- Email: [email protected]
The CTP:phosphocholine cytidylyltransferase encoded by the licC gene of Streptococcus pneumoniae: cloning, expression, purification, and characterization1
Biochimica et Biophysica Acta 1534 (2001) 85^95 www.bba-direct.com
The CTP:phosphocholine cytidylyltransferase encoded by the licC gene of Streptococcus pneumoniae: cloning, expression, puri¢cation, and characterization1 Heidi A. Campbell, Claudia Kent * Department of Biological Chemistry, 4417 Medical Science I, University of Michigan Medical Center, Ann Arbor, MI 48109-0606, USA Received 1 August 2001; received in revised form 13 September 2001; accepted 13 September 2001
Abstract Streptococcus pneumoniae is a member of a small group of bacteria that display phosphocholine on the cell surface, covalently attached to the sugar groups of teichoic acid and lipoteichoic acid. The putative pathway for this phosphocholine decoration is, in its first two enzymes, functionally similar to the CDP-choline pathway used for phosphatidylcholine biosynthesis in eukaryotes. We show that the licC gene encodes a functional CTP:phosphocholine cytidylyltransferase (CCT). The enzyme has been expressed and purified to homogeneity. Assay conditions were optimized, particularly with respect to linearity with time, pH, Mg2 , and ammonium sulfate concentration. The pure enzyme has KM values of 890 þ 240 WM for CTP, and 390 þ 170 WM for phosphocholine. The kcat is 17.5 þ 4.0 s31 . S. pneumoniae CTP :phosphocholine cytidylyltransferase (SpCCT) is specific for CTP or dCTP as the nucleotide substrate. SpCCT is strongly inhibited by Ca2 . The IC50 values for recombinant and native SpCCT are 0.32 þ 0.04 and 0.27 þ 0.03 mM respectively. The enzyme is also inhibited by all other tested divalent cations, including Mg2 at high concentrations. The cloning and expression of this enzyme sets the stage for design of inhibitors as possible antipneumococcal drugs. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Cytidylyltransferase; licC; Phosphocholine; CDP-choline; Teichoic acid; Streptococcus pneumoniae
1. Introduction The pathogen Streptococcus pneumoniae contains
Abbreviations: CCT, CTP:phosphocholine cytidylyltransferase; SpCCT, Streptococcus pneumoniae CTP:phosphocholine cytidylyltransferase; GCT, glycerol-3-phosphate cytidylyltransferase; TMAE, trimethylaminoethyl ; EDTA, ethylenediaminetetraacetic acid * Corresponding author. Fax: +1-734-763-4581. E-mail address: [email protected] (C. Kent). 1 The sequence described herein has been deposited in GenBank with accession number AF402777.
phosphocholine as a constituent of its teichoic and lipoteichoic acid, with the phosphocholine attached to one or both of the N-acetylgalactosamine residues of the pentasaccharide repeating unit [1,2]. This cell surface phosphocholine serves as an anchor for a number of choline-binding proteins [3], as an anchor and activator of both endogenous and bacteriophage autolysins [4], and may interact with the host immune system [5]. Although S. pneumoniae is a choline auxotroph, mutant strains have been isolated that are capable of growing without choline or the similar amino alcohol, ethanolamine [6,7]. It has therefore been postulated that S. pneumoniae is auxotrophic
1388-1981 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 8 - 1 9 8 1 ( 0 1 ) 0 0 1 7 4 - 3
BBAMCB 55834 17-12-01
86
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
for choline, not because any of the roles for cell surface phosphocholine are individually or collectively indispensable, but because there is a checkpoint in the assembly of teichoic and lipoteichoic acid for the presence of phosphocholine [8]. The S. pneumoniae pathway for choline phosphorylation and activation may be similar to the eukaryotic CDP-choline pathway for phosphatidylcholine biosynthesis [9]. The activities of the ¢rst two enzymes of the pathway, choline kinase and CTP:phosphocholine cytidylyltransferase (CCT), have been demonstrated in crude cell extracts, and cells of S. pneumoniae incorporate labelled choline into phosphocholine and CDP-choline [10,11]. The Gram-negative pathogen Haemophilus in£uenzae also produces cell surface phosphocholine, as a constituent of its lipooligosaccharide. The lic1 operon in H. in£uenzae was identi¢ed as functioning in the pathway for addition of phosphocholine to the lipooligosaccharide [12]. A similar operon is found in S. pneumoniae [13]. One of the genes in these operons, licC, has been postulated to encode the CCT [14], based upon its similarity with a large family of sugar-phosphate nucleotidyltransferases. However, the prokaryotic licC gene products have no primary structure similarity to the eukaryotic CCTs, lacking all the signature sequences [15], regulatory domains, and predicted secondary structure.2 In this paper, we report the cloning of the licC gene from S. pneumoniae and the identi¢cation of its product as a CTP:phosphocholine cytidylyltransferase. We have puri¢ed the enzyme and characterized it with regard to its steady state kinetics, substrate speci¢city, and interactions with metals. Detailed analysis of this enzyme will allow a greater understanding of a critical pathway in this important pathogen. 2. Materials and methods 2.1. Chemicals and reagents Oligonucleotide primers, ampicillin, isopropylthio-
2
Secondary structure was predicted and compared with known structures at www.bioimbgu.org.
galactoside, and dithiothreitol were obtained from Gibco-Life Technologies. New England Biolabs supplied Vent polymerase, all restriction enzymes and T4 DNA ligase. Calf intestinal alkaline phosphatase was purchased from Promega. Sigma was the source for CTP, phosphocholine, CDP-choline, S200 gel ¢ltration resin, murexide, and phenylmethylsulfonyl £uoride. Prepared chocolate and blood agar, as well as hemoglobin and IsoVitalex were from BBL. Powder components for Todd^Hewwitt broth and brain-heart infusion broth were from Difco. Qiagen supplied kits for both mini- and midi-scale puri¢cation of plasmid DNA, as well as a kit for the extraction of DNA from agarose gels. The pBluescript plasmid was from Stratagene, and the pET21b plasmid from Novagen. 14 C-labelled phosphocholine (ammonium salt), CDP-choline, and choline were from Amersham, as were unlabelled nucleoside triphosphates and deoxynucleoside triphosphates. 14 C-labelled phosphoethanolamine was from American Radiolabelled Chemicals. Chelex resin, Bradford dye reagent, and the prepacked Superdex 200 column were purchased from Bio-Rad. Trimethylaminoethylfractogel (TMAE-fractogel) resin was obtained from Novagen. The R36A S. pneumoniae strain was from the American Type Culture Collection. Standard laboratory chemicals were from Sigma or Fisher. 2.2. Cloning Genomic DNA was prepared [16] from the type 2 S. pneumoniae R36A for use as the template in PCR. Primers for the cloning of the licC gene were based on the preliminary sequence of a type 4 strain by the Institute for Genomic Research (www.tigr.org). The gene sequence was found in the TIGR database by searching for similarity to the H. in£uenzae licC gene [14]. The forward primer was 5P-C GGG ATC CAT ATG AAA GCC ATT ATC TTA G-3P (BamHI site in bold, NdeI site underlined). The reverse primer was 5P-GGC GAA TTC TTA ATT TTC GTT TTT AAG AAT-3P (EcoRI site underlined). PCR was performed with Vent polymerase, with an initial incubation of 1 min at 94³C, followed by 35 cycles of 1 min at 94³C, 1 min at 55³C, and 2 min at 72³C. A ¢nal elongation period of 10 min at 72³C completed the PCR. The insert was initially moved into pBluescript at the BamHI and EcoRI sites, followed by
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
transfer into pET21b at the NdeI and EcoRI sites by standard cloning procedures [17]. Propagation of plasmids was done in Escherichia coli DH5K cells. The inserts from four independent PCR reactions were sequenced in pBluescript with T3 and T7 primers by the University of Michigan Biomedical Research Core facilities. Additionally, the pET21b clone used for expression was sequenced using primers directed at the T7 promoter and T7 terminator regions. 2.3. Protein expression The licC gene product was expressed in E. coli BL21(DE3) cells from the pET21b construct. A single colony was inoculated into an overnight LB-ampicillin culture no larger than 25 ml.3 1 l cultures were inoculated 1:100 or 1:250 from the overnight culture, and grown to 0.6^0.8 OD600 . Isopropylthiogalactoside was added to 1 mM, and the culture continued for a further 2^3 h. Cell pellets were stored at 380³C prior to lysis. 2.4. Puri¢cation of S. pneumoniae CTP:phosphocholine cytidylyltransferase (SpCCT) All steps of protein puri¢cation were done on ice or at 4³C. Cell pellets from 1 l of culture were resuspended in 10 ml of 10 mM Tris pH 7.4, 100 mM NaCl, 2 mM dithiothreitol, and 1 mM ethylenediaminetetraacetic acid (EDTA) (bu¡er A). Protease inhibitor cocktail [18] and 1 mM phenylmethylsulfonyl were added immediately before the cells were lysed by two passes through a French press at 1100 psi. Lysate was cleared by centrifugation at 90 000Ug for 20 min. Ammonium sulfate was then added to 30% saturation at 4³C, and precipitate removed by centrifugation at 32 000Ug for 30 min. Up to 5 ml of the 30% ammonium sulfate supernatant was loaded onto a 150 ml Sephacryl-200 (S200) col3
It was critical that the initial, uninduced, overnight inoculum was no larger than 25 ml, otherwise large amounts of target protein were seen in the overnight culture, but subsequently, no overexpression occurred with addition of isopropylthiogalactoside. The cells were neither eliminating the plasmid nor excising or mutating the insert.
