A comparative biochemical and structural analysis of the intracellular chorismate mutase (Rv0948c) from Mycobacterium tuberculosis H 37 R v and the secreted chorismate mutase (y2828) from Yersinia pestis Sook-Kyung Kim*, Sathyavelu K. Reddy, Bryant C. Nelson, Howard Robinsonà, Prasad T. Reddy and Jane E. Ladner Biochemical Science Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD, USA Keywords chorismate mutase; Mycobacterium tuberculosis; pathogenesis; shikimate pathway; Yersinia pestis Correspondence P. T. Reddy, Biochemical Science Division, Bldg 227, Rm B244, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA Fax: +1 301 975 8505 Tel: +1 301 975 4871 E-mail: prasad.reddy@nist.gov J. E. Ladner, Biochemical Science Division, Bldg 227, Rm B244, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA Fax: +1 240 314 6225 Tel: +1 240 314 6206 E-mail: jane.ladner@nist.gov Present addresses *Division of Metrology for Quality Life, Korea Research Institute of Standards and Science, Daejeon, Republic of Korea Division of Molecular Biology, Department of Zoology, Sri Venkateswara University, Tirupati, Andhra Pradesh, India àBiology Department, Brookhaven National Laboratory, Upton, NY, USA (Received 9 May 2008, revised 25 July 2008, accepted 30 July 2008) doi:10.1111/j.1742-4658.2008.06621.x The Rv0948c gene from Mycobacterium tuberculosis H 37 R v encodes a 90 amino acid protein as the natural gene product with chorismate mutase (CM) activity. The protein, 90-MtCM, exhibits Michaelis–Menten kinetics with a k cat of 5.5 ± 0.2 s )1 and a K m of 1500 ± 100 lm at 37 °C and pH 7.5. The 2.0 A ˚ X-ray structure shows that 90-MtCM is an all a-helical homodimer (Protein Data Bank ID: 2QBV) with the topology of Escheri- chia coli CM (EcCM), and that both protomers contribute to each catalytic site. Superimposition onto the structure of EcCM and the sequence align- ment shows that the C-terminus helix 3 is shortened. The absence of two residues in the active site of 90-MtCM corresponding to Ser84 and Gln88 of EcCM appears to be one reason for the low k cat . Hence, 90-MtCM belongs to a subfamily of a-helical AroQ CMs termed AroQ d. The CM gene (y2828) from Yersinia pestis encodes a 186 amino acid protein with an N-terminal signal peptide that directs the protein to the periplasm. The mature protein, *YpCM, exhibits Michaelis–Menten kinetics with a k cat of 70±5s )1 and K m of 500 ± 50 lm at 37 °C and pH 7.5. The 2.1 A ˚ X-ray structure shows that *YpCM is an all a-helical protein, and functions as a homodimer, and that each protomer has an independent catalytic unit (Protein Data Bank ID: 2GBB). *YpCM belongs to the AroQ c class of CMs, and is similar to the secreted CM (Rv1885c, *MtCM) from M. tuber- culosis. Abbreviations *MtCM, secreted chorismate mutase from Mycobacterium tuberculosis; *YpCM, secreted chorismate mutase from Yersinia pestis; CM, chorismate mutase; EcCM, chorismate mutase domain of chorismate mutase–prephenate dehydratase from Escherichia coli; IPTG, isopropyl-thio-b- D-galactoside; MtCM, intracellular chorismate mutase from Mycobacterium tuberculosis; PfCM, chorismate mutase from Pyrococcus furiosus; ScCM, allosteric chorismate mutase from Saccharomyces cerevisiae; TSA, transition state analog; TtCM, chorismate mutase from Thermus thermophilus. 4824 FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works Chorismate mutase (CM) (EC 5.4.99.5), a shikimate pathway enzyme [1], catalyzes the pericyclic rearrange- ment of chorismate to prephenate [2]. Subsequent to this conversion, prephenate dehydrogenase and prephenate dehydratase catalyze the biosynthesis of tyrosine and phenylalanine, respectively. As this bio- synthetic pathway is absent in mammals but is essen- tial for the survival of bacteria and fungi, CM is often targeted for the development of inhibitors for micro- bial pathogens. This work was aimed at the character- ization of CM from Mycobacterium tuberculosis H 37 R v , a dreaded pathogen that claims two million human lives annually [3]. Annotation of the genome of M. tuberculosis H 37 R v revealed two genes for CM [4]. The Rv1885c gene encodes a secreted CM (*MtCM) and the Rv0948c gene encodes an intracellular CM (MtCM). Sasso et al. [5], Prakash et al. [6] and Kim et al. [7] have character- ized *MtCM. Kim et al. [7] have shown that *MtCM has in fact an extracellular destination in M. tuber- culosis. Prakash et al. [6] conducted a brief study of the recombinant MtCM. Our work is aimed at the fur- ther characterization of MtCM. The true primary sequence of MtCM is complicated by virtue of a number of presumptive in-frame initiator methionines preceded by a reasonable ribosome-binding sequence. The annotation Rv0948c for MtCM in a laboratory strain of M. tuberculosis H 37 R v would encode a 105 amino acid protein (105-MtCM) [4], whereas the anno- tation MT0975 for MtCM in the CDC1551 strain would encode a 217 amino acid protein (217-MtCM) [8]. Furthermore, alignment of 105-MtCM with the genetically engineered Escherichia coli CM (EcCM) [9] shows that the 105 amino acid protein has extra amino acids beyond the N-terminus of EcCM. Hence, we cloned 90-MtCM beginning with Met16 in Rv0948c. We determined the 3D structure of the 90-MtCM and kinetic parameters of all three proteins. The 90-MtCM is an AroQ class CM and the protein functions as a dimer. In this article, we also report on the cloning of the gene encoding the secreted CM from Yersinia pestis (*YpCM, y2828), purification of the protein, investigation of the properties of the enzyme, and the crystal structure analysis of the protein. Results and Discussion Annotation of CMs in M. tuberculosis H 37 R v The difference in annotation of MtCM arose from an in-frame initiator