Purine-specific nucleoside N-ribohydrolase from Trypanosoma brucei brucei. Purification, specificity, and kinetic mechanism.

Trypanosomes have no de novo purine biosynthesis and thus depend upon salvage pathways to obtain purines for their metabolic pathways and for the biosynthesis of nucleic acids. An inosine-adenosine-guanosine preferring nucleoside hydrolase (IAG-nucleoside hydrolase) from the African trypanosome Trypanosoma brucei brucei represents approximately 0.2% of the soluble protein in this organism. The enzyme has been purified over 400-fold to >95% homogeneity from the bloodstream form of this parasite. IAG-nucleoside hydrolase is a dimer of Mr 36,000 subunits. The kcat/Km for inosine, adenosine, and guanosine are 1.9 x 10(6), 1.2 x 10(6), and 0.83 x 10(6) M -1 s-1, respectively. The kinetic mechanism with inosine as substrate is rapid equilibrium with random product release. The turnover rate for inosine at 30 degrees C is 34 s-1. Pyrimidine nucleosides are poor substrates with kcat/Km values of approximately 10(3) M -1 s-1. Deoxynucleosides are also poor substrates with kcat/Km values near 10(2) M -1 s-1. AMP is not a detectable substrate and there is no measurable purine nucleoside phosphorylase activity. 3-Deazaadenosine, 7-deazaadenosine (tubercidin), and formycin B are competitive inhibitors with Kis of 1.8, 59, and 13 microM, respectively. The Km shows a slight dependence on pH with a pH optimum around 7. The Vmax/Km data indicate there are two ionizable enzymatic groups, pKa 8.6, required for the formation of the Michaelis complex. The Vmax data indicate three ionizable groups involved in catalysis. Two essential groups exhibit pKa values of 8.8, and a third group with a pKa of 6.5 increases the Vmax an additional 10-fold. All three groups must be protonated for optimum catalytic activity.

Trypanosomiasis, caused by tsetse-transmitted hemoprotozoan parasites, continues to be a major threat to human and animal health across vast areas of sub-Sahara Africa, with over 50 million people at risk and 20,000 new cases reported annually (Vickerman et al., 1993). Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense are the causitive agents of sleeping sickness in human beings while Trypanosoma congolense, Trypanosoma vivax, and Trypanosoma brucei brucei cause the disease in domestic animals and wildlife. Trypanosomiasis is arguably the most important disease of domestic livestock in Africa, with approximately 44 million cattle at risk of infection (FAO/WHO/OIE, 1982).
The metabolism of purines differs greatly between the trypanosomes and their mammalian hosts (Hammond and Gutteridge, 1984;Marr and Docampo, 1986). Mammals obtain purines primarily by de novo synthesis and/or the purine salvage pathways. In contrast, trypanosomes have no de novo purine synthesis and must depend upon the host to provide it with either purines or purine ribosides. Enzymes involved in the purine salvage pathways include phosphoribosyl transferases, adenine, and adenosine deaminases and purine nucleoside kinases (Gutteridge and Davies, 1981;Fish et al., 1982;Davies et al., 1983).
To date, none of the purine salvage enzymes from T. brucei brucei or other African trypanosomes have been purified to homogeneity and characterized. However, the purification and extensive characterization of two nucleoside hydrolases has been reported from Crithidia fasciculata, a trypanosome of the mosquito Parkin et al., 1991;Estupiñ á n and Schramm, 1994). The nonspecific enzyme catalyzes the hydrolysis of both purine and pyrimidine nucleosides and has been named inosine-uridine (IU) nucleoside hydrolase. The more specific enzyme preferred purine nucleosides and was named guanosine-inosine (GI) nucleoside hydrolase.
