mutagenesis techniques (pdf)

This booklet is brought to you by the AAAS/Science Business Office

Sponsored by

Over 20 Years of InnovationAs a global leader in the development of innovative technologies,

our products simplify, accelerate, and improve drug discovery and

drug development research. Since 1984, our reagents, kits, software,

and instrumentation systems have been used in pharmaceutical and

biotechnology laboratories worldwide to determine the molecular

mechanisms of health and disease as well as to search for new drug

therapies and diagnostic tests.

To order a copy of our current product catalog, visit

www.stratagene.com/catalog.

AMPLIFICATION

CELL BIOLOGY

CLONING

MICROARRAYS

NUCLEIC ACID ANALYSIS

PROTEIN FUNCTION & ANALYSIS

QUANTITATIVE PCR

SOFTWARE SOLUTIONS

Corporate_Ad_Q407_Science Bookle1 1 8/21/2007 3:52:12 PM

Contents

1© 2007 by The American Association for the Advancement of Science. All rights reserved.

2 Introductions: Bringing About Change Sean Sanders

The Power of Mutagenesis Patricia Sardina

4 Bypassing a Kinase Activity with an ATP-Competitive Drug Feroz R. Papa, Chao Zhang, Kevan Shokat, Peter Walter Science 28 November 2003 302: 1533-1537

12 Human Catechol-O-Methyltransferase Haplotypes Modulate Protein Expression by Altering mRNA Secondary Structure

A. G. Nackley, S.A. Shabalina, I. E. Tchivileva, K. Satterfield, O. Korchynskyi, S. S. Makarov, W. Maixner, L. DiatchenkoScience 22 December 2006 314: 1930-1933

17 An mRNA Surveillance Mechanism That Eliminates Transcripts Lacking Termination Codons

Pamela A. Frischmeyer, Ambro van Hoof, Kathryn O’Donnell, Anthony L. Guerrerio, Roy Parker, Harry C. DietzScience 22 March 2002 295: 2258–2261

24 Direct Demonstration of an Adaptive ConstraintStephen P. Miller, Mark Lunzer, Antony M. DeanScience 20 October 2006 314: 458–461

30 Technical Notes: Large Insertions: Two Simple Steps Using Quikchange® II Site-Directed Mutagenesis Kits

About the Cover:γ-Exonuclease binds to an end of double-stranded DNA and degrades one of the strands in a highly processive manner. The structure of the exonuclease consists of a toroidal trimer (~94 angstroms in diameter) and is presumed to enclose its substrate in the manner illustrated. [Image from the cover of Science 277 (19 September 1997)]

Design and Layout: Amy HardcastleCopy Editor: Robert Buck

Introductions

2

Bringing About ChangeThis new method of mutagenesis has considerable potential in genetic studies.… [S]pecific mutation within protein structural genes will allow the production of proteins with specific amino acid changes. This will allow precise studies of structure-function relationships, for example the role of specific amino acids in enzyme active sites, and the more convenient production of proteins whose action is species specific, for example the conversion of a cloned gene coding for a lower vertebrate hormone to that for a human.— closing paragraph of the seminal paper from Michael Smith’s lab first describing site-directed mutagenesis1

The art of life lies in a constant readjustment to our surroundings.— Okakura Kakuzo, Japanese scholar

It is a truism that, to survive, we must be able to change. This, too, is the essence of evolution on the most fundamental biological level: without the innate ability for DNA to undergo mutation, it would not be possible for organisms to adapt and survive

changes in their environment. Since humankind achieved an understanding of this muta-genesis process and its power, we have been trying to control it and bend it to our will.

First described in 1978 by Michael Smith at the University of British Columbia in Vancouver,1 site-directed mutagenesis (SDM) has become an invaluable tool for molecu-lar biologists. Originally named oligonucleotide-directed mutagenesis, the first experi-ments made use of short, 12-nucleotide strands of synthetic DNA containing a mismatch to an intrinsic reporter gene of the bacteriophage ΦX174. Using these oligonucleotides, together with DNA polymerase I from E. coli and T4 DNA ligase, Smith and his team demonstrated that it was possible to introduce a permanent mutation in the circular DNA of the phage, resulting in a phenotypic change. The prescient quote above shows the au-thors’ understanding of the manifold applications for the new methodology.

Since its development, a number of modifications and improvements have been made to the SDM technique. The greatest advance came with the invention of the polymerase chain reaction (PCR) in 1983. Since then, SDM and PCR have been inextricably linked, a circumstance reflected in the 1993 Nobel in Chemistry being shared by Kary Mullis, the pioneer of PCR, and Michael Smith. This advance made SDM both quicker and more flexible and has enabled it to become a standard and necessary technique employed in labs worldwide.

In this booklet, we bring together a collection of influential papers that all depended on SDM for their success. As one of the most used techniques in molecular biology, it plays an often unacknowledged–but essential–role in many research projects. There is lit-tle doubt that future refinements and improvements will be made to the technique, allow-ing for an even broader range of applications. As these modifications are commercialized and made available to all scientists, new and innovative uses for them will be discovered, facilitating enrichment of our knowledge in many areas of research, from basic cellular processes to complex diseases such as neurodegenerative disorders and cancer.

Sean Sanders, Ph.D.Commercial Editor, Science

1 C. A. Hutchison III, et al. Mutagenesis at a specific position in a DNA sequence. J. Biol. Chem. 253:6551-6560 (1978).

3

Introductions

The Power of Mutagenesis

In vitro mutagenesis is a very powerful tool for studying protein structure-function relationships, altering protein activity, and for modifying vector sequences to incor-porate affinity tags and correct frame shift errors. A variety of mutagenesis techniques

have been developed over the past decade that allow researchers to generate point mu-tations, modify (insert, delete or replace) one or more codons, swap domains between related gene sequences, and create diverse collections of mutant clones.

Mutagenesis strategies can be divided into two main types, random or site-directed. With random mutagenesis, point mutations are introduced at random positions in a gene-of-interest, typically through PCR employing an error-prone DNA polymerase (error-prone PCR). Randomized sequences are then cloned into a suitable expression vector, and the resulting mutant libraries can be screened to identify mutants with altered or im-proved properties. By correlating changes in function to alterations in protein sequence, researchers can create a preliminary map of protein functional domains.

Site-directed mutagenesis is the method of choice for altering a gene or vector se-quence at a selected location. Point mutations, insertions, or deletions are introduced by incorporating primers containing the desired modification(s) with a DNA polymerase in an amplification reaction. In more complex protein engineering experiments, researchers can design mutagenic primers to incorporate degenerate codons (site-saturation muta-genesis).

Stratagene, now an Agilent Technologies company, is a worldwide leader in develop-ing innovative products and technologies for life science research. Our expertise in en-zyme engineering has translated into the most efficient and easy-to-use mutagenesis kits available today. Our QuikChange® Site-Directed Mutagenesis products provide the most reliable and accurate method for performing targeted mutagenesis, saving time and valu-able resources. These products have been cited in thousands of publications, including the papers featured in this booklet. In addition, our GeneMorph® Random Mutagenesis products offer an easy, robust method for creating diverse mutant collections.

Our mutagenesis product offering has been substantially enhanced by our newest product introduction, the QuikChange® Lightning Site-Directed Mutagenesis Kit. This kit offers significant time-savings by incorporating new, exclusive fast enzymes which allow researchers to shorten the duration of the thermal cycling and digestion steps while maintaining ultra-high fidelity and mutation efficiency.

The efficacy of mutagenesis as a tool for studying protein function and structure may lead to new experimental therapeutic approaches to diseases such as cancer. Precise map-ping of protein-protein interactions may ultimately lead to a greater understanding of disease mechanics. This Science Mutagenesis Techniques Collection booklet illustrates the utility of mutagenesis in a wide variety of research studies. We are pleased to have the opportunity to sponsor this publication.

Patricia SardinaProduct Manager Stratagene, An Agilent Technologies Company

4

Secretory and transmembrane pro-teins traversing the endoplasmic reticulum (ER) during their biogen-

esis fold to their native states in this com-partment (1). This process is facilitated by a plethora of ER-resident activities (the protein folding machinery) (2). Insuffi-cient protein folding capacity causes an accumulation of unfolded proteins in the ER lumen (a condition referred to as ER stress) and triggers a transcriptional pro-gram called the UPR. In yeast, UPR tar-gets include genes encoding chaperones, oxido-reductases, phospholipid biosyn-thetic enzymes, ER-associated degrada-tion components, and proteins functioning downstream in the secretory pathway (3). Together, UPR target activities afford pro-teins passing through the ER an extended opportunity to fold and assemble properly, dispose of unsalvageable unfolded poly-peptides, and increase the capacity for ER export.

The yeast UPR is signaled through the ER stress sensor Ire1, a single-spanning ER transmembrane protein with three

functional domains (4). The most N-ter-minal domain, which resides in the ER lumen, senses elevated levels of unfolded ER proteins. Dissociation of ER chaper-ones from Ire1 as they become engaged with unfolded proteins is thought to trig-ger Ire1 activation (3).

Ire1’s most C-terminal domain is a reg-ulated endoribonuclease (RNase), which has a single known substrate in yeast: the HAC1u mRNA (u stands for uninduced), which encodes the Hac1 transcriptional activator necessary for activation of UPR targets (5). HAC1u mRNA is constitutively transcribed, but not translated, because it contains a nonconventional translation-in-hibitory intron (6). Upon activation, Ire1’s RNase cleaves HAC1u mRNA at two spe-cific sites, excising the intron (7). The 5′ and 3′ exons are rejoined by tRNA ligase (8), resulting in spliced HAC1i mRNA (i stands for induced), which is actively translated to produce the Hac1 transcrip-tional activator, in turn upregulating UPR target genes (5).

A functional kinase domain precedes the RNase domain on the cytosolic side of the ER membrane (9). Activation of Ire1 leads to its oligomerization in the ER membrane, followed by trans-auto-phosphorylation (9, 10). Mutations of catalytically essential kinase active-site

Bypassing a Kinase Activity with anATP-Competitive Drug

Feroz R. Papa,1,3* Chao Zhang,2 Kevan Shokat,2 Peter Walter3,4

Unfolded proteins in the endoplasmic reticulum cause trans-autophosphoryl-ation of the bifunctional transmembrane kinase Ire1, which induces its endo-ribonuclease activity. The endoribonuclease initiates nonconventional splicingof HAC1 messenger RNA to trigger the unfolded-protein response (UPR). Weexplored the role of Ire1’s kinase domain by sensitizing it through site-directedmutagenesis to the ATP-competitive inhibitor 1NM-PP1. Paradoxically, ratherthan being inhibited by 1NM-PP1, drug-sensitized Ire1 mutants required 1NM-PP1 as a cofactor for activation. In the presence of 1NM-PP1, drug-sensitizedIre1 bypassed mutations that inactivate its kinase activity and induced a fullUPR. Thus, rather than through phosphorylation per se, a conformationalchange in the kinase domain triggered by occupancy of the active site with aligand leads to activation of all known downstream functions.

Secretory and transmembrane proteins tra-versing the endoplasmic reticulum (ER) dur-ing their biogenesis fold to their native statesin this compartment (1). This process is fa-cilitated by a plethora of ER-resident activi-ties (the protein folding machinery) (2). In-sufficient protein folding capacity causes anaccumulation of unfolded proteins in the ERlumen (a condition referred to as ER stress)and triggers a transcriptional program calledthe UPR. In yeast, UPR targets include genesencoding chaperones, oxido-reductases,phospholipid biosynthetic enzymes, ER-associated degradation components, and pro-teins functioning downstream in the secretorypathway (3). Together, UPR target activitiesafford proteins passing through the ER anextended opportunity to fold and assembleproperly, dispose of unsalvageable unfoldedpolypeptides, and increase the capacity forER export.

The yeast UPR is signaled through the ERstress sensor Ire1, a single-spanning ERtransmembrane protein with three functionaldomains (4). The most N-terminal domain,which resides in the ER lumen, senses ele-vated levels of unfolded ER proteins. Disso-ciation of ER chaperones from Ire1 as theybecome engaged with unfolded proteins isthought to trigger Ire1 activation (3).

Ire1’s most C-terminal domain is a regu-lated endoribonuclease (RNase), which has asingle known substrate in yeast: the HAC1u

mRNA (u stands for uninduced), which en-codes the Hac1 transcriptional activator nec-essary for activation of UPR targets (5).HAC1u mRNA is constitutively transcribed,but not translated, because it contains a non-conventional translation-inhibitory intron(6). Upon activation, Ire1’s RNase cleavesHAC1u mRNA at two specific sites, excisingthe intron (7). The 5� and 3� exons are re-joined by tRNA ligase (8), resulting inspliced HAC1i mRNA (i stands for induced),which is actively translated to produce theHac1 transcriptional activator, in turn up-regulating UPR target genes (5).

A functional kinase domain precedes theRNase domain on the cytosolic side of the ERmembrane (9). Activation of Ire1 leads to itsoligomerization in the ER membrane, fol-lowed by trans-autophosphorylation (9, 10).Mutations of catalytically essential kinaseactive-site residues or mutations of residuesknown to become phosphorylated preventHAC1u mRNA splicing and abrogate UPRsignaling, demonstrating that Ire1’s kinasephosphotransfer function is essential forRNase activation (4, 9, 11). Thus, Ire1 com-municates an unfolded-protein signal fromthe ER to the cytosol using the lumenal do-main as the sensor and the RNase domain asthe effector. Although clearly important forthe circuitry of this machine, it is unknownwhy and how Ire1’s kinase is required foractivation of its RNase. To dissect the role ofIre1’s kinase function, we used a recentlydeveloped strategy that allows us to sensitizeIre1 to specific kinase inhibitors (12).Unexpected behavior of Ire1 mutants

sensitized to an ATP-competitive drug.The mutation of Leu745—situated at a conservedposition in the adenosine 5�-triphosphate (ATP)–binding site—to Ala or Gly is predicted to sen-

sitize Ire1 to the ATP-competitive drug 1-tert-butyl-3-naphthalen-1-ylmethyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylemine (1NM-PP1) by creatingan enlarged active-site pocket not found in anywild-type kinase (13). Most kinases are not af-fected by these mutations, but others arepartially or severely impaired (14, 15). ForIre1, these substitutions markedly decreasedUPR signaling (assayed through an invivo reporter), which is indicative of de-creased kinase activity. Ire1(L7453A745)[Ire1(L745A) (16)] showed a 40% decreaserelative to the wild-type kinase, whereasIre1(L745G) showed a �90% decrease,approaching that of the �ire1 control(Fig. 1A).

Unexpectedly, the addition of 1NM-PP1to cells expressing partially activeIre1(L745A) caused no inhibition, even atconcentrations that completely inhibit other1NM-PP1–sensitized kinases (15, 17). In-stead, 1NM-PP1 increased reporter activityslightly (Fig. 1A, bar 4). To our further sur-prise, 1NM-PP1 provided significantly re-stored signaling to the severely crippledIre1(L745G) (Fig. 1A, bar 6). This resultpresents a paradox: An ATP-competitivecompound that was expected to function asan inhibitor of the rationally engineered mu-tant enzyme instead permits activation.

UPR activation strictly required inductionby dithiothreitol (DTT) (Fig. 1) or tunicamy-cin (18); 1NM-PP1 by itself had no effect,indicating that it does not induce the UPR bydirectly affecting ER protein folding. More-over, the effects of 1NM-PP1 were specific:other structurally similar compounds, 2-naphthylmethyl PP1, which is a regio-isomer,and 1-naphthyl PP1, which lacks the methyl-ene group between the heterocyclic ring andthe naphthyl group, neither activated nor in-hibited wild-type or mutant Ire1 (18).

Activation rather than inhibition by anATP-competitive inhibitor raised the ques-tion of whether the kinase activity is neces-sary in the 1NM-PP1–sensitized mutants. Tothis end, we combined the L745A or L745Gmutation with a “kinase-dead” variantD828A, which lacks an active-site Asp resi-due needed to chelate Mg2� (11, 19). Where-as the Ire1(L745A,D828A) double mutantwas detectable in cells at levels near those ofwild-type cells, Ire1(L745G,D828A) was un-detectable, probably because of instability,and could not be assayed. As expected,Ire1(L745A,D828A) was unable to inducethe UPR reporter with DTT alone, but theaddition of 1NM-PP1 allowed activation tolevels approaching 80% of the wild-type ac-tivation level (Fig. 1B). This suggests that thebinding of 1NM-PP1 to 1NM-PP1–sensitizedIre1 renders the kinase activity dispensable.

1Department of Medicine, 2Department of Cellularand Molecular Pharmacology, 3Department of Bio-chemistry and Biophysics, 4Howard Hughes MedicalInstitute, University of California, San Francisco, CA94143–2200, USA.

*To whom correspondence should be addressed. E-mail: [email protected]

RESEARCH ARTICLE

www.sciencemag.org SCIENCE VOL 302 28 NOVEMBER 2003 1533

Bypassing a Kinase Activity with an ATP-Competitive DrugFeroz R. Papa,1,3* Chao Zhang,2 Kevan Shokat,2 Peter Walter3,4

Unfolded proteins in the endoplasmic reticulum cause trans-autophosphorylation of the bifunctional transmembrane kinase Ire1, which induces its endoribonuclease activity. The endoribonuclease initiates nonconventional splicing of HAC1 messenger RNA to trigger the unfolded protein response (UPR). We explored the role of Ire1’s kinase domain by sensitizing it through site-directed mutagenesis to the ATP-competitive inhibitor 1NM-PP1. Paradoxically, rather than being inhibited by 1NM-PP1, drug-sensitized Ire1 mutants required 1NM-PP1 as a cofactor for activation. In the presence of 1NM-PP1, drug-sensitized Ire1 bypassed mutations that inactivate its kinase activity and induced a full UPR. Thus, rather than through phosphorylation per se, a conformational change in the kinase domain triggered by occupancy of the active site with a ligand leads to activation of all known downstream functions.

5

residues or mutations of residues known to become phosphorylated prevent HAC1u mRNA splicing and abrogate UPR sig-naling, demonstrating that Ire1’s kinase phosphotransfer function is essential for RNase activation (4, 9, 11). Thus, Ire1 communicates an unfolded-protein signal from the ER to the cytosol using the lume-nal domain as the sensor and the RNase domain as the effector. Although clearly important for the circuitry of this machine, it is unknown why and how Ire1’s kinase is required for activation of its RNase. To dissect the role of Ire1’s kinase function, we used a recently developed strategy that allows us to sensitize Ire1 to specific ki-nase inhibitors (12).

Unexpected behavior of Ire1 mu-tants sensitized to an ATP-competitive drug. The mutation of Leu745—situated at a conserved position in the adenosine 5′-triphosphate (ATP)–binding site—to Ala or Gly is predicted to sensitize Ire1 to the ATP-competitive drug 1-tert-butyl-3-naphthalen-1-ylmethyl-1H-pyrazolo[3,4- d]pyrimidin-4-ylemine (1NM-PP1) by creating an enlarged active-site pocket not found in any wild-type kinase (13). Most kinases are not affected by these muta-tions, but others are partially or severely impaired (14, 15). For Ire1, these substi-tutions markedly decreased UPR signal-ing (assayed through an in vivo reporter), which is indicative of decreased kinase ac-tivity. Ire1(L745→A745) [Ire1(L745A) (16)] showed a 40% decrease relative to the wild-type kinase, whereas Ire1(L745G) showed a >90% decrease, approaching that of the Δire1 control (Fig. 1A)

Unexpectedly, the addition of 1NM-PP1 to cells expressing partially active Ire1(L745A) caused no inhibition, even at concentrations that completely inhibit other 1NM-PP1–sensitized kinases (15, 17). Instead, 1NM-PP1 increased reporter activity slightly (Fig. 1A, bar 4). To our further surprise, 1NM-PP1 provided sig-nificantly restored signaling to the severe-ly crippled Ire1(L745G) (Fig. 1A, bar 6).

This result presents a paradox: An ATP-competitive compound that was expected to function as an inhibitor of the rationally engineered mutant enzyme instead per-mits activation.

UPR activation strictly required induc-tion by dithiothreitol (DTT) (Fig. 1) or tunicamycin (18); 1NM-PP1 by itself had no effect, indicating that it does not induce the UPR by directly affecting ER protein folding. Moreover, the effects of 1NM-PP1 were specific: other structurally simi-lar compounds, 2-naphthylmethyl PP1, which is a regio-isomer, and 1-naphthyl PP1, which lacks the methylene group be-tween the heterocyclic ring and the naph-thyl group, neither activated nor inhibited wild-type or mutant Ire1 (18).

Activation rather than inhibition by an ATP-competitive inhibitor raised the question of whether the kinase activ-ity is necessary in the 1NM-PP1–sensi-tized mutants. To this end, we combined the L745A or L745G mutation with a “kinase-dead” variant D828A, which lacks an active-site Asp residue needed to chelate Mg2+(11, 19). Whereas the Ire1(L745A,D828A) double mutant was detectable in cells at levels near those of wild-type cells, Ire1(L745G,D828A) was undetectable, probably because of instabil-ity, and could not be assayed. As expected, Ire1(L745A,D828A) was unable to induce the UPR reporter with DTT alone, but the addition of 1NM-PP1 allowed acti-vation to levels approaching 80% of the wild-type activation level (Fig. 1B). This suggests that the binding of 1NM-PP1 to 1NM-PP1–sensitized Ire1 renders the ki-nase activity dispensable.

Trans-autophosphorylation by Ire1 of two activation-segment Ser residues (S840 and S841) is necessary for signaling (9), prompting the question of whether this requirement is also bypassed in 1NM-PP1–sensitized mutants. As expected, the unphosphorylatable Ire1 (S840A,S841A) mutant was inactive in the absence and presence of 1NM-PP1 (Fig. 1C). In

6

contrast, the 1NM-PP1–sensitized Ire1 (L745A,S840A,S841A) and Ire1(L745G, S840A,S841A) mutants were significantly activated by 1NM-PP1, indicating that phosphorylation of S840 and S841 can be bypassed. Two trends are apparent: With DTT alone, relative to the parent mutant Ire1(S840A,S841A), signaling decreased as the side chain of residue 745 was pro-gressively reduced from Leu to Ala to Gly (Fig. 1C, bars 3, 5, and 7). This is con-sistent with the results of Fig. 1A, which

show an increasing sensitivity of the ac-tive-site to side-chain reduction at residue 745. Reciprocally, in the presence of 1NM-PP1, activation of 1NM-PP1–sensitized Ire1(S840A,S841A) mutants increased in the same order (Fig. 1C, bars 4, 6, and 8), perhaps because of progressively enhanced binding by 1NM-PP1 as the residue 745 side chain was trimmed back

1NM-PP1–sensitized Ire1 requires 1NM-PP1 as a cofactor for HAC1 mRNA splicing. We next confirmed that activa-

Trans-autophosphorylation by Ire1 of twoactivation-segment Ser residues (S840 and S841)is necessary for signaling (9), prompting thequestion of whether this requirement is alsobypassed in 1NM-PP1–sensitized mutants.As expected, the unphosphorylatable Ire1(S840A,S841A) mutant was inactive in the ab-sence and presence of 1NM-PP1 (Fig. 1C).In contrast, the 1NM-PP1–sensitized Ire1(L745A,S840A,S841A) and Ire1(L745G,S840A,S841A) mutants were significantly acti-vated by 1NM-PP1, indicating that phosphoryl-ation of S840 and S841 can be bypassed. Twotrends are apparent: With DTT alone, relative tothe parent mutant Ire1(S840A,S841A), signalingdecreased as the side chain of residue 745 wasprogressively reduced from Leu to Ala to Gly(Fig. 1C, bars 3, 5, and 7). This is consistent withthe results of Fig. 1A, which show an increasingsensitivity of the active-site to side-chain reduc-tion at residue 745. Reciprocally, in the presenceof 1NM-PP1, activation of 1NM-PP1–sensitizedIre1(S840A,S841A) mutants increased in thesame order (Fig. 1C, bars 4, 6, and 8), perhapsbecause of progressively enhanced binding by1NM-PP1 as the residue 745 side chain wastrimmed back.

1NM-PP1–sensitized Ire1 requires1NM-PP1 as a cofactor for HAC1 mRNAsplicing. We next confirmed that activation by1NM-PP1–sensitized Ire1 mutants followed theexpected pathway of activating HAC1 mRNAsplicing and Hac1 protein production. As expect-ed, UPR induction of wild-type cells with DTTcaused splicing of HAC1 mRNA, as revealed bynear-complete conversion of HAC1u to HAC1i

mRNA (Fig. 2A, lanes 1 and 2) (5); 1NM-PP1had no effect by itself or with DTT (Fig. 2A,lanes 3 and 4). In strict correlation with HAC1mRNA splicing, Hac1 protein accumulated sig-nificantly (Fig. 2C, lanes 2 and 4).

In contrast, cells expressing 1NM-PP1–sen-sitized Ire1(L745G) spliced HAC1 mRNAweakly with DTT alone, but displayed a signif-icant increase when 1NM-PP1 was also provid-ed (Fig. 2A, lanes 5 to 8). 1NM-PP1 alone didnot activateHAC1mRNA splicing. As expected,Hac1 protein levels correlated with splicing (e.g.,Fig. 2C, lanes 6 and 8). Activation by 1NM-PP1was even more pronounced in cells expressingthe Ire1(L745A,D828A) double mutant, whichneither spliced HAC1 mRNA nor producedHac1 protein with DTT alone, but exhibitedrobust splicing and Hac1 protein productionwhen provided both 1NM-PP1and DTT (Fig. 2,A and C, compare lanes 10 and 12).Importantly, steady-state levels of Ire1 pro-

teins were comparable in all experiments, ex-cluding as a trivial explanation for the observedeffects the possibility that 1NM-PP1–sensitizedIre1 mutants are only stably expressed in thepresence of 1NM-PP1 (Fig. 2D). Taken togeth-er, our data show that 1NM-PP1 permits butdoes not instruct 1NM-PP1–sensitized Ire1 mu-tants to splice HAC1 mRNA in response to ER

stress. Thus, in the genetic background of1NM-PP1–sensitized IRE1 alleles, 1NM-PP1is, in effect, a cofactor for signaling in the UPR.

1NM-PP1 uncouples the kinase andRNase activities of 1NM-PP1–sensitizedIre1. To rule out indirect effects, we assayedthe activities of the Ire1 mutants describedabove in vitro. The cytosolic portion of Ire1consisting of the kinase domain and RNasedomain (Ire1*) was expressed and purifiedfrom Escherichia coli (7). Using [�-32P]-la-beled ATP, we assayed wild-type Ire1* andIre1*(L745G) for autophosphorylation in thepresence or absence of 1NM-PP1. Wild-typeIre1* exhibited strong autophosphorylation,which was not inhibited by 1NM-PP1 (Fig. 3A,lanes 1 and 2). In contrast, Ire1*(L745G) ex-hibited markedly diminished autophosphoryl-ation, consistent with poor signaling byIre1(L745G) in vivo (Figs. 1 and 2), which wascompletely extinguished by 1NM-PP1 (Fig.3A, lane 3 and 4). This suggested that

Ire1*(L745G) is diminished in its ability to useATP as substrate and that 1NM-PP1 is, aspredicted, an ATP competitor.

