Vrije Universiteit Brussel
Enzyme-Substrate Interactions in the Purine-Specific Nucleoside Hydrolase from
Trypanosoma vivax
Versees, Wim; Decanniere, Klaas; Van Holsbeke, Els; Devroede, Neel; Steyaert, Jan
Published in:
J. Biol. Chem.
DOI:
10.1074/jbc.M111735200
Publication date:
2002
Document Version:
Final published version
Link to publication
Citation for published version (APA):
Versees, W., Decanniere, K., Van Holsbeke, E., Devroede, N., & Steyaert, J. (2002). Enzyme-Substrate
Interactions in the Purine-Specific Nucleoside Hydrolase from Trypanosoma vivax. J. Biol. Chem., 277(18),
15938-15946. https://doi.org/10.1074/jbc.M111735200
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Enzyme-Substrate Interactions in the Purine-specific Nucleoside
Hydrolase from Trypanosoma vivax*
Received for publication, December 10, 2001, and in revised form, February 15, 2002 Published, JBC Papers in Press, February 19, 2002, DOI 10.1074/jbc.M111735200 Wim Verse´es‡, Klaas Decanniere, Els Van Holsbeke, Neel Devroede, and Jan Steyaert§
From the Department of Ultrastructure, Vlaams Interuniversitair Instituut voor Biotechnologie, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 Sint-Genesius-Rode, Belgium
Nucleoside hydrolases are key enzymes in the purine salvage pathway of Trypanosomatidae and are consid-ered as targets for drug design. We previously reported the first x-ray structure of an inosine-adenosine-guanosine preferring nucleoside hydrolase (IAG-NH) from Trypanosoma vivax (1). Here we report the 2.0-Å crystal structure of the slow D10A mutant in complex with the inhibitor 3-deaza-adenosine and the 1.6-Å crystal structure of the same enzyme in complex with a genuine substrate inosine. The enzyme-substrate complex shows the substrate bound to the enzyme in a different confor-mation from 3-deaza-adenosine and provides a snapshot along the reaction coordinate of the enzyme-catalyzed reaction. The chemical groups on the substrate important for binding and catalysis are mapped. The 2ⴕ-OH, 3ⴕ-OH, and 5ⴕ-OH contribute 4.6, 7.5, and 5.4 kcal/mol to kcat/Km,
respectively. Specific interactions with the exocyclic groups on the purine ring are not required for catalysis. Site-directed mutagenesis indicates that the purine spec-ificity of the IAG-NHs is imposed by a parallel aromatic stacking interaction involving Trp83and Trp260. The pH
profiles of kcatand kcat/Kmindicate the existence of one or
more proton donors, possibly involved in leaving group activation. However, mutagenesis of the active site resi-dues around the nucleoside base and an alanine scan of a flexible loop near the active site fail to identify this gen-eral acid. The parallel aromatic stacking seems to provide the most likely alternative mechanism for leaving group activation.
Nucleoside hydrolases (NHs)1 catalyze the irreversible
hy-drolysis of the N-glycosidic bond in ribonucleosides following
the reaction scheme: -purine (or pyrimidine) nucleoside ⫹ H2O 3 purine (or pyrimidine) base ⫹ ribose. The NHs are
members of a broader class of N-ribohydrolases and trans-ferases including the clinically important purine-nucleoside phosphorylases (2) and phosphoribosyl transferases (3, 4). Nu-cleoside hydrolases play a key role in the purine salvage path-way of protozoan parasites. This salvage pathpath-way is vital for these organisms because they lack a de novo purine biosynthe-sis pathway (5–7). Since neither nucleoside hydrolase activity, nor the encoding genes, have been found in mammals they constitute an attractive target for drug design against a variety of pathogens. On the basis of their substrate specificity the nucleoside hydrolases have been divided into three subclasses: the base aspecific inosine-uridine preferring nucleoside hydro-lases (IU-NH) (8), the purine-specific inosine-adenosine-guanosine preferring nucleoside hydrolases (IAG-NH) (1, 9), and the 6-oxo-purine specific inosine-guanosine preferring nu-cleoside hydrolases (IG-NH) (10).
The IU-NHs were scrutinized by V. Schramm and co-workers (11, 12). For this subclass of enzymes an Sn1 reaction
mecha-nism with a ribosyl oxocarbenium-like transition state was established by kinetic isotope effect studies (11, 12). Protona-tion of N-7 of the purine ring before reaching the transiProtona-tion state was proposed to lower the electron density in the purine ring, hence facilitating its departure. The x-ray crystal struc-tures of IU-NHs from Crithidia fasciculata (13) and Leishma-nia major (14) show a histidine (His241in the enzyme from C.
fasciculata) as the most likely candidate for the role of general acid (15). The general base activating the nucleophilic water molecule was identified as Asp10. The presence of a general
base abstracting a proton from the nucleophilic water and a general acid activating the leaving group is in accordance with the mechanism of the inverting O-glycosidases (16).
We recently reported the cloning, basic characterization, and crystal structure of the nucleoside hydrolase from Trypano-soma vivax (1), a protozoan parasite that is one of the most important causative agents of trypanosomiasis in African and South American cattle (17). This nucleoside hydrolase is the first thoroughly investigated representative of the subclass of the IAG-NHs. The structure of the enzyme in complex with the inhibitor 3-deaza-adenosine showed that all the residues inter-acting with ribose are fairly well conserved among the IU-NHs and the IAG-NHs. The aspartate residue (Asp10) functioning as
general base is also conserved. A much greater diversity is observed in the residues surrounding the nucleoside base, as could be expected from the different substrate specificity to-ward the leaving group. A striking feature of the IAG-NH structure, however, is the apparent lack of specific interactions with the purine base. No evident candidates were found for the role of proton donor, postulated to be essential for catalysis in the nucleoside hydrolases. The histidine proposed to fulfil this function in the IU-NHs is replaced by a tryptophan (Trp260) in
the IAG-NH, involved, together with Trp83, in aromatic stack-* This work was supported in part by Vlaams Interuniversitair
In-stituut voor Biotechnologie and the Nationaal Fonds voor Wetenschap-pelijk Onderzoek-Vlaanderen. Travel to and accommodation in Ham-burg was supported by European Union Grant HPRI-CT-1999 – 00017. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The atomic coordinates and structure factors (code 1KIE for the D10A-3-deaza-adenosine complex and 1KIC for the D10A-inosine complex) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).
‡ Supported by a grant from the FWO-Vlaanderen.
§ To whom correspondence should be addressed: ULTR, Ultrastruc-ture Department, IMOL, Instituut voor Moleculaire Biologie, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 Sint-Genesius-Rode, Belgium. Tel.: 32-2-3590248; Fax: 32-2-3590289; E-mail: jsteyaer@ vub.ac.be.
1The abbreviations used are: NH, nucleoside hydrolase; IAG-NH,
inosine-adenosine-guanosine preferring nucleoside hydrolase; IU-NH, uridine preferring nucleoside hydrolase; IG-NH, inosine-guanosine preferring nucleoside hydrolase; MES, 4-morpholineethane-sulfonic acid; ACES, N-(2-acetamido)-2-aminoethane4-morpholineethane-sulfonic acid; HPLC, high performance liquid chromatography.
