FEMS Microbiology Letters, 362, 2015, fnv101

doi: 10.1093/femsle/fnv101Advance Access Publication Date: 22 June 2015Research Letter

RESEARCH LETTER –Physiology & Biochemistry

Simultaneous presence of fhs and purT genesis disadvantageous for the fitness of Escherichiacoli growthSrinivas Aluri1, Kervin Rex1 and Umesh Varshney1,2,∗

1Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore 560012, Indiaand 2Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore 560064, India∗Corresponding author: Department of Microbiology and Cell Biology, Indian Institute of Science, CNR Rao Circle, Bangalore 560012, India.Tel: +918022932686; Fax: +918023602697; E-mail: varshney@mcbl.iisc.ernet.in; uvarshney@gmail.comOne sentence summary: Using Escherichia coli model, we show that the mutual exclusion of purT and fhs, observed in majority (∼94%) of bacteriaconfers fitness advantage.Editor: Andre Klier

ABSTRACT

In bacteria, alternate mechanisms are known to synthesize N10-formyltetrahydrofolate (N10-formyl-THF) and formylglycinamide ribotide (fGAR), which are important in purine biosynthesis. In one of the mechanisms, a direct transfer of onecarbon unit from formate allows Fhs to convert tetrahydrofolate to N10-formyl-THF, and PurT to convert glycinamideribotide (GAR) to fGAR. Our bioinformatics analysis of fhs and purT genes (encoding Fhs and PurT) showed that in a majorityof bacteria (∼94%), their presence was mutually exclusive. A large number of organisms possessing fhs lacked purT and viceversa. The phenomenon is so penetrating that even within a genus (Bacillus) if a species possessed fhs it lacked purT andvice versa. To investigate physiological importance of this phenomenon, we used Escherichia coli, which naturally lacks fhs(and possesses purT) as model. We generated strains, which possessed fhs and purT genes in singles or together. Deletion ofpurT from E. coli in the presence or absence of fhs did not confer a detectable growth disadvantage in pure cultures. However,growth competition assays revealed that the strains possessing either of the single genes outcompeted those possessingboth the genes suggesting that mutual exclusion of purT and fhs in organisms confers fitness advantage in mixed cultures.

Keywords: anti-correlative presence of genes; growth fitness advantage; growth competitions

INTRODUCTION

Fhs catalyzes transfer of one carbon unit from formate to THFby utilizing ATP as the source of energy donor for its ligase re-action and contributes to the synthesis of the cellular pool ofN10-formyltetrahydrofolate (N10-formyl-THF) (Fig. 1) (Himes andRabinowitz 1962; Rabinowitz and Pricer 1962). N10-formyl-THF,besides its use in other metabolic reactions, is employed as acofactor at two steps in purine biosynthetic pathway, namelythe PurN reaction and the PurH reaction (Voet, Voet and Pratt2013). PurN reaction catalyzes the formation of formyl glyci-

namide ribotide (fGAR) from glycinamide ribotide (GAR) (Ingleseet al. 1990). In some bacteria, PurN alone catalyzes this reaction.However, in some others it could also be catalyzed by PurT (Ny-gaard and Smith 1993). PurT is a formate-dependent GAR trans-formylase or phosphoribosylglycinamide formyltransferase II.PurT uses formate as cofactor and ligates formate to GAR, whichinvolves ATP hydrolysis (Fig. 1). When both PurN and PurT arepresent in an organism, radioactive tracing experiments showedthat in a completely synthesized purine, 50% of the carbon in-corporated at C8 position was through the PurN reaction and

Received: 19 April 2015; Accepted: 17 June 2015C© FEMS 2015. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