In contrast, when provided with [�-32P]-la-beled N-6 benzyl ATP, an ATP analog with abulky substituent (Fig. 3E), Ire1*(L745G) dis-played enhanced autophosphorylation activityrelative to wild-type Ire1* (Fig. 3B, lane 1 and2). Therefore, the L745G mutation does not de-stroy Ire1’s catalytic function, but rather alters itssubstrate specificity from ATP to ATP analogshaving complementary features that permit bind-ing in the expanded active site.

To ask how the RNase activity of Ire1 isaffected by manipulation of its kinase domain,we incubated wild-type Ire1* and Ire1*(L745G)with a truncated in vitro transcribed HAC1 RNAsubstrate (HAC1600) and monitored the produc-tion of cleavage products (7). Wild-type Ire1*cleaved HAC1600 RNA at both splice junctions,producing the expected fragments correspondingto the singly and doubly cut species (Fig. 3C,

Fig. 1. (A to D) In vivo UPR assays using a UPR element–driven lacZ reporter. The relevantgenotypes and the presence or absence of 1NM-PP1 and the UPR inducer DTT are indicated. For(A) to (C), DTT was used in all assays. wt, wild type; a.u., arbitrary units.

Fig. 2. In vivo HAC1mRNA splicing andHac1 protein produc-tion. Northern blotsshowing (A) in vivoHAC1 mRNA splicingand (B) SCR1 ribosom-al RNA (rRNA) asloading control. West-ern blots showing (C)Hac1 and (D) Ire1 pro-tein levels. The arrow in(C), lane 6, calls atten-tion to the smallamount of Hac1 proteinpresent in these cells.

R E S E A R C H A R T I C L E

28 NOVEMBER 2003 VOL 302 SCIENCE www.sciencemag.org1534

7

tion by 1NM-PP1–sensitized Ire1 mutants followed the expected pathway of activat-ing HAC1 mRNA splicing and Hac1 pro-tein production. As expected, UPR induc-tion of wild-type cells with DTT caused splicing of HAC1 mRNA, as revealed by near-complete conversion of HAC1u to HAC1i mRNA (Fig. 2A, lanes 1 and 2) (5); 1NM-PP1 had no effect by itself or with DTT (Fig. 2A, lanes 3 and 4). In strict cor-relation with HAC1 mRNA splicing, Hac1 protein accumulated significantly (Fig. 2C, lanes 2 and 4)

In contrast, cells expressing 1NM-PP1–sensitized Ire1(L745G) spliced HAC1 mRNA weakly with DTT alone, but dis-played a significant increase when 1NM-PP1 was also provided (Fig. 2A, lanes 5 to 8). 1NM-PP1 alone did not activate HAC1 mRNA splicing. As expected, Hac1 protein levels correlated with splicing (e.g., Fig. 2C, lanes 6 and 8). Activation by 1NM-PP1 was even more pronounced in cells expressing the Ire1(L745A,D828A) dou-ble mutant, which neither spliced HAC1 mRNA nor produced Hac1 protein with DTT alone, but exhibited robust splicing and Hac1 protein production when pro-vided both 1NM-PP1and DTT (Fig. 2, A and C, compare lanes 10 and 12).

Importantly, steady-state levels of Ire1 proteins were comparable in all experi-ments, excluding as a trivial explanation for the observed effects the possibility that 1NM-PP1–sensitized Ire1 mutants are only stably expressed in the presence of 1NM-PP1 (Fig. 2D). Taken together, our data show that 1NM-PP1 permits but does not instruct 1NM-PP1–sensitized Ire1 mutants to splice HAC1 mRNA in re-sponse to ER stress. Thus, in the genetic background of 1NM-PP1–sensitized IRE1 alleles, 1NM-PP1 is, in effect, a cofactor for signaling in the UPR

1NM-PP1 uncouples the kinase and RNase activities of 1NM-PP1–sensi-tized Ire1. To rule out indirect effects, we assayed the activities of the Ire1 mutants described above in vitro. The cytosolic

portion of Ire1 consisting of the kinase domain and RNase domain (Ire1*) was expressed and purified from Escherichia coli (7). Using γ-32P]-labeled ATP, we as-sayed wild-type Ire1* and Ire1*(L745G) for autophosphorylation in the presence or absence of 1NM-PP1. Wild-type Ire1* ex-hibited strong autophosphorylation, which was not inhibited by 1NM-PP1 (Fig. 3A, lanes 1 and 2). In contrast, Ire1*(L745G) exhibited markedly diminished autophos-phorylation, consistent with poor signal-ing by Ire1(L745G) in vivo (Figs. 1 and 2), which was completely extinguished by 1NM-PP1 (Fig. 3A, lane 3 and 4). This suggested that Ire1*(L745G) is dimin-ished in its ability to use ATP as substrate and that 1NM-PP1 is, as predicted, an ATP competitor.

In contrast, when provided with [γ-32P]-labeled N-6 benzyl ATP, an ATP analog with a bulky substituent (Fig. 3E), Ire1*(L745G) displayed enhanced auto-phosphorylation activity relative to wild-type Ire1* (Fig. 3B, lane 1 and 2). There-fore, the L745G mutation does not destroy Ire1’s catalytic function, but rather alters its substrate specificity from ATP to ATP ana-logs having complementary features that permit binding in the expanded active site.

To ask how the RNase activity of Ire1 is affected by manipulation of its kinase domain, we incubated wild-type Ire1* and Ire1*(L745G) with a truncated in vitro transcribed HAC1 RNA substrate (HAC1

600) and monitored the production

of cleavage products (7). Wild-type Ire1* cleaved HAC1

600 RNA at both splice junc-

tions, producing the expected fragments corresponding to the singly and doubly cut species (Fig. 3C, lane 2). The cleavage re-action was not affected by 1NM-PP1 (Fig. 3C, lane 3). In contrast, Ire1*(L745G) was completely inactive, whereas the addition of 1NM-PP1, at the same concentration that was shown in Fig. 3A (lane 4) to com-pletely inhibit the kinase activity, restored RNase activity significantly (Fig. 3C, lanes 4 and 5), strongly supporting the no-

8

tion that 1NM-PP1 uncouples the kinase and RNase activities of 1NM-PP1–sensi-tized Ire1.

We have previously shown that cleav-age of HAC1 mRNA by wild-type Ire1* is stimulated by adenosine 5′-diphosphate (ADP) (7). This is consistent with the no-tion derived from in vivo studies that Ire1 requires an active kinase domain, because we know that recombinant wild-type Ire1* is already phosphorylated, thereby allevi-ating a necessity for de novo phosphory-lation in vitro. ADP efficiently stimulated the RNase activity of Ire1* (Fig. 3D, lane 1) but not that of Ire1*(L745G) (Fig. 3D, lane 2).

For wild-type Ire1 and 1NM-PP1–sen-sitized Ire1(L745G), ADP and 1NM-PP1, respectively, function as stimulatory co-factors. Therefore, we asked whether a

conformational change occurs in response to cofactor bind-ing by analyzing Ire1* and Ire1*(L745G) through glycerol-gradient velocity sedimentation. Ire1* sedimented as a broad peak

in the gradient (Fig. 4). Upon the addition of ADP (Fig. 4B), but not 1NM-PP1 (Fig. 4C), the peak of Ire1* shifted as a result of faster sedimentation. Conversely, the peak of Ire1*(L745G) shifted by a correspond-ing amount upon the addition of 1NM-PP1 (Fig. 4F) but not upon the addition of ADP (Fig. 4E). This shift may be indicative of a conformational change and/or change in the oligomeric state of the enzymes in response to the binding of their respective cofactors.

1NM-PP1–sensitized, kinase-dead Ire1 signals a canonical UPR. To ask whether bypassing the Ire1 kinase activity with 1NM-PP1 allows the induction of a full UPR, we profiled mRNA expression on yeast genomic microarrays. To this end, mRNA from wildtype IRE1, Δire1, and IRE1(L745A,D828A) cells was isolated

lane 2). The cleavage reaction was not affectedby 1NM-PP1 (Fig. 3C, lane 3). In contrast,Ire1*(L745G) was completely inactive, whereasthe addition of 1NM-PP1, at the same concen-tration that was shown in Fig. 3A (lane 4) tocompletely inhibit the kinase activity, restoredRNase activity significantly (Fig. 3C, lanes 4 and5), strongly supporting the notion that 1NM-PP1uncouples the kinase and RNase activities of1NM-PP1–sensitized Ire1.

We have previously shown that cleavage ofHAC1 mRNA by wild-type Ire1* is stimulatedby adenosine 5�-diphosphate (ADP) (7). This isconsistent with the notion derived from in vivostudies that Ire1 requires an active kinase do-main, because we know that recombinant wild-type Ire1* is already phosphorylated, therebyalleviating a necessity for de novo phosphoryl-ation in vitro. ADP efficiently stimulated theRNase activity of Ire1* (Fig. 3D, lane 1) but notthat of Ire1*(L745G) (Fig. 3D, lane 2).

For wild-type Ire1 and 1NM-PP1–sensitizedIre1(L745G), ADP and 1NM-PP1, respectively,function as stimulatory cofactors. Therefore, weasked whether a conformational change occursin response to cofactor binding by analyzingIre1* and Ire1*(L745G) through glycerol-gradi-

ent velocity sedimentation. Ire1* sedimented asa broad peak in the gradient (Fig. 4). Upon theaddition of ADP (Fig. 4B), but not 1NM-PP1(Fig. 4C), the peak of Ire1* shifted as a result offaster sedimentation. Conversely, the peak ofIre1*(L745G) shifted by a correspondingamount upon the addition of 1NM-PP1 (Fig. 4F)but not upon the addition of ADP (Fig. 4E). Thisshift may be indicative of a conformationalchange and/or change in the oligomeric state ofthe enzymes in response to the binding of theirrespective cofactors.

1NM-PP1–sensitized, kinase-dead Ire1signals a canonical UPR. To ask whetherbypassing the Ire1 kinase activity with 1NM-PP1 allows the induction of a full UPR, weprofiled mRNA expression on yeast genomicmicroarrays. To this end, mRNA from wild-type IRE1, �ire1, and IRE1(L745A,D828A)cells was isolated at different points in timeafter the addition of 1NM-PP1 and DTT. cD-NAs derived from these mRNA populationswere labeled with different fluorescent dyes,mixed pairwise, and hybridized to microarrays(20). Scatter plots of pairwise comparisons ofmRNA expression levels are shown inFig. 5; the x-y scatter plots represent relative

expressions for every mRNA between the spec-ified genotypes.

The comparison of wild-type IRE1-expressing cells with �ire1 cells revealed adistinct set of up-regulated UPR target genes inwild-type IRE1-expressing cells (Fig. 5A;genes that were up-regulated more than twofoldare shown in pink). These IRE1-dependentgenes include previously described UPR tran-scriptional targets (21, 22). Notably, the sameset of (pink-colored) UPR target genes wasup-regulated to a similar magnitude inIRE1(L745A,D828A)-expressing cells relativeto �ire1 cells (Fig. 5B), suggesting that a ca-nonical UPR was induced in the 1NM-PP1–sensitized, kinase-dead mutant. This conclusionwas confirmed by the tight superimposition ofthe scatter diagonals of both UPR targets andthe rest of the transcriptome evident in theexpression profiles of wild-type IRE1- com-pared with IRE1(L745A,D828A)-expressingcells (Fig. 5C).

Building an instructive UPR switch.The1NM-PP1–sensitized IRE1mutants described sofar act permissively, because activation requiresboth 1NM-PP1 and protein-misfolding agents.This indicates that an unfolded-protein signalneeds to be received by the ER lumenal domainfor activation. We asked if we could bypass thisrequirement using a constitutively on version ofIRE1 (called IRE1C) isolated previously throughrandom mutagenesis of the ER lumenal domain(18, 21). IRE1C significantly induced the UPRwithout DTT, resulting in about 75% of themaximal activity of wild-type IRE1 with DTT(Fig. 1D, bars 2 and 5). We predicted that achimera of the IRE1C and 1NM-PP1–sensitized,kinase-dead IRE1(L745A,D828A) mutantswould activate the UPR with 1NM-PP1 alone,even in the absence of DTT. The data in Fig. 1D(bars 13 to 16) show that this prediction holds

Fig. 3. In vitro assays of Ire1 mutant proteins. Ire1* autophosphorylationassays with [�-32P] ATP (A) and [�-32P]-labeled N-6 benzyl ATP (B). Phos-phorylated Ire1* proteins (wild-type and L745G) are indicated. Ire1* RNaseassays (against in vitro transcribed HAC1600 RNA) with stimulatory cofac-tors 1NM-PP1 (C) and ADP (D). Icons in (C) and (D) indicate HAC1600 RNAcleavage products, denoting all combinations of singly and doubly cut RNA.(E) The chemical structures of 1NM-PP1 and N-6 benzyl ATP are shown.

Fig. 4. Glycerol-densitygradient velocity sedi-mentation of wild-typeand mutant Ire1*. Ire1*proteins (A to C) andIre1*(L745G) proteins(D to F) were incubatedin the presence of ADP[(B) and (E)] or 1NM-PP1[(C) and (F)] and subject-ed to velocity sedimen-tation analysis. Sedi-mentation is from leftto right.



























Fig. 4. Glycerol-density gradient velocity sedimentation of wild-type and mutant Ire1*. Ire1* proteins (A to C) and Ire1*(L745G) proteins (D to F) were incubated in the presence of ADP [(B) and(E)] or 1NM-PP1 [(C) and(F)] and subjected to velocity sedimentation analysis. Sedimentation is from left to right.

Fig. 3. In vitro assays of Ire1 mutant proteins. Ire1* autophosphorylation assays with [γ-32P] ATP (A) and[ γ-32P]-labeled N-6 benzyl ATP (B). Phosphorylated Ire1* proteins (wild-type and L745G) are indicated. Ire1* RNase assays (against in vitro transcribed HAC1

600 RNA) with

stimulatory cofactors 1NM-PP1 (C) and ADP (D). Icons in (C) and (D) indicate HAC1

600 RNA cleavage

products, denoting all combinations of singly and doubly cut RNA. (E) The chemical structures of 1NM-PP1 and N-6 benzyl ATP are shown.

9

at different points in time after the addi-tion of 1NM-PP1 and DTT. cDNAs de-rived from these mRNA populations were

labeled with different fluorescent dyes, mixed pairwise, and hybridized to mi-croarrays (20). Scatter plots of pairwise

true. Whereas wild-type IRE1 requires DTT forinduction and is immune to 1NM-PP1 (Fig. 1D,bars 1 to 4), and IRE1(L745A,D828A) requiresboth DTT and 1NM-PP1 (Fig. 1D, bars 9 to 12),the chimeric IRE1C (L745A,D828A) requiresonly 1NM-PP1 for activation, irrespective of theaddition of DTT (Fig. 1D, bars 13 to 16). Thus,IRE1C (L745A,D828A) is an instructive switchthat turns on the UPR upon the addition of1NM-PP1 alone (i.e., even in the absence ofER stress).

Summary and implications. Since ourdiscovery that Ire1’s RNase initiates splic-ing of HAC1 mRNA, the function of itskinase domain has remained an enigma (7 ).Although mutational analyses have indicat-ed a requirement for an enzymatically ac-tive kinase and have defined activation-segment phosphorylation sites (9), to dateno targets besides Ire1 itself have beenidentified. Here, we report that both thekinase activity and activation-segment phos-

phorylation can be bypassed if the smallATP-mimic 1NM-PP1 is provided to a mu-tant enzyme to which it can bind. Instead ofinhibiting the function served by ATP,1NM-PP1 rectifies the signaling defect of1NM-PP1–sensitized Ire1 mutants, leadingto the induction of a canonical UPR (21).Thus, ligand binding to the kinase domain,rather than a phosphotransfer function me-diated by the kinase activity, is required forRNase activation. The kinase module ofIre1, therefore, uses an unprecedentedmechanism to propagate the UPR signal.One model that could explain our data is

proposed in Fig. 6B. According to thismodel, autophosphorylation of wild-typeIre1 occurs in trans between Ire1 moleculesafter they have been released from theirchaperone anchors and have oligomerizedby lateral diffusion in the plane of the ERmembrane (9, 23). Trans-autophosphoryl-ation of the activation segment locks theactivation segment in the “swung-out” state(24, 25 ), and fully opens the active site,allowing ADP or ATP to bind efficiently,leading in turn to conformational changesthat activate the RNase domain.In contrast, the binding of 1NM-PP1 to

the active site of 1NM-PP1–sensitized Ire1occurs even in the absence of phosphoryl-ation of the activation segment. It is plau-sible that because of its small size, 1NM-PP1 can bypass the “closed” activation seg-ment, which is proposed to occlude accessto the active site for the larger ADP andATP molecules (Fig. 6A). Alternatively,1NM-PP1 may enter the active site whenthe unphosphorylated activation segmentopens transiently, which occurs with lowfrequency in other kinases (26). If the off-rate of 1NM-PP1 binding to 1NM-PP1–sensitized Ire1 is slow compared to that ofADP and ATP, most mutant Ire1 moleculescould be trapped in a drug-bound state.Based on either scenario, we propose thatdrug binding alone causes conformationalrearrangements in the kinase that activatethe RNase. The nucleotide-bound state ofphosphorylated wild-type Ire1, therefore,mimics the 1NM-PP1–bound state of un-phosphorylated 1NM-PP1–sensitized Ire1.The model discussed so far does not explain

why 1NM-PP1 should not bind to individual1NM-PP1–sensitized Ire1 molecules and acti-vate them, even without oligomerization inducedby activation of the UPR through the ER lume-nal domain. Two possibilities could account forthe observations: (i) The binding of 1NM-PP1may be enhanced in oligomerized, chaperone-free mutants, or (ii) 1NM-PP1 may bind equallywell to chaperone-anchored and chaperone-free(or constitutive-on) Ire1, but oligomerizationmay be required because interaction betweenneighboring molecules is required for the con-formational change that activates the RNase

Fig. 5. (A to C) x-y scatter plot analysis of mRNA abundance. The plots (log10) compare yeastgenomic microarray analyses between the genotypes indicated. In each case, mRNA was purifiedfrom cells that were provided both DTT and 1NM-PP1. Pink dots represent genes with expressionmore than two times as high in wild-type cells as in �ire1 cells.

Fig. 6.Model of activation of 1NM-PP1–sensitized and wild-type Ire1. For both proteins, chaperonedissociation resulting from unfolded-protein accumulation causes oligomerization in the plane ofthe ER membrane, which is a precondition for further activation. For mutant Ire1 (A), 1NM-PP1binding to the inactive kinase domain active site causes a conformational change, which activatesthe RNase. For wild-type Ire1 (B), trans-autophosphorylation, with ATP, of the activation segmentfully opens the kinase domain for binding of ADP or ATP, which activates the RNase.















true. Whereas wild-type IRE1 requires DTT forinduction and is immune to 1NM-PP1 (Fig. 1D,bars 1 to 4), and IRE1(L745A,D828A) requiresboth DTT and 1NM-PP1 (Fig. 1D, bars 9 to 12),the chimeric IRE1C (L745A,D828A) requiresonly 1NM-PP1 for activation, irrespective of theaddition of DTT (Fig. 1D, bars 13 to 16). Thus,IRE1C (L745A,D828A) is an instructive switchthat turns on the UPR upon the addition of1NM-PP1 alone (i.e., even in the absence ofER stress).

Summary and implications. Since ourdiscovery that Ire1’s RNase initiates splic-ing of HAC1 mRNA, the function of itskinase domain has remained an enigma (7 ).Although mutational analyses have indicat-ed a requirement for an enzymatically ac-tive kinase and have defined activation-segment phosphorylation sites (9), to dateno targets besides Ire1 itself have beenidentified. Here, we report that both thekinase activity and activation-segment phos-

phorylation can be bypassed if the smallATP-mimic 1NM-PP1 is provided to a mu-tant enzyme to which it can bind. Instead ofinhibiting the function served by ATP,1NM-PP1 rectifies the signaling defect of1NM-PP1–sensitized Ire1 mutants, leadingto the induction of a canonical UPR (21).Thus, ligand binding to the kinase domain,rather than a phosphotransfer function me-diated by the kinase activity, is required forRNase activation. The kinase module ofIre1, therefore, uses an unprecedentedmechanism to propagate the UPR signal.One model that could explain our data is

proposed in Fig. 6B. According to thismodel, autophosphorylation of wild-typeIre1 occurs in trans between Ire1 moleculesafter they have been released from theirchaperone anchors and have oligomerizedby lateral diffusion in the plane of the ERmembrane (9, 23). Trans-autophosphoryl-ation of the activation segment locks theactivation segment in the “swung-out” state(24, 25 ), and fully opens the active site,allowing ADP or ATP to bind efficiently,leading in turn to conformational changesthat activate the RNase domain.In contrast, the binding of 1NM-PP1 to

the active site of 1NM-PP1–sensitized Ire1occurs even in the absence of phosphoryl-ation of the activation segment. It is plau-sible that because of its small size, 1NM-PP1 can bypass the “closed” activation seg-ment, which is proposed to occlude accessto the active site for the larger ADP andATP molecules (Fig. 6A). Alternatively,1NM-PP1 may enter the active site whenthe unphosphorylated activation segmentopens transiently, which occurs with lowfrequency in other kinases (26). If the off-rate of 1NM-PP1 binding to 1NM-PP1–sensitized Ire1 is slow compared to that ofADP and ATP, most mutant Ire1 moleculescould be trapped in a drug-bound state.Based on either scenario, we propose thatdrug binding alone causes conformationalrearrangements in the kinase that activatethe RNase. The nucleotide-bound state ofphosphorylated wild-type Ire1, therefore,mimics the 1NM-PP1–bound state of un-phosphorylated 1NM-PP1–sensitized Ire1.The model discussed so far does not explain

why 1NM-PP1 should not bind to individual1NM-PP1–sensitized Ire1 molecules and acti-vate them, even without oligomerization inducedby activation of the UPR through the ER lume-nal domain. Two possibilities could account forthe observations: (i) The binding of 1NM-PP1may be enhanced in oligomerized, chaperone-free mutants, or (ii) 1NM-PP1 may bind equallywell to chaperone-anchored and chaperone-free(or constitutive-on) Ire1, but oligomerizationmay be required because interaction betweenneighboring molecules is required for the con-formational change that activates the RNase

Fig. 5. (A to C) x-y scatter plot analysis of mRNA abundance. The plots (log10) compare yeastgenomic microarray analyses between the genotypes indicated. In each case, mRNA was purifiedfrom cells that were provided both DTT and 1NM-PP1. Pink dots represent genes with expressionmore than two times as high in wild-type cells as in �ire1 cells.

Fig. 6.Model of activation of 1NM-PP1–sensitized and wild-type Ire1. For both proteins, chaperonedissociation resulting from unfolded-protein accumulation causes oligomerization in the plane ofthe ER membrane, which is a precondition for further activation. For mutant Ire1 (A), 1NM-PP1binding to the inactive kinase domain active site causes a conformational change, which activatesthe RNase. For wild-type Ire1 (B), trans-autophosphorylation, with ATP, of the activation segmentfully opens the kinase domain for binding of ADP or ATP, which activates the RNase.



10

comparisons of mRNA expression levels are shown in Fig. 5; the x-y scatter plots represent relative expressions for every mRNA between the specified genotypes.

The comparison of wild-type IRE1- expressing cells with Δire1 cells revealed a distinct set of up-regulated UPR target genes in wild-type IRE1-expressing cells (Fig. 5A; genes that were up-regulated more than twofold are shown in pink). These IRE1-dependent genes include previously described UPR transcriptional targets (21, 22). Notably, the same set of (pink-colored) UPR target genes was up-regulated to a similar magnitude in IRE1(L745A,D828A)-expressing cells relative to Δire1 cells (Fig. 5B), suggest-ing that a canonical UPR was induced in the 1NM-PP1–sensitized, kinase-dead mutant. This conclusion was confirmed by the tight superimposition of the scatter diagonals of both UPR targets and the rest of the transcriptome evident in the expres-sion profiles of wild-type IRE1- compared with IRE1(L745A,D828A)-expressing cells (Fig. 5C).

Building an instructive UPR switch.The 1NM-PP1–sensitized IRE1 mutants described so far act permissively, be-cause activation requires both 1NM-PP1 and protein-misfolding agents. This in-dicates that an unfolded-protein signal needs to be received by the ER lumenal domain for activation. We asked if we could bypass this requirement using a constitutively on version of IRE1 (called IRE1C) isolated previously through ran-dom mutagenesis of the ER lumenal do-main (18, 21). IRE1C significantly induced the UPR without DTT, resulting in about 75% of the maximal activity of wild-type IRE1 with DTT (Fig. 1D, bars 2 and 5). We predicted that a chimera of the IRE1C

and 1NM-PP1–sensitized, kinase-dead IRE1(L745A,D828A) mutants would acti-vate the UPR with 1NM-PP1 alone, even in the absence of DTT. The data in Fig. 1D (bars 13 to 16) show that this prediction holds true. Whereas wild-type IRE1 re-

quires DTT for induction and is immune to 1NM-PP1 (Fig. 1D, bars 1 to 4), and IRE1(L745A,D828A) requires both DTT and 1NM-PP1 (Fig. 1D, bars 9 to 12), the chimeric IRE1C (L745A,D828A) requires only 1NM-PP1 for activation, irrespective of the addition of DTT (Fig. 1D, bars 13 to 16). Thus, IRE1C (L745A,D828A) is an instructive switch that turns on the UPR upon the addition of 1NM-PP1 alone (i.e., even in the absence of ER stress).

Summary and implications. Since our discovery that Ire1’s RNase initiates splicing of HAC1 mRNA, the function of its kinase domain has remained an enigma (7). Although mutational analyses have indicated a requirement for an enzymati-cally active kinase and have defined acti-vation-segment phosphorylation sites (9), to date no targets besides Ire1 itself have been identified. Here, we report that both the kinase activity and activation-segment phosphorylation can be bypassed if the small ATP-mimic 1NM-PP1 is provided to a mutant enzyme to which it can bind. In-stead of inhibiting the function served by ATP, 1NM-PP1 rectifies the signaling de-fect of 1NM-PP1–sensitized Ire1 mutants, leading to the induction of a canonical UPR (21). Thus, ligand binding to the ki-nase domain, rather than a phosphotrans-fer function mediated by the kinase activ-ity, is required for RNase activation. The kinase module of Ire1, therefore, uses an unprecedented mechanism to propagate the UPR signal.

