Expression, Purification, and Characterization of Inosine 5′-Monophosphate Dehydrogenase from Borrelia burgdorferi *

Inosine 5′-monophosphate dehydrogenase (IMPDH) is the rate-limiting enzyme in de novo guanine nucleotide biosynthesis. IMPDH converts IMP to xanthosine 5′-monophosphate with concomitant conversion of NAD+ to NADH. All IMPDHs characterized to date contain a 130-residue “subdomain” that extends from an N-terminal loop of the α/β barrel domain. The role of this subdomain is unknown. An IMPDH homolog has been cloned fromBorrelia burgdorferi, the causative agent of Lyme disease (Margolis, N., Hogan, D., Tilly, K., and Rosa, P. A. (1994)J. Bacteriol. 176, 6427–6432). This homolog has replaced the subdomain with a 50-residue segment of unrelated sequence. We have expressed and characterized the B. burgdorferi IMPDH homolog. This protein has IMPDH activity, which unequivocally demonstrates that the subdomain is not required for catalytic activity. The monovalent cation and dinucleotide binding sites of B. burgdorferi IMPDH are significantly different from those of human IMPDH. Therefore, these sites are targets for the design of specific inhibitors for B. burgdorferi IMPDH. Such inhibitors might be new treatments for Lyme disease.

Inosine 5Ј-monophosphate dehydrogenase catalyzes the conversion of IMP to XMP 1 with the concomitant reduction of NAD to NADH. This reaction is the rate-limiting step in guanine nucleotide biosynthesis, and is therefore a target for numerous chemotherapeutic agents (1). IMPDH inhibitors are used clinically in antiviral (ribavirin) and immunosuppressive therapies (mycophenolate mofetil and mizoribine) (2)(3)(4). In addition, IMPDH inhibitors have anti-tumor and antibiotic activity (5,6). Mammalian and bacterial IMPDHs have significantly different kinetic properties and inhibitor sensitivities, which suggests that species-specific IMPDH inhibitors can be developed that will be useful in treating bacterial and parasitic infections (7)(8)(9). Indeed, many studies have shown that purine metabo-lism plays an important role in bacterial virulence (10 -13).
The spirochete Borrelia burgdorferi is the causative agent of Lyme disease (14). This disease is transmitted by ticks of the Ixodes ricinus complex and is found worldwide. The genes encoding GMP synthase (guaA) and IMPDH (guaB) are located on a 26-kb circular plasmid (cp26) in B. burgdorferi (15). Genes carried by plasmids generally confer selective advantage in a particular environmental niche. The unique plasmid location of these housekeeping genes in B. burgdorferi may be related to their role in the transmission cycle between ticks and mammals. In ticks, guanine is the major nitrogenous waste product and therefore accumulates to high levels. However, in mammals, purine levels are low and limiting for bacterial growth. Therefore expression of the gua genes would be unnecessary for the survival of B. burgdorferi in ticks, but critical for survival in a mammalian host. Consistent with an adaptive role of cp26 in the mammalian environment, the gua genes are linked with ospC, which is induced during tick feeding. ospC encodes a protein that appears on the outer surface of the spirochete immediately preceding transmission to the mammal (16).
The identity of B. burgdorferi guaA was confirmed by complementation of GMP synthase-deficient Escherichia coli (15); however, the guaB homolog did not complement IMPDHdeficient E. coli. The failure to observe complementation by B. burgdorferi guaB most likely results from incompatible promoters or unstable protein. Alternatively, the guaB homolog may not encode an active IMPDH. Indeed, the guaB homolog, with a predicted molecular mass of 44 kDa, is 10 kDa smaller than typical IMPDHs. This difference in size results from the loss of 130 residues (residues 110 -244, Chinese hamster IMPDH numbering) in the middle of IMPDH. These residues have been replaced with 50 residues of unrelated sequence (15).
