). Since
the phenotype appears to be divergent between Salmonella and the closely related species E. coli,
I swapped the E. coli dctA promoter into the Salmonella chromosome using an upstream
chloramphenicol resistance cassette to select for successful recombination. As a control for
artefacts of introducing the chloramphenicol resistance cassette, I conducted the same swap but
inserted Salmonella’s native dctA promoter and found no difference from wild-type Salmonella
when grown in minimal media with succinate as the sole carbon source. In contrast, I found that
replacing the Salmonella dctA promoter with that of E. coli was sufficient to abolish
Salmonella’s ability to repress its uptake of dicarboxylic acids, and this strain grew readily in
succinate media (Figure 29B and D).
89
A)
B)
C) D)
Figure 29: Expression of dctA induces growth using succinate. Growth of Salmonella in
MOPS minimal media with 0.2% succinate as the sole carbon source. A) Overexpression of dctA
from a plasmid induces Salmonella growth using succinate. WT pLacZ and pDctA indicate wild-
type Salmonella containing a pXG10sf plasmid encoding full length lacZ or dctA with
expression driven by the constitutively active PLtet0-1 promoter. The rpoS mutant is shown for
comparison. Data is representative of three biological replicates. B) The E. coli dctA promoter
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(6
00
nm
)
Time (hours)
WT
∆rpoS
WT pXG10sf
WT pXG10sf-DctA
WT
ΔrpoS
WT pLacZ
WT pDctA
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50
OD
(60
0n
m)
Time (hours)
WT
CmR-PdctAsalm in chrom
CmR-PdctAcoli in chrom
∆iraP
S. Typhimurium
PdctA (Salm.)
PdctA (E. coli)
ΔiraP
0
10
20
30
40
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
0
10
20
30
40
Tim
e (h
ou
rs)
to
reac
h O
D ≥
0.1
∆rpoSWT WT pLacZ
WT pDctA
PdctA
(Salm.)WT PdctA
(E. coli)∆iraP
90
(PdctA E. coli) with an upstream chloramphenicol resistance cassette for selection was inserted
into the Salmonella chromosome by lambda red recombination to replace the native dctA
promoter. Replacement with Salmonella’s native dctA promoter controls for effects due to the
chloramphenicol resistance cassette (PdctA Salm.). ΔiraP is shown for comparison. Data shown is
representative of greater than three biological replicates. C and D) Graphs showing the average
time in hours that wild-type (WT) or mutant Salmonella takes to reach an OD of 0.1 as an analog
of the length of lag phase using succinate as the sole carbon source. The data shows the average
of at least three biological replicates and error bars show one standard deviation.
91
5.3 Discussion
In this chapter I discovered not just a PYE mutant phenotype but also a broader and more
interesting characteristic of wild-type Salmonella itself and how it regulates its metabolism in
response to environmental conditions that it perceives as stressful. I chose to follow up this
avenue of research and, through my work, I found that wild-type Salmonella restricts its growth
when using dicarboxylates as a sole carbon source and exhibits an extended lag phase lasting
over thirty hours. This phenotype does not reflect a metabolic inability of Salmonella to utilize
dicarboxylates, as I have discovered multiple mutations in regulatory genes that allow the cells to
grow early in succinate media. It appears that Salmonella employs RpoS, IraP and to some
degree RelA or SpoT to sense this environment as a stress and shut down expression of dctA to
restrict its import of dicarboxylic acids. Interestingly, growth is also stimulated by the addition of
small amounts of proline or citrate, suggesting that these supplements can alleviate growth
repression. Subsequent growth repression following proline or citrate induction required ppGpp.
The stringent response and ppGpp have been demonstrated to impact the expression of both rpoS
and iraP, suggesting that RelA or SpoT may be involved in the initial sensing of succinate media
as a stress and activating the IraP- and RpoS-mediated shutdown of growth17.
Importantly, I demonstrate that this phenotype is divergent between Salmonella and the
closely related (but non-intracellular) bacteria, E. coli. I discuss this phenotype as a trait that
Salmonella has acquired since diverging from E. coli, however it should be noted that it remains
possible that the opposite is true and E. coli has lost the ability to repress the uptake of
dicarboxylates since diverging from Salmonella. The trend held true amongst most tested strains
from the SGSC and ECOR collections, which sample a wide range of the genetic diversity of
these species. Yet there are exceptions, including a number of Salmonella strains that grew early
despite having a catalase positive phenotype. In all instances where multiple strains from the
same Salmonella serovar were tested, at least one exhibited an extended lag phase
(Supplementary DataTable 2). Thus the earlier growth does not appear to be a trait of particular
serovars but rather individual strains seem to have lost the delayed growth phenotype. This could
occur by mutations in genes other than rpoS, such as efp or iraP, that grant early growth in
succinate while remaining catalase positive. Alternatively, some of these strains may never have
acquired this adaptation as opposed to lost it due to mutation. This particularly applies to the
more phylogenetically distinct Salmonella that do not belong to subspecies I (subsp. enterica),
92
where 11 of the 14 tested strains grew earlier than the Salmonella average of 29 hours
(Supplementary DataTable 2). This suggests that delayed growth using succinate may be an
adaptation that primarily benefits subsp. enterica Salmonella, which includes the majority of
serovars responsible for human infections153. However, this is confounded by the fact that 5 of
these strains tested catalase negative, suggesting that their result may simply reflect a mutation in
rpoS. Combined with the low number of non-subsp. enterica Salmonella strains tested, there is
insufficient data to support the subspecies I concept conclusively.
Other exceptions include some Salmonella strains that display an extended lag phase
despite having no catalase activity. It is possible that these strains contain a mutation in their
catalase gene itself as opposed to rpoS. Alternatively, in addition to mutations altering RpoS
activity, these strains may also contain suppressor mutations in genes such as the dicarboxylate-
sensing two-component system dcuSR, which could delay growth in this assay. There also
remains the possibility that these strains employ an alternate RpoS-independent mechanism of
regulating the delayed growth phenotype
Finally, there are multiple E. coli strains that are catalase positive yet show an extended
lag in their growth using succinate. While it is possible that these have mutations in genes
required for the uptake of succinate (such as dctA or dcuSR), these may be genuine variations in
how E. coli strains respond to succinate. Since a major difference between Salmonella and E.
coli is their virulence, and the ECOR collection contains a number of strains that have been
shown to demonstrate indicators of virulence, I compared whether these potentially virulent E.
coli strains correlated with growth in succinate. Seven of the 72 E. coli strains averaged longer
than 24h to reach an OD of 0.1 in succinate media and four of these averaged longer than 29h
(the average time for all tested Salmonella strains) (Supplementary DataTable 2). Of these seven,
only ECOR 52 and to a lesser degree ECOR 48 demonstrated hemolytic and cytotoxic activity in
the literature, suggesting that growth using succinate is not correlated with indicators of
virulence in these E. coli strains154.
I discovered that replacing the Salmonella dctA promoter with that of E. coli is sufficient
to abolish the ability of Salmonella to repress its uptake of succinate. This result strongly
suggests that, since diverging from E. coli, Salmonella has obtained a regulatory element in its
dctA promoter that allows it to be repressed under these conditions. My work and microarray
93
data suggest that RpoS is involved in this regulation; yet RpoS is a transcriptional activator. The
lack of growth on dicarboxylates suggests that dctA is tightly repressed, suggesting against solely
sigma factor competition for RNA polymerase, but rather that RpoS likely acts via an
intermediate and yet unknown transcription factor (potentially a protein factor or a small RNA)
(Figure 30). Based on my chromosomal promoter swap experiment, the binding site of this
repressor is likely in a region that differs between the Salmonella and E. coli dctA promoters. An
alignment of these promoters demonstrates a sequence identity score of 77.9%, indicating that
overall the promoter is fairly well conserved. However there are specific distinctions such as a 9
bp deletion (in the Salmonella promoter) adjacent to a region of dissimilarity approximately 100
bp upstream of the conserved -35 box of the promoter (Figure 31). This distinct sequence could
be sufficient to recruit a repressor to the Salmonella promoter and would be eliminated if
replaced by the E. coli promoter. However, my preliminary analyses using prediction software
(such as BPROM) and examination of transcription factor consensus sequence databases (such as
RegulonDB) did not yielded significant leads regarding a specific regulator that could bind to
this sequence.
The difference in the response of Salmonella and E. coli to dicarboxylic acids may offer
important clues to identifying the evolutionary advantage conveyed by this adaptation. Since
many of the traits that Salmonella has acquired since their divergence are related to its
pathogenic lifestyle, it follows that this phenotype may reflect a situation that Salmonella
encounters during infection of a host. The recent finding that succinate accumulates to high
levels in activated macrophage suggests that Salmonella’s intracellular survival in macrophage
may represent this crucial selective environment155. It is conceivable that Salmonella recognizes
the succinate produced by activated macrophage and restricts its uptake of this dicarboxylate in
response. In the next chapter I will expand upon this hypothesis.
94
Figure 30: Working model of Salmonella’s response to succinate as the sole carbon source.
IraP becomes induced and stabilizes rpoS. Since rpoS is an activator of transcription, it follows
that its activity increases expression of an unknown repressor that turns off expression of dctA.
Since replacing the Salmonella dctA promoter with the E. coli one induces growth on succinate,
it is likely that this repressor binds to a region that is divergent between the two promoters.
Without DctA, the cells do not import their only carbon source (succinate) and do not grow. The
stringent response second messenger ppGpp and supplemented proline or citrate influence
growth using succinate; however, their mechanism and position in the signalling cascade remains
uncertain.
RpoS(σS)
dctA geneSuccinate
import and growth
Succinate as sole carbon source
IraP
Repressor?
Divergent between Salmonella and E. coli
?ppGppProline, Citrate
95
Figure 31: Alignment of Salmonella and E. coli dctA promoters conducted using ClustalΩ
software156. The promoters span 500bp each, which was the length swapped into the Salmonella
chromosome in Figure 29B and D. Sequence from E. coli K12 is on top, sequence from
Salmonella 14028s is on bottom. Asterisks indicate identity and the two promoters share 77.9%
identity across the 500bp. Red rectangles indicate the 5’ UTR identified by RegulonDB157 and in
references 158,159. Blue rectangles indicate the predicted -10 and -35 boxes identified by
RegulonDB157 and reference 158.
ATG start codon
96
6 Salmonella versus macrophage-produced dicarboxylic
acids
Acknowledgements
I would especially like to thank Ryan Gaudet from the lab of Dr. Gray-Owen for his
generous donation of the THP-1 monocyte cell line as well as for his guidance and training for
working with them. I would also like to thank Dr. Teresa O’Meara from the lab of Dr. Leah
Cowen for the generous donation of the J774 macrophage cell line.
97
6.1 Overview
In this chapter I probe the question of why Salmonella may have acquired the trait of
blocking dicarboxylate utilization and what evolutionary advantage it may gain by it. Succinate
levels were recently shown to increase significantly in activated macrophage and Salmonella (but
not E. coli) has adapted to survive effectively in this environment; moreover, survival in
macrophage is required for Salmonella virulence30,31,155. I examined the possibility that activated
macrophage produce succinate not only as an inflammatory signal but also as a weapon to inhibit
phagocytosed bacteria, and Salmonella has evolved to reduce its import of dicarboxylic acids as
a defence mechanism. Using dctA overexpression strains and Salmonella containing the E. coli
dctA promoter to induce uptake of succinate, I demonstrate that Salmonella’s repression of
succinate uptake does not influence its survival in either acidified succinate or in macrophage.
Another dicarboxylic acid, itaconate, is also heavily induced in activated macrophage and
inhibits the bacterial glyoxylate shunt160-163. Moreover, itaconate degradation enzymes have been
characterized, and Salmonella encodes orthologs of these genes in an operon164. I examine
Salmonella’s response to this dicarboxylic acid, identifying the regulator of the degradation
operon and its specific response to itaconate. Using this regulator and its target promoter, I
construct an itaconate biosensor and examine the Salmonella itaconate response in macrophage.
However, despite its potent response to itaconate, deletion of the itaconate degradation operon or
its regulator does not influence Salmonella survival in macrophage under the conditions tested.
These data suggest that degradation of itaconate or repressing succinate uptake may offer
advantages other than enhanced survival in macrophage.
6.2 Results
6.2.1 Hypothesis and rationale that Salmonella may repress uptake of dicarboxylic acids to
improve survival in activated macrophage
Macrophage activated by LPS drastically increase their levels of two dicarboxylic acids,
succinate and itaconate, potentially creating a selective niche where repressing uptake of
dicarboxylic acids may provide Salmonella with a survival advantage155,160,162. Succinate in
98
macrophage was demonstrated to play a role in inflammatory signalling via stabilization of HIF-
1α, leading to increased IL-1β synthesis155. Meanwhile itaconate can inhibit the bacterial
glyoxylate shunt, which is required for growth using carbon sources with two or fewer carbons
or when using fatty acids as a carbon source161,163. Moreover, a cluster of genes have been
reported that degrade itaconate, and Salmonella Typhimurium contains orthologs of this operon
in a pathogenicity island (SPI-13)164.
