Identification of Functional Conserved Residues of CTP:glycerol-3-phosphate Cytidylyltransferase
ROLE OF HISTIDINES IN THE CONSERVED HXGH IN CATALYSIS*

(Received for publication, March 3, 1997, and in revised form, April 11, 1997)

Young Seo Park Dagger , Patricia Gee , Subramaniam Sanker , Eric J. Schurter §, Erik R. P. Zuiderweg § and Claudia Kent

From the Department of Biological Chemistry and the § Biophysics Research Division, The University of Michigan, Ann Arbor, Michigan 48109

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

The CTP:glycerol-3-phosphate cytidylyltransferase (GCT) of Bacillus subtilis has been shown to be similar in primary structure to the CTP:phosphocholine cytidylyltransferases of several organisms. To identify the residues of this cytidylyltransferase family that function in catalysis, the conserved hydrophilic amino acid residues plus a conserved tryptophan of the GCT were mutated to alanine. The most dramatic losses in activity occurred with H14A and H17A; these histidine residues are part of an HXGH sequence similar to that found in class I aminoacyl-tRNA synthetases. The kcat values for H14A and H17A were decreased by factors of 5 × 10-5 and 4 × 10-4, respectively, with no significant change in Km values. Asp-11, which is found near the HXGH sequence in the cytidylyltransferases but not aminoacyl-tRNA synthetases, was also important for activity, with the D11A mutation decreasing activity by a factor of 2 × 10-3. Several residues found in the sequence RTEGISTT, a signature sequence for this cytidylyltransferase family, as well as other isolated residues were also shown to be important for activity, with kcat values decreasing by factors of 0.14-4 × 10-4. The Km values of three mutant enzymes, D38A, W74A, and D94A, for both CTP and glycerol-3-phosphate were 6-130-fold higher than that of the wild-type enzyme. Mutant enzymes were analyzed by two-dimensional NMR to determine if the overall structures of the enzymes were intact. One of the mutant enzymes, D66A, was defective in overall structure, but several of the others, including H14A and H17A, were not. These results indicate that His-14 and His-17 play a role in catalysis and suggest that their role is similar to the role of the His residues in the HXGH sequence in class I aminoacyl-tRNA synthetases, i.e. to stabilize a pentacoordinate transition state.


INTRODUCTION

The CTP:glycerol-3-phosphate cytidylyltransferase (GCT)1 from Bacillus subtilis catalyzes the formation of CDP-glycerol and pyrophosphate from CTP and glycerol-3-phosphate. CDP-glycerol then serves as a principal precursor required for biosynthesis of poly(glycerol phosphate), the major teichoic acid found in the bacterial cell wall (1, 2). GCT appears to be a member of a cytidylyltransferase family that includes CTP:phosphocholine cytidylyltransferases (CCT), a key regulatory enzyme in the CDP-choline pathway for phosphatidylcholine biosynthesis in higher eukaryotes (3-5). Fig. 1 shows a comparison of the deduced amino acid sequences of several cytidylyltransferases: presumed GCTs from Staphylococcus aureus and Streptomyces wedmorensis, GCT from B. subtilis, CCTs from a variety of organisms, and ethanolamine phosphate cytidylyltransferase (ECT) from yeast. This comparison reveals a number of residues that are conserved among these three enzymes (6). These conserved amino acids are contained within the catalytic core of the CCTs (7). Conservation of these amino acids suggests that they may be required for functions such as catalysis, recognition of substrates, structural integrity, or association of subunits.


Fig. 1. Alignment of catalytic regions of cytidylyltransferases. Alignments were made of the conserved regions of the enzymes with the programs PILEUP, BESTFIT, and GAP of the Wisconsin Package Version 9.0, Genetics Computer Group, Madison, WI. Residues that are shaded in black are identical or highly conserved (S = T, K = R, V = I = L) in nine of the ten proteins. The following sequences (followed by the GenBank accession number) were used: Gct-Sw, a sequence from S. wedmorensis (D38561[GenBank]) in which the N-terminal portion is similar to GCT and the C-terminal portion is a presumed phosphoenolpyruate phosphomutase; GCT-Bs, GCT from B. subtilus (M57497[GenBank]); GCT-Sa, presumed GCT from S. aureus (X87105[GenBank]); CCT-At, presumed CCT from Arabidopsis thaliana (U50451[GenBank]); CCT-Ce, presumed CCT from Caenorhabditis elegans (U29378[GenBank]); CCT-R, CCT from rat (L13245[GenBank]); CCT-y, CCT from the yeast Saccharomyces cerevisiae (M36827[GenBank]); CCT-Pf, CCT from Plasmodium falciparum (X84041[GenBank]); ECT-Y, ECT from the yeast S. cerevisiae (D50644[GenBank]); ECT-H, ECT from humans (D84307[GenBank]). ECT-Y and ECT-H actually have two catalytic domains, but the similarity of the second domain sequence ceases soon after the HXGH sequence and was not included in the figure. Other mammalian cytidylyltransferase sequences from human (L28957[GenBank]), mouse (Z12302[GenBank]), and CHO cells (L13244[GenBank]) are 95-99% identical to the rat sequence and therefore were not included here. Other plant sequences from pea (Y09101[GenBank]) and rapeseed (D58404[GenBank], D63166[GenBank], D63167[GenBank], D63168[GenBank]) are greater than 80% identical to CCT-At and were not included here. All shaded residues in the figure were conserved in all mammalian and plant CCTs.
[View Larger Version of this Image (92K GIF file)]