87
umn, running at 0.2 ml/min, that had been equilibrated with bu¡er A plus 25% ammonium sulfate, and 15 min fractions were collected. The ¢nal ammonium sulfate concentration was brought to 70% in the pooled peak fractions, and subjected to centrifugation. The supernatant containing the pure enzyme was stored at 320³C. Protein concentrations were determined by the Bradford assay [19], with bovine serum albumin as the standard. In two of the ¢ve enzyme preparations used in this paper, cleared lysate was ¢rst diluted 10-fold with bu¡er A lacking NaCl, and passed over a 0.5 ml TMAE-fractogel column. In those cases, £ow-through was concentrated for the subsequent steps. 2.5. Preparation of sodium phosphocholine A 0.5 M solution of commercial calcium phosphocholine was passed over a 50 g sodium-charged Chelex column. The resin was regenerated to the sodium form according to the manufacturer's protocol, and the phosphocholine fractions were again passed over the column. Fractions were assayed for calcium by a colorimetric assay with murexide [20]. For each sample, 200 Wl of undiluted sample was made alkaline by the addition of 5 Wl 1 N NaOH. 15 Wl of a fresh 2 mg/ ml solution of murexide was added, and color observed within 10 min. Total phosphate was assayed according to the method of Ames [21]. Pooled fractions were lyophilized overnight, and the resultant powder was stored at 320³C. A portion was dissolved and tested again for calcium and total phosphate. 2.6. Enzyme assay and kinetic analysis Standard SpCCT assays were performed routinely by a modi¢cation of a charcoal-binding assay [22]. Standard assay conditions were 20 mM Tris, pH 8.0, 6 mM sodium phosphocholine (1^4% (v/v) of ¢nal assay was from 55 mmol/mCi [14 C]phosphocholineNH4 ), 6 mM CTP, and 9 mM magnesium acetate in a 100 Wl ¢nal volume. Ammonium sulfate was kept at a ¢nal concentration of 7%. No lipid was included in the standard assay. Bu¡er and substrates were prewarmed at 37³C for 1^5 min, followed by initiation of the reaction by addition of 2 Wg enzyme. Incubation proceeded for 30 s, followed by immedi-
BBAMCB 55834 17-12-01
88
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
ate addition of 10% trichloroacetic acid, 150 mM calcium phosphocholine to stop the reaction. Samples were then placed on ice and 0.5 ml of 10 mg/ml charcoal was added. Samples were incubated with charcoal at 4³C from 15 min to overnight. Charcoal
was pelleted by spinning at maximum speed in a microcentrifuge for 10 min at room temperature. Samples were washed twice with 0.6 ml water, then 0.5 ml of 10% acetic acid was added before transfer to 4.6 ml of scintillation £uid. The scintillation vials were mixed, and then centrifuged at 4000Ug for 10 min prior to counting. In some cases thin layer chromatography was used instead of charcoal to resolve the assay products. In that case, assay conditions were identical to the charcoal-binding assay, except that the reaction was terminated by the addition of 100 mM EDTA [23]. 30 Wl of sample was spotted on a silica gel G plate and resolved in methanol:0.5% NaCl:ammonium hydroxide 50:50:1. Samples were compared to 14 C-labelled choline, phosphocholine, and CDP-choline standards by autoradiography and liquid scintillation counting of the scraped silica. Primary kinetic data were ¢t to the Michaelis^ Menten^Henri equation using Kaleidagraph (Synergy Software). Apparent Vmax values were determined while varying one substrate and holding the other at several ¢xed concentrations. Activity was determined with phosphocholine at 0.1, 0.5, 1, 2, 4, 6, 8, and 10 mM, and CTP at 0.1, 0.5, 1, 2, and 4 mM. These apparent Vmax values were used in secondary plots, that were again ¢t to the Michaelis^ Menten^Henri equation to determine true Vmax and 6
Fig. 1. Protein and DNA alignments of S. pneumoniae CCT. (A) Full length sequences of SpCCT from strain R36A (this paper) and a type 4 strain (TIGR) are aligned with the entire sequence of H. in£uenzae licC (Hi licC), residues 74^304 of Treponema pallidum licC-licA fusion protein (Tp licCA), the entire length of a putative glucose-1-phosphate cytidylyltransferase from Archaeoglobus fulgidus (Af rfbF), and residues 1^57 of E. coli GlmU (Ec GlmU). This alignment was performed by Clustal with the gap penalty set to 20. Gray boxes highlight residues that are conserved among all the licCs and one or both of the sugar nucleotidyltransferases. Black boxes indicate those residues shared amongst the licCs. The one di¡erence at the protein level between the two S. pneumoniae strains is boxed. The following residues were considered equivalent when highlighting conserved residues: R = K, E = D, S = T, V = I = L, and F = Y = W. The ATG start site is indicated by an asterisk. (B) Nine genomic di¡erences occur between the nucleotide sequences from S. pneumoniae R36A (this paper) and type 4 (TIGR). Only the C to T change at position 528 (asterisk) results in an amino acid di¡erence. Four independent PCR products were cloned and sequenced to verify this result.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
89
KM values. IC50 values were determined by ¢tting % activity values to the equation y = 1003(%Vmax U[I])/ (IC50 +[I]), again in Kaleidagraph. In all ¢gures, error bars represent the range observed for duplicate data points. Replicate experiments were performed with independent preparations of enzyme. 2.7. Culture of S. pneumoniae S. pneumoniae was routinely cultured on modi¢ed chocolate agar, prepared with 7.5 g acid hydrolysate of casein, 7.5 g bactopeptone, 1.0 g corn starch, 4.0 g K2 HPO4 , 1.0 g KH2 PO4 , 5.0 g NaCl, 12 g agar, 10 g hemoglobin, and 10 ml IsoVitalex. Liquid cultures were grown in a semi-de¢ned medium [24], and incubated aerobically at 37³C with no agitation. 3. Results 3.1. Cloning of S. pneumoniae licC The licC gene was cloned from a type 2 strain of S. pneumoniae, R36A. As a starting point for the cloning, we searched the incomplete genome of type 4 S. pneumoniae with the sequence of licC from H. in£uenzae. Upon inspection, two possible start sites were present in the type 4 S. pneumoniae licC gene. A standard ATG start site begins with nucleotides ATG ACT GAA AAT, encoding amino acids MTENT. 48 nucleotides upstream and in-frame with the rest of the gene is a possible GTG start site [25], beginning GTG AAA GCC ATT, encoding MKAIT (Fig. 1). We reasoned that GTG was the correct start site, as its usage would allow the inclusion of residues that are highly conserved among licC gene products and the family of sugar-phosphate nucleotidyltransferases, where some of these residues are critical for catalysis [26]. We have, nevertheless, examined each of the possible start sites by cloning and expressing the protein products. As is detailed below, the product of the gene that begins with the GTG site was soluble and active. The gene product from the ATG start site was expressed well, but was neither soluble nor active (data not shown). Preliminary results using two polyclonal antibodies raised in rabbits against puri¢ed SpCCT showed a single im-
Fig. 2. Analysis of puri¢cation steps by SDS^gel electrophoresis. Protein was overexpressed in E. coli, and puri¢ed by chromatography on a Sephadex 200 column, followed by 70% ammonium sulfate precipitation. The S200 column was run in a bu¡er with 25% ammonium sulfate, necessitating a preliminary 30% ammonium sulfate cut. A 12% denaturing gel was run and stained with Coomassie brilliant blue R-250. Five separate enzyme preparations gave similar results.
