UMP kinase from the Gram-positive bacterium Bacillus subtilis is strongly dependent on GTP for optimal activity Cristina Gagyi 1 , Nadia Bucurenci 2 , Ovidiu Sı ˆ rbu 1 , Gilles Labesse 3 , Mihaela Ionescu 1 , Augustin Ofiteru 2 , Liliane Assairi 1 , Ste ´ phanie Landais 1 , Antoine Danchin 4 , Octavian Ba ˆ rzu 1 and Anne-Marie Gilles 1 1 Laboratoire de Chimie Structurale des Macromole ´ cules, 4 Unite ´ de Ge ´ ne ´ tique des Ge ´ nomes Bacte ´ riens, Institut Pasteur, Paris, France, 2 Laboratory of Enzymology and Applied Microbiology, Cantacuzino Institute, Bucharest, Romania, 3 Centre de Biochimie Structurale, Faculte ´ de Pharmacie, Universite ´ de Montpellier I, Montpellier, France The gene encoding Bacillus subtilis UMP kinase (pyrH/ smbA) is transcribed in vivo into a functional enzyme, which represents approximately 0.1% of total soluble proteins. The specific activity of the purified enzyme under optimal conditions is 25 unitsÆmg )1 of protein. In the absenceofGTP,theactivityofB. subtilis enzyme is less than 10% of its maximum activity. Only dGTP and 3¢-anthraniloyl-2¢-deoxyguanosine-5¢-triphosphate (Ant- dGTP) can increase catalysis significantly. Binding of Ant- dGTP to B. subtilis UMP kinase increased the quantum yield of the fluorescent analogue by a factor of more than three. UTP and GTP completely displaced Ant-dGTP, whereas GMP and UMP were ineffective. UTP inhibits UMP kinase of B. subtilis with a lower affinity than that shown towards the Escherichia coli enzyme. Among nucleoside monophosphates, 5-fluoro-UMP (5F-UMP) and 6-aza-UMP were actively phosphorylated by B. subtilis UMP kinase, explaining the cytotoxicity of the corresponding nucleosides towards this bacterium. A structural model of UMP kinase, based on the conser- vation of the fold of carbamate kinase and N-acetyl- glutamate kinase (whose crystals were recently resolved), was analysed in the light of physicochemical and kinetic differences between B. subtilis and E. coli enzymes. Keywords: UMP kinase; B. subtilis; molecular modelling; GTP activation; fluorescent markers. Phosphorylation of UMP and CMP in eukaryotes is carried out by a single protein. UMP/CMP kinases from Saccharo- myces cerevisiae, Dictyostelium discoideum, Arabidopsis thaliana or pig muscle are monomers that resemble adeny- late kinase from muscle cytosol [1–5]. Enteric bacteria contain separate CMP and UMP kinases, and mutants of Escherichia coli or Salmonella typhimurium defective in the corresponding genes (mssA/cmk and pyrH/smbA, respect- ively) were isolated and characterized many years ago [6–8]. Recombinant CMP and UMP kinases from E. coli have been characterized in detail [9–15]. The CMP kinase from E. coli is a monomer, acting preferentially on CMP and dCMP [11]. Despite the little overall sequence identity with other known nucleoside monophosphate (NMP) kinases, CMP kinase from E. coli has, in common with these enzymes, a central parallel b-sheet, the strands of which are connected by a-helices [13]. In contrast, the UMP kinase from E. coli is a homohexamer whose primary structure diverges from that of other NMP kinases, and is controlled allosterically by GTP (activator) and UTP (inhibitor) [9]. Attempts, in the past, to isolate a specific UMP kinase from Bacillus subtilis were unsuccessful. It was suggested that phosphorylation of UMP in this bacterium is accom- plished by a CMP kinase with a broader specificity for pyrimidine nucleotides than the enzyme from E. coli [16]. The deleterious effect of disruption of cmk/jofC gene in B. subtilis [17], and the kinetic properties of the correspond- ing recombinant protein, were in line with this interpre- tation. Thanks to genome sequencing programs, the pyrH gene has been identified in all bacteria investigated, inclu- ding B. subtilis.ThepyrH gene from Lactococcus lactis, a bacterium similar to B. subtilis in the metabolism of pyrimidine nucleotides, complements a temperature-sensi- tive pyrH mutation in E. coli, demonstrating the ability of the encoded protein to synthesize UDP [18]. These observations reopened the question of the role played by UMP kinase in the metabolism of B. subtilis, and in Gram-positive organisms in general, and prompted us to clone the pyrH gene from B. subtilis and to examine the structural and catalytic properties of the recombinant protein. When compared with the E. coli UMP kinase, several striking characteristics of the B. subtilis enzyme were noticed. Thus, in either crude extract or in purified form, the enzyme is unstable in the absence of UTP. On the other hand, the activity of B. subtilis UMP kinase is very low in the absence of GTP, which explains why Correspondence to A M. Gilles, Laboratoire de Chimie Structurale des Macromole ´ cules, Institut Pasteur, 28, rue du Dr Roux, 75724 Paris cedex 15, France. Fax: +33 1 40 61 39 63, Tel.: +33 1 45 68 89 68, E-mail: amgilles@pasteur.fr Abbreviations: Ant-dGTP, 3¢-anthraniloyl-2¢-deoxyguanosine-5¢- triphosphate; CK-like CPSpf, carbamate kinase-like carbamoyl phosphate synthase from Pyrococcus furiosus; CKef, carbamate kinase from Enterococcus faecalis; 5F-UMP, 5-fluoro-UMP; NAGKec, N-acetylglutamate kinase from Escherichia coli. (Received 10 March 2003, revised 20 May 2003, accepted 3 June 2003) Eur. J. Biochem. 270, 3196–3204 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03702.x previous attempts to isolate the enzyme from this organism were unsuccessful. Materials and methods Chemicals Nucleotides, restriction enzymes, T4 DNA ligase, T7 DNA polymerase and coupling enzymes were from Roche- Applied Science or from Sigma. NDP kinase from D. dis- coideum (2000 UÆmg )1 of protein) was kindly provided by I. Lascu. 