The accumulation of succinate and itaconate in activated macrophage could have a
function beyond the characterized inflammation signalling and inhibition of the glyoxylate shunt.
Both compounds are dicarboxylic acids, indicating that they have two carboxyl groups with
characteristic acid dissociation constants (pKa). Indeed the Salmonella-containing vacuole within
macrophage is about pH 5.0, but the low estimates reach pH 4.4, which is similar to the pKa’s of
succinate and itaconate (Figure 32A)34,165. This suggests that in the acidified phagosome,
succinate and itaconate could become protonated either on both carboxyl groups (di-protonated)
or on one (hemi-protonated). In the di-protonated form, these molecules lose their negative
charge, allowing them to traverse the bacterial membrane freely (Figure 32B). Upon reaching the
more neutral pH of the cytoplasm, the equilibrium shifts to favor release of the proton, thereby
acidifying the bacterial cytoplasm. Thus, acidified organic acids such as succinate and itaconate
can act as proton shuttles to force bacteria to acidify in an acidic environment such as the
phagosome. Indeed I have demonstrated that addition of succinate or itaconate to acidifed LPM
media significantly reduces Salmonella’s survival in vitro (Figure 32C).
In the hemi-protonated form, which would account for the majority of molecules at the
estimated pH of the Salmonella containing vacuole, succinate and itaconate would still require a
transporter such as DctA to gain access to the cytoplasm. Yet they would shuttle a proton and
contribute to acidification. As well, DctA symports dicarboxylates with two protons, resulting in
further acidification (Figure 32B). This suggests that in the context of a macrophage phagosome,
Salmonella may benefit from repressing DctA to exclude succinate (and potentially itaconate) in
order to prevent acidification of its cytoplasm. The repression of growth in succinate media
described in the previous chapter may represent such an adaptation. Moreover the growth lag
phenotype was mediated by repression of dctA via the stress response sigma factor RpoS, which
is heavily involved in the organic acid stress response152,166.
99
A)
B)
C)
Figure 32: Dicarboxylic acids can act as proton shuttles in acidified conditions. A) Depiction
of the chemical structures of succinate, fumarate, malate and itaconate in their acidifed form.
Their pKa’s are indicated. B) Depiction of the rationale behind acidifed dicarboxylic acids acting
as proton shuttles using succinate as an example. At pH above the lower pKa the succinate is
hemi-protonated and remains charged and so cannot freely pass through the membrane without a
transporter. C) Survival of wild-type Salmonella is reduced when succinate or itaconate are
added to acidified LPM media. Data shows percent survival at 3h relative to at 0h and is the
average of at least three biological replicates. Error bars indicate one standard deviation.
pKa = 5.6
pKa = 4.2
pKa = 5.5
pKa = 3.8
Succinate ItaconateFumarate Malate
pKa = 4.4
pKa = 3.0
pKa = 5.1
pKa = 3.4
--
pKa = 5.6
pKa = 4.2
pH = 7
pH ~ 7
DctA
H+
H+
Succinate
-
H+
H+ H+H+ H+
H+
pH = 5
pH ~ 7
DctA
2 H+
Succinate3 H+
Succinate2 H+
Succinate2 H+
Cytoplasm
Periplasm
IM
H+
H+ H+H+
H+
H+
H+
H+
H+
H+
pH = 3
pH ~ 7
Diprotonated succinate
Succinate2 H+
0.001
0.01
0.1
1
10
100
LPM mediapH 4.4
+ 0.4%Succinate
+ 0.4%Itaconate
Perc
ent
surv
ival
af
ter
3 h
ou
rs
100
6.2.2 Constitutive expression of dctA from a plasmid reduces Salmonella survival in acidified
succinate and in macrophage
To assess whether repression of dctA granted Salmonella a survival advantage by
excluding succinate under acidic conditions, I overexpressed dctA from a plasmid, which I
previously demonstrated would induce Salmonella uptake and growth using succinate as a sole
carbon source (Figure 29). I subjected strains constitutively expressing dctA or lacZ to acidified
(pH 4.4) LPM media containing 0.4% succinate as well as tested their survival in the human
THP-1 macrophage cell line. I found that overexpression of dctA but not lacZ led to decreased
survival relative to wild-type Salmonella under both conditions (Figure 33A-B).
At first this result was encouraging, as it suggested that forcing Salmonella to express
dctA and import succinate was detrimental to survival under acidic conditions, including in a
macrophage phagosome. However, overexpression of dctA has been demonstrated to be toxic to
E. coli and so there remained the possibility that the decreased survival was not due to succinate
import at all but rather due to cellular stress introduced by constitutive dctA expression150,167. To
address this concern I generated point mutants in dctA, N301A and S380D, which are defective
for succinate transport168. Despite their lack of transport function, these mutants still conveyed
toxicity when overexpressed and yielded smaller colonies than wild-type Salmonella (Figure
33C). The strain expressing pDctAS380D appeared to be more toxic than wild-type DctA and
yielded a mix of very small colonies or colonies that grew as well as wild-type Salmonella and
likely acquired a mutation in the plasmid’s dctA gene within one overnight growth cycle. In
contrast, the strain expressing pDctAN301A yielded similar sized colonies as the strain expressing
wild-type dctA, suggesting the N301A mutant as a valid candidate for further testing. When
subjected to acidified succinate media, survival of the strain overexpressing the dctA N301A
mutant was as low as the strain overexpressing wild-type dctA, implicating that the reduced
survival was not due to succinate uptake but rather an artefact of overexpression of the dctA gene
to toxic levels.
101
Figure 33: Overexpression of DctA from a plasmid decreases survival in macrophage due
to toxicity. A) Infection of THP-1 macrophage examining survival of wild-type Salmonella with
no plasmid (WT) compared to isogenic strains constitutively expressing LacZ (pLacZ) or DctA
(pDctA) from the pXG10sf plasmid. Survival of E. coli and ∆phoP mutant Salmonella are
included as controls for comparison. Data shows CFU recovered at 24 hours post-infection
relative to wild-type Salmonella and is the average of three biological replicates. Error bars show
one standard deviation. B) Survival of wild-type or dctA mutant Salmonella treated with LPM
media containing 0.4% succinate and acidified to pH 4.4. CFU recovered at 3 hours were
normalized to input CFU at 0h and expressed as percent survival. Data shows the average across
three biological replicates and error bars indicate one standard deviation. Where indicated, genes
were expressed from the ampicillin resistant version of the pXG10sf plasmid under the control of
the constitutively active PLtet0-1 promoter. C) Representative image showing overnight growth
of wild-type Salmonella on an LB agar plate comparing colony size when constitutively
expressing DctA point mutants. Black arrow indicates a small colony of the strain expressing
DctAS380D; white arrow indicates a large colony likely resulting from a suppressor mutation.
0.001
0.01
0.1
1
10
100
WT ∆dctA pLacZ ∆dctA pDctA ∆dctA pDctA N301A
Perc
ent
surv
ival
afte
r 3
ho
urs
pDctAS380D No plasmid
pDctAN301A pDctA
00.20.40.60.8
11.21.41.61.8
E coli K12 14028s WT ∆phoP WT pXG10sf-LacZ
WT pXG10sf-DctA
CFU
rec
ove
red
at
24
h
(rel
ativ
e to
WT)
E. coliK12
WT14028s
∆phoP WTpLacZ
WTpDctA
WT ∆dctApLacZ
∆dctApDctA
∆dctApDctAN301A
A)
B) C)
102
6.2.3 Chromosomal E. coli dctA promoter does not reduce Salmonella survival
Since I found that overexpression of dctA from a plasmid reduced Salmonella survival
due to an inherent toxicity artefact, the impact of succinate import in acidified media or in
macrophage remained uncertain. I therefore employed my PdctA swap strain wherein I replaced
the Salmonella chromosomal dctA promoter with that of E. coli. This strain does not repress dctA
and grows readily in succinate media (Figure 29), yet does not constitutively overexpress dctA
from a plasmid and so does not exhibit the associated toxic effects. Thus any reduced survival of
this strain would be due to increased succinate import rather than an artefact of toxicity. I
examined survival of this strain compared to a control with the native Salmonella dctA promoter
swapped into the chromosome by the same method. Underwhelmingly, the strain with the E. coli
dctA promoter showed no significant reduction in survival in acidified succinate or in either
human THP-1 or mouse J774 macrophage cell lines (Figure 34). Moreover, deletion of dctA or
iraP genes in Salmonella did not appear to significantly influence survival in THP-1 macrophage
(Figure 34B). Cumulatively, this data suggests that the Salmonella phenotype of repressing dctA
in minimal media with dicarboxylates as a sole carbon source does not convey a survival
advantage in acidified succinate or in macrophage.
103
Figure 34: Replacement of the Salmonella dctA promoter does not influence survival in
acidified succinate or macrophage despite growing precociously in succinate media as
demonstrated in Figure 29. A) Survival of Salmonella treated with LPM media containing 0.2%
succinate and acidified to pH 4.4. Wild-type (WT) is compared to Salmonella with its
chromosomal dctA promoter replaced with the E. coli dctA promoter (PdctA E. coli) or
Salmonella’s native dctA promoter as a control (PdctA Salm). An rpoS mutant is shown as a
positive control. CFU recovered at 3 hours were normalized to input CFU at 0h and expressed as
percent survival. Data shows the average across three biological replicates and error bars indicate
one standard deviation. B) Infection of THP-1 human macrophage comparing survival of wild-
type Salmonella (WT) with various mutant strains and the dctA promoter swap strain. A phoP
mutant is included as a positive control. Data shows a logarithm of CFU recovered at 24 hours
post infection and is the average of three biological replicates. Error bars show one standard
deviation. C) As in (B) but showing CFU recovered from mouse J774 macrophage.
4.5
5
5.5
6
6.5
Log
(CFU
/mL
reco
vere
d a
t 2
4h
)
Human THP-1 Macrophage
PdctA
(Salm)PdctA
(E. coli)WT ∆iraP ∆rpoS ∆dctA ∆rpoS
∆dctA∆phoP
0.001
0.01
0.1
1
10
100
Perc
ent
surv
ival
afte
r 3
ho
urs
LPM pH 4.4 + 0.2% Succinate
PdctA
(Salm)PdctA
(E. coli)WT ∆rpoS
5
6
7
8
Log
(CFU
/mL
reco
vere
d a
t 2
4h
)
Mouse J774 Macrophage
PdctA
(Salm)PdctA
(E. coli)WT ∆phoP
A)
B)
C)
104
6.2.4 Salmonella contains genes that can degrade itaconate
In addition to succinate, LPS-activated macrophage synthesize another dicarboxylate
called itaconate via the immune-responsive gene 1 (Irg1) protein160,162. Itaconate (methenylated
succinate) is similar in structure to succinate (Figure 32A) and has been demonstrated to inhibit
the bacterial glyoxylate shunt161,163. Recently, a cluster of Yersinia pestis genes were
demonstrated to encode proteins that can degrade itaconate in vitro164. This “itaconate
degradation operon” has orthologs in a number of bacteria, including Salmonella where it is
encoded by STM3120-STM3118 (LT2 nomenclature) (Figure 35). Moreover, these genes are
encoded in a Salmonella pathogenicity island (SPI-13), are amongst the most highly induced
genes of Salmonella during infection of macrophage, and were important for virulence of
Salmonella Gallinarum in an in vivo chick infection model169-171. In addition to STM3120-18,
there is another gene, STM3117, which appears to be encoded in the same operon. As well, the
adjacent gene, STM3121, is encoded in the reverse direction and is annotated as a predicted
lysR-type transcriptional regulator.
105
Figure 35: The itaconate degradation operon. Figure adapted from Pujol et al. (2005)172
comparing Yersinia and Salmonella operons. I have added in the operon from S. Typhimurium
for comparison. Sasikaran et al. (2014)164 demonstrated that these Yersinia proteins (STM3120-
3118 in Salmonella) and orthologs in Pseudomonas can break down itaconate in vitro.
Furthermore, the RNAseq analysis conducted by Srikumar et al. (2015)171 demonstrated a
massive upregulation (over 100 fold) of this operon in Salmonella infecting macrophage.
106
6.2.5 The itaconate degradation operon is induced by STM3121 in the presence of itaconate
To elucidate the role of the itaconate degradation operon in Salmonella and characterize
its induction, I examined the STM3120 promoter that drives expression of the operon. I
generated a plasmid-borne transcriptional fusion reporter of the 333bp upstream of the STM3120
start codon (P3120) driving expression of lacZ fused to sfGFP. Using this reporter I determined
that the STM3120 promoter exhibits very little basal expression but is strongly inducible by
itaconate (but not succinate) in various media (Figure 36A).