munoreactive band in S. pneumoniae extracts that comigrated with recombinant SpCCT from the GTG start site (data not shown). We concluded, therefore, that the GTG site was much more likely to be the true start site for translation of this gene. Sequencing of the cloned PCR product of the type 2 licC gene revealed several di¡erences from the type 4 sequence in the TIGR database. There were eight di¡erences in the DNA sequence that did not result in a change of the amino acid sequence, and there was one non-silent change: valine-177 in the type 4 sequence, alanine-177 in R36A (Fig. 1). 3.2. Expression and puri¢cation of SpCCT The licC gene from type 2 S. pneumoniae was expressed from the pET21b plasmid in BL21(DE3) E. coli. The gene was highly expressed, resulting in target protein levels that constituted a major portion of total soluble protein (Fig. 2). The speci¢c activity of recombinant SpCCT in crude E. coli extracts was
BBAMCB 55834 17-12-01
90
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
19.3 Wmol/min/mg protein (Table 1), whereas the enzyme activity in crude extracts of cells of S. pneumoniae was 49.8 þ 3.2 (n = 4) nmol/min/mg protein. Puri¢cation of SpCCT was accomplished by a combination of gel ¢ltration and ammonium sulfate precipitation. The enzyme was not only highly soluble in ammonium sulfate, it appeared to be stabilized by the salt. For this reason the ¢rst step of puri¢cation was a 30% ammonium sulfate precipitation, so that gel ¢ltration on Sephacryl-200 could be run in the presence of 25% ammonium sulfate. The active fractions from the gel ¢ltration column were then subjected to 70% ammonium sulfate precipitation, with the enzyme remaining soluble. The enzyme preparation was judged homogeneous by sodium dodecyl sulfate (SDS)^polyacrylamide gel electrophoresis (Fig. 2). Chromatography on TMAE-fractogel initially appeared to be a useful puri¢cation step, and was included as a ¢rst step in two of the ¢ve enzyme preparations used in this paper. It proved to bind SpCCT only very weakly, however, and most enzyme was in the £ow-through, with no increase in the speci¢c activity (data not shown). Stored in 70% ammonium sulfate at 0.2^0.6 mg/ml at 320³C, the enzyme was stable with respect to solubility and activity for at least 8 months. 3.3. Native size of SpCCT Chromatography on the Sephacryl-200 resin could not be used in size estimation of the enzyme. The enzyme eluted as if it were 8 kDa in size when compared with standards of known size, indicating that the enzyme probably binds to this resin. Gel ¢ltration on a Superdex 200 FPLC column revealed a
single peak eluting at 25 400 þ 1000 Da, similar to the calculated size of 26 874 Da for a monomeric protein. 3.4. Enzymatic assay of SpCCT Several assay conditions including assay time, pH, and protein, Mg2 , and ammonium sulfate concentrations were optimized. It was determined that the activity was linear with protein up to about 10 Wg of pure enzyme (not shown) and with assay time for less than 1 min (Fig. 3A). Because the enzyme was extremely stable under storage conditions, was stable at 0.2^0.6 mg/ml in 70% ammonium sulfate for at least 16 h at 37³C, and retained almost half its activity after 2 weeks at room temperature, it was surprising that the enzyme lost activity so rapidly under assay conditions. Various additives were screened for ability to stabilize the enzyme in the assay, but none were successful. These included bovine serum albumin, glycerol, dithiothreitol, betaine, choline, glycine, proline, PEG-6000, PEG-20 000, Ficoll-400 [27], and phosphatidylcholine:oleate [28]. Enzyme instability was observed across the pH spectrum. The enzyme was also assayed at 0³C, where it had approximately half the activity seen at 37³C, but still exhibited the loss of activity under assay conditions with the same time scale as at 37³C. Reaction products from both temperatures were shown to comigrate with authentic CDP-choline standards by thin layer chromatography. Pre-incubation of the enzyme under assay conditions with one of the substrates or Mg2 neither stabilized nor caused any further loss of activity in the assay, relative to that seen without pre-incubation (data not shown). Therefore it appears that the loss of activity is related to dilution of the pure
Table 1 Puri¢cation of SpCCT
Cleared lysate 30% AS sup. S200 fractions 70% AS sup.
Protein (mg)
Enzyme units (Wmol/min)
Speci¢c activity (Wmol/min/mg)
% Protein
% Activity
Fold puri¢cation
167 121 51 54
3.22U103 2.87U103 1.89U103 2.01U103
19.3 23.8 37.2 37.0
100 72 30 32
100 89 59 62
1.0 1.2 1.9 1.9
Protein and activity were followed throughout the course of the puri¢cation. Protein was measured by the Bradford assay, and activity by the charcoal-binding assay. Standard assay conditions as reported in Section 2 were used. These experiments were performed a total of ¢ve times.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
91
Fig. 3. Assay optimization. (A) SpCCT is linear with time for about 60 s. 2 Wg of puri¢ed enzyme were used in the assay shown, with 6 mM CTP, 6 mM phosphocholine, and 9 mM magnesium acetate. This experiment was performed eight times. (B) SpCCT has negligible activity at pH 5.5 and below, intermediate activity at pH 6.0^6.5, and full activity at pH 7.0 and higher. This experiment was performed three times with independent enzyme preparations. (C) Mg2 in slight excess of CTP gives maximal activity. Closed circles show a titration of Mg2 into assays with 6 mM phosphocholine and 1 mM CTP. Open circles represent assays with 6 mM phosphocholine and 5 mM CTP. This experiment was performed twice. (D) Ammonium sulfate activates SpCCT. Ammonium sulfate was removed from the enzyme by dialysis in 100 mM NaCl, 10 mM Tris pH 7.4, 2 mM DTT, and 1 mM EDTA. Ammonium sulfate was then titrated into assays. This experiment has been performed three times.
enzyme and/or to the turnover process. This rapid loss of activity with time did not occur with either crude overexpressed enzyme or native enzyme from S. pneumoniae cells. SpCCT was maximally active in neutral and alkaline conditions (Fig. 3B). Standard assays were therefore conducted at pH 8. The ratio of CTP to Mg2 was examined, and determined to be optimal with a 3 mM excess of Mg2 over CTP (Fig. 3C). Preliminary evidence indicates that Mg2 at 20 mM or higher in an otherwise standard assay is inhibitory (not shown). Because the enzyme is routinely stored in 70% ammonium sulfate, some ammonium sulfate is necessarily carried over into the assay. Maximal SpCCT activity occurred with a ¢nal concentration of 5^
10% ammonium sulfate in the assay, with no activity observed when the ¢nal ammonium sulfate was v 30% (Fig. 3D). Ammonium sulfate did not a¡ect the charcoal-binding assay with respect to assay background or CDP-choline recovery, nor did it affect enzyme stability. As expected, lipids had no effect on this enzyme. 3.5. Kinetic constants Kinetic parameters were determined with three different pure preparations of SpCCT. Data were collected at several ¢xed concentrations of one substrate while the other substrate was varied, and the true KM and Vmax values were derived from secondary plots [29]. Titration data were collected with up to 10 mM
BBAMCB 55834 17-12-01
92
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
Fig. 4. Enzyme kinetics. Kinetic assays were done with three independent puri¢cations of enzyme, with similar results. A representative assay is shown. (A) CTP was titrated while phosphocholine was held steady. Concentrations of phosphocholine were: b 0.1 mM, F 0.5 mM, R 1.0 mM, S 2 mM, a 4 mM, E 6 mM, O 8 mM, and P 10 mM. (B) Transformation of the data to show phosphocholine titration while CTP is held steady. CTP concentrations were: b 0.1 mM, F 0.5 mM, R 1.0 mM, S 2 mM, a 4 mM, E 6 mM, O 8 mM, and P 10 mM.
of each substrate (Fig. 4). CTP was inhibitory, however, at higher concentrations, so we included data only through 4 mM CTP for determining the kinetic constants. The KM values for CTP and phosphocholine were 0.89 þ 0.24 mM and 0.39 þ 0.17 mM, respectively. The Vmax value was 39.1 þ 8.9 Wmol/min/ mg protein, which corresponds to a kcat of 17.5 þ 4.0 s31 . The Vmax and KM values for both substrates were similar to those observed for recombinant rat [28] or Saccharomyces cerevisiae CCT [30]. The substrate speci¢city of the enzyme was examined. The pure enzyme utilized only CTP and dCTP e¡ectively as nucleotide substrates (Fig. 5), although the crude, overexpressed enzyme utilized other nucleoside triphosphates (data not shown). Perhaps the other nucleoside triphosphates served to prevent hydrolysis of endogenous CTP levels in the crude extracts, or nucleotide kinases were transferring phosphate groups to endogenous CMP and CDP. The enzyme was also able to utilize phosphoethanolamine, although much less e¡ectively than phosphocholine (not shown).
was used for initial kinetic analysis (not shown). This inhibition could be mimicked by addition of extra Ca2 , or relieved by the addition of EGTA (data not shown). The sodium salt of phosphocholine was prepared by passing the substrate over a sodium-charged Chelex column. The limits of detection for the murexide assay indicate that the remaining
3.6. Inhibition by Ca 2+ and other metal ions In the process of determining optimal assay conditions, it became apparent that SpCCT is inhibited by Ca2 . Phosphocholine is commercially available only as the calcium salt, and the calcium inhibition became apparent when the calcium phosphocholine
Fig. 5. Nucleotide speci¢city of SpCCT. Various nucleotide triphosphates at 10 mM were tested as SpCCT substrates. Only CTP and dCTP were found to support appreciable turnover. CTP* indicates CTP from Sigma. All others were obtained as 100 mM solutions from Amersham. This experiment was performed a total of three times.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
93
Fig. 6. Calcium inhibition of SpCCT. (A) CaCl2 inhibits puri¢ed, recombinant SpCCT with an IC50 of 0.32 þ 0.04 mM. This experiment was done three times. (B) CaCl2 inhibits the CCT activity of S. pneumoniae extracts with an IC50 of 0.27 þ 0.03 mM. This experiment was done twice. Standard assay conditions were used in both sets of experiments, except that assay time for S. pneumoniae extracts was extended to 60 min.
