Southeast Asian J Trop Med Public Health

EFFICIENT HETEROLOGOUS EXPRESSION AND ONE-STEP PURIFICATION OF FULLY ACTIVE C-TERMINAL HISTIDINE-TAGGED FROM MYCOBACTERIUM TUBERCULOSIS

Praweenuch Penpassakarn1, Pimchai Chaiyen2 and Prasit Palittapongarnpim3,4

1Department of Biotechnology, 2Department of Biochemistry and Center of Excellence in Protein Structure and Function, 3Department of Microbiology, Faculty of Science, Mahidol University, Bangkok; 4National Science and Technology Development Agency, Pathum Thani, Thailand

Abstract. Tuberculosis has long been recognized as one of the most significant public health problems. Finding novel antituberculous drugs is always a neces- sary approach for controlling the disease. Mycobacterium tuberculosis pyrH gene (Rv2883c) encodes for uridine monophosphate kinase (UMK), which is a key en- zyme in the uridine nucleotide interconversion pathway. The is essential for M. tuberculosis to sustain growth and hence is a potential drug target. In this study, we have developed a rapid protocol for production and purification ofM. tuberculosis UMK by cloning pyrH (Rv2883c) of M. tuberculosis H37Rv with the addition of 6-histidine residues to the C-terminus of the protein, and expressing in E. coli BL21-CodonPlus®(DE3)-RIPL using an auto-induction medium. The enzyme was efficiently purified by a single-step TALON® cobalt affinity chroma- tography with about 8 fold increase in specific activity, which was determined by a coupled assay with the and lactate dehydrogenase. The molecular mass of monomeric UMK was 28.2 kDa and that of the native enzyme was 217 kDa. The enzyme uses UMP as a but not CMP and TMP and

activity was enhanced by GTP. Measurements of revealed the kcat value of 7.6 ± 0.4 U mg-1 or 0.127 ± 0.006 sec-1.The protocol reported here can be used for expression of M. tuberculosis UMK in large quantity for formulating a high throughput target-based assay for screening anti-tuberculosis UMK compounds. Keywords: Mycobacterium tuberculosis, uridine monophosphate kinase, antituber- culous drugs, pyrimidine, protein purification

INTRODUCTION become more worrisome than in the past mainly due to the emergence of multi- The widespread of tuberculosis, drug resistant strains (Dheda et al, 2010). caused by Mycobacterium tuberculosis, has The cases are often untreatable, signifying an urgent need for new antituberculous Correspondence: Prasit Palittapongarnpim, agents. Department of Microbiology, Faculty of Science, Mahidol University, Rama VI Road, Bangkok Current antituberculous drugs are 10400, Thailand. known to act on a limited number of drug Tel: 66 (0) 81 867 4202 targets, while recent investigations have E-mail: [email protected] shown hundreds of potential drug targets