One model that could explain our data is proposed in Fig. 6B. According to this model, autophosphorylation of wild-type Ire1 occurs in trans between Ire1 mol-ecules after they have been released from their chaperone anchors and have oligo-merized by lateral diffusion in the plane of the ER membrane (9, 23). Trans-autophos-phorylation of the activation segment locks the activation segment in the “swung-out” state (24, 25), and fully opens the active site, allowing ADP or ATP to bind effi-ciently, leading in turn to conformational

11

changes that activate the RNase domain. In contrast, the binding of 1NM-PP1

to the active site of 1NM-PP1–sensitized Ire1 occurs even in the absence of phos-phorylation of the activation segment. It is plausible that because of its small size, 1NMPP1 can bypass the “closed” activa-tion segment, which is proposed to occlude access to the active site for the larger ADP and ATP molecules (Fig. 6A). Alterna-tively, 1NM-PP1 may enter the active site when the unphosphorylated activation seg-ment opens transiently, which occurs with low frequency in other kinases (26). If the offrate of 1NM-PP1 binding to 1NM-PP1– sensitized Ire1 is slow compared to that of ADP and ATP, most mutant Ire1 molecules could be trapped in a drug-bound state. Based on either scenario, we propose that drug binding alone causes conformational rearrangements in the kinase that activate the RNase. The nucleotide-bound state of phosphorylated wild-type Ire1, therefore, mimics the 1NM-PP1–bound state of un-phosphorylated 1NM-PP1–sensitized Ire1.

The model discussed so far does not explain why 1NM-PP1 should not bind to individual 1NM-PP1–sensitized Ire1 molecules and activate them, even without oligomerization induced by activation of the UPR through the ER lumenal domain. Two possibilities could account for the ob-servations: (i) The binding of 1NM-PP1 may be enhanced in oligomerized, chap-erone-free mutants, or (ii) 1NM-PP1 may bind equally well to chaperone-anchored and chaperone-free (or constitutive-on) Ire1, but oligomerization may be required because interaction between neighboring molecules is required for the conforma-tional change that activates the RNase func-tion. Currently, our data do not allow us to distinguish between these possibilities.

Our findings provoke the question of what the “natural” stimulatory ligand of Ire1’s kinase domain may be. One likely candidate is ADP, which is generated by Ire1 from ATP during the autophosphory-lation reaction. In vitro, ADP is a better

activator of Ire1 than ATP (7). Ire1 would therefore be “self priming,” generating an efficient activator already bound to its ac-tive site. In many professional secretory cells such as the β-cells of the endocrine pancreas (which contain high concentra-tions of Ire1), ADP levels rise (and ATP levels decline) temporarily in proportion to nutritional stress (27). ADP therefore is physiologically poised to serve as a co-factor that could signal a starvation state. It is known that protein folding becomes inefficient as the nutritional status of cells declines, triggering the UPR (28). Thus, in the face of ATP depletion, ADP-mediated conformational changes might increase the dwell time of activated Ire1. As such, Ire1 could have evolved this regulatory mechanism to monitor the energy balance of the cell and couple this information to activation of the UPR.

Although unprecedented in proteins containing active kinase domains, our findings are consistent with previous ob-servations of RNase L, which is closely related to Ire1, bearing a kinase-like do-main followed C-terminally by an RNase domain (29, 30). Similar to Ire1, RNase L is activated by dimerization (31). How-ever, RNase L’s kinase domain is naturally inactive (29), whereas its RNase activity is stimulated by adenosine nucleotide bind-ing to the kinase domain. Therefore, like Ire1, the ligand-occupied kinase domain of RNase L serves as a module that par-ticipates in activation and regulation of the RNase function. Insights gained from Ire1 and RNase L may extend to other proteins containing kinase or enzymatically inac-tive pseudokinase domains (32).

function. Currently, our data do not allow us todistinguish between these possibilities.

Our findings provoke the question of whatthe “natural” stimulatory ligand of Ire1’s ki-nase domain may be. One likely candidate isADP, which is generated by Ire1 from ATPduring the autophosphorylation reaction. Invitro, ADP is a better activator of Ire1 thanATP (7). Ire1 would therefore be “self prim-ing,” generating an efficient activator alreadybound to its active site. In many professionalsecretory cells such as the �-cells of theendocrine pancreas (which contain high con-centrations of Ire1), ADP levels rise (andATP levels decline) temporarily in proportionto nutritional stress (27). ADP therefore isphysiologically poised to serve as a cofactorthat could signal a starvation state. It isknown that protein folding becomes ineffi-cient as the nutritional status of cells declines,triggering the UPR (28). Thus, in the face ofATP depletion, ADP-mediated conforma-tional changes might increase the dwell timeof activated Ire1. As such, Ire1 could haveevolved this regulatory mechanism to moni-tor the energy balance of the cell and couplethis information to activation of the UPR.

Although unprecedented in proteins con-taining active kinase domains, our findingsare consistent with previous observations ofRNase L, which is closely related to Ire1,bearing a kinase-like domain followed C-terminally by an RNase domain (29, 30).

Similar to Ire1, RNase L is activated bydimerization (31). However, RNase L’s ki-nase domain is naturally inactive (29), where-as its RNase activity is stimulated by adeno-sine nucleotide binding to the kinase domain.Therefore, like Ire1, the ligand-occupied ki-nase domain of RNase L serves as a modulethat participates in activation and regulationof the RNase function. Insights gained fromIre1 and RNase L may extend to other pro-teins containing kinase or enzymatically in-active pseudokinase domains (32).

References and Notes1. M. J. Gething, J. Sambrook, Nature 355, 33 (1992).2. F. J. Stevens, Y. Argon, Semin. Cell Dev. Biol. 10, 443(1999).

3. C. Patil, P. Walter, Curr. Opin. Cell Biol. 13, 349 (2001).4. J. S. Cox, C. E. Shamu, P. Walter, Cell 73, 1197 (1993).5. J. S. Cox, P. Walter, Cell 87, 391 (1996).6. U. Ruegsegger, J. H. Leber, P. Walter, Cell 107, 103(2001).

7. C. Sidrauski, P. Walter, Cell 90, 1031 (1997).8. C. Sidrauski, J. S. Cox, P. Walter, Cell 87, 405 (1996).9. C. E. Shamu, P. Walter, EMBO J. 15, 3028 (1996).10. A. Weiss, J. Schlessinger, Cell 94, 277 (1998).11. K. Mori, W. Ma, M. J. Gething, J. Sambrook, Cell 74,

743 (1993).12. K. Shah, Y. Liu, C. Deirmengian, K. M. Shokat, Proc.

Natl. Acad. Sci. U.S.A. 94, 3565 (1997).13. A. C. Bishop et al., Curr. Biol. 8, 257 (1998).14. E. L. Weiss, A. C. Bishop, K. M. Shokat, D. G. Drubin,

Nature Cell Biol. 2, 677 (2000).15. A. C. Bishop et al., Nature 407, 395 (2000).16. Single-letter abbreviations for the amino acid resi-

dues are as follows: A, Ala; D, Asp; G, Gly; L, Leu; andS, Ser.

17. A. S. Carroll, A. C. Bishop, J. L. DeRisi, K. M. Shokat, E. K.O’Shea, Proc. Natl. Acad. Sci. U.S.A. 98, 12578 (2001).

18. F. R. Papa, C. Zhang, K. Shokat, P. Walter, data not shown.19. M. Huse, J. Kuriyan, Cell 109, 275 (2002).20. J. L. DeRisi, V. R. Iyer, P. O. Brown, Science 278, 680 (1997).21. K. J. Travers et al., Cell 101, 249 (2000).22. Materials and methods are available as supporting

material on Science Online.23. F. Urano, A. Bertolotti, D. Ron, J. Cell Sci. 113, 3697 (2000).24. S. R. Hubbard, M. Mohammadi, J. Schlessinger, J. Biol.

Chem. 273, 11987 (1998).25. T. Schindler et al., Mol. Cell 3, 639 (1999).26. M. Porter, T. Schindler, J. Kuriyan, W. T. Miller, J. Biol.

Chem. 275, 2721 (2000).27. F. Schuit, K. Moens, H. Heimberg, D. Pipeleers, J. Biol.

Chem. 274, 32803 (1999).28. R. J. Kaufman et al., Nature Rev. Mol. Cell Biol. 3, 411

(2002).29. B. Dong, M. Niwa, P. Walter, R. H. Silverman, RNA 7,

361 (2001).30. S. Naik, J. M. Paranjape, R. H. Silverman, Nucleic Acids

Res. 26, 1522 (1998).31. B. Dong, R. H. Silverman,Nucleic Acids Res. 27, 439 (1999).32. M. Kroiher, M. A. Miller, R. E. Steele, Bioessays 23, 69

(2001).33. We thank M. Stark for constructing the IRE1C allele, S.

Ullrich for his initial effort in constructing Ire1 mu-tants, and J. Kuriyan and members of the Walter andShokat labs for valuable discussions. Supported bygrants from the NIH to P.W. (GM32384) and K.S.(AI44009), and postdoctoral support from the UCSFMolecular Medicine Program (to F.P.). P.W. receivessupport as an investigator of the Howard HughesMedical Institute.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/1090031/DC1Materials and MethodsMicroarray Excel Spreadsheets S1 to S3References and Notes

4 August 2003; accepted 1 October 2003Published online 16 October 2003;10.1126/science.1090031Include this information when citing this paper.

REPORTSObservation of the Inverse

Doppler EffectN. Seddon* and T. Bearpark

We report experimental observation of an inverse Doppler shift, in which thefrequency of a wave is increased on reflection from a receding boundary. Thiscounterintuitive effect has been produced by reflecting a wave from a movingdiscontinuity in an electrical transmission line. Doppler shifts produced by thissystem can be varied in a reproducible manner by electronic control of thetransmission line and are typically five orders of magnitude greater than thoseproduced by solid objects with kinematic velocities. Potential applicationsinclude the development of tunable and multifrequency radiation sources.

The Doppler effect is the well-known phe-nomenon by which the frequency of a wave isshifted according to the relative velocity of

the source and the observer (1, 2). Our con-ventional understanding of the Doppler ef-fect, from the schoolroom to everyday expe-rience of passing vehicles, is that increasedfrequencies are measured when a source andobserver approach each other. Applicationsof the effect are widely established and in-clude radar, laser vibrometry, blood flowmeasurement, and the search for new astro-

nomical objects. The inverse Doppler effectrefers to frequency shifts that are in the op-posite sense to those described above; forexample, increased frequencies would bemeasured on reflection of waves from a re-ceding boundary. Demonstration of thiscounterintuitive phenomenon requires a fun-damental change in the way that radiation isreflected from a moving boundary, and, de-spite a wide range of theoretical work thatspans the past 60 years, the effect has notpreviously been verified experimentally.

Although inverse Doppler shifts havebeen predicted to occur in particular disper-sive media (3–8) and in the near zone ofthree-dimensional dipoles (9–11), theseschemes are difficult to implement experi-mentally and have not yet been realized.However, recent advances in the design ofcomposite condensed media (metamaterials)offers new and exciting possibilities for thecontrol of radiation. In particular, materialswith a negative refractive index (NRI) (12–14) and the use of shock discontinuities inphotonic crystals (15) and transmission lines

Optics and Laser Technology Department, AdvancedTechnology Centre, BAE Systems, Post Office Box 5,Filton, Bristol BS34 7QW, UK.




12

statistically and experimentally (1). How-ever, characterizing polymorphisms locat-










































































































creased, and bodymasswas lower when squirrels

were feeding on these items as compared with

feeding on seed (26, 27). Lastly, food supple-

mentation experiments of American reds have

failed to produce increases in litter size or a

second litter, providing strong evidence that the

increased reproductive investment observed in

our study cannot be triggered by increased

energy availability alone (17, 28, 29). We hy-

pothesize that, rather than following a resource-

tracking strategy where reproductive investment

is determined by current resource amounts, re-

productive rates are driven by future fitness

payoffs. During years of low seed production,

competition among juveniles for available re-

sources is intense, and, although litter augmenta-

tion experiments in American reds show that

females are capable of supporting larger litters

(30), they refrain from doing so because off-

spring recruitment is low (30). However, when

mast years occur, competition among juveniles is

reduced, and females produce more offspring,

which successfully recruit into the population

(31, 32). Further, the increased production of young

by females in mast years does not come with

any obvious cost to the female, because over-

winter survival is not reduced after years of

increased reproductive investment (for Amer-

ican reds, offspring production in the previous

year versus proportion of adult females surviving

to spring has slope = –0.023 ± 0.034, t15 = –0.7,

and P = 0.51; for Eurasian reds, proportion of

estrous females in the previous year versus adult

female survival to spring has slope = 0.37 ± 0.27,

t15 = 1.4, and P = 0.19).

If masting has evolved as a swamp-and-

starve adaptation against seed predation, then

anticipatory reproduction and population growth

represent a potent counteradaptation by the pred-

ators. Given that increased reproductive output

in these systems coincides with low current but

high future resources, our results suggest that

reproductive investment in these systems is more

responsive to future fitness prospects than present

energetic constraints. The evolution of seedmast-

ing in trees is also driven by the survival pros-

pects for progeny rather than simple resource

tracking, suggesting an intriguing parallel in re-

productive strategies of trees and the predators

that consume their seed.

References and Notes1. J. L. Gittleman, S. D. Thompson, Am. Zool. 28, 863

(1988).

2. T. H. Clutton-Brock, Reproductive Success (Univ. of

Chicago Press, Chicago, 1988).

3. T. H. Clutton-Brock, The Evolution of Parental Care

(Princeton Univ. Press, Princeton, NJ, 1999).

4. G. Caughley, in Theoretical Ecology, R. M. May, Ed.

(Blackwell, Oxford, 1981), ch. 6.

5. P. Bayliss, D. Choquenot, Proc. R. Soc. London Ser. B

357, 1233 (2002).

6. C. G. Jones, R. S. Ostfeld, M. P. Richard, E. M. Schauber,

J. O. Wolff, Science 279, 1023 (1998).

7. R. S. Ostfeld, F. Keesing, Trends Ecol. Evol. 15, 232 (2000).

8. D. Kelly, V. L. Sork, Annu. Rev. Ecol. Syst. 33, 427

(2002).

9. L. M. Curran, M. Leighton, Ecol. Monogr. 70, 101 (2000).

10. W. D. Koenig, J. M. H. Knops, Am. Sci. 93, 340 (2005).

11. D. H. Janzen, Annu. Rev. Ecol. Syst. 2, 465 (1971).

12. A. F. Hedlin, For. Sci. 10, 124 (1964).

13. M. J. McKone, D. Kelly, A. L. Harrison, J. J. Sullivan,

A. J. Cone, N.Z. J. Zool. 28, 89 (2001).

14. P. R. Wilson, B. J. Karl, R. J. Toft, J. R. Beggs, Biol. Conserv.

83, 175 (1998).

15. T. Ruf, J. Fietz, W. Schlund, C. Bieber, Ecology 87, 372

(2006).

16. F. H. Bronson, Reproductive Biology (Univ. of Chicago

Press, Chicago, 1989).

17. Materials and methods are available on Science Online.

18. J. O. Wolff, J. Mammal. 77, 850 (1996).

19. A. S. McNeilly, Reprod. Fertil. Dev. 13, 583 (2001).

20. J. D. Ligon, Nature 250, 80 (1974).

21. P. J. Berger, N. C. Negus, E. H. Sanders, P. D. Gardner,

Science 214, 69 (1981).

22. S. Eis, J. Inkster, Can. J. For. Res. 2, 460 (1972).

23. B. M. Fitzgerald, M. J. Efford, B. J. Karl, N.Z. J. Zool. 31,

167 (2004).

24. M. C. Smith, J. Wildl. Manag. 32, 305 (1968).

25. D. A. Rusch, W. G. Reeder, Ecology 59, 400 (1978).

26. L. A. Wauters, C. Swinnen, A. A. Dhondt, J. Zool. 227, 71

(1992).

27. L. A. Wauters, A. A. Dhondt, J. Zool. 217, 93 (1989).

28. C. D. Becker, Can. J. Zool. 71, 1326 (1993).

29. K. W. Larsen, C. D. Becker, S. Boutin, M. Blower,

J. Mammal. 78, 192 (1997).

30. M. M. Humphries, S. Boutin, Ecology 81, 2867

(2000).

31. A. G. McAdam, S. Boutin, Evolution 57, 1689 (2003).

32. L. A. Wauters, E. Matthysen, F. Adriaensen, G. Tosi,

J. Anim. Ecol. 73, 11 (2004).

33. We thank J. Herbers, C. Aumann, J. LaMontagne,

M. Andruskiw, J. Lane, R. Boonstra, K. McCann,

B. Thomas, Kluane Red Squirrel Project 2005 annual

meeting participants, and three anonymous reviewers for

very useful comments. We are indebted to the many field

workers who helped collect the data and to A. Sykes and

E. Anderson for management of the long-term data.

Financed by Natural Sciences and Engineering Research

Council of Canada, NSF (American reds), a Concerted

Action of the Belgian Ministry of Education, and a

Training and Mobility of Researchers grant from the

Commission of the European Union (Eurasian reds). This

is publication 30 of the Kluane Red Squirrel Project.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/314/5807/1928/DC1

Materials and Methods

SOM Text

Fig. S1

Table S1

References

25 September 2006; accepted 15 November 2006

10.1126/science.1135520

Human Catechol-O-MethyltransferaseHaplotypesModulate Protein Expressionby Altering mRNA Secondary StructureA. G. Nackley,1 S. A. Shabalina,2 I. E. Tchivileva,1 K. Satterfield,1 O. Korchynskyi,3

S. S. Makarov,4 W. Maixner,1 L. Diatchenko1*

Catechol-O-methyltransferase (COMT) is a key regulator of pain perception, cognitive function, andaffective mood. Three common haplotypes of the human COMT gene, divergent in two synonymousand one nonsynonymous position, code for differences in COMT enzymatic activity and areassociated with pain sensitivity. Haplotypes divergent in synonymous changes exhibited the largestdifference in COMT enzymatic activity, due to a reduced amount of translated protein. The majorCOMT haplotypes varied with respect to messenger RNA local stem-loop structures, such that themost stable structure was associated with the lowest protein levels and enzymatic activity. Site-directed mutagenesis that eliminated the stable structure restored the amount of translatedprotein. These data highlight the functional significance of synonymous variations and suggest theimportance of haplotypes over single-nucleotide polymorphisms for analysis of genetic variations.

The ability to predict the downstream

effects of genetic variation is critically

important for understanding both the

evolution of the genome and the molecular

basis of human disease. The effects of non-

synonymous polymorphisms have been wide-

ly characterized; because these variations

directly influence protein function, they are

relatively easy to study statistically and

experimentally (1). However, characterizing

polymorphisms located in regulatory regions,

which are much more common, has proved to

be problematic (2). Here, we focus on the

mechanism whereby polymorphisms of the

cathechol-O-methyltransferase (COMT ) gene

regulate gene expression.

COMT is an enzyme responsible for de-

grading catecholamines and thus represents a

critical component of homeostasis maintenance

(3). The human COMT gene encodes two

distinct proteins: soluble COMT (S-COMT)

and membrane-bound COMT (MB-COMT)

through the use of alternative translation initia-

1Center for Neurosensory Disorders, University of NorthCarolina, Chapel Hill, NC 27599, USA. 2National Center forBiotechnology Information, National Institutes of Health,Bethesda, MD 20894, USA. 3Thurston Arthritis Center,University of North Carolina, Chapel Hill, NC 27599, USA.4Attagene, Inc., 7030 Kit Creek Road, Research Triangle Park,NC 27560, USA.

*To whom correspondence should be addressed. E-mail:[email protected]

22 DECEMBER 2006 VOL 314 SCIENCE www.sciencemag.org1930

REPORTS

The ability to predict the downstream effects of genetic variation is criti-cally important for understanding

both the evolution of the genome and the molecular basis of human disease. The ef-fects of nonsynonymous polymorphisms have been widely characterized; because these variations directly influence protein function, they are relatively easy to study

creased, and bodymasswas lower when squirrels

were feeding on these items as compared with

feeding on seed (26, 27). Lastly, food supple-

mentation experiments of American reds have

failed to produce increases in litter size or a

second litter, providing strong evidence that the

increased reproductive investment observed in

our study cannot be triggered by increased

energy availability alone (17, 28, 29). We hy-

pothesize that, rather than following a resource-

tracking strategy where reproductive investment

is determined by current resource amounts, re-

productive rates are driven by future fitness

payoffs. During years of low seed production,

competition among juveniles for available re-

sources is intense, and, although litter augmenta-

tion experiments in American reds show that

females are capable of supporting larger litters

(30), they refrain from doing so because off-

spring recruitment is low (30). However, when

mast years occur, competition among juveniles is

reduced, and females produce more offspring,

which successfully recruit into the population

(31, 32). Further, the increased production of young

by females in mast years does not come with

any obvious cost to the female, because over-

winter survival is not reduced after years of

increased reproductive investment (for Amer-

ican reds, offspring production in the previous

year versus proportion of adult females surviving

to spring has slope = –0.023 ± 0.034, t15 = –0.7,

and P = 0.51; for Eurasian reds, proportion of

estrous females in the previous year versus adult

female survival to spring has slope = 0.37 ± 0.27,

t15 = 1.4, and P = 0.19).

If masting has evolved as a swamp-and-

starve adaptation against seed predation, then

anticipatory reproduction and population growth

represent a potent counteradaptation by the pred-

ators. Given that increased reproductive output

in these systems coincides with low current but

high future resources, our results suggest that

reproductive investment in these systems is more

responsive to future fitness prospects than present

energetic constraints. The evolution of seedmast-

ing in trees is also driven by the survival pros-

pects for progeny rather than simple resource

tracking, suggesting an intriguing parallel in re-

productive strategies of trees and the predators

that consume their seed.

References and Notes1. J. L. Gittleman, S. D. Thompson, Am. Zool. 28, 863

(1988).

2. T. H. Clutton-Brock, Reproductive Success (Univ. of

Chicago Press, Chicago, 1988).

3. T. H. Clutton-Brock, The Evolution of Parental Care

(Princeton Univ. Press, Princeton, NJ, 1999).

4. G. Caughley, in Theoretical Ecology, R. M. May, Ed.

(Blackwell, Oxford, 1981), ch. 6.

5. P. Bayliss, D. Choquenot, Proc. R. Soc. London Ser. B

357, 1233 (2002).

6. C. G. Jones, R. S. Ostfeld, M. P. Richard, E. M. Schauber,

J. O. Wolff, Science 279, 1023 (1998).

7. R. S. Ostfeld, F. Keesing, Trends Ecol. Evol. 15, 232 (2000).

8. D. Kelly, V. L. Sork, Annu. Rev. Ecol. Syst. 33, 427

(2002).

9. L. M. Curran, M. Leighton, Ecol. Monogr. 70, 101 (2000).

10. W. D. Koenig, J. M. H. Knops, Am. Sci. 93, 340 (2005).

11. D. H. Janzen, Annu. Rev. Ecol. Syst. 2, 465 (1971).

12. A. F. Hedlin, For. Sci. 10, 124 (1964).

13. M. J. McKone, D. Kelly, A. L. Harrison, J. J. Sullivan,

A. J. Cone, N.Z. J. Zool. 28, 89 (2001).

14. P. R. Wilson, B. J. Karl, R. J. Toft, J. R. Beggs, Biol. Conserv.

83, 175 (1998).

15. T. Ruf, J. Fietz, W. Schlund, C. Bieber, Ecology 87, 372

(2006).

16. F. H. Bronson, Reproductive Biology (Univ. of Chicago

Press, Chicago, 1989).

17. Materials and methods are available on Science Online.

18. J. O. Wolff, J. Mammal. 77, 850 (1996).

19. A. S. McNeilly, Reprod. Fertil. Dev. 13, 583 (2001).

20. J. D. Ligon, Nature 250, 80 (1974).

21. P. J. Berger, N. C. Negus, E. H. Sanders, P. D. Gardner,

Science 214, 69 (1981).

22. S. Eis, J. Inkster, Can. J. For. Res. 2, 460 (1972).

23. B. M. Fitzgerald, M. J. Efford, B. J. Karl, N.Z. J. Zool. 31,

167 (2004).

24. M. C. Smith, J. Wildl. Manag. 32, 305 (1968).

25. D. A. Rusch, W. G. Reeder, Ecology 59, 400 (1978).

26. L. A. Wauters, C. Swinnen, A. A. Dhondt, J. Zool. 227, 71

(1992).

27. L. A. Wauters, A. A. Dhondt, J. Zool. 217, 93 (1989).

28. C. D. Becker, Can. J. Zool. 71, 1326 (1993).

29. K. W. Larsen, C. D. Becker, S. Boutin, M. Blower,

J. Mammal. 78, 192 (1997).

30. M. M. Humphries, S. Boutin, Ecology 81, 2867

(2000).

31. A. G. McAdam, S. Boutin, Evolution 57, 1689 (2003).

32. L. A. Wauters, E. Matthysen, F. Adriaensen, G. Tosi,

J. Anim. Ecol. 73, 11 (2004).

33. We thank J. Herbers, C. Aumann, J. LaMontagne,

M. Andruskiw, J. Lane, R. Boonstra, K. McCann,

B. Thomas, Kluane Red Squirrel Project 2005 annual

meeting participants, and three anonymous reviewers for

very useful comments. We are indebted to the many field

workers who helped collect the data and to A. Sykes and

E. Anderson for management of the long-term data.

Financed by Natural Sciences and Engineering Research

Council of Canada, NSF (American reds), a Concerted

Action of the Belgian Ministry of Education, and a

Training and Mobility of Researchers grant from the

Commission of the European Union (Eurasian reds). This

is publication 30 of the Kluane Red Squirrel Project.



SOM Text

Fig. S1

Table S1

References

25 September 2006; accepted 15 November 2006

10.1126/science.1135520

Human Catechol-O-MethyltransferaseHaplotypesModulate Protein Expressionby Altering mRNA Secondary StructureA. G. Nackley,1 S. A. Shabalina,2 I. E. Tchivileva,1 K. Satterfield,1 O. Korchynskyi,3

S. S. Makarov,4 W. Maixner,1 L. Diatchenko1*

Catechol-O-methyltransferase (COMT) is a key regulator of pain perception, cognitive function, andaffective mood. Three common haplotypes of the human COMT gene, divergent in two synonymousand one nonsynonymous position, code for differences in COMT enzymatic activity and areassociated with pain sensitivity. Haplotypes divergent in synonymous changes exhibited the largestdifference in COMT enzymatic activity, due to a reduced amount of translated protein. The majorCOMT haplotypes varied with respect to messenger RNA local stem-loop structures, such that themost stable structure was associated with the lowest protein levels and enzymatic activity. Site-directed mutagenesis that eliminated the stable structure restored the amount of translatedprotein. These data highlight the functional significance of synonymous variations and suggest theimportance of haplotypes over single-nucleotide polymorphisms for analysis of genetic variations.

The ability to predict the downstream

effects of genetic variation is critically

important for understanding both the

evolution of the genome and the molecular

basis of human disease. The effects of non-

synonymous polymorphisms have been wide-

ly characterized; because these variations

directly influence protein function, they are

relatively easy to study statistically and

experimentally (1). However, characterizing

polymorphisms located in regulatory regions,

which are much more common, has proved to

be problematic (2). Here, we focus on the

mechanism whereby polymorphisms of the

cathechol-O-methyltransferase (COMT ) gene

regulate gene expression.