Considering the high degree of similarity of GCT with CCT and ECT, it would be useful to use the GCT as a model cytidylyltransferase for studying structure-function relationships. We have expressed the GCT in Escherichia coli, purified it in large quantities, and examined its physicochemical characteristics (6). The availability of recombinant GCT expressed in E. coli permits the use of site-directed mutagenesis to search for catalytically important amino acids. Site-directed mutagenesis has allowed direct, quantifiable assessment of the contributions of individual amino acids toward enzyme specificity and catalysis.

In the present study, conserved hydrophilic residues, a conserved tryptophan, and a non-conserved cysteine residue were chosen for mutagenesis. The mutated GCTs were expressed and purified as histidine-tagged forms and subjected to kinetic and structural analyses to evaluate the contributions of those residues in the GCT to catalytic parameters.


EXPERIMENTAL PROCEDURES

Materials

Restriction endonucleases and DNA-modifying enzymes were from Life Technologies, Inc. Sequenase sequencing kit was from U. S. Biochemical Corp. CDP-[methyl-14C]choline, L-[U-14C]glycerol-3-phosphate, [35S]dATPalpha S, and the oligonucleotide-directed in vitro mutagenesis kit were from Amersham Corp. Oligonucleotides for site-directed mutagenesis were synthesized by the DNA Synthesis Core Facility, University of Michigan. Isopropyl-1-thio-beta -D-galactopyranoside (IPTG), 5-bromo-4-chloro-3-indolyl-beta -D-galactopyranoside (X-gal), CTP, L-alpha -glycerophosphate, CDP-choline, and imidazole were supplied from Sigma Co. Acrylamide and N,N'-methylenebisacrylamide were from Boehringer Mannheim. SDS-polyacrylamide gel electrophoresis molecular weight standards were from Bio-Rad. Ni2+-nitrilotriacetic acid (Ni2+-NTA) resin was from Qiagen, Inc.

Bacterial Strains and Plasmids

E. coli TG1 (8) was used to prepare M13 single-stranded DNA template for mutagenesis. E. coli DH5alpha (9) was used for the screening and propagation of plasmids. E. coli M15[pREP4] (10) was used to express the histidine-tagged wild-type and mutant enzymes. M13mp18 was used as a vector for site-directed mutagenesis. The plasmid pQE30 (Qiagen Inc.) (11), containing the highly efficient T5 promoter and His6 tag coding sequence, was used as the expression vector to overexpress the wild-type and the mutated GCT-coding genes for kinetic analysis. Plasmid pET11a, from Novagen, was used for expression for NMR analysis.

Oligonucleotide-directed Mutagenesis

A 0.4-kilobase PstI-XbaI fragment from pBGCT (6), containing the entire coding region of GCT, was subcloned into the polylinker region of M13mp18 at the PstI and XbaI sites to produce M13mp18GCT. A single-stranded form of this plasmid was prepared by the transformation of M13mp18GCT into E. coli TG1 cells by the method of Tabor et al. (12) and used as a template for site-directed mutagenesis. Site-directed mutagenesis was performed with synthetic oligonucleotides according to the method of Sayers et al. (13) using a kit from Amersham Corp. Mutagenic oligonucleotides were 18-20 nucleotides in length, with the mutagenic codon approximately in the middle. The codon GCT was used for conversions to Ala, GAA was used for conversions to Glu, and AAA was used for conversions to Lys. Mutations were identified by sequencing of single-stranded DNA from M13 transformants. DNA sequencing was carried out by the dideoxy chain termination method of Sanger et al. (14). To ensure that the desired mutation was the only mutation in the sequence, the entire region of the inserted DNA was sequenced.

Vectors for Protein Expression

The pQE30 plasmid was used as an expression vector for all the kinetic studies. To subclone mutagenized GCT genes into the vector, an NdeI Site (at base 145) and a BamHI site (at base 181) were introduced by site-directed mutagenesis within the polylinker region of the vector. The oligonucleotides for introducing NdeI and BamHI sites (underlined sequences) were 5'-CACCATCACCATATGGCATGCGAG-3' and 5'-TCGACCGGATCCAAGCTTA-3', respectively. The reconstructed sequence of the polylinker region was 5'-CATATGGCATGCGAGCTCGGTACCCCGGGTCGACCTGGATCCAAGCTT-3'. The additional NdeI site in the pQE30 plasmid, located at position 1399, was removed by site-directed mutagenesis with oligonucleotide, 5'-GTGCACCAAATGCGGTA-3'. The resultant plasmid was named as pQE30NB. The 400-base pair NdeI-BamHI fragments of wild-type and mutant GCT genes were subcloned from M13mp18GCT into the NdeI-BamHI site of pQE30NB and transformed into E. coli DH5alpha . The recombinant DNA was isolated, and its structure was analyzed by restriction enzyme digestion. The identified recombinant DNA was then transformed into E. coli M15[pREP4] to overexpress the wild-type and mutant proteins.