ATG codon in MT0975 (217-MtCM) and in Rv0948c (105-MtCM). Furthermore, the N-ter- minus of 105-MtCM has 22 more residues than the CM domain of E. coli prephenate dehydratase (Fig. 1). There is an in-frame methionine at position 16 of 105-MtCM and a purine-rich sequence analogous to the Shine–Dalgarno sequence about 10 nucleotides upstream of Met16. We reasoned that this Met16 could be the real initiator and consequently would pro- duce a 90 amino acid protein (90-MtCM). We charac- terized all three versions of the putative intracellular CM, i.e. 217-MtCM, 105-MtCM, and 90-MtCM. In a recent publication, Schneider et al. [10] observed simi- lar ambiguity about the translation start site(s) in the gene MSMEG5513 for an intracellular CM, a homolog of Rv0948c, from the annotation of the Mycobacte- rium smegmatis genome. They determined the transla- tion start site by translation fusion with the b-galactosidase gene, and showed that the first methio- nine in MSMEG5513 is the ‘real initiator’. Schneider et al. [10] did not determine the translation start site for Rv0948c. Production and purification of MtCM The 105-MtCM was overproduced as a fusion protein with the subtilisin prodomain (Fig. 2, lane 2). The fusion protein was completely soluble (Fig. 2, lane 3). Cleavage of 105-MtCM from the prodomain was trig- gered by fluoride-induced subtilisin activity (Fig. 2, lane 5). We observed three major protein products at this stage of purification: intact fusion protein (per- haps not very tightly bound), 105-MtCM, and an unidentified lower molecular mass protein. Hence, 105-MtCM was further purified by molecular sieve chromatography to near homogeneity (Fig. 2, lane 6). The molecular mass of 105-MtCM determined by MALDI-TOF MS was 11 771 Da (theoretical mass = 11 771 Da), and established that the protein is intact. Similarly, 90-MtCM was overproduced as a fusion protein with the subtilisin prodomain, and the protein was purified to homogeneity (Fig. 3, lane 5). As can be seen in lane 5 of Fig. 3, 90-MtCM migrated as a  6 kDa protein. Hence, we determined the molecular mass of 90-MtCM by MALDI-TOF MS as 10 090 ± 1 Da, which is identical to the theoretical mass of 10 090 Da. The yield of 105-MtCM and 90-MtCM was 1 mg per 1 L of culture. Activity mea- surements for CM using these two proteins showed that both proteins catalyze the conversion of choris- mate to prephenate (see kinetic measurements for k cat ). The 217-MtCM, purified from the subtilisin column, had 1 ⁄ 50th of the CM activity of 90-MtCM and 105- MtCM at the same stage of purification. Hence, we did not further purify or characterize 217-MtCM. We conclude that the annotation of the MT0975 gene was S. -K. Kim et al. AroQ chorismate mutases FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works 4825 misled by an upstream in-frame initiator methionine preceded by the Shine–Dalgarno sequence. Purification of *YpCM from the periplasmic fluid of E. coli *YpCM production was induced with 10 lm isopro- pyl-thio-b-d-galactoside (IPTG). Periplasmic proteins were isolated as described for *MtCM [7]. The peri- plasmic fluid was concentrated to  500 lL in a Milli- pore centrifugal tube with a 5000 Da molecular mass cutoff. *YpCM was purified on a 210 mL Biosep SEC-3000 HPLC column (Phenomenex, Torrance, CA, USA), equilibrated and eluted with 50 mm Tris ⁄ HCl (pH 7.5), 1 mm EDTA, and 100 mm NaCl. Fig. 1. Alignment of 90-MtCM with AroQ a CMs. The alignment begins with amino acid 13 in 90-MtCM (the numbering begins with amino acid 1 in the 90 amino acid protein). Amino acids 1–12 were not seen in the electron density map; their sequence is shown above the align- ment. In the structural alignment of MtCM, EcCM, PfCM, and TtCM by MATRAS [12], the top line indicates the average secondary structure (AVE_SECSTR); H, helical; T hydrogen-bonded turn; C, coil; S, bend. Capital letters indicate agreement for all structures. The active site resi- dues in EcCM are highlighted, and are shadowed when similar in the other sequences. C-terminal residues that were not visible in the struc- tures are shown as lower-case letters for MtCM, PfCM, and TtCM. At the top of the figure, the 15 N-terminal residues of the 105-MtCM construct are shown. 1 2 3 4 5 6 7 kDa 32.5 Subtilisin prodomain:105 aa MtCM fusion protein 25.0 16.5 105 aa MtCM monomer 6.5 Fig. 2. SDS ⁄ PAGE (16%) of the production and purification of 105-MtCM. Lane 1: uninduced cell-free extract of E. coli BL21(DE3) harboring the pG58–105-MtCM clone (25 lg of protein). Lane 2: induced cell-free extract of E. coli BL21(DE3) harboring the pG58– 105-MtCM clone (25 lg of protein). Lane 3: 48 000 g supernatant of induced cells (same volume as used in lane 2). Lane 4: flow through from subtilisin column (same volume as used in lane 2). Lane 5: 10 lg of protein(s) eluted after equilibration with 100 m M sodium fluoride. Lane 6: 5 lg of purified 105-MtCM from a Sepha- dex G-75 column. Lane 7: molecular mass markers. 