© 2002 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
This paper is available on line at http://www.jbc.org
15938
by guest, on October 27, 2012
www.jbc.org
ing with the purine base. Only an aspartate (Asp40) located at
3.51 Å from the purine N-9 or a residue from a flexible loop not visible in the structure remain as candidates for general acid. In this article we focus on the substrate specificity of the T. vivax IAG-NH. The crystal structure, at 1.6-Å resolution, of a mutant IAG-NH with the natural substrate inosine bound in its active site is presented. This structure shows a different binding mode for the natural substrate from the inhibitor 3-deaza-adenosine. We quantified the contribution to binding and catalysis of the different chemical groups on the substrate. Elaborate site-directed mutagenesis studies are used to deter-mine the molecular basis of the specificity toward the purine ring. Finally, the question of the identity of the proton donor is addressed.
EXPERIMENTAL PROCEDURES
Production of the Mutant IAG-NHs—The T. vivax
inosine-adenosine-guanosine preferring nucleoside hydrolase (iagnh) open reading frame cloned into the BamHI-PstI restriction sites of the pQE-30 plasmid (Qia-gen) was used as the template for all in vitro mutagenic studies (1). Mutants were created using a primer-extension overlap PCR technique (18). A forward primer and a complementary reverse primer containing the mutated codon were used together with a general reverse and forward primer for the pQE-30 plasmid to generate the 3⬘ and 5⬘ fragments of the mutated open reading frame. A PCR with the general primers was used to link both fragments together. This last PCR product was recloned in the pQE-30 plasmid using the BamHI and HindIII restriction sites. A list of forward mutagenesis primers is given in Table I.
Expression and Purification—Expression and purification of the wild
type and mutant T. vivax IAG-NHs were performed as described pre-viously (1). E. coli cells (WK6) containing the iagnh open reading frame cloned in the pQE-30 expression vector were used to express the pro-tein. The presence of an N-terminal His6-tag allowed for a two-step
purification scheme, consisting of a Ni-NTA affinity chromatography step (Qiagen) and gel filtration on a Superdex-200 column (Amersham Bioscience).
Kinetic Analysis of Substrates—The kinetic properties of the IAG-NH
for the substrates xanthosine and purine riboside were determined with the reducing sugar assay as described previously (1, 9). Hydrolysis of adenosine, 2⬘-deoxyadenosine, and 5⬘-deoxyadenosine was monitored spectrophotometrically using the difference in absorption between the nucleoside and the purine base. A⌬⑀ value of ⫺1.4 mM⫺1cm⫺1was used at a wavelength of 276 nm. The conversion of 3⬘-deoxyadenosine to 3⬘-deoxyribose and adenine was followed on a C18 HPLC column (100 ⫻ 4.6-mm ODS HYPERSIL RP-C18, 5m) attached to a Waters HPLC system. Substrate and products were eluted with a linear gradient of acetonitrile⫹ 0.08% trifluoroacetic acid in 10 mMammonium acetate and monitored spectrophotometrically at 260 nm. A linear gradient from 4.5 to 15% acetonitrile over an elution volume of 6 ml led to an optimal separation of substrate and products. All measurements were carried out at 35 °C in a 50 mMphosphate buffer, pH 7.0. The data were
fitted to the Michaelis-Menten equation using the program Origin (Mi-crocal). All kinetic parameters were calculated per active site making them independent of the multimerization state of the enzyme.
Analysis of pH Data—The pH dependence of the steady state kinetic
parameters between pH 4.0 and 10.0 of the T. vivax IAG-NH was determined with inosine as a substrate. The steady state parameters for inosine were measured spectrophotometrically or using the HPLC method as described above. A standard curve for the⌬⑀ of inosine at 280 nm was determined as a function of pH. To maintain a constant ionic strength of 0.1Mover the whole pH range the measurements were done in a mixed buffer system. In the pH range 3.6 to 9 a buffer containing 50 mMacetic acid, 50 mMMES, and 100 mMTRIS was used. In the higher pH range the mixed buffer was composed of 52 mMTRIS, 52 mM ethanolamine, and 100 mMACES (19). The pH dependence of kcatand
kcat/Kmwere fitted to their respective equations,
kcat(obs)⫽ kcat(m)⫻ H2⫹ kcat(i)⫻ H ⫻ Ka1 H2⫹ H ⫻ K a1⫹ Ka1⫻ Ka2 (Eq. 1) kcat/KM(obs)⫽
kcat/KM(m)⫻ H2⫻ Ka1⫹ kcat/KM(i)⫻ H ⫻ Ka1⫻ Ka2
H3⫹ H2⫻ K
a1⫹ H ⫻ Ka1⫻ Ka2⫹ Ka1⫻ Ka2⫻ Ka3
(Eq. 2) where Kavalues are the acid dissociation constants, H is the hydrogen
ion concentration, kcat(m)and kcat/Km(m)are the maximal kcatand kcat/Km
values, and kcat(i)and kcat/Km(i)are the kcatand kcat/Kmvalues at the
intermediary plateau of the pH profiles.
Crystallization, Data Collection, and Processing—Crystals of the
D10A IAG-NH were grown using the hanging drop vapor diffusion method. Equal volumes of protein solution (8 mg/ml) and precipitant containing 1.6Mammonium sulfate in 100 mMTRIS buffer, pH 8.5, were mixed and equilibrated at 20 °C. For the 3-deaza-adenosine com-plex the crystals were soaked in a cryo-solution containing 40% PEG 6000 in a 100 mMHEPES buffer, pH 7.5, and 3 mM3-deaza-adenosine. After several hours crystals were transferred to the cryostream (100 K) and data were collected to a resolution of 2.0 Å on beamline BW7A (EMBL, DESY, Hamburg). In the case of the inosine complex the crystals were soaked in the same cryo-solution, now containing 50 mM inosine. After 3 min crystals were transferred to the cryostream and data were collected to a resolution of 1.6 Å on beamline ID-14b (ESRF, Grenoble).
The diffraction data were integrated, scaled, and merged with the HKL package (20). Intensities were converted to structure factors using TRUNCATE (21). Relevant statistics for both datasets are summarized in Table II.
Structure Determination and Refinement—The coordinates of the
wild type IAG-NH (pdb 1HOZ) were used as a search model to solve the structure of the D10A-inosine complex by molecular replacement using AMORE (22). After a first round of simulated annealing refinement (23), several cycles of positional and B-factor refinement using CNS (24) were alternated with manual rebuilding in TURBO-FRODO (25). A total of four inosine molecules, one in each active site of the dimer and two others in crystal packing interfaces, were found per asymmetric unit in 2Fo⫺ Fcand Fo⫺ Fcmaps. The inosine molecule bound in the
active site of monomer A of the dimer was refined with full occupancy on the ribose moiety of the nucleoside and with 30% occupancy on the hypoxanthine base. The inosine molecule bound in the other active site (monomer B) was refined with full occupancy on the ribose and 50% occupancy on hypoxanthine. This active site is better resolved and therefore used for further interpretation. The substitution of alanine for Asp10 was confirmed. Amino acids 247–256 were excluded from the
model in both subunits in the asymmetric unit, due to very weak electron density in this region. After refinement an R factor of 17.16% was obtained (free R factor of 19.39%).