1452 Vol 42 No. 6 November 2011 Purification of M. tuberculosis Uridylate Kinase

(Duncan, 2004). It is therefore necessary Pyrimidine monophosphate to exploit these new drug targets for the found in bacteria are different from those development of new antituberculous of eukaryotes (Jones, 1980). There are compounds. three distinct prokaryotic pyrimidine Nucleotides are essential substrates monophosphate kinases, UMK, CMK and for bacterial growth, making interruption TMK, each specific for an individual py- of their metabolism being a viable strat- rimidylate. In contrast, phosphorylation egy in developing antituberculous drugs. of UMP and CMP in eukaryotic organisms Most bacteria obtain their nucleotides by is carried out by a bifunctional UMP/CMP both de novo biosynthesis and salvage kinase, which phosphorylates both UMP pathways. The pyrimidine de novo biosyn- and CMP with comparable efficiencies. thesis pathway appears to be nearly iden- The amino acid sequences of bacterial tical in all bacteria (Evrin et al, 2007). The CMK and eukaryotic UMP/CMP kinases six-membered pyrimidine is synthesized are similar to those of adenylate kinases, first from the reaction between carbamoyl while bacterial UMK is not homologous phosphate and aspartate, catalyzed by to any eukaryotic (Gagyi et al, aspartate transcarbamoylase. Dihydro- 2003). This suggests that UMK can serve orotase then catalyzes the closure of the as a potential anti-tuberculous drug target pyrimidine ring, which is subsequently as a chemical compound inhibiting M. oxidized to orotate by dihydroorotate tuberculosis enzyme should not affect the dehydrogenase using NAD+ as an electron human enzymes. Biochemical and cata- acceptor. Phosphoribosyl pyrophosphate lytic properties of the enzyme has recently (PRPP) is attached by the enzyme orotate been reported and the data are useful for phosphorybosyl , resulting in developing specific inhibitors against the orotidylate, which is decarboxylated to enzyme (Rostirolla et al, 2011), but the yield uridine 5’-monophosphate (UMP), production yield is low (20 mg/2g of wet which is subsequently phosphorylated by cell paste and 21% enzyme yield). In or- uridine monophosphate kinase (UMK) der to improve the yield, which would be to yield uridine 5’-diphosphate (UDP). beneficial for high-throughput inhibitor UDP is converted to uridine 5’-triphos- screen assays, we report the development phate (UTP) by a non-specific nucleoside of an expression and purification protocol diphosphate kinase, and subsequently for production of fully active C-terminal forms cytidine 5’-triphosphate (CTP) by histidine-tagged M. tuberculosis UMP cytidylate synthetase. In the salvage path- (MtUMKHis6) in high yield. The purified way, pyrimidine bases are converted to enzyme had the expected biochemical uracil and subsequently to UMP by UMK. properties of M. tuberculosis UMK, indi- To our knowledge, no viable bacterial cating that the heterologously expressed deletion mutant of UMK has ever been enzyme can be used in high throughput reported. Mushegian and Koonin (1996) drug screening assays. identify pyrH as a member of the minimal gene set a highly conserved set of genes MATERIALS AND METHODS supposed to provide minimal essential functions of bacterial life. Its essentiality Strains and plasmids for M. tuberculosis has been experimen- M. tuberculosis H37Rv was used for tally confirmed (Robertson et al, 2007). genomic DNA extraction. E. coli XL-1

Vol 42 No. 6 November 2011 1453 Southeast Asian J Trop Med Public Health blue (Invitrogen, Carlsbad, CA) was used compared. E. coli BL21-CodonPlus®(DE3)- for production of recombinant plasmid RIPL cells harboring pMtUMK were and E. coli BL21-CodonPlus®(DE3)-RIPL grown at 37ºC in 2xYT broth (16 g/l (Invitrogen, Carlsbad, Germany) for over- bacto-tryptone, 10 g/l yeast extract, and expression of pyrH. 0.5 g/l NaCl) supplemented with 50 µg/ml µ Cloning of M. tuberculosis pyrH coding kanamycin and 34 g/ml chlorampheni- sequence col. ZYP-5052 medium is composed of 10 g/l soluble enzymatic digest of casein, A 786 bp fragment containing the 5 g/l yeast extract, 50 mM Na HPO , 50 entire coding region of pyrH of M. tuber- 2 4 mM KH PO , 25 mM(NH ) SO , 2 mM culosis H37Rv was amplified by PCR using 2 4 4 2 4 MgSO , 0.5 g/l glucose and 2 g/l lactose pfu DNA with two primers, 4 and 0.2x trace metals (Studier, 2005) and containing NdeI and XhoI restriction sites supplemented with 50 µg/ml kanamy- (underlined), (pyrHF: 5’- CGTTAAC- cin and 34 µg/ml chloramphenicol. For CATATGACAGAGCCCGAT -3’, pyrHR: isopropyl-b-d-1-thiogalactopyranoside 5’-TTTCTCGAGGGTGGTGACCAGCGT (IPTG) induction, cells were grown until -3’). PCR was carried out in a 50 µl final the absorbance at 600 nm (OD ) reached volume containing 5 µl of 10x buffer 600 about 1.0. Then, induction was performed (Fermentas AB, Vilnius, Lituania), 25 µM by adding IPTG (Fermentas) to a final MgSO , 100 µM each dNTP, 0.5 µM each 4 concentration of 1 mM and was shaken primer, 2.0 U pfu DNA polymerase (Fer- at 20ºC, 250 rpm for 24 hours. For over- mentas) and 1 µg of M. tuberculosis H37Rv expression in auto-induction medium, DNA, using the following thermocycling E. coli BL21-CodonPlus®(DE3)-RIPL parameters: 5 minutes at 95ºC, 30 cycles of cells harboring pMtUMK were grown 95ºC for 60 seconds, 69.6ºC for 60 seconds at 37ºC in ZYP-5052. Cells were grown and 72ºC for 90 seconds with the final step until OD reached about 1.0 and then, of 5 minutes at 72ºC. PCR amplicons were 600 without adding IPTG, the culture was analyzed by electrophoresis in 1.5% (w/v) shaken at 20ºC, 250 rpm for 24 hours. agarose gel and stained with ethidium Bacteria were harvested by centrifugation bromide. The amplicon was inserted and at 4,000g for 10 minutes at 4ºC and stored ligated into pET-24b plasmid (EMB biosci- at -80ºC until used. Cell pastes from both ence, Madison, WI) at NdeI and XhoI sites methods of expression were suspended in using T4 DNA polymerase (Fermentas), 50 mM sodium phosphate buffer pH 7.4 resulting in the plasmid pMtUMK, which containing 1 mg/ml lysozyme and 1 mM was used for expression of the recom- phenyl methyl sulfonyl fluoride. Cells binant protein fused to His -tag at the 6 were lysed by sonication (6 x 10 s pulses) C-terminus. The whole sequence of pyrH on ice. Clarified cell lysate was obtained in the pMtUMK plasmid was sequenced by centrifugation at 20,800g for 30 min- (Macrogen, Seoul, Korea) to verify the utes at 4ºC. The recombinant protein was absence of any mutation. purified from the supernatant by affinity Overexpression and purification of recom- chromatography using 15 ml TALON® binant MtUMKHis6 Metal affinity column in ÄKTA™ auto- Overexpression of MtUMK in cells mated liquid chromatography system (GE grown in yeast extract tryptone (2xYT) and Healthcare, Uppsala, Sweden) employing in auto-induction medium (ZYP-5052) was a linear gradient of 5 to 400 mM imidazole.