1 2 3 4 5 6 kDa 32.5 25.0 16.5 Subtilisin prodomain: 90 aa MtCM fusion protein 6.5 90 aa MtCM monomer Fig. 3. SDS ⁄ PAGE (16%) of the production and purification of 90-MtCM. Lane 1: induced cell-free extract of E. coli BL21(DE3) harboring the pG58–90-MtCM clone (25 lg of protein). Lane 2: 48 000 g supernatant of induced cells (same volume as used in lane 1). Lane 3: flow through from subtilisin column (same volume as used in lane 1). Lane 4: 10 lg of protein(s) eluted after equilibra- tion with 100 m M sodium fluoride. Lane 5: 13 lg of purified 90-MtCM from a Sephadex G-75 column. Lane 6: molecular mass markers. AroQ chorismate mutases S. -K. Kim et al. 4826 FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works Kinetic measurements Assays for CM were performed with 90-MtCM and 105-MtCM at chorismate concentrations of 100 lm to 4mm. Both MtCMs catalyzed the conversion of chorismate to prephenate with a k cat of 5.5 ± 0.2 s )1 and a K m of 1500 ± 100 lm at 37 °C and pH 7.5 (Table 1). The k cat for MtCM is about 14-fold lower than that reported for EcCM (72 s )1 ) [11]. The K m of 1500 lm chorismate for MtCMs is five times higher than that observed for EcCM. One obvious reason for the low k cat and high K m exhibited by MtCM is the absence of two of the substrate-binding residues found in the C-terminus of EcCM (Fig. 1). In contrast, *YpCM, in which all the catalytic site residues are pre- served, exhibits a high k cat of 70 ± 5 s )1 , similar to that for EcCM. Crystal structure of 90-MtCM The crystal structure of 90-MtCM shows clearly that the molecule is an all a-helical homodimer (Protein Data Table 1. Comparison of the catalytic properties of MtCM, EcCM, and *YpCM. MtCM proteins and *YpCM were purified as described in Experimental procedures. One microgram of MtCM or 200 ng of *YpCM in 10 lL was used in each assay of 0.3 mL of buffer. The buffer consisted of 50 m M Tris ⁄ HCl (pH 7.5), 0.5 mM EDTA, 0.1 mg BSAÆmL )1 , and 10 mM b-mercaptoethanol. Choris- mate concentrations were varied from 0.25 to 4 m M. Assays were performed at 37 °C for 5 min [34], and the reaction was stopped with 0.3 mL of 1 M HCl. After further incubation for 10 min at 37 °C to convert prephenate to phenylpyruvate, 0.6 mL of 2.5 M NaOH was added. The absorbance of the phenylpyruvate chromo- phore was read at 320 nm. Blanks without the enzyme were main- tained for each of the chorismate concentrations to account for the nonenzymatic conversion of chorismate to prephenate. Results are the average of three experiments. The kinetic data for EcCM were from the literature [11], measured at 30 °C. Enzyme k cat (s )1 ) K m (lM) 90-MtCM 5.5 ± 0.2 1500 ± 100 105-MtCM 5.5 ± 0.2 1500 ± 100 217-MtCM 0.1 Not determined EcCM 72 296 ± 19 *YpCM 70 ± 5 500 ± 50 A B C Fig. 4. Crystal structure of 90-MtCM. (A) The homodimer is shown as a stereo cartoon on the top with one polypeptide chain in blue and the other in green. (B) The superimposition of a single chain of TtCM, PfCM and EcCM onto 90-MtCM is shown, with the TSA from EcCM marking the active site. The approximate positions of the N-termini and C-termini are labeled in the same color as the polypeptide chain. The three helices are labeled H1, H2, and H3. (C) Helix 3 from each of the four structures is shown. The helices were taken from the superimposed structures and then separated by translating each horizontally in the plane of the page. The figures were drawn with PYMOL (http://www.pymol.org). S. -K. Kim et al. AroQ chorismate mutases FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works 4827 Bank ID: 2QBV; Fig. 4). The polypeptide chain has one long 36 residue helix (helix 1), an eight residue loop, an 11 residue helix (helix 2), a two residue loop, and a 15 residue helix (helix 3). The buried surface area of the dimer is 3810 A ˚ 2 . This crystal form has one protomer in the asymmetric unit; the complete molecule is generated by a crystallographic two-fold. The data and refinement statistics are shown in Table 2. The Ramachandran plot has 96.9% of the residues in the most favored region and 3.1% in the additional allowed region. Five residues were modeled with alternative conformations. No inter- pretable density was observed for the first 12 residues or for the last five residues. When the model is numbered according to the 90 residue protein, residues 13–85 are seen in the electron density map. Using matras [12] to compare the structure with a representative library of structures, three structures stood out as very similar; these are Protein Data Bank IDs 2D8D, 1YBZ and 1ECM. 2D8D and 1YBZ are annotated as CMs from Thermus thermophilus (TtCM) and Pyrococcus furiosus (PfCM), respectively, from Structural Genomics pro- jects on these organisms. 1ECM [13] is a genetically engineered 109 amino acid CM domain from E. coli. The dimer of 90-MtCM is shown in Fig. 4A, and the superimposition of the four structures for a single poly- peptide chain is shown in Fig. 4B. It is apparent from the structural alignment that helix 3 of 90-MtCM is a shorter version of helix 3 of EcCM, TtCM and PfCM lacking two of the binding site residues (Fig. 4C) corre- sponding to Ser84 and Gln88 of EcCM as discussed below. During the preparation of this article, we were made aware of another deposited Protein Data Bank file for 90-MtCM, 2VKL (M. Okvist, K. Roderer, S. Sasso, P. Kast, and U. Krengel, unpublished data). In the structure of 90-MtCM in 2VKL, there is a malate ion from the buffer in the active site of the enzyme. Malate mimics endo-oxabicyclic dicarboxylic acid, the transition state analog (TSA), in much the same fashion as citrate that we observed in our *YpCM structure (Protein Data Bank ID: 2GBB). We crystallized 90-MtCM in the pres- ence of citrate but did not see citrate in the active site. However, we studied the effect of citrate on 90-MtCM activity, and found citrate to have some kind of inhibi- tory effect from preliminary results (data not shown). The inhibition is not a salt effect, because sodium chlo- ride and sodium acetate had no effect on the activity. The rmsd on C-alphas between the two 90-MtCM structures is 0.69 A ˚ . One difference between the struc- tures is that five residues at the C-terminal end are dis- ordered in our structure (2QBV) and only one residue is disordered in 2VKL. In fact, although the space group is the same for both structures, the c-axis is 10 A ˚ shorter in 2VKL. This difference is due to crystal packing, which allows the C-terminal residues