The crystal structure of the D10A mutant in complex with 3-deaza-adenosine is isomorphous to the previous one. Therefore, refinement was started from the D10A-inosine model (without solvent and ligands), taking care to use the same set of reflections for cross-validation. The refinement procedure used was the same as stated above. The 2Fo⫺ Fc
and Fo⫺ Fcmaps in both active sites in the asymmetric unit showed
unambiguous electron density for 3-deaza-adenosine. As in the previous structure, amino acids 246 –257 were omitted from the model. The final model provided an R factor of 15.64% (free R factor 21.53%).
The quality of the models was checked with PROCHECK (26). Fig-ures were made with MOLSCRIPT (27) and CONSCRIPT (28).
RESULTS
Substrate Specificity—The steady state kinetic parameters of the T. vivax IAG-NH for the common nucleosides and some
TABLE I
List of mutagenesis primers
Only the “forward” mutagenesis primers are given. The “reverse” primers are the complement of these primers. The codon shown bold specifies the site of mutation.
Mutant Primer (5⬘ to 3⬘) D10A GGACCATGCTGGAAATCTAG N12A CTGGACCATGATGGAGCTCTAGACGACTTTGTC D40A CTCTGTACCGATGCGGCTTGCTTCGTTGAGAAT W83A CCCTTCCCGAAAGAAGCGCGGTGCCTGGCCAAG C245A GGTACAATGTGGGCAATGGCCACACACTGTGAGTTATTG T246A TGGGCAATGTGCGCACACTGTGAGTTAT H247A GTGGGCAATGTGCACAGCCTGTGAGTTATTGCGTG C248A ATGTGCACACACGCAGAGTTATTGCGTGAT E249A TGCACACACTGTGCCTTATTGCGTGAT R252A CACTGTGAGTTATTGGCTGATGGGGATGGCTACTAC D253A GAGTTATTGCGTGCAGGGGATGGCTAC D255A TTGCGTGATGGGGCAGGCTACTACGCCT Y257A GGGGATGGCGCCTACGCCTGGGAGGCACTC Y257F CGTGATGCGGATGGCTTCTACGCCTGGGACGCACTC Y258A GATGGGGATGGCTACGCCGCCTGGGACGCACTCAC Y258F GATGGGGATGGCTACTTCGCCTGGGACGCACTCAC W260A GATGGCTACTACGCCGCGGACGCACTCACTGCC
IAG-NH from T. vivax
15939
by guest, on October 27, 2012
www.jbc.org
nucleoside analogues are summarized in Table III. The kcat/Km
ratios show that the enzyme is 1,000 –10,000 times more spe-cific toward the naturally occurring purine nucleosides then toward the pyrimidine nucleosides (1). Xanthosine is the only exception to this rule. While its turnover is comparable with the other purine nucleosides, it binds to the enzyme with a Km
that is almost 3 orders in magnitude higher. Purine riboside, which lacks all exocyclic groups on the purine ring, is as good a substrate as the three natural purine nucleosides.
The kinetic constants of the deoxynucleosides show the crit-ical importance in catalysis of all three hydroxyl groups of the ribose moiety of the substrate. Removing the 2⬘-OH, 3⬘-OH, or the 5⬘-OH lowers kcat/Kmby factors of 1700, 2.3 ⫻ 105, and
7100, respectively. While the 2⬘-OH is contributing to catalysis only, the 3⬘-OH and the 5⬘-OH are important for both catalysis and substrate binding.
Effect of pH on the Kinetic Constants of the IAG-nucleoside Hydrolase—The pH dependence of kcatand kcat/Kmwas
deter-mined between pH 4.0 and 10.0 using inosine as a substrate (Fig. 1). The turnover rate of the enzyme-catalyzed reaction decreases between pH 4 and 10, with a plateau near pH 7. The data are consistent with the existence of two groups on the enzyme-substrate complex needed in a protonated state for full catalytic activity. Deprotonation of a first group with a pKaof 5.6⫾ 0.2 leads to a factor of 10 decrease in kcat. Deprotonation
of the second group with a pKa8.6⫾ 0.2 completely abolishes
the enzymatic activity. The pH profile of kcat/Km shows an
optimum near pH 5.0. Two groups on the free enzyme, with a pKaof 5.0⫾ 0.1 and 7.7 ⫾ 0.1, respectively, are needed in a
protonated state for optimal formation of the catalytically com-petent Michaelis complex. An additional third group with a pKa
of 5.1⫾ 0.1 needed in a deprotonated state appears in the plot of kcat/Kmversus pH. The Kmvalue is nearly constant between
pH 6 and 8 but increases quickly below and above this pH range.
Structure of Inhibitor and Substrate Complexes—In the pres-ent study two x-ray crystal structures of the D10A mutant (see next paragraph) of the T. vivax IAG-NH were solved: one in complex with the active site inhibitor 3-deaza-adenosine and one in complex with the substrate inosine. Since we previously solved the structure of the wild type enzyme in complex with 3-deaza-adenosine (1) this approach enabled us to determine
which features in the D10A-inosine complex were due to the mutation and which were caused by the substrate. For a de-tailed description of the overall fold and the active site archi-tecture of the wild type IAG-NH from T. vivax we refer to Ref. 1.
Structure of the D10A Mutant in Complex with the Inhibitor 3-Deaza-adenosine—The structure of the D10A mutant in com-plex with 3-deaza-adenosine consists of a homodimer with each subunit containing 10-strands, 12 ␣-helices, and three short 310helices (see Fig. 2). The overall fold of the protein is nearly
identical to the wild type enzyme in complex with 3-deaza-adenosine (Protein Data Bank entry code 1HP0, see Ref. 1), with an root mean square deviation of only 0.512 Å between a subunit of both enzymes (considering main chain atoms only). The 3-deaza-adenosine molecules found in both active sites of the dimer interact with a Ca2⫹ion at the bottom of the active
site through their 2⬘- and 3⬘-hydroxyl groups. The ribose adopts an O-1⬘ exo envelope conformation in one active site (0E, P⫽
274°) and a conformation between4E and 0
4T (P⫽ 245°) in the
other active site. In both active sites the purine adopts a syn-conformation toward the ribose. This syn-conformation is analo-gous to the conformation of 3-deaza-adenosine bound to the wild type IAG-NH (1). A superposition of the active sites of the wild type-3-deaza-adenosine complex and the D10A-3-deaza-adenosine complex is shown in Fig. 3A.
Structure of the D10A Mutant in Complex with the Natural Substrate Inosine—The turnover number of the D10A IAG-NH with inosine as a substrate is 0.00058 s⫺1at pH 7.0 (see next paragraph). This means that it takes the enzyme about 28 min to convert an enzyme bound substrate to the products. This allowed us to collect a dataset of the mutant enzyme in complex with inosine at 100 K. At this temperature the enzymatic reaction is stopped and binding and product release are also inhibited (29).