1454 Vol 42 No. 6 November 2011 Purification of M. tuberculosis Uridylate Kinase

-3 -1 -1 Fractions containing MtUMKHis6 were = 6.22x10 M cm ), compared to the pooled, concentrated and desalted using control in which UMK was absent. One Sephadex G-25 gel filtration column previ- unit of UMK is defined as the amount of ously equilibrated with 50 mM Tris-Cl pH enzyme that converts 1 µmol of UMP to

7.4, 50 mM KCl and 2 mM MgCl2. Protein UDP per minute. Values of kcat and Km concentration was measured by Bradford were obtained by fitting the initial rates at Assay (Bradford, 1976). various concentrations of UMP and ATP Molecular mass determination to the Michaelis-Menten equation. For investigating the effects of GTP, assays of The subunit molecular mass of MtUMK at the GTP concentration of MtUMK was estimated by 12 % SDS- His6 His6 0, 0.1, 0.2, and 0.5 mM were conducted. PAGE. Estimation of the native molecular mass was based on the elution volume of RESULTS MtUMKHis6 from a Superdex™ 200 analy- tical gel filtration column in 50 mM Tris pH E. coli expression of MtUMK 7.4, 150 mM NaCl, operated by ÄKTA™ His6 M. tuberculosis pyrH was successfully automated liquid chromatography system cloned into E. coli BL21 CodonPlus®(DE3)- (GE Healthcare, Uppsala, Sweden). The RIPL. The recombinant plasmid, pM- following proteins with known molecu- tUMK, was isolated and sequenced. The lar mass were used as standard markers: plasmid contained the intact pyrH gene ribonuclease A, 15.6 kDa; oval albumin, including 6 histidine codons as expected 48.9 kDa; bovine serum albumin, 67 kDa; (data not shown). aldolase, 158 kDa; and ferrin 440 kDa (GE Healthcare, Uppsala, Sweden). Typically in our expression condi- tions, the 2YT medium yielded about Enzyme activity measurement 5-10 g of wet cell pellet per 1 liter while MtUMKHis6 activity was measured the auto-induction medium yielded about with pyruvate kinase (PK)/ lactate de- 20-25 g of wet cell pellet per 1 liter. The hydrogenase (LDH) coupling reaction amount of UMK per gram of wet cell paste as previously described (Blondin, 1994). purified from the IPTG-induction method Formation of ADP and UDP from ATP and in this study (7.2 mg) was slightly less UMP catalyzed by UMK is coupled to the than a previous report (10 mg) (Rostirolla + formation of lactate and NAD catalyzed et al, 2011). As shown in the protein puri- by the reactions of PK and LDH in the fication table (Table 1), the auto-induction presence of phosphenolpyruvate (PEP) medium produced 4-5 times more wet cell and NADH. paste and total protein than did the IPTG Enzyme activity was measured in induction medium from the same amount a total volume of 500 µl at 25ºC using a of culture medium while the specific Shimadzu UV-2501PC spectrophotometer. enzymatic activity of the total proteins The assay reaction contained 0.3-0.45 nM was also slightly higher. The His-tagged MtUMKHis6, 50 mM Tris HCl (pH7.5), 50 enzyme from both preparations could be mM KCl, 2 mM MgCl2, 3 mM PEP, 0.2 purified to the same final specific activity mM NADH, 3 mM ATP, 3 mM UMP, and of about 7.6 U/mg, which is very similar 28 and 16 units of PK and LDH, respec- to the value recently reported for the re- tively. Initial rate was calculated from the combinant UMK without a tag (Rostirolla e decrease in absorbance at 340 nm ( nadh et al, 2011).