COMT is an enzyme responsible for de-

grading catecholamines and thus represents a

critical component of homeostasis maintenance

(3). The human COMT gene encodes two

distinct proteins: soluble COMT (S-COMT)

and membrane-bound COMT (MB-COMT)

through the use of alternative translation initia-

1Center for Neurosensory Disorders, University of NorthCarolina, Chapel Hill, NC 27599, USA. 2National Center forBiotechnology Information, National Institutes of Health,Bethesda, MD 20894, USA. 3Thurston Arthritis Center,University of North Carolina, Chapel Hill, NC 27599, USA.4Attagene, Inc., 7030 Kit Creek Road, Research Triangle Park,NC 27560, USA.



REPORTS







































































13

group demonstrated that three common haplotypes of the human COMT gene are associated with pain sensitivity and the likelihood of developing temporoman-dibular joint disorder (TMJD), a common chronic musculoskeletal pain condition (4). Three major haplotypes are formed by four single-nucleotide polymorphisms (SNPs): one located in the S-COMT pro-moter region (A/G; rs6269) and three in the S- and MB-COMT coding region at codons his62his (C/T; rs4633), leu136leu (C/G; rs4818), and val158met (A/G; rs4680) (Fig. 1A). On the basis of subjects’ pain responsiveness, haplotypes were designat-ed as low (LPS; GCGG), average (APS; ATCA), or high (HPS; ACCG) pain sen-

ed in regulatory regions, which are much more common, has proved to be problem-atic (2). Here, we focus on the mechanism whereby polymorphisms of the cathechol-O-methyltransferase (COMT) gene regu-late gene expression.

COMT is an enzyme responsible for degrading catecholamines and thus represents a critical component of ho-meostasis maintenance (3). The human COMT gene encodes two distinct pro-teins: soluble COMT (S-COMT) and membrane-bound COMT (MB-COMT) through the use of alternative translation initiation sites and promoters (3). Re-cently, COMT has been implicated in the modulation of persistent pain (4–7). Our

tion sites and promoters (3). Recently, COMT

has been implicated in the modulation of

persistent pain (4–7). Our group demonstrated

that three common haplotypes of the human

COMT gene are associated with pain sensitivity

and the likelihood of developing temporoman-

dibular joint disorder (TMJD), a common

chronic musculoskeletal pain condition (4).

Three major haplotypes are formed by four

single-nucleotide polymorphisms (SNPs): one

located in the S-COMT promoter region (A/G;

rs6269) and three in the S- and MB-COMT

coding region at codons his62his (C/T; rs4633),

leu136

leu (C/G; rs4818), and val158

met (A/G;

rs4680) (Fig. 1A). On the basis of subjects’

pain responsiveness, haplotypes were desig-

nated as low (LPS; GCGG), average (APS;

ATCA), or high (HPS; ACCG) pain sensitive.

Individuals carrying HPS/APS or APS/APS

diplotypes were nearly 2.5 times as likely to

develop TMJD. Previous data further suggest

that COMT haplotypes code for differences in

COMT enzymatic activity (4); however, the

molecular mechanisms involved have remained

unknown.

A common SNP in codon 158 (val158met)

has been associated with pain ratings and

m-opioid system responses (7) as well as

addiction, cognition, and common affective

disorders (3, 8–10). The substitution of valine

(Val) by methionine (Met) results in reduced

thermostability and activity of the enzyme (3).

It is generally accepted that val158met is the

main source of individual variation in human

COMT activity; numerous studies have iden-

tified associations between the low-activity

met allele and several complex phenotypes

(3, 8, 10). However, observed associations

between these conditions and the met allele

are often modest and occasionally inconsistent

(3). This suggests that additional SNPs in the

COMT gene modulate COMT activity. Indeed,

we found that haplotype rather than an

individual SNP better accounts for variability

in pain sensitivity (4). The HPS and LPS

haplotypes that both code for the stable val158

variant were associated with the two extreme-

pain phenotypes; thus, the effect of haplotype

on pain sensitivity in our study cannot be

explained by the sum of the effects of

functional SNPs. Instead, the val158

met SNP

interacts with other SNPs to determine phe-

notype. Because variation in the S-COMT

promoter region does not contribute to pain

phenotype (Fig. 1A), we suggest that the rate

of mRNA degradation or protein synthesis is

affected by the structural properties of the

haplotypes, such as haplotype-specific mRNA

secondary structure.

Previous reports have shown that polymor-

phic alleles can markedly affect mRNA sec-

ondary structure (11, 12), which can then have

functional consequences on the rate of mRNA

degradation (11, 13). It is also plausible that

polymorphic alleles directly modulate protein

translation through alterations in mRNA sec-

ondary structure, because protein translation

efficiency is affected bymRNA secondary struc-

ture (14–16). To test these possibilities, we

evaluated the affect of LPS, APS, and HPS

haplotypes on the stability of the corresponding

mRNA secondary structures (17).

Secondary structures of the full-length LPS,

APS, and HPS mRNA transcripts were pre-

dicted by means of the RNAMfold (18, 19) and

Afold (20) programs. The mRNA folding

analyses demonstrated that the major COMT

haplotypes differ with respect to mRNA

secondary structure. The LPS haplotype codes

for the shortest, least stable local stem-loop

structure, and the HPS haplotype codes for the

longest, most stable local stem-loop structure

in the val158 region for both S-COMT and

MB-COMT. Gibbs free energy (DG) for the stem-

loop structure associatedwith theHPS haplotype is

~17 kcal/mol less than that associated with the

LPS haplotype for both S-COMTandMB-COMT

(Fig. 1B and fig. S1A). Additional evidence sup-

porting predicted RNA folding structures was

obtained by generating consensus RNA sec-

ondary structures based on comparative analysis

of COMT sequences from eight mammalian

species (fig. S2). The consensus RNA folding

structures were LPS-like and did not contain

highly stable local stem-loop structures analo-

gous to the human HPS-like form. Thus, sub-

stantial deviation from consensus structure, as

observed for the HPS haplotype, should have

notable functional consequences. Additional

studies were conducted to test this molecular

modeling.

We constructed full-length S- andMB-COMT

cDNA clones in mammalian expression vec-

tors that differed only in three nucleotides

corresponding to the LPS, APS, and HPS hap-

lotypes (17, 21). Rat adrenal (PC-12) cells

were transiently transfected with each of these

six constructs. COMT enzymatic activity, pro-

tein expression, and mRNA abundance were

measured. Relative to the LPS haplotype, the

HPS haplotype showed a 25- and 18-fold

reduction in enzymatic activity for S- and

MB-COMT constructs, respectively (Fig. 1C

and fig. S1B). The HPS haplotype also exhib-

ited marked reductions in S- and MB-COMT

protein expression (Fig. 1D and fig. S1C). The

APS haplotype displayed a moderate 2.5- and 3-

fold reduction in enzymatic activity for S- and

MB-COMT constructs, respectively, while pro-

Fig. 1. Common haplo-types of the human COMTgene differ with respecttomRNA secondary struc-ture and enzymatic ac-tivity. (A) A schematicdiagram illustrates COMTgenomic organization andSNP composition for thethree haplotypes. Per-cent frequency of eachhaplotype in a cohort ofhealthy Caucasian fe-males, and percent inde-pendent SNP contributionto pain sensitivity, areindicated. (B) The localstem-loop structures as-sociated with each ofthe three haplotypes areshown. Relative to theLPS and APS haplotypes,the HPS local stem-loopstructure had a higherfolding potential. (C andD) The LPS haplotype ex-hibited the highest, whilethe HPS haplotype exhib-ited the lowest enzymaticactivity and protein lev-els in cells expressingCOMT. ***P < 0.001, ≠LPS. +++P< 0.001,≠ APS.

www.sciencemag.org SCIENCE VOL 314 22 DECEMBER 2006 1931

REPORTS

tion sites and promoters (3). Recently, COMT

has been implicated in the modulation of

persistent pain (4–7). Our group demonstrated

that three common haplotypes of the human

COMT gene are associated with pain sensitivity

and the likelihood of developing temporoman-

dibular joint disorder (TMJD), a common

chronic musculoskeletal pain condition (4).

Three major haplotypes are formed by four

single-nucleotide polymorphisms (SNPs): one

located in the S-COMT promoter region (A/G;

rs6269) and three in the S- and MB-COMT

coding region at codons his62his (C/T; rs4633),

leu136

leu (C/G; rs4818), and val158

met (A/G;

rs4680) (Fig. 1A). On the basis of subjects’

pain responsiveness, haplotypes were desig-

nated as low (LPS; GCGG), average (APS;

ATCA), or high (HPS; ACCG) pain sensitive.

Individuals carrying HPS/APS or APS/APS

diplotypes were nearly 2.5 times as likely to

develop TMJD. Previous data further suggest

that COMT haplotypes code for differences in

COMT enzymatic activity (4); however, the

molecular mechanisms involved have remained

unknown.

A common SNP in codon 158 (val158met)

has been associated with pain ratings and

m-opioid system responses (7) as well as

addiction, cognition, and common affective

disorders (3, 8–10). The substitution of valine

(Val) by methionine (Met) results in reduced

thermostability and activity of the enzyme (3).

It is generally accepted that val158met is the

main source of individual variation in human

COMT activity; numerous studies have iden-

tified associations between the low-activity

met allele and several complex phenotypes

(3, 8, 10). However, observed associations

between these conditions and the met allele

are often modest and occasionally inconsistent

(3). This suggests that additional SNPs in the

COMT gene modulate COMT activity. Indeed,

we found that haplotype rather than an

individual SNP better accounts for variability

in pain sensitivity (4). The HPS and LPS

haplotypes that both code for the stable val158

variant were associated with the two extreme-

pain phenotypes; thus, the effect of haplotype

on pain sensitivity in our study cannot be

explained by the sum of the effects of

functional SNPs. Instead, the val158

met SNP

interacts with other SNPs to determine phe-

notype. Because variation in the S-COMT

promoter region does not contribute to pain

phenotype (Fig. 1A), we suggest that the rate

of mRNA degradation or protein synthesis is

affected by the structural properties of the

haplotypes, such as haplotype-specific mRNA

secondary structure.

Previous reports have shown that polymor-

phic alleles can markedly affect mRNA sec-

ondary structure (11, 12), which can then have

functional consequences on the rate of mRNA

degradation (11, 13). It is also plausible that

polymorphic alleles directly modulate protein

translation through alterations in mRNA sec-

ondary structure, because protein translation

efficiency is affected bymRNA secondary struc-

ture (14–16). To test these possibilities, we

evaluated the affect of LPS, APS, and HPS

haplotypes on the stability of the corresponding

mRNA secondary structures (17).

Secondary structures of the full-length LPS,

APS, and HPS mRNA transcripts were pre-

dicted by means of the RNAMfold (18, 19) and

Afold (20) programs. The mRNA folding

analyses demonstrated that the major COMT

haplotypes differ with respect to mRNA

secondary structure. The LPS haplotype codes

for the shortest, least stable local stem-loop

structure, and the HPS haplotype codes for the

longest, most stable local stem-loop structure

in the val158 region for both S-COMT and

MB-COMT. Gibbs free energy (DG) for the stem-

loop structure associatedwith theHPS haplotype is

~17 kcal/mol less than that associated with the

LPS haplotype for both S-COMTandMB-COMT

(Fig. 1B and fig. S1A). Additional evidence sup-

porting predicted RNA folding structures was

obtained by generating consensus RNA sec-

ondary structures based on comparative analysis

of COMT sequences from eight mammalian

species (fig. S2). The consensus RNA folding

structures were LPS-like and did not contain

highly stable local stem-loop structures analo-

gous to the human HPS-like form. Thus, sub-

stantial deviation from consensus structure, as

observed for the HPS haplotype, should have

notable functional consequences. Additional

studies were conducted to test this molecular

modeling.

We constructed full-length S- andMB-COMT

cDNA clones in mammalian expression vec-

tors that differed only in three nucleotides

corresponding to the LPS, APS, and HPS hap-

lotypes (17, 21). Rat adrenal (PC-12) cells

were transiently transfected with each of these

six constructs. COMT enzymatic activity, pro-

tein expression, and mRNA abundance were

measured. Relative to the LPS haplotype, the

HPS haplotype showed a 25- and 18-fold

reduction in enzymatic activity for S- and

MB-COMT constructs, respectively (Fig. 1C

and fig. S1B). The HPS haplotype also exhib-

ited marked reductions in S- and MB-COMT

protein expression (Fig. 1D and fig. S1C). The

APS haplotype displayed a moderate 2.5- and 3-

fold reduction in enzymatic activity for S- and

MB-COMT constructs, respectively, while pro-

Fig. 1. Common haplo-types of the human COMTgene differ with respecttomRNA secondary struc-ture and enzymatic ac-tivity. (A) A schematicdiagram illustrates COMTgenomic organization andSNP composition for thethree haplotypes. Per-cent frequency of eachhaplotype in a cohort ofhealthy Caucasian fe-males, and percent inde-pendent SNP contributionto pain sensitivity, areindicated. (B) The localstem-loop structures as-sociated with each ofthe three haplotypes areshown. Relative to theLPS and APS haplotypes,the HPS local stem-loopstructure had a higherfolding potential. (C andD) The LPS haplotype ex-hibited the highest, whilethe HPS haplotype exhib-ited the lowest enzymaticactivity and protein lev-els in cells expressingCOMT. ***P < 0.001, ≠LPS. +++P< 0.001,≠ APS.


REPORTS

14

sitive. Individuals carrying HPS/APS or APS/APS diplo-types were nearly 2.5 times as likely to develop TMJD. Previous data fur-ther suggest that COMT haplotypes code for differences in COMT enzymatic activity (4); however, the molecular mechanisms involved have remained unknown.

A common SNP in codon 158 (val158met) has been associated with pain ratings and μ-opioid system responses (7) as well as addiction, cognition, and common affec-tive disorders (3, 8–10). The substitution of valine (Val) by methionine (Met) results in reduced thermostability and activity of the enzyme (3). It is generally accepted that val158met is the main source of indi-vidual variation in human COMT activity; numerous studies have identified associa-tions between the low-activity met allele and several complex phenotypes (3, 8, 10).

However, observed associations between these conditions and the met allele are of-ten modest and occasionally inconsistent (3). This suggests that additional SNPs in the COMT gene modulate COMT activ-ity. Indeed, we found that haplotype rather than an individual SNP better accounts for variability in pain sensitivity (4). The HPS and LPS haplotypes that both code for the stable val158 variant were associated with the two extreme-pain phenotypes; thus, the effect of haplotype on pain sensitiv-ity in our study cannot be explained by the sum of the effects of functional SNPs. Instead, the val158met SNP interacts with other SNPs to determine phenotype. Be-cause variation in the S-COMT promoter

tein expression levels did not differ. The mod-

erate reduction in enzymatic activity produced

by the APS haplotype is most likely due to the

previously reported decrease in protein thermo-

stability coded by the met158 allele (3). These

data illustrate that the reduced enzymatic activity

corresponding to the HPS haplotype is paralleled

by reduced protein levels, an effect that could be

mediated by local mRNA secondary structure at

the level of protein synthesis and/or mRNA

degradation. Because total RNA abundance and

RNA degradation rates did not parallel COMT

protein levels (fig. S2), differences in protein

translation efficiency likely results from differ-

ences in the local secondary structure of cor-

responding mRNAs.

To directly assess this hypothesis, we

performed site-directed mutagenesis (17).

The stable stem-loop structure of S- and

MB-COMT mRNA corresponding to the HPS

haplotype is supported by base pairs between

several critical nucleotides, including 403C

and 479G in S-COMT and 625C and 701G in

MB-COMT (Fig. 2A and fig. S3A). Mutation

of 403C to G in S-COMT or 625C to G in

MB-COMT destroys the stable stem-loop struc-

ture and converts it into a LPS haplotype–like

structure (HPS Lsm). Double mutation of

mRNA in position 403C to G and 479G to C in

S-COMT or 625C to G and 701G to C in MB-

COMT reconstructs the original long stem-loop

structure (HPS dm). The single- and double-

nucleotide HPS mutants (HPS Lsm and HPS

dm, respectively) were transiently transfected to

PC-12 cells. As predicted by the mRNA

secondary-structure folding analyses, the HPS

Lsm exhibited increased COMT enzymatic

activity and protein levels equivalent to those

of the LPS haplotype, whereas the HPS dm

exhibited reduced enzymatic activity and pro-

tein levels equivalent to those of the original

HPS haplotype (Fig. 2, B and C; fig. S3, B and

C). These data rule out the involvement of RNA

sequence recognition motifs or codon usage in

the regulation of translation. In contrast to the

HPS haplotype, protein levels did not parallel

COMTenzymatic activity for the APS haplotype

and site-directed mutagenesis confirmed that the

met158 allele, not a more stable mRNA sec-

ondary structure, drives the reduced enzymatic

activity observed for the APS haplotype (fig.

S4). This difference is moderate relative to the

mRNA structure–dependent difference coded by

LPS and HPS haplotypes. These results were

verified by an alternate approach of modifying

mRNA secondary structure (fig. S5).

Our data have very broad evolutionary and

medical implications for the analysis of

variants common in the human population.

The fact that alterations in mRNA secondary

structure resulting from synonymous changes

have such a pronounced effect on the level of

protein expression emphasizes the critical role

of synonymous nucleotide positions in main-

taining mRNA secondary structure and sug-

gests that the mRNA secondary structure,

rather than independent nucleotides in the

synonymous positions, should undergo sub-

stantial selective pressure (22). Furthermore,

our data stress the importance of synonymous

SNPs as potential functional variants in the

area of human medical genetics. Although

nonsynonymous SNPs are believed to have

the strongest impact on variation in gene

function, our data clearly demonstrate that

haplotypic variants of common synonymous

SNPs can have stronger effects on gene func-

tion than nonsynonymous variations and

play an important role in disease onset and

progression.

References and Notes

1. L. Y. Yampolsky, F. A. Kondrashov, A. S. Kondrashov,

Hum. Mol. Genet. 14, 3191 (2005).

2. J. C. Knight, J. Mol. Med. 83, 97 (2005).

3. P. T. Mannisto, S. Kaakkola, Pharmacol. Rev. 51, 593

(1999).

4. L. Diatchenko et al., Hum. Mol. Genet. 14, 135

(2005).

5. J. J. Marbach, M. Levitt, J. Dent. Res. 55, 711

(1976).

6. T. T. Rakvag et al., Pain 116, 73 (2005).

7. J. K. Zubieta et al., Science 299, 1240 (2003).

8. G. Winterer, D. Goldman, Brain Res. Brain Res. Rev. 43,

134 (2003).

9. B. Funke et al., Behav. Brain Funct. 1, 19 (2005).

10. G. Oroszi, D. Goldman, Pharmacogenomics 5, 1037

(2004).

11. J. Duan et al., Hum. Mol. Genet. 12, 205

(2003).

12. L. X. Shen, J. P. Basilion, V. P. Stanton Jr., Proc. Natl.

Acad. Sci. U.S.A. 96, 7871 (1999).

13. I. Puga et al., Endocrinology 146, 2210 (2005).

14. K. Mita, S. Ichimura, M. Zama, T. C. James, J. Mol. Biol.

203, 917 (1988).

15. T. D. Schmittgen, K. D. Danenberg, T. Horikoshi,

H. J. Lenz, P. V. Danenberg, J. Biol. Chem. 269, 16269

(1994).

16. A. Shalev et al., Endocrinology 143, 2541

(2002).

17. Materials and methods are available as supporting

material on Science Online.

18. D. H. Mathews, J. Sabina, M. Zuker, D. H. Turner, J. Mol.

Biol. 288, 911 (1999).

19. M. Zuker, Nucleic Acids Res. 31, 3406 (2003).

20. A. Y. Ogurtsov, S. A. Shabalina, A. S. Kondrashov,

M. A. Roytberg, Bioinformatics 22, 1317 (2006).

21. Sequence accession numbers: S-COMT clones BG290167,

CA489448, and BF037202 represent LPS, APS, and HPS

haplotypes, respectively. Clones corresponding to all

three haplotypes and including the transcriptional start

site were constructed by use of the unique restriction

enzyme Bsp MI. The gene coding region containing all

three SNPs was cut and inserted into the plasmids

through use of the S-COMT (BG290167) and MB-COMT

(BI835796) clones containing the entire COMT 5′ and 3′

ends as a backbone vector.

22. S. A. Shabalina, A. Y. Ogurtsov, N. A. Spiridonov, Nucleic

Acids Res. 34, 2428 (2006).

Fig. 2. Site-directed mu-tagenesis that destroysthe stable stem-loop struc-ture corresponding to theHPS haplotype restoresCOMT enzymatic activityand protein expression.(A) The mRNA structurecorresponding to the HPShaplotype was convertedto an LPS haplotype-likestructure (HPS Lsm) bysingle mutation of 403Cto G. The original HPShaplotype structure (HPSdm) was restored by dou-ble mutation of interact-ing nucleotides 403C to Gand 479G to C. (B and C)The HPS Lsm exhibitedCOMT enzymatic activityand protein levels equiv-alent to those of the LPShaplotype, whereas theHPS dm exhibited re-duced enzymatic activity.***P < 0.001, ≠ LPS.


REPORTS
















responding mRNAs.
































































progression.



Hum. Mol. Genet. 14, 3191 (2005).



(1999).


(2005).


(1976).




134 (2003).



(2004).


(2003).


Acad. Sci. U.S.A. 96, 7871 (1999).



203, 917 (1988).



(1994).


(2002).




Biol. 288, 911 (1999).















Acids Res. 34, 2428 (2006).



REPORTS

15

region does not contribute to pain pheno-type (Fig. 1A), we suggest that the rate of mRNA degradation or protein synthesis is affected by the structural properties of the haplotypes, such as haplotype-specific mRNA secondary structure.

Previous reports have shown that polymorphic alleles can markedly affect mRNA secondary structure (11, 12), which can then have functional consequences on the rate of mRNA degradation (11, 13). It is also plausible that polymorphic al-leles directly modulate protein translation through alterations in mRNA secondary structure, because protein translation ef-ficiency is affected by mRNA secondary structure (14–16). To test these possibili-ties, we evaluated the effect of LPS, APS, and HPS haplotypes on the stability of the corresponding mRNA secondary struc-tures (17).

Secondary structures of the full-length LPS, APS, and HPS mRNA transcripts were predicted by means of the RNA Mfold (18, 19) and Afold (20) programs. The mRNA folding analyses demon-strated that the major COMT haplotypes differ with respect to mRNA secondary structure. The LPS haplotype codes for the shortest, least stable local stem-loop structure, and the HPS haplotype codes for the longest, most stable local stem-loop structure in the val158 region for both S-COMT and MB-COMT. Gibbs free en-ergy (ΔG) for the stem-loop structure as-sociated with the HPS haplotype is ~17 kcal/mol less than that associated with the LPS haplotype for both S-COMT and MB-COMT (Fig. 1B and fig. S1A). Ad-ditional evidence supporting predicted RNA folding structures was obtained by generating consensus RNA secondary structures based on comparative analysis of COMT sequences from eight mamma-lian species (fig. S2). The consensus RNA folding structures were LPS-like and did not contain highly stable local stem-loop structures analogous to the human HPS-like form. Thus, substantial deviation

from consensus structure, as observed for the HPS haplotype, should have notable functional consequences. Additional stud-ies were conducted to test this molecular modeling.

We constructed full-length S- and MB-COMT cDNA clones in mammalian ex-pression vectors that differed only in three nucleotides corresponding to the LPS, APS, and HPS haplotypes (17, 21). Rat adrenal (PC-12) cells were transiently transfected with each of these six constructs. COMT enzymatic activity, protein expression, and mRNA abundance were measured. Rela-tive to the LPS haplotype, the HPS haplo-type showed a 25- and 18-fold reduction in enzymatic activity for S- and MB-COMT constructs, respectively (Fig. 1C and fig. S1B). The HPS haplotype also exhibited marked reductions in S- and MB-COMT protein expression (Fig. 1D and fig. S1C). The APS haplotype displayed a moderate 2.5- and 3-fold reduction in enzymatic activity for S- and MB-COMT constructs, respectively, while protein expression lev-els did not differ. The moderate reduction in enzymatic activity produced by the APS haplotype is most likely due to the previ-ously reported decrease in protein ther-mostability coded by the met158 allele (3). These data illustrate that the reduced en-zymatic activity corresponding to the HPS haplotype is paralleled by reduced protein levels, an effect that could be mediated by local mRNA secondary structure at the lev-el of protein synthesis and/or mRNA deg-radation. Because total RNA abundance and RNA degradation rates did not parallel COMT protein levels (fig. S2), differences in protein translation efficiency likely re-sults from differences in the local second-ary structure of corresponding mRNAs

To directly assess this hypothesis, we performed site-directed mutagenesis (17). The stable stem-loop structure of S- and MB-COMT mRNA corresponding to the HPS haplotype is supported by base pairs between several critical nucleotides, in-cluding 403C and 479G in S-COMT and

16

625C and 701G in MB-COMT (Fig. 2A and fig. S3A). Mutation of 403C to G in S-COMT or 625C to G in MB-COMT destroys the stable stem-loop structure and converts it into a LPS haplotype–like structure (HPS Lsm). Double mutation of mRNA in position 403C to G and 479G to C in S-COMT or 625C to G and 701G to C in MB-COMT reconstructs the original long stem-loop structure (HPS dm). The single- and double-nucleotide HPS mu-tants (HPS Lsm and HPS dm, respective-ly) were transiently transfected to PC-12 cells. As predicted by the mRNA second-ary-structure folding analyses, the HPS Lsm exhibited increased COMT enzymat-ic activity and protein levels equivalent to those of the LPS haplotype, whereas the HPS dm exhibited reduced enzymatic activity and protein levels equivalent to those of the original HPS haplotype (Fig. 2, B and C; fig. S3, B and C). These data rule out the involvement of RNA sequence recognition motifs or codon usage in the regulation of translation. In contrast to the HPS haplotype, protein levels did not parallel COMT enzymatic activity for the APS haplotype and site-directed muta-genesis confirmed that the met158 allele, not a more stable mRNA secondary struc-ture, drives the reduced enzymatic activ-ity observed for the APS haplotype (fig. S4). This difference is moderate relative to the mRNA structure–dependent differ-ence coded by LPS and HPS haplotypes. These results were verified by an alternate approach of modifying mRNA secondary structure (fig. S5).

Our data have very broad evolutionary and medical implications for the analysis of variants common in the human population. The fact that alterations in mRNA second-ary structure resulting from synonymous changes have such a pronounced effect on the level of protein expression emphasizes the critical role of synonymous nucleotide positions in maintaining mRNA second-ary structure and suggests that the mRNA secondary structure, rather than indepen-
















responding mRNAs.
































































progression.



Hum. Mol. Genet. 14, 3191 (2005).



(1999).


(2005).