Ca2 was present at less than 1 ion Ca2 per 800 molecules phosphocholine. All data shown in this paper utilized the sodium phosphocholine. The sodium phosphocholine behaved as a standard substrate (Fig. 4B), with no inhibition apparent up to 50 mM (data not shown). Ca2 inhibition was observed with sodium phosphocholine as the substrate with pure (Fig. 6A) or crude recombinant enzyme (not shown), as well as in R36A lysates (Fig. 6B). IC50 values were identical within error for puri¢ed recombinant enzyme and S. pneumoniae extracts, 0.32 þ 0.04 and 0.27 þ 0.03 mM respectively. Other metal ions were tested for inhibition, activation, or stabilization of the enzyme. All tested divalent metals, other than Mg2 , severely inhibited SpCCT (Fig. 7). Tested monovalent cations, including Na , K , Li , and NH 4 either had no e¡ect on activity, or were activating (Fig. 3D, and data not shown). Ca2 does not a¡ect assay background or CDP-choline recovery. 4. Discussion This report con¢rms the presence of a CTP:phosphocholine cytidylyltransferase in extracts of S. pneumoniae, and establishes that the gene encoding this enzyme is gene C of the lic operon. Like its eukaryotic counterparts, SpCCT is speci¢c for CTP or deoxyCTP. However, there is no obvious similarity between the sequences of SpCCT and the eukaryotic
CCTs. Instead, licC CCT is a member of the GlmU/RmlA nucleotidyltransferase family which consists primarily of sugar phosphate nucleotidyltransferases. Members of this family for which three-dimensional structures have been determined include N-acetylglucosaminyl-1-phosphate uridylyltransferase (GlmU) from E. coli [31] and S. pneumoniae [26], and glucose-1-phosphate thymidylyltransferase (RmlA) of Pseudomonas aeruginosa [32]. The GlmU/RmlA family has a consensus sequence, GXG(T/S)RX4 PK [26]. In SpCCT, the corresponding sequence is 9 GLGTRX8 PK. The conservation of these residues and their importance for activity of GlmU supports the use of the GTG start site rather than the downstream ATG. S. pneumoniae R36A contains an alanine residue at position 177, whereas the type 4 strain sequenced by TIGR contains a valine residue. This strain di¡erence may prove to be important. For example, an alanine to valine mutation is found as a polymorphism in human methyltetrahydrofolate reductase. The analogous mutation in the E. coli enzyme resulted in a tilt of an entire alpha-helix, increased propensity to lose the £avin cofactor, and a signi¢cantly less active enzyme under folate- or £avin-limiting conditions [33]. Based on the structure of RmlA and secondary structure predictions, the di¡erence in bulk at position 177 in SpCCT will be at the C-terminal end of a long helix (relative to RmlA), near the CTP-binding site and solvent accessibility site. In searching for the gene that encodes the SpCCT,
BBAMCB 55834 17-12-01
94
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
more than 50% at 1 mM (Fig. 7). Ca2 inhibition began at levels as low as 40 WM (Fig. 5). Total intracellular concentrations of elemental Ca, Mg, and Zn in E. coli are 14, 29, and 0.6 mM respectively [37]. Although metal storage systems in cells undoubtedly reduce the free metal levels, these levels are high enough to warrant the possibility of inhibition of SpCCT by metals. Ca2 is particularly interesting, given its role in the competence for transformation of S. pneumoniae [38], a process in which the presence of phosphocholine on the cell surface is intimately linked [39]. Because these metals inhibit at concentrations far below the Mg2 or CTP levels in the assay, it is assumed that this inhibition is not competitive with respect to the presumed Mg2 ^ CTP interaction. Fig. 7. All tested divalent cations inhibit SpCCT. Except for Fe, striped bars indicate assays in which 1 mM of indicated metal was added to a standard assay. Stippled bars show activity in presence of 5 mM of indicated metal in the standard assay. Solid bars show the level of turnover supported by 5 mM of the indicated metal in the absence of Mg2 . Fe was considered separately because under aerobic conditions, a mixture of Fe2 and Fe3 was present. The striped bar for Fe represents 1.25 mM Fe in a standard assay, and the stippled bar represents 6.25 mM Fe. The solid bar shows results with 6.25 mM Fe in the absence of Mg2 . These experiments were performed four times.
we also considered a gene, originally termed Sp232D by TIGR, that encodes a protein with limited similarity to eukaryotic CCTs. The sequence of the protein product of Sp232D has high similarity to eukaryotic CCTs in the HXGH region and some similarity in the RTEGIST sequence. We cloned this gene and expressed it, but it had no CCT activity. It is possible that this gene encodes a phosphopantetheine adenylyltransferase, due to its overall similarity to the kdtB gene of E. coli [34]. At least one strain of S. pneumoniae encodes a protein that is highly related to the eukaryotic CCTs, and probably encodes CTP:glycerol-3-phosphate cytidylyltransferase (GCT) [35], which has been used as a model for this cytidylyltransferase family [36]. The same locus encoding the probable GCT was sequenced in many strains, including type 2 and type 4, but the GCT homolog has been found only in type 18C. The metal inhibition may be physiologically relevant, because all tested metals inhibited activity by
Acknowledgements We would like to thank Joshua A. Newsted for work done with the N-terminal truncation of the enzyme, Jeanne Stuckey for structural modelling, Zhaohui Xu for use of the Superdex-200 chromatography system, and the Institute for Genomic Research for preliminary sequence data. Funding was provided by NIH grant RO1 GM60510 (C.K.), NIH training grant GM07767 (H.A.C.), and institutional NIH grants P60DK20572 and MO1-RR00042.
References [1] A. Tomasz, Science 157 (1967) 694^697. [2] H.J. Jennings, C. Lugowski, N.M. Young, Biochemistry 19 (1980) 4712^4719. [3] E. Tuomanen, Curr. Opin. Microbiol. 2 (1999) 35^39. [4] A. Tomasz, M. Westphal, E.B. Briles, P. Fletcher, J. Supramol. Struct. 3 (1975) 1^16. [5] E.S. Lysenko, J. Gould, R. Bals, J.M. Wilson, J.N. Weiser, Infect. Immun. 68 (2000) 1664^1671. [6] A. Severin, D. Horne, A. Tomasz, Microb. Drug Resist. 3 (1997) 391^400. [7] J. Yother, K. Leopold, J. White, W. Fischer, J. Bacteriol. 180 (1998) 2093^2101. [8] W. Fischer, Res. Microbiol. 151 (2000) 421^427. [9] E.P. Kennedy, S.B. Weiss, J. Biol. Chem. 222 (1956) 193^ 214. [10] I.R. Poxton, D.J. Leak, J. Gen. Microbiol. 100 (1977) 23^29. [11] B. Bean, A. Tomasz, J. Bacteriol. 130 (1977) 571^574.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95 [12] J.N. Weiser, J.M. Love, E.R. Moxon, Cell 59 (1989) 657^ 665. [13] J.R. Zhang, I. Idanpaan-Heikkila, W. Fischer, E.I. Tuomanen, Mol. Microbiol. 31 (1999) 1477^1488. [14] J.N. Weiser, M. Shchepetov, S.T. Chong, Infect. Immun. 65 (1997) 943^950. [15] C. Kent, Biochim. Biophys. Acta 1348 (1997) 79^90. [16] C.H. Collins, P.M. Lyne, J.M. Grange, Microbiological Methods, 7th Edn., Butterworths, London, 1995. [17] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, 2nd Edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. [18] J.I. MacDonald, C. Kent, Protein Expr. Purif. 4 (1993) 1^ 7. [19] M.M. Bradford, Anal. Biochem. 72 (1976) 248^254. [20] D.C. Harris, Quantitative Chemical Analysis, 4th Edn., Freeman, New York, 1995. [21] B.N. Ames, Methods Enzymol. 8 (1966) 115^118. [22] J.N. Morand, C. Kent, J. Biol. Chem. 264 (1989) 13785^ 13792. [23] A. Lykidis, K.G. Murti, S. Jackowski, J. Biol. Chem. 273 (1998) 14022^14029. [24] S. Lacks, R.D. Hotchkiss, Biochim. Biophys. Acta 39 (1960) 508^517. [25] S. Shinedling, M. Gayle, D. Pribnow, L. Gold, Mol. Gen. Genet. 207 (1987) 224^232. [26] G. Sulzenbacher, L. Gal, C. Pene¡, F. Fassy, Y. Bourne, J. Biol. Chem. 276 (2001) 11844^11851.
95
[27] F. Anjum, V. Rishi, F. Ahmad, Biochim. Biophys. Acta 1476 (2000) 75^84. [28] J.A. Friesen, H.A. Campbell, C. Kent, J. Biol. Chem. 274 (1999) 13384^13389. [29] I.H. Segel, Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems, Wiley, New York, 1975. [30] J.A. Friesen, Y. Seo Park, C. Kent, Protein Expr. Purif. 21 (2001) 141^148. [31] K. Brown, F. Pompeo, S. Dixon, D. Mengin-Lecreulx, C. Cambillau, Y. Bourne, EMBO J. 18 (1999) 4096^4107. [32] W. Blankenfeldt, M. Asuncion, J.S. Lam, J.H. Naismith, EMBO J. 19 (2000) 6652^6663. [33] B.D. Guenther, C.A. Sheppard, P. Tran, R. Rozen, R.G. Matthews, M.L. Ludwig, Nat. Struct. Biol. 6 (1999) 359^ 365. [34] A. Geerlof, A. Lewendon, W.V. Shaw, J. Biol. Chem. 274 (1999) 27105^27111. [35] S.M. Jiang, L. Wang, P.R. Reeves, Infect. Immun. 69 (2001) 1244^1255. [36] Y.S. Park, T.D. Sweitzer, J.E. Dixon, C. Kent, J. Biol. Chem. 268 (1993) 16648^16654. [37] F.C. Kung, J. Raymond, D.A. Glaser, J. Bacteriol. 126 (1976) 1089^1095. [38] H. Seto, A. Tomasz, J. Bacteriol. 126 (1976) 1113^1118. [39] A. Tomasz, Proc. Natl. Acad. Sci. USA 59 (1968) 86^93.