5-Fluoro-UMP (5F-UMP) and 6-aza-UMP were synthesized from the corresponding nucleosides with GTP as phophoryl donor, creatine phosphate as regenerating system and recombinant E. coli uridine kinase and rabbit muscle creatine kinase as catalysts. The reaction medium in H 2 O, adjusted to pH 7.0, contained: 40 m M 5F-uridine or 6-azauridine, 0.5 m M GTP, 45 m M creatine phosphate, 20 m M MgCl 2 , 2 unitsÆmL )1 of uridine kinase and 10 unitsÆmL )1 of creatine kinase. When conversion of 6-aza- uridine and 5F-uridine to nucleoside monophosphates was >90%, the reaction medium was heated to precipitate proteins. After desalting, 5-mg samples of 5F-UMP and 6-aza-UMP were purified by reverse-phase HPLC. 3¢-Anthraniloyl-2¢-deoxyguanosine-5¢-triphosphate (Ant- dGTP) was prepared from dGTP and isatoic anhydride essentially by the procedure described in Hiratsuka [19], for the synthesis of Ant-dATP. The analogue was purified by chromatography on LiChroprep RP-18 (25–40 lm) using 1m M triethylammonium acetate as eluent [20]. Bacterial strains, plasmids, growth conditions and DNA manipulations General DNA manipulations were performed as described by Sambrook [21]. ORFs from the pyrH gene from B. subtilis were PCR amplified from chromosomal DNA using the strain 168 [22] as the template and with the following primers: 5¢-pyrH,5¢-GGGGCATATGGAA AAACCAAAATACAAACGTATCGTATTA-3¢;and 3¢-pyrH,5¢-CCCCCTCGAGTTATTTCCCCCTCACGA TCGTTCCGATTGATTCAC-3¢. The product was inser- ted between the NdeIandXhoI restriction sites of plasmid pET24a (Novagen). The resulting plasmid (pSL13) was introduced into strain BL21(DE3)/pDIA17 [23] to overpro- duce the UMP kinase. The recombinant strain was grown in 2YT medium supplemented with antibiotics to an absor- bance (A) of 1.0 at 600 nm, then overproduction was triggered by isopropyl-b- D -thiogalactoside induction (1 m M final concentration) for 3 h, and bacteria were harvested by centrifugation. Purification of UMP kinase and activity assays E. coli overproducing the UMP kinase from B. subtilis were disrupted by sonication in 50 m M Tris/HCl (pH 7.4) containing 2 m M UTP. The bacterial extract was heated for 10 min at 65 °C, then precipitated proteins were removed by centrifugation at 10 000 g for 30 min. The supernatant was concentrated by ultrafiltration, then applied to a Sephacryl S-300 HR column (2.5 · 110 cm), equilibrated with 50 m M Tris/HCl (pH 7.4), 0.1 M NaCl and 2 m M UTP, at a flow rate of 10 mLÆh )1 . The peak fraction containing >95% pure UMP kinase was concentrated to  5mgÆmL )1 of protein. A His-tagged form of UMP kinase was purified by Nickel-nitriloacetic acid affinity chromato- graphy using the QIA express system [24]. The UMP kinase activity was determined at 30 °C using a coupled spectro- photometric assay (0.5-mL final volume) on an Eppendorf PCP6121 photometer [25]. The reaction medium contained 50 m M Tris/HCl (pH 7.4), 50 m M KCl, 2 m M MgCl 2 ,1m M phosphoenolpyruvate, 0.2 m M NADH, 2 m M ATP, 0.5 m M GTP and 2 units each of lactate dehydrogenase, pyruvate kinase and NDP kinase. The crude or pure preparation of UMP kinase was then added, followed 2 min later by 1 m M UMP. The decrease in absorbance (A) at 334 nm was then recorded and corrected for secondary reactions, occurring in the absence of UMP. One unit of UMP kinase corresponds to 1 lmol of product formed per min. The thermal stability was tested by incubating the purified enzyme (1 mgÆmL )1 ) in 50 m M Tris/HCl (pH 7.4), containing 0.1 M NaCl, at temperatures between 30 and 80 °C for 10 min, in the presence or absence of various nucleotides. Results were expressed as percentage of activity as compared with unincubated controls. Analytical procedures Protein concentration was measured according to Brad- ford [26]. Ion spray mass spectra were recorded on a quadrupole mass spectrometer, API-365 (Perkin-Elmer), equipped with an ion spray (nebulizer-assisted electro- spray) source. The sample ( 2pmolÆlL )1 ), dissolved in 20% acetonitrile in water and 0.1% HCOOH, was delivered to the source at a flow rate of 5 lLÆmin )1 . SDS/PAGE was performed as described by Laemmli [27]. The protein bands from SDS/PAGE were electroblotted into a Problot membrane filter (Applied Biosystems) and detected by Coomassie Blue staining. The N-terminal amino acid sequence of the protein from the excised band was determined by a protein sequencer (Applied Biosys- tems, Inc.). Fluorescence experiments were performed on a Jasco FP-750 spectrofluorimeter thermostated at 25 °C. Emission spectra of UMP kinase (kexc ¼ 297 nm; band width ¼ 5 nm) were recorded from 305–400 nm. Equili- brium ultracentifugation was performed at 20 °Cona Beckman Optima XLA ultracentrifuge using a AN60Ti rotor and a cell with a 12-mm optical length. Samples (100 lL) at 0.1 mgÆmL )1 of protein were centrifuged at 10 000 r.p.m. Data were analysed by using the programs IDEAL 1and IDEAL 2, supplied by Beckman. Sequence comparison and molecular modelling Protein sequence database searches were performed using the PSI - BLAST , version 2.0.5, program [28] with default parameters. Protein structure database searches were per- formed using a metaserver dedicated to sequence–structure comparison at low sequence identity level (15–20%) [29]. Alignment refinement was performed manually and assessed by using the program TITO [30], while molecular modelling was performed using MODELLER 6.2 with loop optimization and long molecular dynamics [31]. Models were assessed using PROSA [32], VERIFY 3 D [33] and ERRAT Ó FEBS 2003 UMP kinase from B. subtilis (Eur. J. Biochem. 270) 3197 [34]. Trace evolution is as analysed using CONSURF [35], aiming to delimit the active site as well as the monomer– monomer interface. Results Cloning of the pyrH gene from B. subtilis The pyrH gene from B. subtilis was cloned by PCR into the expression vector pET24a, and sequenced. The resulting ORF showed two differences when compared with the published sequence [22]: one additional T at bp 170; and one missing A at bp 185. As a consequence, the ORF of the pyrH gene displays a double frameshift of 14-bp, resulting in four amino acid residue changes, as follows: 57LeuTrpArg- Gly60 instead of TyrGlyAlaGlu in the original sequence. These differences are not trivial; Arg and Gly are two strictly conserved residues in bacterial UMP kinases, the sequences of which are available in the gene databank. Substitution of the equivalent Arg in E. coli UMP kinase (Arg62His) has pleiotropic effects on stability, catalysis or allosteric regu- lation [12] and is responsible for the altered morphological phenotype of E. coli under nonpermissive conditions, such as cold-sensitive growth or