The overall fold of the D10A-inosine structure is the same as that found for the structure in complex with 3-deaza-adenosine (Fig. 2). No major conformational changes appear upon binding of the substrate inosine and all the amino acid residues in the active sites of both complexes superimpose nearly perfectly (see Fig. 3B).
The inosine molecules in the active sites are bound to the enzyme with the ribosyl adopting a C-4⬘ endo envelope
confor-TABLE II
Data collection and refinement statistics
D10A-inosine D10A-3deaza-adenosine Diffraction data Space group P21 P21 a (A˚ ) 54.35 53.27 b (A˚ ) 74.62 72.96 c (A˚ ) 73.02 71.53  (deg) 98.31 98.18 Resolution range (A˚ )a 50.0 – 1.6 (1.66 – 1.60) 20.0 – 2.0 (2.07 – 2.00) Rsym(%) a 4 (19.5) 6.5 (34.9) I/Ia 28.4 (6.1) 13.5 (3.5) Completeness (%)a 92.7 (60.4) 97.3 (95.2) Structure refinement Rcryst(%) 17.61 15.64 Rfree(%) 19.39 21.53
Rmsd for bond lengths (A˚ ) 0.0156 0.0101
Rmsd for bond angles (deg) 1.780 1.859
Ramachandran plot (% most favored, allowed, generously allowed, disallowed residues)
91.2, 8.6, 0.2, 0.0 87.9, 10.8, 1.3, 0.0
No. atoms per atomic unit 5651 5515
Average B-factors (A˚2
) (protein, water) 19.49, 33.43 24.95, 36.60
a
Values in parentheses are for the highest resolution shell
Rsym⫽ ⌺ 兩 Ii(hkl)⫺ (Ii(hkl)兩 / ⌺Ii(hkl), where Ii(hkl) are the intensities of multiple measurements and (Ii(hkl)) is the average of the measured
intensities for the ith reflection.
Rcryst⫽ 100 ⫻ (⌺h兩Fobs,h⫺ Fcalc,h兩 / ⌺hFobs,h), where Fobsand Fcalcare observed and calculated structure amplitudes, respectively.
Rfree⫽ Rcrystcalculated for the test set of reflections not used in refinement.
by guest, on October 27, 2012
www.jbc.org
mation (P ⫽ 236°) (see Fig. 4). In contrast with the 3-deaza-adenosine ligands, the purine ring is oriented in an anti-con-formation toward the ribose ( ⫽ ⫺176°) (see Fig. 3B). When the structure was refined without restraints on the scissile C-1⬘-N-9 bond length this bond was extended to an aberrantly long 1.67 Å.
The ribose is involved in enzyme-substrate interactions in the same way as in the D10A-3-deaza-adenosine complex (Fig. 5). Apart from the interaction with the Ca2⫹ion in the bottom
of the active site, the 2⬘-OH also interacts with Asp14and the
3⬘-OH forms contacts with Asp261 and Asn186. The 5⬘-OH in
turn hydrogen bonds to Asn173and Glu184. As in the
3-deaza-adenosine complex, the purine base (in this case hypoxanthine) is stacked between the side chains of Trp83and Trp260.
How-ever, the nearly 180° rotation (anti versus syn) of the nucleoside
base causes some new interactions to appear: Asn12is located
2.55 Å from the N-3 of the hypoxanthine base and Asp40is
located within 3.02 Å of the anomeric carbon. A water molecule is found to hydrogen bond with the N-7 of the nucleoside base. No interactions are made between the enzyme and the N-1 or the exocyclic keto group at position six of the purine base.
In addition to the two inosine molecules found in the active sites, two other inosine molecules, bound in crystal packing in-terfaces, were incorporated in the model. In both of these mole-cules the ribose adopts a C-4⬘ endo, C-3⬘ exo twist conformation (34T), while the nucleoside base is oriented in a syn-conformation
toward the ribose. When refined with low energy restraints on the N-glycosidic bond length, values of 1.51 Å and 1.42 Å are found for the C-1⬘-N-9 bond length for both inosines, respectively. Site-directed Mutagenesis—Several active site residues visi-ble in our crystal structures were replaced by alanine via site-directed mutagenesis as described under “Experimental Proce-dures.” In Table IV the kinetic parameters of these mutants with inosine as a substrate are compared with the correspond-ing values for the wild type enzyme at pH 7.0 and/or 5.0 (i.e. the optima for Kmand kcat). Asp10 has been proposed to be the
general base in the reaction mechanism of the nucleoside hy-drolases (1). Upon mutation to alanine the Kmvalue is
rela-tively unchanged while kcatdecreases by a factor of 1000 (at pH 5.0) to 9000 (at pH 7.0). Four residues, Asn12, Asp40, Trp83, and
Trp260, are situated within interaction distance of the purine-TABLE III
Steady state kinetic parameters for the wild type T. vivax nucleoside hydrolase
Substrate/Inhibitor kcat KM Klc s⫺1 M M Inosinea 5.19⫾ 0.08d 5.37⫾ 0.42 Adenosine 2.58⫾ 0.18 8.00⫾ 1.80 Guanosinea,b 1.9 3.8 Xanthosine 2.35⫾ 0.24 1760⫾ 353 Purine riboside 3.92⫾ 0.14 3.79⫾ 1.11 Cytidinea 0.338⫾ 0.005 925⫾ 39 Uridinea 0.022⫾ 0.002 586⫾ 150 p-Nitrophenylribosidea,c 0.206⫾ 0.005 257⫾ 18 2⬘-Deoxyadenosine 0.0043⫾ 0.0002 22.64⫾ 4.32 3⬘-Deoxyadenosine 0.00031⫾ 0.000005 223.36⫾ 14.84 5⬘-Deoxyadenosine 0.030⫾ 0.002 656.26⫾ 119.12 3-Deaza-adenosinea 0.2⫾ 0.02 7-Deaza-adenosinea 356.5⫾ 30.3 a
Data from Ref. 1.
bErrors on k
catand KMcould not be determined (see Ref. 1).
cp-Nitrophenyl--D-ribofuranoside.
dAll parameters were measured in 50 mMphosphate buffer, pH 7.0, at 35 °C.
FIG. 1. Effect of pH on the kinetic constants of the IAG-NH from T. vivax, using inosine as a substrate. a, variation of kcatwith
pH. b, variation of kcat/Kmwith pH. The pH dependence of Kmis not
shown. The ordinate scale is s⫺1for kcatandM⫺1s⫺1for kcat/Km. The
data were fitted to the appropriate equations (see “Experimental Pro-cedures”). The best fit parameters are given in the text.
FIG. 2. Structure of the D10A mutant of the IAG-NH from T.
vivax in complex with inosine. The inosine molecules located in each
active site of the IAG-NH dimer are shown as ball-and-stick models, the calcium ions are depicted as blue spheres. Amino acids 245–256 were excluded from the model. Arrows indicate the position of the flexible loop containing these amino acids in one of the subunits of the IAG-NH dimer.
