DOI: 10.1126/science.1229017, 85 (2013);339 Science et al.Lili K. Doerfel

Consecutive Proline ResiduesEF-P Is Essential for Rapid Synthesis of Proteins Containing

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): January 4, 2013 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/339/6115/85.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2012/12/12/science.1229017.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/339/6115/85.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/339/6115/85.full.html#ref-list-1, 13 of which can be accessed free:cites 33 articlesThis article

http://www.sciencemag.org/content/339/6115/85.full.html#related-urls1 articles hosted by HighWire Press; see:cited by This article has been

registered trademark of AAAS. is aScience2013 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Janu

ary

4, 2

013

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

References and Notes1. M. Bailly, V. de Crécy-Lagard, Biol. Direct 5, 3 (2010).2. N. C. Kyrpides, C. R. Woese, Proc. Natl. Acad. Sci. U.S.A.

95, 224 (1998).3. B. R. Glick, M. C. Ganoza, Proc. Natl. Acad. Sci. U.S.A.

72, 4257 (1975).4. B. R. Glick, S. Chládek, M. C. Ganoza, Eur. J. Biochem.

97, 23 (1979).5. G. Blaha, R. E. Stanley, T. A. Steitz, Science 325, 966

(2009).6. T. Yanagisawa, T. Sumida, R. Ishii, C. Takemoto,

S. Yokoyama, Nat. Struct. Mol. Biol. 17, 1136 (2010).7. W. W. Navarre et al., Mol. Cell 39, 209 (2010).8. H. Roy et al., Nat. Chem. Biol. 7, 667 (2011).9. J. H. Park et al., J. Biol. Chem. 287, 2579 (2012).

10. L. Peil et al., Nat. Chem. Biol. 8, 695 (2012).11. E. de Crécy et al., Appl. Microbiol. Biotechnol. 77,

489 (2007).12. S. B. Zou et al., J. Bacteriol. 194, 413 (2012).13. K. Kaniga, M. S. Compton, R. Curtiss 3rd, P. Sundaram,

Infect. Immun. 66, 5599 (1998).

14. S. M. Bearson, B. L. Bearson, M. A. Rasmussen,Appl. Environ. Microbiol. 72, 2829 (2006).

15. W. T. Peng, L. M. Banta, T. C. Charles, E. W. Nester,J. Bacteriol. 183, 36 (2001).

16. S. M. Bearson, B. L. Bearson, B. W. Brunelle,V. K. Sharma, I. S. Lee, Foodborne Pathog. Dis.8, 725 (2011).

17. S. Y. Meng, G. N. Bennett, J. Bacteriol. 174, 2670(1992).

18. G. Fritz et al., J. Mol. Biol. 393, 272 (2009).19. Material and methods are available as supplementary

materials on Science Online.20. H. Muto, K. Ito, Biochem. Biophys. Res. Commun. 366,

1043 (2008).21. M. Y. Pavlov et al., Proc. Natl. Acad. Sci. U.S.A. 106, 50

(2009).22. I. Wohlgemuth, S. Brenner, M. Beringer, M. V. Rodnina,

J. Biol. Chem. 283, 32229 (2008).23. Y. Shimizu et al., Nat. Biotechnol. 19, 751 (2001).24. M. N. Neely, E. R. Olson, J. Bacteriol. 178, 5522

(1996).

25. J. Garmendia, C. R. Beuzón, J. Ruiz-Albert, D. W. Holden,Microbiology 149, 2385 (2003).

26. R. L. Lucas et al., J. Bacteriol. 182, 1872 (2000).

Acknowledgments: This work was supported by theExcellence Cluster “Center for integrated Protein Science”Munich (Exc114/1), the European Molecular BiologyOrganization young investigator program (to D.N.W.) andgrants JU270/5-3 (K.J.) and WI3285/1-1 (D.N.W.) from theDeutsche Forschungsgemeinschaft. A.L.S. is funded by anAXA Postdoctoral fellowship.