hypersensitivity to SDS [36]. On the other hand, Trp58 of the B. subtilis enzyme is conserved in Streptococcus pneumoniae, Strep. mutans, Enterococ- cus faecalis and Clostridium acetobutylicum UMP kinase and may be a ÔsignatureÕ for Gram-positive organisms (Fig. 1). This amino acid is substituted in the UMP kinase from Gram-negative bacteria by a Phe residue. When present in UMP kinases from Gram-negative organisms, Trp is located in the middle of the sequence (Trp119 in E. coli). As the environment of Trp in UMP kinases from Gram-positive and Gram-negative organisms is apriori different, we expected to find differences in protein intrinsic fluorescence properties, which was indeed the case. Harboured on high-copy number vectors, the B. subtilis pyrH gene complemented the thermosensitive phenotype of strain KUR1244 (pyrH88ts)ofE. coli [37], indicating that it was functional. Complementation experiments performed using E. coli strain MC4100-42-14:40 (car::lacZpyrH42), in which expression of the car::lacZ fusion is repressed in the presence of wild-type UMP kinase activity, showed that in high copy-number, the pyrH gene from B. subtilis resulted in a significant repression of b-galactosidase activity. From the specific activity of UMP kinase in crude extracts of strain 168 (0.03 UÆmg )1 of protein) and the specific activity of the pure recombinant protein under identical experimental conditions, we assumed a protein abundance in B. subtilis extracts of 0.1%, a figure close to that found in E. coli. Sequence comparison and molecular modelling Database screening using the program PSI - BLAST [28], with UMP kinase from B. subtilis as a query sequence, first showed sequence similarities with UMP kinases from other bacteria. Among 26 UMP kinases examined for sequence homology (12 from Gram-positive bacteria, 14 from Gram- negative bacteria), 33 strictly conserved amino acid residues, i.e. 14% of the whole sequence, were noticed. The most frequently represented residues were Gly, Asp and Arg. Site-directed mutagenesis or analysis of the phenotypically characterized mutants [36,37] showed that several strictly conserved amino acids were essential for thermodynamic stability, catalysis or allosteric regulation [12]. Further iterations of the PSI - BLAST search revealed sequence similar- ity of bacterial UMP kinases with pyrroline-5-carboxylate synthase, glutamate-5-kinase, aspartokinases, N-acetylglu- tamate kinases and carbamate kinases (Fig. 1). The latter appeared to be 18% identical with UMP kinases, while N-acetylglutamate kinases were  15% identical (over a region of 200 amino acids). Several motifs appeared to be well conserved among these kinases. At the N-terminus, a glycine-rich sequence motif (r///k/sGxA/: upper case stands for strictly conserved residue; /, for hydrophobic amino acid; and, x, for any amino acid) and a second glycine-rich motif (///GgGnx/r) appeared conserved. These motifs comprise the predicted b-strands (b1andb2) of UMP kinases, both of which might be involved in phosphate binding, according to the crystal structures of carbamate and N-acetylglutamate kinases. In the first motif, the position marked ÔkÕ (Lys12 in UMP kinase from B. subtilis) is occupied either by an alanine (carbamate kinases only) or by a lysine in all other kinases belonging to this superfamily. The side-chain of this residue points toward the leaving phosphate group of ATP. This suggested that carbamate kinases constitute a divergent subfamily with a slightly distinct catalytic mech- anism. A third conserved stretch (r////aaGxgn), is present at the end of the fourth predicted b-strand [38]. In UMP kinases, it is immediately followed by the motif ffttDs, including a catalytic aspartate [9,12,38]. The next motif (txvdGvftadPk) followed the predicted b5-strand (ead///) andprovidesaThrandaValtotheactivesiteofUMP kinases, as well as in aspartokinases and glutamate kinases. This motif comprises the aspartates D168 and D174 (UMP kinase from E. coli numbering) whose role in enzyme activity was also probed by site-directed mutagenesis [9,12,38]. A motif (/k//Dxta) conserved among UMP kinases (including D201 in the UMP kinase from E. coli) [31] would correspond to the N-terminus of the putative helix, a7, which is predicted to contribute residues to the active site. The last conserved motif (gtx/) seems to stabilize the specific structure of the ATP-binding site, rather than directly participating in the binding itself, in the two solved crystal structures [39,40]. The observed conservation of these motifs suggested that a similar mode of ATP binding is used by the enzymes of this superfamily. Despite clear homologies shared by these kinases, proper sequence alignment requires further refinement, based on sequence- structure comparison, owing to the low overall sequence identity level ( 15–20%). Threading of UMP kinase from B. subtilis, using the new meta-server [29], revealed significant fold compatibility (>95% confidence) with carbamate kinase from Ent. fae- calis (CKef) [41] and carbamate kinase-like carbamoyl phosphate synthase from Pyrococcus furiosus (CK-like CPSpf) [39], as well as with N-acetylglutamate kinase from E. coli (NAGKec) [40]. Sequence-to-structure alignment was refined using TITO [30] in order to gather the majority of insertions/deletions in loop regions. CKef and CK-like CPSpf are more than 60 amino acids longer than the UMP kinase from B. subtilis owing to the presence of a small 3198 C. Gagyi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 subdomain of unknown function. The latter module is also absent in NAGKec. The three-dimensional structure of NAGKec and CK-like CPSpf were subsequently used for molecular modelling, taking advantage of their cocrystallization with ATP analogues. However, sequence conservation in the catalytic site also suggested closer relatedness of UMP kinases and N-acetylglutamate kinases (e.g: conservation of a buried lysine; K12 in B. subtilis). In CK-like CPSpf and NAGKec, the C-terminal region participating in the binding of the phosphate donor shows some structural and sequence variations [39,40]. Refined sequence comparison in the vicinity of this