The purine salvage pathways are targets for the development of a new generation of chemotherapeutic drugs, since they possess two important criteria. Foremost, the purine salvage pathway is essential for the viability of the parasite and second there are enzymes involved in the salvage pathway which differ from those of the host (Wang, 1984). With resistance developing to the currently available drugs and little prospect of new antiparasitic agents, a renewed effort to develop target specific drugs for the enzymes of the purine salvage pathway could be important in the control of parasitic diseases (Aldhous, 1994).
Reported in this paper is the purification and characterization of a nucleoside hydrolase of the purine salvage pathway from the bloodstream form of T. brucei brucei. During this study, two nucleoside hydrolases were identified: an inosineadenosine-guanosine preferring nucleoside hydrolase (IAG-nucleoside hydrolase) and an inosine-guanosine preferring nucleoside hydrolase (IG-nucleoside hydrolase). The purification scheme, substrate specificity, inhibitor studies, and a kinetic mechanism for the IAG-nucleoside hydrolase are discussed.

MATERIALS AND METHODS
Growth of Bloodstream Forms of T. brucei brucei (ILTaT 1.1)-T. brucei brucei ILTaT 1.1 is a cloned population of trypanosomes derived from an isolate collected from a naturally infected cow in Utembo, Kenya (Miller and Turner, 1981). Bloodstream forms of this clone were grown in sublethally irradiated (650 rads) Sprague-Dawley rats after inoculation with cryopreserved stabilates. The trypanosomes were then isolated from blood using DEAE-cellulose chromatography (Lanham and Godfrey, 1970). After separation from the blood elements, the parasites were concentrated in phosphate-saline glucose, pH 8.0, containing protease inhibitors (5 g/ml each of E-64, leupeptin, chymostatin, and antipain), then frozen in liquid nitrogen. p-Nitrophenyl ␤-Dribofuranoside (p-nitrophenylriboside) was a generous gift from Dr. Vern L. Schramm.
Colorimetric Assays for Nucleoside Hydrolase-Catalytic activity for the nucleoside hydrolase was determined by the formation of ribose (reducing sugar) from 2.5 mM inosine in 50 mM phosphate, pH 7.2. The assay volumes were 1 ml. The reaction was terminated by the addition of 0.3 ml of reagent containing 4% Na 2 CO 3 , 1.6% glycine, 0.045% CuSO 4 ⅐5H 2 0 and 0.3 ml of 0.12% 2,9-dimethyl-1-10-phenanthroline, in a total volume of 1.6 ml. The color was developed in boiling water for 10 min and the optical absorbance measured at 450 nm. The normal procedure is to develop the color at 95°C for 8 min. However, due to the high altitude of Nairobi (6000 ft), 10 min was used. A standard ribose curve was used to determine the extinction coefficient under the assay conditions. A modified assay procedure, which removes excess protein, was used during the initial stages of the purification procedure. Protein solutions of up to 25 l were added to 100 l of the assay mixture and incubated for the desired time. The assay was terminated by the addition of 100 l of 0.1 M ZnSO 4 . Excess Zn 2ϩ was precipitated by the addition of 100 l of 0.1 M NaOH. After centrifugation, 200 l of supernatant was removed and assayed for ribose as described above.
Spectrometric Assay for Nucleoside Hydrolase-The 1-ml reaction mixture contained substrate at the desired concentration, 50 mM phosphate buffer, pH 7.2, and the desired concentration of inhibitor. For the pH profile studies a mixed buffer system consisting of 25 mM each of citrate, phosphate, and pyrophosphate, adjusted to the desired pH with HCl or NaOH, was used.
⌬ Millimolar Extinction Coefficients (⌬⑀)-The conversion of a specified concentration of substrate to product was accomplished by the addition of sufficient IAG-nucleoside hydrolase to completely convert substrate to product. The solution was scanned, before and after incubation with enzyme, to determine ⌬⑀. All values of the starting substrates were determined at pH 7.2 unless otherwise stated. The ⌬⑀ (mM Ϫ1 cm Ϫ1 ) is: inosine, Ϫ0.92 at 280 nm; adenosine, Ϫ1.4 at 276 nm; guanosine, 0.16 at 308 nm; purine riboside, 4 at 281 nm; p-nitrophenylriboside, 12.2 at 400 nm. For the pH profile a standard curve of the ⌬⑀ of inosine at 280 nm or adenosine at 276 nm was determined as a function of pH.