Supplementary Materialswww.sciencemag.org/cgi/content/full/science.1228985/DC1Materials and MethodsFigs. S1 to S2Tables S1 to S5References (27–55)

17 August 2012; accepted 29 October 2012Published online 13 December 2012;10.1126/science.1228985

EF-P Is Essential for Rapid Synthesisof Proteins Containing ConsecutiveProline ResiduesLili K. Doerfel,1* Ingo Wohlgemuth,1,2* Christina Kothe,1 Frank Peske,1

Henning Urlaub,2,3 Marina V. Rodnina1†

Elongation factor P (EF-P) is a translation factor of unknown function that has been implicatedin a great variety of cellular processes. Here, we show that EF-P prevents ribosome from stallingduring synthesis of proteins containing consecutive prolines, such as PPG, PPP, or longer prolinestrings, in natural and engineered model proteins. EF-P promotes peptide-bond formation andstabilizes the peptidyl–transfer RNA in the catalytic center of the ribosome. EF-P is posttranslationallymodified by a hydroxylated b-lysine attached to a lysine residue. The modification enhances thecatalytic proficiency of the factor mainly by increasing its affinity to the ribosome. We proposethat EF-P and its eukaryotic homolog, eIF5A, are essential for the synthesis of a subset of proteinscontaining proline stretches in all cells.

Bacterial elongation factor P (EF-P) and itsarchaeal/eukaryotic homolog, initiationfactor 5A (a/eIF5A), are universally con-

served proteins with unknown cellular function.EF-P was reported to modulate cell viability,growth, virulence, motility, and sensitivity to lowosmolarity, detergents, and antibiotics (1, 2).The EF-P binding site on the ribosome is locatedat the interface of the 30S and 50S ribosomalsubunits between the binding sites for peptidyl-tRNA (P site) and the exiting tRNA (E site). TheN-terminal domain of EF-P interacts with the ac-ceptor stem of the P site–bound tRNA near thepeptidyl transferase center (3). Both EF-P and

eIF5A are posttranslationally modified (4–8).In Escherichia coli EF-P, Lys34 is modified bythe action of three enzymes, YjeK, YjeA, andYcfM (5, 7–9). The proposed cellular functionof EF-P/eIF5A is to optimize the ribosomes formore productive interactions with tRNA and re-lease factors (3, 10–13). However, the magnitudeof the reported effects appeared too small (lessthan twofold) for a universal function. We clarifythe functional role of EF-P by investigating therole of the factor in the translation of variousmRNA sequences.

The activity of EF-P in accelerating peptidebond formation was originally identified by itsability to increase the yield of formylmethionyl-puromycin (fMet-Pmn) synthesis (12, 13). Becausereaction rates vary by ~1000-fold for differentC-terminal amino acids in the peptidyl-tRNA(14), we examined whether EF-P may specifical-ly accelerate the product formation with poorlyreactive peptidyl-tRNAs, such as those containingC-terminal proline residues (Fig. 1A and fig. S1).The formation of the tripeptide fMetPro-Pmnwasaccelerated by almost 90-fold in the presence of

EF-P, whereas in the other cases the reaction wasstimulated by less than fivefold. To explore theeffect of the amino acid following a proline resi-due, we testedGly- and Pro-tRNAs that are knownto be slow in peptide bond formation and cancontribute to ribosome stalling (15–17). WhereasEF-P enhanced peptide-bond formation by lessthan twofold for most combinations, large effectswere observed for the formation of fMetProGly(fMPG, eightfold) and fMetProPro (16-fold) (Fig.1B).When longer strings of Pro and Gly residues,such as fMPPG, fMPPGF, or fMPPPF, were ex-amined, practically no correct product was formedin the absence of EF-P, whereas the addition ofEF-P rescued the rapid production of the respec-tive peptide (Fig. 1C). The low yield of productformation in the absence of EF-P can be explainedby the loss of peptidyl-tRNA from the stalled ribo-somes (e.g., fMPP-tRNA; fig. S2A). These resultssuggested a specific effect of EF-P in acceleratingthe reaction with poorly reactive substrates thatotherwise cause ribosome stalling.