of neighboring molecules (not in the same dimer) to inter- act and the tail of one protomer to almost reach the active site of the neighbor. Active site of 90-MtCM The structure of EcCM includes the TSA, which clearly identifies the active site. From structural and sequence homology with EcCM, the active site residues of 90-MtCM can be identified, and are shown in Fig. 5. The striking difference between EcCM and 90- MtCM is that the EcCM residues Ser84 and Gln88 are absent in 90-MtCM (Fig. 1). Structurally, Ser84 of EcCM lines up with Gly84 of 90-MtCM. The final five residues, GRLGH, of 90-MtCM are not seen in the electron density map. However, none of these residues is a candidate for performing the role of Gln88 in EcCM. Of the other two structures, PfCM has the conserved Ser70 and Gln74, and TtCM has Ser81 and Glu85. There are two Protein Data Bank files for TtCM; in the file 2D8D, Glu85 is only seen in one of the two chains in the structure, which has a dimer in the asymmetric unit, and in the file 2D8E, which has one chain in the asymmetric unit, all of the C-terminal residues are seen. Another difference is the loop orien- Table 2. Diffraction data and refinement statistics showing over- all ⁄ high-resolution shell (2.18–2.10 A ˚ ) values where appropriate. 90-MtCM *YpCM SeMet Diffraction data Space group P4 3 2 1 2 C222 1 Cell dimensions (a, b, c)(A ˚ ) 59.9, 59.9, 47.5 89.0, 144.1, 116.6 Resolution limit (A ˚ ) 2.0 2.1 No. of measured intensities 91 040 566 550 No. of unique reflections 6192 ⁄ 815 43 510 ⁄ 4106 Mean redundancy 14.7 ⁄ 14.8 13.0 ⁄ 8.0 R merge 0.056 ⁄ 0.323 0.113 ⁄ 0.469 0% Completeness 100.0 ⁄ 100.0 99.4 ⁄ 95.3 Average I ⁄ r 25.3 ⁄ 7.0 12.0 ⁄ 2.0 Mosaicity 1.37 0.77 Radiation wavelength 1.54 0.979 Refinement Resolution limits (A ˚ ) 20.0–2.0 20.0–2.1 R-factor (95% of the data) 0.219 0.207 R free (5% of the data) 0.298 0.258 No. of water molecules 47 193 Bond length rmsd (A ˚ ) 0.021 0.019 Bond angle rmsd (°) 1.98 1.86 Average B (main chain ⁄ side chain) (A ˚ 2 ) 42.2 ⁄ 44.0 34.3 ⁄ 36.5 Average B for water (A ˚ 2 ) 41.5 34.5 AroQ chorismate mutases S. -K. Kim et al. 4828 FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works tation between the first long helix and the second helix. Figure 4B shows that for the EcCM, TtCM and PfCM structures, this loop aligns very well, but that the 90-MtCM structure is significantly altered; however, examination of the surface (not shown) demonstrates that even with this change, the active site remains bur- ied. We superimposed the EcCM structure with the TSA onto 90-MtCM to see the TSA in the active site of 90-MtCM (Figs 5 and 6). As the residues corre- sponding to Ser84 and Gln88 of EcCM are absent in 90-MtCM, chorismate is unlikely to be as well stabi- lized in its active site. This at least partly explains the low k cat for 90-MtCM (Table 1). In an attempt to substantiate the notion that the lower k cat and higher K m are due to the missing sub- strate-binding residues, we engineered a modified version of helix 3 in 90-MtCM. We replaced the C-ter- minal seven amino acids GRGRLGH in 90-MtCM with amino acids SVLTQQALL or SVLTEQALL, cor- responding to the C-terminus of EcCM, thus providing Ser84 and Gln88 ⁄ Glu88 in corresponding positions in 90-MtCM. Production of glutamine and glutamic acid H 2 N H 2 N NH 2 NH Arg18′ (Arg127) (Lys54) (Asp63) (Gln66) (Arg43) Arg58 Arg35 (Ala99) Ser 84 (Gln 103) (Glu67) Gln 88 Residues in EcCM missing in 90-MtCM Arg46 Val55 Glu59 NH O HN OH HN N H H N H H H O O O + - - - + + + O O HO O O O O NH 2 NH 2 H 2 N H 2 N H 2 N Fig. 6. The active site of 90-MtCM: a diagrammatic view of the active site of 90-MtCM is shown, with the superimposition of the TSA from the EcCM structure. The corresponding residue numbers for *YpCM are shown in parentheses. A B C Fig. 5. Stereo view of the active sites of EcCM, 90-MtCM, and *YpCM: a stereo view of the active site of EcCM is shown in (A), and the corresponding view of 90-MtCM is shown in (B). The active site residues are shown in stick form, and the rest of the structure is in cartoon form. In EcCM, one polypeptide chain is gray and the other is rose. In 90-MtCM, one polypeptide chain is blue and the other is green. The TSA from the EcCM structure is shown with yellow carbon atoms in the 90-MtCM structure for orientation. The active site residues are labeled, and the N-terminus of the chain that contributes one residue (R11¢ in EcCM and R18¢ in 90- MtCM) to the active site is labeled. In 90-MtCM, the observed C-terminus of the structure visible in the electron density is indi- cated for the second chain. The citrate from the crystal structure of *YpCM is shown in the active site with yellow carbon atoms in (C). All of the active site residues belong to the same chain. The figure was drawn with PYMOL (http://www.pymol.org). S. -K. Kim et al. AroQ chorismate mutases FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works 4829 variants of the 90-MtCM clones resulted in inclusion bodies of the overproduced protein(s) under various conditions of growth and induction. Thus, we could not experimentally verify our interpretation of the lower k cat and higher K m. In an analysis of active site residues in EcCM by site-directed mutagenesis, Liu et al. [11] observed lower activity for the Q88A mutant. They proposed that the side chain of Gln88 in EcCM hydrogen bonds with O7 of the transition state analog, endo-oxabicyclic dicarboxylic acid (Fig. 6). This experi- mental evidence reinforces the low k cat that we observed for 90-MtCM, which has leucine instead of glutamine in the corresponding position. Crystal structure of *YpCM The crystal structure of the secreted, mature, dimeric *YpCM with a citrate ion in the active