(Received for publication, December 2, 1996, and in revised form, January 21, 1997)
From the Department of Biochemistry and Molecular Biology, Oregon Health Sciences University, Portland, Oregon 97201-3098
Crystal structures of hypoxanthine-guanine phosphoribosyltransferase (HGPRT) proteins have implied that the translocation of a flexible loop containing a highly conserved Ser-Tyr dipeptide is necessary for the protection of the proposed oxocarbonium ion transition state of the enzyme (Eads, J. C., Scapin, G. T., Xu, Y., Grubmeyer. C., and Sacchettini, J. C. (1994) Cell 78, 325-334; Schumacher, M. A., Carter, D., Roos, D. S., Ullman, B., and Brennan, R. G. (1996) Nature Struct. Biol. 3, 881-887). An essential role for this Ser-Tyr dyad in HGPRT catalysis has now been verified biochemically and genetically for the Leishmania donovani HGPRT employing a combination of protein modifying reagents and site-directed mutagenesis. Incubation of HGPRT with either tetranitromethane or diethyl pyrocarbonate inactivated the enzyme completely, and peptide sequence analysis revealed that tetranitromethane treatment modified the Tyr residue within the Ser95-Tyr96 dipeptide. Analysis of site-directed mutants confirmed that both amino acids were vital for phosphoribosylation activity. Mutant HGPRTs, S95A, S95E, Y96F, and Y96V, exhibited dramatic reductions in their catalytic capabilities of 2-3 orders of magnitude, whereas HGPRTs containing conservative substitutions, S95C and S95T, displayed only a 2-3-fold decrease in kcat. Km values for the substrates of the forward and reverse reactions were largely unchanged for all HGPRT constructs, except for a 4-5-fold decrease in the Km value of the Y96F and Y96V mutants for phosphoribosylpyrophosphate. Expression of L. donovani hgprt constructs in Escherichia coli indicated that wild type and S95T HGPRTs complemented bacterial phosphoribosyltransferase deficiencies, whereas the S95A and S95C mutants complemented weakly, and the S95E, Y96F, and Y96V HGPRT did not support bacterial growth. These data authenticate that the Ser-Tyr dipeptide that is conserved among all members of the HGPRT family is essential for phosphoribosylation of purine nucleobases by HGPRT.
Protozoan parasites constitute a devastating public health and socioeconomic burden in the tropical and subtropical regions of the world. In the absence of effective vaccines, empirically derived drugs are the only means available to treat parasitic infections. However, the current arsenal of antiparasitic drugs is far from ideal, as these agents are moderately to highly toxic, possibly mutagenic and/or carcinogenic, and require protracted treatment regimens. The control of parasitic diseases has also been complicated by the emergence of drug-resistant strains (1), further underscoring the need for novel antiparasitic agents. Rational development of new drugs that selectively target protozoan pathogens requires the exploitation of biochemical or metabolic differences between parasite and host. As protozoan parasites, unlike mammalian cells, are auxotrophic for purines and consequently rely on purine appropriation from the host for survival and proliferation (2), the purine salvage pathway has stimulated considerable interest as a paradigm for antiparasitic drug development. One enzyme that is central to purine acquisition in many parasites (2) is hypoxanthine-guanine phosphoribosyltransferase (HGPRT)1 which catalyzes the phosphoribosylpyrophosphate (PRPP)-dependent phosphoribosylation of hypoxanthine and guanine to the nucleotide level. In some parasites, HGPRT also recognizes xanthine as a substrate (3) and is, therefore, termed a hypoxanthine-guanine-xanthine phosphoribosyltransferase (HGXPRT). Genes and cDNAs encoding HG(X)PRT (HG(X)PRT) from a number of protozoan parasites have now been cloned, sequenced, and overexpressed in Escherichia coli, and the recombinant HG(X)PRT enzymes have been purified and characterized kinetically (3). Multisequence alignments demonstrate several short regions of amino acid homology among HG(X)PRT family members that are flanked by much longer regions without significant sequence similarity. The crystal structures for the product-bound form of the human HGPRT (4) and the Tritrichomonas foetus HGXPRT (5) have revealed, however, a common core framework within the tertiary structure that is conserved among other members of the phosphoribosyltransferase (PRT) family (6-8). These structures have established functional roles for all of the conserved regions in HG(X)PRT proteins, except one motif that is located on a flexible loop distal to the active site. This flexible loop encompasses a Ser-Tyr dipeptide that is found in all HG(X)PRTs for which sequence is available. The recently determined crystal structures of the HGXPRT from Toxoplasma gondii in the absence of ligand and in the presence of product have demonstrated that this flexible loop in the apoenzyme is shifted toward the active site (9). Schumacher et al. (9) further conjectured that this loop in the apoenzyme is poised to cover the catalytic pocket to protect the transition state intermediate from hydrolysis by the bulk solvent.