Since the adjacent gene, STM3121, is annotated as a putative lysR-type transcriptional
regulator, I examined its potential to regulate the expression of the itaconate degradation
operon’s STM3120 promoter. Using the same reporter system I found that itaconate induction of
P3120 was abolished in an STM3121 deletion mutant of Salmonella (Figure 36B). This
demonstrates that STM3121 is acting as an itaconate-sensing transcriptional activator that is
necessary for induction of the itaconate degradation operon. Furthermore, to test whether
STM3121 is sufficient for itaconate-induced expression, I generated an alternate version of the
transcriptional fusion plasmid that includes the STM3121 ORF encoded in the reverse direction
of the P3120-GFP fusion (pSTM3121-P3120). I found that inclusion of the STM3121 gene on the
plasmid was able to complement the Salmonella STM3121 mutant (Figure 36C). Furthermore,
the plasmid including STM3121 was able to exhibit itaconate-induced P3120-GFP expression in E.
coli K12, which lacks the itaconate degradation operon and regulator. Cumulatively, these data
demonstrate that the promoter of STM3120 is induced in the presence of itaconate, and the
transcriptional activator STM3121 is both necessary and sufficient for this induction.
107
Figure 36: The STM3120 promoter is inducible by itaconate and STM3121 is necessary
and sufficient for this induction. Plasmid-borne transcriptional fusion of P3120 driving
expression of sfGFP in wild-type (WT) or STM3121 knockout (∆STM3121) Salmonella. Figures
show GFP fluorescence normalized to optical density at 600nm (OD600) after 16h of growth in
the indicated media. Where indicated, glycerol, glucose, succinate or itaconate were added to
0.2% (w/v). Data are the average of at least three biological replicates. A) P3120 is inducible by
itaconate (but not succinate). B) STM3121 is necessary for itaconate induction of P3120. C)
STM3121 encoded on the same plasmid as the transcriptional fusion (pSTM3121-P3120) is
sufficient to complement the ΔSTM3121 strain and to allow itaconate induction in E. coli K12.
0
5000
10000
15000
20000
GFP
Flu
ore
scen
ce /
OD
60
0
No supplement
+ Itaconate
+ Succinate
MOPS Glycerol
MOPS Glucose
LB LPM pH 5.8
0
5000
10000
15000
GFP
Flu
ore
scen
ce /
OD
60
0
WT
∆STM3121
MOPS Glycerol
+ Itaconate
0
5000
10000
15000
20000
25000
GFP
Flu
ore
scen
ce /
OD
60
0
LB
LB + Itaconate
WT ∆STM3121 E. coli WT ∆STM3121 E. coli
pP3120 pSTM3121-P3120
A)
B)
C)
108
6.2.6 Deletion of the itaconate degradation operon does not reduce Salmonella survival
My discovery that Salmonella induces expression of the STM3120-17 operon in the
presence of itaconate ties together the findings in the literature that activated macrophage
accumulate itaconate and that Salmonella induces this operon when infecting
macrophage160,162,170,171. Indeed the itaconate-specific induction makes a lot of sense, as the
operon encodes homologs of genes shown to degrade itaconate164. Collectively, these data
suggest that Salmonella encounters itaconate in the phagosome of macrophage and that its
degradation contributes to bacterial survival in this environment. Since itaconate, like succinate,
is a dicarboxylic acid with the potential to shuttle protons across the bacterial membrane, I
examined survival of Salmonella in acidifed media containing itaconate. Though itaconate
reduced Salmonella survival compared to media without the weak acid (Figure 32C), mutants
lacking the itaconate degradation operon or its regulator showed no further reduction in survival
(Figure 37A). As well, the mutants also did not show reduced survival at 24 hours post-infection
in either human THP-1 or mouse J774 macrophage cell lines (Figure 37B and C). This data
suggests that, although the itaconate degradation operon is induced in macrophage and responds
to itaconate, it does not convey a survival advantage in acidified itaconate or macrophage cell
lines at the tested time points.
109
Figure 37: Deletion of the Salmonella itaconate degradation operon or its regulator does
not influence survival in acidified itaconate or macrophage. A) Indicated bacterial strains
were incubated in acidified LPM media containing 0.4% itaconate. CFU recovered at 3 hours
were normalized to CFU at 0h and expressed as percent survival. Data shows the average across
three biological replicates and error bars indicate one standard deviation. B) and C) Infection of
THP-1 (B) or J774 (C) macrophage comparing survival of wild-type Salmonella (WT) with
mutant strains lacking the itaconate degradation operon (∆STM3120-17) or its regulator
(∆STM3121). A phoP mutant is included as a control. Data shows a logarithm of CFU recovered
at 24 hours post infection and is the average of three biological replicates. Error bars show one
standard deviation.
0.0001
0.001
0.01
0.1
1
Perc
ent
surv
ival
afte
r 3
ho
urs
LPM pH 4.4 + 0.4% Itaconate
WT ∆STM3120-17 ∆STM3121
4.5
5
5.5
6
6.5
Log
(CFU
/mL
reco
vere
d a
t 2
4h
)
Human THP-1 Macrophage
∆STM3120-17∆STM3121WT ∆phoP
5
5.5
6
6.5
7
7.5
8
Log
(CFU
/mL
reco
vere
d a
t 2
4h
)
Mouse J774 Macrophage
∆STM3120-17∆STM3121WT ∆phoP
A)
B)
C)
110
6.2.7 The itaconate degradation operon is induced in J774 but not THP-1 macrophage
I found that deletion of the itaconate degradation operon had no influence on CFU
recovered at 24h from either human THP-1 or mouse J774 macrophage cell lines. A possible
explanation for this lack of phenotype is that the cell lines do not produce itaconate under the
growth conditions employed and thus do not induce the STM3120 promoter. To assess this
possibility I modified the P3120 transcriptional fusion plasmid to include a constitutively active
mCherry gene in an independent region of the plasmid from the P3120-GFP reporter. When used
for fluorescence microscopy, this new plasmid (pICM for independent constitutive mCherry)
illuminates all bacteria red while fluorescing green in response to induction of P3120 (likely by
itaconate based on my earlier findings). I examined fluorescence in wild-type or STM3121
mutant Salmonella at 8 hours post-infection of human THP-1 and mouse J774 macrophage cell
lines. I found that P3120 is strongly induced in J774 macrophage, suggesting that the cells are
capable of producing itaconate as well as introducing it into the Salmonella-containing vacuole
where it can be recognized by the bacteria (Figure 38). In agreement with my previous data, this
induction in macrophage is dependent on STM3121. Cumulatively this suggests that the absence
of a survival defect upon deletion of the itaconate degradation operon cannot be explained by
lack of itaconate production in J774 macrophage, implying that STM3120-17 are not required for
Salmonella survival in these cells at 24h. In contrast, the P3120 reporter did not fluoresce in THP-
1 cells under the tested conditions (Figure 38). This may reflect an inability of these cells to
produce itaconate or import it into the phagosome, or may suggest that additional stimulus
beyond LPS and bacterial infection is required to stimulate itaconate production in these cells.
111
Figure 38: The STM3120 promoter is induced in mouse but not human macrophage and
induction requires STM3121. A) Representative fluorescence microscopy images of macrophage
infected for 8 hours with wild-type (left two columns) or ΔSTM3121 (right column) Salmonella
containing pICM-PSTM3120. This plasmid expresses mCherry constitutively and sfGFP expression
is driven by the STM3120 promoter. Blue is DAPI stain for DNA. The 5µm scale bar in the
bottom right is applicable to all images. Three biological replicates have been conducted with
similar results. B) Quantification using ImageJ software of fluorescence intensity of greater than
125 bacteria per condition across three biological replicates. Data shows the GFP/mCherry
fluorescence ratio as a boxplot wherein the center line indicates the median, boxes indicate the
25-75th percentile, whiskers show the 5-95th percentile, and dots indicate individual quantified
bacteria beyond that range. An unpaired t-test with Welch’s correction indicated a p-value <
0.0001 comparing WT Salmonella in J774 macrophage compared to either of the other
conditions.
DAPI
mCherry(constitutively expressed
from pICM)
sfGFP(Expression driven by STM3120 promoter)
Merge
J774 (Mouse)∆STM3121 Salm.THP-1 (Human) J774 (Mouse)
J774WT Salm
J774∆STM3121
THP-1WT Salm.
A)
B)
112
6.3 Discussion
In this chapter I attempted to elucidate the evolutionary advantage conveyed by
Salmonella’s repression of dicarboxylate import. Since the phenotype is significantly less
prominent in E. coli, and LPS-activated macrophage accumulate the dicarboxylates succinate
and itaconate to high levels, I hypothesized that the selective pressure may involve Salmonella’s
adaptation to intracellular growth. Indeed, I demonstrated that these organic acids can exacerbate
acid stress in vitro and may contribute to acidification of phagocytosed bacteria within
macrophage. However, Salmonella with the E. coli dctA promoter (which grows readily in
succinate media), or Salmonella lacking the itaconate degradation operon or its regulator,
displayed no decrease in survival in acidified dicarboxylate media or in either human THP-1 or
mouse J774 macrophage cell lines. This result narrows the range of potential selective pressures
that could have contributed to the development of this phenotype in Salmonella and suggests
alternative hypotheses beyond survival within macrophage:
For succinate, overexpression of dctA from a plasmid was toxic to the cells and led to
reduced survival. It is conceivable that Salmonella represses this gene in order to eliminate any
toxic effects when the cell encounters stress. Though replacement of the dctA promoter with that
of E. coli yielded enough expression to instigate growth using succinate, it is possible that the
small degree of toxicity was masked by alternate cellular stress coping mechanisms that allow
the cell to maintain full survival. It is therefore possible that the repression of dctA may still
relate to reducing toxic expression due to stress, but only to a small degree that I could not detect
by the methods employed in this chapter. However, dctA is also toxic when overexpressed in E.
coli, and so the fact that E. coli does not instigate this repression to the same degree argues
against this hypothesis.
An alternate hypothesis is that Salmonella’s repression of dicarboxylate import may not
relate to survival per se, but may reflect a growth advantage relating to Salmonella’s induction of
inflammation. The inflammation induced by S. Typhimurium leads to an oxidative burst and the
oxidation of thiosulfate to tetrathionate, a compound that Salmonella is distinctively equipped to
use as a terminal electron acceptor39,40. Inflammation therefore provides Salmonella with the
potential to respire in the anaerobic gut and employ this ability to outcompete other microbiota
by utilizing alternate carbon sources such as ethanolamine that require the electron transport
113
chain to provide energy41. This important breakthrough demonstrated an evolutionary advantage
for why non-typhoidal Salmonella induces inflammation. This relates to dicarboxylates, since
succinate accumulation in activated macrophage was demonstrated to play a role in induction of
inflammation155. Thus, if Salmonella “wants” to induce inflammation, it would benefit from
maximizing succinate levels in macrophage by repressing succinate uptake and degradation as a
carbon source. Though none of the data presented here tests this theory, it will be the subject of
further research in the Navarre lab.
Finally, it also remains a possibility that Salmonella’s repression of dicarboxylate uptake
may not relate to succinate accumulation in activated macrophage but may reflect an advantage
in another extracellular environment.
Regarding itaconate, my findings that itaconate is produced in J774 macrophage yet
deletion of the Salmonella itaconate degradation operon yielded no reduction in survival could
suggest a number of possibilities. One is that the primary advantage of the itaconate degradation
operon may relate to some extracellular environment, although the strong induction of this
operon in macrophage strongly argues for an important role in this context. Indeed, my
identification of the specific induction of the STM3120-17 operon by itaconate likely explains
the lack of expression of this operon in a variety of tested growth conditions and the high degree
of induction in mouse macrophage170,171,173. Another hypothesis is that Salmonella possesses
redundant mechanisms to protect itself against itaconate in macrophage, such as exclusion of
itaconate uptake. Such a mechanism could not be induced by STM3121, since deletion of this
itaconate-responding transcriptional activator also conveyed no survival defect. It is possible that
Salmonella employs nutrient sources in macrophage that do not require the glyoxylate shunt and
so are not inhibited by itaconate. It follows that induction of the itaconate degradation operon
could allow Salmonella to supplement its metabolism by degrading itaconate to a useable carbon
source rather than to prevent itaconate poisoning. Finally, it is also possible that itaconate plays a
critical role in controlling bacterial growth during long term intracellular growth. Since I only
tested survival at 24 hour post-infection, any effects on longer timescales would not be identified
by my method.