region suggested that the ATP-binding site of UMP kinases more closely resembles that of NAGKec than CKs (Fig. 2). Molecular models were built using the program MODEL- LER [31]. It had a pseudo-energy, according to PROSA [32], of )0.7 kcal ()1.9 and )1.8 for the structures of NAGKec and CK-like CPSpf, respectively). VERIFY 3 D scores (dimeric/ monomeric model +0.33/+0.29) were also satisfactory (+0.55 for both crystal structures). ERRAT scores confirmed Fig. 1. Sequence alignment of 10 representative bacterial UMP kinases (bacsu, Bacillus subtilis; strpn, Streptococcus pneumoniae; staau, Staphylo- coccus aureus; entfe, E nt erococcus faecalis; myctu, Mycobacterium tuberculosis; neime, Neisseria meningitidis; pseae, Pseudomonas aeruginosa; ecoli, Escherichia coli; haein, Haemophilus influenzae; yerpe, Yersinia pestis). ThesequenceofN-acetylglutamate kinase from E. coli, as well as of carbamate kinase-like carbamoyl phosphate synthase from Pyrococcus furiosus, whose structures were solved (PDB1gs5 and PDB1e19, respect- ively), are at the bottom. Strictly conserved residues are indicated in the red box. The secondary structure elements, as assigned after structure modelling, are indicated on the top of the sequences. Motifs discussed in the text are written below the alignment. The figure was drawn using ESPRIPT [49]. Ó FEBS 2003 UMP kinase from B. subtilis (Eur. J. Biochem. 270) 3199 the quality at the atomic level (correctness scores of 73% for UMP kinase from B. subtilis vs. 98% and 99% for the structures of NAGKec and CK-like CPSpf, respectively). Similar values were also obtained for various other UMP kinases. The proposed interface for dimerization is in agreement with the evolutionary trace computed by CON- SURF [35]. The latter highlighted a second interface closer to the active site that might correspond to the dimer–dimer interactions. The latter would encompass the region of Trp119 and Pro141 in the UMP kinase from E. coli, in agreement with fluorescence and mutagenesis data, suggest- ing that this region is implied in allosteric regulation [9,38]. Further experimental validation will be necessary to verify these hypotheses. The present model gathers the sequence stretches that are well conserved among UMP kinases on one face of each monomer. These faces are mostly composed of b-strand C-termini and a-helice N-termini, and the connecting loops. The model identifies the C-terminal segment in bacterial UMP kinases as the ATP-binding site, while the N-terminal domain would contain the cosubstrate and the effector binding sites. The dimerization interface is composed of a-helices and a b-strand from the N-terminal end. The current model suggests that Trp58 is readily accessible to the solvent in a pocket opening on the active site, while Cys206 is buried by the very C-terminus (residues Gly239–Lys240) at  15 A ˚ from the ATP-binding site and 30 A ˚ from Trp58. ATP would be in close contact with Ne of Lys12, a residue strictly conserved in bacterial UMP kinases as well as in the other members of this superfamily of catalysts, such as carbamate kinases. Substitution of this conserved Lys residue in UMP kinase from Streptococcus pneumoniae yielded a completely inactive enzyme (L. Assairi, unupub- lished results). The catalytic Asp146 identified in UMP kinase from E. coli corresponds to Asp143 in the corresponding B. subtilis enzyme (Fig. 2). In the vicinity of the active site, the main difference between the latter enzyme and UMP kinases from the Gram-negative organisms is the insertion of one residue (Asn164) lying in the vicinity of ATP. The putative role of this one-residue insertion remains unexplained in our current structure modelling. Purification and molecular properties of recombinant UMP kinase from B. subtilis UMP kinase from B. subtilis, overproduced in strain BL21(DE3)/pDIA17, was purified as described in the Materials and methods, i.e. a heating step followed by gel-permeation chromatography. The molecular mass of B. subtilis UMP kinase (26 084.2 ± 1.5 Da), as measured by ESI-MS, was in agreement with that calculated (26 083 Da) from the sequence. Gel-permeation chromato- graphy yielded a peak consistent with an oligomeric enzyme (six subunits/oligomer) that was preceded by a shoulder. Ultracentrifugation analysis by sedimentation equilibrium indicated that the dominant species corresponded to the hexameric enzyme (154 kDa), even though oligomers of a higher molecular mass (283 kDa) were also identified. They correspond most probably to the association of two hexamers. In parallel, an N-terminal His-tagged form was produced and purified by Nickel-nitriloacetic chromato- graphy. The ESI-MS-determined molecular mass, of 28 114.6 ± 1.2 Da, was lower than that calculated from the sequence (28 246.41 Da), the difference (131.9 Da) accounting for the missing N-terminal Met in the His- tagged recombinant enzyme. The specific activity of the native and His-tagged UMP kinase under optimal assay conditions was 26 UÆmg )1 of protein and 25 UÆmg )1 of protein, respectively, which correspond to molar activities (mol productÆs )1 Æmol enzyme )1 ) of 11.3 and 11.8, respect- ively. UTP (2 m M ) significantly stabilized the bacterial UMP kinase, which can be stored at room temperature for 2weeks in 50m M Tris/HCl (pH 7.4), containing 0.1 M NaCl and 2 m M UTP, with no loss of activity. UTP also increased the thermal stability of B. subtilis UMP kinase, the half-maximal inactivation being shifted from 42 °Cin the absence to >70 °C in the presence (1 m M )ofthe nucleotide. The protective effect of UTP against thermal denaturation is specific for this nucleotide, ATP and GTP being ineffective. UTP was demonstrated also to increase the thermal stability of E. coli UMP kinase [9,10]. The single Cys residue (Cys206) of UMP kinase from B. subtilis is conserved in the enzyme from S. aureus, Pseudomonas aeruginosa, Bordetella pertussis, Mycobacterium tuberculosis, Neisseria meningitidis, Chla- mydia trachomatis and M. pneumoniae, but not in the UMP kinase from E. coli, Sal. typhi, Yersinia pestis or Haemophilus influenzae. This residue reacted with DTNB under native conditions. The kinetics was fitted to a single- exponential equation except for an initial missing amplitude, which corresponds probably to the reaction with DTNB of the partially unfolded fraction of enzyme (Fig. 3). The k obs varied between 10 )3 Æs )1 (in the presence of UTP) and Fig. 2. Modelled structure of UMP kinase from Bacillus subtilis. One monomer is shown in CPK representation. Colours (from blue to violet) indicate residue conservation (weakly to strongly) as computed by CONSURF [35] and visualized using RASMOL (http://www.umass.edu/ microbio/rasmol/). The second monomer in the model is shown as green ribbon. ATP molecules are shown in yellow. 