These structures, however, have not established which amino acid residues are critical to the catalytic mechanism of HG(X)PRT proteins. To identify amino acids requisite for the phosphoribosylation reaction, we have employed chemical modification and site-directed mutagenesis to definitively establish a critical catalytic role for the Ser95-Tyr96 dipeptide of the HGPRT from the protozoan parasite Leishmania donovani.
Acetic anhydride, diethyl pyrocarbonate (DEPC), diisopropyl fluorophosphate, ethyl dimethylaminopropyl carbodiimide, iodoacetate, p-chloromercuric benzoate, phenyl glyoxal, tetranitromethane (TNM), guanosine monophosphate dialdehyde, inosine monophosphate (IMP), PRPP, hypoxanthine, and guanine were obtained from Sigma. Radiolabeled [14C]guanine (56 mCi/mmol) was purchased from Moravek (Brea, CA). All other reagents were of the highest quality commercially available.
HGPRT Expression in E. coli and Purification of Recombinant ProteinL. donovani HGPRT was ligated into the pBAce
expression vector and transformed into S606 (
pro-gpt-lac,
thi, hpt) E. coli as described (10, 11). This strain
lacks both the bacterial xanthine-guanine and hypoxanthine
phosphoribosyltransferase activities (11). HGPRT expression
was induced in low phosphate induction (LPI) medium as reported (10).
One-liter bacterial cultures were harvested by centrifugation, and the
cell pellets were resuspended in 15 ml of 20 mM Tris-HCl,
pH 8.0, 10 mM MgCl2, 2.0 mM
dithiothreitol (TMD) buffer and lysed with two passes through a French
Press. Lysates were clarified by centrifugation at 31,000 × g for 30 min and supernatants applied to a mixed bed column
of DEAE-cellulose and AG 50-8X (2.7 × 10 cm) pre-equilibrated
with TMD buffer. Void volume fractions containing HGPRT activity, as
determined radiometrically (10), were pooled and solid
(NH4)2SO4 added to 40% saturation. Precipitates were removed by centrifugation, and the supernatants were
brought to 80% saturation with
(NH4)2SO4 to recover HGPRT. The
protein pellets were dissolved in 4.0 ml of TMD buffer and dialyzed
overnight at 4 °C against 4 liters of 20 mM sodium
carbonate, pH 9.8, 10 mM MgCl2, 2.0 mM dithiothreitol. Dialysates were clarified by
centrifugation and applied to a Q-Sepharose column (1.5 × 30 cm)
equilibrated with 50 mM sodium carbonate, pH 9.8, 10 mM MgCl2 and developed with a 500-ml linear
gradient of 0-0.4 M NaCl in the equilibration buffer.
Column eluates were monitored at 280 nm and protein fractions assayed
for HGPRT activity (10) and analyzed for purity by SDS-polyacrylamide
gel electrophoresis (12). Fractions containing a homogeneous 24-kDa
polypeptide were pooled, adjusted to pH 8.0 with 1.0 M
Tris-HCl, pH 8.0, and concentrated to 10 mg/ml using a Centriprep
concentrator (Amicon, Beverly, MA). All enzyme preparations were frozen
in dry ice/acetone and stored at
70 °C, except the S95C mutant,
which was used immediately after purification. The S95C HGPRT was
relatively unstable and lost activity upon storage. All other HGPRT
mutants, as well as the wild type enzyme, were stable at
70 °C for
up to 6 months.