IAG-NH from T. vivax
15941
by guest, on October 27, 2012
www.jbc.org
leaving group in the structure of the D10A-inosine complex (Fig. 5). Mutation of Asn12and Asp40to alanine causes only a
minor increase in Kmand no change in kcat. Both tryptophans
are, however, essential for the functioning of the IAG-NH. While Trp83only contributes a factor of 20 to k
catand a factor
of 270 to Km, replacement of Trp260with alanine causes a loss
of almost a factor of 85 in turnover and an increase in Kmof a
factor of 3700. The turnover rate with the pyrimidine nucleo-side uridine is 0.00029 s⫺1⫾ 0.00002 s⫺1and 0.00013 s⫺1⫾ 0.00002 s⫺1for W83A and W260A, respectively, while the Km
values for this substrate are 1013.3 ⫾ 273.2M and 510.2⫾ 228.0Mfor W83A and W260A, respectively.
In all the three-dimensional structures of IAG-NHs deter-mined thus far, 12 amino acids (Cys245to Tyr257) located near the
active site were omitted because of weak electron density (see above and Fig. 2). In accordance with, among others, the catalyt-ically related phosphoribosyl transferase (30, 31), a role in catal-ysis was previously assigned to this loop in a type of induced fit mechanism (1). To test this hypothesis all amino acids containing functional side chains in this loop were mutated to alanines. Tyr257and Tyr258were additionally mutated to phenylalanines.
The results of this study are summarized in Table IV. Five
mutants, C245A, T246A, R252A, Y257A, and Y258A, show a significant increase in Km. The effect of both tyrosines on Kmis
clearly due to the aromatic moiety of the side chain since muta-tion to phenylalanine has no effect. None of the “flexible loop” mutants seem to have a negative effect on catalysis. On the contrary replacing Arg252 with alanine increases the turnover
almost 5-fold. The same phenomenon is observed, although to a lesser extend, in the Y257A and Y258F mutants.
DISCUSSION
Inhibitor Versus Substrate Binding in the Crystal Structure of the T. vivax IAG-NH
Previous reports of x-ray structures of nucleoside hydrolases dealt with uncomplexed enzymes or enzymes in complex with an inhibitor (e.g. IU-NH from C. fasciculata in complex with p-aminophenyliminoribitol (13); or IAG-NH from T. vivax in complex with 3-deaza-adenosine (1)). Since these inhibitors differ, by definition, to a greater or lesser extent from the real substrates, far-reaching conclusions concerning catalysis drawn from such structures should be regarded with caution.
Here we present the structure of a slow mutant enzyme, D10A, in complex with inosine, the natural substrate of the enzyme. Since Asp10 corresponds to the general base in the
reaction mechanism of the nucleoside hydrolases and increases the nucleophilicity of the attacking water molecule, replacing Asp10by an alanine will slow the attack of the water molecule
on the oxocarbenium-like transition state. The 9000-fold de-crease in kcat of the D10A mutant allowed us to solve its structure in complex with inosine. A second structure of the same mutant in complex with the inhibitor 3-deaza-adenosine allowed for a straightforward interpretation of the changes observed between the D10A-inosine complex and the previ-ously reported structure of the wild type IAG-NH in complex with 3-deaza-adenosine.
Comparison of the structures of the wild type enzyme and the D10A mutant complexed with 3-deaza-adenosine shows a nearly identical position and conformation of the ligand in the
FIG. 3. A, superposition of the active sites of the wild type T. vivax IAG-NH in complex with the inhibitor 3-deaza-adenosine (PDB 1HP0) and the D10A mutant of the same enzyme in complex with 3-deaza-adenosine. The active site residues and the inhibitor are shown in light gray for the wild type enzyme and in dark gray for the mutant. The Ca2⫹ion is shown in dark blue (transparent for the wild type IAG-NH) and the
nucleophilic water is shown as a red sphere for the wild type IAG-NH and as a light blue sphere for the D10A mutant. Replacing of Asp10by an
alanine causes the catalytic water to move away from the scissile bond and causes a shift of the side chains of Asp40and Trp83. B, superposition
of the active site of the D10A mutant in complex with 3-deaza-adenosine and the active site of the same mutant in complex with the substrate inosine. The same color codes as above are used for the 3-deaza-adenosine complex. The active site residues of the D10A IAG-NH in complex with inosine are shown in yellow. The carbon atoms of inosine are colored orange. The active site residues superimpose nearly perfectly. However, the nucleoside base of inosine adopts an anti-conformation as compared with the syn-conformation of the base in 3-deaza-adenosine.
FIG. 4. Foⴚ Fcmap around an inosine in one of the active sites
of the D10A IAG-NH. A, Fo⫺ Fcmap contoured at 3. B, Fo⫺ Fcmap
contoured at 4.5. The C-4⬘ endo envelope conformation of the ribose is shown.
by guest, on October 27, 2012
www.jbc.org
active sites of both structures. However, deleting the carboxyl of Asp10 in the D10A mutant causes the proposed catalytic
water to shift about 0.5 Å away from the scissile bond (Fig. 3A). This causes in turn a shift of the Asp40side chain of about 2.23
Å. The void left by the Asp40side chain is filled by the Trp83
side chain which has moved 3.1 Å compared with the wild type structure. This structural rearrangement may contribute to the dramatic effect of the D10A mutation on kcat.
No conformational changes in the enzyme can be observed between the inosine and 3-deaza-adenosine complexed mutant structures. Striking, however, is the difference in ligand bind-ing. While the 3-deaza-adenine base is bound in a syn-confor-mation toward the ribose, the hypoxanthine base of inosine is bound in the anti-conformation (Fig. 3B). In this conformation the N-3 of the hypoxanthine base is located 2.55 Å from the carboxamide group of Asn12, while Asn12is not in the proximity
of the nucleoside base in the D10A-3-deaza-adenosine struc-ture. One explanation for the conformational change is that a catalytically important hydrogen bond between Asn12and N-3
keeps the nucleoside base in the anti-conformation in common substrates. Substitution of the N-3 by a carbon atom, as in 3-deaza-adenosine, would abolish this hydrogen bond and would lead to a nucleoside base bound in the syn-conformation. If this hypothesis were correct substituting Asn12would have
the same effect as replacing N-3 in the nucleoside base. How-ever, the N12A mutation does not affect kcatand has only a
small effect on Km(Table IV), indicating that Asn12is close to
N-3 but does not strongly interact with it. Another possible explanation for the 3-deaza-adenosine conformation and inhib-itory effect is that a negative interaction (i.e. a steric clash) between Asn12and the hydrogen atom (H-3) connected to C-3
in 3-deaza-adenosine forces binding in the, presumably cata-lytically incompetent, syn-conformation. In both the syn- and anti-conformations the nucleoside base is stacked between the side chains of Trp83and Trp260accounting for the strong
bind-ing of 3-deaza-adenosine (KI⫽ 0.2M).