Analysis of pH Data-The pH dependence of V max , V max /K m , and K m were fitted to their respective equations (Cleland, 1979), where K a values are the acid dissociation constants, H ϩ is the hydrogen ion concentration, K m(L) and K m(H) , or V max(L) and V max(H) are the limiting K m values or V max values at low (L) and high (H) pH. Determination of K m(app) and V max(app) for 2Ј or 3Ј-Deoxyadenosine-Since the deoxynucleoside substrates compete with inosine and the rate of hydrolysis was less than 0.1% of inosine, the K m(app) values were assumed to be equivalent to K i(app) for the deoxynucleosides. K i(app) for 2Ј-or 3Ј-deoxyadenosine was determined using the following equation: where v o and v i are the initial rates without and with inhibitor, respectively. S is the substrate concentration, I is the inhibitor concentration, K m is the Michaelis constant for the substrate, and K i(app) is the apparent dissociation constant for the inhibitor (deoxynucleosides). The V max -(app) was then determined by setting the concentration of the deoxynucleoside to 10 ϫ K m(app) and measuring the initial rate.
Purification of IAG-nucleoside Hydrolase-Approximately 100 g of frozen T. brucei brucei ILTaT 1.1 (bloodstream form) cells were thawed and suspended in 100 ml of 50 mM sodium phosphate, pH 7.5, containing 1 mM MgCl 2 , 0.1 mM EDTA, 0.1 mM dithiothreitol, 1 mM benzamidine, 1 mM 1,10-phenanthroline, 0.1 mM phenylmethylsulfonyl fluoride, 50 mg/liter soybean trypsin inhibitor, 1.25 mg/liter leupeptin, and 6.5 mg/liter E-64. The suspension was sonicated, using a Branson Sonifier fitted with a 1 ⁄2-inch standard tip, 4 ϫ 30 s on a continuous setting with cooling to 5°C between sonications. All the following steps were performed at 4°C. The disrupted cell suspension was then centrifuged for 2 h at 111,000 ϫ g av in a Ti-42.1 fixed angle rotor.
Solid ammonium sulfate was added to the supernatant to give a 0.55 saturation at 4°C, stirred for 30 min, and centrifuged for 10 min at 20,000 ϫ g. The resulting pellet was dissolved in a minimum volume of 50 mM Hepes, pH 7.5, then dialyzed exhaustively overnight against the same buffer.
The dialyzed fraction was applied to a Pharmacia Q-Sepharose Hi-Load 26/10 column, equilibrated with 50 mM Hepes, pH 7.5, connected to a Pharmacia FPLC system. The column was washed with 100 ml of buffer and the enzymes eluted with a 470-ml linear gradient of 0 -1 M KCl in the same buffer. IAG-nucleoside hydrolase elutes near 0.25 M KCl and the IG-nucleoside hydrolase elutes near 0.5 M KCl. The fractions which contained the IAG-nucleoside hydrolase activity were pooled and concentrated to about 5 ml using a collodion bag concentrator which retained proteins with M r greater than 12,000. The fractions containing the IG-nucleoside hydrolase activity were also pooled and concentrated, as described above, then stored at Ϫ70°C until further use.
The concentrated IAG-nucleoside hydrolase from Q-Sepharose was applied to a HiLoad 26/60 Superdex-200 (Pharmacia Biotech Inc.) equilibrated with 50 mM Hepes, pH 7.5, containing 0.25 M KCl and eluted at a flow rate of 2 ml/min using the same buffer. The resulting fractions which contained IAG-nucleoside hydrolase activity were pooled and then concentrated using a collodion bag as described above.