To verify whether proline-containing sequencesas such induce ribosome stalling that can be res-cued by EF-P, we engineered PG, PP, PPG, andPPP sequences into an N-terminal fragment ofprotein PrmC, which originally does not containsuch sequences and is rapidly synthesized inde-pendent of the presence of EF-P (Fig. 2). Intro-ducing a glycine residue after Pro20 resulted in theribosome pausing, as shown by the accumulationof a peptide of about 20 amino acids; however,pausing was transient and not affected by EF-P.When a second proline residue preceding Pro20

was engineered, the ribosome paused at the ProProsequence, and EF-P reduced the pause time from20 to 10 s. When PPG or PPP sequences wereintroduced, the ribosomes were stalled, and es-sentially no full-length product was produced. Theribosomes stalled upon translation of the PPG se-quence that contained a ProPro-ending peptidyl-tRNA in the P site and a Gly-tRNA in the A site(fig. S2B). Addition of EF-P rescued the synthesisof full-length product by preventing or alleviatingribosome stalling (Fig. 2 and fig. S2C).

1Department of Physical Biochemistry, Max Planck Institutefor Biophysical Chemistry, 37077 Goettingen, Germany. 2Bio-analytical Mass Spectrometry Group, Max Planck Institute forBiophysical Chemistry, 37077 Goettingen, Germany. 3Bio-analytics, Department of Clinical Chemistry, University MedicalCenter Goettingen, 37075 Goettingen, Germany.

*These authors contributed equally to this work.†To whom correspondence should be addressed. E-mail:rodnina@mpibpc.mpg.de

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 85

REPORTS

on

Janu

ary

4, 2

013

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

Fig. 1. EF-P promotes the synthesis of Pro- andGly-containing peptides on the ribosome. (A) Rates(kobs) of peptide-bond formation between fMet-tRNAfMet or fMetX-tRNAX and Pmn, where X standsfor different amino acids, as indicated, in the ab-sence (white) or presence (black) of EF-P (3 mM). Errorbars indicate standard deviations. Single-letter ab-breviations for the amino acid residues are as fol-lows: D, Asp; E, Glu; F, Phe; G, Gly; K, Lys; M, Met;P, Pro; Q, Gln; R, Arg; V, Val; and W, Trp. (B) Ratesof peptide-bond formation between fMet-tRNA(fM) or fMetPro-tRNA (fMP) in the P site and Gly-tRNAGly, Phe-tRNAPhe, or Pro-tRNAPro (G, F, and P,respectively) in the A site. (C) Formation of modeloligopeptides in a reconstituted translation systemin the absence (open) and presence (solid) of EF-P.

Fig. 2. EF-P prevents ribosome stalling on PPG andPPP sequences engineered into PrmC. Translationof the N-terminal domain of PrmC (75 amino acids)with wild-type (WT) or mutant sequences containingPG, PP, PPG, or PPP in the absence or presence ofEF-P. Peptides were separated by SDS–polyacrylamidegel electrophoresis (SDS-PAGE) and visualized bythe fluorescence of BODIPY-FL attached to the Nterminus of the peptides. M1 and M2 are peptidemarkers for PrmC fragments of the indicated num-ber of amino acids. Single-letter abbreviations forthe amino acid residues are as follows: 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; and Y, Tyr.

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org86

REPORTS

on

Janu

ary

4, 2

013

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

The sequences PPG and PPP, as well as lon-ger proline stretches, are found in a large numberof cellular proteins, raising the question of wheth-er they require EF-P for their synthesis.We testedthe effect of EF-P on the synthesis of some ofthese proteins, such as TonB, YafD, Rz1, andAmiB (Fig. 3). In the absence of EF-P, translationwas stalled at the consecutive prolines, and theaddition of EF-P alleviated ribosome stalling, re-sulting in the rapid synthesis of the respectivefull-length peptide. When the Pro stretches wereparticularly long, as inAmiB or Rz1, very little full-length product was synthesized in the absence ofEF-P, because translation was halted at the stallingsite. In all cases, the addition of EF-P resulted in theefficient synthesis of full-length protein. Transientpauses that occurred at sequences other than PPPand PPG were not alleviated by EF-P (Fig. 3).