(1976).




134 (2003).



(2004).


(2003).


Acad. Sci. U.S.A. 96, 7871 (1999).



203, 917 (1988).



(1994).


(2002).




Biol. 288, 911 (1999).















Acids Res. 34, 2428 (2006).



REPORTS

dent nucleotides in the synonymous posi-tions, should undergo substantial selective pressure (22). Furthermore, our data stress the importance of synonymous SNPs as potential functional variants in the area of human medical genetics. Although non-synonymous SNPs are believed to have the strongest impact on variation in gene function, our data clearly demonstrate that haplotypic variants of common synony-mous SNPs can have stronger effects on gene function than nonsynonymous varia-tions and play an important role in disease onset and progression.

17

23. We are grateful to the National Institute of Child

Health and Human Development, National Institute of

Neurological Disorders and Stroke, and National

Institute of Dental and Craniofacial Research at the NIH

for financial support of this work. This research was

also supported by the Intramural Research Program of

the NIH, National Center for Biotechnology

Information.



SOM Text

Figs. S1 to S6

References

13 June 2006; accepted 16 November 2006

10.1126/science.1131262

Lineages of AcidophilicArchaea Revealed by CommunityGenomic AnalysisBrett J. Baker,1 Gene W. Tyson,2 Richard I. Webb,3 Judith Flanagan,2*

Philip Hugenholtz,2† Eric E. Allen,2‡ Jillian F. Banfield1,2§

Novel, low-abundance microbial species can be easily overlooked in standard polymerase chainreaction (PCR)–based surveys. We used community genomic data obtained without PCR orcultivation to reconstruct DNA fragments bearing unusual 16S ribosomal RNA (rRNA) and protein-coding genes from organisms belonging to novel archaeal lineages. The organisms are minorcomponents of all biofilms growing in pH 0.5 to 1.5 solutions within the Richmond Mine,California. Probes specific for 16S rRNA showed that the fraction less than 0.45 micrometers indiameter is dominated by these organisms. Transmission electron microscope images revealed thatthe cells are pleomorphic with unusual folded membrane protrusions and have apparent volumesof <0.006 cubic micrometer.

Our understanding of the variety of

microorganisms that populate natural

environments was advanced by the

development of polymerase chain reaction

(PCR)–based, cultivation-independent methods

that target one or a small number of genes (1–3).

Genomic analyses of DNA sequence fragments

derived from multispecies consortia (4–6) and

whole environments (7–9) have provided new

information about diversity and metabolic

potential. However, PCR-based methods have

limited ability to detect organisms whose genes

are significantly divergent relative to gene

sequences in databases, and most cultivation-

independent genomic sequencing approaches

are relatively insensitive to organisms that

occur at low abundance. Consequently, it is

likely that low-abundance microorganisms dis-

tantly related to known species will be undetected

members of natural consortia, even in low

complexity systems such as acid mine drainage

(AMD) (10).

An important way in which microorganisms

affect geochemical cycles is by accelerating the

dissolution of minerals. For example, micro-

organisms can derive metabolic energy by ox-

idizing iron released by the dissolution of

pyrite (FeS2). The ferric iron by-product pro-

motes further pyrite dissolution, leading to AMD

generation. AMD solutions forming under-

ground in the Richmond Mine at Iron Mountain,

California, are warm (30° to 59°C), acidic (pH

~0.5 to 1.5), metal-rich [submolar Fe2+ and mi-

cromolar As and Cu (11)] and host active

microbial communities. Extensive cultivation-

independent sequence analysis of functional and

rRNA genes (11, 12) revealed that biofilms con-

tain a significant number of Archaea, but the

diversity reported to date has been limited to the

order Thermoplasmatales (10). Current models

for AMD generation thus include only these

species.

The genomes of the five dominant members

of one biofilm community from the “5-way”

region of the Richmond Mine (fig. S1) were

largely reconstructed through the assembly of

76 Mb of shotgun genomic sequence (4). Pre-

viously unreported is a genome fragment that

encodes part of the 16S rRNA gene of a novel

archaeal lineage: Archaeal Richmond Mine

Acidophilic Nanoorganism (ARMAN-1). Using

an expanded data set that now comprises more

than 100 Mb of genomic sequence, we recon-

structed a contiguous 4.2-kb fragment adjacent

to this gene. A second 13.2-kb genome fragment

encoding a 16S rRNA gene from an organism

that is related to ARMAN-1 (ARMAN-2) was

reconstructed from 117 Mb of community ge-

nomic sequence derived from a biofilm from the

A drift (fig. S1). Within the data sets from each

site, results to date indicate that each ARMAN

population is near-clonal.

Comparison of the ARMAN-1 and -2 DNA

fragments revealed some gene rearrangements,

insertions, and deletions (Fig. 1). Genes present

in both organisms encode putative inorganic

pyrophosphatases, a transcription regulator, and

a gene shown to be an arsenate reductase (13).

Comparative analysis of these genes with se-

quences in the public databases consistently

1Department of Earth and Planetary Sciences, University ofCalifornia, Berkeley, CA 94720, USA. 2EnvironmentalScience, Policy, and Management, University of California,Berkeley, CA 94720, USA. 3Centre for Microscopy andMicroanalysis and Department of Microbiology and Para-sitology, University of Queensland, Brisbane 4072,Australia.

*Present address: University of California, San Francisco,CA 94143, USA.†Present address: Department of Energy Joint GenomeInstitute, Walnut Creek, CA 94598, USA.‡Present address: University of California, San Diego, La Jolla,CA 92093, USA.§To whom correspondence should be addressed. E-mail:[email protected]

1 13,236

ARMAN-1

ARMAN-2

16S rRNAarsCmercuric reductasetransposase GTP cyclohydrolases ribosomal proteins

31% 85%69% 90% 97%

pyrophosphatase insert

4,2261

Fig. 1. Comparison of syntenous genomic regions of ARMAN-1 [from the“5-way” (CG) community (4)] and ARMAN-2 (from the UBA community).Orthologs and their protein identity are indicated by the red bands. The

percentage similarity for the 16S rRNA gene sequences is also shown.Numbers at ends indicate length (number of nucleotides); predicted openreading frames for hypothetical proteins are indicated by boxes.


REPORTS








Information.



SOM Text

Figs. S1 to S6

References


10.1126/science.1131262

























(AMD) (10).




















species.































1 13,236

ARMAN-1

ARMAN-2


31% 85%69% 90% 97%


4,2261




REPORTS








Information.



SOM Text

Figs. S1 to S6

References


10.1126/science.1131262

























(AMD) (10).




















species.































1 13,236

ARMAN-1

ARMAN-2


31% 85%69% 90% 97%


4,2261




REPORTS

Eukaryotes have evolved surveillance mechanisms that are intimately linked to translation to eliminate er-

rors in mRNA biogenesis. The decay of transcripts containing PTCs by the NMD pathway effectively prevents expression of deleterious truncated proteins. In pro-karyotes, protein products encoded by transcripts lacking termination codons are marked for degradation by the addition of a COOH-terminal tag encoded by tmRNA (1, 2). Thus, both the presence and context of translational termination can regulate

gene expression. In order to determine whether the presence of translational ter-mination influences mRNA stability, we assayed PGK1 transcripts in Saccharo-myces cerevisiae derived from the follow-ing constructs (3): wild-type PGK1 (WT-PGK1), a nonsense form of PGK1 harbor-ing a PTC at codon 22 [PTC(22)-PGK1], and nonstop-PGK1 that was created by removing the bona fide termination codon and all in-frame termination codons in the 3′ UTR (untranslated region) from the WT-PGK1 transcript. Nonstop-PGK1 tran-scripts were as labile as their nonsense-containing counterparts (Fig. 1). At least three trans-acting factors (Upf1p, Upf2p, and Upf3p) are essential for NMD in S. cerevisiae (4, 5). Remarkably, nonstop transcripts were not stabilized in strains lacking Upf1p, distinguishing the pathway








Information.



SOM Text

Figs. S1 to S6

References


10.1126/science.1131262

























(AMD) (10).




















species.































1 13,236

ARMAN-1

ARMAN-2


31% 85%69% 90% 97%


4,2261




REPORTS








Information.



SOM Text

Figs. S1 to S6

References


10.1126/science.1131262

























(AMD) (10).




















species.































1 13,236

ARMAN-1

ARMAN-2


31% 85%69% 90% 97%


4,2261




REPORTS

An mRNA Surveillance Mechanism That Eliminates Transcripts Lacking Termination CodonsPamela A. Frischmeyer,1 Ambro van Hoof,3,4 Kathryn O’Donnell,1 Anthony L. Guerrerio,2 Roy Parker,3,4 Harry C. Dietz1,4*

Translation is an important mechanism to monitor the quality of messenger RNAs (mRNAs), as exemplified by the translation-dependent recognition and degradation of transcripts harboring premature termination codons (PTCs) by the nonsense-mediated mRNA decay (NMD) pathway. We demonstrate in yeast that mRNAs lacking all termination codons are as labile as nonsense transcripts. Decay of “nonstop” transcripts in yeast requires translation but is mechanistically distinguished from NMD and the major mRNA turnover pathway that requires deadenylation, decapping, and 5′-to-3′ exonucleolytic decay. These data suggest that nonstop decay is initiated when the ribosome reaches the 3′ terminus of the message. We demonstrate multiple physiologic sources of nonstop transcripts and conservation of their accelerated decay in mammalian cells. This process regulates the stability and expression of mRNAs that fail to signal translational termination.

culosis challenge. In contrast, the unvaccinatedmacaques developed lethargy, anorexia, andwasting; they subsequently died and theyshowed evidence of miliary tuberculosis 4 to 6weeks after M. tuberculosis aerosol challenge.We realize that the BCG-vaccinated, M. tuber-culosis-infected monkeys may progress to asubclinical form of tuberculosis, because theydid exhibit detectable M. tuberculosis and itsmRNA in BAL cells (Fig. 4B). Nevertheless,this potential outcome would not negate ourobservation that a rapid recall response ofV�2V�2� T cells coincided with immune pro-tection against the acutely fatal tuberculosis inmonkeys.

Our studies provide strong evidence thatV�2V�2� T cells, like �� T cells, contributeto adaptive immune responses in mycobacterialinfections. The adaptive immune responses ofV�2V�2� T cells are indeed driven by BCGnonpeptide antigens (19). The contribution ofthese cells to vaccine protection against tuber-culosis was demonstrated in the juvenile rhesusmodel. BCG-mediated protection against thefatal form of tuberculosis has been reported inchildren, although its protective efficacy forchronic pulmonary tuberculosis in adults andmonkeys is controversial (22–29). It seems thatthe antigen specificity, TCR diversity, and re-call features of V�2V�2� T cells separate this�� T cell subset from innate cells includingperipheral mononuclear cells (PMN), mono-cytes, and natural killer (NK) cells as well asthose �� T cells that express invariant �� TCR(30, 31). The unique ability of V�2V�2� Tcells to mount rapid and large expansions inmycobacterial infections suggests that vaccine-elicited V�2V�2� T cell immunity may be bothpossible and useful. Thus, V�2V�2� T cellsmay broadly contribute to both innate and ac-quired immunity against microbial infections.

References and Notes1. K. Pfeffer, B. Schoel, H. Gulle, S. H. Kaufmann, H.

Wagner, Eur. J. Immunol. 20, 1175 (1990).2. Y. Tanaka et al., Proc. Natl. Acad. Sci. U.S.A. 91, 8175

(1994).3. M. R. Burk, L. Mori, G. De Libero, Eur. J. Immunol. 25,

2052 (1995).4. Y. Poquet et al., Res. Immunol. 147, 542 (1996).5. Y. Tanaka, C. T. Morita, E. Nieves, M. B. Brenner, B. R.

Bloom, Nature 375, 155 (1995).6. P. Constant et al., Science 264, 267 (1994).7. J. F. Bukowski, C. T. Morita, M. B. Brenner, Immunity11, 57 (1999).

8. V. Kunzmann, E. Bauer, M. Wilhelm, N. Engl. J. Med.340, 737 (1999).

9. C. T. Morita et al., Immunity 3, 495 (1995).10. C. T. Morita, R. A. Mariuzza, M. B. Brenner, Springer

Semin. Immunopathol. 22, 191 (2000).11. P. Mombaerts, J. Arnoldi, F. Russ, S. Tonegawa, S. H.

Kaufmann, Nature 365, 53 (1993).12. C. H. Ladel et al., Eur. J. Immunol. 25, 838 (1995).13. K. Hiromatsu et al., J. Exp. Med. 175, 49 (1992).14. C. D. D’Souza et al., J. Immunol. 158, 1217 (1997).15. R. Sciammas, P. Kodukula, Q. Tang, R. L. Hendricks,

J. A. Bluestone, J. Exp. Med. 185, 1969 (1997).16. T. A. Moore, B. B. Moore, M. W. Newstead, T. J.

Standiford, J. Immunol. 165, 2643 (2000).17. D. Zhou et al., J. Immunol. 162, 2204 (1999).18. Z. W. Chen et al., J. Med. Primatol. 29, 143 (2000).19. Supplemental material is available on Science Online

at www.sciencemag.org/cgi/content/full/295/5563/2255/DC1.

20. Y. Shen et al., Infect. Immun. 70, 869 (2002).21. Y. Shen et al., J. Virol. 75, 8690 (2001).22. L. C. Rodrigues, V. K. Diwan, J. G. Wheeler, Int. J.

Epidemiol. 22, 1154 (1993).23. G. A. Colditz et al., JAMA 271, 698 (1994).24. W. R. Barclay et al., Am. Rev. Respir. Dis. 107, 351

(1973).25. G. A. Colditz et al., Pediatrics 96, 29 (1995).26. B. W. Janicki, R. C. Good, P. Minden, L. F. Affronti,

W. F. Hymes, Am. Rev. Respir. Dis. 107, 359 (1973).27. E. Ribi et al., J. Infect. Dis. 123, 527 (1971).

28. D. N. McMurray, Clin. Infect. Dis. 30 (suppl. 3), S210(2000).

29. J. A. Langermans et al., Proc. Natl. Acad. Sci. U.S.A.98, 11497 (2001).

30. A. Mukasa, W. K. Born, R. L. O’Brien, J. Immunol. 162,4910 (1999).

31. S. Itohara et al., Nature 343, 754 (1990).32. Supported by NIH RO1 grants HL64560 and RR13601

(to Z.W.C.). We thank members at Viral Pathogenesisfor technical support and W. R. Jacobs at AlbertEinstein College of Medicine for providing M. tuber-culosis H37Rv used in the study.

10 December 2001; accepted 7 February 2002

An mRNA SurveillanceMechanism That Eliminates

Transcripts Lacking TerminationCodons

Pamela A. Frischmeyer,1 Ambro van Hoof,3,4

Kathryn O’Donnell,1 Anthony L. Guerrerio,2 Roy Parker,3,4

Harry C. Dietz1,4*

Translation is an important mechanism to monitor the quality of messengerRNAs (mRNAs), as exemplified by the translation-dependent recognition anddegradation of transcripts harboring premature termination codons (PTCs) bythe nonsense-mediatedmRNAdecay (NMD) pathway.We demonstrate in yeastthat mRNAs lacking all termination codons are as labile as nonsense transcripts.Decay of “nonstop” transcripts in yeast requires translation but is mechanis-tically distinguished from NMD and the major mRNA turnover pathway thatrequires deadenylation, decapping, and 5�-to-3� exonucleolytic decay. Thesedata suggest that nonstop decay is initiated when the ribosome reaches the 3�terminus of the message. We demonstrate multiple physiologic sources ofnonstop transcripts and conservation of their accelerated decay in mammaliancells. This process regulates the stability and expression of mRNAs that fail tosignal translational termination.

Eukaryotes have evolved surveillancemechanisms that are intimately linked totranslation to eliminate errors in mRNAbiogenesis. The decay of transcripts con-taining PTCs by the NMD pathway effec-tively prevents expression of deleterioustruncated proteins. In prokaryotes, proteinproducts encoded by transcripts lacking ter-mination codons are marked for degrada-tion by the addition of a COOH-terminaltag encoded by tmRNA (1, 2). Thus, boththe presence and context of translationaltermination can regulate gene expression.In order to determine whether the presenceof translational termination influencesmRNA stability, we assayed PGK1 tran-

scripts in Saccharomyces cerevisiae de-rived from the following constructs (3):wild-type PGK1 (WT-PGK1), a nonsenseform of PGK1 harboring a PTC at codon 22[PTC(22)-PGK1], and nonstop-PGK1 thatwas created by removing the bona fidetermination codon and all in-frame termi-nation codons in the 3� UTR (untranslatedregion) from the WT-PGK1 transcript.Nonstop-PGK1 transcripts were as labile astheir nonsense-containing counterparts(Fig. 1). At least three trans-acting factors(Upf1p, Upf2p, and Upf3p) are essential forNMD in S. cerevisiae (4, 5). Remarkably,nonstop transcripts were not stabilized instrains lacking Upf1p, distinguishing thepathway of decay from NMD (Fig. 1).

The turnover of normal mRNAs requiresdeadenylation followed by Dcp1p-mediat-ed decapping and degradation by the major5�-to-3� exonuclease Xrn1p (6, 7 ). NMD isdistinguished in that these events occurwithout prior deadenylation (8, 9). Non-stop-PGK1 transcripts showed rapid decay

1Institute for Genetic Medicine, 2Department of Bio-physics and Biophysical Chemistry, Johns HopkinsUniversity School of Medicine, Baltimore, MD 21205,USA. 3University of Arizona, Department of Molecularand Cellular Biology, Tucson, AZ 85721, USA.4Howard Hughes Medical Institute.


R E P O R T S

22 MARCH 2002 VOL 295 SCIENCE www.sciencemag.org2258

on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

18

of decay from NMD (Fig. 1). The turnover of normal mRNAs re-

quires deadenylation followed by Dcp1p-mediated decapping and degradation by the major 5′-to-3′ exonuclease Xrn1p (6, 7). NMD is distinguished in that these events occur without prior deadenylation (8, 9). Nonstop-PGK1 transcripts showed rapid decay in strains lacking Xrn1p or Dcp1p activity (Fig. 1), providing fur-ther evidence that they are not subject to NMD. This result was surprising since both of these factors are also required for the turnover of normal mRNAs after deadenylation by the major deadenylase Ccr4p (10). Nonstop decay was also un-

altered in a strain lacking Ccr4p (Fig. 1). Therefore, degradation of nonstop-PGK1 transcripts requires none of the factors in-volved in the pathway for degradation of wild-type or nonsense mRNAs.

Additional experiments were performed to assess the role of translation in nonstop decay. Treatment with cycloheximide (CHX) or depletion of charged tRNAs in yeast harboring the conditional cca1-1 al-lele (11) grown at the nonpermissive tem-perature substantially increased the stabil-ity of nonstop-PGK1 transcripts (Fig. 1) (12). It has been shown that translation into the 3′ UTR of selected transcripts can displace bound trans-factors that are posi-

tive determinants of message stability (13, 14). To determine whether this mechanism is relevant to nonstop decay, we examined into the 3′ UTR of selected transcripts can

Fig. 1. Nonstop-PGK1 transcripts are rapidly degraded by a novel mechanism. Half-lives (t

1/2) of WT-PGK1, PTC(22)-PGK1, nonstop-

PGK1, and Ter-poly(A)-PGK1 transcripts are shown (27). Half-lives were performed in a wild-type yeast strain ( WT ), strains deleted for Upf1p (upf1Δ), Xrn1p (xrn1Δ), or Ccr4p (ccr4Δ), a strain lacking activity of the decapping enzyme Dcp1p (dcp1-2), and a WT yeast strain treated for 1 hour with 100 μg/ml of cycloheximide (+CHX ).

19

tive determinants of message stability (13, 14). To determine whether this mechanism is relevant to nonstop decay, we examined the performance of transcript Ter-poly(A)- PGK1 which contains a termination codon inserted one codon upstream of the site of polyadenylate [poly(A)] addition (15) in transcript nonstop-PGK1. The addition of a termination codon at the 3′ end of the 3′ UTR of a nonstop transcript resulted in substantial (threefold) stabilization (Fig. 1). Unlike nonstop-PGK1 transcripts, the half-life of Ter-poly(A)-PGK1 was in-creased in the absence of Xrn1p (Fig. 1), indicating that this transcript is degraded by the pathway for normal mRNA and not by the nonstop pathway. These data sug-gest that the instability of nonstop-PGK1 transcripts cannot fully be explained by ribosomal displacement of factors bound to the 3′ UTR.

There is an emerging view that the 5′and 3′ ends of mRNAs interact to form a closed loop and that this conformation is required for efficient translation initiation and normal mRNA stability. Participants in the interaction include components of the translation initiation complex eIF4F as well as Pab1p. The absence of a termina-tion codon in nonstop-PGK1 is predicted to allow the ribosome to continue trans-lating through the 3′ UTR and poly(A) tail, potentially resulting in displacement of Pab1p, disruption of normal mRNP structure, and consequently accelerated degradation of the transcript. However, while mRNAs in pab1Δ strains do under-go accelerated decay, the rapid turnover is a result of premature decapping and can be suppressed by mutations in XRN1 (16), allowing distinction from nonstop decay. Lability of nonstop transcripts might also be a consequence of ribosomal stalling at the 3′ end of the transcript. It is tempting to speculate that 3′-to-5′ exonucleolytic activity underlies the decapping- and 5′-to-3′ exonuclease–independent, transla-tion-dependent accelerated decay of non-stop transcripts. An appealing candidate

is the exosome, a collection of proteins with 3′-to-5′ exoribonuclease activity that functions in the processing of 5.8S RNA, rRNA, small nucleolar RNA (snoRNA), small nuclear RNA (snRNA), and other transcripts. Data presented in van Hoof et al. (17) validate this prediction.

in strains lacking Xrn1p or Dcp1p activity(Fig. 1), providing further evidence thatthey are not subject to NMD. This resultwas surprising since both of these factorsare also required for the turnover of normalmRNAs after deadenylation by the majordeadenylase Ccr4p (10). Nonstop decaywas also unaltered in a strain lacking Ccr4p(Fig. 1). Therefore, degradation of nonstop-PGK1 transcripts requires none of the fac-tors involved in the pathway for degrada-tion of wild-type or nonsense mRNAs.

Additional experiments were performedto assess the role of translation in nonstopdecay. Treatment with cycloheximide(CHX) or depletion of charged tRNAs inyeast harboring the conditional cca1-1 al-lele (11) grown at the nonpermissive tem-perature substantially increased the stabili-ty of nonstop-PGK1 transcripts (Fig. 1)(12). It has been shown that translation intothe 3� UTR of selected transcripts can dis-place bound trans-factors that are positivedeterminants of message stability (13, 14 ).To determine whether this mechanism isrelevant to nonstop decay, we examined theperformance of transcript Ter-poly(A)-PGK1 which contains a termination codon

inserted one codon upstream of the site ofpolyadenylate [poly(A)] addition (15) intranscript nonstop-PGK1. The addition of atermination codon at the 3� end of the 3�UTR of a nonstop transcript resulted insubstantial (threefold) stabilization (Fig. 1).Unlike nonstop-PGK1 transcripts, the half-life of Ter-poly(A)-PGK1 was increased inthe absence of Xrn1p (Fig. 1), indicatingthat this transcript is degraded by the path-way for normal mRNA and not by thenonstop pathway. These data suggest thatthe instability of nonstop-PGK1 transcriptscannot fully be explained by ribosomal dis-placement of factors bound to the 3� UTR.

There is an emerging view that the 5�and 3� ends of mRNAs interact to form aclosed loop and that this conformation isrequired for efficient translation initiationand normal mRNA stability. Participants inthe interaction include components of thetranslation initiation complex eIF4F as wellas Pab1p. The absence of a terminationcodon in nonstop-PGK1 is predicted to allowthe ribosome to continue translating throughthe 3� UTR and poly(A) tail, potentially re-sulting in displacement of Pab1p, disruptionof normal mRNP structure, and consequently

accelerated degradation of the transcript.However, while mRNAs in pab1� strains doundergo accelerated decay, the rapid turnoveris a result of premature decapping and can besuppressed by mutations in XRN1 (16 ), al-lowing distinction from nonstop decay. La-bility of nonstop transcripts might also be aconsequence of ribosomal stalling at the 3�end of the transcript. It is tempting to specu-late that 3�-to-5� exonucleolytic activityunderlies the decapping- and 5�-to-3� exo-nuclease–independent, translation-dependentaccelerated decay of nonstop transcripts. Anappealing candidate is the exosome, a collec-tion of proteins with 3�-to-5� exoribonucleaseactivity that functions in the processing of5.8S RNA, rRNA, small nucleolar RNA

Fig. 1. Nonstop-PGK1 transcripts are rapidlydegraded by a novel mechanism. Half-lives(t1/2) of WT-PGK1, PTC(22)-PGK1, nonstop-PGK1, and Ter-poly(A)-PGK1 transcripts areshown (27). Half-lives were performed in awild-type yeast strain ( WT ), strains deleted forUpf1p (upf1�), Xrn1p (xrn1�), or Ccr4p(ccr4�), a strain lacking activity of the decap-ping enzyme Dcp1p (dcp1-2), and a WT yeaststrain treated for 1 hour with 100 �g/ml of cycloheximide (�CHX ).

Fig. 2. Nonstop decay is conserved in mamma-lian cells. (A) Steady-state abundance of �-glu-curonidase transcripts derived from wild type–( WT ), nonsense-, nonstop-, and Ter-poly(A)-�gluc minigene constructs after transienttransfection into HeLa cells (28). Zeocin (zeo)transcripts served as a control for loading andtransfection efficiency. For each construct, the�gluc/Zeo transcript ratio is expressed relativeto that observed for WT-�gluc. (B) Steady-state abundance of WT-�gluc, nonstop-�gluc,and nonsense-�gluc transcripts were deter-mined in the nuclear (N) and cytoplasmic (C)fractions of transfected cells (28). The ratio ofnormalized nonstop- or nonsense-�gluc tran-script levels to WT-�gluc transcript levels areshown for each cellular compartment. Reportedresults are the average of three independentexperiments.

R E P O R T S

www.sciencemag.org SCIENCE VOL 295 22 MARCH 2002 2259

on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

20

In order to determine whether nonstop decay functions in mammals, we assessed the performance of transcripts derived from β-glucuronidase (βgluc) minigene constructs containing exons 1, 10, 11, and 12 separated by introns derived from the endogenous gene (18). The abundance of transcripts derived from a nonstop-βgluc version of this minigene was significantly reduced relative to WT-βgluc, suggesting that a termination codon is also essential for normal mRNA stability in mamma-lian cells (Fig. 2A). Addition of a stop codon one codon upstream from the site of poly(A) addition in nonstop-βgluc [Ter-poly(A)-βgluc (15)] increased the abundance of this transcript to near–wild-type levels (Fig. 2A). Thus, all functional characteristics of nonstop transcripts in yeast appear to be relevant to mammalian systems.