HGPRT preparations were diluted to 10 µg/ml in 100 mM Tris phosphate, 10 mM MgCl2 (either pH 5.0, 7.0, or 9.0), and protein modifying agents were added to a final concentration of 10 mM, with the exception of guanosine monophosphate dialdehyde which was used at a concentration of 100 µM. Samples were incubated at 20 or 0 °C for 60 min. HGPRT activity was measured by adding 2.0 µl of the enzyme solution to 18 µl of a reaction mixture containing 40 µM [14C]guanine (56 mCi/mmol), 1.0 mM PRPP, 100 mM Tris-HCl, 10 mM MgCl2, and 2.0 mM dithiothreitol, pH 8.0, and incubating at 37 °C for 4 min. The HGPRT assay was linear with time and protein under these conditions. Fifteen-µl aliquots of the reaction mixtures were applied to DEAE-Whatman filters (Maidstone, United Kingdom) and washed with dH2O on a Millipore manifold (Bedford, MA) to remove unreacted [14C]guanine. [14C]GMP reaction product was quantitated by liquid scintillation.
Quantitation and Isolation of TNM-labeled PeptidesHGPRT
(2.3 nmol) was dissolved in 1.0 ml of 100 mM Tris
phosphate, 10 mM MgCl2 in the presence or
absence of 10 mM PRPP and preincubated at 20 °C for 5 min. TNM (230 nmol) dissolved in 5 µl of methanol was then added to
each sample, and the mixture was incubated for an additional 20 min at
20 °C. The reaction was then terminated by the addition of
trichloroacetic acid to a final concentration of 15%. The HGPRT
precipitate was dissolved in 1.0 ml of 100 mM sodium
acetate, pH 4.7, and dialyzed overnight against 2.0 liters of 20 mM acetic acid to remove residual nitroformate chromophore
(13). The dialyzed protein was lyophilized and dissolved in 200 µl of
50 mM ammonium bicarbonate, pH 8.0, and the levels of
nitrotyrosine were quantitated spectrophotometrically at 428 nm using
an extinction coefficient of 4100 M1
cm
1 (13). The nitrated protein was then digested with the
endoproteinase Glu-C (Sigma) at a molar ratio of 20 µl (HGPRT:Glu-C)
for 20 h at 37 °C. Nitrated peptides were purified by high
performance liquid chromatography (HPLC) using a reversed phase
C18 column (Bio-Rad) equilibrated with 20 mM
ammonium bicarbonate, pH 8.0, and developed with a 1%/min linear
gradient of 50:50 methanol, 1-propanol. Fractions exhibiting 428 nm
absorbance were collected, lyophilized, and subjected to N-terminal
sequence analysis on a Applied Biosystems 473 pulsed liquid phase
sequenator equipped with a 120 PTH analyzer (Foster City, CA).
To determine molecular masses of nitrated peptides, the peptides were fractionated by HPLC on a Brownlee C8 column (2.1 × 100 mm) (Applied Biosystems) and developed with a 1%/min linear gradient of 1% acetic acid, 2-propanol at a flow rate of 0.2 ml/min. The HPLC column effluent was interfaced via an electrospray source to a Fisons VG Quattro mass spectrometer (Cheshire, United Kingdom) to determine parent ion masses.
Site-directed Mutagenesis of the L. donovani HGPRTMutations within L. donovani HGPRT were generated by subcloning the EcoRI-XbaI fragment from the HGPRT-pBAce construct described by Allen et al. (10) into the corresponding sites in the pSelect mutagenesis vector (Promega). This EcoRI-XbaI fragment encompasses the entire HGPRT protein coding region, as well as the phoA promoter, thus enabling the expression of the mutated genes in E. coli without further subcloning. Single amino acid substitutions were introduced into HGPRT with 25-base pair mutagenic oligonucleotides containing the appropriate base replacements centrally located in the primer following the protocols outlined in the brochure provided by Promega.
Complementation AnalysisWild type and mutant
HGPRT-pSelect constructs were transformed into S609
(
pro-gpt-lac, hpt, thi, pup, purHJ) E. coli, a purine auxotroph that lacks the same PRT activities as the SN606 cells
(11). Transformants were isolated on LB plates containing 100 µg/ml
ampicillin and 50 µg/ml streptomycin. A single bacterial colony was
then picked and resuspended in 200 µl of LPI medium supplemented with
100 µg/ml ampicillin and 50 µg/ml streptomycin (10). Twenty-µl
aliquots were then dispensed into 3.0 ml of LPI medium containing both
antibiotics and adenine, adenine/hypoxanthine, adenine/guanine, or
adenine/guanosine. Each nucleobase or nucleoside was present at a
concentration of 125 µM. Cultures were incubated at
37 °C with vigorous shaking (300 rpm) for 8 h, and cell
densities were measured spectrophotometrically at 600 nm.