Posttranslational modification of EF-P ap-pears crucial for the factor’s function (9, 18).Translation of the fMPPG sequence, which wasabolished in the absence of EF-P, was fully re-stored with EF-P carrying the b-lysine modifica-tion at Lys34; further hydroxylation did not alterthe activity, whereas the absence of modificationsreduced the activity of the factor (Fig. 4). A sim-ilar effect was observed for the translation ofthe PPG sequence in PrmC (fig. S4). Analysis ofthe reaction kinetics indicated that the modifica-tion increased the affinity of EF-P binding to theribosome 30-fold (Fig. 4). The maximum rate offMPPG synthesis decreased ~fourfold with un-modified compared with modified EF-P, resultingin a >100-fold difference in kcat/KM (where kcatis the maximum reaction rate obtained by hyper-bolic fitting and KM is the Michaelis constant).The low catalytic proficiency of unmodified EF-Pmay explain why deletions of the yjeA or yjeKgenes that code for the modification enzymeslead to phenotypes that are similar to, or onlysomewhat milder than, the deletion of efp, thegene coding for EF-P (1, 2). On the other hand,even in the absence of active EF-P, small amountsof proteins with proline stretches are slowlyformed (Fig. 3); this may be sufficient to supportthe viability of strains in which the efp, yieA, andyieK genes are deleted.

Our data indicate that EF-P is a translationfactor that promotes the synthesis of proteinscontaining PPG, PPP, and longer polyprolinestretches by preventing ribosome stalling during

Fig. 3. EF-P alleviates PPP- and PPG-induced stalling during synthesis of natural proteins. Translationproducts of TonB (239 amino acids), YafD (75 amino acids from the N terminus), Rz1 (62 amino acids),and AmiB (159 amino acids from the N terminus) were separated by SDS-PAGE. M3 and M4 are peptidemarkers containing TonB fragments of the indicated lengths. M5, M6, and M7 are peptide markers of theindicated lengths of YafD, RZ1, and AmiB sequences, respectively.

Fig. 4. Contribution of Lys34 modi-fications to EF-P activity. (A) Timecourses of fMPPG synthesis. Theformation of the tetrapeptide wasmeasured in the absence of EF-P(open circles), in the presence ofunmodified EF-P (open squares),overexpressed lysinylated but un-hydroxylated EF-P (solid triangles),and lysinylated/hydroxylated overexpressed EF-P (solid circles) or native EF-P (opentriangles). The extent of the respective modification was verified by mass spec-trometry (fig. S3). (B) Yield of fMPPG peptide as a function of EF-P concentrationwith unmodified (open circles; KM = 2.4 T 0.5 mM) or fully modified (solid circles;

KM = 0.08 T 0.02 mM) EF-P. KM is the EF-P concentration at which 50% of themaximum yield is reached. (C) EF-P concentration dependence of fMPPG synthesisrate with unmodified EF-P (open circles; kcat = 0.12 T 0.03 s−1) and fully modifiedEF-P (solid circles; kcat = 0.65 T 0.02 s−1). Error bars indicate standard deviations.

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 87

REPORTS

on

Janu

ary

4, 2

013

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

the formation of proline-proline or proline-glycinepeptide bonds. Thus, EF-P (and likely eIF5A) canbe considered a third universal translation elon-gation factor that acts on the ribosome, in ad-dition to the two factors, EF-Tu and EF-G, thatcatalyze the mRNA decoding and tRNA trans-location steps, respectively. In contrast to thelatter two factors, which act in each elongationcycle irrespective of the amino acid incorporated,EF-P and eIF5A are responsible for the elonga-tion of a subset of nascent proteins and are theonly factors identified so far that augment thepeptidyl transferase activity of the ribosome.

The pyrrolidine ring of proline restricts thenumber of accessible conformations for this res-idue and may impose structural constraints onthe positioning of the amino acid in the peptidyltransferase center, resulting in slow peptide-bondformation and thereby ribosome stalling. EF-Pfacilitates peptide-bond formation by stabilizingpeptidyl-tRNA on the ribosome and possibly bypromoting optimal positioning of the substrates.The cellular fate of the nascent peptide on stalledribosomes is likely to depend on its length: Shorterpeptides may be lost because peptidyl-tRNA dropoff the ribosome (16,19),whereas ribosomes stalledafter the synthesis of longer peptides might berescued through SsrA-, ArfA-, or YaeJ-dependentpathways (20–23).