1

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from

2 FEMS Microbiology Letters, 2015, Vol. 362, No. 14

Figure 1. (A) Schematic of the synthesis of purine ring. Purine synthesis starts with phophoribosyl pyrophosphate (1) and by a series of reactions results into the

synthesis of inosine monophosphate (6). Among the enzymes shown, Fhs synthesizes N10- formyl THF from formate and THF. PurN converts GAR (2) to fGAR (3) usingN10-formyl THF as cofactor. The reaction is also carried out by PurT using formate as cofactor. PurH catalyzes the conversion of aminoimidazole carboxamide ribotide(4) to formyl aminoimidazole carboxamide ribotide (5) using N10-formyl THF as cofactor. Fhs, Formyltetrahydrofolate synthetase; PurN, phosphoribosylglycinamideformyltransferase; PurT, phosphoribosylglycinamide formyltransferase II; and PurH, Amino imidazole carboxamide ribotide formyl transferase. (B) Pie diagram show-

ing the occurrence of fhs and purT in different genera. (C) The occurrence of fhs and purT in Bacillus genus. Species harboring fhs alone or purT alone are indicated.Analysis also shows that the glyA, folD and purN are ubiquitously present.

the other 50% through the PurT reaction (Dev and Harvey 1982).Although, the contribution from PurT, for purine synthesis isabout 50%, mutation in purT does not confer growth defect inEscherichia coli (Nygaard and Smith 1993). In E. coli, it was re-ported that formate released by PurU catalyzed deformylationof N10-formyl-THF serves as the source for PurT reaction (Nagy,McCorkle and Zalkin 1993).

The reactions catalyzed by Fhs and PurT share many sim-ilarities. Firstly, both the reactions carry out a ligase function;secondly, in both the reactions, formate is the common sub-strate and, thirdly, both the reactions require ATP hydrolysis forcatalysis. The end products also share similarities. The productformed by Fhs is used as cofactor for purine biosynthesis, andthe product formed by PurT is a purine biosynthesis intermedi-ate. Though both the proteins share many similarities in theirfunction to synthesize purines, their presence and distributionin various organisms have not been investigated.

We have earlier shown that fhs is present in some genera ofbacteria, and its presence confers growth advantage under hy-poxia (Sah et al. 2015). While analyzing distribution of fhs across

genomes, we found that a large number of organisms possess-ing fhs lacked purT, a phenomenon called anticorrelative pres-ence of the genes. Earlier, preliminary data were available forthe distribution of purT, where the authors reported that PurTactivitieswere present in Gram-positive andGram-negative bac-teria, and in archaea e.g. Sulfolobus shibatae. But the exact corre-lation between the distribution of fhs and purT was unknown.In this report, by using bioinformatics tools we have analyzedthe genomes for the presence of fhs and purT in different generaand found that fhs and purT indeed show a strong anticorrela-tion for their occurrence in various organisms. The anticorrela-tion between the occurrences of fhs and purT is so striking thateven within the genus of Bacillus, if one species possess fhs itlacks purT and vice versa. Since we found such a strong anti-correlation for the occurrence of fhs and purT, it was of interestto understand its physiological importance in the organisms. Toinvestigate the biological significance of this phenomenon, weused growth competition experiments using E. coli, a facultativeanaerobe that grows under both the microaerobic and aerobicconditions. Our studies reveal that a simultaneous presence of

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from

Aluri et al. 3

Table 1. Strains and plasmids used in this study.

Escherichia coli strains/plasmids Genotype References

TG1 E. coli K-12 supE thi-1 �(lac-ProAB) �(mcrB-hsdSM) Sambrook and Russell (2001)JW1838-1 E. coli F−, �(araD-araB)567, �lacZ4787(::rrnB-3), λ−, �purT774::kan, rph-1,

�(rhaD-rhaB)568, hsdR514Baba et al. (2006)

TG1�purT E. coli TG1 deleted for purT This studyTG1/pACDHfhs E. coli TG1 harboring pACDHfhs Sah et al. (2015)TG1�purT/pACDHfhs E. coli TG1�purT harboring pACDHfhs This studypACDHfhs fhs gene open reading frame from Clostridium perfringens was PCR

amplified and cloned into pACDH vector for expression in E. coliSah et al. (2015)

both the genes confers growth disadvantage compared to whenpresent in single.