Initial rate measurements for the forward
reaction were determined using a 1.0-cm path length on a Beckman DU640
spectrophotometer equipped with a kinetic software package. Reactions
were carried out in 100 mM Tris-HCl, pH 8.0, 10 mM MgCl2, 2.0 mM dithiothreitol at
37 °C either in 1.0 mM PRPP and 1-100 µM
guanine or hypoxanthine to determine the Km for the
base or at 40 µM guanine and 25-800 µM
PRPP to ascertain the Km for PRPP. IMP and GMP
formation were monitored at 243 and 257 nm using extinction coefficients of 2200 M1 cm
1 and
4200 M
1 cm
1, respectively. The
reverse reaction in which IMP in the presence of PPi is
converted to hypoxanthine and PRPP was monitored using the xanthine
oxidase-coupled enzyme assay previously reported (14).
To identify amino acid residues within HGPRT that
are essential for catalytic activity, the purified enzyme was treated
with a battery of protein modifying reagents, including iodoacetate, p-chloromercuric benzoate, acetic anhydride, ethyl
dimethylaminopropyl carbodiimide, diisopropyl fluorophosphate, phenyl
glyoxal, diethyl pyrocarbonate, and guanosine monophosphate dialdehyde.
The specificity of these protein modifying reagents is indicated in
Table I. Surprisingly, the only reagent among these that
effectively reduced L. donovani HGPRT activity at 20 °C
was DEPC (Table I). Although DEPC reacts preferentially with the
imidazole side chains of His residues at pH 5.0-7.0 (15), the reagent
can also modify to a lesser extent the -amino group of Lys and the
hydroxyl group of Tyr residues under alkaline conditions (16).
Incubation of HGPRT with 10 mM DEPC at pH 9.0 resulted in
>95% loss in activity after 60 min at 0 °C (Table I). No loss in
HGPRT activity was observed with DEPC at either pH 5.0 or 7.0, implying
that a His residue is not requisite for HGPRT activity (Table I). Since acetic anhydride, a Lys modifying agent, failed to abrogate enzyme activity, this suggested that DEPC probably modified an essential Tyr
residue. That Tyr was the amino acid within HGPRT being modified by
DEPC was substantiated by changes in the aromatic region of the UV
spectrum after treatment with DEPC. These spectral changes included a
decrease in the 278 nm absorbance with a concomitant increase in the
absorbances at 260 and 290 nm, implying a modification of a phenolic
hydroxyl group (17). As DEPC carboethoxylations are relatively unstable
(18), identification of the specific Tyr residue within the L. donovani HGPRT that was inactivated by DEPC was not attempted.
|
To establish that a Tyr residue is essential for HGPRT catalysis, the L. donovani enzyme was reacted with the Tyr-specific protein modifying reagent TNM, which nitrates Tyr residues at the meta positions in a pH-dependent manner (19). Incubation of HGPRT with a 100-fold molar excess of TNM at 0 °C abolished >80% of the enzymatic activity when the reaction was performed at pH >9.0, while no significant loss of activity by TNM was observed at either pH 5.0 or 7.0 (Table I).
To ascertain if the Tyr residue(s) modified by TNM were located
proximal to the PRPP or nucleobase-binding sites of HGPRT (4),
protection experiments were conducted at 20 °C in the presence of
either PRPP or IMP. HGPRT inactivation in the presence of PRPP
exhibited a concentration-dependent biphasic profile not observed for the TNM-treated enzyme in the absence of PRPP and resulted
in a 50% reduction in specific activity after 5 min, followed by a
much slower rate of inactivation (Fig. 1). Conversely, IMP did not protect HGPRT from TNM inactivation, even at a
concentration as high as 10 mM, although the rate of
inactivation at the higher IMP concentrations was slightly less (Fig.
1).