Substrate Binding Versus Transition State Binding
Both inosines in the active sites of the D10A-inosine complex are bound with their ribosyl group adopting a C-4⬘ endo-con-formation (Fig. 4), which is energetically less favorable than the commonly observed C-3⬘ endo and C-2⬘ endo-conformations (32). This conformation is, however, very close to the C-3⬘ exo-conformation expected to be adopted by the oxocarbenium transition state of an Sn1 mechanism (11). A nearly identical
C-4⬘ endo-conformation was observed for the iminoribitol moi-ety of p-aminophenyliminoribitol bound in the active site of the IU-NH of Crithidia fasciculata. A pseudorotation phase angle, P, equal to 236° is observed for inosine in our structure, while P is 233° for p-aminophenyliminoribitol in the IU-NH (13). p-Aminophenyliminoribitol is a proposed transition state ana-logue where the ribose part of the nucleoside is replaced with an iminoribitol group mimicking the main features of the oxo-carbenium ion.
FIG. 5. Close up stereoview showing the active site of the D10A mutant of the IAG-NH in complex with inosine. The carbon atoms of inosine are shown in
orange. Trp83and Trp260, which stack to
the purine base of the substrate, are shown in yellow. The side chains of Ala10,
Asn12, and Asp40are shown in dark gray.
The other active site residues are shown in light gray. The Ca2⫹ion in the bottom of the active site is depicted as a dark blue
sphere. The nucleophilic water and the
water molecule hydrogen bonded to the N-7 of the nucleoside base are depicted as
light blue spheres.
TABLE IV
Steady state kinetic constants of mutants with inosine as substrate
Mutant pH 5.0
a pH 7.0b
kcat KM kcat KM
s⫺1 M s⫺1 M
Active site mutants
Wild type 20.81⫾ 0.63 35.71⫾ 3.77 5.19⫾ 0.08 5.37⫾ 0.42 D10A 0.019⫾ 0.0009 74.62⫾ 10.08 0.00058⫾ 0.00002 28.14⫾ 2.69 N12A 19.09⫾ 1.29 110.81⫾ 21.30 8.50⫾ 0.68 37.75⫾ 9.01 D40A 12.16⫾ 0.91 113.30⫾ 21.38 3.53⫾ 0.12 34.9⫾ 4.5 W83A NAc NA 0.27⫾ 0.006 1470.5⫾ 71.6 W260A NA NA 0.061⫾ 0.003 19802.6⫾ 2662.6
Flexible loop mutants
C245A 19.28⫾ 1.50 323.91⫾ 66.03 NA NA T246A 16.31⫾ 0.71 154.42⫾ 20.35 NA NA H247A 21.65⫾ 1.21 39.47⫾ 7.12 NA NA C248A 21.00⫾ 1.97 25.49⫾ 7.45 NA NA E249A 13.45⫾ 2.12 62.12⫾ 26.26 NA NA R252A 95.86⫾ 4.58 193.43⫾ 28.61 NA NA D253A 14.16⫾ 1.63 43.74⫾ 16.31 NA NA D255A 17.44⫾ 0.91 27.43⫾ 4.85 NA NA Y257A 74.68⫾ 6.33 104.15⫾ 30.31 NA NA Y257F 11.93⫾ 0.58 25.57⫾ 5.25 NA NA Y258A 16.50⫾ 1.43 414.05⫾ 111.03 NA NA Y258F 36.60⫾ 2.38 86.60⫾ 13.02 NA NA
aMeasured in 50 mMacetate buffer, pH 5.0, at 35 °C. bMeasured in 50 mMphosphate buffer, pH 7.0, at 35 °C. c
NA, not applicable.
IAG-NH from T. vivax
15943
by guest, on October 27, 2012
www.jbc.org
Another interesting structural feature is the length of the N-glycosidic bond. When the inosines in the active sites of IAG-NH were refined with a fixed C-1⬘-N-9 bond length of 1.46 Å, the hypoxanthine base did not fit the density well. When the restraints on this bond length were relaxed, the distance be-tween C-1⬘ and N-9 increased to 1.67 Å, while the correspond-ing bonds in the “control” inosines, situated in the crystal packing, were unaffected. In the transition state of the nucle-oside hydrolases the C-1⬘-N-9 bond is expected to be nearly broken with the N-9 located at 2.0 Å from the anomeric carbon (11). These data suggest that the inosines are bound in a reactive conformation in the active sites of our structure. Such a conformation may be considered as a species along the reac-tion coordinate of the IAG-NH-catalyzed reacreac-tion prior to reaching the transition state.
Substrate Specificity
Ribose Specificity—The energy needed to force a ribose into the C-4⬘ endo-conformation as observed in the active sites of the D10A-IAG-NH would be of the order of 2.5–3 kcal/mol (33). The steady state kinetic parameters of 2⬘-deoxyadenosine, 3⬘-deoxyadenosine, and 5⬘-deoxyadenosine indicate that all hy-droxyl groups are involved in stabilizing this intrinsically un-favorable conformation (Table III). The 2⬘-OH is within interacting distance of Asp14(Fig. 5). Removal of this hydroxyl
increases the activation energy to be surmounted by⬃3.9 kcal/ mol, while binding of the ground state remains roughly unaf-fected. This means that the Asp14-2⬘-OH interaction is involved
only in catalysis and is used to distort the ribose ring. The 3⬘-OH, which interacts with Asn186and Asp261, and the 5⬘-OH,
which interacts with Asn173and Glu184, are extremely
impor-tant for catalysis and substrate binding. With the 2⬘-OH con-tributing 4.6 kcal/mol, the 3⬘-OH contributing 7.5 kcal/mol, and the 5⬘-OH contributing 5.4 kcal/mol to the catalytic efficiency (kcat/Km) we can state that the three hydroxyl groups are
indispensable in the mechanism of the IAG-nucleoside hydrolases.
Molecular Basis of the Purine Specificity—While the IU-NHs do not discriminate between purine and pyrimidine nucleo-sides, the IAG-NHs are truly purine specific (see Table III). Yet, the only requirement for efficient catalysis seems to be the presence of the purine ring, since the hydrolysis of purine riboside, which lacks all exocyclic groups on the purine ring, is as effectively catalyzed as the other purine nucleosides. The only exception to this rule is xanthosine which is negatively charged at physiological pH (34). This charge will certainly hamper the binding of this compound in the highly negatively charged active site pocket of the IAG-NHs.
From the crystal structures it is clear that the IAG-NH from T. vivax has only four amino acids, Asn12, Asp40, Trp83, and
Trp260, at its disposal to discriminate between purine and
pyrimidine nucleosides (Fig. 5). Replacing Asn12 and Asp40
with alanines has very little effect on turnover and substrate binding with inosine as a substrate (Table IV), making it un-likely that these residues are responsible for purine specificity. On the contrary, the W83A and W260A mutants are severely impaired in catalysis (Table IV). The indole rings of these tryptophans are involved in a face-to-face aromatic stacking interaction with the purine base. Trp83 is aligned perfectly
parallel to the purine base, while the indole ring of Trp260is
somewhat tilted (Fig. 5). Since such a parallel stacking inter-action is much more favorable with heterocyclic purines than with monocyclic pyrimidines (35, 36) these interactions could contribute to the purine specificity. In this regard it should be noted that in the base-aspecific IU-NHs, Trp83is replaced with
an isoleucine and Trp260with a histidine (1). We compared the
steady state kinetic parameters of the wild type IAG-NH with a purine nucleoside (inosine) and a pyrimidine nucleoside (uri-dine) with the same parameters for the W83A and W260A mutants and obtained the thermodynamic cycles shown in Fig. 6. The wild type enzyme is a factor 25,760 more specific (i.e. a 25,760 times higher kcat/Km) toward inosine than to uridine.