Both the abundance and stability of most nonsense transcripts, including non-sense-βgluc (Fig. 2B), are reduced in the nuclear fraction of mammalian cells (19, 20). This has been interpreted to suggest that translation and NMD initiate while the mRNA is still associated with (if not within) the nuclear compartment. If translation is initiated on nucleus-associ-ated transcripts, so might nonstop decay. Subcellular fractionation studies localized mammalian nonstop decay to the cyto-plasmic compartment, providing further distinction from NMD (Fig. 2B).

The conservation of nonstop decay in yeast and mammals suggests that the pathway serves an important biologic role. There are many potential physiologic sources of nonstop transcripts that warrant consideration. Mutations in bona fide ter-mination codons would not routinely initi-ate nonstop decay due to the frequent oc-currence of in-frame termination codons in the 3′ UTR and could not plausibly provide the evolutionary pressure for maintenance of nonstop decay. In contrast, alternative use of 3′-end processing signals embedded in coding sequence has been documented

in many genes including CBP1 in yeast and the growth hormone receptor (GHR) gene in fowl (21–23). Moreover, a com-puter search of the human mRNA and S. cerevisiae open reading frame (ORF) data-bases revealed many additional genes that contain a strict consensus sequence for 3′ end cleavage and polyadenylation within their coding region (Fig. 3A). Utilization of these premature signals would direct formation of truncated transcripts that might be substrates for the nonstop decay pathway. Indeed, analysis of 3425 random yeast cDNA clones sequenced from the 3′ end (i.e., 3′ ESTs) revealed that 40 showed apparent premature polyadenylation with-in the coding region (24).

The truncated GHR transcript in fowl is apparently translated, as evidenced by its association with polysomes (22). As predicted for a substrate for nonstop de-cay, the ratio of truncated–to–full-length GHR transcripts was dramatically in-creased after treatment of cultured chick-en hepatocellular carcinoma cells with the translational inhibitor emetine (Fig. 3B). Other truncated transcripts, both bigger and smaller than the predicted 0.7-kb non-stop transcript, did not show a dramatic increase in steady-state abundance upon translational arrest suggesting that they are derived from other mRNA processing events, perhaps alternative splicing (Fig. 3B) (12). To directly test whether the non-stop mRNA pathway degrades premature-ly polyadenylated mRNAs, we analyzed CBP1 transcripts in yeast. The CBP1 gene produces a 2.2-kb full-length mRNA and a 1.2-kb species with premature 3′ end processing and polyadenylation within the coding region (25). As expected from the observation that nonstop decay requires the exosome in yeast (17), deletion of the gene encoding the exosome component Ski7p stabilized the prematurely polyad-enylated mRNA but had no effect on the stability of the full-length mRNA (Fig. 3C). These data indicate that physiologic transcripts arising from premature poly-

21

Fig. 3. Many physiologic transcripts are substrates for nonstop decay. (A) 239 human mRNAs contain a polyadenylation signal consisting of either of the two most common words for the positioning, cleavage, and downstream elements (29), separated by optimal distances (30), where the cleavage site occurs between the start and stop codons as determined from the cDNA coding sequence (CDS). Similarly 52 S. cerevisiae ORFs contained the yeast polyadenylation signal consisting of optimal upstream, positioning, and cleavage signals followed by any of nine common “U-rich” signals (24) with optimal spacing of the elements. Computer search was performed using the human mRNA database and the S. cerevisiae ORF database and an analysis program written in PERL that is available upon request. (B) Abundance of nonstop and full-length cGHR transcripts in chicken hepatocellular carcinoma cells (CRL-2117 purchased from the American Type Culture Collection, Manassas, VA) treated with 100 μg/ml of emetine for 6 hours. Northern analysis was performed as described (22). The relative ratio of each transcript in untreated and treated cells is reported. (C) Half-lives (t

1/2) of

the full-length and nonstop CBP1 mRNAs measured in wild-type–(WT) and SKI7-deleted (ski7Δ) yeast strains carrying plasmid pG::-26 (25, 31). (D) Half-lives of WT-PGK1 and Ter-poly(A)-PGK1 transcripts in wild-type–(WT ) and SKI7-deleted (ski7Δ) yeast strains treated with the indicated dose of paromomycin (mg/ml) (PM) for 20 hours. For cycloheximide (CHX ) experiments, yeast were treated identically except 100 μg/ml of CHX was added during the last hour before half-life determination. Half-lives were determined as described (27).

22

developmental and/or homeostatic func-tions that are regulated by nonstop decay (21, 26).

Many processes contribute to the pre-cise control of gene expression including transcriptional and translational control mechanisms. In recent years, mRNA sta-bility has emerged as a major determinant of both the magnitude and fidelity of gene expression. Perhaps the most striking and comprehensively studied example is the accelerated decay of transcripts harboring PTCs by the NMD pathway. Nonstop de-cay now serves as an additional example of the critical role that translation plays in monitoring the fidelity of gene expres-sion, the stability of aberrant or atypical transcripts, and hence the abundance of truncated proteins.

adenylation are subject to nonstop mRNA decay. The cytoplasmic localization of ski7p is consistent with our observation that nonstop decay occurs within the cyto-plasm (Fig. 2B).

Any event that diminishes translational fidelity and promotes readthrough of ter-mination codons could plausibly result in the generation of substrates for nonstop decay. In view of recent attempts to treat genetic disorders resulting from PTCs with long-term and high-dose aminogly-coside regimens, this may achieve medical significance. As a proof-of-concept exper-iment, we examined the performance of transcripts with one [Ter-poly(A)-PGK1] or multiple (WT-PGK1) in-frame termi-nation codons (including those in the 3′ UTR) in yeast strains after treatment with paromomycin, which induces ribosomal readthrough. Remarkably, both transcripts showed a dose-dependent decrease in sta-bility that could be reversed by inhibiting translational elongation with CHX or pre-vented by deletion of the gene encoding Ski7p (Fig. 3D). These data suggest that nonstop decay can limit the efficiency of therapeutic strategies aimed at enhancing nonsense suppression and might contrib-ute to the toxicity associated with amino-glycoside therapy.

The degradation of nonstop transcripts may be regulated. The relative expression level of truncated GHR transcripts com-pared to full-length transcripts varies in a tissue-, gender-, and developmental stage-specific manner (22) and the relative abun-dance of truncated CBP1 transcripts varies with growth condition (21). Data present-ed here warrant the hypothesis that regula-tion may occur at the level of nonstop tran-script stability rather than production. The conservation of the GHR coding sequence 3′- end processing signal throughout avian phylogeny and conservation of premature polyadenylation of the yeast RNA14 tran-script and its fruitfly homolog su(f ) sup-port speculation that nonstop transcripts or derived protein products serve essential

prematurely polyadenylated mRNA but hadno effect on the stability of the full-lengthmRNA (Fig. 3C). These data indicate thatphysiologic transcripts arising from prema-ture polyadenylation are subject to nonstopmRNA decay. The cytoplasmic localizationof ski7p is consistent with our observationthat nonstop decay occurs within the cyto-plasm (Fig. 2B).

Any event that diminishes translationalfidelity and promotes readthrough of termi-nation codons could plausibly result in thegeneration of substrates for nonstop decay.In view of recent attempts to treat geneticdisorders resulting from PTCs with long-term and high-dose aminoglycoside regi-mens, this may achieve medical signifi-cance. As a proof-of-concept experiment,we examined the performance of transcriptswith one [Ter-poly(A)-PGK1] or multiple(WT-PGK1) in-frame termination codons(including those in the 3� UTR) in yeaststrains after treatment with paromomycin,which induces ribosomal readthrough. Re-markably, both transcripts showed a dose-dependent decrease in stability that couldbe reversed by inhibiting translational elon-gation with CHX or prevented by deletionof the gene encoding Ski7p (Fig. 3D).These data suggest that nonstop decay canlimit the efficiency of therapeutic strategiesaimed at enhancing nonsense suppressionand might contribute to the toxicity associ-ated with aminoglycoside therapy.

The degradation of nonstop transcripts maybe regulated. The relative expression level oftruncated GHR transcripts compared to full-length transcripts varies in a tissue-, gender-,and developmental stage-specific manner (22)and the relative abundance of truncated CBP1transcripts varies with growth condition (21).Data presented here warrant the hypothesis thatregulation may occur at the level of nonstoptranscript stability rather than production. Theconservation of the GHR coding sequence 3�-end processing signal throughout avian phylog-eny and conservation of premature polyadenyl-ation of the yeast RNA14 transcript and itsfruitfly homolog su(f ) support speculation thatnonstop transcripts or derived protein productsserve essential developmental and/or homeo-static functions that are regulated by nonstopdecay (21, 26).

Many processes contribute to the precisecontrol of gene expression including tran-scriptional and translational control mecha-nisms. In recent years, mRNA stability hasemerged as a major determinant of both themagnitude and fidelity of gene expression.Perhaps the most striking and comprehen-sively studied example is the accelerated de-cay of transcripts harboring PTCs by theNMD pathway. Nonstop decay now serves asan additional example of the critical role thattranslation plays in monitoring the fidelity of

gene expression, the stability of aberrant oratypical transcripts, and hence the abundanceof truncated proteins.

References and Notes1. A. W. Karzai, E. D. Roche, R. T. Sauer, Nature Struct.

Biol. 7, 449 (2000).2. K. C. Keiler, P. R. Waller, R. T. Sauer, Science 271, 990

(1996).3. The nonstop-PGK1 construct was generated from

WT-PGK1 [pRP469 (8)] by creating three point mu-tations, using site-directed mutagenesis (Quik-Change Site-Directed Mutagenesis Kit, Stratagene)which eliminated the bona fide termination codonand all in-frame termination codons in the 3� UTR.The Ter-poly(A)-PGK1 construct was created fromthe nonstop-PGK1 construct, using site-directed mu-tagenesis (Quik-Change Site-Directed MutagenesisKit, Stratagene). All changes were confirmed by se-quencing. Primer sequences are available upon re-quest. The PTC(22)-PGK1 construct is described in (8)(pRP609).

4. P. Leeds, J. M. Wood, B. S. Lee, M. R. Culbertson, Mol.Cell. Biol. 12, 2165 (1992).

5. Y. Cui, K. W. Hagan, S. Zhang, S. W. Peltz, Genes Dev.9, 423 (1995).

6. D. Muhlrad, C. J. Decker, R. Parker, Genes Dev. 8, 855(1994).

7. C. A. Beelman et al., Nature 382, 642 (1996).8. D. Muhlrad, R. Parker, Nature 370, 578 (1994).9. G. Caponigro, R. Parker, Microbiol. Rev. 60, 233

(1996).10. M. Tucker et al., Cell 104, 377 (2001).11. C. L. Wolfe, A. K. Hopper, N. C. Martin, J. Biol. Chem.271, 4679 (1996).

12. P. A. Frischmeyer, H. C. Dietz, data not shown.13. I. M. Weiss, S. A. Liebhaber, Mol. Cell. Biol. 14, 8123

(1994).14. X. Wang, M. Kiledjian, I. M. Weiss, S. A. Liebhaber,

Mol. Cell. Biol. 15, 1769 (1995).15. We performed 3� rapid amplification of cDNA ends

(RACE) using the Marathon cDNA Amplification Kit(Clontech) and the Advantage 2 Polymerase Mix(Clontech) according to the manufacturer’s instruc-tions. Poly(A) RNA isolated from cells (OligotexmRNA midi kit, Qiagen) transiently transfected withWT-�gluc or from a wild-type yeast strain carrying aplasmid expressing WT-PGK1 was used as template.A single RACE product was generated, cloned ( TOPOTA 2.1 kit, Invitrogen), and sequenced.

16. G. Caponigro, R. Parker, Genes Dev. 9, 2421 (1995).17. A. van Hoof, P. A. Frischmeyer, H. C. Dietz, R. Parker,

Science 295, 2262 (2002).18. To create the WT-�gluc minigene construct, mouse

genomic DNA was used as template for amplifyingthree different portions of the �-glucuronidase gene:from nucleotide 2313 to 3030 encompassing exon 1and part of intron 1 (including the splice donor se-quence), from nucleotide 11285 to 13555 encompass-ing part of intron 9 (including the splice acceptor se-quence ), exon 10, and part of intron 10 (including thesplice donor sequence), and from 13556 to 15813which included the remainder of intron 10, exon 11,intron 11, and exon 12. All three fragments were TAcloned ( Topo TA 2.1 kit, Invitrogen), sequenced, andthen cloned into pZeoSV2 (Invitrogen). A single base-pair deletion corresponding to the gusmps allele wasgenerated by site-directed mutagenesis (Quik-ChangeSite-Directed Mutagenesis Kit, Stratagene) of the WT-�gluc construct to create nonsense-�gluc. Nonstop-�gluc was generated from WT-�gluc by two rounds ofsite-directed mutagenesis (Quik-Change Site-DirectedMutagenesis Kit, Stratagene), which removed the bonafide termination codon and all in-frame terminationcodons in the 3� UTR. Ter-poly(A)-�gluc was generatedfrom nonstop-�gluc via a single point mutation, whichcreated a stop one codon upstream of the poly(A) tail.All changes were confirmed by sequencing. Primer se-quences are available upon request.

19. P. Belgrader, J. Cheng, X. Zhou, L. S. Stephenson, L. E.Maquat, Mol. Cell. Biol. 14, 8219 (1994).

20. O. Kessler, L. A. Chasin, Mol. Cell. Biol. 16, 4426(1996).

21. K. A. Sparks, C. L. Dieckmann, Nucleic Acids Res. 26,4676 (1998).

22. E. R. Oldham, B. Bingham, W. R. Baumbach, Mol.Endocrinol. 7, 1379 (1993).

23. C. Hilger, I. Velhagen, H. Zentgraf, C. H. Schroder,J. Virol. 65, 4284 (1991).

24. J. H. Graber, C. R. Cantor, S. C. Mohr, T. F. Smith,Nucleic Acids Res. 27, 888 (1999).

25. K. A. Sparks, S. A. Mayer, C. L. Dieckmann, Mol. Cell.Biol. 17, 4199 (1997).

26. A. Audibert, M. Simonelig, Proc. Natl. Acad. Sci.U.S.A. 95, 14302 (1998).

27. All PGK1 minigene constructs are under the control ofthe GAL1 upstream activation sequence (UAS). Cultureswere grown to mid-log phase in synthetic completemedia–uracil (SC-ura) containing 2% galactose (tran-scription on). Cells were pelleted and resuspended inSC-ura media. Glucose was then added (transcriptionoff ) to a final concentration of 2% and aliquots wereremoved at the indicated time points. Approximately10 �g of total RNA isolated with the hot phenol meth-od (4) was electrophoresed on a 1.2% agarose formal-dehyde gel, transferred to a nylon membrane (Gene-Screen Plus, NEN), and hybridized with a 23 bp oligo-nucleotide end-labeled radioactive probe that specifi-cally detects the polyG tract in the 3� UTR of the PGK1transcripts (8). Half-lives were determined by plottingthe percent of mRNA remaining versus time on a semi-log plot. All half-lives were determined at 25°C exceptfor the ccr4� (performed at 30°C) and dcp1-2 [per-formed after shift to the nonpermissive temp (37°C) for1 hour] experiments.

28. HeLa cells [maintained in DMEM with 10% fetalbovine serum (FBS)] were grown to �80 to 90%confluency, trypsinized, and resuspended in 300 �l ofRPMI 1640 without FBS, 10 mM glucose and 0.1 mMdithiothreitol (DTT). Plasmid DNA (10 �g) was addedand cells were electroporated (Bio-Rad, Gene Pulser IIElectroporation System) at a voltage of 0.300 kV anda capacitance of 500 �F. Poly(A) RNA (Oligotex midikit, Qiagen, 2 �g), isolated 36 hours after transfec-tion, was separated on a 1.6% agarose formaldehydegel, transferred to a nylon membrane, and hybridizedwith a polymerase chain reaction (PCR) product com-prising exons 10 through 12 of the �-gluc gene thathad been radioactively labeled by nick translation(Random Primed DNA Labeling Kit, Boehringer Mann-heim). The membrane was subsequently strippedwith boiling 0.5% SDS and probed with radioactivelylabeled zeocin cDNA. Northern blot results werequantitated using an Instant Imager (Packard). Forsubcellular fractionation, transiently transfectedHeLa cells were trypsinized, washed in cold 1� phos-phate buffered saline, and pelleted by centrifuging at2040g for 5 min at 4°C. Cells were then resuspendedin 200 �l of 140 mM NaCl, 1.5 mM MgCl2, 10 mMTris-HCl (pH 8.6), 0.5% NP-40, and 1 mM DTT,vortexed, and incubated on ice for 5 min. Nuclei werepelleted by centrifuging at 12,000g for 45 s at 4°Cand subsequently separated from the aqueous (cyto-plasmic) phase.

29. J. H. Graber, C. R. Cantor, S. C. Mohr, T. F. Smith, Proc.Natl. Acad. Sci. U.S.A. 96, 14055 (1999).

30. F. Chen, C. C. MacDonald, J. Wilusz, Nucleic Acids Res.23, 2614 (1995).

31. Plasmid pG-26 (25) containing the CBP1 gene un-der the control of the GAL promoter and the LEU2selectable marker was introduced into wild-typeand SKI7-deleted strains. Yeast were grown in me-dia lacking leucine and containing 2% galactose at30°C to an OD600 of 0.5. Cells were pelleted andresuspended in media lacking a carbon source.Glucose was then added to a final concentration of2% at time 0 and time points were taken. RNA wasextracted and analyzed by northern blotting with alabeled oligo probe oRP1067 (5�-CTCGGTCCTG-TACCGAACGAGACGAGG-3�).

32. We thank N. Martin, A. Atkin, C. Dieckmann, and M.Sands for the provision of valuable reagents. Thiswork was supported by a grant from NIH (GM55239)(H.C.D.), the Howard Hughes Medical Institute(H.C.D, R.P.), and the Medical Scientist Training Pro-gram (P.A.F.).

22 October 2001; accepted 31 January 2002

R E P O R T S


on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

23





































R E P O R T S


on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from





































R E P O R T S


on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from





































R E P O R T S


on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from





































R E P O R T S


on

June

29,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

24

The old notion of natural selection as an omnipotent force in biologi-cal evolution has given way to one

where adaptive processes are constrained by physical, chemical, and biological exi-gencies (1–4). Whether constraint and/or stabilizing selection explain phenotypic stasis, in the fossil record and in phylog-enies, remains an open question (5). Direct experimental tests of constraint are scarce (6–9). Even tight correlations among traits, at once suggestive of constraint, can be broken by artificial selection to produce new phenotypic combinations (8, 9). De-spite all circumstantial evidence, results from direct experimental tests imply that selection is largely unconstrained.

The direct experimental test for con-straint is conceptually simple. A pheno-type is subjected to selection (natural or artificial) in an attempt to break the pos-tulated constraint (6–9). A response to se-lection indicates a lack of constraint. No response to selection indicates the pres-ence of a constraint. However, the cause of a constraint is rarely specified because the etiologies of most phenotypes are not well understood, their relationships to fitness

are usually opaque, and a lack of response to selection may reflect nothing more than a lack of heritable variation (4, 7). If the cause of a constraint is to be elucidated, it must be for a simple phenotype whose relationship to fitness is understood.

Coenzyme use by β-isopropylmalate dehydrogenase (IMDH) is a simple phe-notype whose etiology and relationship to fitness are understood (10, 11). IMDHs catalyze the oxidative decarboxylation of β-isopropylmalate to α-ketoisocaproate during the biosynthesis of leucine, an es-sential amino acid. All IMDHs use nico-tinamide adenine dinucleotide (NAD) as a coenzyme (cosubstrate). This invariance of function among IMDHs hints at the pres-ence of ancient constraints, even though some related isocitrate dehydrogenases (IDHs) use NADP instead (12, 13).

Structural comparisons with related NADP-using IDHs identify amino ac-ids controlling coenzyme use (14–16) (Fig. 1A). Introducing five replacements (Asp236 → Arg, Asp289 → Lys, Ile290→ Tyr, Ala296→ Val, and Gly337→ Tyr) into the coenzyme-binding pocket of Escherichia coli leuB–encoded IMDH by site-directed mutagenesis causes a complete reversal in specificity (10, 11): NAD performance (kNAD

cat /KNADm ,where K

m is the Michae-

lis constant) is reduced by a factor of 340, from 68 × 103 M–1 s–1 to 0.2 × 103

the CCR5 agonist RANTES induced interleukin-

6 (IL-6) release from wild-type but not CCR5D32

iDCs (Fig. 4C). Furthermore, pretreatment of

iDCs with pertussis toxin or TAK-779 inhibited

the ability of peptide-loaded myHsp70 to stimu-

late DCs to generate influenza peptide–specific

CTLs (Fig. 4D). To examine the effect of CCR5-

mediated signaling in mycobacterial infection, we

used the model pathogenM. bovis BCG-lux (12).

As reported for murine DCs (13), human DCs

from immune people were unable to kill inter-

nalized mycobacteria, even in the presence of

autologous T cells. TAK-779 inhibition of CCR5

led to a dose-dependent enhancement of intra-

cellular mycobacterial replication (Fig. 4E and fig.

S6A) at all concentrations of mycobacteria tested

(fig. S6B), suggesting an important role for this

receptor in controlling mycobacterial infection.

The identification of CCR5 as the critical re-

ceptor for myHsp70-mediated DC stimulation has

implications for both mycobacterial infection and

the therapeutic use of myHsp70. CCR5 is impor-

tant in immune cell cross talk. Interaction with its

naturally occurring ligand MIP-1b promotes the

recruitment of cells to sites of inflammation (14),

facilitates immune synapse formation (15), orches-

trates T cell interactions within lymph nodes

(16), and controls the activation and differentiation

of T cells (17). Our finding that a mycobacterial

lysate, as well as purified myHsp70, stimulated a

CCR5-dependent calcium response indicates a

further connection between the innate and adaptive

immune responses during mycobacterial infection.

The cellular aggregation induced by myHsp70

signaling through CCR5 may play an important

role in the formation of granulomas, the hallmark

of mycobacterial infection.

Microbial-induced DC responses need to be

highly regulated, reflecting a balance between a

rapid and appropriate response to invading mi-

crobes and the inducement of immunopathology

(2)—a particular problem in mycobacterial in-

fection. An increasing number of human patho-

gens, including HIV (18), toxoplasma (19), and

M. tuberculosis (as described here), target the

CCR5 receptor. This role of CCR5 as a pattern-

recognition receptor for myHsp70 may, at least

in part, be responsible for the maintenance of the

high CCR5D32 allele frequency (10 to 15%) in

Northern European populations (20) and may

alter the pattern of disease seen in people with

the CCR5D32 allele.

References and Notes1. J. L. Flynn, J. Chan, Trends Microbiol. 13, 98 (2005).

2. J. L. Flynn, J. Chan, Annu. Rev. Immunol. 19, 93 (2001).

3. Y. Wang et al., J. Immunol. 169, 2422 (2002).

4. M. J. Gething, J. Sambrook, Nature 355, 33 (1992).

5. P. Srivastava, Annu. Rev. Immunol. 20, 395 (2002).

6. P. A. MacAry et al., Immunity 20, 95 (2004).

7. M. Baba et al., Proc. Natl. Acad. Sci. U.S.A. 96, 5698 (1999).

8. M. Benkirane, D. Y. Jin, R. F. Chun, R. A. Koup,

K. T. Jeang, J. Biol. Chem. 272, 30603 (1997).

9. J. E. Hildreth, F. M. Gotch, P. D. Hildreth, A. J. McMichael,

Eur. J. Immunol. 13, 202 (1983).

10. P. Revy, M. Sospedra, B. Barbour, A. Trautmann,

Nat. Immunol. 2, 925 (2001).

11. M. L. Dustin, T. G. Bivona, M. R. Philips, Nat. Immunol. 5,

363 (2004).

12. B. Kampmann et al., J. Infect. Dis. 182, 895 (2000).

13. K. A. Bodnar, N. V. Serbina, J. L. Flynn, Infect. Immun.

69, 800 (2001).

14. J. G. Cyster, Annu. Rev. Immunol. 23, 127 (2005).

15. M. Nieto et al., J. Exp. Med. 186, 153 (1997).

16. F. Castellino et al., Nature 440, 890 (2006).

17. S. A. Luther, J. G. Cyster, Nat. Immunol. 2, 102 (2001).

18. T. Dragic et al., Nature 381, 667 (1996).

19. J. Aliberti et al., Nat. Immunol. 4, 485 (2003).

20. E. de Silva, M. P. Stumpf, FEMS Microbiol. Lett. 241, 1 (2004).

21. We are grateful to M. Mahaut-Smith for the use of the Cairn

Spectrophotometer and A. Betz for CCR5–/– mice. Funded by

the Wellcome Trust (P.J.L.), the Medical Research Council

(R.A.F.), the Swiss National Science Foundation ( J.P.), and

FEBS (E.H.). P.J.L. holds a Lister Institute Research Prize.



Figs. S1 to S7

Table S1

Videos S1 to S4

References

26 May 2006; accepted 8 September 2006

10.1126/science.1133515

Direct Demonstration of anAdaptive ConstraintStephen P. Miller,1 Mark Lunzer,1 Antony M. Dean1,2*

The role of constraint in adaptive evolution is an open question. Directed evolution of an engineeredb-isopropylmalate dehydrogenase (IMDH), with coenzyme specificity switched from nicotinamideadenine dinucleotide (NAD) to nicotinamide adenine dinucleotide phosphate (NADP), always producesmutants with lower affinities for NADP. This result is the correlated response to selection for relieffrom inhibition by NADPH (the reduced form of NADP) expected of an adaptive landscape subject tothree enzymatic constraints: an upper limit to the rate of maximum turnover (kcat), a correlation inNADP and NADPH affinities, and a trade-off between NAD and NADP usage. Two additional constraints,high intracellular NADPH abundance and the cost of compensatory protein synthesis, have ensuredthe conserved use of NAD by IMDH throughout evolution. Our results show that selective mechanismsand evolutionary constraints are to be understood in terms of underlying adaptive landscapes.

The old notion of natural selection as an

omnipotent force in biological evolution

has given way to one where adaptive pro-

cesses are constrained by physical, chemical, and

biological exigencies (1–4). Whether constraint

and/or stabilizing selection explain phenotypic

stasis, in the fossil record and in phylogenies, re-

mains an open question (5). Direct experimental

tests of constraint are scarce (6–9). Even tight

correlations among traits, at once suggestive of

constraint, can be broken by artificial selection to

produce new phenotypic combinations (8, 9). De-

spite all circumstantial evidence, results from

direct experimental tests imply that selection is

largely unconstrained.

The direct experimental test for constraint is

conceptually simple. A phenotype is subjected to

selection (natural or artificial) in an attempt to

break the postulated constraint (6–9). A response

to selection indicates a lack of constraint. No re-

sponse to selection indicates the presence of a

constraint. However, the cause of a constraint is

rarely specified because the etiologies of most

phenotypes are not well understood, their relation-

ships to fitness are usually opaque, and a lack of

response to selection may reflect nothing more

than a lack of heritable variation (4, 7). If the

cause of a constraint is to be elucidated, it must

be for a simple phenotype whose relationship to

fitness is understood.