As nitration of Tyr side chains generates a chromophore with a
max at 428 nm, it is possible to quantitate
spectroscopically the number of Tyr residues modified within a protein
after TNM treatment (13). Treatment of L. donovani HGPRT
with a 100-fold molar excess of TNM for 20 min at 20 °C in the
absence of ligand resulted in the nitration of 1.7 ± 0.4 mol of
Tyr/mol of HGPRT. Inclusion of 10 mM PRPP in the reaction
mixture decreased the amount of Tyr nitration to 0.84 ± 0.10 mol
of Tyr/mol of HGPRT. To determine the number of Tyr residues required
for HGPRT catalysis, the extent of HGPRT inactivation as a function of
TNM incubation time was correlated with the level of Tyr nitration. As
demonstrated in Fig. 2, HGPRT inactivation was directly
proportional to the number of modified Tyr residues in the protein
after TNM treatment, and the nitration of one Tyr residue resulted in
>80% loss of HGPRT activity.
Characterization of Nitrotyrosine Containing Peptides
To
determine which Tyr residues in the L. donovani HGPRT were
modified by TNM, TNM-treated enzyme was digested with endoproteinase Glu-C, and the nitrated peptides were resolved by HPLC and sequenced. As illustrated in Fig. 3, TNM treatment of HGPRT in the
absence of PRPP yielded five 428 nm absorbing peaks diagnostic of
nitrotyrosine-containing peptides. Untreated HGPRT did not contain any
428 nm absorbing peptides. Peptide NP1 was a tetrapeptide,
Ser-Tyr-Arg-Glu, corresponding to residues 189-192 of the L. donovani HGPRT (10) (Table II). The second Edman
degradation cycle yielded two phenylthiohydantoin (PTH)-derivatives,
one eluting with the PTH-Tyr standard. The second was a novel
PTH-derivative, presumably PTH-nitrotyrosine (PTH-Tyr(NO2)). This PTH-Tyr(NO2) derivative
was found in all five 428-nm absorbing peptides obtained from Glu-C
digestion of TNM-treated HGPRT. Peptides NP2 and NP3 both
exhibited the same N-terminal Tyr-Tyr-Glu sequence. NP3 was the
tripeptide, while NP2 was an incomplete Glu-C digestion product
coinciding with residues 203-208 of the L. donovani HGPRT
(Table II). Peptides NP4 and NP5 revealed the N-terminal
sequence,
Phe-Ile-Xaa-Ala-Ser-Ser-(Tyr/Tyr(NO2))-Gly-Thr, that
corresponded to residues 90-98 of the enzyme. Mass spectroscopy analysis of NP4 and NP5 yielded parent molecular ions (M + H+) of 1281 and 1327 mass units, respectively. These masses
approximate the molecular mass of the putative Glu-C fragment of this
region of the protein with the sequence
Phe-Ile-Cys-Ala-Ser-Ser-Tyr-Gly-Thr-Gly-Val-Glu (1234 Da). The
differences between the calculated and observed masses for peptides NP4
and NP5 were 46 and 92 Da, respectively, which can be ascribed to the
mononitration and dinitration of Tyr96.
|
Of the four Tyr residues that are modified by TNM treatment of the
L. donovani HGPRT, Tyr96, Tyr190,
Tyr203, and Tyr204, multisequence alignments
reveal that only Tyr96 is conserved in all members of the
HG(X)PRT family, as well as in the xanthine-guanine PRT E. coli (20) and the guanine PRT from Giardia lamblia (21)
(Fig. 4). Tyr190, Tyr203, and
Tyr204 are all located in the C-terminal portion of the
L. donovani HGPRT (10) and are not conserved among all
HG(X)PRT family members. This multisequence alignment implies that the
Tyr96 that is nitrated in peptides NP4 and NP5 is important
for HGPRT catalysis, making this Tyr residue a primary candidate for
further study by site-directed mutagenesis.
Complementation Analysis
To validate the hypothesis that the
conserved Ser95-Tyr96 dipeptide plays an
integral role in the catalytic mechanism of the enzyme, a series of
site-directed mutations was introduced into the L. donovani
HGPRT at these positions. Ser95 was mutated to Ala,
Cys, Glu, or Thr, whereas Tyr96 was changed to either Phe
or Val (Fig. 5). These mutated constructs were then
assessed for their ability to complement the bacterial gpt
and hpt lesions within the S609 E. coli
genetic background. Allen et al. (10) have previously
demonstrated that the wild type L. donovani HGPRT can
complement these bacterial prt lesions.
