(Received for publication, March 3, 1997, and in revised form, April 11, 1997)
From the Department of Biological Chemistry and the § Biophysics Research Division, The University of Michigan, Ann Arbor, Michigan 48109
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 × 105
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.
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.
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.
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]dATPS, 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-
-D-galactopyranoside (IPTG),
5-bromo-4-chloro-3-indolyl-
-D-galactopyranoside (X-gal), CTP, L-
-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.
E. coli TG1 (8)
was used to prepare M13 single-stranded DNA template for mutagenesis.
E. coli DH5 (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.
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 ExpressionThe 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 DH5
. 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.
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.
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 ElectrophoresisSDS-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 DeterminationProtein 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 AnalysisFor 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).
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 -carbon without imposing severe constraints on secondary structure and tertiary conformation.
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.
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.
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.
|
Mutations in three
His residues resulted in a decrease in kcat
values by more than 3 orders of magnitude, H14A (5 × 105), 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.
|
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
103, 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.
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.
|
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.
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.
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 -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).
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 kcatAsp-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 KmOnly 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 ActivityBecause 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 RemarksIn 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.
The authors thank Joel Clement, Jon Friesen, and James Peliska for helpful discussions.