This corresponds to a free energy difference of 6.24 kcal/mol, with the difference in turnover rate accounting for 3.36 kcal/ mol, and the difference in binding for 2.88 kcal/mol. For the W83A mutant (Fig. 6a) the free energy difference between inosine and uridine is reduced to 3.97 kcal/mol, which corre-sponds to a loss in purine specificity of 2.27 kcal/mol. This difference can be totally assigned to substrate binding. The difference in turnover rate between inosine and uridine is unaffected by the mutation. The same effect is observed for the W260A mutant (Fig. 6b). For this mutant the free energy difference between inosine and uridine is reduced to 1.54 kcal/
FIG. 6. “2-Enzyme-2-substrate” thermodynamic cycles show-ing the contribution of Trp83and Trp260to the purine specificity of the IAG-NHs. a, thermodynamic cycle using the wild type IAG-NH and the W83A mutant in combination with a purine nucleoside (in-osine) and a pyrimidine nucleoside (uridine). b, thermodynamic cycle using the wild type IAG-NH and the W260A mutant in combination with inosine and uridine. The coupling ⌬⌬G gives the difference in purine specificity (subdivided in a catalytic and binding part) between the wild type IAG-NH and the mutant IAG-NH. kcatvalues are given in
s⫺1, K
m values inMand kcat/Kmvalues inM⫺1s⫺1. The Gibbs free
energy differences (⌬G and ⌬⌬G) are given in kcal/mol. The abbrevia-tions used are⌬Gcat, catalytic free energy difference (i.e. related to kcat
or activation energy difference);⌬Gbind, binding free energy difference
(i.e. related to Km);⌬Gtotal, total free energy difference (i.e. related to the
specificity constant kcat/Km).
by guest, on October 27, 2012
www.jbc.org
mol. The loss in purine specificity now amounts to 4.71 kcal/ mol. The difference in turnover between inosine and uridine remains unaffected by the mutation, but this mutant has by far the lowest Kmfor the pyrimidine nucleoside. In conclusion, the
parallel aromatic stacking between Trp83, the nucleoside base,
and Trp260contributes significantly to the turnover of purine
and pyrimidine nucleosides, but discriminates between purines and pyrimidines at the level of substrate binding. The discrim-ination by the IAG-NHs between purine and pyrimidine nucleosides on the level of turnover remains unaccounted for.
Is a General Acid Involved in the Enzyme Mechanism?
Kinetic isotope effect studies on the IU-NH of C. fasciculata have established that in the reaction mechanism of these en-zymes the N-7 of the purine ring is protonated prior to reaching the transition state (11). Protonation of the leaving group low-ers the electron density in the purine ring, hence pulling the electrons of the N-glycosidic bond. Protonation of the leaving group is considered to be a general feature in the mechanism of many N-glycosidic bond cleaving enzymes (e.g. purine nucleo-side phosphorylases (2), AMP nucleosidases (37), and the re-verse reaction of phosphoribosyl transferases (38)). For the IU-NHs a histidine was proposed to function as the general acid (15).
The pH dependence of kcatfor the T. vivax IAG-NH shows
that two residues must be protonated for full enzymatic activ-ity. In the pH profile of kcat/Kmapparently the same two groups
are observed together with an additional residue needed in a deprotonated state. The most likely candidates for this last residue are Glu184or Asp261which are expected to contribute to
the substrate binding via interaction with the 5⬘-OH and 3⬘-OH, respectively (Fig. 5). The identity of the two protonated residues is less obvious. In analogy with the IU-NHs, a role as general acids in protonating the purine leaving group, was ascribed to them previously (1, 9). For the “low pKa” residue
(pKa⫽ 5.6) an aspartate or glutamate with a raised pKaor a histidine residue with reduced pKacould be considered. A
his-tidine, a tyrosine, or a cysteine residue could account for the “high pKa” group. However, only Asn12, Asp40, Trp83, and
Trp260are within interaction distance of the purine base. Also,
a fixed water molecule is hydrogen bonded to the N-7 of the hypoxanthine base in the D10A-inosine structure (Fig. 5). His241 in the IU-NH, the proposed general acid in these
en-zymes, is replaced with Trp260in the IAG-NH. Asp40, which a
priori would be a good candidate for the low pKageneral acid,
can be ruled out since the D40A mutation has little effect on kcator Km(see Table IV).
A fixed water molecule such as the one hydrogen bonded to the N-7 of the nucleoside base has been proposed to act as general acid in the thymidylate synthase from Lactobacillus casei (39). However, in that enzyme the water molecule is itself hydrogen bonded to a histidine, which might lower its pKa. In the IAG-NH the water molecule hydrogen bonded to N-7 inter-acts only with other water molecules. During the hydrolysis of the N-glycosidic bond the pKaof the N-7 changes from 2.3 in
the substrate inosine to 8.5 in the product hypoxanthine (31). It is difficult to imagine a water molecule, itself with a pKaof
15.7, as a general acid without any aid from an amino acid. It is possible that the proton donor(s) required for chemical turnover are located in the part of the structure that is poorly defined in the electron density map. This region contains a number of ionizable residues (Cys245, His247, Cys248, Glu249,
Arg252, Asp253, Asp255, Tyr257, and Tyr258) that could serve
catalytic functions. The loop would then have to undergo a conformational change to bring these groups into position for catalysis. An analogous mechanism was recently described in
the APRTase of Saccharomyces cerevisiae, where a catalytically important glutamate was identified in a flexible loop via mu-tagenesis (30). To test this hypothesis for the IAG-NH, we performed an alanine scan on all the amino acids bearing a functional side chain in the flexible loop (see Table IV). None of these mutants was significantly impaired in its catalytic abil-ities, ruling out any of these residues as general acids.
All these data, although contradictory to the pH profile and the higher kcat with purines as compared with pyrimidines, seem to suggest that the IAG-NH from T. vivax functions without the involvement of a general acid activating the leav-ing group. More complex mechanisms, such as conformational changes, could underlie the observed pH profile. As an alter-native we propose a mechanism in which the leaving group is partly activated without the involvement of a genuine proton donor, through the parallel aromatic stacking with Trp83and
Trp260. Apart from their role in substrate binding and purine
specificity, these tryptophan residues are also important for turnover (Table IV). A face-to-face aromatic stacking geometry arises from the interaction between the highest occupied mo-lecular orbital of the donor ring (Trp) and the lowest unoccu-pied molecular orbital of the acceptor ring (purine). Using an indole-adenine system it has been shown that protonation of the nucleoside base lowers the lowest unoccupied molecular orbital energy, leading to a strengthening of the highest occu-pied molecular orbital-lowest unoccuoccu-pied molecular orbital mixing and thus a reinforcement of the stacking interaction (36, 40). As such, the parallel stacking interaction would strengthen while the IAG-NH-catalyzed reaction proceeds from an unprotonated ground state to a tentative protonated tran-sition state. This would have the effect of increasing the pKaof
the purine base and would explain the contribution of Trp83
and Trp260to catalysis.