As shown in Fig. 5, S609 cells transformed with wild type
HGPRT can grow in minimal medium supplemented with either
hypoxanthine or guanine. If either hypoxanthine or guanine is omitted
from the medium, S
609 cells cannot survive because of the genetic defect (purHJ) in the purine biosynthetic pathway that
confers purine auxotrophy. The adenine present in this minimal medium (see "Experimental Procedures") cannot serve as a source of
guanylate nucleotides in the presence of high histidine levels in the
media (11). In contrast, the S
609 cells transformants containing mutant hgprt genes exhibited three distinct growth
phenotypes (Fig. 5). The S95T transformant exhibited essentially a wild
type growth rate when hypoxanthine was provided as a purine and a
slightly reduced growth rate when guanine was added to the medium. The second phenotype was associated with the S95A and S95C mutations. These
transformants complemented the bacterial prt mutations
weakly, i.e. the cells grew, albeit slowly when either
hypoxanthine or guanine was the 6-oxypurine in the medium. The third
growth phenotype was observed with the transformants in which the
L. donovani hgprt contained S95E, Y96F, or a Y96V mutation.
These transformants were incapable of growth when either hypoxanthine
or guanine was provided as a purine source (Fig. 5). All strains grew
equivalently in minimal medium supplemented with guanosine, as the
nucleoside can be converted to the nucleotide level via a route
independent of HGPRT metabolism. Thus, this growth condition served as
a positive control for each transformant. The distinctive growth
phenotypes observed among the wild type and mutant transformants could
not be ascribed to differences in protein production as all
transformants grown in LPI medium supplemented with adenine plus
guanosine produced similar amounts of HGPRT/hgprt protein, as assessed
by Coomassie staining of cell lysates fractionated by
SDS-polyacrylamide gel electrophoresis (data not shown).
The effects of these single site mutations on the kinetics of the forward and reverse reactions of the enzyme were also characterized. Wild type HGPRT exhibited Km values of 6.1 ± 1.0 and 5.1 ± 0.8 µM for hypoxanthine and guanine, respectively. These values agree with those reported previously for the native (22, 23) and recombinant L. donovani HGPRT (10) enzymes. Although this enzyme exhibited similar affinities for both nucleobase substrates, the phosphoribosylation reaction was ~4 times more efficient with guanine than with hypoxanthine (Table III). The Km values of all of the mutant hgprt enzymes for hypoxanthine and guanine were similar to those obtained with the wild type protein and ranged from 4.7 ± 1.0 to 10.8 ± 2.4 µM for hypoxanthine and from 5.1 ± 0.8 to 12.2 ± 3.4 µM for guanine (Table III).
|
Km values for the PRPP substrate were also determined. A Km value for PRPP of 134 ± 22 µM was obtained for wild type HGPRT, a value similar to that obtained for the recombinant human HGPRT (24). All the mutations at Ser95 did not significantly alter the PRPP Km value (Table III). Surprisingly, however, mutations at Tyr96 dropped the Km value of the enzyme for PRPP ~4-5 fold. Km values of 28 ± 3.0 and 30 ± 7.5 µM were obtained for the Y96F and Y96V mutations, respectively.
The most striking aspect of the kinetics of the forward reaction,
however, was the dramatic decrease in the kcat
values observed with hgprt enzymes containing
Ser95-Tyr96 mutations. The
kcat for the wild type enzyme was 8.5 ± 0.8 s1 for hypoxanthine and 41.3 ± 1.8 s
1 for guanine (Table III). The conservative S95C and
S95T substitutions decreased the kcat of the
enzyme for both hypoxanthine and guanine ~2- and ~4-fold,
respectively. However, the hgprts encompassing the S95A or S95E
mutations exhibited a marked diminution in the turnover rate of the
enzyme of 1-2 orders of magnitude (Table III). Most remarkable were
the kcat values determined for the Y96F and Y96V
mutations. The kcat values determined for the
Y96F enzymes were 0.09 ± 0.015 and 0.12 ± 0.08 s
1 for hypoxanthine and guanine, respectively, while
those for Y96V were 0.072 ± 0.01 s
1 for
hypoxanthine and 0.04 ± 0.01 s
1 for guanine.