MATERIALS AND METHODSBioinformatics

Genome sequences available for different bacteria were col-lected and analyzed using SEED (Overbeek et al. 2005) andUniprot databases. Using computational analyses, genomeswere screened for the presence or absence of fhs and purT.Genomes positive for fhs and purT were pooled and distributedin two groups fhs group and purT group. In the genomes wherefhs was present, status of purT was checked by BLAST analyses.Likewise, the genomes where purT was present were analyzedfor the status of fhs.

Plasmids and strains

Generation of pACDHfhs, a medium copy plasmid harboringClostridium perfringens fhs open reading frame for expression inE. coli has been described earlier (Sah et al. 2015). Escherichia colistrain deleted for purT (�purT::kan, Table 1) (Baba et al. 2006) wasobtained fromCGSC and the �purT::kan locus from it wasmovedinto E. coliTG1 by P1 phage-mediated transduction. purT deletionwas confirmed by PCR with Taq DNA polymerase using PurT KOC FP (5′-CAATAAAGACACACGCAAACG-3′) and PurT KO C RP (5′-GCGCGTGAAGCTGTAGAAG-3′) primers (∼10 pmoles each). PCRinvolved 30 cycles of initial denaturation at 94◦C for 1 min, an-nealing at 50◦C for 30 s and extension at 70◦C for 2 min.

Growth competitions

Exponential phase cultures (OD600 ∼0.6–0.7) of various strainsof E. coli TG1, E. coli TG1�purT::kan (KanR), E. coli TG1/pACDHfhs(TetR) and E. coli TG1�purT::kan/pACDHfhs (KanRTetR) weregrown in LB (2 mL) with appropriate antibiotics and washedwith plain M9 minimal (antibiotic free) and resuspended in thesamemedium (1mL) whichwas used as seeding culture. Growthcompetition experiments were set up by mixing the competingstrains in different ratios. Vitamin C (10mM)was added to screwcap flat bottom tubes containing media to create hypoxic envi-ronment (Taneja et al. 2010). Hypoxic cultures were grown withslow stirring on amultipointmagnetic board at 37◦C andmethy-lene blue was used as an oxygen indicator. Cultures grown forvarious periods were analyzed for total viable counts at varioustimes by dilution plating on LB plates. To determine the iden-tity of the individual colonies, the colonies were picked ran-domly and patched on LB plates lacking any antibiotic or con-taining Tet, Kan or Kan plus Tet to distinguish between E. coli

TG1, E. coli TG1�purT (KanR), E. coli TG1/pACDHfhs (TetR) and E.coli TG1�purT/pACDHfhs (KanRTetR). A colony growing only onthe no antibiotic plate was identified as E. coli TG1, and a colonygrowing on all plates as E. coli TG1�purT/pACDHfhs. Likewise, acolony growing only on no antibiotic plate and Tet plate as E. coliTG1/pACDHfhs (TetR), and the one growing only on no antibi-otic and Kan plate as E. coli TG1�purT (KanR). Positive coloniesof each strain were counted and plotted using Graphpad prism.

RESULTSAnalysis of the presence of fhs and purT

In bacteria, fhs and purT are present as single genes. Our initialanalyses revealed that a large number of genomes possessingfhs lacked purT. To better investigate this phenomenon, we col-lected genome records fromUniprot and Seed Viewer databases.Among the 176 genera that were analyzed in this study, 94% con-curred with the anticorrelative presence of fhs and purT genesand a mere 6% deviated from the hypothesis (Fig. 1B). Our anal-yses revealed that purT is absent from most of the firmicutes.However, it is present in some species of Bacillus. Interestingly,within the genus of Bacillus when a species possessed purT itlacked fhs, showing a strong anticorrelation of the presence ofthe two genes even within a genus (Fig. 1C). As control, whenwe analyzed the genomes for the presence of glyA, folD and purNgenes, these were found in all the species (Fig. 1C).