Interestingly, using my itaconate biosensor I found that P3120 was induced strongly in
mouse J774 but not in human THP-1 macrophage. Though it remains possible that this reflects
114
an artefact of the cell line, my data is supported by the work of Michelucci et al (2013), who
demonstrated that, following LPS-activation, itaconate levels were approximately 125-fold lower
in human PBMC-derived macrophage than in mouse RAW264.7 macrophage160. Moreover, I
employed the Prokaryotic Genome Analysis Tool (PGAT)112 to examine conservation of the
itaconate operon within Salmonella enterica and found that all strains in the database from
Salmonella serovars Typhi and Paratyphi A exclusively lack this operon (Figure 39).
Importantly, these serovars are typhoidal and specifically human-adapted24. In contrast, other
typhoidal but not human-specific serovars such as Paratyphi B and C still retained the itaconate
degradation operon. Interestingly, a non-typhoidal and not human-specific serovar, S. Agona,
also appears to lack the itaconate degradation operon. Though there is only one strain of S.
Agona in the PGAT database, I manually searched another S. Agona complete genome on
GenBank (CP006876.1) and found a similar lack of the operon174. How S. Agona (the only non-
human-specific serovar lacking the itaconate degradation operon in the PGAT database) is
affected by itaconate during infections remains an unanswered question.
Cumulatively, these data suggest that at least in response to Salmonella infection, human
macrophage may be less able to synthesize itaconate than other species. If verified, this
fundamental difference in the innate immune system of humans compared to model organisms
such as mice must be considered when investigating the innate immune response to pathogens. It
is also possible that human macrophage require additional inflammatory cues beyond LPS – such
as interferon gamma (IFN-γ) – to fully induce itaconate synthesis. Ultimately, if the human
immune system does not employ itaconate and human-adapted pathogens do not encode
defences against it, the inhibitory properties of itaconate could propose it as a potential novel
antimicrobial compound in humans.
115
15 17 18 19 20 21 22 23 24
S. enterica 4 5 12 i 08-1736
S. enterica Agona SL483
S. enterica arizonae 62:z4,z23:--
S. enterica Choleraesuis A50
S. enterica Choleraesuis SC-B67
S. enterica Dublin CT_02021853
S. enterica Dublin SD3246
S. enterica Enteritidis P125109
S. enterica Gallinarum 287/91
S. enterica Gallinarum SG9
S. enterica Gallinarum/pullorum RKS5078
S. enterica Heidelberg B182
S. enterica Heidelberg SL476
S. enterica Newport SL254
S. enterica Paratyphi A AKU_12601
S. enterica Paratyphi A ATCC 9150
S. enterica Paratyphi B SPB7
S. enterica Paratyphi C RKS4594
S. enterica Schwarzengrund CVM19633
S. enterica Typhi CT18
S. enterica Typhi P-stx-12
S. enterica Typhi Ty2
S. enterica Typhimurium 14028S
S. enterica Typhimurium 415DRC
S. enterica Typhimurium 798
S. enterica Typhimurium A130
S. enterica Typhimurium BC_2557
S. enterica Typhimurium BC_2558
S. enterica Typhimurium BC_2559
S. enterica Typhimurium BC_2560
S. enterica Typhimurium BC_2561
S. enterica Typhimurium BC_2562
S. enterica Typhimurium BC_2563
S. enterica Typhimurium BC_2564
S. enterica Typhimurium BC_2565
S. enterica Typhimurium BC_2566
S. enterica Typhimurium BC_2567
S. enterica Typhimurium BC_2568
S. enterica Typhimurium C13184
S. enterica Typhimurium D23580
S. enterica Typhimurium LT2
S. enterica Typhimurium LT7
S. enterica Typhimurium M1175849
S. enterica Typhimurium M1776464
S. enterica Typhimurium PB1
S. enterica Typhimurium PB2
S. enterica Typhimurium PB3
S. enterica Typhimurium SARA10
S. enterica Typhimurium SARA12
S. enterica Typhimurium SARA4
S. enterica Typhimurium SARA9
S. enterica Typhimurium SL1344
S. enterica Typhimurium SOHS02-20
S. enterica Typhimurium SOHS02-68
S. enterica Typhimurium SOHS03-1
S. enterica Typhimurium SOHS04-44
S. enterica Typhimurium ST13
S. enterica Typhimurium ST14
S. enterica Typhimurium ST15
S. enterica Typhimurium ST16
S. enterica Typhimurium ST17
S. enterica Typhimurium ST18
S. enterica Typhimurium ST19
S. enterica Typhimurium ST24
S. enterica Typhimurium ST25
S. enterica Typhimurium ST29
S. enterica Typhimurium ST32
S. enterica Typhimurium ST33
S. enterica Typhimurium ST34
S. enterica Typhimurium ST35
S. enterica Typhimurium ST36
S. enterica Typhimurium ST4/74
S. enterica Typhimurium ST40
S. enterica Typhimurium T000240
S. enterica Typhimurium UK-1
STM31__A)
116
B)
Figure 39: Salmonella serovars Typhi, Paratyphi A and Agona uniquely lack the itaconate
degradation operon. A) Comparison of the presence (green) or absence (red) of the itaconate
degradation operon (STM3120-STM3117), its regulator (STM3121), and the neighbouring genes
in the 69 Salmonella strains in the Prokaryotic Genome Analysis Tool (PGAT) database using
the Ortholog Search Tool112. Numbers at the top indicate gene loci (S. Typhimurium LT2
nomenclature) from STM3115 to STM3124. STM3116 is pheV encoding tRNA-Phe (GAA
codon) rather than a protein and so was not analyzed by PGAT. B) The itaconate degradation
operon region comparing S. Typhimurium to S. Typhi and S. Agona serovars using PGAT’s
Synteny Mapper tool. Asterisk indicates STM3116 encoding tRNA-Phe. P indicates a
pseudogene was identified by the software.
117
7 Discussion and Future Directions
7.1 Summary and Discussion
At the start of my graduate research little was known about the mechanism of EF-P
except that it had been shown to modestly stimulate first peptide bond synthesis in vitro and was
demonstrated to bind to the ribosome between the P and E sites77,79. As well, efp mutants exhibit
pleiotropic phenotypes ranging from overactive metabolism using specific nutrient sources to
hyper-susceptibility to a variety of different cellular stressors74,84. I contributed to this work in
the first months of my graduate research by demonstrating that efp mutants also display a
permeability defect, which could contribute to some of the phenotypes. Yet the mechanism of
how EF-P influenced translation remained unknown.
In this thesis I employed SILAC to demonstrate in vivo that EF-P is not a general
translation factor but rather impacts the synthesis of a specific subset of proteins in Salmonella.
This was in agreement with previously conducted 2D-DIGE analyses conducted by a former
graduate student in the lab, Dr. Betty Zou74,84. Shortly after the completion of my SILAC
experiment, two papers were published in Science that demonstrated that the primary role of EF-
P is to rescue ribosomes stalled at polyproline motifs114,115. Specifically, the motifs PPP and PPG
were heavily dependent on EF-P for their efficient translation in vitro. I compared the presence
of these motifs to my in vivo SILAC data and discovered that polyproline motifs are not always
necessary and are not solely sufficient to render a protein dependent on EF-P for its efficient
expression. I further went on to identify a novel EF-P-dependent motif, GSCGPG, in PoxB, and
employed the model proteins AtpA and AtpD to demonstrate that the residues upstream of
polyproline motifs play a significant role in influencing ribosome stalling. Moreover, I found that
the rate of translation initiation strongly effects whether or not a polyproline motif will reduce
protein synthesis in the absence of EF-P. I rationalized that this reflects the rate limiting step of
translation and that ribosome stalls will only impact protein synthesis if they become more rate
limiting than translation initiation. Thus, a transcript with a low rate of initiation is protected
against ribosome stalls influencing protein output, since initiation remains the rate limiting step.
This innovation provides a parsimonious explanation for the differences in protein abundance
observed for proteins with identical polyproline motifs.
118
Collectively, my work on EF-P demonstrates that the impact of EF-P on protein synthesis
is determined by the stall strength (influenced not only by a polyproline motif but also by
upstream residues) and, moreover, by the rate of translation initiation. This overarching concept
is significant for understanding EF-P-dependent translation in vivo and is furthermore applicable
to other kinds of ribosome stalls as well, such as at rare codons. Indeed, bacteria may employ
weak ribosome binding sites to reduce translation initiation rate in order to protect against
downstream ribosome stalling. This would allow for functional stalls, such as for N-terminal
sequence recognition or for regulation of downstream genes, without costing the cell resources
due to ribosome jamming reducing translation efficiency.
In addition to my work elucidating the mechanism of how EF-P influences translation, I
also investigated how deletion of efp effects Salmonella. Many of the EF-P-dependent genes
identified by my SILAC analysis can provide an explanation for certain efp mutant phenotypes.
For instance, levels of GgT were strongly reduced in the efp mutant, and this could explain the
inability of PYE mutants to use γ-glutamyl-glycine as a nitrogen source and their resistance to
the nitrogen donating comound S-Nitrosoglutathione (GSNO)74,84,110. Altered expression of a
number of metabolic proteins, including the catalytic subunit of ATP Synthase, AtpD, may lead
to the increased metabolic activity of the PYE mutants and potentially contribute to their hyper-
susceptibility to stress74,84.
While investigating the hyperactive metabolism of efp mutants, I discovered that wild-
type Salmonella exhibits a drastically long lag phase in response to media with a dicarboxylic
acid as the sole carbon source. I determined that Salmonella’s growth shutdown was instigated
via repression of the primary dicarboxylate importer, DctA, and required the stress response
sigma factor, RpoS, and an RpoS-stabilizer, IraP. Furthermore, I found that the closely related
species E. coli does not exhibit this phenotype and that replacing the dctA promoter with that of
E. coli abolishes the ability of Salmonella to repress its growth. This demonstrates that under
these conditions Salmonella employs a transcriptional repressor to bind to a region of the dctA
promoter that is different between Salmonella and E. coli. This is significant, as it implies that
Salmonella has developed this adaptation since diverging from E. coli. Many such differences
pertain to Salmonella’s virulent lifestyle, including acquisition of the Salmonella pathogenicity
islands (SPI) and regulatory alterations such as implementing the PhoPQ two-component system
to recognize the intracellular environment encountered by Salmonella19,32,33. Thus, the difference
119
between Salmonella and E. coli may suggest that repression of dicarboxylate uptake may reflect
an adaptation for virulence.
The potential for the dicarboxylate exclusion phenotype to play a role in virulence is
supported by the finding that activated macrophage accumulate succinate as well as another
dicarboxylate, itaconate155,160,162. Acidification of the Salmonella containing vacuole and the
pKa’s of these dicarboxylic acids suggested the hypothesis that macrophage may employ these
weak organic acids to shuttle protons into phagocytosed bacteria – forcing them to acidify.
However, when I abolished Salmonella’s ability to repress dctA by swapping in the E. coli
promoter, I found that the bacteria showed no reduction in survival in acidified succinate media
or in macrophage. This suggests that the evolutionary advantage of the Salmonella dicarboxylate
phenotype is either relevant at other time points beyond 24 hour post macrophage infection,
grants only a modest survival advantage below the detectable limits of my methods, or does not
involve survival in the macrophage phagosome. It is also possible that dicarboxylate repression
is useful to Salmonella in extracellular environments for survival or metabolic prioritization, and
the accumulation of succinate in activated macrophage is an unrelated coincidence.
An alternate hypothesis is that Salmonella does not import and degrade succinate in
macrophage in order to maximize succinate levels leading to HIF-1α stabilization, IL-1β
synthesis and the induction of inflammation155. This would not grant Salmonella an advantage in
survival in macrophage per se, but would give the Salmonella remaining in the gut lumen a
growth advantage by maximizing the abundance of tetrathionate produced by the immune
system’s oxidative burst39-41. If verified, such a result would link the growth repression that I
identified for Salmonella using dicarboxylates as a carbon source to the virulent lifestyle specific
to Salmonella. Furthermore, it would represent a novel mechanism exemplifying a bacterium
modulating the host immune response by manipulating metabolite concentration rather than
employing injected effector proteins. Cumulatively, my work examining Salmonella’s repression
of growth using the dicarboxylic acid succinate makes significant progress towards
characterizing this extended lag phase phenotype. My findings lend insight into the mechanism
underlying this response and further narrow the scope of why Salmonella has developed this
adaptation and what evolutionary advantage it gains from it.