An Ado-agarose affinity gel was prepared from AMP-agarose (Sigma, A1271). The AMP-agarose was suspended in 50 mM Ches, 1 pH 9.2, containing 1 mM MgCl 2 and 20 units/ml alkaline phosphatase. This mixture was incubated at 30°C overnight. Approximately 12 ml of the Ado-agarose was packed into a Pharmacia XK 16/20 custom FPLC column and attached to the Pharmacia FPLC system. The column was equilibrated with 20 mM pyrophosphate, pH 8.5, and the IAG-nucleoside hydrolase applied to the column at a flow rate of 0.5 ml/min. The IAG-nucleoside hydrolase activity was retarded slightly and eluted after the contaminating proteins. The fractions containing the enzymatic activity were pooled and concentrated to about 1 ml using a collodion bag. Purity of the protein was estimated by electrophoresis in polyacrylamide gels containing SDS. The enzyme was then flash-frozen in dry ice/ethanol and stored at Ϫ70°C.
Molecular Weight of IAG-Nucleoside Hydrolase-The elution position of active IAG-nucleoside hydrolase relative to proteins of known molecular weight was determined by HiLoad 26/60 Superdex-200 gel chromatography using the coelution positions of chymotrypsinogen A, ovalbumin, bovine serum albumin, aldolase, catalase, ferritin, and bovine thyroglobulin. blue dextran 2000 was used to determine void volume (V o ). The column was eluted with 50 mM Hepes, pH 7.5, containing 0.25 M KCl at a flow rate of 2 ml/min. The IAG-nucleoside hydrolase was detected by activity and the other proteins detected by A 280 . The M r of the IAG-nucleoside hydrolase was determined from the elution profile of the Superdex 200 chromatography purification step. The ratio of elution position of the proteins (V e ) relative to void volume (V o ) was determined graphically, and the molecular weight for IAGnucleoside hydrolase was estimated relative to the molecular weight standards. The subunit molecular weight was estimated on denaturing SDS-polyacrylamide gel electrophoresis and mass spectroscopy. The purified IAG-nucleoside hydrolase was subjected to electrospray ionization analysis in a Sciex API-III mass spectrometer. Samples of 1-5 g/l were introduced through a 1 ϫ 50-mm C-8 HPLC column equilibrated with 20% acetonitrile in H 2 O containing 0.1% trifluoroacetic acid.

RESULTS
Purine Nucleoside Phosphorylase Activity-The cell-free extract was dialyzed against 50 mM Hepes, pH 7.5, and assayed for the presence of a purine nucleoside phosphorylase activity. The activity was determined using the spectrophotometric assay in either 50 mM Pipes, pH 7, or 50 mM phosphate, pH 7, containing one of the following substrates; 0.5 mM inosine, 0.2 mM adenosine, or 0.2 mM guanosine. The activity in the absence or presence of inorganic phosphate was, respectively, as follows; 5.2 Ϯ 0.2 and 4.9 Ϯ 0.1 units/ml using inosine as substrate, 2.0 Ϯ 0.1 and 1.9 Ϯ 0.1 units/ml using adenosine, and 4.7 Ϯ 0.2 and 3.8 Ϯ 0.1 units/ml using guanosine. These data show that the purine nucleoside phosphorylase activity in cellfree extracts of T. brucei brucei is negligible relative to the nucleoside hydrolase activity.