site has been determined to 2.1 A ˚ resolution, using data collected at a single wavelength for the selenomethionine derivative of the protein. The protein crystallized in the space group C222 1 with two homodimers (A ⁄ B and C ⁄ D) in the asymmetric unit. The protomers superimpose with average rmsd values in the Ca coordinates of less than 0.8 A ˚ . The final model for *YpCM includes all 155 residues for chains A and C and 154 residues for chains B and D, where the initial residue, Gln31, is not ordered. The model also includes four citrate ions, one in each active site, 13 sulfate ions with 11 modeled at 0.5 occupancy, and 193 water molecules. [Correction added on 28 August 2008 after first online publication: in the preceding sentence, ‘13 sulfate ion, with 11’ was corrected to ‘13 sulfate ions with 11’]. In the Ramachandran plots, 95.1% of the residues are in the most favored regions, 4.5% in the additional allowed regions, and 0.4% in the generously allowed regions. The structure is all a-helical, and the protomer has the fold of the EcCM dimer with an inserted loop connecting the two chains. Each protomer of *YpCM has one active site, and the molecule forms a homo- dimer. In this crystal form, citrate ions from the crys- tallization solution are present in all the active sites. This is the same fold as for *MtCM [7,14,15]. The superimposition of *MtCM on *YpCM aligns 132 resi- dues and yields an rmsd for Ca atoms of 1.8 A ˚ for both Protein Data Bank files 2F6L and 2FP2. The dimer is also formed in the same manner as that of *MtCM. There is only 23% sequence identity over the aligned residues. As in *MtCM, the active site has residues mainly from the N-terminal half of the chain, and the region that would correspond to a second active site by analogy to the EcCM dimer is closed off by a disulfide bond. In *MtCM, the disulfide bond between Cys160 and Cys193 links helices that correspond to helix 2 and helix 3 of the second EcCM protomer; in *YpCM, the disulfide bond is between the third residue of the mature protomer and the bottom of helix 1 of the second EcCM protomer, Cys33 and Cys148. Classification of MtCM A diverse array of CMs occur in nature: AroQ class CMs such as EcCM [13], CM from Methanococcus jann- aschii [16], and allosteric CM from Saccharomyces cere- visiae (ScCM) [17]; *AroQ class CMs such as Erwinia herbicola CM (*EhCM) [16], *MtCM [5–7], and *YpCM; and AroH class CMs such as Bacillus subtilis CM (BsCM) [18]. [Correction added on 28 August 2008 after first online publication: in the preceding sentence, ‘*Erwinia herbicola (*EhCM) [16], *MtCM [5-7], and *YpCM; and AroH class CMs such as Bacillus subtilis (BsCM) [18]’ was corrected to ‘Erwinia herbicola CM (*EhCM) [16], *MtCM [5-7], and *YpCM; and AroH class CMs such as Bacillus subtilis CM (BsCM) [18]’]. AroQ class and *AroQ class CMs function as dimers, whereas AroH class CMs function as trimers. In addi- tion, ScCM has a domain for regulation of the activity by tryptophan and tyrosine [19], whereas *MtCM [7,14] and *YpCM do not have such a regulatory domain. Furthermore, structural motifs differ among the AroQ and AroH classes of CMs. AroQ and *AroQ CMs exhi- bit all a-helical bundles, whereas AroH CMs contain both a-helices and b-sheets. The active site in EcCM is formed by residues from all three helices of one pro- tomer and by a residue from the N-terminal long helix of the second protomer. In contrast, the active site in ScCM [17], *MtCM [7,14,15] and *YpCM is formed within a single protomer. Further subclassification of AroQ CMs on the basis of their distinct structural prototypes was proposed by Okvist et al. [14] (Fig. 7). EcCM-like proteins whose catalytic site is formed with residues from both protomers are denoted as AroQ a . ScCM-like proteins in which the catalytic site is formed within a single protomer with a domain for regulation of activity by tryptophan and tyrosine are denoted as AroQ b . Secreted CMs such as *MtCM and *YpCM, in which the catalytic site is formed within a single protomer but without an apparent regulatory domain, are denoted as AroQ c . A fourth subclass of CMs denoted AroQ d was proposed by Okvist et al. [14], on the basis of the primary sequence alone. The structural motif of AroQ d CMs resembles that of AroQ a , with the notable difference of a shortened third helix that lacks two substrate-binding site residues. Here we describe the first 3D structure of such a protein, MtCM. AroQ chorismate mutases S. -K. Kim et al. 4830 FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works Experimental procedures Materials All the reagents used in this work were obtained from the specified sources [7]. A selenomethionine auxotroph of E. coli strain B834(DE3) was obtained from EMD Biosciences Inc. (Madison, WI, USA). The M9 salts growth medium (Cat. No. MD045003) for the incorporation of selenomethionine into *YpCM was purchased from Medici- lon Inc. (Chicago, IL, USA). E. coli strains and plasmids E. coli strains NovaBlue and BL21(DE3) were used for cloning the target gene and expression of the cloned gene, respectively. The plasmid vector pG58 and subtilisin column were kindly provided by P. N. Bryan (Center for Advanced Research in Biotechnology, University of Mary- land Biotechnology Institute). Engineering of the fusion protein production vector pG58 was described by Ruan et al. [20]. Briefly, pG58 was designed to produce a target gene product as a fusion protein with the subtilisin prodo- main. The fusion protein would be bound to a resin cou- pled with a stable variant of subtilisin protease. Next, equilibration with fluoride anion will trigger the cleavage by subtilisin between the prodomain and the target protein, thus releasing the