The crystal structure of IMPDH provides little information about the role of this missing region (17). IMPDH is an ␣/␤ barrel protein, with the active site located in the loops on the C-terminal ends of the ␤ sheets. Residues 100 -244 form an unusual insertion in a loop on the N-terminal end of the ␣/␤ barrel. Therefore, this "subdomain" is on the opposite end of the ␣/␤ barrel from the active site. The subdomain is present in all of the IMPDHs characterized to date, although the subdomain of the guaB1 homolog from Mycobacterium leprae is missing 90 residues (GenBank TM accession no. U00015). This region is also absent in the homologous enzyme GMP reductase (18). A recent report claims that activity is retained when the subdomain is deleted from Chinese hamster IMPDH, although the experimental details were not presented (17). The subdomain has no sequence or structural similarities to proteins other than IMPDH. Thus the role of the subdomain in IMPDH is unknown. Intriguingly, the subdomain is also found in the guaB homolog from the relapsing fever spirochete Borrelia * This work was supported in part by a grant from the Lucille P. Markey Charitable Trust to Brandeis University and the Searle Scholar Program (to L. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  To whom correspondence should be addressed. Tel.: 617-736-2333; Fax: 617-736-2349. 1 The abbreviations used are: XMP, xanthosine 5Ј-monophosphate; IMPDH, inosine 5Ј-monophosphate dehydrogenase; CH 2 -TAD, ␤-methylene thiazole-4-carboxamide adenine dinucleotide; IPTG, isopropyl-1thio-␤-galactopyranoside; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis.
hermsii, 2 which suggests a species-specific modification of IMPDH in Lyme disease spirochetes.
We have expressed the B. burgdorferi guaB homolog in E. coli. This protein has IMPDH activity, which demonstrates that the subdomain is not required for IMPDH activity. B. burgdorferi IMPDH has significantly different kinetic properties from human IMPDH. Therefore, B. burgdorferi IMPDH may be a target for the development of new treatments for Lyme disease.
Construction of an Expression Plasmid for B. burgdorferi IMPDH-Plasmid p68, containing the guaB gene of B. burgdorferi B31, was previously isolated from a genomic library constructed in the vector ZapII (Stratagene, La Jolla, CA) (15). PCR was used to insert convenient restriction sites at the beginning and end of the guaB gene. The  (19). The resulting construct is designated pB9. The guaB coding sequence of pB9 was completely sequenced using a PRISM Dyedeoxy Terminator cycle sequencing kit (ABI) and an Applied Biosystems 373A DNA sequencer at the Brandeis DNA Facility. No unwanted mutations were introduced in the PCR reaction.
Expression and Purification of B. burgdorferi IMPDH-pB9 was transformed into E. coli strain H712, which contains a partial deletion of the E. coli guaB gene (20). An overnight culture of cells was diluted 200-fold into fresh LB broth containing 100 g/ml ampicillin. After 1 h at 37°C, 1 mM IPTG was added to induce expression of IMPDH. The cells were harvested after 13 h by centrifugation, resuspended in buffer A (50 mM Tris, pH 7.4, 1 mM dithiothreitol, 10% glycerol), and frozen at Ϫ20°C. All of the following manipulations were performed at 4°C. The cells were thawed and disrupted by sonication. Debris was removed by centrifugation at 12,000 ϫ g for 25 min. The supernatant was applied to a Cibacron blue Sepharose column previously equilibrated in buffer A. IMPDH was eluted in a linear gradient of 0 to 2 M KCl in buffer A. Fractions containing IMPDH activity were pooled and diluted 4-fold with buffer A. IMP resin was added to the enzyme solution until no activity remained in the supernatant. The resin was then poured into a column and washed with buffer A. IMPDH was eluted with 1 mM IMP, 500 mM KCl in buffer A (IMPDH does not elute in the absence of KCl). Table I summarizes the purification. Protein concentration was measured using the Bio-Rad assay with IgG as a standard. Active sites were titrated with 5-ethynyl-1-␤-D-ribofuranosylimidazole-4-carboxamide 5Јmonophosphate (9). N-terminal sequencing was performed by the Tufts Medical School Protein Sequencing Facility. Electrospray ionization mass spectroscopy was performed by the Harvard University Mass Spectroscopy Laboratory.
Analytical Ultracentrifugation-Sedimentation equilibrium experiments were performed on a Beckman model XL-A analytical ultracentrifuge monitoring absorbance at 280 nm. The rotor speed was 9000 rpm and the temperature was 20°C. Samples contained 0.2-1.5 mg/ml enzyme, 50 mM Tris, pH 7.5, and 1 M dithiothreitol in the presence and absence of 0.1 M KCl. Data were fit to the following equation, which describes the sedimentation of a single component system, using Origin Technical Graphics and Data Analysis (MicroCal): where A 0 and A are the absorbance at the meniscus (r 0 ) and radius r respectively, M is the molecular weight, is the partial specific volume of IMPDH, is solvent density, is the angular velocity, R is the gas constant, and T is temperature. The partial specific volume ϭ 0.75 for B. burgdorferi IMPDH was calculated from the amino acid sequence. Enzyme Kinetics-Assays were performed in 100 mM KCl, 1 mM dithiothreitol, 50 mM Tris, pH 8.0. Activity was routinely assayed in the presence of 250 M IMP and 500 M NAD at 25°C. The production of XMP was monitored in crude extracts by absorbance at 290 nm (⑀ ϭ 5.4 mM Ϫ1 cm Ϫ1 ). The production of NADH was monitored at 340 nm for purified enzyme assays (⑀ ϭ 6.2 mM Ϫ1 cm Ϫ1 ). Initial velocity data were fit to the following equations using KinetAsyst and Kaleidagraph software, Michaelis-Menten equation Competitive inhibition equation Uncompetitive inhibition equation where v is the initial velocity, V m is the maximal velocity, S is substrate, I is inhibitor, K m is the Michaelis constant, and K ii and K is are the intercept and slope inhibition constants, respectively.