purT deletion does not reveal growth defectin E. coli possessing fhs

To understand the physiological importance of the anticor-relative presence of the purT and fhs genes, we used E. coli,which lacks fhs as a model system. For this purpose, we moved�purT::kan locus (also called as �purT) into E. coli TG1 and con-firmed the genetic change by colony PCR (Figs S1A and B, Sup-porting Information). We then carried out growth curve analy-ses for E. coli TG1 expressing fhs and E. coli TG1�purT expressingfhs. Consistent with an earlier report (Nygaard and Smith 1993),when grown as pure cultures, deletion of purT did not revealany significant impact either in the presence or absence of fhs(Fig. S2, Supporting Information). Given that deletions of non-essential genes often do not result in significant growth defects,particularly under nutrient-rich conditions, this was not an un-expected phenotype.More importantly, this observation allowedus to investigate if the fhs and purT genes conferred any fitnessadvantage to the strains harboring them by investigating theirrelative survival in mixed cultures where the strains competefor the availability of the nutrients.

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from

4 FEMS Microbiology Letters, 2015, Vol. 362, No. 14

Figure 2.Growth competition experiment between E. coliTG1 (in green) and E. coli

TG1�purT (in red). Escherichia coli TG1 and E. coli TG1�purT cultures were mixed

and grown under aerobic or hypoxic conditions. Percent occurrence (survival)of the strains was determined at different times using the distinctive antibioticresistance phenotypes of the strains (section Materials and Methods).

Growth competition experiments

Recently, we reported that the expression of fhs confers competi-tive growth advantage under hypoxic conditions (Sah et al. 2015).This may be because the presence of Fhs could detoxify the for-mate formed under the hypoxic conditions of growth. Becauseboth PurT and Fhs commonly utilize formate for their functions,we carried out growth competition experiments by mixing cul-tures of various strains expressing PurT and/or Fhs in differ-ent combinations, and growing them under aerobic and hypoxicconditions. Abundance of individual strains at different times ofcompetitive growth was assessed as described in Materials andMethods section. As shown in Fig. 2, deletion of purT from theE. coli genome conferred growth disadvantage under both theaerobic and hypoxic conditions (compare E. coli TG1 and E. coliTG1�purT). The growth defect of purT deletion was more promi-nent under hypoxic conditions (known to generate formate). Ina mixed culture starting with approximately equal populationsof the two strains, E. coli TG1�purT was outcompeted nearlycomprehensively by E. coli TG1 in about 25 h. Likewise, E. coliTG1�purT was at a disadvantage in the growth competition ex-periments with E. coli TG1�purT/pACDHfhs also (Fig. 3). In thisexperiment, starting with its equal presence in the beginning ofthe experiment, E. coli TG1�purT was eliminated from the cul-

Figure 3. Growth competition experiment between E. coli TG1�purT (in red)and E. coli TG1�purT/pACDHfhs (in purple). Escherichia coli TG1�purT and E. coli

TG1�purT/pACDHfhs cultures were mixed and grown under aerobic or hypoxicconditions. Percent occurrence (survival) of the strains was determined at dif-ferent times using the distinctive antibiotic resistance phenotypes of the strains

(section Materials and Methods).

ture in about 25 h both under the aerobic and hypoxic condi-tions of growth suggesting that the presence of Fhs on a low-copy plasmid was more effective in compensating for the defi-ciency of PurT. In fact, E. coli TG1�purT/pACDHfhs outcompetedeven E. coli TG1 with the same efficiency both under the aero-bic and hypoxic conditions (Fig. 4). We then decided to set upa growth competition assay between E. coli TG1�purT and E.coli TG1/pACDHfhs. In this experiment, a clear advantage of thepresence of Fhs was seen only under the hypoxic conditions ofgrowth (Fig. 5). However, to better understand the impact of thesimultaneous presence of both PurT and Fhs, we then set up yetanother competition assay between E. coli TG1�purT/pACDHfhsand E. coli TG1/pACDHfhs (Fig. 6). This combination offered uswith a control where presence of Fhs was common in both thecompeting pairs but one possessed deletion of purT and theother did not. In this experiment, E. coli TG1�purT/pACDHfhsoutcompeted E. coli TG1/pACDHfhs both under the aerobic andhypoxic conditions of growth, with its decisive predominancein about 25h. This experiment also provided a control thatdeletion of purT in itself was not disadvantageous; it is ratherthe simultaneous presence of both purT and fhs, which is