Vol 42 No. 6 November 2011 1455 Southeast Asian J Trop Med Public Health

kDa

66.2

1.00 8.65 1.00 8.20 45.0

35.0 Purification fold MtUMKHis6 25.0

18.4 66 93 100 100 Yield (%) Yield

1 2 3 4 5 6 mg) /

Fig 1–Expression of MtUMKHis6 in E. coli BL21(DE3)-RIPL using IPTG-induction

0.88 7.59 0.93 7.66 and auto-induction analyzed by 12 % SDS-PAGE. The bacteria were grown in 2xYT or ZYP-5052 media, respectively until OD 600 mm reached 1.0 and then Specific activity (U at 20oC for 24 hours with or without IPTG. Lane 1, molecular weight markers

(kDa); lane 2, purified MtUMKHis6 from Table 1 Table IPTG induction (1 µg); lane 3, purified from IPTG induction and auto-induction. µ MtUMKHis6 from auto-induction (1 g); His6 lane 4, clarified lysate of the control, E. 370 243

1,168 1,087 coli BL21(DE3)-RIPL harboring pET24b without MtUMK insert (17 µg); lane 5, clarified lysate ofE. coli BL21(DE3)-RIPL harboring pMtUMK induced by IPTG Total enzyme activity (U) Total (15 µg); lane 6, clarified lysate of E. coli From wet cell pellet 10.2 g (0.5 liter culture) From

b BL21(DE3)-RIPL harboring pMtUMK in auto-induction medium (15 µg). 32 421 142 1,250 Purification table of MtUMK Total protein (mg) protein Total The purified enzyme contained a single 28-kDa protein band (Fig 1) as expected from the sequence analysis, b a confirming the purity of the enzyme. Af- ter removing imidazole from the purified UMK by exchanging buffer in a desalting

Crude extract Crude IMAC column extract Crude IMAC column column and kept at -80ºC, the frozen en- zyme has maintained its original activity Step IPTG induction Auto-induction From wet cell pellet 4.4 g (1 liter culture); wet cell pellet 4.4 g (1 liter culture); From a for at least 6 months.

1456 Vol 42 No. 6 November 2011 Purification of M. tuberculosis Uridylate Kinase

Molecular mass determination Table 2

A subunit molecular mass of MtUM- MtUMKHis6 activity in the presence of GTP. K was calculated based on its sequence His6 GTP (mM) Specific activity Relative as 28.2 kDa, consistent with that of the (U/mg) activity purified protein estimated by SDS-PAGE (Fig 1). The native molecular mass of No GTP 7.54 1.00

MtUMKHis6 was estimated to be about 217 0.1 mM 12.35 1.64 kDa based in Superdex S-200 gel filtration. 0.2 mM 14.23 1.89 Therefore, the oligomeric state of MtUM- 0.5 mM 14.14 1.87

KHis6 should be around 8.