The possibility that the decreases in the turnover rates observed for the mutated L. donovani hgprts could be attributed to an increased affinity for one of the reaction products was assessed by studying the kinetics of the reverse reaction, i.e. the formation of hypoxanthine and PRPP from IMP and PPi. Using the xanthine oxidase-coupled enzyme assay described originally by Giacomello and Salerno (14), Km values for IMP and pyrophosphate of 90 ± 24 and 103 ± 16 µM were obtained for wild type HGPRT (Table IV). Similar Km values for both substrates were obtained for the S95E, S95T, Y96F, and Y96V hgprts (Table IV). As with the forward reaction, the most significant consequence of the site-directed mutations within the Ser-Tyr dipeptide was the ~2-450 fold decrease in the kcat for pyrophosphorolysis of IMP. The extent of the reduction in catalytic capability of an individual mutant in the reverse direction was roughly proportional to the decrease in kcat for the forward reaction (Tables III and IV).
|
The impact of protein modifying reagents and site-directed mutagenesis on the activity of L. donovani HGPRT has established an essential function for the Ser-Tyr dipeptide that is conserved among all members of the HG(X)PRT family. Both TNM and DEPC introduce bulky moieties onto Tyr side chains, presumably increasing steric hindrance at the active site cleft and thereby precipitating a loss of enzyme function. Moreover, addition of a NO2 group to the Tyr ring profoundly decreases the pKa of the phenolic hydroxyl from ~10.5 to ~7.2 (25-27). This pKa reduction at the Tyr96 of the L. donovani HGPRT would markedly disrupt the hydrogen bonding potential of this residue (26, 27) and also introduces a negative charge on the flexible loop at physiologic pH that could either impede PRPP entry into the catalytic pocket or prevent the loop from covering the active site during catalysis.
The requirement of Tyr96 for HGPRT activity was confirmed by mutational analysis of this residue. Mutant hgprts in which Tyr96 was changed to either Phe or Val exhibited substantial diminutions in kcat values for both the forward and reverse reactions and failed to complement bacterial hpt and gpt lesions. A similar study of Ser95 mutants also revealed a prominent role of Ser95 in HGPRT catalysis, although the effects of mutations at this position on the kcat value were not as dramatic as those observed for the Tyr96 mutations. Moreover, several of the Ser95 mutants could complement bacterial hpt and gpt lesions. That Ser95 is important for HGPRT activity is not particularly surprising in view of the fact that a S103R mutation at the analogous site in the human enzyme results in a marked diminution of enzyme activity and leads to the clinical manifestation of gout (28). Although no changes in the Km values of any of the Leishmania hgprt mutants for nucleobases were observed, it is interesting to note that the S103R mutation in the human enzyme increased the Km value for hypoxanthine 74-fold (28).
The kinetic data on the site-directed mutants revealed that the genetic
alterations created within the L. donovani HGPRT did not
affect the overall substrate specificity of the enzyme nor was the
affinity of the enzyme for nucleobase substrates appreciably affected,
although a consistent but small decrease in the PRPP Km value of the Tyr96 mutants was
obtained. Clearly, the kinetic parameter most significantly affected by
the mutations was the turnover rate, implying that this highly
conserved motif is integral to the catalytic mechanism. Gapp values (29) calculated from the
specificity constants (kcat/Km obtained for
guanine) revealed differences between wild type HGPRT and mutant hgprts
in the stabilization energy for enzyme-substrate complex. The
destabilizing effects of the mutations in the L. donovani
HGPRT can be separated into two groups. The first includes the S95C and
S95T mutants that exhibited
Gapp values of
3.2 and 4.3 kJ/mol, respectively. These small energetic changes can be
attributed to non-ideal geometries for hydrogen bond formation or to
diminished hydrogen bond dissociation energies (30). The second group
of destabilizing mutations includes the S95A, S95E, Y96F, and Y96V
mutants for which
Gapp values of 11.5, 10.5, 15.8, and 20.7 kJ/mol, respectively, were calculated. The
magnitudes of these free energy changes are commensurate with loss of a
hydrogen bond to a charged acceptor group (31), intimating that the
hydroxyl groups of both Ser95 and Tyr96 play
important roles in transition state stabilization.
The findings obtained with the L. donovani HGPRT incubated with protein modifying reagents contrast with previous studies on the human and Schistosoma mansoni HGPRT enzymes. The human and S. mansoni HGPRTs were inactivated by sulfydryl reagents (32, 33), whereas the L. donovani HGPRT was refractory to inactivation by either iodoacetate or p-chloromercuric benzoate. Covalent labeling experiments with [14C]iodoacetate identified the Cys22 and Cys25 residues of the human and schistosomal HGPRTs, respectively, as crucial for enzyme activity (33, 34). The insensitivity of the L. donovani enzyme to sulfydryl inactivation does not, however, support a general role for Cys residues in HGPRT catalysis.