While expression from the pQE30 plasmid was successful for our kinetic studies, considerable difficulty with expression was encountered when the proteins were being expressed for NMR studies. The reason for the difference may be due to growth of the bacteria in rich medium for the kinetic studies and in minimal medium for the NMR analyses. To overcome this problem, we transferred the wild-type and mutant constructs to the pET11a expression vector for producing protein for NMR studies. The plasmid PQE-30 conferred the sequence MRGS(H)7 to the N terminus of the GCT sequence. To produce the identical wild-type and mutant proteins that had been used for the kinetic studies, the MRGS(H)7 sequence was added to the N terminus of the proteins in the pET11a vector by PCR. The 5' PCR primer was designed to incorporate the MRGS(H)7 at the N terminus as well as to create an NdeI site at the 5' end of the construct for cloning. In addition, the codon CAT coding for the seventh His of the tag was modified to CAC to disrupt the NdeI site present there. The nucleotide sequence of the 5' primer was 5'-CGGGAATTCCATATGAGAGGATCGCATCACCATCACCATCACCACATGAAAAAAGTTATCACATACGG-3'. The 3' primer, which incorporated a BamHI site for cloning into pET11a was 5'-CGCGGATCCGCGTTATAAACCAGC-3'. The PCR products were ligated into pET11a that had been digested with NdeI and BamHI. The coding regions of all resulting constructs were sequenced to confirm that the expected sequence was obtained.

Protein Purification

Single colonies of cells harboring recombinant DNA were picked from cells plated on LB plates containing 100 µg/ml ampicillin and 25 µg/ml kanamycin and used to inoculate 5-ml cultures. The 5-ml cultures were grown overnight at 37 °C and then used to inoculate 50-ml culture in LB medium containing the antibiotics. Cultures were grown at 37 °C until the absorbance at 600 nm reached a value of about 0.8. IPTG was then added to a final concentration of 1 mM, and the culture was grown for 4 h at 30 °C (D11A and R55A) or 37 °C (wild-type and all other mutants). The cells from an IPTG-induced culture were harvested by centrifugation at 5,000 × g at 4 °C for 7 min and resuspended in 5 ml of 10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% (v/v) glycerol. Cells were disrupted by sonication, and the lysate was centrifuged at 12,000 × g for 20 min. The supernatant was applied to a 9 × 0.8-cm column containing 0.5 ml of Ni2+-NTA resin and equilibrated with 10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% (v/v) glycerol. The column was washed sequentially with 20 ml of 10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% (v/v) glycerol and 10 ml of 20 mM imidazole, 10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% (v/v) glycerol until the absorbance at 280 nm reached below 0.02. The wild-type and mutant GCTs were eluted and fractionated with 1 ml of 200 mM imidazole, 10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% (v/v) glycerol. Fractions containing GCT were pooled and dialyzed against 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM dithiothreitol to remove the imidazole prior to use. Histidine-tagged wild-type and mutant GCT were stored at -4 °C.

Kinetic Analysis of Purified GCT Mutants

Purified GCT mutants were subjected to analysis by steady-state kinetics. Reactions were performed under the standard conditions described previously (6) for the determination of activity; the concentration of both substrates varied between 0.125 and 4 mM. The kinetic data were fitted to the Michaelis-Menten equation, and the kinetic parameters kcat and Km values for both substrates were determined from secondary plots of the slopes and intercepts versus the reciprocal of the fixed substrate concentration.

Gel Electrophoresis

SDS-polyacrylamide gel electrophoresis was carried out in 1.5-mm-thick 12.5 or 14% slab gels as described by Laemmli (15). Gels were stained with Coomassie Brilliant Blue R-250. Phosphorylase b (Mr = 97,400), bovine serum albumin (Mr = 66,200), ovalbumin (Mr = 45,000), carbonic anhydrase (Mr = 31,000), soybean trypsin inhibitor (Mr = 21,500), and lysozyme (Mr = 14,400) were used as molecular weight standards.

Protein Concentration Determination

Protein concentration was determined by the Bradford method (16) with reagent purchased from Bio-Rad. Bovine serum albumin was used as a standard.