target protein in its native form, begin- ning with the initiator methionine. Cloning of Rv0948c and MT0975 genes The Rv0948c ORF for the 105 amino acid protein was ampli- fied by PCR from M. tuberculosis H 37 R v genomic DNA. Oligonucleotide pair 1 with specific restriction recognition sequences for cloning into pG58 was: 5 ¢-GCTACG TTTAAAGCGATGATGAGACCAGAACCCCCACATCA CG-3¢ (forward primer with DraI site underlined) and 5¢-CG GAATTCTTAGTGACCGAGGCGGCCCCTGCC-3¢ (reverse primer with EcoRI site underlined). Similarly, the Rv0948c ORF for the 90 amino acid protein beginning with Met16 was amplified with oligonucleotide pair 2: 5¢-GCTACG TTTAAAGCGATGATGAACCTGGAAATG CTCGAGTCC-3¢ (forward primer with DraI site underlined) and the same reverse primer. Oligonucleotide pair 3 for amplification of MT0975 (217 amino acid protein – another annotation for MtCM) for cloning into pG58 was: 5¢-GCTACG TTTAAAGCGATGATGGACCGGGAGGCT TGGCG-3¢ (forward primer with DraI site underlined) and the same reverse primer as above. Amplification conditions with all sets of primers were: 95 ° C for 5 min for initial melting of DNA, followed by 30 cycles of amplification, with each cycle consisting of melting at 95 °C for 60 s, annealing at 50 ° C for 60 s, and polymerization at 72 °C for 60 s. Polymerization was continued at the end for 10 min at A B C D Fig. 7. Four subclasses of AroQ CMs: the four subclasses of AroQ CMs are shown with cartoon drawings of representative structures. All AroQ CMs are homodimers. One chain is blue and the other chain is green. The distinguishing features are emphasized by indi- cating the active sites with red circles. The regulatory sites of the AroQ b class are highlighted with red squares. The shortened third helices of AroQ d are pointed to with red arrows. (A) AroQ a is EcCM. (B) AroQ b is ScCM. (C) AroQ c is *YpCM. (D) AroQ d is 90-MtCM with the TSA from the superimposition of EcCM. S. -K. Kim et al. AroQ chorismate mutases FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works 4831 72 °C. One hundred nanograms of M. tuberculosis H 37 R v genomic DNA (generously provided by J. Belisle and P. Brennan, Colorado State University) and 100 ng of prim- ers were used in the amplification. The amplified DNA obtained with oligonucleotide pairs 1, 2 and 3 was digested with Dra I and EcoRI for cloning into the respective sites of the pG58 plasmid. A recombinant was isolated from E. coli Novablue and introduced into E. coli BL21(DE3) for protein production. Overproduction of the proteins E. coli BL21(DE3) harboring either pG58–Rv0948c (105 amino acids), pG58–Rv0948c (90 amino acids) or pG58–MT0975 (217 amino acids) recombinant plasmid was grown in 25 mL of LB medium containing ampicillin (100 lgÆmL )1 )at37°CtoanA 600 nm  0.5. Protein pro- duction was induced with 30 lm IPTG overnight at 24 °C, except for the pG58–MT0975, clone which was induced overnight at 15 °C to eliminate the formation of inclusion bodies of the protein. All three fusion proteins were pro- duced in fully soluble form under these conditions. Purification of native MtCM Cells from 1 L of induced culture of BL21(DE3) harboring pG58–Rv0948c (encoding either the 105 amino acid protein or the 90 amino acid protein) were suspended in 40 mL of lysis buffer (10 mm potassium phosphate, pH 7.4, 15 mm NaCl), to which a tablet of protease inhibitor cocktail was added. The cell suspension was passed through a French press twice at 10 000 lbÆin )2 , and the extract was centrifuged at 48 000 g for 1 h. Supernatant containing the subtilisin prodomain–MtCM fusion protein was loaded onto a 5 mL subtilisin column at a flow rate of 0.5 mLÆmin )1 . The resin was washed with 60 mL of the lysis buffer at a flow rate of 1mLÆmin )1 . The resin was further washed with 50 mL of 1 m sodium acetate in the lysis buffer. Next, the cleavage of MtCM from the prodomain was triggered by flushing the resin at a flow rate of 1 mLÆmin )1 with 20 mL of 100 mm sodium fluoride in the lysis buffer and equilibration for 30 min. The resin was washed with 25 mL aliquots of 100 mm sodium fluoride in the lysis buffer. Effluent fractions containing CM, as judged by SDS ⁄ PAGE and by activity, were pooled and concentrated to 5 mL in an Amicon cell using a 5000 Da molecular mass cut-off membrane. MtCM (105 amino acids ⁄ 90 amino acids) was further purified by molecular sieve chromatography on a 480 mL Sephadex G-75 superfine column, which was equilibrated and eluted with 50 mm Tris ⁄ HCl (pH 7.5), 1 mm EDTA, and 100 mm NaCl. Effluent fractions containing pure MtCM (105 amino acids ⁄ 90 amino acids) were concentrated for protein determi- nation and biochemical analysis. The 217-MtCM was simi- larly purified with the subtilisin column. Further purification was not pursued, as it exhibited extremely low CM activity. Cloning and expression of the *YpCM gene (y2828) in E. coli The gene y2828 from the genome sequence of Y. pestis strain Kim10+ [21] was annotated as CM. [Correction added on 28 August 2008 after first online publication: in the preceding sentence, ‘The gene y2828 from the genome sequence of Y. pestis strain Kim10+ (21) as CM’ was corrected to ‘The gene y2828 from the genome sequence of Y. pestis strain Kim10+ (21) was annotated as CM’]. The full-length *YpCM gene coding sequence, including the signal peptide, was amplified by PCR using the forward primer 5¢-GG AATTC CATATGCAACCCACTCATACGCTAACAAG-3¢ (with the NdeI restriction recognition sequence underlined) and the reverse primer 5¢-CG GGATCCTTATTTTAATT TTACCTGATTGAAGGTTGAG-3¢ (with the BamHI restriction recognition sequence underlined). Amplification conditions were: 95 °C for 60 s for initial melting of DNA, followed by 30 cycles of amplification, with each cycle