In view of the multiplicity of conserved motifs among HG(X)PRT proteins, the failure of other protein modifying reagents to inactivate the L. donovani HGPRT is somewhat surprising (3). Several of these motifs are embedded within the core of the protein and may not be accessible to chemical modifying reagents (4). The Glu-Asp dipeptide that is situated within the PRPP binding motif that is found among all PRT family members (35) is, however, surface-exposed (4, 5, 9). Yet, the carboxyl-modifying reagent ethyl dimethylaminopropyl carbodiimide did not inactivate the protein. From the three-dimensional structures of HG(X)PRT proteins, it has been suggested that this Glu-Asp dipeptide functions in the binding of the ribose moiety of PRPP via a Mg2+ bridge and in the stabilization of the transition state complex (4, 9). It is possible that the free carboxyl moieties in the Glu-Asp dipeptide are shielded from chemical modification by either coordination to Mg2+ ions or by formation of internal salt bridges. The recently determined crystal structure of the Salmonella typhimurium OPRT (36) and the T. gondii HGXPRT (9) have revealed a Mg2+ ion complexed to this dyad.
The data with protein modifying reagents and mutants clearly confirm an
essential role of the Ser-Tyr dipeptide in HGPRT catalysis. The human
HGPRT (4) and T. gondii (9) HGXPRT crystal structures reveal
that this Ser-Tyr is located on a solvent-exposed loop that exhibits
weak electron density, intimating that this motif possesses a high
degree of flexibility. This loop consists of residues 103-117 of the
human enzyme and residues 96-106 of the T. gondii HGXPRT.
It has been conjectured that this loop functions in a structural role
to protect the oxocarbonium ion transition state of the enzyme from
nucleophilic attack by water (4). A similar protective function has
been proposed for an analogous loop (residues 98-119) in the S. typhimurium OPRT. Precedence for such a flexible loop functioning
in a gating capacity to occlude solvent molecules from catalytic
pockets has been observed for a variety of other proteins, including
triose-phosphate isomerase (39), human immunodeficiency virus type I
protease (38), and tyrosyl-tRNA synthetase (31). A second and perhaps
more significant role for the Ser-Tyr dipeptide may be in the formation
of the charge transfer complex in the transition state intermediate. Several lines of evidence support this idea. First, the structure of
OPRT co-crystallized with orotate and PRPP reveals that
Lys100 and Lys103 within the flexible loop of
OPRT are translocated to within 6 Å of the active site of the
enzyme-substrate complex (36) and may form contacts with the
PPi moiety of PRPP (39). The crystal structures of the
T. gondii HGXPRT apoenzyme and product-bound form also
reveal a shift of this loop toward the active site (9). Moreover, the
Arg100 and Arg102 residues within the human and
S. mansoni HGPRT proteins, respectively, can be specifically
cross-linked by the product analog 2,3
-GMP dialdehyde indicating that
these residues are close to the active site (33). These observations
coupled with the present data on the L. donovani enzyme
imply that the hydroxyl groups of the Ser and Tyr residues may function
as hydrogen bond donors. Alternatively, the Tyr96 may serve
as a general acid that can protonate the PPi moiety of
PRPP, thereby facilitating the scission of the C1-O bond of PRPP with
the concomitant formation of the C1-N9 amino glycoside linkage of the
nucleotide product. This latter mechanism may be similar to that
observed for the acid-catalyzed hydrolysis of glucose 1-phosphate (40).
Precedence for Tyr hydroxyls participating as general acids in reaction
mechanisms has been obtained for the ketosteroid isomerase of
Pseudomonas testosteroni (41) and the tyrosine
phenol-lyase of Citrobacter freundii (42). The role of a
general acid in PRT catalysis has also been proposed for the highly
conserved Lys103 of OPRTs, since mutation of this residue
in the S. typhimurium enzyme caused a 1000-fold decrease in
the kcat value of the enzyme without affecting
the Km values for either the forward or reverse
reaction (39).
We thank Dustin Lippert and Sandy Kielland of the University of Victoria Microchemistry Center for their technical expertise and assistance in performing the protein microsequencing and mass spectroscopy analysis.