Nuclear Magnetic Resonance Analysis

For NMR analysis, mutant and wild-type enzymes were expressed by growth in M9 minimal medium supplemented with 15NH4Cl. The wild-type NMR spectrum was recorded on a Bruker AMX-600 spectrometer while the spectra of the mutants were obtained on a Bruker AMX-500 spectrometer, both using a gradient triple resonance 15N/13C/1H 5-mm probe. The sample concentrations were typically 0.2 mM in monomer (the protein is a 28-kDa dimer at this concentration) in 90% H2O, 10% 2H2O, 5 mM phosphate buffer, pH 7.8, with the temperature regulated at 25 °C. Gradient-enhanced 15N-HSQC spectra (17, 18) were acquired with 2048 and 128 complex points in the t2 and t1 dimensions, respectively, for total experimental times of 14 h each. The data were processed on a Silicon Graphics personal Iris workstation using the program Felix 2.0 (a gift of Dr. D. R. Hare, Hare, Inc., Woodinville, WA).


RESULTS

Site-directed Mutation of GCT

The residues originally chosen for site-directed mutagenesis were based on a comparison of the Bacillus GCT with yeast and mammalian CCT, which were the only sequences in this family known at that time (6). The conserved charged and hydroxy amino acids were targeted for mutagenesis because the side chains of these residues are likely to be involved in catalysis and to make major contributions to the recognition of ligands. In addition, they are likely to occupy exposed positions in the tertiary structure (19, 20), and substituting these residues may not disrupt the overall conformation of the molecule. The single cysteine residue in GCT was mutated because rat CCT (21) and GCT2 are inactivated by sulfhydryl reagents such as p-chloromercuribenzoate and N-ethylmaleimide (21). Substitution to alanine was chosen because it eliminates the side chain beyond the beta -carbon without imposing severe constraints on secondary structure and tertiary conformation.

Expression and Purification of Wild-type and Mutant GCTs

Wild-type GCT and the 24 alanine mutants were expressed as histidine-tagged forms of the enzyme. The purified wild-type and mutant enzymes were essentially homogeneous by the criterion of SDS-gel electrophoresis (Fig. 2). 3-7 mg of the purified enzymes were recovered from 50-ml cultures. The apparent molecular mass of the GCT was about 16 kDa, which was in agreement with the sum of the molecular mass of the native GCT and that of the attached 11 residues.


Fig. 2. SDS-polyacrylamide gel electrophoresis of purified wild-type and mutant GCT. Purified enzymes (10 µg each) were loaded on 12.5% SDS-polyacrylamide gels and electrophoresed under denaturing conditions. The gels were stained with Coomassie Brilliant Blue R-250. Molecular mass standards (MW) of the indicated sizes are shown in kDa. WT, wild type
[View Larger Version of this Image (45K GIF file)]

Specific Activities of Mutant Enzymes

A preliminary characterization of the mutant enzymes was carried out with purified enzymes of mutated clones. Activity was measured under Vmax conditions previously established for the wild-type enzyme. Among the 24 mutant enzymes, 11 were highly active, with specific activities greater than 20% of the wild-type value (Fig. 3). Eight mutant enzymes were quite inactive, with specific activities less than 1% of the wild-type value. At the standard protein concentration used to measure wild-type activity, these mutants activities could not be distinguished from the background of the assay. To obtain measurable activities for these mutants, it was necessary to assay up to 500 times more of the mutant enzyme proteins and to use longer reaction times than those used for the assay of the wild-type enzyme.


Fig. 3. Specific activity of alanine mutants of GCT. Specific activities of purified alanine mutants of GCT were expressed as percentages of wild-type activity. WT, wild type.
[View Larger Version of this Image (29K GIF file)]

Kinetic Analysis of Mutants

The purified 24 mutant enzymes were analyzed by steady-state kinetics. Both the kcat and Km values for the histidine-tagged wild-type enzyme determined in this study showed only a 2.5-3-fold decrease compared with those for non-histidine-tagged enzyme (6) so that the catalytic efficiency, kcat/Km, was not altered appreciably by the addition of the histidine tag. The kinetic parameters of the mutants with kcat values greater than 20% of the wild-type value are shown in Table I. The residues altered in the mutants shown in this table are not likely to be as important in catalysis as those altered in other mutants with more dramatic effects, as discussed below.

Table I. Kinetic parameters of wild-type and mutant enzymes that have similar kinetic constants to the wild-type enzyme

The nomenclature used for mutants is the amino acid residue substituted in GCT, followed by the residue number, followed by the amino acid replacement. Values represent the averages of at least three independent experiments.