BBAMCB 55834 17-12-01
The CTP:phosphocholine cytidylyltransferase encoded by the licC gene of Streptococcus pneumoniae: cloning, expression, puri¢cation, and characterization1 Heidi A. Campbell, Claudia Kent * Department of Biological Chemistry, 4417 Medical Science I, University of Michigan Medical Center, Ann Arbor, MI 48109-0606, USA Received 1 August 2001; received in revised form 13 September 2001; accepted 13 September 2001
Abstract Streptococcus pneumoniae is a member of a small group of bacteria that display phosphocholine on the cell surface, covalently attached to the sugar groups of teichoic acid and lipoteichoic acid. The putative pathway for this phosphocholine decoration is, in its first two enzymes, functionally similar to the CDP-choline pathway used for phosphatidylcholine biosynthesis in eukaryotes. We show that the licC gene encodes a functional CTP:phosphocholine cytidylyltransferase (CCT). The enzyme has been expressed and purified to homogeneity. Assay conditions were optimized, particularly with respect to linearity with time, pH, Mg2 , and ammonium sulfate concentration. The pure enzyme has KM values of 890 þ 240 WM for CTP, and 390 þ 170 WM for phosphocholine. The kcat is 17.5 þ 4.0 s31 . S. pneumoniae CTP :phosphocholine cytidylyltransferase (SpCCT) is specific for CTP or dCTP as the nucleotide substrate. SpCCT is strongly inhibited by Ca2 . The IC50 values for recombinant and native SpCCT are 0.32 þ 0.04 and 0.27 þ 0.03 mM respectively. The enzyme is also inhibited by all other tested divalent cations, including Mg2 at high concentrations. The cloning and expression of this enzyme sets the stage for design of inhibitors as possible antipneumococcal drugs. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Cytidylyltransferase; licC; Phosphocholine; CDP-choline; Teichoic acid; Streptococcus pneumoniae
1. Introduction The pathogen Streptococcus pneumoniae contains
Abbreviations: CCT, CTP:phosphocholine cytidylyltransferase; SpCCT, Streptococcus pneumoniae CTP:phosphocholine cytidylyltransferase; GCT, glycerol-3-phosphate cytidylyltransferase; TMAE, trimethylaminoethyl ; EDTA, ethylenediaminetetraacetic acid * Corresponding author. Fax: +1-734-763-4581. E-mail address: [email protected] (C. Kent). 1 The sequence described herein has been deposited in GenBank with accession number AF402777.
phosphocholine as a constituent of its teichoic and lipoteichoic acid, with the phosphocholine attached to one or both of the N-acetylgalactosamine residues of the pentasaccharide repeating unit [1,2]. This cell surface phosphocholine serves as an anchor for a number of choline-binding proteins [3], as an anchor and activator of both endogenous and bacteriophage autolysins [4], and may interact with the host immune system [5]. Although S. pneumoniae is a choline auxotroph, mutant strains have been isolated that are capable of growing without choline or the similar amino alcohol, ethanolamine [6,7]. It has therefore been postulated that S. pneumoniae is auxotrophic
1388-1981 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 8 - 1 9 8 1 ( 0 1 ) 0 0 1 7 4 - 3
BBAMCB 55834 17-12-01
86
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
for choline, not because any of the roles for cell surface phosphocholine are individually or collectively indispensable, but because there is a checkpoint in the assembly of teichoic and lipoteichoic acid for the presence of phosphocholine [8]. The S. pneumoniae pathway for choline phosphorylation and activation may be similar to the eukaryotic CDP-choline pathway for phosphatidylcholine biosynthesis [9]. The activities of the ¢rst two enzymes of the pathway, choline kinase and CTP:phosphocholine cytidylyltransferase (CCT), have been demonstrated in crude cell extracts, and cells of S. pneumoniae incorporate labelled choline into phosphocholine and CDP-choline [10,11]. The Gram-negative pathogen Haemophilus in£uenzae also produces cell surface phosphocholine, as a constituent of its lipooligosaccharide. The lic1 operon in H. in£uenzae was identi¢ed as functioning in the pathway for addition of phosphocholine to the lipooligosaccharide [12]. A similar operon is found in S. pneumoniae [13]. One of the genes in these operons, licC, has been postulated to encode the CCT [14], based upon its similarity with a large family of sugar-phosphate nucleotidyltransferases. However, the prokaryotic licC gene products have no primary structure similarity to the eukaryotic CCTs, lacking all the signature sequences [15], regulatory domains, and predicted secondary structure.2 In this paper, we report the cloning of the licC gene from S. pneumoniae and the identi¢cation of its product as a CTP:phosphocholine cytidylyltransferase. We have puri¢ed the enzyme and characterized it with regard to its steady state kinetics, substrate speci¢city, and interactions with metals. Detailed analysis of this enzyme will allow a greater understanding of a critical pathway in this important pathogen. 2. Materials and methods 2.1. Chemicals and reagents Oligonucleotide primers, ampicillin, isopropylthio-
2
Secondary structure was predicted and compared with known structures at www.bioimbgu.org.
galactoside, and dithiothreitol were obtained from Gibco-Life Technologies. New England Biolabs supplied Vent polymerase, all restriction enzymes and T4 DNA ligase. Calf intestinal alkaline phosphatase was purchased from Promega. Sigma was the source for CTP, phosphocholine, CDP-choline, S200 gel ¢ltration resin, murexide, and phenylmethylsulfonyl £uoride. Prepared chocolate and blood agar, as well as hemoglobin and IsoVitalex were from BBL. Powder components for Todd^Hewwitt broth and brain-heart infusion broth were from Difco. Qiagen supplied kits for both mini- and midi-scale puri¢cation of plasmid DNA, as well as a kit for the extraction of DNA from agarose gels. The pBluescript plasmid was from Stratagene, and the pET21b plasmid from Novagen. 14 C-labelled phosphocholine (ammonium salt), CDP-choline, and choline were from Amersham, as were unlabelled nucleoside triphosphates and deoxynucleoside triphosphates. 14 C-labelled phosphoethanolamine was from American Radiolabelled Chemicals. Chelex resin, Bradford dye reagent, and the prepacked Superdex 200 column were purchased from Bio-Rad. Trimethylaminoethylfractogel (TMAE-fractogel) resin was obtained from Novagen. The R36A S. pneumoniae strain was from the American Type Culture Collection. Standard laboratory chemicals were from Sigma or Fisher. 2.2. Cloning Genomic DNA was prepared [16] from the type 2 S. pneumoniae R36A for use as the template in PCR. Primers for the cloning of the licC gene were based on the preliminary sequence of a type 4 strain by the Institute for Genomic Research (www.tigr.org). The gene sequence was found in the TIGR database by searching for similarity to the H. in£uenzae licC gene [14]. The forward primer was 5P-C GGG ATC CAT ATG AAA GCC ATT ATC TTA G-3P (BamHI site in bold, NdeI site underlined). The reverse primer was 5P-GGC GAA TTC TTA ATT TTC GTT TTT AAG AAT-3P (EcoRI site underlined). PCR was performed with Vent polymerase, with an initial incubation of 1 min at 94³C, followed by 35 cycles of 1 min at 94³C, 1 min at 55³C, and 2 min at 72³C. A ¢nal elongation period of 10 min at 72³C completed the PCR. The insert was initially moved into pBluescript at the BamHI and EcoRI sites, followed by
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
transfer into pET21b at the NdeI and EcoRI sites by standard cloning procedures [17]. Propagation of plasmids was done in Escherichia coli DH5K cells. The inserts from four independent PCR reactions were sequenced in pBluescript with T3 and T7 primers by the University of Michigan Biomedical Research Core facilities. Additionally, the pET21b clone used for expression was sequenced using primers directed at the T7 promoter and T7 terminator regions. 2.3. Protein expression The licC gene product was expressed in E. coli BL21(DE3) cells from the pET21b construct. A single colony was inoculated into an overnight LB-ampicillin culture no larger than 25 ml.3 1 l cultures were inoculated 1:100 or 1:250 from the overnight culture, and grown to 0.6^0.8 OD600 . Isopropylthiogalactoside was added to 1 mM, and the culture continued for a further 2^3 h. Cell pellets were stored at 380³C prior to lysis. 2.4. Puri¢cation of S. pneumoniae CTP:phosphocholine cytidylyltransferase (SpCCT) All steps of protein puri¢cation were done on ice or at 4³C. Cell pellets from 1 l of culture were resuspended in 10 ml of 10 mM Tris pH 7.4, 100 mM NaCl, 2 mM dithiothreitol, and 1 mM ethylenediaminetetraacetic acid (EDTA) (bu¡er A). Protease inhibitor cocktail [18] and 1 mM phenylmethylsulfonyl were added immediately before the cells were lysed by two passes through a French press at 1100 psi. Lysate was cleared by centrifugation at 90 000Ug for 20 min. Ammonium sulfate was then added to 30% saturation at 4³C, and precipitate removed by centrifugation at 32 000Ug for 30 min. Up to 5 ml of the 30% ammonium sulfate supernatant was loaded onto a 150 ml Sephacryl-200 (S200) col3
It was critical that the initial, uninduced, overnight inoculum was no larger than 25 ml, otherwise large amounts of target protein were seen in the overnight culture, but subsequently, no overexpression occurred with addition of isopropylthiogalactoside. The cells were neither eliminating the plasmid nor excising or mutating the insert.