3200 C. Gagyi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 7 · 10 )3 Æs )1 (in its absence). GTP, ATP and Mg 2+ did not significantly affect the reaction-rate constant of UMP kinase with DTNB. Mg 2+ , when present in excess over UTP, reverses its protective effect. The fluorescence spectrum of UMP kinase from B. sub- tilis upon excitation at 295 nm exhibited a maximum at 348 nm, indicating that Trp58 is fully exposed to the solvent. UTP shifted the fluorescence maximum to 331 nm, with no change in intensity. Mg 2+ , in excess over UTP, reversed this effect. Binding of UTP to the enzyme exhibits a slight cooperative effect (Hill number 1.5, K d 22 l M ). GTP binding, which decreased the fluorescence intensity at 348 nm without shift in the maximum (K d  30 l M ), was independent of Mg 2+ ions (results not shown). Kinetic and nucleotide-binding properties of UMP kinase from B. subtilis The UMP kinase activity with various nucleoside triphos- phates and UMP at fixed concentrations (1 m M ) indicated unusually low specific activities for this class of enzymes. The maximal rate was found with ATP and dATP. When NTPs were used in the mixture, the highest specific activity was obtained with ATP + GTP, indicating a requirement for GTP for expression of full catalytic activity (Table 1). When the ATP concentration was varied in the presence (0.5 m M ) or absence of GTP at a single concentration of UMP (1 m M ), the apparent K m for ATP was 0.9 m M .When ATP was constant (1 m M ), the kinetics with variable concentrations of UMP was dependent on GTP. Thus, in the absence of GTP, the rates were maximal at 50–70 l M UMP (K m for UMP  8 l M ) and declined upon further increases in UMP. In the presence of GTP, saturation was attained at 0.2 m M UMP and the apparent K m for UMP was 30 l M without inhibition by excess nucleoside mono- phosphate. GTP showed the most important activating effect, the half-maximum activation being reached at  0.1 m M . dGTP was also effective, but with lower affinity. Esterification of 3¢-OH in dGTP with the anthraniloyl group increased, by one order of magnitude, the affinity of the parent nucleotide for B. subtilis UMP kinase. ITP was a rather weak activator, whereas GMP was ineffective (Fig. 4A). These results are significantly different from those obtained with E. coli UMP kinase, in which GMP, cGMP and even guanosine exerted activation. UTP antagonized the activation by GTP (Fig. 4B). In the absence of GTP, UTP decreased the reaction rate with an I 50 of approxi- mately 60 l M , where I 50 represents the concentration of UTP at which half-inhibition was observed. Of the nucle- oside monophosphates tested, 5F-UMP and 6-aza-UMP showed the most interesting properties. Both analogues were actively phosphorylated by UMP kinase from B. sub- tilis and E. coli (Table 2), which is in line with the cytotoxic effect of the corresponding nucleosides on these organisms [42,43]. Kinetic and intrinsic fluorescence studies of UMP kinase from B. subtilis suggested that the allosteric effectors, GTP and UTP, bind to identical or largely overlapping sites. However, in the absence of binding studies, ambiguities, as regards the identity of the ÔactivatingÕ or ÔinhibitoryÕ sites, still persist. Therefore, we explored several analogues of UTP and GTP capable of binding to allosteric site(s) of UMP kinase and giving an appropriate fluorescence signal. Ant-dGTP, which activated UMP kinase with a higher affinity than that of the parent nucleotide, exhibited strong fluorescent properties in aqueous solution upon excitation at 330 nm with a maximum at 425 nm. The addition of enzyme in excess over Ant-dGTP increased its fluorescence intensity by more than threefold. UTP and GTP, and, to a lesser extent, the corresponding nucleoside diphosphates, displaced the fluorescent analogue from UMP kinase. Under identical experimental conditions, ATP competed weakly with Ant-dGTP binding, whereas UMP and GMP were ineffective (Fig. 4C). Ant-dUTP behaved similarly, although the enhancement in fluorescence intensity of the analogue, upon addition of the enzyme, was lower (by a factor of 1.8). Both GTP and UTP competed effectively with Ant-dUTP, whereas ATP was almost completely ineffective (data not shown). Fig. 3. Reaction of UMP kinase from Bacillus subtilis with 5,5¢-di- thiobis (2-nitrobenzoic acid) (DTNB) under native conditions. Enzyme (25 l M in terms of monomer), in 50 m M Tris/HCl (pH 8) containing 100 m M NaCl, was thermostated at 25 °C and treated with DTNB (0.2 m M final concentration). The absorbance (A) was then read at 412 nm for 20 min. The molar ratio of thiol reacted was calculated using the mass of His-tagged protein of 28.1 kDa, and e of thio- nitrobenzoate anion of 13.6Æm M )1 . Table 1. Reaction rate of UMP kinase from Bacillus subtilis with vari- ous nucleoside triphosphates (NTPs). The reaction medium (0.5 mL final volume) was the same as that described in the legend to Fig. 4. The concentrations of NTPs and of UMP were constant (1 m M ). The reaction rate with ATP + GTP (14.2 lmolÆmin )1 Æmg )1 of protein) is considered to be 100%. NTP Reaction rate (%) ATP 7 dATP 8 GTP 0.7 ITP <0.07 CTP <0.07 ATP + GTP 100 dATP + GTP 94 ATP + ITP 27 dATP + ITP 41 ATP + CTP 6.3 ITP + GTP 0.7 Ó FEBS 2003 UMP kinase from B. subtilis (Eur. J. Biochem. 