Mutant kcat Km
kcat/Km
CTP GP CTP GP

s-1 mM s-1 mM-1
Wild-type 18.8 1.39 1.09 13.5 17.2
K19A 24.9 2.19 1.78 11.4 14.0
E22A 23.6 1.80 1.37 13.1 17.2
K25A 18.1 1.44 1.06 12.6 17.1
E39A 12.3 1.25 0.85 9.86 14.5
E67A 12.0 2.63 2.00 4.55 5.98
D87A 23.8 2.30 1.72 10.4 13.9
D93A 26.2 2.25 1.85 11.7 14.2
K103A 22.3 1.74 1.46 12.8 15.3
C106A 13.9 1.33 1.19 10.5 11.7
E115A 21.6 6.04 4.34 3.57 4.97

Mutants with Low kcat Values

Mutations in three His residues resulted in a decrease in kcat values by more than 3 orders of magnitude, H14A (5 × 10-5), H17A (4 × 10-4), and H84A (4 × 10-4) (Table II). Mutations that caused kcat to decrease to values between 0.1 and 0.3% were D11A (0.2%), R55A (0.3%), and D66A (0.1%). The R63A, R113A, T114A, S118A, and T119A mutations decreased the kcat values to approximately 2-14% of that of wild-type. None of these mutations to Ala caused more than a 2-fold increase in Km, with the exception of D66A, in which the Km for glycerol-3-phosphate was increased 4-fold.

Table II. Kinetic parameters of GCT mutants with low kcat values

Values represent the averages of at least three independent experiments.

Mutant kcat Km
kcat/Km
CTP GP CTP GP

s-1 mM s-1 mM-1
Wild-type 18.8 1.39 1.09 13.5 17.2
D11A 0.03 1.02 1.67 0.03 0.02
D11E 0.02 0.73 0.85 0.03 0.02
H14A 0.001 1.03 1.26 0.001 0.001
H17A 0.007 1.27 2.22 0.01 0.003
R55A 0.05 0.72 1.72 0.07 0.03
R55K 4.32 4.57 11.4 0.95 0.38
R63A 2.70 1.07 0.79 2.52 3.42
D66A 0.01 2.02 4.58 0.006 0.003
D66E 3.05 0.63 0.45 4.84 6.78
H84A 0.007 0.50 0.53 0.02 0.01
R113A 0.33 1.02 2.60 0.32 0.13
R113K 0.01 1.40 3.93 0.008 0.003
T114A 1.63 1.62 1.30 1.01 1.25
S118A 1.23 1.14 0.89 0.79 1.38
T119A 1.45 1.64 1.56 0.88 0.93

To ask if the charges of the Asp and Arg side chains were critical for catalysis, some of these residues were changed to Glu and Lys, respectively (Table II). The kcat value of the D11E mutant enzyme was as low as that of the D11A mutant, suggesting that the precise position of the carboxylate group, and not simply its charge, is essential for catalysis. When Arg-55 and Asp-66 were mutated to Lys and Glu, respectively, kcat values of both mutant enzymes were restored to values close to that of wild-type. The Km values for R55K, however, were increased by 3-10-fold. These results indicate that a charged residue at position 55 is important for catalysis and that the precise position of the charge may affect substrate binding. The observation that the R113K mutation decreased the kcat value by a factor of 10-3, much greater than the decrease in kcat observed for the R113A mutation, was unexpected and may be due to the Lys causing a mispositioning of the substrate in the transition state.

Mutants with High Km Values

Three mutants, D38A, W74A, and D94A, had Km values that were 6-129-fold higher than that of the wild type enzyme (Table III). The kcat values for these mutant enzymes were at least 8% of that of wild-type or higher. It is interesting that no mutation resulted in alteration of the Km for only one substrate, suggesting that the binding sites for the two substrates are quite close or that Km does not actually represent a true binding equilibrium. When Asp-38 and Asp-94 were mutated to Glu, Km values were fully restored to normal, which suggests that a carboxylate group at these positions is essential. Unexpectedly, the kcat value for D94E was much lower than that of either the wild-type or D94A.

Table III. Kinetic parameters of GCT mutants with high Km values

Values represent the averages of at least three independent experiments.

Mutant kcat Km
kcat/Km
CTP GP CTP GP

s-1 mM s-1 mM-1
Wild-type 18.8 1.39 1.09 13.5 17.2
D38A 1.55 8.88 8.12 0.17 0.19
D38E 24.1 1.00 0.85 24.1 28.4
W74A 9.37 7.94 14.2 1.18 0.66
D94A 3.41 93.5 141 0.04 0.02
D94E 0.02 0.63 1.01 0.03 0.02

Structural Analysis

To determine if the observed defects in catalytic activity caused by amino acid substitutions were due to extensive conformational defects, two-dimensional NMR analysis was carried out. The 15N-HSQC spectra were used to provide direct evidence for any disruptions in the tertiary structure of GCT caused by the mutations. Comparison of the 15N-HSQC spectra for the wild-type GCT and mutation H17A showed no evidence for a collapse of the 1H and 15N chemical shift dispersion in the spectra of the mutant. Furthermore, comparison of cross-peak chemical shift patterns revealed that the mutations did not cause significant changes in GCT tertiary structure (Fig. 4). Similarly, spectra for mutations H14A, D38A, W74A, and D94A revealed that these mutations do not substantially disrupt the tertiary structure of GCT (data not shown). Chemical shift differences for individual cross-peaks among the mutants and between the mutants and wild type were present. However, the relative pattern of cross-peaks within corresponding regions of the spectra show a high degree of homology, indicating that all the structures are similar.