87
umn, running at 0.2 ml/min, that had been equilibrated with bu¡er A plus 25% ammonium sulfate, and 15 min fractions were collected. The ¢nal ammonium sulfate concentration was brought to 70% in the pooled peak fractions, and subjected to centrifugation. The supernatant containing the pure enzyme was stored at 320³C. Protein concentrations were determined by the Bradford assay [19], with bovine serum albumin as the standard. In two of the ¢ve enzyme preparations used in this paper, cleared lysate was ¢rst diluted 10-fold with bu¡er A lacking NaCl, and passed over a 0.5 ml TMAE-fractogel column. In those cases, £ow-through was concentrated for the subsequent steps. 2.5. Preparation of sodium phosphocholine A 0.5 M solution of commercial calcium phosphocholine was passed over a 50 g sodium-charged Chelex column. The resin was regenerated to the sodium form according to the manufacturer's protocol, and the phosphocholine fractions were again passed over the column. Fractions were assayed for calcium by a colorimetric assay with murexide [20]. For each sample, 200 Wl of undiluted sample was made alkaline by the addition of 5 Wl 1 N NaOH. 15 Wl of a fresh 2 mg/ ml solution of murexide was added, and color observed within 10 min. Total phosphate was assayed according to the method of Ames [21]. Pooled fractions were lyophilized overnight, and the resultant powder was stored at 320³C. A portion was dissolved and tested again for calcium and total phosphate. 2.6. Enzyme assay and kinetic analysis Standard SpCCT assays were performed routinely by a modi¢cation of a charcoal-binding assay [22]. Standard assay conditions were 20 mM Tris, pH 8.0, 6 mM sodium phosphocholine (1^4% (v/v) of ¢nal assay was from 55 mmol/mCi [14 C]phosphocholineNH4 ), 6 mM CTP, and 9 mM magnesium acetate in a 100 Wl ¢nal volume. Ammonium sulfate was kept at a ¢nal concentration of 7%. No lipid was included in the standard assay. Bu¡er and substrates were prewarmed at 37³C for 1^5 min, followed by initiation of the reaction by addition of 2 Wg enzyme. Incubation proceeded for 30 s, followed by immedi-
BBAMCB 55834 17-12-01
88
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
ate addition of 10% trichloroacetic acid, 150 mM calcium phosphocholine to stop the reaction. Samples were then placed on ice and 0.5 ml of 10 mg/ml charcoal was added. Samples were incubated with charcoal at 4³C from 15 min to overnight. Charcoal
was pelleted by spinning at maximum speed in a microcentrifuge for 10 min at room temperature. Samples were washed twice with 0.6 ml water, then 0.5 ml of 10% acetic acid was added before transfer to 4.6 ml of scintillation £uid. The scintillation vials were mixed, and then centrifuged at 4000Ug for 10 min prior to counting. In some cases thin layer chromatography was used instead of charcoal to resolve the assay products. In that case, assay conditions were identical to the charcoal-binding assay, except that the reaction was terminated by the addition of 100 mM EDTA [23]. 30 Wl of sample was spotted on a silica gel G plate and resolved in methanol:0.5% NaCl:ammonium hydroxide 50:50:1. Samples were compared to 14 C-labelled choline, phosphocholine, and CDP-choline standards by autoradiography and liquid scintillation counting of the scraped silica. Primary kinetic data were ¢t to the Michaelis^ Menten^Henri equation using Kaleidagraph (Synergy Software). Apparent Vmax values were determined while varying one substrate and holding the other at several ¢xed concentrations. Activity was determined with phosphocholine at 0.1, 0.5, 1, 2, 4, 6, 8, and 10 mM, and CTP at 0.1, 0.5, 1, 2, and 4 mM. These apparent Vmax values were used in secondary plots, that were again ¢t to the Michaelis^ Menten^Henri equation to determine true Vmax and 6
Fig. 1. Protein and DNA alignments of S. pneumoniae CCT. (A) Full length sequences of SpCCT from strain R36A (this paper) and a type 4 strain (TIGR) are aligned with the entire sequence of H. in£uenzae licC (Hi licC), residues 74^304 of Treponema pallidum licC-licA fusion protein (Tp licCA), the entire length of a putative glucose-1-phosphate cytidylyltransferase from Archaeoglobus fulgidus (Af rfbF), and residues 1^57 of E. coli GlmU (Ec GlmU). This alignment was performed by Clustal with the gap penalty set to 20. Gray boxes highlight residues that are conserved among all the licCs and one or both of the sugar nucleotidyltransferases. Black boxes indicate those residues shared amongst the licCs. The one di¡erence at the protein level between the two S. pneumoniae strains is boxed. The following residues were considered equivalent when highlighting conserved residues: R = K, E = D, S = T, V = I = L, and F = Y = W. The ATG start site is indicated by an asterisk. (B) Nine genomic di¡erences occur between the nucleotide sequences from S. pneumoniae R36A (this paper) and type 4 (TIGR). Only the C to T change at position 528 (asterisk) results in an amino acid di¡erence. Four independent PCR products were cloned and sequenced to verify this result.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
89
KM values. IC50 values were determined by ¢tting % activity values to the equation y = 1003(%Vmax U[I])/ (IC50 +[I]), again in Kaleidagraph. In all ¢gures, error bars represent the range observed for duplicate data points. Replicate experiments were performed with independent preparations of enzyme. 2.7. Culture of S. pneumoniae S. pneumoniae was routinely cultured on modi¢ed chocolate agar, prepared with 7.5 g acid hydrolysate of casein, 7.5 g bactopeptone, 1.0 g corn starch, 4.0 g K2 HPO4 , 1.0 g KH2 PO4 , 5.0 g NaCl, 12 g agar, 10 g hemoglobin, and 10 ml IsoVitalex. Liquid cultures were grown in a semi-de¢ned medium [24], and incubated aerobically at 37³C with no agitation. 3. Results 3.1. Cloning of S. pneumoniae licC The licC gene was cloned from a type 2 strain of S. pneumoniae, R36A. As a starting point for the cloning, we searched the incomplete genome of type 4 S. pneumoniae with the sequence of licC from H. in£uenzae. Upon inspection, two possible start sites were present in the type 4 S. pneumoniae licC gene. A standard ATG start site begins with nucleotides ATG ACT GAA AAT, encoding amino acids MTENT. 48 nucleotides upstream and in-frame with the rest of the gene is a possible GTG start site [25], beginning GTG AAA GCC ATT, encoding MKAIT (Fig. 1). We reasoned that GTG was the correct start site, as its usage would allow the inclusion of residues that are highly conserved among licC gene products and the family of sugar-phosphate nucleotidyltransferases, where some of these residues are critical for catalysis [26]. We have, nevertheless, examined each of the possible start sites by cloning and expressing the protein products. As is detailed below, the product of the gene that begins with the GTG site was soluble and active. The gene product from the ATG start site was expressed well, but was neither soluble nor active (data not shown). Preliminary results using two polyclonal antibodies raised in rabbits against puri¢ed SpCCT showed a single im-
Fig. 2. Analysis of puri¢cation steps by SDS^gel electrophoresis. Protein was overexpressed in E. coli, and puri¢ed by chromatography on a Sephadex 200 column, followed by 70% ammonium sulfate precipitation. The S200 column was run in a bu¡er with 25% ammonium sulfate, necessitating a preliminary 30% ammonium sulfate cut. A 12% denaturing gel was run and stained with Coomassie brilliant blue R-250. Five separate enzyme preparations gave similar results.