Of more than 4000 annotated E. coli pro-teins, about 270 contain motifs of three or moreconsecutive prolines or PPG motifs (table S1).Among those, proteins belonging to the basaltranscription-translation machinery are under-represented, whereas metabolic enzymes, trans-porters, and regulatory transcription factors arefrequent. The observed pleiotropic effects of efp,yjeA, or yjeK deletions (1) can be rationalized bythe impaired synthesis of proteins with prolinestretches. For example, decreased synthesis of

TonB, a protein from the EF-P interactome (24)that supplies energy for the function of TonB-dependent transporters, may result in a cumula-tive inhibition of transport processes (25). Thesevere effects of efp, yjeA, or yjeK deletions oncell motility may be explained by the failure tosupport the coordinated expression of flagellarproteins FlhC, FliF, and Flk, all containing PPPand PPG motifs, whereas the deficiency of EF-Pmutants in using g-glutamyl-glycine as a nitrogensource (1) may be caused by impaired synthesisof the corresponding catabolic enzyme, g-glutamyl-transpeptidase, which contains a PPPmotif. Last-ly, EF-P may affect the translation of virulenceproteins, such as protein EspF, which is a keyplayer during the infection of eukaryotic hosts byenterohaemorrhagic and enteropathogenic strainsof E. coli (26, 27). EspF contains several prolineruns, including PPPP sequences, in its function-ally important Src homology 3 binding domains(27). Given that the posttranslational modifica-tions of eukaryotic eIF5A and bacterial EF-P aredifferent, the occurrence of proline runs in manypathogenic proteins makes EF-P and its modifi-cation enzymes promising new targets for de-veloping highly specific, potent antimicrobials.

References and Notes1. S. B. Zou et al., J. Bacteriol. 194, 413 (2012).2. S. B. Zou, H. Roy, M. Ibba, W. W. Navarre, Virulence 2,

147 (2011).3. G. Blaha, R. E. Stanley, T. A. Steitz, Science 325, 966 (2009).4. E. C. Wolff, K. R. Kang, Y. S. Kim, M. H. Park,

Amino Acids 33, 341 (2007).5. T. Yanagisawa, T. Sumida, R. Ishii, C. Takemoto,

S. Yokoyama, Nat. Struct. Mol. Biol. 17, 1136 (2010).6. M. Bailly, V. de Crécy-Lagard, Biol. Direct 5, 3 (2010).7. H. Roy et al., Nat. Chem. Biol. 7, 667 (2011).8. L. Peil et al., Nat. Chem. Biol. 8, 695 (2012).9. W. W. Navarre et al., Mol. Cell 39, 209 (2010).10. B. R. Glick, S. Chládek, M. C. Ganoza, Eur. J. Biochem.

97, 23 (1979).

11. B. R. Glick, M. C. Ganoza, Proc. Natl. Acad. Sci. U.S.A.72, 4257 (1975).

12. P. Saini, D. E. Eyler, R. Green, T. E. Dever, Nature 459,118 (2009).

13. A. P. Gregio, V. P. Cano, J. S. Avaca, S. R. Valentini,C. F. Zanelli, Biochem. Biophys. Res. Commun. 380,785 (2009).

14. I. Wohlgemuth, S. Brenner, M. Beringer, M. V. Rodnina,J. Biol. Chem. 283, 32229 (2008).

15. M. Johansson et al., Proc. Natl. Acad. Sci. U.S.A. 108,79 (2011).

16. M. Y. Pavlov et al., Proc. Natl. Acad. Sci. U.S.A. 106,50 (2009).

17. D. R. Tanner, D. A. Cariello, C. J. Woolstenhulme,M. A. Broadbent, A. R. Buskirk, J. Biol. Chem. 284,34809 (2009).

18. J. H. Park et al., J. Biol. Chem. 287, 2579 (2012).19. V. Heurgué-Hamard et al., EMBO J. 17, 808 (1998).20. Y. Chadani, K. Ono, K. Kutsukake, T. Abo, Mol. Microbiol.