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from

Aluri et al. 5

Figure 4. Growth competition experiment between E. coli TG1 (in green)and E. coli TG1�purT/pACDHfhs (in purple). Escherichia coli TG1 and E. coli

TG1�purT/pACDHfhs cultures were mixed and grown under aerobic or hypoxicconditions. Percent occurrence (survival) of the strains was determined at dif-ferent times using the distinctive antibiotic resistance phenotype of the strains

(section Materials and Methods).

disadvantageous when the strains compete for the commonnutrient pool.

DISCUSSION

Although the enzymes involved in folic acid and purine biosyn-thesis have been extensively studied, the details of the occur-rence of genes that encode them in different genera of bacteriahave remained unexplored. Some of the genes in the one car-bon metabolic pathway, e.g. fhs, are not omnipresent. The phe-nomenon of anticorrelative presence of the genes amongst theorganisms observed in this study is not a first. Earlier studies ob-served such a phenomenon for the presence of YgfA and FT-CD(Jeanguenin et al. 2010). Bacteria lacking YgfA always possessedFT-CD to perform a moon lighting function of YgfA.

We have earlier shown that the occurrence of fhs is largelylimited to bacteria, which are either facultative anaerobes or ob-ligate anaerobes (Sah et al. 2015). As shown in Fig. 1B, the presentstudy showed that among the 176 genera that we analyzed (Ta-ble S1, Supporting Information), a mere 6% possessed both thefhs and purT genes. The observation of anticorrelation of the

Figure 5.Growth competition experiment between E. coli TG1/pACDHfhs (in blue)and E. coli TG1�purT (in red). Escherichia coli TG1/pACDHfhs and E. coli TG1�purT

cultures were mixed and grown under aerobic or hypoxic conditions. Percent

occurrence (survival) of the strains was determined at different times using thedistinctive antibiotic resistance phenotype of the strains (section Materials andMethods).

presence of the two genes even within a single genus (Bacillus)is striking, suggesting that simultaneous presence of both thegenes may be detrimental to the natural fitness of the organ-isms. To test this hypothesis, we made use of E. coli, which nat-urally lacks fhs but possesses purT. We used E. coli TG1 or E. coliTG1�purT strains wherein fhs was either expressed or not ex-pressed. Furthermore, as we had shown that fhs confers growthadvantage under hypoxic growth conditions, the growth compe-tition experiments were performed not only under aerobic con-ditions but also under hypoxic conditions. Growth curve anal-yses revealed that when grown as pure cultures, the presenceof both genes together or either of the genes alone has no de-tectable growth phenotype in LB or M9 minimal media. How-ever, the growth competition experiment between E. coliTG1 andE. coli TG1�purT revealed that E. coli TG1 efficiently outcompetesthe latter under hypoxic conditions, and even under aerobic con-ditions TG1 shows a slightly better fitness over E. coli TG1�purT(Fig. 2). This observation is not surprising as the fitness disad-vantage of purT null strain could be due to the increased re-quirement of PurT during exponential growth phase in mini-mal media (Saxild, Jacobsen and Nygaard 1995). It was shownby the authors that PurT expression increased in exponen-tial phase when grown in minimal media. TG1�purT/pACDHfhs

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from

6 FEMS Microbiology Letters, 2015, Vol. 362, No. 14

Figure 6.Growth competition experiment between E. coli TG1/pACDHfhs (in blue)and E. coli TG1�purT/pACDHfhs (in purple). Escherichia coli TG1/pACDHfhs and E.

coli TG1�purT/pACDHfhs cultures were mixed and grown under aerobic or hy-poxic conditions. Percent occurrence (survival) of the strains was determinedat different times using the distinctive antibiotic resistance phenotype of the

strains (section Materials and Methods).