munoreactive band in S. pneumoniae extracts that comigrated with recombinant SpCCT from the GTG start site (data not shown). We concluded, therefore, that the GTG site was much more likely to be the true start site for translation of this gene. Sequencing of the cloned PCR product of the type 2 licC gene revealed several di¡erences from the type 4 sequence in the TIGR database. There were eight di¡erences in the DNA sequence that did not result in a change of the amino acid sequence, and there was one non-silent change: valine-177 in the type 4 sequence, alanine-177 in R36A (Fig. 1). 3.2. Expression and puri¢cation of SpCCT The licC gene from type 2 S. pneumoniae was expressed from the pET21b plasmid in BL21(DE3) E. coli. The gene was highly expressed, resulting in target protein levels that constituted a major portion of total soluble protein (Fig. 2). The speci¢c activity of recombinant SpCCT in crude E. coli extracts was
BBAMCB 55834 17-12-01
90
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
19.3 Wmol/min/mg protein (Table 1), whereas the enzyme activity in crude extracts of cells of S. pneumoniae was 49.8 þ 3.2 (n = 4) nmol/min/mg protein. Puri¢cation of SpCCT was accomplished by a combination of gel ¢ltration and ammonium sulfate precipitation. The enzyme was not only highly soluble in ammonium sulfate, it appeared to be stabilized by the salt. For this reason the ¢rst step of puri¢cation was a 30% ammonium sulfate precipitation, so that gel ¢ltration on Sephacryl-200 could be run in the presence of 25% ammonium sulfate. The active fractions from the gel ¢ltration column were then subjected to 70% ammonium sulfate precipitation, with the enzyme remaining soluble. The enzyme preparation was judged homogeneous by sodium dodecyl sulfate (SDS)^polyacrylamide gel electrophoresis (Fig. 2). Chromatography on TMAE-fractogel initially appeared to be a useful puri¢cation step, and was included as a ¢rst step in two of the ¢ve enzyme preparations used in this paper. It proved to bind SpCCT only very weakly, however, and most enzyme was in the £ow-through, with no increase in the speci¢c activity (data not shown). Stored in 70% ammonium sulfate at 0.2^0.6 mg/ml at 320³C, the enzyme was stable with respect to solubility and activity for at least 8 months. 3.3. Native size of SpCCT Chromatography on the Sephacryl-200 resin could not be used in size estimation of the enzyme. The enzyme eluted as if it were 8 kDa in size when compared with standards of known size, indicating that the enzyme probably binds to this resin. Gel ¢ltration on a Superdex 200 FPLC column revealed a
single peak eluting at 25 400 þ 1000 Da, similar to the calculated size of 26 874 Da for a monomeric protein. 3.4. Enzymatic assay of SpCCT Several assay conditions including assay time, pH, and protein, Mg2 , and ammonium sulfate concentrations were optimized. It was determined that the activity was linear with protein up to about 10 Wg of pure enzyme (not shown) and with assay time for less than 1 min (Fig. 3A). Because the enzyme was extremely stable under storage conditions, was stable at 0.2^0.6 mg/ml in 70% ammonium sulfate for at least 16 h at 37³C, and retained almost half its activity after 2 weeks at room temperature, it was surprising that the enzyme lost activity so rapidly under assay conditions. Various additives were screened for ability to stabilize the enzyme in the assay, but none were successful. These included bovine serum albumin, glycerol, dithiothreitol, betaine, choline, glycine, proline, PEG-6000, PEG-20 000, Ficoll-400 [27], and phosphatidylcholine:oleate [28]. Enzyme instability was observed across the pH spectrum. The enzyme was also assayed at 0³C, where it had approximately half the activity seen at 37³C, but still exhibited the loss of activity under assay conditions with the same time scale as at 37³C. Reaction products from both temperatures were shown to comigrate with authentic CDP-choline standards by thin layer chromatography. Pre-incubation of the enzyme under assay conditions with one of the substrates or Mg2 neither stabilized nor caused any further loss of activity in the assay, relative to that seen without pre-incubation (data not shown). Therefore it appears that the loss of activity is related to dilution of the pure
Table 1 Puri¢cation of SpCCT
Cleared lysate 30% AS sup. S200 fractions 70% AS sup.
Protein (mg)
Enzyme units (Wmol/min)
Speci¢c activity (Wmol/min/mg)
% Protein
% Activity
Fold puri¢cation
167 121 51 54
3.22U103 2.87U103 1.89U103 2.01U103
19.3 23.8 37.2 37.0
100 72 30 32
100 89 59 62
1.0 1.2 1.9 1.9
Protein and activity were followed throughout the course of the puri¢cation. Protein was measured by the Bradford assay, and activity by the charcoal-binding assay. Standard assay conditions as reported in Section 2 were used. These experiments were performed a total of ¢ve times.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
91
Fig. 3. Assay optimization. (A) SpCCT is linear with time for about 60 s. 2 Wg of puri¢ed enzyme were used in the assay shown, with 6 mM CTP, 6 mM phosphocholine, and 9 mM magnesium acetate. This experiment was performed eight times. (B) SpCCT has negligible activity at pH 5.5 and below, intermediate activity at pH 6.0^6.5, and full activity at pH 7.0 and higher. This experiment was performed three times with independent enzyme preparations. (C) Mg2 in slight excess of CTP gives maximal activity. Closed circles show a titration of Mg2 into assays with 6 mM phosphocholine and 1 mM CTP. Open circles represent assays with 6 mM phosphocholine and 5 mM CTP. This experiment was performed twice. (D) Ammonium sulfate activates SpCCT. Ammonium sulfate was removed from the enzyme by dialysis in 100 mM NaCl, 10 mM Tris pH 7.4, 2 mM DTT, and 1 mM EDTA. Ammonium sulfate was then titrated into assays. This experiment has been performed three times.
enzyme and/or to the turnover process. This rapid loss of activity with time did not occur with either crude overexpressed enzyme or native enzyme from S. pneumoniae cells. SpCCT was maximally active in neutral and alkaline conditions (Fig. 3B). Standard assays were therefore conducted at pH 8. The ratio of CTP to Mg2 was examined, and determined to be optimal with a 3 mM excess of Mg2 over CTP (Fig. 3C). Preliminary evidence indicates that Mg2 at 20 mM or higher in an otherwise standard assay is inhibitory (not shown). Because the enzyme is routinely stored in 70% ammonium sulfate, some ammonium sulfate is necessarily carried over into the assay. Maximal SpCCT activity occurred with a ¢nal concentration of 5^
10% ammonium sulfate in the assay, with no activity observed when the ¢nal ammonium sulfate was v 30% (Fig. 3D). Ammonium sulfate did not a¡ect the charcoal-binding assay with respect to assay background or CDP-choline recovery, nor did it affect enzyme stability. As expected, lipids had no effect on this enzyme. 3.5. Kinetic constants Kinetic parameters were determined with three different pure preparations of SpCCT. Data were collected at several ¢xed concentrations of one substrate while the other substrate was varied, and the true KM and Vmax values were derived from secondary plots [29]. Titration data were collected with up to 10 mM
BBAMCB 55834 17-12-01
92
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
Fig. 4. Enzyme kinetics. Kinetic assays were done with three independent puri¢cations of enzyme, with similar results. A representative assay is shown. (A) CTP was titrated while phosphocholine was held steady. Concentrations of phosphocholine were: b 0.1 mM, F 0.5 mM, R 1.0 mM, S 2 mM, a 4 mM, E 6 mM, O 8 mM, and P 10 mM. (B) Transformation of the data to show phosphocholine titration while CTP is held steady. CTP concentrations were: b 0.1 mM, F 0.5 mM, R 1.0 mM, S 2 mM, a 4 mM, E 6 mM, O 8 mM, and P 10 mM.
of each substrate (Fig. 4). CTP was inhibitory, however, at higher concentrations, so we included data only through 4 mM CTP for determining the kinetic constants. The KM values for CTP and phosphocholine were 0.89 þ 0.24 mM and 0.39 þ 0.17 mM, respectively. The Vmax value was 39.1 þ 8.9 Wmol/min/ mg protein, which corresponds to a kcat of 17.5 þ 4.0 s31 . The Vmax and KM values for both substrates were similar to those observed for recombinant rat [28] or Saccharomyces cerevisiae CCT [30]. The substrate speci¢city of the enzyme was examined. The pure enzyme utilized only CTP and dCTP e¡ectively as nucleotide substrates (Fig. 5), although the crude, overexpressed enzyme utilized other nucleoside triphosphates (data not shown). Perhaps the other nucleoside triphosphates served to prevent hydrolysis of endogenous CTP levels in the crude extracts, or nucleotide kinases were transferring phosphate groups to endogenous CMP and CDP. The enzyme was also able to utilize phosphoethanolamine, although much less e¡ectively than phosphocholine (not shown).
was used for initial kinetic analysis (not shown). This inhibition could be mimicked by addition of extra Ca2 , or relieved by the addition of EGTA (data not shown). The sodium salt of phosphocholine was prepared by passing the substrate over a sodium-charged Chelex column. The limits of detection for the murexide assay indicate that the remaining
3.6. Inhibition by Ca 2+ and other metal ions In the process of determining optimal assay conditions, it became apparent that SpCCT is inhibited by Ca2 . Phosphocholine is commercially available only as the calcium salt, and the calcium inhibition became apparent when the calcium phosphocholine
Fig. 5. Nucleotide speci¢city of SpCCT. Various nucleotide triphosphates at 10 mM were tested as SpCCT substrates. Only CTP and dCTP were found to support appreciable turnover. CTP* indicates CTP from Sigma. All others were obtained as 100 mM solutions from Amersham. This experiment was performed a total of three times.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
93
Fig. 6. Calcium inhibition of SpCCT. (A) CaCl2 inhibits puri¢ed, recombinant SpCCT with an IC50 of 0.32 þ 0.04 mM. This experiment was done three times. (B) CaCl2 inhibits the CCT activity of S. pneumoniae extracts with an IC50 of 0.27 þ 0.03 mM. This experiment was done twice. Standard assay conditions were used in both sets of experiments, except that assay time for S. pneumoniae extracts was extended to 60 min.