Fig. 4. Two-dimensional 15N-HSQC spectra of wild-type GCT (A) and H17A (B). Sample concentration was 0.2 mM monomer in 90% H2O, 10% 2H2O, 5 mM phosphate buffer, pH 7.8, with the temperature regulated at 25 °C.
[View Larger Version of this Image (15K GIF file)]

Due to low sample solubility, it was not possible to obtain two-dimensional 15N-HSQC spectra for mutations H84A, D66A, R55A, and D11A, which limits the structural information available on these mutants. However, the one-dimensional spectra of the H84A and D11A mutants did show good dispersion of the methyl resonances, indicating that these mutations did not cause a complete loss of tertiary structure. In contrast, the same methyl region of the one-dimensional 1H spectrum for the D66A mutant was collapsed, strongly suggesting that the D66A mutation did cause a widespread change in tertiary structure.


DISCUSSION

The HXGH Sequence

It is clear from the mutagenesis and structural studies that His-14 and His-17 are critical for catalytic activity. Mutations of these residues result in a lowering of kcat by several orders of magnitude. These residues are part of the sequence HXGH, which is also found in a nucleotidyltransferase "superfamily" consisting of class I aminoacyl-tRNA synthetases (22, 23) as well as other known and putative adenylyltransferases and cytidylyltransferases (5). The first half-reaction catalyzed by tRNA synthetases is a nucleotidyltransfer of the adenylyl moiety from ATP to form the aminoacyl-adenylate. For the class I aminoacyl-tRNA synthetases, the two His residues have been shown to be involved in stabilizing the pentacoordinate transition state by binding to the phosphates of the ATP (24-26). Alteration of kcat by mutating these residues to Ala in GCT suggests that they are also involved in stabilization of a transition state for GCT. It is reasonable to propose that a pentacoordinate transition state is formed as part of the GCT reaction, with glycerol-3-phosphate as the nucleophile attacking the alpha -phosphorus of the CTP (Fig. 5). In support of a mechanism in which the pentacoordinate transition state involves the two substrates are kinetic studies that indicate a random order reaction mechanism, rather than the formation of a covalent enzyme-bound intermediate (6).


Fig. 5. Proposed formation of pentacoordinate transition state during the cytidylyltransferase reaction.
[View Larger Version of this Image (20K GIF file)]

Further evidence of the importance of the HXGH sequence was provided by Veitch and Cornell (27) who found that mutating the Gly of the HXGH of rat CCT to Ser resulted in a lowering of activity due to a 4-fold decrease in Vmax plus a 25-fold increase in the Km for CTP. Veitch and Cornell (27) proposed that the HXGH sequence is involved in binding of CTP. Our results are more consistent with a role of the His residues for transition-state stabilization than for initial CTP-binding since mutations H14A and H17A do not result in an altered Km. Nevertheless, it is entirely possible that mutation of the Gly residue of the HXGH would perturb initial substrate binding. Indeed, analysis of the crystal structure of glutaminyl-tRNA synthetase indicates that addition of a side chain at this position would cause steric interference with binding of the adenine ring of ATP (25).

Other Perturbations of kcat

Asp-11 lies very close to the HXGH sequence of GCT and is absolutely conserved in all the cytidylyltransferase sequences. Mutation of Asp-11 to Ala or Glu resulted in lowering kcat by 3 orders of magnitude, indicating that it is a critical residue. The structural analysis of the D11A enzyme was not conclusive, so it is not yet clear if the role of this residue is catalytic or structural. It is interesting that class I aminoacyl-tRNA synthetases usually have a Gly at this position (22). The choice of Gly or Asp may reflect differences in the mechanism of binding a cytidylyl as opposed to an adenylyl moiety.

The sequence RTEGISTT is very highly conserved in this cytidylyltransferase family, with Thr-114, Gly-116, and Ser-118 absolutely conserved and only very conservative substitutions found for Ile-117, Thr-119, and Thr-120. This sequence can be considered a signature sequence for this family of cytidylyltransferases; it is not actually found in the cytidylyltransferase "superfamily" discussed by Bork et al.(5), and we have not found it in other proteins in the GenBankTM/EBI Data Bank. Bork et al. (5) proposed that the KMSKS signature sequence of Class I aminoacyl-tRNA synthetases is similar to the STTK sequence of GCT. While the sequence similarity is not obvious, it is possible that these regions have functional similarity. The crystal structure of glutaminyl-tRNA synthetase shows that the KMSKS sequence is in a loop that interacts with the HXGH sequence loop and participates in stabilization of the transition state (25). In methionyl-tRNA synthetase, the KMSKS loop is close to the HXGH loop and is involved in ATP binding (28). In light of the high degree of conservation of the RTEGISTT sequence, it is likely that this sequence plays a similar role in these cytidylyltransferases by interacting with the HXGH loop and portions of the glycerol-3-phospho-CTP transition state.