RESULTS AND DISCUSSION
Expression and Purification of B. burgdorferi IMPDH-PCR was used to insert convenient restriction sites into the B. burgdorferi guaB gene, NdeI at the start site and PstI in the 3Јnoncoding region, and the gene was cloned into pTactac (19). The guaB gene was sequenced to ensure that no mutations were introduced during PCR. This construct, pB9, was transformed into E. coli strain H712, which lacks endogenous IM-PDH activity (20). H712 cells carrying pB9 could grow on minimal medium, while cells carrying the parent plasmid, pTactac, cannot grow (data not shown). This result demonstrates that pB9 can complement the IMPDH deficiency of 2 P. Rosa, unpublished data.  The purification was accomplished in two steps and high yield, using Cibacron blue Sepharose and IMP affinity resins ( Table  I). Purification of B. burgdorferi IMPDH closely resembled the purification of human type II and E. coli IMPDHs (9,21), with the exception that 0.5 M KCl is required in addition to 500 M IMP to elute the enzyme from the IMP resin. This observation suggests that the B. burgdorferi enzyme may have a different requirement for monovalent cations than other IMPDHs. The purified enzyme is Ͼ95% homogeneous as judged by SDS-PAGE (Fig. 1).
Characterization of B. burgdorferi IMPDH-The N-terminal sequence of purified IMPDH is Pro-Asn-Lys-Ile-Thr-Lys as determined by Edman degradation. This sequence corresponds to the predicted N-terminal sequence of B. burgdorferi IMPDH after removal of the first Met. SDS-PAGE analysis of purified IMPDH shows a single band at 44 kDa (Fig. 1), consistent with the molecular mass of 43,637 Da calculated from the deduced amino acid sequence. This molecular mass confirms that the IMPDH activity derives from the expression of B. burgdorferi guaB because IMPDHs typically have a molecular mass of ϳ55,000 Da (22). A single species with a molecular mass of 43,660 Da is observed by electrospray ionization mass spectroscopy (Fig. 2), which further confirms the identity of B. burgdorferi IMPDH. Fig. 3 shows equilibrium sedimentation data for B. burgdorferi IMPDH. These data were fit to an equation describing the sedimentation of a single ideal species with a molecular mass of 43,660 Da (monomer), 87,320 Da (dimer), or 174,640 Da (tetramer). Reasonable fits could not be obtained to monomeric and dimeric species; the fit to a tetrameric species appears to be good (Fig. 3). Nevertheless, the residuals of the tetrameric fit display a nonrandom distribution, which can indicate an associating system. This fit could not be improved by including terms for monomer-tetramer, dimer-tetramer, or monomerdimer-tetramer equilibria. The fit could be improved slightly by including a tetramer-octamer equilibrium. Such higher order aggregates have been observed in IMPDHs from other species (22,23). No difference in sedimentation behavior was observed at different enzyme concentrations (0.2 to 1.5 mg/ml) or in the presence of 0.1 M KCl. These results indicate that B. burgdorferi IMPDH is a tetramer like other IMPDHs (22,24,25) and may form higher order aggregates as also observed for other IMPDHs.
Steady State Kinetic Parameters-Initial velocity data were collected at varying concentrations of IMP (7 to 980 M) and NAD (125 to 5000 M). The initial velocity versus IMP plots at fixed concentrations of NAD follow Michaelis-Menten kinetics (Fig. 4A). In contrast, substrate inhibition is observed at high NAD concentrations (Fig. 4B). Such substrate inhibition is commonly observed in IMPDHs and suggests that product dissociation follows an ordered mechanism where NADH is the first product released. Steady state parameters were derived by first determining the apparent values of V m for the initial velocity versus IMP plots (as in Fig. 4A) and replotting these values against NAD concentration. These data were fit to Equation 3, which describes uncompetitive substrate inhibition: k cat ϭ 2.6 Ϯ 0.3 s Ϫ1 , K m (NAD) ϭ 1100 Ϯ 160 M and K ii (NAD) ϭ 2300 Ϯ 390 M. The value of K m (IMP) was derived by first determining the apparent values of V m for the initial velocity versus NAD plots using Equation 3 (as in Fig. 4B) and replotting these values against IMP concentration. These data were fit to the Michaelis-Menten equation: K m (IMP) ϭ 29 Ϯ 8 M. These values are comparable to those reported for other bacterial IMPDHs (22,26). Thus B. burgdorferi guaB encodes an IMPDH with typical kinetic properties. This observation demonstrates that the subdomain is not required for IMPDH activity.