120
In addition to its response to other dicarboxylates such as succinate, Salmonella also
appears to be able to neutralize the inhibitory effects of itaconate by encoding an itaconate
degradation operon (STM3120-17) in SPI-13161,163,164. These genes are heavily expressed in
mouse macrophage but not in a variety of other tested conditions170,171,173. I demonstrated that
this operon is induced in response to the presence of itaconate in the media and, moreover, that
the neighbouring gene STM3121 is both necessary and sufficient for this induction. However,
upon deletion of the itaconate degradation operon or its regulator, I found no reduction in
Salmonella survival in acidified itaconate media or in macrophage. Since the operon is strongly
induced in mouse macrophage, it is unlikely that it conveys no advantage to Salmonella in this
environment. Therefore, my findings that the operon conveyed no survival advantage at 24 hours
suggests that it may be important for longer infections, may be only one of redundant itaconate
defence mechanisms, or may not be required for survival per se, but may allow Salmonella to use
itaconate in macrophage as a supplementary carbon source.
To further examine itaconate synthesis in macrophage, I employed the itaconate
degradation operon promoter (P3120) to generate a fluorescence-based itaconate biosensor.
Interestingly, I found that in response to Salmonella infection, itaconate was produced strongly
in mouse J774 macrophage but not in the human THP-1 macrophage cell line. This reflects the
work of Michelucci et al. (2013), who found that itaconate synthesis in human PBMC-derived
macrophage was approximately 125-fold lower than in mouse RAW264.7 macrophage160. By
genome comparison I also found that the itaconate degradation operon is present in most
Salmonella strains but is distinctively absent in the human-specific and typhoidal serovars S.
Typhi and S. Paratyphi A, as well as in serovar Agona. The absence of the itaconate degradation
operon in S. Agona is an interesting outlier, since it is non-typhoidal and also not human specific.
Even if itaconate does not greatly accumulate in human cells, this serovar may encounter it in
other species. It is possible that S. Agona includes alternate itaconate defence mechanisms that
may also be present in S. Typhimurium and could account for the lack of survival defect that I
observed upon deletion of STM3120-17. Cumulatively, the lower itaconate levels in human cells
in my data and in the literature, combined with the lack of the itaconate degradation operon in all
human-specific strains in my analysis, may suggest that the human immune system either
requires strong pro-inflammatory stimulus beyond just LPS (such as by cytokines like IFN-γ) or
does not synthesize itaconate to the same degree as other species including mice.
121
If verified, the difference in itaconate synthesis between humans and mice may highlight
important limitations of conducting mouse studies of bacterial pathogenesis. On the other hand,
the lack of defence mechanisms in human-adapted microbes may render them susceptible to
itaconate as an antimicrobial compound. Moreover, its generation by the immune system to some
degree may suggest that it is not toxic to human cells, and it is also readily and cheaply available,
as it has been used in industrial polymer synthesis for decades175. In summary, my findings
demonstrate that Salmonella Typhimurium employs an itaconate-sensing transcriptional activator
to induce the expression of an itaconate degradation operon in SPI-13. Though deletion of this
operon had no effect on survival in macrophage at 24 hours, it may convey a survival advantage
in mouse macrophage at later timepoints or indicate redundant defence mechanisms. Finally, my
finding that human THP-1 macrophage do not induce expression from P3120 under the conditions
tested, and my discovery that human-adapted strains of Salmonella do not encode the itaconate
degradation operon, may propose itaconate as a potential antimicrobial compound that could be
employed against such strains.
7.2 Future Directions
7.2.1 Regulation of EF-P and its potential regulatory activity
My data and work in the literature demonstrate that EF-P is involved in the efficient
translation of a specific subset of proteins, including some involved in central
metabolism110,127,130. As well, Salmonella and E. coli efp mutants are viable but display a variety
of pleiotropic phenotypes, including impaired growth, compared to wild-type cells74,83,84. These
data suggest that efp expression is concomitant with growth, similar to ribosomal genes and other
general translation factors. Interestingly, high throughput analyses in the literature suggest that
efp expression may be governed by the stringent response as well as by catabolite
repression176,177. To demonstrate regulation of efp by these pathways in a targeted manner,
stringent response (∆relA ∆spoT) and catabolite repression (∆crp) mutants could be employed to
assess protein (western blot) and mRNA (RT-qPCR) levels of EF-P, PoxA, YjeK and YfcM
under conditions of amino acid starvation or when growing on alternative carbon sources.
Results would demonstrate whether the expression of efp, like many genes involved in
translation, is linked to cellular metabolism and growth rate.
122
It also remains possible that EF-P itself plays a regulatory role under certain conditions
wherein its activity is reduced to alter the synthesis of specific proteins. Conceivably, storing
stalled ribosomes on particular transcripts could also allow for rapid translation of these proteins
following a change of environment and recovery of EF-P activity. To examine whether EF-P
activity is altered, my EF-P-dependent translational fusions could be employed. Specifically,
highly EF-P-dependent constructs (with a high translation initiation rate) and their mutated
polyproline motif counterparts (EF-P-independent) could be subjected to a variety of growth
conditions. EF-P activity could then be assessed in a high-throughput manner by comparing
expression of the independent versus EF-P-dependent constructs. If the independent construct
retained high expression while the dependent construct showed reduced fluorescence, it would
suggest that EF-P activity is reduced under the tested condition in a manner distinct from general
translation. As a follow-up to this high throughput screen, EF-P, PoxA, YjeK and YfcM protein
and mRNA levels could be assessed by western blot or RT-qPCR. Results from these
experiments would lend insight into how bacteria regulate the expression of efp and,
furthermore, whether EF-P itself plays a regulatory role in certain environments.
7.2.2 Characterize the repression of the Salmonella dctA promoter
In this thesis I demonstrated that RpoS and DctA are involved in Salmonella’s growth
repression in minimal media with succinate as the sole carbon source. Specifically, deletion of
rpoS or replacement of the dctA promoter with that of E. coli reduced the length of the
Salmonella lag phase. Since RpoS is primarily an activator of transcription, this suggests that it
induces an intermediate transcriptional repressor that binds to and shuts down the Salmonella
dctA promoter. To confirm that the effects of RpoS and the dctA promoter are linked, the rpoS
mutation could be introduced into the strain with E. coli’s dctA promoter and additive effects
could be assessed. If the reduction in Salmonella lag time is not additive, it would suggest that
RpoS and the repression of DctA act in the same pathway. Furthermore, the impact of RpoS on
dctA expression could be assessed by examining dctA mRNA levels under conditions of stress
that induce RpoS activity but do not relate to dicarboxylic acids. Examples could include
treatment with hydrogen peroxide or high osmolarity. Results could demonstrate whether or not
123
general induction of RpoS activity is sufficient to repress dctA expression or if there is an
additional dicarboxylate-specific input.
My finding that the E. coli dctA promoter is active in succinate media implies that there is
a transcriptional repressor that binds to the dctA promoter of Salmonella but not of E. coli. The
sequence differences between these promoters could be used to identify potential candidate
regulators by examining databases such as PRODORIC and TEC to find transcription factor
consensus binding sites present in the Salmonella but not the E. coli dctA promoter178,179. An
alternative approach could involve biotinylating the dctA promoter using chemically generated
PCR primers such that it could be pulled down using streptavidin beads. Thus the Salmonella
dctA promoter (or the E. coli one as a control) could be used as bait and combined with cell
lysates generated under growing or growth repressed conditions to pull down DNA-binding
proteins for subsequent mass spectrometry analysis and identification. Findings from these
analyses could be followed-up by generating deletion mutants of candidate regulators and
assessing growth in MOPS minimal medium with succinate as the sole carbon source. The
regulation of these transcription factors by RpoS could also be assessed using RT-qPCR under
RpoS-inducing stress conditions and with the rpoS deletion mutant.
An alternative screen could also be conducted to identify regulators of dctA by
conducting Tn-Seq or mutagenesis followed by genome sequencing. Specifically a mutant
library could be grown in minimal media using succinate as a sole carbon source to select for
mutants that grow readily. To deplete the library of rpoS mutants, the culture could be further
treated with hydrogen peroxide. Surviving cells could be plated for single colonies and
individual clones sequenced to identify the mutation that conveyed growth using succinate.
Results could be confirmed by targeted deletions introduced by lambda red recombination.
Cumulatively, findings from these experiments could elucidate the regulatory pathway that
Salmonella employs to repress dctA.
7.2.3 Elucidate the role of Salmonella dctA repression in macrophage,
Using a strain with the E. coli dctA promoter inserted into the Salmonella chromosome I
demonstrated that the repression of dctA in minimal media with a dicarboxylate as a sole carbon
124
source does not relate to Salmonella survival in macrophage at 24h. However, Salmonella may
still be responding to succinate in macrophage. Macrophage activated by LPS have been
demonstrated to accumulate succinate as an inflammatory signal acting by stabilization HIF-1α
and resulting in IL-1β production155. Since it has been shown that Salmonella serovars that
induce gastroenteritis instigate inflammation in order to acquire the terminal electron acceptor
tetrathionate, it is possible that Salmonella represses its uptake of succinate to maximize
succinate concentration in macrophage and promote inflammation40. To assess this, macrophage
could be infected with Salmonella containing either its native or the E. coli dctA promoter, and
induction of IL-1β could be quantified by RT-qPCR or by enzyme-linked immunosorbent assay
(ELISA). If IL-1β synthesis is reduced in macrophage infected with E. coli dctA promoter
Salmonella, it would suggest that Salmonella uptake and degradation of succinate is detrimental
to inducing inflammation. If verified, this could represent a rationale for the observed phenotype
in succinate media and an important novel aspect of Salmonella’s virulent lifestyle.
7.2.4 Investigate alternate itaconate defence mechanisms
My finding that deletion of the itaconate degradation operon had little effect on survival
of Salmonella in macrophage could suggest that the bacterium encodes alternate mechanisms to
defend against itaconate. Since itaconate would have to penetrate the bacterial cytoplasm in order
to inhibit the glyoxylate shunt, it is conceivable that Salmonella restricts its uptake of itaconate
similarly to my findings in minimal media with other dicarboxylates. Similarly, the use of
itaconate as a carbon source (via degradation by STM3120-17) could be assessed for various
mutants lacking transporters such as DctA. Uptake of itaconate could also be assessed by
examining induction of P3120 in these transporter mutants. If itaconate required dctA (for
instance) to bypass the bacterial membrane, deletion or repression of the transporter would
restrict itaconate uptake and reduce the induction of P3120 via STM3121. A potential caveat of
this induction assay is that it assumes STM3121 recognizes cytoplasmic itaconate to induce its
transcriptional activation effects. STM3121 is predicted to be a cytosolic lysR-type regulator but
it remains possible that it could acquire an activation signal from a transmembrane sensor. If this
were the case, such a sensor would have to be common between Salmonella and E. coli, as
STM3121 was sufficient to allow itaconate-induced expression of P3120 in E. coli.
125
My genomic analysis of Salmonella serovars revealed that the human adapted serovars
Typhi and Paratyphi A do not encode the itaconate degradation operon. A non-typhoidal serovar,
S. Agona, also lacks the itaconate degradation operon, and its SPI-13 resembles that of S. Typhi
and Paratyphi A. The degree of itaconate inhibition of these serovars could be assessed in
minimal media where the only carbon source requires the glyoxylate shunt (for example,
acetate). This assay could determine whether these strains contain alternate mechanisms for
preventing itaconate toxicity or if they are inhibited by itaconate, suggesting that they have lost
the operon because they do not encounter it in their particular growth niches. This may relate to
the reduced inflammation induced by typhoidal Salmonella, or to the reduced itaconate
production observed in human macrophage. If these strains lack any defence against itaconate, it
could propose the use of itaconate as an antimicrobial agent against them.
7.2.5 Scrutinize the induction of itaconate synthesis in human and mouse macrophage
My itaconate biosensor could be employed to further examine how and when
macrophage produce itaconate. The biosensor is fluorescence-based and lends itself readily to
fluorescence microscopy. The dynamics of itaconate synthesis in macrophage could be observed
using live microscopy to reveal the time when infecting Salmonella first encounter itaconate.
Moreover, by comparison to standard curves, the levels of itaconate encountered by Salmonella
in a macrophage phagosome could be semi-quantitatively determined. Itaconate concentration
could be verified using liquid chromatography coupled to mass spectrometry. The biosensor
could also rapidly probe for itaconate synthesis to examine whether it is specific to phagocytes or
can be produced in other cells types as well, such as epithelial cells. This could be extended to
multiple human cells lines and donor-derived macrophage to scrutinize whether the reduced
itaconate production in THP-1 macrophage represents a bona fide phenomenon of human cells. It
is also possible that human cells require additional inflammatory cues to trigger itaconate
production, and my biosensor could be employed to identify such inducers. Various cytokines
such as IFN-γ could be used to stimulate macrophage infected with Salmonella containing the
biosensor to test for itaconate synthesis. Finally, since Salmonella may contain alternative
mechanisms of excluding itaconate and thereby negating the requirement for induction of P3120 in
human cells, the experiment could also be conducted using E. coli with the pSTM3121-P3120
126
version of the itaconate biosensor. Results from these experiments would lend insight into the
mechanisms of itaconate utilization by the immune system and, moreover, could reveal an
important difference between human and mouse phagocytes.