Ca2 was present at less than 1 ion Ca2 per 800 molecules phosphocholine. All data shown in this paper utilized the sodium phosphocholine. The sodium phosphocholine behaved as a standard substrate (Fig. 4B), with no inhibition apparent up to 50 mM (data not shown). Ca2 inhibition was observed with sodium phosphocholine as the substrate with pure (Fig. 6A) or crude recombinant enzyme (not shown), as well as in R36A lysates (Fig. 6B). IC50 values were identical within error for puri¢ed recombinant enzyme and S. pneumoniae extracts, 0.32 þ 0.04 and 0.27 þ 0.03 mM respectively. Other metal ions were tested for inhibition, activation, or stabilization of the enzyme. All tested divalent metals, other than Mg2 , severely inhibited SpCCT (Fig. 7). Tested monovalent cations, including Na , K , Li , and NH 4 either had no e¡ect on activity, or were activating (Fig. 3D, and data not shown). Ca2 does not a¡ect assay background or CDP-choline recovery. 4. Discussion This report con¢rms the presence of a CTP:phosphocholine cytidylyltransferase in extracts of S. pneumoniae, and establishes that the gene encoding this enzyme is gene C of the lic operon. Like its eukaryotic counterparts, SpCCT is speci¢c for CTP or deoxyCTP. However, there is no obvious similarity between the sequences of SpCCT and the eukaryotic
CCTs. Instead, licC CCT is a member of the GlmU/RmlA nucleotidyltransferase family which consists primarily of sugar phosphate nucleotidyltransferases. Members of this family for which three-dimensional structures have been determined include N-acetylglucosaminyl-1-phosphate uridylyltransferase (GlmU) from E. coli [31] and S. pneumoniae [26], and glucose-1-phosphate thymidylyltransferase (RmlA) of Pseudomonas aeruginosa [32]. The GlmU/RmlA family has a consensus sequence, GXG(T/S)RX4 PK [26]. In SpCCT, the corresponding sequence is 9 GLGTRX8 PK. The conservation of these residues and their importance for activity of GlmU supports the use of the GTG start site rather than the downstream ATG. S. pneumoniae R36A contains an alanine residue at position 177, whereas the type 4 strain sequenced by TIGR contains a valine residue. This strain di¡erence may prove to be important. For example, an alanine to valine mutation is found as a polymorphism in human methyltetrahydrofolate reductase. The analogous mutation in the E. coli enzyme resulted in a tilt of an entire alpha-helix, increased propensity to lose the £avin cofactor, and a signi¢cantly less active enzyme under folate- or £avin-limiting conditions [33]. Based on the structure of RmlA and secondary structure predictions, the di¡erence in bulk at position 177 in SpCCT will be at the C-terminal end of a long helix (relative to RmlA), near the CTP-binding site and solvent accessibility site. In searching for the gene that encodes the SpCCT,
BBAMCB 55834 17-12-01
94
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95
more than 50% at 1 mM (Fig. 7). Ca2 inhibition began at levels as low as 40 WM (Fig. 5). Total intracellular concentrations of elemental Ca, Mg, and Zn in E. coli are 14, 29, and 0.6 mM respectively [37]. Although metal storage systems in cells undoubtedly reduce the free metal levels, these levels are high enough to warrant the possibility of inhibition of SpCCT by metals. Ca2 is particularly interesting, given its role in the competence for transformation of S. pneumoniae [38], a process in which the presence of phosphocholine on the cell surface is intimately linked [39]. Because these metals inhibit at concentrations far below the Mg2 or CTP levels in the assay, it is assumed that this inhibition is not competitive with respect to the presumed Mg2 ^ CTP interaction. Fig. 7. All tested divalent cations inhibit SpCCT. Except for Fe, striped bars indicate assays in which 1 mM of indicated metal was added to a standard assay. Stippled bars show activity in presence of 5 mM of indicated metal in the standard assay. Solid bars show the level of turnover supported by 5 mM of the indicated metal in the absence of Mg2 . Fe was considered separately because under aerobic conditions, a mixture of Fe2 and Fe3 was present. The striped bar for Fe represents 1.25 mM Fe in a standard assay, and the stippled bar represents 6.25 mM Fe. The solid bar shows results with 6.25 mM Fe in the absence of Mg2 . These experiments were performed four times.
we also considered a gene, originally termed Sp232D by TIGR, that encodes a protein with limited similarity to eukaryotic CCTs. The sequence of the protein product of Sp232D has high similarity to eukaryotic CCTs in the HXGH region and some similarity in the RTEGIST sequence. We cloned this gene and expressed it, but it had no CCT activity. It is possible that this gene encodes a phosphopantetheine adenylyltransferase, due to its overall similarity to the kdtB gene of E. coli [34]. At least one strain of S. pneumoniae encodes a protein that is highly related to the eukaryotic CCTs, and probably encodes CTP:glycerol-3-phosphate cytidylyltransferase (GCT) [35], which has been used as a model for this cytidylyltransferase family [36]. The same locus encoding the probable GCT was sequenced in many strains, including type 2 and type 4, but the GCT homolog has been found only in type 18C. The metal inhibition may be physiologically relevant, because all tested metals inhibited activity by
Acknowledgements We would like to thank Joshua A. Newsted for work done with the N-terminal truncation of the enzyme, Jeanne Stuckey for structural modelling, Zhaohui Xu for use of the Superdex-200 chromatography system, and the Institute for Genomic Research for preliminary sequence data. Funding was provided by NIH grant RO1 GM60510 (C.K.), NIH training grant GM07767 (H.A.C.), and institutional NIH grants P60DK20572 and MO1-RR00042.
References [1] A. Tomasz, Science 157 (1967) 694^697. [2] H.J. Jennings, C. Lugowski, N.M. Young, Biochemistry 19 (1980) 4712^4719. [3] E. Tuomanen, Curr. Opin. Microbiol. 2 (1999) 35^39. [4] A. Tomasz, M. Westphal, E.B. Briles, P. Fletcher, J. Supramol. Struct. 3 (1975) 1^16. [5] E.S. Lysenko, J. Gould, R. Bals, J.M. Wilson, J.N. Weiser, Infect. Immun. 68 (2000) 1664^1671. [6] A. Severin, D. Horne, A. Tomasz, Microb. Drug Resist. 3 (1997) 391^400. [7] J. Yother, K. Leopold, J. White, W. Fischer, J. Bacteriol. 180 (1998) 2093^2101. [8] W. Fischer, Res. Microbiol. 151 (2000) 421^427. [9] E.P. Kennedy, S.B. Weiss, J. Biol. Chem. 222 (1956) 193^ 214. [10] I.R. Poxton, D.J. Leak, J. Gen. Microbiol. 100 (1977) 23^29. [11] B. Bean, A. Tomasz, J. Bacteriol. 130 (1977) 571^574.
BBAMCB 55834 17-12-01
H.A. Campbell, C. Kent / Biochimica et Biophysica Acta 1534 (2001) 85^95 [12] J.N. Weiser, J.M. Love, E.R. Moxon, Cell 59 (1989) 657^ 665. [13] J.R. Zhang, I. Idanpaan-Heikkila, W. Fischer, E.I. Tuomanen, Mol. Microbiol. 31 (1999) 1477^1488. [14] J.N. Weiser, M. Shchepetov, S.T. Chong, Infect. Immun. 65 (1997) 943^950. [15] C. Kent, Biochim. Biophys. Acta 1348 (1997) 79^90. [16] C.H. Collins, P.M. Lyne, J.M. Grange, Microbiological Methods, 7th Edn., Butterworths, London, 1995. [17] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, 2nd Edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. [18] J.I. MacDonald, C. Kent, Protein Expr. Purif. 4 (1993) 1^ 7. [19] M.M. Bradford, Anal. Biochem. 72 (1976) 248^254. [20] D.C. Harris, Quantitative Chemical Analysis, 4th Edn., Freeman, New York, 1995. [21] B.N. Ames, Methods Enzymol. 8 (1966) 115^118. [22] J.N. Morand, C. Kent, J. Biol. Chem. 264 (1989) 13785^ 13792. [23] A. Lykidis, K.G. Murti, S. Jackowski, J. Biol. Chem. 273 (1998) 14022^14029. [24] S. Lacks, R.D. Hotchkiss, Biochim. Biophys. Acta 39 (1960) 508^517. [25] S. Shinedling, M. Gayle, D. Pribnow, L. Gold, Mol. Gen. Genet. 207 (1987) 224^232. [26] G. Sulzenbacher, L. Gal, C. Pene¡, F. Fassy, Y. Bourne, J. Biol. Chem. 276 (2001) 11844^11851.
95
[27] F. Anjum, V. Rishi, F. Ahmad, Biochim. Biophys. Acta 1476 (2000) 75^84. [28] J.A. Friesen, H.A. Campbell, C. Kent, J. Biol. Chem. 274 (1999) 13384^13389. [29] I.H. Segel, Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems, Wiley, New York, 1975. [30] J.A. Friesen, Y. Seo Park, C. Kent, Protein Expr. Purif. 21 (2001) 141^148. [31] K. Brown, F. Pompeo, S. Dixon, D. Mengin-Lecreulx, C. Cambillau, Y. Bourne, EMBO J. 18 (1999) 4096^4107. [32] W. Blankenfeldt, M. Asuncion, J.S. Lam, J.H. Naismith, EMBO J. 19 (2000) 6652^6663. [33] B.D. Guenther, C.A. Sheppard, P. Tran, R. Rozen, R.G. Matthews, M.L. Ludwig, Nat. Struct. Biol. 6 (1999) 359^ 365. [34] A. Geerlof, A. Lewendon, W.V. Shaw, J. Biol. Chem. 274 (1999) 27105^27111. [35] S.M. Jiang, L. Wang, P.R. Reeves, Infect. Immun. 69 (2001) 1244^1255. [36] Y.S. Park, T.D. Sweitzer, J.E. Dixon, C. Kent, J. Biol. Chem. 268 (1993) 16648^16654. [37] F.C. Kung, J. Raymond, D.A. Glaser, J. Bacteriol. 126 (1976) 1089^1095. [38] H. Seto, A. Tomasz, J. Bacteriol. 126 (1976) 1113^1118. [39] A. Tomasz, Proc. Natl. Acad. Sci. USA 59 (1968) 86^93.
BBAMCB 55834 17-12-01