Monovalent Cation Dependence-B. burgdorferi IMPDH is inactive in the absence of K ϩ (Ͻ1% activity). High concentrations of K ϩ (Ͼ300 mM) inhibit the enzyme. As shown in Table  II, NH 4 ϩ and K ϩ are interchangeable, while Cs ϩ is a much less effective activator. Neither Na ϩ or Li ϩ activate B. burgdorferi IMPDH; both of these cations are competitive inhibitors with respect to K ϩ . While monovalent cation dependence is commonly observed among IMPDHs, the monovalent cation specificity varies significantly among IMPDHs. K ϩ , NH 4 ϩ , Na ϩ , Tl ϩ , and Rb ϩ activate human IMPDH, while Li ϩ has no effect on this enzyme (27). K ϩ , but not NH 4 ϩ , activates cowpea IMPDH (28), while Bacillus subtilis IMPDH will utilize K ϩ , NH 4 ϩ , Na ϩ , Cs ϩ , Li ϩ , Tl ϩ , and Rb ϩ (29). In contrast to these enzymes, Tritrichomonas foetus IMPDH does not require monovalent cations; this IMPDH retains 80% of its activity in the absence of K ϩ (8). These observations suggest that the monovalent cation binding site is a potential target for species specific inhibitors.
Inhibitor Sensitivity-XMP and GMP are competitive inhibitors with respect to IMP, K is ϭ 85 and 6 M, respectively. The value of K is for XMP is similar to those reported for other IMPDHs. In contrast, the K is for GMP is approximately 15-fold lower than those usually observed for IMPDHs from both mammalian and bacterial sources (8,30,31).
Mycophenolic acid binds in the nicotinamide subsite of the dinucleotide site of IMPDH and prevents the hydrolysis of the covalent enzyme-XMP (E-XMP*) intermediate (17,32,33). Consistent with this mechanism of action, mycophenolic acid is an uncompetitive inhibitor of B. burgdorferi IMPDH with respect to both IMP and NAD (Table III). This observation suggests that B. burgdorferi IMPDH has an ordered mechanism of product release where NADH dissociates before hydrolysis of E-XMP*, as observed with other IMPDHs (34). The K ii for mycophenolic acid inhibition of B. burgdorferi IMPDH is 10 3fold greater than that for mammalian IMPDHs, as is typical of microbial IMPDHs (8,7,35). These results suggest that the dinucleotide site of microbial IMPDHs is a target for speciesspecific inhibitors. CH 2 -TAD is a nonhydrolyzable analog of the active metabolite of the anti-tumor drug tiazofurin. CH 2 -TAD is also an uncompetitive inhibitor of B. burgdorferi IMPDH with respect to both IMP and NAD. This observation suggests that CH 2 -TAD also binds to E-XMP*. The K ii of CH 2 -TAD ϭ 1.0 M is approximately 20-fold greater than that of mammalian IMPDHs (36), further demonstrating a difference in the dinucleotide sites of microbial and mammalian IMPDHs.
Summary-We have demonstrated that the guaB homolog of B. burgdorferi encodes IMPDH. This result demonstrates that the subdomain is not required for IMPDH activity. The function of this subdomain is unknown. In addition, we show that the monovalent cation and dinucleotide binding sites of B. burgdorferi IMPDH differ significantly from mammalian IMPDHs. Therefore these sites are targets for the design of species-specific inhibitors of IMPDH. Such inhibitors could be used to treat Lyme disease. a Inhibition is observed at high concentrations of KCl and NH 4 Cl (Ͼ300 mM); the K ii is shown from the fit of the initial velocity data to Equation 2, which describes substrate inhibition. b n.a., not applicable. c No activity was observed in 1 M NaCl (Ͻ1%) or 0.25 M LiCl (Ͻ0.5%). Both Na ϩ and Li ϩ are competitive inhibitors with respect to K ϩ .