con- sisting of melting at 95 °C for 60 s, annealing at 60 °C for 60 s, and polymerization at 72 °C for 60 s. Polymerization was continued at the end for 10 min at 72 °C. Two hundred nanograms of Y. pestis strain KIM10+ chromosomal DNA (kindly provided by R. D. Perry, University of Kentucky) and 100 ng of primers were used in the amplification. The amplified DNA was digested with NdeI and BamHI and cloned into the respective sites of the pET15b plasmid. A recombinant was isolated from E. coli Novablue and intro- duced into BL21(DE3) for protein production. E. coli BL21(DE3) harboring the pET15b–y2828 recombinant plasmid was grown in 100 mL of LB medium containing ampicillin (100 lgÆmL )1 )at37°CtoanA 600 nm  0.6. Protein production was induced with 10 lm IPTG overnight at 15 °C. *YpCM was purified by molecular sieve chroma- tography from the periplasmic fluid of E. coli as described for *MtCM [7]. The production and purification of *YpCM was scaled up for crystallization. Production and purification of selenomethionine *YpCM E. coli B834(DE3), a methionine auxotroph, was trans- formed with the pET15b–y2828 recombinant plasmid. Incor- poration of selenomethionine into *YpCM was performed using the M9 salts ⁄ selenomethionine growth medium, according to the manufacturer’s recommendation. Briefly, cells were grown in 1 L of LB medium containing ampicillin (100 lgÆmL )1 ) overnight at 37 °C. Cells were harvested, washed twice with sterile water, and suspended in 100 mL of M9 salts medium. Four 1 L volumes of M9 salts media con- taining ampicillin were inoculated with 25 mL of the culture per 1 L. Cells were grown at 37 °CtoA 600 nm = 0.4. At this stage, selenomethionine was added and induced with 10 lm IPTG at 15 °C overnight. *YpCM was purified by molecular sieve chromatography from the periplasmic fluid. AroQ chorismate mutases S. -K. Kim et al. 4832 FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works Crystallization of 90-MtCM The 90-MtCM was concentrated to 8.3 mgÆmL )1 in 50 mm Tris ⁄ HCl (pH 7.5), 1 mm EDTA, and 100 m m sodium chlo- ride. Crystallization conditions were surveyed by the sitting drop vapor diffusion method using Emerald BioSystems Wizard Screens I and II. There were several hits. The crystal used for data collection was grown with a well solution of 0.1 m Tris ⁄ HCl (pH 8.6), 0.2 m magnesium chloride, and 20% poly(ethylene glycol) 400. The crystallization drops were made with equal volumes of protein and well solution. Crystallization of *YpCM Crystallization conditions were surveyed by the hanging drop vapor diffusion method using the Wizard II kit from Emerald BioSystems (http://www.emeraldbiosystems.com). The protein concentration was 10 mgÆmL )1 in 50 mm Tris ⁄ HCl (pH 7.5), 1 mm EDTA, 1 mm dithiothreitol, and 200 mm sodium chloride. The original hit was with solu- tion 9 (2 m ammonium sulfate, 0.1 m citrate ⁄ phosphate buffer, pH 4.2). For the refined conditions, a well solution of 1.5–1.6 m ammonium sulfate and 0.1 m citrate ⁄ phos- phate buffer (pH 4.2) was used, and the protein concentra- tion was reduced to 5 mgÆmL )1 . For the selenomethionine protein, the well solution was 1.8–2.0 m ammonium sulfate and 0.1 m citrate ⁄ phosphate buffer (pH 4.2), with a protein concentration of 2.5–5 mgÆmL )1 . Data collection for 90-MtCM Diffraction data were collected using a home source Riga- ku 007 generator and a RAXIS IV ++ image plate detector (Rigaku ⁄ MSC, The Woodlands, TX, USA). The crystal was cooled to 105 K with a Rigaku Xtream 2000 cryocool- er. For cryo-data collection, the crystals were mounted through a layer of paraffin oil placed on top of the crystal- lization drop. The data were collected and processed with crystalclear [22], and the statistics are shown in Table 2. Structure determination for 90-MtCM The structure of 90-MtCM was solved by molecular replacement using phaser [23], with the structure of PfCM (Protein Data Bank ID: 1YBZ). The asymmetric unit of the P4 3 2 1 2 crystal includes a single chain of 90-MtCM. Molecu- lar replacement trials using a single protomer failed. How- ever, when the symmetry was lowered to P4 3 and the dimer was used as the search model, a solution was found. The remainder of the structure determination was carried out in the space group P4 3 2 1 2. refmac5 [24,25] was used to refine the model, and resolve [26] was used to iteratively rebuild the model to remove bias. The final refinement statistics are shown in Table 2. coot [27] was used to view the model graphically and to build portions not built by resolve. The stereochemistry was checked with procheck [28] and with routines inside coot. Data collection for *YpCM Preliminary data were collected on the home source described above, and cryoprotection was accomplished in the same manner as for 90-MtCM. The selenomethionine data for *YpCM were collected on beamline X29A of the National Synchrotron Light Source at wavelength 0.9790 A ˚ with the crystal cooled to 100 K. The statistics are shown in Table 2. *YpCM structure determination The structure of *YpCM was solved using the phasing information from the anomalous data. The positions of the selenium atoms were located with shelxd [29], and the initial phases were calculated with solve [30]. Two dimers in the asymmetric unit cell gives a Matthews coeffi- cient of 2.6 and a solvent content of 52.5%. The initial model was built with resolve [26,31], using iterative rounds of pattern-matching, fragment identification, den- sity modification, and refinement. This model included 78% of the residues and placed 71% of the side chains. The noncrystallographic symmetry was used to combine the four partial chains to produce a more complete model. Then, further cycles of model