7.3 Conclusions
My work determined that EF-P will only impact protein synthesis if ribosome stalls at
polyproline motifs become more rate limiting than translation initiation. This is influenced
primarily by the ribosome binding site and thereby the rate of initiation of translation, and by the
strength of the EF-P dependent stall itself, which is determined by the polyproline motif and
nearby upstream residues. This concept not only clarifies the factors underlying whether a
protein will or will not require EF-P for its efficient translation in vivo, but also can be applied to
other forms of ribosome stalls. Moreover, it proposes a rationale for the evolution of inefficient
ribosome binding sites and poor initiation rates, as they can play a protective role against
downstream ribosome stalling impacting protein synthesis.
In addition, I also discovered a novel Salmonella phenotype wherein it shuts down its
growth for an extended lag phase when using dicarboxylic acids as a sole carbon source. This
repression of dicarboxylate uptake requires a number of stress response regulators and may relate
to the accumulation of succinate and itaconate in activated macrophage. When I abolished
Salmonella’s ability to shut down dicarboxylate uptake or when I deleted the itaconate
degradation operon, my data showed no reduction in survival in acidified dicarboxylic acids or in
macrophage at 24 hours post infection. The advantage conveyed by Salmonella’s dicarboxylate
exclusion phenotype may relate to survival in other conditions, or to modulating the host
inflammatory response for the acquisition of supplementary nutrients such as tetrathionate.
Finally, my findings employing an itaconate biosensor in THP-1 macrophage, combined with
previous findings in the literature, suggest that human cells might not produce itaconate as
robustly as mouse cells. I discovered that human adapted serovars of Salmonella lack the
itaconate degradation operon and this may render them susceptible to the use of itaconate as an
antimicrobial compound. In conclusion, my findings identify a novel Salmonella phenotype and
narrow down the potential evolutionary advantages that led to its development.
127
References
1. Shimizu K. Metabolic regulation and coordination of the metabolism in bacteria in response to
a variety of growth conditions. Adv Biochem Eng Biotechnol. 2016;155:1-54.
2. Görke B, Stülke J. Carbon catabolite repression in bacteria: Many ways to make the most out
of nutrients. Nat Rev Microbiol. 2008;6(8):613-624.
3. Zheng D, Constantinidou C, Hobman JL, Minchin SD. Identification of the CRP regulon using
in vitro and in vivo transcriptional profiling. Nucleic Acids Res. 2004;32(19):5874-5893.
4. Buettner MJ, Spitz E, Rickenberg HV. Cyclic adenosine 3',5'-monophosphate in Escherichia
coli. J Bacteriol. 1973;114(3):1068-1073.
5. You C, Okano H, Hui S, et al. Coordination of bacterial proteome with metabolism by cyclic
AMP signalling. Nature. 2013;500(7462):301-306.
6. Pastan I, Perlman R. Cyclic adenosine monophosphate in bacteria. Science.
1970;169(3943):339-344.
7. Epstein W, Rothman-Denes L, Hesse J. Adenosine 3':5'-cyclic monophosphate as mediator of
catabolite repression in Escherichia coli. Proc Natl Acad Sci U S A. 1975;72(6):2300-2304.
8. Ramseier TM. Cra and the control of carbon flux via metabolic pathways. Res Microbiol.
1996;147(6-7):489-493.
9. Saier MH,Jr. Cyclic AMP-independent catabolite repression in bacteria. FEMS Microbiol Lett.
1996;138(2-3):97-103.
10. Kochanowski K, Volkmer B, Gerosa L, Haverkorn van Rijsewijk BR, Schmidt A,
Heinemann M. Functioning of a metabolic flux sensor in Escherichia coli. Proc Natl Acad Sci U
S A. 2013;110(3):1130-1135.
11. Karp PD, Weaver D, Paley S, et al. The EcoCyc database. EcoSal Plus.
2014;6(1):10.1128/ecosalplus.ESP-0009-2013.
12. Unden G, Bongaerts J. Alternative respiratory pathways of escherichia coli: Energetics and
transcriptional regulation in response to electron acceptors. Biochim Biophys Acta.
1997;1320(3):217-234.
13. Boutte CC, Crosson S. Bacterial lifestyle shapes stringent response activation. Trends
Microbiol. 2013;21(4):174-180.
14. Magnusson LU, Farewell A, Nyström T. ppGpp: A global regulator in Escherichia coli.
Trends Microbiol. 2005;13(5):236-242.
128
15. Dalebroux ZD, Svensson SL, Gaynor EC, Swanson MS. ppGpp conjures bacterial virulence.
Microbiol Mol Biol Rev. 2010;74(2):171-199.
16. Potrykus K, Cashel M. (P)ppGpp: Still magical? Annu Rev Microbiol. 2008;62:35-51.
17. Battesti A, Majdalani N, Gottesman S. The RpoS-mediated general stress response in
Escherichia coli. Annu Rev Microbiol. 2011;65(1):189-213.
18. Dong T, Schellhorn HE. Role of RpoS in virulence of pathogens. Infect Immun.
2010;78(3):887-897.
19. Meysman P, Sánchez-Rodríguez A, Fu Q, Marchal K, Engelen K. Expression divergence
between Escherichia coli and Salmonella enterica serovar Typhimurium reflects their lifestyles.
Mol Biol Evol. 2013;30(6):1302-1314.
20. Porwollik S, Wong RM, McClelland M. Evolutionary genomics of Salmonella: Gene
acquisitions revealed by microarray analysis. Proceedings of the National Academy of Sciences.
2002;99(13):8956-8961.
21. Sabbagh SC, Forest CG, Lepage C, Leclerc JM, Daigle F. So similar, yet so different:
Uncovering distinctive features in the genomes of Salmonella enterica serovars Typhimurium
and typhi. FEMS Microbiol Lett. 2010;305(1):1-13.
22. LaRock DL, Chaudhary A, Miller SI. Salmonellae interactions with host processes. Nat Rev
Microbiol. 2015;13(4):191-205.
23. McClelland M, Sanderson KE, Clifton SW, et al. Comparison of genome degradation in
Paratyphi A and Typhi, human-restricted serovars of Salmonella enterica that cause typhoid. Nat
Genet. 2004;36(12):1268-1274.
24. Tracz DM, Tabor H, Jerome M, Ng LK, Gilmour MW. Genetic determinants and
polymorphisms specific for human-adapted serovars of Salmonella enterica that cause enteric
fever. J Clin Microbiol. 2006;44(6):2007-2018.
25. Majowicz SE, Musto J, Scallan E, et al. The global burden of nontyphoidal Salmonella
gastroenteritis. Clin Infect Dis. 2010;50(6):882-889.
26. Crump JA, Luby SP, Mintz ED. The global burden of typhoid fever. Bull World Health
Organ. 2004;82(5):346-353.
27. Bäumler AJ, Tsolis RM, Ficht TA, Adams LG. Evolution of host adaptation in Salmonella
enterica. Infect Immun. 1998;66(10):4579-4587.
28. van den Beld MJ, Reubsaet FA. Differentiation between shigella, enteroinvasive Escherichia
coli (EIEC) and noninvasive Escherichia coli. Eur J Clin Microbiol Infect Dis. 2012;31(6):899-
904.
129
29. Hansen-Wester I, Hensel M. Salmonella pathogenicity islands encoding type III secretion
systems. Microbes Infect. 2001;3(7):549-559.
30. Fields PI, Swanson RV, Haidaris CG, Heffron F. Mutants of Salmonella Typhimurium that
cannot survive within the macrophage are avirulent. Proc Natl Acad Sci U S A.
1986;83(14):5189-5193.
31. Bäumler AJ, Kusters JG, Stojiljkovic I, Heffron F. Salmonella Typhimurium loci involved in
survival within macrophages. Infect Immun. 1994;62(5):1623-1630.
32. Miller SI, Kukral AM, Mekalanos JJ. A two-component regulatory system (phoP phoQ)
controls Salmonella Typhimurium virulence. Proc Natl Acad Sci U S A. 1989;86(13):5054-5058.
33. Monsieurs P, De Keersmaecker S, Navarre WW, et al. Comparison of the PhoPQ regulon in
Escherichia coli and Salmonella Typhimurium. J Mol Evol. 2005;60(4):462-474.
34. Alpuche Aranda,C.M., Swanson JA, Loomis WP, Miller SI. Salmonella Typhimurium
activates virulence gene transcription within acidified macrophage phagosomes. Proc Natl Acad
Sci U S A. 1992;89(21):10079-10083.
35. Richards SM, Strandberg KL, Conroy M, Gunn JS. Cationic antimicrobial peptides serve as
activation signals for the Salmonella Typhimurium PhoPQ and PmrAB regulons in vitro and in
vivo. Front Cell Infect Microbiol. 2012;2:102.
36. Navarre WW, Halsey TA, Walthers D, et al. Co-regulation of Salmonella enterica genes
required for virulence and resistance to antimicrobial peptides by SlyA and PhoP/PhoQ. Mol
Microbiol. 2005;56(2):492-508.
37. Dalebroux ZD, Miller SI. Salmonellae PhoPQ regulation of the outer membrane to resist
innate immunity. Curr Opin Microbiol. 2014;17:106-113.
38. Tu X, Latifi T, Bougdour A, Gottesman S, Groisman EA. The PhoP/PhoQ two-component
system stabilizes the alternative sigma factor RpoS in Salmonella enterica. Proc Natl Acad Sci U
S A. 2006;103(36):13503-13508.
39. Hensel M, Hinsley AP, Nikolaus T, Sawers G, Berks BC. The genetic basis of tetrathionate
respiration in Salmonella Typhimurium. Mol Microbiol. 1999;32(2):275-287.
40. Winter SE, Thiennimitr P, Winter MG, et al. Gut inflammation provides a respiratory
electron acceptor for Salmonella. Nature. 2010;467(7314):426-429.
41. Thiennimitr P, Winter SE, Winter MG, et al. Intestinal inflammation allows Salmonella to
use ethanolamine to compete with the microbiota. Proc Natl Acad Sci U S A.
2011;108(42):17480-17485.
42. Schmeing TM, Ramakrishnan V. What recent ribosome structures have revealed about the
mechanism of translation. Nature. 2009;461(7268):1234-1242.
130
43. Laursen BS, Sorensen HP, Mortensen KK, Sperling-Petersen HU. Initiation of protein
synthesis in bacteria. Microbiol Mol Biol Rev. 2005;69(1):101-123.
44. Ma J, Campbell A, Karlin S. Correlations between Shine-Dalgarno sequences and gene
features such as predicted expression levels and operon structures. J Bacteriol.
2002;184(20):5733-5745.
45. Salis HM, Mirsky EA, Voigt CA. Automated design of synthetic ribosome binding sites to
control protein expression. Nat Biotech. 2009;27(10):946-950.
46. Chen H, Bjerknes M, Kumar R, Jay E. Determination of the optimal aligned spacing between
the Shine-Dalgarno sequence and the translation initiation codon of Escherichia coli mRNAs.
Nucleic Acids Res. 1994;22(23):4953-4957.
47. Qing G, Xia B, Inouye M. Enhancement of translation initiation by A/T-rich sequences
downstream of the initiation codon in Escherichia coli. J Mol Microbiol Biotechnol. 2003;6(3-
4):133-144.
48. Takahashi S, Furusawa H, Ueda T, Okahata Y. Translation enhancer improves the ribosome
liberation from translation initiation. J Am Chem Soc. 2013;135(35):13096-13106.
49. Vimberg V, Tats A, Remm M, Tenson T. Translation initiation region sequence preferences
in Escherichia coli. BMC Molecular Biology. 2007;8(1):100.
50. Fredrick K, Ibba M. How the sequence of a gene can tune its translation. Cell.
2010;141(2):227-229.
51. Plotkin JB, Kudla G. Synonymous but not the same: The causes and consequences of codon
bias. Nat Rev Genet. 2011;12(1):32-42.
52. Buchan JR, Stansfield I. Halting a cellular production line: Responses to ribosomal pausing
during translation. Biology of the Cell. 2007;99(9):475-487.
53. Elf J, Nilsson D, Tenson T, Ehrenberg M. Selective charging of tRNA isoacceptors explains
patterns of codon usage. Science. 2003;300(5626):1718-1722.
54. Novoa EM, Ribas dP. Speeding with control: Codon usage, tRNAs, and ribosomes. Trends in
Genetics. 2012;28(11):574-581.
55. Cannarozzi G, Schraudolph NN, Faty M, et al. A role for codon order in translation
dynamics. Cell. 2010;141(2):355-367.