MtUMKHis6 kinetics and specificity

MtUMKHis6 activity was determined The availability of the purified by coupling to the reactions catalyzed by MtUMK would certainly facilitate the PK and LDH in the presence of PEP and development of inhibitors by making it NADH. Steady-state kinetics of MtUM- possible to perform a high-throughput

KHis6 with varying UMP or ATP concentra- target-based assay. It is also a preliminary tions at a saturating concentration of ATP step to acquire high-resolution structural or UMP (5 mM) revealed Km for UMP of information of the protein, so that vari- 1.9 mM ± 0.9 mM, Km for ATP of 2.0 mM ous approaches for structure-based drug -1 ± 1.1 mM, and kcat of 7.6 ± 0.4 U mg . designs can be applied. The recombinant enzyme could not In this study, MtUMK has been cloned use CMP or TMP as substrates, even at and expressed as a C-his-tagged protein very high concentration of CMP and and has been shown that it retains the TMP of 10-50 mM (data not shown). We highly specific kinase activity for UMP examined the activity of the enzyme in but not for CMP and TMP, emphasizing the presence of 0.1, 0.2 and 0.5 mM GTP its difference from UMP/CMP eukaryotic and found that the UMK activity reached kinases. It can also be activated by GTP maximum at 0.2 mM GTP with activity indicating that the MtUMKHis6 is allosteri- being about twice the value in the absence cally controlled by GTP, similar to UMK of GTP (Table 2). from E. coli and other bacteria (Gagyi et al, 2003; Evrin et al, 2007; Tu et al, 2009). DISCUSSION While this work is ongoing, pyrH of M. tuberculosis has been cloned and Pyrimidines are essential components expressed in its native form but the of DNA and RNA, making interrupting purification is more tedious, requiring their metabolic pathway being an interest- multiple steps of chromatography. Unlike ing strategy for developing antitubercu- all known prokaryotic UMK, the heter- lous drugs. Many works have been done on ologously expressed native MtUMK was M. tuberculosis (Gasse a tetramer (Rostirolla et al, 2011), which et al, 2008; Familiar et al, 2010), resulting was different from most native bacterial in the discovery of several interesting and archaeal UMK, which are generally compounds. M. tuberculosis CMK (Thum hexamers, such as E. coli UMK, which is et al, 2009) has been recently cloned and 56.4% identical to MtUMKHis6 (Briozzo could be further used for drug discovery. et al, 2005), Ureaplasma parvum (Egeblad-

Vol 42 No. 6 November 2011 1457 Southeast Asian J Trop Med Public Health

Welin et al, 2007), Bacillus anthracis (Meier REFERENCES et al, 2008), Pyrococcus furiosus (Marco- Blondin C, Serina L, Wiesmüller L, Gilles AM, Marin et al, 2005) and Sulfolobus solfataricus Bârzu O. Improved spectrophotometric as- (Jensen, 2007). say of nucleoside monophosphate kinase In contrast, C-His-MtUMK was found activity using the pyruvate kinase/lactate to be an octamer in this study, which was dehydrogenase coupling system. Anal different from a report of a hexameric Biochem 1994; 220: 219-21. structure of recombinant MtUMKHis6 Bradford MM. A rapid and sensitive method for expressed in E. coli (Labesse et al, 2010). the quantitation of microgram quantities The cause of the discrepancy is unknown. of protein utilizing the principle of protein- It may be possible that the enzyme can dye binding. Anal Biochem1976; 72: 248-54. adopt different quaternary structures Briozzo P, Evrin C, Meyer P, et al. Structure of depending on the environment. This war- Escherichia coli UMP kinase differs from rants further study. that of other nucleoside monophosphate In summary, we have developed a kinases and sheds new light on enzyme regulation. J Biol Chem 2005; 280: 25533-40. method for production and purification of MtUMK with higher yield than the Dheda K, Warren RM, Zumla A, Grobusch MP. previously reported protocol while attain- Extensively drug-resistant tuberculosis: epidemiology and management chal- ing the same level of purity and specific lenges. Infect Dis Clin North Am 2010; 24: activity. The purified enzyme exhibited 705-25. the functional characteristics expected Duncan K. Identification and validation of for bacterial UMK, such as, being specific novel drug targets in tuberculosis. Curr to UMP but not catalyzing the reaction Pharm Des 2004; 10: 3185-94. of CMP or TMP, and being enhanced by Egeblad-Welin L, Welin M, Wang L, Eriksson GTP. The preparation of the histidine- S. Structural and functional investigations tagged enzyme significantly simplifies the of Ureaplasma parvum UMP kinase – a purification and also minimizes the pos- potential antibacterial drug target. FEBS J sibility of contamination by E. coli UMK. 2007; 274: 6403-14. The auto-induction protocol results in a Evrin C, Straut M, Slavova-Azmanova N, et al. high yield and also eliminates the need for Regulatory mechanisms differ in UMP IPTG, which significantly reduces the cost kinases from gram-negative and gram- of the enzyme preparation. This should be positive bacteria. J Biol Chem 2007; 282: beneficial for preparation of MtUMK for 7242-53. high-throughput drug-screening assays Familiar O, Munier-Lehmann H, Ainsa JA, in the future. Camarasa MJ, Pérez-Pérez MJ. Design, synthesis and inhibitory activity against ACKNOWLEDGEMENTS Mycobacterium tuberculosis thymidine mo- nophosphate kinase of acyclic nucleoside Praweenuch Penpassakarn is sup- analogues with a distal imidazoquino- ported by Mahidol University Medical linone. Eur J Med Chem 2010; 45: 5910-8. Scholarship (PhD-MD) Program. We Gagyi C, Bucurenci N, Sîrbu O, et al. UMP would like to thank Mahidol University kinase from the Gram-positive bacterium for the financial support of this work. We Bacillus subtilis is strongly dependent on thank Dr Kittipat Sopitthummakhun for GTP for optimal activity. Eur J Biochem his help and suggestions. 2003; 270: 3196-204.