Coenzyme use by b-isopropylmalate dehydro-

genase (IMDH) is a simple phenotype whose

etiology and relationship to fitness are understood

(10, 11). IMDHs catalyze the oxidative decarboxyl-

ation of b-isopropylmalate to a-ketoisocaproate

during the biosynthesis of leucine, an essential

amino acid. All IMDHs use nicotinamide adenine

dinucleotide (NAD) as a coenzyme (cosubstrate).

This invariance of function among IMDHs hints

at the presence of ancient constraints, even though

some related isocitrate dehydrogenases (IDHs)

use NADP instead (12, 13).

Structural comparisons with related NADP-

using IDHs identify amino acids controlling co-

enzyme use (14–16) (Fig. 1A). Introducing five

replacements (Asp236 Y Arg, Asp289 Y Lys,

Ile290Y Tyr, Ala296Y Val, and Gly337Y Tyr)

into the coenzyme-binding pocket of Escherichia

coli leuB–encoded IMDH by site-directed muta-

genesis causes a complete reversal in specificity

(10, 11): NAD performance (kcatNAD/K

mNAD, where

Kmis the Michaelis constant) is reduced by a

factor of 340, from 68 103 M–1 s–1 to 0.2

103 M–1 s–1, whereas NADP performance

(kcatNADP/K

mNADP) is increased by a factor of 70,

from 0.49 103 M–1 s–1 to 34 103 M–1 s–1.

The engineered LeuBERKYVYR^ mutant (the

final R represents Arg341, already present in

wild-type E. coli IMDH) is as active and as spe-

cific toward NADP as the wild-type enzyme is

1BioTechnology Institute, 2Department of Ecology, Evolu-tion and Behavior, University of Minnesota, St. Paul, MN55108, USA.


REPORTS

20 OCTOBER 2006 VOL 314 SCIENCE www.sciencemag.org458

Direct Demonstration of an Adaptive Constraint

the CCR5 agonist RANTES induced interleukin-

6 (IL-6) release from wild-type but not CCR5D32

iDCs (Fig. 4C). Furthermore, pretreatment of

iDCs with pertussis toxin or TAK-779 inhibited

the ability of peptide-loaded myHsp70 to stimu-

late DCs to generate influenza peptide–specific

CTLs (Fig. 4D). To examine the effect of CCR5-

mediated signaling in mycobacterial infection, we

used the model pathogenM. bovis BCG-lux (12).

As reported for murine DCs (13), human DCs

from immune people were unable to kill inter-

nalized mycobacteria, even in the presence of

autologous T cells. TAK-779 inhibition of CCR5

led to a dose-dependent enhancement of intra-

cellular mycobacterial replication (Fig. 4E and fig.

S6A) at all concentrations of mycobacteria tested

(fig. S6B), suggesting an important role for this

receptor in controlling mycobacterial infection.

The identification of CCR5 as the critical re-

ceptor for myHsp70-mediated DC stimulation has

implications for both mycobacterial infection and

the therapeutic use of myHsp70. CCR5 is impor-

tant in immune cell cross talk. Interaction with its

naturally occurring ligand MIP-1b promotes the

recruitment of cells to sites of inflammation (14),

facilitates immune synapse formation (15), orches-

trates T cell interactions within lymph nodes

(16), and controls the activation and differentiation

of T cells (17). Our finding that a mycobacterial

lysate, as well as purified myHsp70, stimulated a

CCR5-dependent calcium response indicates a

further connection between the innate and adaptive

immune responses during mycobacterial infection.

The cellular aggregation induced by myHsp70

signaling through CCR5 may play an important

role in the formation of granulomas, the hallmark

of mycobacterial infection.

Microbial-induced DC responses need to be

highly regulated, reflecting a balance between a

rapid and appropriate response to invading mi-

crobes and the inducement of immunopathology

(2)—a particular problem in mycobacterial in-

fection. An increasing number of human patho-

gens, including HIV (18), toxoplasma (19), and

M. tuberculosis (as described here), target the

CCR5 receptor. This role of CCR5 as a pattern-

recognition receptor for myHsp70 may, at least

in part, be responsible for the maintenance of the

high CCR5D32 allele frequency (10 to 15%) in

Northern European populations (20) and may

alter the pattern of disease seen in people with

the CCR5D32 allele.

References and Notes1. J. L. Flynn, J. Chan, Trends Microbiol. 13, 98 (2005).

2. J. L. Flynn, J. Chan, Annu. Rev. Immunol. 19, 93 (2001).

3. Y. Wang et al., J. Immunol. 169, 2422 (2002).

4. M. J. Gething, J. Sambrook, Nature 355, 33 (1992).

5. P. Srivastava, Annu. Rev. Immunol. 20, 395 (2002).

6. P. A. MacAry et al., Immunity 20, 95 (2004).

7. M. Baba et al., Proc. Natl. Acad. Sci. U.S.A. 96, 5698 (1999).

8. M. Benkirane, D. Y. Jin, R. F. Chun, R. A. Koup,

K. T. Jeang, J. Biol. Chem. 272, 30603 (1997).

9. J. E. Hildreth, F. M. Gotch, P. D. Hildreth, A. J. McMichael,

Eur. J. Immunol. 13, 202 (1983).

10. P. Revy, M. Sospedra, B. Barbour, A. Trautmann,

Nat. Immunol. 2, 925 (2001).

11. M. L. Dustin, T. G. Bivona, M. R. Philips, Nat. Immunol. 5,

363 (2004).

12. B. Kampmann et al., J. Infect. Dis. 182, 895 (2000).

13. K. A. Bodnar, N. V. Serbina, J. L. Flynn, Infect. Immun.

69, 800 (2001).

14. J. G. Cyster, Annu. Rev. Immunol. 23, 127 (2005).

15. M. Nieto et al., J. Exp. Med. 186, 153 (1997).

16. F. Castellino et al., Nature 440, 890 (2006).

17. S. A. Luther, J. G. Cyster, Nat. Immunol. 2, 102 (2001).

18. T. Dragic et al., Nature 381, 667 (1996).

19. J. Aliberti et al., Nat. Immunol. 4, 485 (2003).

20. E. de Silva, M. P. Stumpf, FEMS Microbiol. Lett. 241, 1 (2004).

21. We are grateful to M. Mahaut-Smith for the use of the Cairn

Spectrophotometer and A. Betz for CCR5–/– mice. Funded by

the Wellcome Trust (P.J.L.), the Medical Research Council

(R.A.F.), the Swiss National Science Foundation ( J.P.), and

FEBS (E.H.). P.J.L. holds a Lister Institute Research Prize.



Figs. S1 to S7

Table S1

Videos S1 to S4

References

26 May 2006; accepted 8 September 2006

10.1126/science.1133515

Direct Demonstration of anAdaptive ConstraintStephen P. Miller,1 Mark Lunzer,1 Antony M. Dean1,2*

The role of constraint in adaptive evolution is an open question. Directed evolution of an engineeredb-isopropylmalate dehydrogenase (IMDH), with coenzyme specificity switched from nicotinamideadenine dinucleotide (NAD) to nicotinamide adenine dinucleotide phosphate (NADP), always producesmutants with lower affinities for NADP. This result is the correlated response to selection for relieffrom inhibition by NADPH (the reduced form of NADP) expected of an adaptive landscape subject tothree enzymatic constraints: an upper limit to the rate of maximum turnover (kcat), a correlation inNADP and NADPH affinities, and a trade-off between NAD and NADP usage. Two additional constraints,high intracellular NADPH abundance and the cost of compensatory protein synthesis, have ensuredthe conserved use of NAD by IMDH throughout evolution. Our results show that selective mechanismsand evolutionary constraints are to be understood in terms of underlying adaptive landscapes.

The old notion of natural selection as an

omnipotent force in biological evolution

has given way to one where adaptive pro-

cesses are constrained by physical, chemical, and

biological exigencies (1–4). Whether constraint

and/or stabilizing selection explain phenotypic

stasis, in the fossil record and in phylogenies, re-

mains an open question (5). Direct experimental

tests of constraint are scarce (6–9). Even tight

correlations among traits, at once suggestive of

constraint, can be broken by artificial selection to

produce new phenotypic combinations (8, 9). De-

spite all circumstantial evidence, results from

direct experimental tests imply that selection is

largely unconstrained.

The direct experimental test for constraint is

conceptually simple. A phenotype is subjected to

selection (natural or artificial) in an attempt to

break the postulated constraint (6–9). A response

to selection indicates a lack of constraint. No re-

sponse to selection indicates the presence of a

constraint. However, the cause of a constraint is

rarely specified because the etiologies of most

phenotypes are not well understood, their relation-

ships to fitness are usually opaque, and a lack of

response to selection may reflect nothing more

than a lack of heritable variation (4, 7). If the

cause of a constraint is to be elucidated, it must

be for a simple phenotype whose relationship to

fitness is understood.

Coenzyme use by b-isopropylmalate dehydro-

genase (IMDH) is a simple phenotype whose

etiology and relationship to fitness are understood

(10, 11). IMDHs catalyze the oxidative decarboxyl-

ation of b-isopropylmalate to a-ketoisocaproate

during the biosynthesis of leucine, an essential

amino acid. All IMDHs use nicotinamide adenine

dinucleotide (NAD) as a coenzyme (cosubstrate).

This invariance of function among IMDHs hints

at the presence of ancient constraints, even though

some related isocitrate dehydrogenases (IDHs)

use NADP instead (12, 13).

Structural comparisons with related NADP-

using IDHs identify amino acids controlling co-

enzyme use (14–16) (Fig. 1A). Introducing five

replacements (Asp236 Y Arg, Asp289 Y Lys,

Ile290Y Tyr, Ala296Y Val, and Gly337Y Tyr)

into the coenzyme-binding pocket of Escherichia

coli leuB–encoded IMDH by site-directed muta-

genesis causes a complete reversal in specificity

(10, 11): NAD performance (kcatNAD/K

mNAD, where

Kmis the Michaelis constant) is reduced by a

factor of 340, from 68 103 M–1 s–1 to 0.2

103 M–1 s–1, whereas NADP performance

(kcatNADP/K

mNADP) is increased by a factor of 70,

from 0.49 103 M–1 s–1 to 34 103 M–1 s–1.

The engineered LeuBERKYVYR^ mutant (the

final R represents Arg341, already present in

wild-type E. coli IMDH) is as active and as spe-

cific toward NADP as the wild-type enzyme is

1BioTechnology Institute, 2Department of Ecology, Evolu-tion and Behavior, University of Minnesota, St. Paul, MN55108, USA.


REPORTS


25

M–1 s–1, whereas NADP performance (kNADP

cat /KNADPm ) is increased by a factor of

70, from 0.49 × 103 M–1 s–1 to 34 × 103 M–1 s–1. The engineered LeuB[RKYVYR] mu-tant (the final R represents Arg341, already present in wild-type E. coli IMDH) is as active and as specific toward NADP as the wild-type enzyme is toward NAD. Evi-dently, protein architecture has not con-strained IMDH to use NAD since the last common ancestor.

Despite similar in vitro performances, the NADP-specific LeuB[RKYVYR] mu-

tant is less fit than the NAD-specific wild type (11). IMDHs, wild-type and mutant alike, display a factor of 30 higher affinity for the reduced form of NADP (NADPH) (coproduct) than for NADP (Fig. 1B). We suggested the LeuB[RKYVYR] mutant is subject to intense inhibition by intracellu-lar NADPH, which is far more abundant in vivo than is NADP (11, 17). The inhi-bition slows leucine biosynthesis, reduces growth rate, and lowers Darwinian fitness (Fig. 1C). The wild type retains high fit-ness because its affinity for NADPH is

toward NAD. Evidently, protein architecture has

not constrained IMDH to use NAD since the last

common ancestor.

Despite similar in vitro performances, the

NADP-specific LeuBERKYVYR^ mutant is less

fit than the NAD-specific wild type (11). IMDHs,

wild-type and mutant alike, display a factor of 30

higher affinity for the reduced form of NADP

(NADPH) (coproduct) than for NADP (Fig. 1B).

We suggested the LeuBERKYVYR^mutant is sub-

ject to intense inhibition by intracellular NADPH,

which is far more abundant in vivo than is NADP

(11, 17). The inhibition slows leucine biosynthesis,

reduces growth rate, and lowers Darwinian fitness

(Fig. 1C). Thewild type retains high fitness because

its affinity forNADPH is low,whereas inhibition by

NADH, which is far less abundant than NAD in

vivo, is ineffective. Perhaps as a consequence of

differences in Michaelis complex structure (18),

IDH is not subject to such intense NADPH in-

hibition and hence could evolve NADP use.

We hypothesize that IMDH is constrained to

use NAD because the strong inhibition associated

with NADP use reduces fitness. However, identify-

ing themechanismof selection (NADPHinhibition)

is not synonymous with identifying the causes of

constraint. Mutations that increase kcatNADP (maxi-

mum rate of NADP turnover), that break the

correlation in NADP and NADPH affinities (Fig.

1B), or that eliminate the trade-off in coenzyme

performances (Fig. 1C) could each benefit the

LeuBERKYVYR^ mutant without compromising

its performance with NADP (18). We therefore

hypothesize that IMDH is constrained to use NAD

by three causes: an upper limit to kcatNADP, an

unbreakable correlation in the affinities of NADP

and NADPH, and an inescapable trade-off in

coenzyme performance.

We used directed evolution (targeted random

mutagenesis and selective screening) (19–21) to

test whether NADP-specific IMDHs with higher

fitness could be isolated. Random substitutions

were introduced into leuB[RKYVYR] by means

of error-prone polymerase chain reaction (18).

Mutated alleles were ligated downstream of the

T7 promoter in pETcoco (a stable single-copy

vector) and transformed into strain RFSEDE3^

(leuAþBamCþ, with T7 RNA polymerase ex-

pressed from a chromosomal lacUV5 promoter).

Sequencing unscreened plasmids revealed that,

on average, each 1110–base pair leuB[RKYVYR]

received two nucleotide substitutions. From the

pattern of base substitutions in these mutants and

assuming Poisson statistics, we estimate that only

10.5 (0.4%) of the 2431 possible amino acid re-

placementsweremissing in ourmutant library (18).

Our experimental design used decreased

IMDH expression to provide a simple selective

screen for mutations in leuB[RKYVYR] that in-

crease growth and/or fitness. As predicted from

the phenotype-fitness map (Fig. 2A), selection

against leuB[RKYVYR] intensified as IMDH

expression was lowered in the presence of excess

glucose (Fig. 2B). Using 40 mM isopropyl-b-D-

thiogalactopyranoside (IPTG) to induce a low

level of IMDH expression, we found that cells

harboring leuB[WT] formed large colonies at 24

hours, whereas cells harboring leuB[RKYVYR]

Fig. 1. The molecular anatomy of an adaptive constraint. (A) Structuralalignment of the coenzyme binding pockets of E. coli IMDH (14) (brownmain chain) with bound NAD modeled from Thermus thermophilus IMDH(15) and E. coli IDH (16) (green main chain) with bound NADP. Only keyresidues are shown (gray, carbon; red, oxygen; blue, nitrogen; yellow,phosphorus) with labels designating the amino acid and site number inIMDH followed by the amino acid in IDH. Coenzyme use is determined by Hbonds to NAD (brown lines) and to the 2¶-phosphate (2¶P) of NADP (greenlines). (B) Mutations in the coenzyme binding pocket that stabilize NADPalso stabilize NADPH. The ensuing correlation in affinities for NADP (Km

NADP)

and NADPH (KiNADPH) seen with engineered mutants (circles) (11) is retained

in the screened mutants (dots). (C) The adaptive landscape for coenzyme useby IMDH (11) showing the phenotype-fitness map (blue surface) determinedin chemostat competition and described by equation S5 (11, 18) and thepredicted distribution of performances for 512 mutants in the coenzymebinding pocket (gray dots), with wild type (red dot), LeuB[RKYVYR] (blackdot), and single amino acid replacements in LeuB[RKYVYR] coenzymebinding pocket (pink dots). The fitness of LeuB[RKYVYR] is hypothesized tobe lower than in the wild type because of strong inhibition by abundantintracellular NADPH (11).

Fig. 2. The phenotypic basis of the genetic screen. Lowered expression in the presence of excess glucosebrings IMDH to saturation with isopropylmalate, increasing coenzyme affinities. (A) Lowering IMDHexpression in the chemostat-derived adaptive landscape [lower concentration of E in equation S5 (11, 18)] ispredicted to reduce the fitness of LeuB[RKYVYR] (white sphere) far more than that of the wild type (redsphere). (B) IPTG-controlled expression of IMDHs ligated downstream of the T7 promoter in pETcoco in strainRFS[DE3] (leuAþBamCþ, with T7 RNA polymerase expressed from a chromosomal lacUV5 promoter) confirmsthat lower expression affects growth in minimal glucose medium of the LeuB[RKYVYR] (white) more than inthe wild type (red). Plasmids lacking LeuB (black) are incapable of growth except in the presence of leucine.

REPORTS

www.sciencemag.org SCIENCE VOL 314 20 OCTOBER 2006 459



common ancestor.



































































NADP)




REPORTS


Fig. 1. The molecular anatomy of an adaptive constraint. (A) Structural alignment of the coenzyme binding pockets of E. coli IMDH (14) (brown main chain) with bound NAD modeled from Thermus thermophilus IMDH (15) and E. coli IDH (16) (green main chain) with bound NADP. Only key residues are shown (gray, carbon; red, oxygen; blue, nitrogen; yellow,

phosphorus) with labels designating the amino acid and site number in IMDH followed by the amino acid in IDH. Coenzyme use is determined by H bonds to NAD (brown lines) and to the 2’-phosphate (2’P) of NADP (green lines). (B) Mutations in the coenzyme binding pocket that stabilize NADP also stabilize NADPH. The ensuing correlation in affinities for NADP (KNADP

m ) and NADPH (KNADPH

i ) seen with engineered mutants (circles) (11) is retained in the screened mutants (dots). (C) The adaptive landscape for coenzyme use by IMDH (11) showing the phenotype-fitness map (blue surface) determined in chemostat competition and described by equation S5 (11, 18) and the predicted distribution of performances

for 512 mutants in the coenzyme binding pocket (gray dots), with wild type (red dot), LeuB[RKYVYR] (black dot), and single amino acid replacements in LeuB[RKYVYR] coenzyme binding pocket (pink dots). The fitness of LeuB[RKYVYR] is hypothesized to be lower than in the wild type because of strong inhibition by abundant intracellular NADPH (11)

26

low, whereas inhibition by NADH, which is far less abundant than NAD in vivo, is ineffective. Perhaps as a consequence of differences in Michaelis complex struc-ture (18), IDH is not subject to such in-tense NADPH inhibition and hence could evolve NADP use.

We hypothesize that IMDH is con-strained to use NAD because the strong inhibition associated with NADP use re-duces fitness. However, identifying the mechanism of selection (NADPH inhibi-tion) is not synonymous with identifying the causes of constraint. Mutations that increase kNADP

cat (maximum rate of NADP turnover), that break the correlation in NADP and NADPH affinities (Fig. 1B), or that eliminate the trade-off in coen-zyme performances (Fig. 1C) could each benefit the LeuB[RKYVYR] mutant with-out compromising its performance with NADP (18). We therefore hypothesize that IMDH is constrained to use NAD by three causes: an upper limit to kNADP

cat , an unbreakable correlation in the affinities of NADP and NADPH, and an inescapable trade-off in coenzyme performance.

We used directed evolution (targeted random mutagenesis and selective screen-

ing) (19–21) to test whether NADP-spe-cific IMDHs with higher fitness could be isolated. Random substitutions were introduced into leuB[RKYVYR] by means of error-prone polymerase chain reaction (18). Mutated alleles were ligated down-stream of the T7 promoter in pETcoco (a stable single-copy vector) and transformed into strain RFS[DE3] (leuA+BamC+, with T7 RNA polymerase expressed from a chro-mosomal lacUV5 promoter). Sequencing unscreened plasmids revealed that, on av-erage, each 1110–base pair leuB[RKYVYR] received two nucleotide substitutions. From the pattern of base substitutions in these mutants and assuming Poisson statis-tics, we estimate that only 10.5 (0.4%) of the 2431 possible amino acid replacements were missing in our mutant library (18).

Our experimental design used de-creased IMDH expression to provide a simple selective screen for mutations in leuB[RKYVYR] that increase growth and/or fitness. As predicted from the phenotype-fitness map (Fig. 2A), selec-tion against leuB[RKYVYR] intensified as IMDH expression was lowered in the presence of excess glucose (Fig. 2B). Using 40 μM isopropyl-β-D-thiogalacto-



common ancestor.



































































NADP)




REPORTS


27

pyranoside (IPTG) to induce a low level of IMDH expression, we found that cells harboring leuB[WT] formed large colo-nies at 24 hours, whereas cells harboring leuB[RKYVYR] barely formed pinprick colonies at 48 hours. We chose colony for-mation at 48 hours as the least stringent cri-

terion compatible with reliably identifying beneficial mutations in leuB[RKYVYR]. Longer periods of growth (or higher con-centrations of IPTG) allowed unmutated leuB[RKYVYR] to form colonies, whereas shorter periods (or lower concentrations of IPTG) produced no colonies.

barely formed pinprick colonies at 48 hours. We

chose colony formation at 48 hours as the least

stringent criterion compatible with reliably identi-

fying beneficial mutations in leuB[RKYVYR].

Longer periods of growth (or higher concentrations

of IPTG) allowed unmutated leuB[RKYVYR] to

form colonies, whereas shorter periods (or lower

concentrations of IPTG) produced no colonies.

Of the 100,000 mutated plasmids screened,

134 (representing 107 distinct isolates) formed

colonies within 48 hours (table S1). Each had

either a substitution in the 5¶ leader sequence

upstream of leuB[RKYVYR], an amino acid

replacement in the coenzyme binding pocket, or

both. Upstream substitutions occurred at three

nucleotide positions: –3, –9, and –14 relative to

the leuB AUG start codon (fig. S1). Ten amino

acid replacements were found at three codons in

the coenzyme binding pocket.

At first glance, the positive response to selection

might suggest that coenzyme use by IMDH is

unconstrained. Upstream substitutions in the Shine-

Dalgarno sequence (positions –9 and –14), as well

as a new AUG start codon (position –3) that

replaces the less efficient GUG start codon,

presumably derive their benefits through increases

in expression because their kinetics are unchanged.

Unexpectedly, however, beneficial amino acid

replacements in the coenzyme binding pocket

eliminate H bonds to the 2¶-phosphate of NADP

to cause striking reductions in NADP performance

(Table 1). Although some mutants have improved

NAD performance, others remain unchanged and

several show reduced NAD performance. Isolated

amino acid replacements outside the coenzyme

binding pocket, which might have been expected

to increase kcatNADP or to break the correlation in

affinities between NADP and NADPH (KmNADP

and KiNADPH), have no detectable functional ef-

fects (table S2, those associated with beneficial 5¶-

leader mutations in Table 1). No doubt they,

along with 126 silent substitutions (table S1),

hitchhiked through the genetic screen with the

beneficial mutations.

Our results suggest that increases in expression

are beneficial, whereas increases in NADP per-

formance are not. This seeming paradox is resolved

if there are no mutations capable of breaking the

upper limit to kcatNADP, the correlation in affinities

for NADP and NADPH, or the trade-off in co-

enzyme performance. With these constraints, and

with reduced expression in the genetic screen, the

phenotype-fitness map near LeuBERKYVYR^

remains flat with respect to increases in NADP

performance (Fig. 2A). By contrast, severe losses

of NADP performance are predicted to be benefi-

cial as correlated reductions in the affinities for

NADPHfreeup IMDHforusewith abundantNAD.

As predicted, all beneficial LeuBERKYVYR^

mutants have reduced affinities for both NADP

andNADPH (Table 1). Unaffected by constraints,

increases in expression are unconditionally ben-

eficial (18). These results support the hypothesis

that NADP-specific IMDHs function poorly in

vivo because of strong inhibition by abundant

NADPH. That reductions in NADP performance

and increases in expression are both beneficial are

the predicted consequences of a phenotype-fitness

map constrained by an upper limit to kcatNADP, a

correlation in affinities for NADP and NADPH,

and a trade-off in coenzyme performance.

Breaking any one constraint would allow

NADP-specific IMDHs to evolve. Yet no mutant

increases kcatNADP, nomutant uncouples the affinities

for NADP and NADPH, and nomutant breaks the

trade-off in coenzyme performance. These con-

clusions are not the result of a selective screen that

is too stringent. Of 35 colonies, representing 26

distinct mutants, that appeared after the 48-hour

limit, 17 had no nucleotide substitutions in the 5¶

leader sequence or amino acid replacements in

LeuBERKYVYR^ (table S3). Amino acid replace-

ments in the other mutants neither improved en-

zyme performance nor decreased NADPH

inhibition (table S4). That no additional benefi-

cial mutations were recovered with relaxed cri-

teria demonstrates that the selective screen was

not overly stringent.