56. Ciandrini L, Stansfield I, Romano MC. Ribosome traffic on mRNAs maps to gene ontology:
Genome-wide quantification of translation initiation rates and polysome size regulation. PLoS
Comput Biol. 2013;9(1):e1002866.
57. Chu D, Kazana E, Bellanger N, Singh T, Tuite MF, von der Haar T. Translation elongation
can control translation initiation on eukaryotic mRNAs. EMBO J. 2014;33(1):21-34.
131
58. Kudla G, Murray AW, Tollervey D, Plotkin JB. Coding-sequence determinants of gene
expression in Escherichia coli. Science. 2009;324(5924):255-258.
59. Goodman DB, Church GM, Kosuri S. Causes and effects of N-terminal codon bias in
bacterial genes. Science. 2013;342(6157):475-479.
60. Stewart V, Landick R, Yanofsky C. Rho-dependent transcription termination in the
tryptophanase operon leader region of Escherichia coli K-12. J Bacteriol. 1986;166(1):217-223.
61. Bhushan S, Hoffmann T, Seidelt B, et al. SecM-stalled ribosomes adopt an altered geometry
at the peptidyl transferase center. PLoS Biology. 2010;9(1):e1000581.
62. Chiba S, Ito K. Multisite ribosomal stalling: A unique mode of regulatory nascent chain
action revealed for MifM. Mol Cell. 2012;47(6):863-872.
63. Murakami A, Nakatogawa H, Ito K. Translation arrest of SecM is essential for the basal and
regulated expression of SecA. Proc Natl Acad Sci U S A. 2004;101(33):12330-12335.
64. Sandler P, Weisblum B. Erythromycin-induced ribosome stall in the ermA leader: A
barricade to 5'-to-3' nucleolytic cleavage of the ermA transcript. J Bacteriol. 1989;171(12):6680-
6688.
65. Tsai C, Sauna ZE, Kimchi-Sarfaty C, Ambudkar SV, Gottesman MM, Nussinov R.
Synonymous mutations and ribosome stalling can lead to altered folding pathways and distinct
minima. J Mol Biol. 2008;383(2):281-291.
66. Nakatogawa H, Ito K. Secretion monitor, SecM, undergoes self-translation arrest in the
cytosol. Mol Cell. 2001;7(1):185-192.
67. Yap M, Bernstein HD. The translational regulatory function of SecM requires the precise
timing of membrane targeting. Mol Microbiol. 2011;81(2):540-553.
68. Nakatogawa H, Ito K. The ribosomal exit tunnel functions as a discriminating gate. Cell.
2002;108(5):629-636.
69. Wilson DN, Beckmann R. The ribosomal tunnel as a functional environment for nascent
polypeptide folding and translational stalling. Curr Opin Struct Biol. 2011;21(2):274-282.
70. Tanner DR, Cariello DA, Woolstenhulme CJ, Broadbent MA, Buskirk AR. Genetic
identification of nascent peptides that induce ribosome stalling. J Biol Chem.
2009;284(50):34809-34818.
71. Ito K, Chiba S, Pogliano K. Divergent stalling sequences sense and control cellular
physiology. Biochem Biophys Res Commun. 2010;393(1):1-5.
72. Giudice E, Gillet R. The task force that rescues stalled ribosomes in bacteria. Trends
Biochem Sci. 2013;38(8):403-411.
132
73. Pavlov MY, Watts RE, Tan Z, Cornish VW, Ehrenberg M, Forster AC. Slow peptide bond
formation by proline and other N-alkylamino acids in translation. Proc Natl Acad Sci.
2009;106(1):50-54.
74. Navarre WW, Zou SB, Roy H, et al. PoxA, YjeK, and elongation factor P coordinately
modulate virulence and drug resistance in Salmonella enterica. Mol Cell. 2010;39(2):209-221.
75. Chang YY, Cronan JE,Jr. Genetic and biochemical analyses of Escherichia coli strains
having a mutation in the structural gene (poxB) for pyruvate oxidase. J Bacteriol.
1983;154(2):756-762.
76. Chang YY, Cronan JE,Jr. Mapping nonselectable genes of Escherichia coli by using
transposon Tn10: Location of a gene affecting pyruvate oxidase. J Bacteriol. 1982;151(3):1279-
1289.
77. Glick BR, Ganoza MC. Identification of a soluble protein that stimulates peptide bond
synthesis. Proc Natl Acad Sci. 1975;72(11):4257-4260.
78. Glick BR, Chládek S, Ganoza MC. Peptide bond formation stimulated by protein synthesis
factor EF-P depends on the aminoacyl moiety of the acceptor. Eur J Biochem. 1979;97(1):23-28.
79. Blaha G, Stanley RE, Steitz TA. Formation of the first peptide bond: The structure of EF-P
bound to the 70S ribosome. Science. 2009;325(5943):966-970.
80. Hanawa-Suetsugu K, Sekine S, Sakai H, et al. Crystal structure of elongation factor P from
Thermus thermophilus HB8. Proc Natl Acad Sci. 2004;101(26):9595-9600.
81. Choi S, Choe J. Crystal structure of elongation factor P from Pseudomonas aeruginosa at
1.75 Å resolution. Proteins. 2011;79(5):1688-1693.
82. Bailly M, de Crecy-Lagard V. Predicting the pathway involved in post-translational
modification of elongation factor P in a subset of bacterial species. Biology Direct. 2010;5(1):3.
83. Zou SB, Roy H, Ibba M, Navarre WW. Elongation factor P mediates a novel post-
transcriptional regulatory pathway critical for bacterial virulence. Virulence. 2011;2(2):1-5.
84. Zou SB, Hersch SJ, Roy H, et al. Loss of elongation factor P disrupts bacterial outer
membrane integrity. J Bacteriol. 2012;194(2):413-425.
85. Roy H, Zou SB, Bullwinkle TJ, et al. The tRNA synthetase paralog PoxA modifies
elongation factor-P with (R)-β-lysine. Nat Chem Biol. 2011;7(10):667-669.
86. Peil L, Starosta AL, Virumäe K, et al. Lys34 of translation elongation factor EF-P is
hydroxylated by YfcM. Nat Chem Biol. 2012;8(8):695-697.
87. Bullwinkle TJ, Zou SB, Rajkovic A, et al. (R)-β-lysine-modified elongation factor P
functions in translation elongation. J Biol Chem. 2013;288(6):4416-4423.
133
88. Bochner BR, Gadzinski P, Panomitros E. Phenotype MicroArrays for high-throughput
phenotypic testing and assay of gene function. Genome Res. 2001;11(7):1246-1255.
89. Peng W, Banta LM, Charles TC, Nester EW. The chvH locus of Agrobacterium encodes a
homologue of an elongation factor involved in protein synthesis. J Bacteriol. 2001;183(1):36-45.
90. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli
K-12 using PCR products. Proc Natl Acad Sci. 2000;97(12):6640-6645.
91. Cooper CA, Zhang K, Andres SN, et al. Structural and biochemical characterization of SrcA,
a multi-cargo type III secretion chaperone in Salmonella required for pathogenic association with
a host. PLoS Pathog. 2010;6(2):e1000751.
92. Schmieger H. A method for detection of phage mutants with altered transducing ability.
Molecular and General Genetics MGG. 1971;110(4):378-381.
93. Corcoran CP, Podkaminski D, Papenfort K, Urban JH, Hinton JCD, Vogel J. Superfolder
GFP reporters validate diverse new mRNA targets of the classic porin regulator, MicF RNA. Mol
Microbiol. 2012;84(3):428-445.
94. Beltran P, Plock SA, Smith NH, Whittam TS, Old DC, Selander RK. Reference collection of
strains of the Salmonella Typhimurium complex from natural populations. Microbiology.
1991;137(3):601-606. doi: 10.1099/00221287-137-3-601.
95. Boyd EF, Wang F, Beltran P, Plock SA, Nelson K, Selander RK. Salmonella reference
collection B (SARB): Strains of 37 serovars of subspecies I. Microbiology. 1993;139(6):1125-
1132. doi: 10.1099/00221287-139-6-1125.
96. Boyd EF, Wang FS, Whittam TS, Selander RK. Molecular genetic relationships of the
Salmonellae. Appl Environ Microbiol. 1996;62(3):804-808.
97. Ochman H, Selander RK. Standard reference strains of Escherichia coli from natural
populations. J Bacteriol. 1984;157(2):690-693.
98. Baba T, Ara T, Hasegawa M, et al. Construction of Escherichia coli K-12 in-frame, single-
gene knockout mutants: The keio collection. Mol Syst Biol. 2006;2.
99. Pédelacq J, Cabantous S, Tran T, Terwilliger TC, Waldo GS. Engineering and
characterization of a superfolder green fluorescent protein. Nat Biotech. 2006;24(1):79-88.
100. Gibson DG, Young L, Chuang R, Venter JC, Hutchison CA, Smith HO. Enzymatic
assembly of DNA molecules up to several hundred kilobases. Nat Meth. 2009;6(5):343-345.
101. Cherepanov PP, Wackernagel W. Gene disruption in Escherichia coli: TcR and KmR
cassettes with the option of flp-catalyzed excision of the antibiotic-resistance determinant. Gene.
1995;158(1):9-14.
134
102. Hancock REW. Alterations in outer membrane permeability. Annu Rev Microbiol.
1984;38(1):237-264.
103. Loh B, Grant C, Hancock RE. Use of the fluorescent probe 1-N-phenylnaphthylamine to
study the interactions of aminoglycoside antibiotics with the outer membrane of Pseudomonas
aeruginosa. Antimicrobial Agents Chemother. 1984;26(4):546-551.
104. Vaara M, Plachy WZ, Nikaido H. Partitioning of hydrophobic probes into
lipopolysaccharide bilayers. Biochimica et Biophysica Acta (BBA) - Biomembranes.
1990;1024(1):152-158.
105. Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins
from silver-stained polyacrylamide gels. Anal Chem. 1996;68(5):850-858.
106. Chan QWT, Howes CG, Foster LJ. Quantitative comparison of caste differences in
honeybee hemolymph. Molecular & Cellular Proteomics. 2006;5(12):2252-2262.
107. de Godoy, Lyris M. F., Olsen JV, Cox J, et al. Comprehensive mass-spectrometry-based
proteome quantification of haploid versus diploid yeast. Nature. 2008;455(7217):1251-1254.
108. Cox J, Matic I, Hilger M, et al. A practical guide to the MaxQuant computational platform
for SILAC-based quantitative proteomics. Nat Protocols. 2009;4(5):698-705.
109. Dennis G, Sherman B, Hosack D, et al. DAVID: Database for annotation, visualization, and
integrated discovery. Genome Biol. 2003;4(5):P3.
110. Hersch SJ, Wang M, Zou SB, et al. Divergent protein motifs direct elongation factor P-
mediated translational regulation in Salmonella enterica and Escherichia coli. mBio. 2013;4(2).
111. Loewen PC, Triggs BL. Genetic mapping of katF, a locus that with katE affects the
synthesis of a second catalase species in Escherichia coli. J Bacteriol. 1984;160(2):668-675.
112. Brittnacher MJ, Fong C, Hayden HS, Jacobs MA, Radey M, Rohmer L. PGAT: A
multistrain analysis resource for microbial genomes. Bioinformatics. 2011;27(17):2429-2430.
113. Bearson SMD, Bearson BL, Brunelle BW, Sharma VK, Lee IS. A mutation in the poxA
gene of Salmonella enterica serovar Typhimurium alters protein production, elevates
susceptibility to environmental challenges, and decreases swine colonization. Foodborne
Pathogens and Disease. 2011;8(6):725-732.
114. Ude S, Lassak J, Starosta AL, Kraxenberger T, Wilson DN, Jung K. Translation elongation
factor EF-P alleviates ribosome stalling at polyproline stretches. Science. 2013;339(6115):82-85.
115. Doerfel LK, Wohlgemuth I, Kothe C, Peske F, Urlaub H, Rodnina MV. EF-P is essential
for rapid synthesis of proteins containing consecutive proline residues. Science.
2013;339(6115):85-88.
135
116. Murata T, Tseng W, Guina T, Miller SI, Nikaido H. PhoPQ-mediated regulation produces a
more robust permeability barrier in the outer membrane of Salmonella enterica serovar
Typhimurium. J Bacteriol. 2007;189(20):7213-7222.
117. Nikaido H. Molecular basis of bacterial outer membrane permeability revisited.
Microbiology and Molecular Biology Reviews. 2003;67(4):593-656.
118. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: Paths toward the
comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009;37(1):1-13.
119. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene
lists using DAVID bioinformatics resources. Nat Protocols. 2008;4(1):44-57.