1458 Vol 42 No. 6 November 2011 Purification of M. tuberculosis Uridylate Kinase

Gasse C, Douguet D, Huteau V, Marchal G, nucleotide-. J Mol Biol 2008; Munier-Lehmann H, Pochet S. Substituted 381: 1098-105. benzyl-pyrimidines targeting thymidine Mushegian AR, Koonin EV. A minimal gene set monophosphate kinase of Mycobacterium for cellular life derived by comparison of tuberculosis: synthesis and in vitro anti- complete bacterial genomes. Proc Natl Acad mycobacterial activity. Bioorg Med Chem Sci USA 1996; 93: 10268-73. 2008; 16: 6075-85. Robertson D, Carroll P, Parish T. Rapid recom- Jensen, KS, Johansson E, Jensen, KF. Structural bination screening to test gene essential- and enzymatic investigation of the Sul- ity demonstrates that pyrH is essential in folobus solfataricus uridylate kinase shows Mycobacterium tuberculosis. Tuberculosis competitive UTP inhibition and the lack 2007; 87: 450-8. of GTP stimulation. Biochemistry 2007; 46: 2745-57. Rostirolla DC, Breda A, Rosado LA, Palma MS, Basso LA, Santos DS. UMP kinase from Jones ME. Pyrimidine nucleotide biosynthesis Mycobacterium tuberculosis: Mode of action in animals: Genes, enzymes, and regula- and allosteric interactions, and their likely tion of UMP biosynthesis. Ann Rev Bio- role in pyrimidine metabolism regulation. chem 1980; 49: 253-79. Arch Biochem Biophys 2011; 505: 202-12. Labesse G, Benkali K, Salard-Arnaud I, Gilles AM, Munier-Lehmann H. Structural and Studier FW. Protein production by auto-in- functional characterization of the Mycobac- duction in high density shaking cultures. terium tuberculosis uridine monophosphate Protein Expr Purif 2005; 41: 207-34. kinase: insights into the allosteric regula- Thum C, Schneider CZ, Palma MS, Santos DS, tion. Nucleic Acids Res 2010; 39: 3458-72. Basso LA. The Rv1712 Locus from Myco- Marco-Marin C, Gil-Ortiz F, Rubio V. The bacterium tuberculosis H37Rv codes for a crystal structure of Pyrococcus furiosus functional CMP kinase that preferentially UMP kinase provides insight into phosphorylates dCMP. J Bacteriol 2009; and regulation in microbial pyrimidine 191: 2884-7. nucleotide biosynthesis. J Mol Biol 2005; Tu JL, Chin KH, Wang AH, Chou SH. Unique 352: 438-54. GTP-binding pocket and allostery of Meier C, Carter LG, Sainsbury S, et al. The crys- uridylate kinase from a gram-negative tal structure of UMP kinase from Bacillus phytopathogenic bacterium. J Mol Biol anthracis (BA1797) reveals an allosteric 2009; 385: 1113-26.

Vol 42 No. 6 November 2011 1459