Conclusions
In this paper we reported the crystal structures of the D10A mutant of the IAG-NH from T. vivax in complex with the inhibitor 3-deaza-adenosine and the substrate inosine. The D10A-inosine complex shows a snapshot of a pre-transition state species along the reaction coordinate of the enzyme-cat-alyzed reaction. The danger of overinterpreting enzyme-inhib-itor complexes is pointed out, as the natural substrate inosine is bound in a different conformation to the active site as the substrate analogue inhibitor, 3-deaza-adenosine. We also ex-amined the substrate specificity of the IAG-NH. While all three hydroxyls of the ribose contribute significantly to catalysis, the presence of a purine ring seems to be the only requirement for an effective leaving group. The purine specificity of the IAG-NHs is imposed by a parallel aromatic stacking interaction involving Trp83and Trp260. Mutating one of these tryptophans
to alanine drastically reduces the preference of the IAG-NHs for the purine nucleoside inosine over the pyrimidine nucleo-side uridine. The lack of specific interactions between the en-zyme and the nucleoside base of the substrate is striking. The presence of a general acid involved in protonating the leaving group was inferred from the pH profile and previous studies with the IU-NH from C. fasciculata. Site-directed mutagenesis of the active site residues and an alanine scan of a flexible loop in the vicinity of the active site failed to identify this proton donor. An alternative mechanism to activate the leaving group, by increasing its pKa, through the parallel aromatic stacking interaction is proposed.
Acknowledgments—We thank Maia De Kerpel and Elke Brosens for
excellent technical assistance. We acknowledge the use of beamlines BW7A at the EMBL, Hamburg, and ID-14b at the ESRF, Grenoble.
IAG-NH from T. vivax
15945
by guest, on October 27, 2012
www.jbc.org
REFERENCES
1. Verse´es, W., Decanniere, K., Pelle´, R., Depoorter, J., Brosens, E., Parkin, D. W., and Steyaert, J. (2001) J. Mol. Biol. 307, 1363–1379
2. Bzowska, A., Kulikowska, E., and Shugar, D. (2000) Pharmacol. Ther. 88, 349 – 425
3. Seegmiller, J. E., Rosenbloom, F. M., and Kelley, W. N. (1967) Science 155, 1682–1684
4. Aronov, A. M., Munagala, N. R., de Montellano, P. R. O., Kuntz, I. D., and Wang, C. C. (2000) Biochemistry 39, 4684 – 4691
5. Ogbunude, P. O. J., Ikediobi, C. O., and Ukoha, A. I. (1985) Ann. Trop. Med.
Parasit. 79, 7–11
6. Hammond, D. J., and Gutteridge, W. E. (1984) Mol. Biochem. Parasit. 13, 243–261
7. Berens, R. L., Krug, E. D., and Marr, J. J. (1995) in Biochemistry and
Molec-ular Biology of Parasites (Marr, J. J., and Mu¨ ller, M., eds) pp. 89 –117, Academic Press, New York
8. Parkin, D. W., Horenstein, B. A., Abdulah, D. R., Estupin˜ a´n, B., and Schramm, V. L. (1991) J. Biol. Chem. 266, 20658 –20665
9. Parkin, D. W. (1996) J. Biol. Chem. 271, 21713–21719
10. Estupin˜ a´n, B., and Schramm, V. L. (1994) J. Biol. Chem. 269, 23068 –23073 11. Horenstein, B. A., Parkin, D. W., Estupin˜a´n, B., and Schramm, V. L. (1991)
Biochemistry 30, 10788 –10795
12. Horenstein, B. A., and Schramm, V. L. (1993) Biochemistry 32, 7089 –7097 13. Degano, M., Almo, S. C., Sacchettini, J. C., and Schramm, V. L. (1998)
Bio-chemistry 37, 6277– 6285
14. Shi, W., Schramm, V. L., and Almo, S. C. (1999) J. Biol. Chem. 274, 21114 –21120
15. Gopaul, D. N., Meyer, S. L., Degano, M., Sacchettini, J. C., and Schramm, V. L. (1996) Biochemistry 35, 5963–5970
16. Sinnott, M. L. (1990) Chem. Rev. 90, 1171–1202
17. Gardiner, P. R., and Wilson, A. J. (1987) Parasitol. Today 2, 255–257 18. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989)
Gene (Amst.) 77, 51–59
19. Ellis, K. J., and Morrison, J. F. (1982) Methods Enzymol. 87, 405– 426 20. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307–326 21. French, S., and Wilson, K. (1978) Acta Crystallogr. Sect. A 34, 517–525 22. Navaza, J. (1994) Acta Crystallogr. Sect. A 50, 157–163
23. Bru¨ nger, A. T., Krukowski, A., and Erickson, J. W. (1990) Acta Crystallogr.
Sect. A 46, 585–593
24. Bru¨ nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grose-Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, N., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr.
Sect. D 54, 905–921
25. Roussel, A., Fontecilla-Camps, J. C., and Cambillau, C. (1990) XV IUCr
Con-gress Abstracts, pp. 66 – 67, International Union of Crystallography,
Bor-deaux, France
26. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993)
J. Appl. Crystallogr. 26, 283–291
27. Kraulis, P. J. (1991) J. Appl. Crystallogr. 24, 946 –950
28. Lawrence, M. C., and Bourke, P. (2000) J. Appl. Crystallogr. 33, 990 –991 29. Petsko, G. A., and Ringe, D. (2000) Curr. Opin. Chem. Biol. 4, 89 –94 30. Shi, W., Tanaka, K. S. E., Crother, T. R., Taylor, M. W., Almo, S. C., and
Schramm, V. L. (2001) Biochemistry 40, 10800 –10809
31. He´roux, A., White, E. L., Ross, L. J., Kuzin, A. P., and Borhani, D. W. (2000)
Structure 8, 1309 –1318
32. Altona, C., and Sundaralingam, M. (1972) J. Am. Chem. Soc. 94, 8212 33. Olson, W. K., and Sussman, J. L. (1982) J. Am. Chem. Soc. 104, 270 –278 34. Roy, K. B., and Miles, H. T. (1983) Nucleoside Nucleotide 2, 231–242 35. Burley, S. K., and Petsko, G. A. (1988) Adv. Protein Chem. 39, 125–189 36. Ishida, T., Tarui, M., In, Y., Ogiyama, M., Doi, M., and Inoue, M. (1993) FEBS
Lett. 333, 214 –216
37. Schramm, V. L. (1998) Annu. Rev. Biochem. 67, 693–720 38. Xu, Y., and Grubmeyer, C. (1998) Biochemistry 37, 4114 – 4124
39. Liu, L., and Santi, D. V. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8604 – 8608 40. Ishida, T., Doi, M., Ueda, H., Inoue, M., and Scheldrick, G. M. (1988) J. Am.
Chem. Soc. 110, 2286 –2294
by guest, on October 27, 2012
www.jbc.org