outcompeted E. coli TG1, E. coli TG1/pACDHfhs and E. coliTG1�purT under both the aerobic and hypoxic conditions.We have earlier shown that E. coli TG1 outcompetes E. coliTG1/pACDHfhs under aerobic conditions but the opposite is trueunder hypoxic conditions (Sah et al. 2015). However, when purTis deleted in E. coli TG1/pACDHfhs (i.e. TG1�purT/pACDHfhs) itoutcompetes E. coli TG1 under both the conditions tested. Theseobservations are consistent with the hypothesis that the pres-ence of both purT and fhs genes is disadvantageous for the fit-ness of the organism. Our analyses show that purT is absent inmost of the anaerobic and facultative anaerobic bacteria wherefhs is predominantly present. However, how this anticorrelativepresence of purT and fhs is beneficial to bacteria for its fitnessremains unclear.

SUPPLEMENTARY DATA

Supplementary data is available at FEMSLE online.

ACKNOWLEDGEMENTS

We thank our laboratory colleagues for their suggestions on themanuscript.

FUNDING

This workwas supported by grants from the Departments of Sci-ence andTechnology (DST), and Biotechnology (DBT), NewDelhi.UV is a J. C. Bose fellow of DST. SA was supported a senior re-search fellowship of the Council of Scientific and Industrial Re-search, New Delhi.

Conflict of interest. None declared.

REFERENCES

Baba T, Ara T, Hasegawa M, et al. Construction of Escherichia coliK-12 in-frame, single-gene knockout mutants: the Keio col-lection. Mol Syst Biol 2006;2:1–11.

Dev IK, Harvey RJ. Sources of one-carbon units in the folate path-way of Escherichia coli. J Biol Chem 1982;257:1980–6.

Himes RH, Rabinowitz JC. Formyltetrahydrofolate synthetase.III. Studies on the mechanism of the reaction. J Biol Chem1962;237:2915–25.

Inglese J, Johnson DL, Shiau A, et al. Subcloning, characteriza-tion, and affinity labeling of Escherichia coli glycinamide ri-bonucleotide transformylase. Biochemistry 1990;29:1436–43.

Jeanguenin L, Lara-Nunez A, Pribat A, et al. Moonlightingglutamate formiminotransferases can functionally re-place 5-formyltetrahydrofolate cycloligase. J Biol Chem2010;285:41557–66.

Nagy PL, McCorkle GM, Zalkin H. purU, a source of formate forpurT-dependent phosphoribosyl-N-formylglycinamide syn-thesis. J Bacteriol 1993;175:7066–73.

Nygaard P, Smith JM. Evidence for a novel glycinamide ri-bonucleotide transformylase in Escherichia coli. J Bacteriol1993;175:3591–7.

Overbeek R, Begley T, Butler RM, et al. The subsystems approachto genome annotation and its use in the project to annotate1000 genomes. Nucleic Acids Res 2005;33:5691–702.

Rabinowitz JC, Pricer WE, Jr. Formyltetrahydrofolate synthetase.I. Isolation and crystallization of the enzyme. J Biol Chem1962;237:2898–902.

Sah S, Aluri S, Rex K, et al. One-carbon metabolic pathwayrewiring in Escherichia coli reveals an evolutionary advantageof 10-formyltetrahydrofolate synthetase (Fhs) in survival un-der hypoxia. J Bacteriol 2015;197:717–26.

Sambrook J, Russell DW. Molecular Cloning: A Laboratory Manual.New York: Cold Spring Harbor Laboratory Press, 2001.

Saxild HH, Jacobsen JH, Nygaard P. Functional analysis of theBacillus subtilis purT gene encoding formate-dependentglycinamide ribonucleotide transformylase. Microbiology1995;141:2211–8.

Taneja NK, Dhingra S, Mittal A, et al. Mycobacterium tuberculo-sis transcriptional adaptation, growth arrest and dormancyphenotype development is triggered by vitamin C. PLoS One2010;5:0010860.

Voet D, Voet JG, Pratt CW. Fundamentals of Biochemistry: Life at theMolecular Level. Hoboken, NJ: Wiley, 2013.

by guest on July 21, 2016http://fem

sle.oxfordjournals.org/D

ownloaded from