80, 772 (2011).21. Y. Chadani et al., Mol. Microbiol. 78, 796 (2010).22. C. S. Hayes, B. Bose, R. T. Sauer, J. Biol. Chem. 277,

33825 (2002).23. T. Sunohara, T. Abo, T. Inada, H. Aiba, RNA 8, 1416 (2002).24. L. Salwinski et al., Nucleic Acids Res. 32, D449 (2004).25. N. Noinaj, M. Guillier, T. J. Barnard, S. K. Buchanan,

Annu. Rev. Microbiol. 64, 43 (2010).26. N. A. Sallee et al., Nature 454, 1005 (2008).27. A. Holmes, S. Mühlen, A. J. Roe, P. Dean, Infect. Immun.

78, 4445 (2010).

Acknowledgments: We thank W. Wintermeyer for criticallyreading the manuscript; W. Holtkamp and J. Mittelstaet for thetranslation assay and M1 and M2 peptide markers; D. Görlichand C. Enke for the antibody against EF-P; and O. Geintzer,S. Kappler, T. Wiles, M. Zimmermann, T. Uhlendorf, A. Bursy,and U. Plessmann for expert technical assistance. The work wasfunded by the Max Planck Society and by grants from theDeutsche Forschungsgemeinschaft.

Supplementary Materialswww.sciencemag.org/cgi/content/full/science.1229017/DC1Materials and MethodsFigs. S1 to S4Table S1References (28–33)

20 August 2012; accepted 26 October 2012Published online 13 December 2012;10.1126/science.1229017

Para-Aminosalicylic Acid Acts as anAlternative Substrate of Folate Metabolismin Mycobacterium tuberculosisSumit Chakraborty,1,2 Todd Gruber,3 Clifton E. Barry III,3 Helena I. Boshoff,3 Kyu Y. Rhee1,2*

Folate biosynthesis is an established anti-infective target, and the antifolate para-aminosalicylicacid (PAS) was one of the first anti-infectives introduced into clinical practice on the basis oftarget-based drug discovery. Fifty years later, PAS continues to be used to treat tuberculosis. PASis assumed to inhibit dihydropteroate synthase (DHPS) in Mycobacterium tuberculosis by mimickingthe substrate p-aminobenzoate (PABA). However, we found that sulfonamide inhibitors of DHPSinhibited growth of M. tuberculosis only weakly because of their intracellular metabolism. Incontrast, PAS served as a replacement substrate for DHPS. Products of PAS metabolism at thisand subsequent steps in folate metabolism inhibited those enzymes, competing with theirsubstrates. PAS is thus a prodrug that blocks growth of M. tuberculosis when its active formsare generated by enzymes in the pathway they poison.

Domagk’s discovery of the sulfonamideProntosil, a synthetic prodrug inhibitorof dihydropteroate synthase (DHPS),

led to the first commercially available antibioticand established folate biosynthesis as the firstbiologically selective target for anti-infectives.

However, sulfonamides had little activity againstMycobacterium tuberculosis, the single leadingcause of death from bacterial infection (1). Shortlyafter Waksman’s and Schatz’s discovery of strepto-mycin, Lehmann developed para-aminosalicylicacid (PAS), a close structural analog of the folateprecursor p-aminobenzoate (PABA) (2). PAS be-came the second drug for tuberculosis (TB). Hitch-ings subsequently developed antibacterial inhibitorsof dihydrofolate reductase, validating a secondenzyme of folate metabolism as a therapeuticallyselective target (3). In parallel, studies of PAS incombination with streptomycin and eventuallyother anti-TB drugs laid the foundation for mod-ern TB chemotherapy.

1Department of Medicine, Weill Cornell Medical College, NewYork, NY 10065, USA. 2Department of Microbiology and Im-munology, Weill Cornell Medical College, New York, NY 10065,USA. 3Tuberculosis Research Section, Laboratory of ClinicalInfectious Diseases, National Institute for Allergy and Infec-tious Diseases (NIAID), National Institutes of Health (NIH),Bethesda, MD 20892, USA.

*To whom correspondence should be addressed. E-mail:kyr9001@med.cornell.edu

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org88

REPORTS

on

Janu

ary

4, 2

013

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om