270) 3201 Discussion Bacterial UMP kinases are the most intriguing members of the NMP kinase family, for several reasons. These enzymes do not exhibit any sequence homology with other NMP kinases described so far, exist in oligomeric form (hexamers) and are subject to complex regulation by nucleotides. In addition, the pyrH gene product from enteric bacteria directly participates in the pyrimidine-specific control of the carAB operon [44]. For these reasons we undertook a systematic analysis of the pyrH gene product in different bacteria with the aim of correlating differences in enzyme structure with substrate specificity, stability and the capa- bility of organisms to adapt to environmental conditions. In this respect, UMP kinases from E. coli (a Gram-negative bacterium) and B. subtilis (a Gram-positive bacterium) have some characteristics in common, i.e. high sequence identity, identical quaternary structure and similar catalytic proper- ties, such as activation by GTP and inhibition by UTP. They also exhibit significant differences in stability and physicochemical properties that might be rationalized in structural terms [38]. According to the current modelling study, UMP kinase from B. subtilis is made of one domain comprising a central, mostly parallel, b-sheet surrounded by a-helices. The N-terminus (residues 5–142) of protein would also build up the dimer interface. The position of ATP in the model was deduced from the position of the AMP-PNP (adenosine b,c-imido-5¢-triphosphate) and that of ADP in the crystal structures of NAGKec (PDB1gs5) and of CK-like CPSpf (PDB1e19), respectively. It suggests that the C-terminus (residues 143–240) is probably the phos- phate donor-binding site. The UMP- and the allosteric binding sites remain to be identified. Our model suggested that Trp58 in the B. subtilis UMP kinase is 10 A ˚ from the active site and close to a solvent-exposed groove. The latter is formed by the helix a2 (residues 55–67), the following loop (68–74) and a second loop (137–141). The equivalent region in the E. coli UMP kinase contains several conserved residues whose substitutions (Arg62His, Asp77Asn) affect the kinase activity and especially its allosteric regulation [12]. The fluorescence of Trp119 in the E. coli UMP kinase, modelled in the same region, is also affected by allosteric effectors [9]. The corresponding region (residues 100–150) of the related glutamate-5-kinase from E. coli also contained an allosteric site according to two characterized mutations [45]. The unique cysteine (Cys206) of B. subtilis UMP kinase was assigned at the C-terminus of helix a7 facing the strand b8. The reactivity Fig. 4. Interaction of UMP kinase from Bacillus subtilis with various nucleotides acting as allosteric effectors. The reaction medium (0.5 mL final volume) buffered with 50 m M Tris/HCl (pH 7.4) contained 50 m M KCl, 2 m M MgCl 2 , 1m M phosphoenolpyruvate, 0.2 m M NADH, 2m M ATP, 1 m M UMP and different concentrations of GTP, dGTP, 3¢-anthraniloyl-2¢-deoxyguanosine-5¢-triphosphate (Ant-dGTP), ITP or GMP, as indicated, and 2 units of each of pyruvate kinase, NDP kinase and lactate dehydrogenase. The reaction was started with pure UMP kinase. (Top) Ant-dGTP (h); GTP (j); dGTP (m); ITP (s); GMP (d). (Middle) The variable nucleotide was UTP, in the presence of constant concentrations of GTP: 0.1 m M (j); 0.2 m M (m); 0.5 m M (h). (Bottom) Binding of Ant-dGTP (2 l M ) to UMP kinase of B. subtilis (20 l M ) as determined at 420 nm (excitation k at 330 nm) and displacement by UTP (d), GTP (j), ATP (h)andGMP(n). Table 2. Comparative kinetic parameters of UMP kinase from Bacillus subtilis and Escherichia co li with UMP analogues. The reaction medium is the same as that described in the legend to Fig. 4. The concentrations of ATP (2 m M ) and GTP 0.5 (mM) were constant. The concentration of UMP and of its analogues varied between 0.02 and 2 m M . V m is expressed as lmolÆmin )1 Æmg )1 of protein. 5F-UMP, 5-fluoro-UMP. NMP B. subtilis E. coli V m K m (l M ) Relative V m /K m V m K m (l M ) Relative V m /K m UMP 25 30 100 126 50 100 5F-UMP 24 120 24 162 110 58 6 aza-UMP 0.6 140 5 67 710 4 3202 C. Gagyi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 to DTNB is affected specifically by UTP. This effect comes in line with the effect of UTP on the protein stability, as the flexibility decrease would maintain the buried Cys206 as almost completely inaccessible. Another striking characteristic of B. subtilis UMP kinase is its high requirement for GTP for full catalytic activity, an effect antagonized by UTP. This property is shared by UMP kinase from other Gram-positive organisms such as Strep. pneumoniae or Staph. aureus (N. Bucurenci and M. Straut, unpublished data). At physiological pH, GTP activates UMP kinase from E. coli or H. influenzae by only three- to fourfold. Binding studies with Ant-dGTP and Ant- dUTP, and displacement by the natural effectors GTP and UTP,suggestthattheÔactivatorÕ and ÔinhibitorÕ sites are identical. Although chemical modification experiments on E. coli UMP kinase favoured also the idea that the GTP and UTP sites overlap [9], direct proof that the allosteric effectors compete for the same site in bacterial UMP kinases was still missing. We propose, by analogy with aspartate transcarbamylase, which binds at the same regulatory-site CTP (inhibitor) and ATP (activator) [46], that UMP kinase requires essentially the same residues for interacting with both UTP and GTP. The ÔredistributionÕ of hydrogen bond network upon binding of closely related nucleotides has been previously shown with other enzymatic systems [15]. It remains to be demonstrated that the opposite effects on conformation and catalysis, accompanying the binding of GTP and UTP, result primarily from interactions at the level of the heterocycle. Whatever the mechanism of activation may be under in vitro conditions, a question arises: can the concentration of GTP and UTP play a role in the metabolism of B. subtilis in vivo? From the total concentration of different NTPs and Mg 2+ in B. subtilis, [47,48], it is conceivable that GTP is the major player besides the two substrates ATP and UMP, as it interacts with UMP kinase both under Mg 2+ complexed form or as a free nucleotide. Mg 2+ -UTP, which is the major form of the nucleotide in the bacterial cell, is only a weak competitive inhibitor for ATP (K i >2m M ), whereas Mg 2+ -free UTP (1–5 l M ) is unable to inhibit UMP kinase from B. subtilis in vivo, via the allosteric site. This picture is different for UMP kinase from Gram-negative organisms, which exhibit a K d for UTP in the micromolar or nanomolar range [9] (Gilles et al. unpublished results). A last point, worthy of mention, is the involvement of UMP kinase in the metabolism of cytotoxic nucleobases or nucleosides, which act as phosphorylated derivatives inter- fering either with DNA or RNA synthesis or by inhibiting key enzymes in the formation of nucleoside triphosphates [42,43]. In this respect, both 5F-uridine and 6-azauridine are readily phosphorylated to 5F-UMP and 6-aza-UMP by bacterial uridine kinase. UMP kinase from B. subtilis and E. coli use both nucleotides with relatively high efficiency, contributing, with the nonspecific NDP kinase, to the formation of the highly toxic compounds F-UTP and 6-aza-UTP. Acknowledgements We thank Jan Neuhard for many inspiring discussions, Nicolas Glansdorff for the kind gift of E. coli 14:40-42, Sylvie Pochet for help in the synthesis of 5F-UMP and 6-Aza-UMP, Hiroshi Sakamoto and Tudor Borza for cloning and complementation experiments, Jean-Claude Rousselle for mass spectrometry measure- ments, Dominique Douguet and Susan Michelson for constructive criticism, He ´ le ` ne Munier-Lehmann for fruitful comments, and Re ´ gine Lambrecht for excellent secretarial help. This work was supported by grants from the Centre National de la Recherche Scientifique (URA2185), Institut Pasteur (AC02) and AstraZeneca R &D,Boston,Inc. References 1. Dreusicke, D., Karplus, P.A. & Schulz, G.E. (1988) Refined structure of porcine cytosolic adenylate kinase at 2.1 A ˚ resolution. J. Mol. Biol. 199, 359–371. 2. Mu ¨ ller-Dieckmann, H J. & Schulz, G.E. (1994) The structure of uridylate kinase with its substrates, showing the transition state geometry. J. Mol. Biol. 236, 361–367. 3. Okajima, T., Goto, S., Tanizawa, K., Tagaya, M., Fukui, T., Shimofuruya, H. & Suzuki, J. (1995) Cloning, sequencing, and expression in Escherichia coli of cDNA encoding porcine brain UMP-CMP kinase. J. Biochem. 117, 980–986. 4. Scheffzek, K., Kliche, W., Wiesmu ¨ ller, L. & Reinstein, J. (1996) Crystal structure of the complex of UMP/CMP kinase from Dictyostelium discoideum and the bisubstrate inhibitor P 1 -(5¢- adenosyl) P 5 -(5¢-uridyl) pentaphosphate (UP 5 A) and Mg 2+ at 2.2 A ˚ : implications for water-mediated specificity. Biochemistry 35, 9716–9727. 5. Zhou, L., Lacroute, F. & Thornburg, R. (1998) Cloning, expres- sion in Escherichia coli,andcharacterizationofArabidopsis thali- ana UMP/CMP kinase. Plant Physiol. 117, 245–254. 6. Ingraham, J.L. & Neuhard, J. (1972) Cold-sensitive mutants of Salmonella typhimurium defective in uridine monophosphate kinase (pyrH). J. Biol. Chem. 247, 6259–6265. 7. Beck, C.F., Neuhard, J., Thomassen, E., Ingraham, J.L. & Kleker, E. (1974) Salmonella typhimurium mutants defective in cytidine monophosphate kinase (cmk). J. Bacteriol. 120, 1370–1379. 8. Pie ´ rard, A., Glansdorff, N., Gigot, D., Crabeel, M., Halleux, P. & Thiry, L. (1976) Repression of Escherichia coli carbamoyl- phosphate synthase: relationships with enzyme synthesis in the arginine and pyrimidine pathways. J. Bacteriol. 127, 291–301. 9. Serina, L., Blondin, C., Krin, E., Sismeiro, O., Danchin, A., Sakamoto, H., Gilles, A M. & Baˆ rzu, O. (1995) Escherichia coli UMP-kinase, a member of the aspartokinase family, is a hexamer regulated by guanine nucleotides and UTP. Biochemistry 34, 5066–5074. 10. Serina, L., Bucurenci, N., Gilles, A M., Surewicz, W.K., Fabian, H., Mantsch, H.H., Takahashi, M., Petrescu, I., Batelier, G. & Baˆ rzu, O. (1996) Structural properties of UMP-kinase from Escherichia coli: modulation of protein solubility by pH and UTP. Biochemistry 35, 7003–7011. 11. Bucurenci, N., Sakamoto, H., Briozzo, P., Palibroda, N., Serina, L., Sarfati, R.S., Labesse, G., Briand, G., Danchin, A., Baˆ rzu, O. & Gilles, A M. (1996) CMP kinase from Escherichia coli is structurally related to other nucleoside monophosphate kinases. J. Biol. Chem. 271, 2856–2862. 12. Bucurenci, N., Serina, L., Zaharia, C., Landais, S., Danchin, A. & Baˆ rzu, O. (1998) Mutational analysis of UMP kinase from Escherichia coli. J. Bacteriol. 180, 473–477. 13. Briozzo, P., Golinelli-Pimpaneau, B., Gilles, A M., Gaucher, J F., Burlacu-Miron, S., Sakamoto, H., Janin, J. & Baˆ rzu, O. (1998) Structures of Escherichia coli CMP kinase alone and in complex with CDP: a new fold of the nucleoside monophosphate binding domain and insights into cytosine nucleotide specificity. Structure 6, 1517–1527. 14. Landais, S., Gounon, P., Laurent-Winter, C., Mazie ´ , J C., Dan- chin, A., Baˆ rzu, O. & Sakamoto, H. (1999) Immunochemical Ó FEBS 2003 UMP kinase from B. subtilis (Eur. J. Biochem. 270) 3203 analysis of UMP kinase from Escherichia coli. J. Bacteriol. 181, 833–840. 15. Bertrand, T., Briozzo, P., Assairi, L., Ofiteru, A., Bucurenci, N., Munier-Lehmann, H., Golinelli-Pimpaneau, B., Baˆ rzu, O. & Gilles, A M. (2002) Sugar specificity of bacterial CMP kinases revealed by crystal structures and mutagenesis of Escherichia coli enzyme. J. Mol. Biol. 315, 1099–1110. 16. Waleh, N.S. & Ingraham, J.L. (1976) Pyrimidine ribonucleoside monophosphokinase and the mode of RNA turnover in Bacillus subtilis. Arch. Microbiol. 110, 49–54. 17. Sorokin, A., Serror, P., Pujic, P., Azevedo, V. & Ehrlich, S.D. (1995) The Bacillus subtilis chromosome region encoding homo- logues of the Escherichia coli mssA and rpsA gene products. Microbiology 141, 311–319. 18. Wadskov-Hansen, S.L., Martinussen, J. & Hammer, K. (2000) The pyrH gene of Lactococcus lactis subsp. cremoris encoding UMP kinase is transcribed as part of an operon including the frr1 gene encoding ribosomal recycling factor 1. Gene 241, 157–166. 19. Hiratsuka, T. (1983) New ribose-modified fluorescent analogs of adenine and guanine nucleotides available as substrates for var- ious enzymes. Biochim. Biophys. Acta 742, 496–508. 20. Sarfati, R.S., Kansal, V.K., Munier, H., Glaser, P., Gilles, A M., Labruye ` re, E., Mock, M., Danchin, A. & Baˆ rzu, O. (1990) Binding of 3¢-anthraniloyl-2¢deoxy-ATP to calmodulin-activated adenylate cyclase from Bordetella pertussis and Bacillus anthracis. J. Biol. Chem. 265, 18902–18906. 21. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 22. Kunst, F., Ogasawara, N., Moszer, I., Albertini, A.M., Alloni, G. et al. (1997) The complete genome sequence of the Gram-positive bacterium Bacillus subtilis. Nature 390, 249–256. 23. Munier, H., Gilles, A M., Glaser, P., Krin, E., Danchin, A., Sarfati, R.S. & Baˆ rzu, O. (1991) Isolation and characterization of catalytic and calmodulin-binding domains of Bordetella pertussis adenylate cyclase. Eur. J. Biochem. 196, 469–474. 24. Crowe, J., Dobeli, H., Gentz, R., Hochuli, E., Stuber, D. & Henco, K. (1994) 6xHis-Ni-NTA chromatography as a superior technique in recombinant protein expression/purification. Meth- ods Mol. Biol. 31, 371–387. 25. Blondin, C., Serina, L., Wiesmu ¨ ller, L., Gilles, A M. & Baˆ rzu, O. (1994) Improved spectrophotometric assay of nucleoside mono- phosphate kinase activity using the pyruvate kinase/lactate dehy- drogenase coupling system. Anal. Biochem. 220, 219–221. 26. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye-binding. Anal. Biochem. 72, 248–254. 27. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680–685. 28. Altschul, S.F., Madden, T.L., Schaeffer, A.A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D.J. (1997) Gapped BLAST and PSI - BLAST : a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. 29. Douguet, D. & Labesse, G. (2001) Easier threading efficiency through Web-based comparisons and cross-validations. Bioinfor- matics 17, 752–753. 30. Labesse, G. & Mornon, J P. (1998) A Tool for Incremental Threading Optimization ( TITO ) to help alignment and modelling of remote homologs. Bioinformatics 14, 206–211. 31. Sali, A. & Blundell, T.L. (1993) Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815. 32. Sippl, M.J. (1993) Recognition of errors in three-dimensional structures of proteins. Proteins 17, 355–362. 33. Colovos, C. & Yeates, T.O. (1993) Verification of protein struc- tures: patterns of nonbonded atomic interactions. Protein Sci. 2, 1511–1519. 34. Eisenberg, D., Luthy, R. & Bowie, J.U. (1997) VERIFY 3 D : assess- ment of protein models with three-dimensional profiles. Methods Enzymol. 277, 396–404. 35. Armon, A., Graur, D. & Ben-Tal, N. (2001) CONSURF :analgo- rithmic tool for the identification of functional regions in proteins by surface mapping of phylogenetic information. J. Mol. Biol. 307, 447–463. 36. Yamanaka, K., Ogura, T., Niki, H. & Hiraga, S. (1992) Identification and characterization of the smbA gene, a suppressor of the mukB null mutant of Escherichia coli. J. Bacteriol. 174, 7517–7526. 37. Smallshaw, J. & Kelln, R.A. (1992) Cloning, nucleotide sequence and expression of the Escherichia coli K-12 pyrH gene encoding UMP kinase. Genetics (Life Sci. Adv.) 11, 59–65. 38. Labesse, G., Bucurenci, N., Douguet, D., Sakamoto, H., Lan- dais, S., Gagyi, C., Gilles, A.M. & Barzu, O. (2002) Compara- tive modelling and immunochemical reactivity of Escherichia coli UMP kinase. Biochem. Biophys. Res. Commun. 294, 173– 179. 39. Ramon-Maiques, S., Marina, A., Uriarte, M., Fita, I. & Rubio, V. (2000) The 1.5-A ˚ resolution crystal structure of the carbamate kinase-like carbamoyl phosphate synthetase from the hyper- thermophilic Archaeon Pyrococcus furiosus, bound to ADP, confirms that this thermostable enzyme is a carbamate kinase, and provides insight into substrate binding and stability in carbamate kinases. J. Mol. Biol. 299, 463–476. 40. Ramon-Maiques, S., Marina, A., Gil-Ortiz, F., Fita, I. & Rubio, V. (2002) Structure of acetylglutamate kinase, a key enzyme for arginine biosynthesis and a prototype for the amino acid kinase enzyme family, during catalysis. Structure 10, 329–342. 41. Marina, A., Alzari, P.M., Bravo, J., Uriarte, M., Barcelona, B., Fita, I. & Rubio, V. (1999) Carbamate kinase: new structural machinery for making carbamoyl phosphate, the common pre- cursor of pyrimidines and arginine. Protein Sci. 8, 934–940. 42. Saunders, P.P., Bass, R.E. & Saunders, G.F. (1968) Properties of 5-fluorouracil-containing ribonucleic acid and ribosomes from Bacillus subtilis. J. Bacteriol. 96, 525–532. 43. Beranek, J. (1986) A study on structure–activity relationships of nucleoside analogues. DrugsExp.Clin.Res.12, 355–367. 44. Kholti, A., Charlier, D., Gigot, D., Huysveld, N., Roovers, M. & Glansdorrf, N. (1998) pyrH-encoded UMP-kinase directly parti- cipates in pyrimidine-specific modulation of promoter activity in Escherichia coli. J. Mol. Biol. 280, 571–582. 45. Csonka, L.N., Gelvin, S.B., Goodner, B.W., Orser, C.S., Siemie- niak, D. & Slightom, J.L. (1988) Nucleotide sequence of a muta- tion in the proB gene of Escherichia coli that confers proline overproduction and enhanced tolerance to osmotic stress. Gene 664, 199–205. 46. Lipscomb, W.N. (1994) Aspartate transcarbamylase from Escherichia coli: activity and regulation. In Advances in Enzymo- logy (A. Meister, ed), Vol. 68, pp. 67–151. J. Wiley and Sons, Inc., New York. 47. Molgaard, H. (1980) Deoxyadenosine/deoxycytidine kinase from Bacillus subtilis: purification, characterization and physiological functions. J. Biol. Chem. 255, 8216–8220. 48. Snavely, N.D. (1990) Magnesium transport in procaryote cells. In Metals in Biological Systems, Vol. 26 (Sigel, H. & Sigel, A., eds), pp. 155–175. Marcel Dekker, Inc., New York. 49. Gouet, P., Courcelle, E., Stuart, D.I. & Metoz, F. (1999) ESPRIPT : analysis of multiple sequence aligments in Post Script. Bioinfor- matics 15, 305–308. 3204 C. Gagyi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 . UMP kinase from the Gram-positive bacterium Bacillus subtilis is strongly dependent on GTP for optimal activity Cristina Gagyi 1 , Nadia Bucurenci 2 , Ovidiu. demonstrating the ability of the encoded protein to synthesize UDP [18]. These observations reopened the question of the role played by UMP kinase in the metabolism of B. subtilis, and in Gram-positive. UMP kinase from B. subtilis is conserved in the enzyme from S. aureus, Pseudomonas aeruginosa, Bordetella pertussis, Mycobacterium tuberculosis, Neisseria meningitidis, Chla- mydia trachomatis