The R63A mutation resulted in a relatively modest change in kcat. Arg-63 of GCT corresponds to Arg-140 of mammalian CCT, which is mutated to His in a temperature-sensitive CCT from the mutant strain 58 of Chinese hamster ovary cells (29-31). CCT with the R140H mutation appears to be active but is present at much lower levels than the wild-type enzyme (31). While R63A had a mild effect on kcat, the temperature sensitivity of the R140H enzyme suggests that Arg-63 is more important for structural stability than catalysis. Arg-55 and His-84 also appear important for activity although it is not clear if they play catalytic or structural roles. It is noteworthy that His-84 is not conserved in all cytidylyltransferases, suggesting its role may not be catalytic. Arg-55, on the other hand, remains absolutely conserved.

Perturbations of Km

Only three mutations were shown to modify Km by more than 4-fold, D38A, W74A, and D94A. Each mutation also causes a modest alteration in kcat. Structural studies on these mutants reveal that each is not grossly altered from wild-type. Thus Asp-38, Trp-74, and Asp-94 all appear to be involved in catalysis, perhaps at the level of substrate binding. The residues substituted in these mutants may be candidates for those that interact directly with CTP and/or glycerol-3-phosphate, or they may be an important structural feature of a loop that interacts with both substrates. Alternatively, since Km is not always a true measure of binding affinities, the mutations could be affecting a kinetic step after substrate binding that alters the apparent binding constant for both substrates.

Mutations That Do Not Perturb CT Activity

Because the rationale for mutagenesis was based on alignment of only the three sequences that were available at the time these studies began, there was an overestimate of the conserved residues. With the addition of further sequences shown in Fig. 1, several residues that were chosen for mutation are no longer conserved. Among these are Lys-19, Glu-22, and Glu-39, and so it is not surprising that mutation of these residues did not alter activity. Cys-106 was not conserved even in the early alignments. We chose to mutate it, however, because we found that sulfhydryl reagents such as N-ethylmaleimide inhibit activity2 as has also been shown for rat CCT (21). The lack of effect of the C106A mutation indicates that this thiol is not critical for activity.

Mutations of Lys-25, Glu-67, Asp-87, Lys-103, and Glu-115 did not appreciably affect activity in GCT. These residues are conserved through most, but not all, sequences in Fig. 1. Why there is such strong conservation of these five residues through most of the sequences can only be speculated, but these residues may be involved in such functions as regulation or stabilization of the enzymes.

Whether or not Asp-93 and Asp-94 are conserved depends on the gaps allowed in the alignment of the proteins (Fig. 1; see Ref. 5), but it is clear that at least one Asp is present in all cytidylyltransferases in this region. In our studies, Asp-93 could be mutated without appreciably affecting activity while Asp-94 appeared critical.

Concluding Remarks

In summary, these studies have defined several residues that are critical for catalysis by GCT; His-14 and His-17 may participate in stabilization of a proposed pentacoordinate transition state, and Asp-38, Trp-74, and Asp-94 may be important for proper substrate binding. Several other residues, Asp-11, Arg-55, His-84, Arg-113, Thr-114, Ser-118, and Thr-119, are important for activity although whether their roles are catalytic or structural has not yet been determined. The complete interpretation of the functional roles played by the residues identified in this study requires the analysis of the three-dimensional structure of the enzyme. Crystals of GCT have been obtained and are being subjected to x-ray diffraction analysis.3 The availability of the three-dimensional structure of the GCT should facilitate an understanding of the roles these residues play in catalysis.


FOOTNOTES

*   This work was supported in part by a grant from the American Cancer Society, BE-126, and National Institutes of Health Grant CA64159. Sequence analysis was supported in part by National Institutes of Health Grant M01-RR0042 to the General Clinical Research Center, University of Michigan Medical Center.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Present address: Dept. of Food and Bioengineering, College of Engineering, Kyungwon University, Sungnam 461-701, Korea.
   To whom correspondence should be addressed: Dept. of Biological Chemistry, 4417 Medical Science I, Box 0606, University of Michigan, Medical School, Ann Arbor, MI 48109-0606. Tel.: 313-747-3317; Fax: 313-7634581.
1   The abbreviations used are: GCT, CTP:glycerol-3-phosphate cytidylyltransferase; CCT, CTP:phosphocholine cytidylyltransferase; ECT, CTP:phosphoethanolamine cytidylyltransferase; IPTG, isopropyl-1-thio-beta -D-galactopyranoside; X-gal, 5-bromo-4-chloro-3-indolyl-beta -D-galactopyranoside; NTA, nitrilotriacetic acid; HSQC, heteronuclear single quantum correlation.
2   Y. S. Park, unpublished data.
3   C. Weber, Y. S. Park, C. Kent, and M. Ludwig, unpublished data.