building and refinement were performed using xtalview [32] and refmac5 [24]. The final refinement statistics are shown in Table 2. The stereochemistry was checked with procheck [28] and with molprobity [33]. Four residues in the A chain and one in the D chain were modeled with alternative side-chain con- formations. No interpretable electron density was observed for the first residues (residue 31) of chains B and D. The residues between Cys148 and Asp155 are somewhat disor- dered, particularly in the C and D chains, and conse- quently have increased B-values. Other methods CM was assayed by the method of Davidson & Hudson [34], essentially as described in our previous study [7]. One microgram of MtCM protein or 200 ng of *YpCM protein were used in the assay. Protein concentration was deter- mined by the Micro BCA method with BSA as the stan- dard (Pierce, Rockford, IL, USA). The monomeric molecular mass of the native MtCM was determined by MALDI-TOF MS. Mass spectra were collected and analyzed using an Applied Biosystems Voyager-DE STR Biospectrometry Workstation (Foster City, CA, USA). The DNA sequence of the cloned genes was confirmed by the dideoxy sequencing method [35], as adopted for the Applied Biosystems model 3130 Genetic Analyzer. S. -K. Kim et al. AroQ chorismate mutases FEBS Journal 275 (2008) 4824–4835 Journal compilation ª 2008 FEBS. No claim to original US government works 4833 [...]... laboratory strains J Bacteriol 184, 5479–5490 Stewart J, Wilson DB & Ganem B (1990) A genetically engineered monofunctional chorismate mutase J Am Chem Soc 112, 4582–4584 Schneider CZ, Parish T, Basso LA & Santos DS (2008) The two chorismate mutases from both Mycobacterium tuberculosis and Mycobacterium smegmatis: biochemical analysis and limited regulation of promoter activity by aromatic amino acids... P, Aruna B, Sardesai AA & Hasnain SE (2005) Purified recombinant hypothetical protein coded by open reading frame Rv1885c of Mycobacterium tuberculosis exhibits a monofunctional AroQ class of periplasmic chorismate mutase activity J Biol Chem 280, 19641–19648 7 Kim S-K, Reddy SK, Nelson BC, Vasquez GB, Davis A, Howard AJ, Patterson S, Gilliland GL, Ladner JE & Reddy PT (2006) Biochemical and structural. .. Escherichia coli chorismate mutase J Am Chem Soc 117, 3627–3628 Okvist M, Dey R, Sasso S, Grahn S, Kast P & ˚ Krengel U (2006) 1.6 A Crystal structure of the secreted chorismate mutase from Mycobacterium tuberculosis: novel fold topology revealed J Mol Biol 357, 1483–1499 Qamra R, Prakash P, Aruna B, Hasnain SE & Mande ˚ SC (2006) The 2.15 A crystal structure of Mycobacterium tuberculosis chorismate mutase. .. 2.2 -A resolution Proc Natl Acad Sci USA 91, 10814–10818 Chook YM, Ke H & Lipscomb WN (1993) Crystal structure of the monofunctional chorismate mutase from Bacillus subtilis and its complex with a transition state analog Proc Natl Acad Sci USA 90, 8600–8603 Schnappauf G, Strater N, Lipscomb WN & Braus GH (1997) A glutamate residue in the catalytic center of the yeast chorismate mutase restricts enzyme activity... reveals an unexpected gene duplication and suggests a role in host–pathogen interactions Biochemistry 45, 6997–7005 MacBeath G, Kast P & Hilvert D (1998) A small, thermostable, and monofunctional chorismate mutase from the archaeon Methanococcus jannaschii Biochemistry 37, 10062–10073 Xue Y, Lipscomb WN, Graf R, Schnappauf G & Braus G (1994) The crystal structure of allosteric chorismate ˚ mutase at...AroQ chorismate mutases S -K Kim et al Disclaimer Certain commercial equipment or materials are identified in this article in order to specify adequately the experimental procedure Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, and nor does it imply that the materials or equipment identified are necessarily the best available... characterization of the secreted chorismate mutase (Rv1885c) from Mycobacterium tuberculosis H37Rv: an *AroQ enzyme not regulated by the aromatic amino acids J Bacteriol 188, 8638–8648 8 Fleischmann RD, Alland D, Eisen JA, Carpenter L, White O, Peterson J, DeBoy R, Dodson R, Gwinn M, 4834 14 15 16 17 18 19 20 21 Haft D et al (2002) Whole-genome comparison of Mycobacterium tuberculosis clinical and laboratory... R, Parkhill J, Garnier T, Churcher C, Harris D, Gordon SV, Eiglmeier K, Gas S, Barry CE III et al (1998) Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence Nature 393, 537–544 5 Sasso S, Ramakrishnan C, Gamper M, Hilvert D & Kast P (2005) Characterization of the secreted chorismate mutase from the pathogen Mycobacterium tuberculosis FEBS J 272, 375–389 6 Prakash... Bacteriol 190, 122–134 Liu DR, Cload ST, Pastor RM & Schultz PG (1996) Analysis of active site residues in Escherichia coli chorismate mutase by site-directed mutagenesis J Am Chem Soc 118, 1789–1790 Kawabata T (2003) MATRAS: a program for protein 3D structure comparison Nucleic Acids Res 31, 3367– 3369 Lee AY, Karplus PA, Ganem B & Clardy J (1995) Atomic structure of the buried catalytic pocket of. .. best available for the purpose 9 10 Acknowledgements 11 We thank John Belisle and Patrick Brennan, Colorado State University, for generously providing Mycobacterium tuberculosis H37Rv DNA We also thank Robert Perry, University of Kentucky, for generously providing Yersinia pestis Kim10+ DNA We are grateful to the anonymous reviewers of this paper for their constructive critique and valuable suggestions . A comparative biochemical and structural analysis of the intracellular chorismate mutase (Rv0948c) from Mycobacterium tuberculosis H 37 R v and the secreted chorismate. prephenate dehydrogenase and prephenate dehydratase catalyze the biosynthesis of tyrosine and phenylalanine, respectively. As this bio- synthetic pathway is absent