120. Rees DM, Montgomery MG, Leslie AGW, Walker JE. Structural evidence of a new
catalytic intermediate in the pathway of ATP hydrolysis by F1-ATPase from bovine heart
mitochondria. Proceedings of the National Academy of Sciences. 2012;109(28):11139-11143.
121. Biter AB, Lee S, Sung N, Tsai FTF. Structural basis for intersubunit signaling in a protein
disaggregating machine. Proceedings of the National Academy of Sciences. 2012;109(31):12515-
12520.
122. Niwa H, Tsuchiya D, Makyio H, Yoshida M, Morikawa K. Hexameric ring structure of the
ATPase domain of the membrane-integrated metalloprotease FtsH from Thermus thermophilus
HB8. Structure. 2002;10(10):1415-1424.
123. Gong F, Yanofsky C. Instruction of translating ribosome by nascent peptide. Science.
2002;297(5588):1864-1867.
124. Ramu H, Vázquez-Laslop N, Klepacki D, et al. Nascent peptide in the ribosome exit tunnel
affects functional properties of the A-site of the peptidyl transferase center. Mol Cell.
2011;41(3):321-330.
125. Woolstenhulme CJ, Parajuli S, Healey DW, et al. Nascent peptides that block protein
synthesis in bacteria. Proceedings of the National Academy of Sciences. 2013;110(10):E878-
E887.
126. Hersch SJ, Elgamal S, Katz A, Ibba M, Navarre WW. Translation initiation rate determines
the impact of ribosome stalling on bacterial protein synthesis. J Biol Chem. 2014;289(41):28160-
28171.
127. Elgamal S, Katz A, Hersch SJ, et al. EF-P dependent pauses integrate proximal and distal
signals during translation. PLoS Genet. 2014;10(8):e1004553.
128. Senior AE. The proton-translocating ATPase of Escherichia coli. Annu Rev Biophys
Biophys Chem. 1990;19(1):7-41.
129. von Ballmoos C, Cook GM, Dimroth P. Unique rotary ATP synthase and its biological
diversity. Annu Rev Biophys. 2008;37(1):43-64.
136
130. Peil L, Starosta AL, Lassak J, et al. Distinct XPPX sequence motifs induce ribosome
stalling, which is rescued by the translation elongation factor EF-P. Proc Natl Acad Sci.
2013;110(38):15265-15270.
131. Voss NR, Gerstein M, Steitz TA, Moore PB. The geometry of the ribosomal polypeptide
exit tunnel. J Mol Biol. 2006;360(4):893-906.
132. Urban J, Vogel J. A green fluorescent protein (GFP)-based plasmid system to study post-
transcriptional control of gene expression in vivo. In: Serganov A, ed. Vol 540. Humana Press;
2009:301-319. 10.1007/978-1-59745-558-9_22.
133. Lu P, Vogel C, Wang R, Yao X, Marcotte EM. Absolute protein expression profiling
estimates the relative contributions of transcriptional and translational regulation. Nat Biotech.
2007;25(1):117-124.
134. Rychkova A, Mukherjee S, Bora RP, Warshel A. Simulating the pulling of stalled elongated
peptide from the ribosome by the translocon. Proceedings of the National Academy of Sciences.
2013;110(25):10195-10200.
135. Wilson DN. Peptides in the ribosomal tunnel talk back. Mol Cell. 2011;41(3):247-248.
136. Woolstenhulme C, Guydosh N, Green R, Buskirk A. High-precision analysis of
translational pausing by ribosome profiling in bacteria lacking EFP. Cell Reports. 2015;11(1):13-
21.
137. Shah P, Ding Y, Niemczyk M, Kudla G, Plotkin J. Rate-limiting steps in yeast protein
translation. Cell. 2013;153(7):1589-1601.
138. Goldman E, Rosenberg AH, Zubay G, Studier WF. Consecutive low-usage leucine codons
block translation only when near the 5′ end of a message in Escherichia coli. J Mol Biol.
1995;245(5):467-473.
139. Ikemura T. Codon usage and tRNA content in unicellular and multicellular organisms. Mol
Biol Evol. 1985;2(1):13-34.
140. Stoletzki N, Eyre-Walker A. Synonymous codon usage in Escherichia coli: Selection for
translational accuracy. Mol Biol Evol. 2007;24(2):374-381.
141. Ingolia NT, Ghaemmaghami S, Newman JRS, Weissman JS. Genome-wide analysis in vivo
of translation with nucleotide resolution using ribosome profiling. Science. 2009;324(5924):218-
223.
142. Tuller T, Carmi A, Vestsigian K, et al. An evolutionarily conserved mechanism for
controlling the efficiency of protein translation. Cell. 2010;141(2):344-354.
143. Angov E. Codon usage: Nature's roadmap to expression and folding of proteins.
Biotechnology Journal. 2011;6(6):650-659.
137
144. Kemp AJ, Betney R, Ciandrini L, Schwenger ACM, Romano MC, Stansfield I. A yeast
tRNA mutant that causes pseudohyphal growth exhibits reduced rates of CAG codon translation.
Molecular Microbiology. 2013;87(2):284-300.
145. Chartier M, Gaudreault F, Najmanovich R. Large-scale analysis of conserved rare codon
clusters suggests an involvement in co-translational molecular recognition events.
Bioinformatics. 2012;28(11):1438-1445.
146. Zhang G, Hubalewska M, Ignatova Z. Transient ribosomal attenuation coordinates protein
synthesis and co-translational folding. Nat Struct Mol Biol. 2009;16(3):274-280.
147. Oh E, Becker A, Sandikci A, et al. Selective ribosome profiling reveals the cotranslational
chaperone action of trigger factor in vivo. Cell. 2011;147(6):1295-1308.
148. Jensen PR, Michelsen O. Carbon and energy metabolism of atp mutants of Escherichia coli.
J Bacteriol. 1992;174(23):7635-7641.
149. Butlin JD, Cox GB, Gibson F. Oxidative phosphorylation in Escherichia coli K 12.
mutations affecting magnesium ion- or calcium ion-stimulated adenosine triphosphatase.
Biochem J. 1971;124(1):75-81.
150. Boogerd FC, Boe L, Michelsen O, Jensen PR. Atp mutants of Escherichia coli fail to grow
on succinate due to a transport deficiency. J Bacteriol. 1998;180(22):5855-5859.
151. Chen G, Patten CL, Schellhorn HE. Positive selection for loss of RpoS function in
Escherichia coli. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis.
2004;554(1):193-203.
152. Hengge-Aronis R. Signal transduction and regulatory mechanisms involved in control of the
σS (RpoS) subunit of RNA polymerase. Microbiology and Molecular Biology Reviews.
2002;66(3):373-395.
153. McQuiston JR, Herrera-Leon S, Wertheim BC, et al. Molecular phylogeny of the
Salmonellae: Relationships among salmonella species and subspecies determined from four
housekeeping genes and evidence of lateral gene transfer events. J Bacteriol.
2008;190(21):7060-7067.
154. Lai X, Wang S, Uhlin BE. Expression of cytotoxicity by potential pathogens in the standard
Escherichia coli collection of reference (ECOR) strains. Microbiology. 1999;145(11):3295-3303.
doi: 10.1099/00221287-145-11-3295.
155. Tannahill GM, Curtis AM, Adamik J, et al. Succinate is an inflammatory signal that induces
IL-1β through HIF-1α. Nature. 2013;496(7444):238-242.
156. Sievers F, Wilm A, Dineen D, et al. Fast, scalable generation of high-quality protein
multiple sequence alignments using clustal omega. Molecular systems biology. 2011;7:539.
138
157. Salgado H, Peralta-Gil M, Gama-Castro S, et al. RegulonDB v8.0: Omics data sets,
evolutionary conservation, regulatory phrases, cross-validated gold standards and more. Nucleic
Acids Res. 2012;41:D203-D213.
158. Abo-Amer A, Munn J, Jackson K, et al. DNA interaction and phosphotransfer of the C4-
dicarboxylate- responsive DcuS-DcuR two-component regulatory system from Escherichia coli.
J Bacteriol. 2004;186(6):1879-1889.
159. Kröger C, Dillon SC, Cameron ADS, et al. The transcriptional landscape and small RNAs
of Salmonella enterica serovar Typhimurium. Proceedings of the National Academy of Sciences.
2012;109(20):E1277-E1286.
160. Michelucci A, Cordes T, Ghelfi J, et al. Immune-responsive gene 1 protein links
metabolism to immunity by catalyzing itaconic acid production. Proceedings of the National
Academy of Sciences. 2013;110(19):7820-7825.
161. McFadden BA, Purohit S. Itaconate, an isocitrate lyase-directed inhibitor in Pseudomonas
indigofera. J Bacteriol. 1977;131(1):136-144.
162. Strelko CL, Lu W, Dufort FJ, et al. Itaconic acid is a mammalian metabolite induced during
macrophage activation. J Am Chem Soc. 2011;133(41):16386-16389.
163. Patel TR, McFadden BA. Caenorhabditis elegans and Ascaris suum: Inhibition of isocitrate
lyase by itaconate. Exp Parasitol. 1978;44(2):262-268.
164. Sasikaran J, Ziemski M, Zadora PK, Fleig A, Berg IA. Bacterial itaconate degradation
promotes pathogenicity. Nat Chem Biol. 2014;10(5):371-377.
165. Rathman M, Sjaastad MD, Falkow S. Acidification of phagosomes containing Salmonella
Typhimurium in murine macrophages. Infect Immun. 1996;64(7):2765-2773.
166. Bearson BL, Wilson L, Foster JW. A low pH-inducible, PhoPQ-dependent acid tolerance
response protects Salmonella Typhimurium against inorganic acid stress. J Bacteriol.
1998;180(9):2409-2417.
167. Baker KE, Ditullio KP, Neuhard J, Kelln RA. Utilization of orotate as a pyrimidine source
by Salmonella Typhimurium and Escherichia coli requires the dicarboxylate transport protein
encoded by dctA. J Bacteriol. 1996;178(24):7099-7105.
168. Steinmetz PA, Wörner S, Unden G. Differentiation of DctA and DcuS function in the
DctA/DcuS sensor complex of Escherichia coli: Function of DctA as an activity switch and of
DcuS as the C4-dicarboxylate sensor. Molecular Microbiology. 2014;94(1):218-229.
169. Shah DH, Lee MJ, Park JH, et al. Identification of Salmonella Gallinarum virulence genes
in a chicken infection model using PCR-based signature-tagged mutagenesis. Microbiology.
2005;151(Pt 12):3957-3968.
139
170. Eriksson S, Lucchini S, Thompson A, Rhen M, Hinton JC. Unravelling the biology of
macrophage infection by gene expression profiling of intracellular Salmonella enterica. Mol
Microbiol. 2003;47(1):103-118.
171. Srikumar S, Kröger C, Hébrard M, et al. RNA-seq brings new insights to the intra-
macrophage transcriptome of Salmonella Typhimurium. PLoS Pathog. 2015;11(11):e1005262.
172. Pujol C, Grabenstein JP, Perry RD, Bliska JB. Replication of Yersinia pestis in interferon γ-
activated macrophages requires ripA, a gene encoded in the pigmentation locus. Proc Natl Acad
Sci U S A. 2005;102(36):12909-12914.
173. Kröger C, Colgan A, Srikumar S, et al. An infection-relevant transcriptomic compendium
for Salmonella enterica serovar Typhimurium. Cell Host Microbe. 2013;14(6):683-695.
174. McCusker MP, Hokamp K, Buckley JF, Wall PG, Martins M, Fanning S. Complete genome
sequence of Salmonella enterica serovar Agona pulsed-field type SAGOXB.0066, cause of a
2008 pan-european outbreak. Genome Announc. 2014;2(1):10.1128/genomeA.01219-13.
175. Cordes T, Michelucci A, Hiller K. Itaconic acid: The surprising role of an industrial
compound as a mammalian antimicrobial metabolite. Annu Rev Nutr. 2015;35:451-473.
176. Traxler MF, Summers SM, Nguyen H, et al. The global, ppGpp-mediated stringent response
to amino acid starvation in Escherichia coli. Mol Microbiol. 2008;68(5):1128-1148.
177. Gosset G, Zhang Z, Nayyar S, Cuevas WA, Saier MH. Transcriptome analysis of crp-
dependent catabolite control of gene expression in Escherichia coli. J Bacteriol.
2004;186(11):3516-3524.
178. Ishihama A, Shimada T, Yamazaki Y. Transcription profile of Escherichia coli: Genomic
SELEX search for regulatory targets of transcription factors. Nucleic Acids Res.
2016;44(5):2058-2074.
179. Münch R, Hiller K, Barg H, et al. PRODORIC: Prokaryotic database of gene regulation.
Nucleic Acids Res. 2003;31(1):266-269.