ACKNOWLEDGEMENTS

The authors thank Joel Clement, Jon Friesen, and James Peliska for helpful discussions.


REFERENCES

  1. Mauel, C., Young, M., and Karamata, D. (1991) J. Gen. Microbiol. 137, 929-941 [Medline] [Order article via Infotrieve]
  2. Pooley, H. M., Abellan, F.-X., and Karamata, D. (1991) J. Gen. Microbiol. 137, 921-928 [Medline] [Order article via Infotrieve]
  3. Kent, C. (1990) Prog. Lipid Res. 29, 87-105 [Medline] [Order article via Infotrieve]
  4. Vance, D. E. (1989) in Phosphatidylcholine Metabolism (Vance, D. E., ed), pp. 33-45, CRC Press, Inc., Boca Raton, FL
  5. Bork, P., Holm, L., Koonin, E. V., and Sander, C. (1995) Proteins 22, 259-266 [Medline] [Order article via Infotrieve]
  6. Park, Y. S., Sweitzer, T. D., Dixon, J. E., and Kent, C. (1993) J. Biol. Chem. 268, 16648-16654 [Abstract/Free Full Text]
  7. Kalmar, G. B., Kay, R. J., Lachance, A., Aebersold, R., and Cornell, R. B. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 6029-6033 [Abstract]
  8. Gibson, T. T. (1984) Studies on the Epstein-Barr Virus Genome. Ph.D. thesis, Cambridge University, Cambridge, UK
  9. Hanahan, D. (1983) J. Mol. Biol. 166, 557-580 [Medline] [Order article via Infotrieve]
  10. Villarejo, M. R., and Zabin, I. (1974) J. Bacteriol. 120, 466-474 [Medline] [Order article via Infotrieve]
  11. Bujard, H., Gentz, R., Lanzer, M., Stüber, D., Müller, M., Ibrahimi, I., Häuptle, M. T., and Dobberstein, B. (1987) Methods Enzymol. 155, 416-433 [Medline] [Order article via Infotrieve]
  12. Tabor, S., and Richardson, C. C. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 4767-4771 [Abstract]
  13. Sayer, J. R., Krekel, C., and Eckstein, F. (1992) BioTechniques 13, 592-596 [Medline] [Order article via Infotrieve]
  14. Sanger, F., Nicklen, S., and Coulson, A. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 5463-5467 [Abstract]
  15. Laemmli, U. K. (1970) Nature 227, 680-685 [Medline] [Order article via Infotrieve]
  16. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 [CrossRef][Medline] [Order article via Infotrieve]
  17. Bodenhausen, G., and Ruben, D. J. (1980) Chem. Phys. Lett. 69, 185-190 [CrossRef]
  18. Hurd, R. E., and John, B. K. (1991) J. Magn. Reson. 92, 648-653
  19. Chothia, C. (1976) J. Mol. Biol. 105, 1-14 [Medline] [Order article via Infotrieve]
  20. Janin, J. (1979) Nature 277, 491-492 [Medline] [Order article via Infotrieve]
  21. Weinhold, P. A., Rounsifer, M. E., and Feldman, D. A. (1986) J. Biol. Chem. 261, 5104-5110 [Abstract/Free Full Text]
  22. Delarue, M., and Moras, D. (1993) Bioessays 15, 675-687 [Medline] [Order article via Infotrieve]
  23. Jones, M. D., Lowe, D. M., Burgford, T., and Fersht, A. R. (1986) Biochemistry 25, 1887-1891 [Medline] [Order article via Infotrieve]
  24. Leatherbarrow, R. J., Fersht, A. R., and Winter, G. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 7840-7844 [Abstract]
  25. Perona, J. J., Rould, M. A., and Steitz, T. A. (1993) Biochemistry 32, 8758-8771 [Medline] [Order article via Infotrieve]
  26. Brunie, S., Zelwer, C., and Risler, J.-L. (1990) J. Mol. Biol. 216, 411-424 [Medline] [Order article via Infotrieve]
  27. Veitch, D. P., and Cornell, R. B. (1996) Biochemistry 35, 10743-10750 [CrossRef][Medline] [Order article via Infotrieve]
  28. Nureki, O., Vassylyev, D. G., Katayanagi, K., Shimizu, T., Sekine, S., Kigawa, T., Miyazawa, T., Yokoyama, S., and Morikawa, K. (1995) Science 267, 1958-1965 [Medline] [Order article via Infotrieve]
  29. Esko, J. D., and Raetz, C. R. H (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 5192-5196 [Abstract]
  30. Esko, J. D., Wermuth, M. M., and Raetz, C. R. H. (1981) J. Biol. Chem. 256, 7388-7393 [Free Full Text]
  31. Sweitzer, T. D., and Kent, C. (1994) Arch. Biochem. Biophys. 311, 107-116 [CrossRef][Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.