Nor are the results a consequence of inadequate

sampling of protein sequence space. Natural

adaptive evolution fixes advantageous mutations

sequentially (22–24). Indeed, experimental evo-

lution demonstrates that advantageous double

mutants in the evolved b-galactosidase of E.

coli are not evolutionarily accessible and per-

force must be accumulated as sequential ad-

vantageous mutations (25). Hence, screening

Table 1. Kinetic effects of amino acid replacements in LeuB[RKYVYR] isolated at 48 hours growth.Standard errors are G13% of estimates. Single-letter abbreviations for amino acid residues: A, Ala;C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R,Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

Enzyme

NADP NAD

Performance

(A)

Performance

(B)

Preference

(A/B)

KiNADPH

(mM)

Km(mM)

kcat(s–1)

kcat/Km(mM–1 s–1)

Km(mM)

kcat(s–1)

kcat/Km(mM–1 s–1)

Site

DDIAGR (wild type) 8400 4.10 0.4900 101 6.90 68.3200 0.007 254.30

RKYVYR 183 6.20 33.8800 4108 0.83 0.2000 169.400 9.90

Beneficial replacements

K235N 3919 5.95 1.5200 6356 0.50 0.0800 19.000 129.00

K235N,–3(M),A60V 1744 5.06 2.9000 5395 0.44 0.0800 35.700 56.00

K235R 554 6.39 11.5000 3292 0.36 0.1100 106.100 26.00

K235R,E63V 658 6.75 10.3000 3748 0.46 0.1200 84.100 28.00

K289E* 3449 2.71 0.7900 4897 1.98 0.4000 2.000 133.50

K289E*,D317E 3554 3.62 1.0200 5580 2.09 0.3700 2.800 112.00

K289E*,D87Y,S182F 2551 2.54 1.0000 5750 3.65 0.6300 1.600 105.20

K289E*,F102L 3891 3.66 0.9400 4823 1.10 0.2300 4.100 131.90

K289M 2216 4.13 1.8600 4034 0.48 0.1200 15.500 78.60

K289M,F170L 1945 2.50 1.2900 2947 0.48 0.1600 8.100 87.20

K289N* 1141 6.32 5.5400 4997 1.14 0.2300 24.300 44.00

K289N*,R187H 1146 5.83 5.0900 3069 0.79 0.2600 19.600 44.70

K289N*,H367Q 1248 5.27 4.2200 4555 1.05 0.2300 18.300 57.80

K289T 1596 6.18 3.8700 4056 0.78 0.1900 20.400 71.60

K289T,Q157H 1696 5.66 3.3400 5038 0.93 0.1800 18.600 72.60

Y290C 988 5.21 5.2700 3792 1.02 0.2700 19.600 49.00

Y290C,R152C 910 3.15 3.4600 3844 0.90 0.2300 14.800 43.00

Y290D 10213 3.13 0.3100 9040 0.82 0.0900 3.400 344.00

Y290F* 1658 5.59 3.3700 3110 0.65 0.2100 16.000 59.70

Y290F,P97S,D314E 1097 4.64 4.2300 4158 0.71 0.1700 24.800 57.40

Y290N,N52T 2381 3.44 1.4400 10660 0.26 0.0200 59.500 90.00

Replacements with beneficial 5¶-leader mutations

RKYVYR 183 6.20 33.8800 4108 0.83 0.2000 169.400 9.90

–3(M)†,E66D 208 2.97 14.2800 5002 0.30 0.0600 238.000 11.90

–3(M)†,E66D,E331D 267 7.99 29.9000 3403 0.34 0.1000 296.900 14.00

E82D 221 5.44 24.6200 4735 0.35 0.0700 351.700 14.10

K100R,A229T,L248M 137 5.58 40.3700 4079 0.45 0.1100 370.300 11.50

G156R,K289R,Y311H 292 5.40 18.4900 4464 0.54 0.1200 144.800 17.40

E173D 219 6.39 29.1800 5304 0.89 0.1700 171.600 6.60

Y337N,G131 171 5.71 33.3900 3367 0.24 0.0700 468.500 8.00

Y337H,P85S,E181K 318 5.21 16.3800 2409 0.22 0.0900 183.600 12.80

*Switch from NADP to NAD requires two base substitutions in one codon. This is a possible transitional amino acid produced bya single base substitution (11). †The –3(M) designates a G to A substitution at base –3 that creates a new start codon.

REPORTS


28

Of the 100,000 mutated plasmids screened, 134 (representing 107 distinct isolates) formed colonies within 48 hours (table S1). Each had either a substitution in the 5′ leader sequence upstream of leuB[RKYVYR], an amino acid replace-ment in the coenzyme binding pocket, or both. Upstream substitutions occurred at three nucleotide positions: –3, –9, and –14 relative to the leuB AUG start codon (fig. S1). Ten amino acid replacements were found at three codons in the coenzyme binding pocket.

At first glance, the positive response to selection might suggest that coenzyme use by IMDH is unconstrained. Upstream sub-stitutions in the Shine-Dalgarno sequence (positions –9 and –14), as well as a new AUG start codon (position –3) that re-places the less efficient GUG start codon, presumably derive their benefits through increases in expression because their ki-netics are unchanged. Unexpectedly, how-ever, beneficial amino acid replacements in the coenzyme binding pocket eliminate H bonds to the 2′-phosphate of NADP to cause striking reductions in NADP perfor-mance (Table 1). Although some mutants have improved NAD performance, others remain unchanged and several show re-duced NAD performance. Isolated amino acid replacements outside the coenzyme binding pocket, which might have been expected to increase kNADP

cat or to break the correlation in affinities between NADP and NADPH (KNADP

m and KNADPHi ), have

no detectable functional effects (table S2, those associated with beneficial 5′-leader mutations in Table 1). No doubt they, along with 126 silent substitutions (table S1), hitchhiked through the genetic screen with the beneficial mutations.

Our results suggest that increases in ex-pression are beneficial, whereas increases in NADP performance are not. This seem-ing paradox is resolved if there are no mutations capable of breaking the upper limit to kNADP

cat the correlation in affinities for NADP and NADPH, or the trade-off in

coenzyme performance. With these con-straints, and with reduced expression in the genetic screen, the phenotype-fitness map near LeuB[RKYVYR] remains flat with respect to increases in NADP perfor-mance (Fig. 2A). By contrast, severe loss-es of NADP performance are predicted to be beneficial as correlated reductions in the affinities for NADPH free up IMDH for use with abundant NAD. As predicted, all beneficial LeuB[RKYVYR] mutants have reduced affinities for both NADP and NADPH (Table 1). Unaffected by con-straints, increases in expression are un-conditionally beneficial (18). These results support the hypothesis that NADP-specific IMDHs function poorly in vivo because of strong inhibition by abundant NADPH. That reductions in NADP performance and increases in expression are both ben-eficial are the predicted consequences of a phenotype-fitness map constrained by an upper limit to kNADP

cat , a correlation in affini-ties for NADP and NADPH, and a trade-off in coenzyme performance.

Breaking any one constraint would al-low NADP-specific IMDHs to evolve. Yet no mutant increases kNADP

cat , no mutant uncouples the affinities for NADP and NADPH, and no mutant breaks the trade-off in coenzyme performance. These con-clusions are not the result of a selective screen that is too stringent. Of 35 colonies, representing 26 distinct mutants, that ap-peared after the 48-hour limit, 17 had no nucleotide substitutions in the 5′ leader sequence or amino acid replacements in LeuB[RKYVYR] (table S3). Amino acid replacements in the other mutants nei-ther improved enzyme performance nor decreased NADPH inhibition (table S4). That no additional beneficial mutations were recovered with relaxed criteria dem-onstrates that the selective screen was not overly stringent.

Nor are the results a consequence of inadequate sampling of protein sequence space. Natural adaptive evolution fixes advantageous mutations sequentially (22–

29

24). Indeed, experimental evolution dem-onstrates that advantageous double mu-tants in the evolved ß-galactosidase of E. coli are not evolutionarily accessible and perforce must be accumulated as sequen-tial advantageous mutations (25). Hence, screening all single substitutions is a sufficient sampling of protein sequence space for robust evolutionary conclusions. We estimate that only 0.04 advantageous amino acid replacements are missing from the mutant leuB[RKYVYR] library (18). We conclude that mutations capable of breaking the limit, the correlation, or the trade-off are unlikely to ever be fixed in populations because they are exceedingly rare (they may not exist), because they are minimally advantageous, or both.

The two remaining ways to evolve an NADP-specific IMDH are to reduce intra-cellular NADPH pool and, as our results show, to increase expression. Reducing in-tracellular NADPH relieves the inhibition but, as experiments deleting sources of NADPH show (13), the disruption to the rest of metabolism costs far more than the

all single substitutions is a sufficient sampling of

protein sequence space for robust evolutionary

conclusions. We estimate that only 0.04 advan-

tageous amino acid replacements are missing

from themutant leuB[RKYVYR] library (18). We

conclude that mutations capable of breaking the

limit, the correlation, or the trade-off are unlikely

to ever be fixed in populations because they are

exceedingly rare (they may not exist), because

they are minimally advantageous, or both.

The two remaining ways to evolve an NADP-

specific IMDH are to reduce intracellular NADPH

pool and, as our results show, to increase expres-

sion. Reducing intracellular NADPH relieves the

inhibition but, as experiments deleting sources of

NADPH show (13), the disruption to the rest of

metabolism costs far more than the benefit to be

gained. The phenotype-fitness map (Fig. 1C)

imposes a law of diminishing returns such that

LeuBERKYVYR^ must be expressed above wild-

type levels by a factor of 100 to overcome the

inhibition by NADPH (18). Diverting resources

away from other metabolic needs toward compen-

satory protein synthesis would impose a protein

burden (26–29) sufficient to prevent the evolu-

tion of NADP-specific IMDHs.

The production of unnatural phenotypes, by

artificial selection or molecular engineering, is not

sufficient to conclude that evolutionary constraints

are absent entirely. Rather, potential constraints

underlying a conserved phenotype can be identified

from the relationships among genotype, phenotype,

and fitness that define an adaptive landscape.

Experimental evolution can then be used to test

their existence. Using this approach, we have

shown how certain structure-function relationships

in IMDHhave constrained its coenzymephenotype

since the last common ancestor.

References and Notes1. S. J. Gould, R. C. Lewontin, Proc. R. Soc. London Ser. B

205, 581 (1979).

2. J. Antonovics, P. H. van Tienderen, Trends Ecol. Evol. 6,

166 (1991).

3. M. R. Rose, G. V. Lauder, Adaptation (Academic Press,

San Diego, CA, 1996).

4. M. Pigliucci, J. Kaplan, Trends Ecol. Evol. 15, 66 (2000).

5. N. Eldredge et al., Paleobiology 31, 133 (2005).

6. M. Travisano, J. A. Mongold, A. F. Bennett, R. E. Lenski,

Science 267, 87 (1995).

7. H. Teotonio, M. R. Rose, Nature 408, 463 (2000).

8. P. Beldade, K. Koops, P. M. Brakefield, Nature 416, 844

(2002).

9. W. A. Frankino, B. J. Zwaan, D. L. Stern, P. M. Brakefield,

Science 307, 718 (2005).

10. R. Chen, A. Greer, A. M. Dean, Proc. Natl. Acad. Sci.

U.S.A. 93, 12171 (1996).

11. M. Lunzer, S. P. Miller, R. Felsheim, A. M. Dean, Science

310, 499 (2005).

12. A. M. Dean, G. B. Golding, Proc. Natl. Acad. Sci. U.S.A.

94, 3104 (1997).

13. G. Zhu, G. B. Golding, A. M. Dean, Science 307, 1279

(2005); published online 13 January 2005 (10.1126/

science.1106974).

14. G. Wallon et al., J. Mol. Biol. 266, 1016 (1997).

15. J. H. Hurley, A. M. Dean, Structure 2, 1007 (1994).

16. J. H. Hurley, A. M. Dean, D. E. Koshland Jr., R. M. Stroud,

Biochemistry 30, 8671 (1991).

17. T. Penfound, J. W. Foster, in Escherichia coli and Salmonella:

Cellular and Molecular Biology, F. C. Neidhardt, Ed. (ASM

Press, Washington, DC, ed. 2, 1996), pp. 721–730.

18. See supporting material on Science Online.

19. O. Kuchner, F. H. Arnold, Trends Biotechnol. 58, 1 (1997).

20. F. H. Arnold, P. L. Wintrobe, K. Miyazaki, A. Gershenson,

Trends Biochem. Sci. 26, 100 (2001).

21. L. G. Otten, W. J. Quax, Biomol. Eng. 22, 1 (2005).

22. S. Wright, in Proceedings of the Sixth International

Genetics Congress, D. F. Jones, Ed. (Brooklyn Botanic

Garden, Menasha, WI, 1932), pp. 356–366.

23. J. Maynard Smith, Nature 225, 563 (1970).

24. D. M. Weinreich, N. F. Delaney, M. A. Depristo, D. L. Hartl,

Science 312, 111 (2005).

25. B. G. Hall, Antimicrob. Agents Chemother. 46, 3035

(2002).

26. S. Zamenhof, H. H. Eichhorn, Nature 216, 456 (1967).

27. K. J. Andrews, G. D. Hegeman, J. Mol. Evol. 8, 317 (1976).

28. D. E. Dykhuizen, Evolution 32, 125 (1978).

29. A. L. Koch, J. Mol. Evol. 19, 455 (1983).

30. Supported by NIH grant GM060611 (A.M.D.). We thank

three anonymous reviewers whose constructive criticisms

we found invaluable.



SOM Text

Figs. S1 to S3

Tables S1 to S5

References

4 August 2006; accepted 20 September 2006

10.1126/science.1133479

An Essential Role for LEDGF/p75in HIV IntegrationManuel Llano, Dyana T. Saenz, Anne Meehan, Phonphimon Wongthida, Mary Peretz,

William H. Walker, Wulin Teo, Eric M. Poeschla*

Chromosomal integration enables human immunodeficiency virus (HIV) to establish a permanentreservoir that can be therapeutically suppressed but not eradicated. Participation of cellularproteins in this obligate replication step is poorly understood. We used intensified RNA interferenceand dominant-negative protein approaches to show that the cellular transcriptional coactivator lensepithelium–derived growth factor (LEDGF)/p75 (p75) is an essential HIV integration cofactor. Themechanism requires both linkages of a molecular tether that p75 forms between integrase andchromatin. Fractionally minute levels of endogenous p75 are sufficient to enable integration,showing that cellular factors that engage HIV after entry may elude identification in less intensiveknockdowns. Perturbing the p75-integrase interaction may have therapeutic potential.

Integration enables human immunodeficiency

virus type 1 (HIV-1) to establish a perma-

nent genetic reservoir that can initiate new

virion production, evade immune surveillance,

and replicate through mitosis. Integrated pro-

viruses that persist in long-lived T cells ensure

rapid HIV recrudescence if antiviral drugs are

withdrawn. Integration is catalyzed by the viral

integrase (IN). When expressed as a free protein in

cells rather than within its normal context as an

intravirion cleavage product of the HIV Gag-Pol

precursor, IN becomes tethered to chromatin by

cellular lens epithelium–derived growth factor/p75

(p75) (1–3), which is a transcriptional coactivator

(4). Accordingly, both proteins display tight colo-

calization with chromatin throughout the cell

cycle; short hairpin RNA (shRNA)–mediated

knockdown of p75 untethers IN, redistributing it

from an entirely nuclear to an entirely cyto-

plasmic location (3). Molecular tethering results

from specific linkages formed by p75_s discrete

functional modules: the N-terminal Pro-Trp-Trp-

Pro (PWWP) and A/T-hook elements bind to

chromatin (5), and a C-terminal integrase-binding

domain (IBD) binds to IN (6, 7). p75 also protects

the HIV-1 IN protein from rapid degradation in

the 26S proteasome (8). In the bona fide viral

context, drastic knockdown of p75 changed the

genomic pattern of HIV-1 integration by reducing

the viral bias for active genes, which suggests that

p75 influences integration targeting (9). However,

changes in overall levels of HIV integration and

replication have been either absent or modest, and

single-cycle infection analyses in cell lines have

consistently detected no effect, which has led to

questions about the overall importance of p75 in

the viral life cycle (3, 7, 9–12).

Previously, we observed that a nuclear lo-

calization signal–mutant p75 protein became con-

stitutively chromatin-trapped in stable cell lines

(7). In the present work, we hypothesized the

existence in previous severely RNA interfer-

ence (RNAi)–depleted HIV-susceptible cells of

a very small yet virologically potent chromatin-

associated p75 residuum. We reasoned that a

fractionally minute residual pool with a spatially

favorable location (colocalized with chromatin)

could explain the inability to demonstrate substan-

tial, reproducible impairments in integration or

viral replication in cells lacking detectable p75.

Such a reservoir would be inadequate to affect

observable properties of ectopically expressed IN

but might be sufficient to engage the vastly less

abundant incoming viral preintegration complex.

To test this hypothesis, we performed sub-

cellular fractionation and interrogated chromatin,

using a deoxyribonuclease (DNase) I– and salt-

based extraction protocol (13). These methods

Molecular Medicine Program, Mayo Clinic College ofMedicine, Rochester, MN 55905, USA.


REPORTS









































205, 581 (1979).


166 (1991).






Science 267, 87 (1995).



(2002).


Science 307, 718 (2005).


U.S.A. 93, 12171 (1996).


310, 499 (2005).


94, 3104 (1997).



science.1106974).


















Science 312, 111 (2005).


(2002).




29. A. L. Koch, J. Mol. Evol. 19, 455 (1983).






SOM Text

Figs. S1 to S3

Tables S1 to S5

References


10.1126/science.1133479
































































REPORTS


benefit to be gained. The phenotype-fitness map (Fig. 1C) imposes a law of diminish-ing returns such that LeuB[RKYVYR] must be expressed above wild-type levels by a factor of 100 to overcome the inhibi-tion by NADPH (18). Diverting resources away from other metabolic needs toward compensatory protein synthesis would im-pose a protein burden (26–29) sufficient to prevent the evolution of NADP-specific IMDHs.

The production of unnatural pheno-types, by artificial selection or molecular engineering, is not sufficient to conclude that evolutionary constraints are absent entirely. Rather, potential constraints un-derlying a conserved phenotype can be identified from the relationships among genotype, phenotype, and fitness that de-fine an adaptive landscape. Experimental evolution can then be used to test their existence. Using this approach, we have shown how certain structure-function re-lationships in IMDH have constrained its coenzyme phenotype since the last com-mon ancestor.








































205, 581 (1979).


166 (1991).






Science 267, 87 (1995).



(2002).


Science 307, 718 (2005).


U.S.A. 93, 12171 (1996).


310, 499 (2005).


94, 3104 (1997).



science.1106974).


















Science 312, 111 (2005).


(2002).




29. A. L. Koch, J. Mol. Evol. 19, 455 (1983).






SOM Text

Figs. S1 to S3

Tables S1 to S5

References


10.1126/science.1133479
































































REPORTS


30

Tech Notes

plasmid template at efficiencies greater than 80% (Figure 2). The QuikChange II and QuikChange II XL kits feature our highest fidelity DNA polymerase, our gold standard for accuracy, and a linear amplifi-cation strategy. Together, they reduce the mutation frequency due to incorporation errors typically caused by using an expo-nential PCR amplification approach. The result is high-efficiency mutagenesis with-out unwanted errors.

Ideal for Large Insertions and Deletions We have used the QuikChange kit method for small insertions and deletions (~12 bp) and observed greater than 80% efficiency2. Many of our customers have used the QuikChange method or have modified it to introduce deletions of 31 bp3, 87 bp4, and 3 kb5 or insertions of 31 bp3 and up to ~1 kb5. In order to achieve large inser-tions of up to 1 kb, megaprimers must be generated from an initial PCR reaction. To ensure the highest mutagenesis efficiency, we recommend gel purification with our StrataPrep® PCR Purification Kit. To create the megaprimers, a minimum of 20 bp upstream and downstream of the insertion sequence should be complementary to your template vector that will be used in the QuikChange kit reaction. Therefore, each oligo used in the initial PCR reac-tion will require 20 bp overhangs on the 5’ ends.

Using the QuikChange kit for large dele-tions is even easier since oligos can be synthesized, rendering the initial PCR reac-tion and fragment purification unneces-sary. Again, 20 to 30 bp may be required to bind to your template DNA both up-stream and downstream of the region that you want to delete. However, the largest oligo required should not exceed 60 bp.

This Technical Note describes performing

Abstract This Technical Note describes a modified method to introduce large insertions in plasmids in two simple steps using the QuikChange® II Site-Directed Mutagen-esis Kit. Dr. Wenge Wang in Dr. Wafik S. El-Deiry’s lab at the University of Pennsyl-vania has used the QuikChange kit and a modification of the method described by Geiser et al. (2001)1 to perform a large insertion of a PCR product into a target plasmid. He had inserted the enhanced green fluorescent protein (EGFP) open reading frame (760 bp) following the C-terminus of death receptor TRAIL recep-tor 1/death receptor 4 (TRAIL-R1/DR4) right after the signal sequence, using the QuikChange II site-directed mutagenesis kit with some modifications to the basic protocol. Dr. Wang transfected cells with the recombined plasmid and it generated a fusion protein with the correct molecular size (Figure 1).

Background Dr. Wafik S. El-Deiry’s lab is focused on studying the mechanism of action of the tumor suppressor p53 and the contribu-tion of its downstream target genes to control cell growth. Their lab has identi-fied a number of genes that are directly regulated by p53 and which can inhibit cell cycle progression (p21WAF1), induce apoptosis (KILLER/DR5, Bid, caspase 6, Traf4 and others) or activate DNA repair (DDB2). The research has provided knowl-edge into the tissue specificity of the DNA damage response in vivo and into the mechanism by which wild-type p53 sensi-tizes cells to killing by anti-cancer drugs.

High-Efficiency Mutagenesis Method All QuikChange® kitsa offer a one-day method to introduce point mutations, ami-no acid substitutions, small insertions, and deletions in virtually any double-stranded

Large Insertions: Two Simple Steps Using QuikChange® II Site-Directed Mutagenesis Kits

31

Tech Notes

Fig. 2. The QuikChange® II One-Day Site-Directed Mutagenesis Method. 1. Mutant strand synthesis. 2. Dpn I Digestion of parental DNA template. 3. Trans-formation of the resulting annealed double-stranded nicked DNA molecules. After transformation, the XL-1 Blue E. coli cell repairs nicks in the plasmid.

Fig. 1 Enhanced green fluorescent protein (EGFP) was inserted in the open reading frame (760 bp) fol-lowing the C-terminus of death receptor TRAIL recep-tor 1/death receptor 4 (TRAIL-R1/DR4) right after the signal sequence. Cell staining shows the member protein fused with EGFP.

1. Mutant Strand Synthesis Perform thermal cycling to: • Denature DNA template • Anneal mutagenic primers containing desired mutation • Extend and incorporate primers with high-fidelity DNA polymerase

2. Dpn I Digestion of Template Digest parental methylated and hemimethylated DNA with Dpn I

3. Transformation Transform mutated molecule into competant cells for nick repair

a large insertion that can be completed in two simple steps. First, PCR-amplify the megaprimer. Second, add your gel purified megaprimer to our QuikChange site-directed mutagenesis kit (Figure 2). This strategy avoids tedious and time-con-suming sub-cloning. With our QuikChange kits, you will have confidence in generat-ing an error free clone while saving valu-able time.

Materials & Methods

Part I. PCR reaction to generate the megaprimer

Reaction and cycling conditions.PCR reagents: PfuUltra® II Fusion HS DNA Polymerase (Stratagene)

Thermal Cycler used: TECHNE: Touch-gene® Gradient

• The forward primer: GACTCCGAATCCCGGGAGCGCAGCGGTGAGCAAGGGCGAGGAG, the first 25 bp overlaps the target vector, the remaining primer sequence is complementary to EGFP. The primers were resuspended in 50 mM NaCl.

• The reverse primer: CGCGGCTGCCTCTGTCCCACTCTTGTACAGCTCGTCCATGCC, the first 22 bp targets the vector and the remaining primer sequence is complementary to EGFP. The primers were resuspended in 50 mM NaCl.

• Template for PCR reaction: 200 ng of plasmid DNA (pEGFP-N1; Clontech)

• PfuUltra® II Fusion HS DNA Polymer-ase (Stratagene)

• PCR product was purified with QIAGEN’S QIAEX II Gel Extraction Kit

• The quantity of the PCR product was estimated on agarose gel with Ethidium Bromide

Note: The size of the PCR product is 760 bp, which includes the insert (EGFP) of 714 bp plus overlapping sequence of the target vector.

32

Tech Notes

Part II. The QuikChange II mutagenesis reaction can be set up using the follow-ing guidelines (Tables 1 and 2).

• Primer for the inser-tion reaction 300-500 ng of PCR product (760 bp EGFP) to serve as a megaprimer.

TABLE 2. Cycling Method

Segment Cycles Temperature Time

1 1 95°C 30 seconds


52°C 1 minute

68°C 1 min/kb of plasmid* (7 min)


55°C 1 minute

68°C 1 min/kb of plasmid* (7 min)

* For example, a 7 kb plasmid would have a 7 minute extension time

References1. Geiser, M., et al. (2001) BioTechniques 31: 88-92.2. Papworth, C., Braman, J., and Wright, D.A. (1996) Strategies 9: 3-4.3. Wang, W. and Malcolm, B. (1999) BioTech-niques 26: 680-682.4. Burke, T., et al. (1998) Oncogene 16: 1031-1040.5. Makarova, O., et al. (2000) BioTechniques 29: 970-972.

LegalQuikChange® and StrataPrep® are registered

trademarks of Stratagene, an Agilent Technolo-gies company.QIAGEN® is a registered trademark of QIAGEN. Touchgene® is the registered trademark of Tech-ne Incorporated.Clontech© logo is the trademark of Clontech Laboratories, Inc.a. U.S. Patent Nos. 7,176,004; 7,132,265; 7,045,328; 6,734,293; 6,489,150; 6,444,428; 6,391,548; 6,183,997; 5,948,663; 5,932,419; 5,866,395; 5,789,166; 5,545,552, and patents pending.

Note: Follow reaction with Dpn I digestion and transformation per the QuikChange® II manual. For PfuUltra® II Fusion HS DNA Polymerase infor-mation please refer to the manual.

Results The results of this mu-tagenesis experiment demonstrate that large PCR fragments can be suc-cessfully inserted into the target plasmid DNA. The 760 bp PCR fragment had at each end only short regions of homology where the insertion occurred. There were approximately 100 colonies produced. About 20 colonies were selected for further analysis of mutagenesis efficiency, and they were 100% positive by sequencing.

Amount Component per reactionDistilled water (dH2O) X µl10x QuikChange reaction buffer 5.0 µldNTP mix 1 µlDNA template (50 ng) X µlMegaprimer (300 - 500 ng) X µlHigh Fidelity DNA polymerase (2.5 U/µl) 1.0 µl (2.5U) Total reaction volume 50 µl

TABLE 1. Reaction Set-up

Acknowledgments This Technical Note was written based on data provided by Wenge Wang, M.D., Ph.D., in Dr. Wafik S. El-Deiry’s laboratory, Division of Hematology and Oncology in the Department of Medicine at the University of Pennsylvania.

Tech Notes

s

u.s. and canada 1-800-424-5444 x3europe 00800-7000-7000www.stratagene.com

10%Value 20% 30% 40% 50% 60% 70% 80%

Mutazyme® IIDNA Polymerase

Mutazyme® IDNA Polymerase

TaqDNA Polymerase

AT to N GC to N

Our GeneMorph®

II Kits include Mutazyme®

IIPolymerase, which delivers a balanced mutationalspectrum.

U.S. Patent No. 7,045,328; 6,803,216; 6,734,293; 6,489,150; 6,444,428; 6,183,997; 5,489,523 and patents pending

GeneMorph® and Mutazyme® are registered trademarks of Stratagene, an Agilent Technologies Company,

in the United States.

With GeneMorph® II mutagenesis kits, a balanced spectrum of mutations is right at your fingertips

C

M

Y

CM

MY

CY

CMY

K

GeneeMorph_Ad_Q407_Science InserPage 1 8/21/2007 3:26:25 PM

QuikChange® is a registered trademark of Stratagene, an Agilent Technologies Company, in the United States.

*U.S. Patent Nos 6,734,293; 6,489,150; 6,444,428; 6,391,548; 6,183,997; 5,948,663;

5,932,419; 5,866,395; 5,789,166; 5,545,552; 6,713,285 and patents pending.

2X Faster

u.s. and canada 1-800-424-5444 x3europe 00800-7000-7000

www.stratagene.com

C

M

Y

CM

MY

CY

CMY

K

QuikChange Lightning_Ad_Q407_SciPage 1 8/21/2007 3:35:29 PM

mutagenesis techniques (pdf)

Documents