Purification of IAG-Nucleoside Hydrolase-Two distinct nucleoside hydrolase activities coprecipitate in the 0.55 saturated ammonium sulfate fractionation. These two activities were subsequently determined to be the inosine-adenosineguanosine preferring nucleoside hydrolase (IAG-nucleoside hydrolase) and the inosine-guanosine preferring nucleoside hydrolase (IG-nucleoside hydrolase). The Q-Sepharose column step gave a 28-fold increase in specific activity of the IAGnucleoside hydrolase and completely resolved the IAG-nucleoside hydrolase activity from the IG-nucleoside hydrolase. Superdex 200 size exclusion chromatography gave a 2-fold purification. The final Ado-agarose column provided a protein of Ͼ95% purity with a specific activity of 50 units/mg protein (Fig. 1). The enzyme did not bind tightly to the Ado-agarose column. However, the enzyme activity was retarded sufficiently enough to separate it from contaminating proteins. In total, the IAG-nucleoside hydrolase was purified over 400-fold from cell extracts with a 16% yield as summarized in Table I. IAG-nucleoside hydrolase represents about 0.2% of the total soluble protein of bloodstream form of T. brucei brucei.
Molecular Weight and Subunit Structure-The molecular weight of IAG-nucleoside hydrolase was estimated using Hi-Load 26/60 Superdex-200 gel chromatography, by denaturing SDS-gel electrophoresis using proteins of known molecular weights as standards, and by electrospray ionization mass spectroscopy. IAG-nucleoside hydrolase elutes with a V e /V o of 1.85 which corresponds to a M r of 71,000 (Fig. 2). Denaturing polyacrylamide gel electrophoresis gave a single band with an apparent subunit molecular weight of 41,000 (Fig. 2). Electron spray mass spectrometry of IAG-nucleoside hydrolase gave a molecular mass of 35,759 Ϯ 4. These values are consistent with the IAG-nucleoside hydrolase being composed of two subunits of identical, or similar, molecular weight.
Substrate Specificity of IAG-Nucleoside Hydrolase-The commonly occurring purine nucleosides are all substrates of IAG-nucleoside hydrolase with k cat /K m values from 0.83 to 1.9 ϫ 10 6 M Ϫ1 s Ϫ1 (Table II). Inosine is the best naturally occurring substrate for IAG-nucleoside hydrolase with a V max of 50 mol min Ϫ1 mg Ϫ1 , resulting in a turnover number of 34 s Ϫ1 with a K m of 18 M to give a k cat /K m , catalytic efficiency, of 1.9 ϫ 10 6 M Ϫ1 s Ϫ1 . The K m values for the purines and pyrimi-  Nucleoside Hydrolase from T. brucei brucei dines are similar, ranging from 15 M for adenosine to 106 M for cytidine. The pyrimidines, however, are poorer substrates with V max of 0.02 and 0.46 mol min Ϫ1 mg Ϫ1 for uridine and cytidine, respectively. The k cat /K m values for uridine and cytidine are 0.13 ϫ 10 3 and 3.0 ϫ 10 3 M Ϫ1 s Ϫ1 , respectively. Based on k cat /K m values, a 1000-fold decrease in catalytic efficiency exists between the purines and pyrimidines. Deoxyadenosines in the 2Ј or 3Ј position are poor substrates. When compared to adenosine, 2Ј-deoxyadenosine has a 1000-fold decreased k cat(app) , 3Ј-deoxyadenosine has a 40-fold increased K m(app) and a 1000-fold decreased k cat(app) , and 5Ј-deoxyadenosine has a 600-fold increased K m and a 25-fold decreased k cat . Thus the deoxyadenosines have a decreased catalytic efficiency of 4 to 5 orders of magnitude compared to adenosine or other purine nucleosides.
Additional nucleoside analogs were tested as substrates. Purine riboside is almost as efficient a substrate as the other purine nucleosides, with an affinity the same as inosine but a slower turnover number resulting in a slightly less k cat /K m value of 0.46 ϫ 10 6 M Ϫ1 s Ϫ1 . 5-Methyluridine has an affinity about the same as the purines. However, catalysis is significantly slower, lowering the catalytic efficiency to the same level as the other pyrimidines. When compared to inosine the synthetic substrate p-nitrophenylriboside has a 30-fold increased K m and a 20-fold decreased k cat resulting in a decrease in catalytic efficiency of 2 orders of magnitude. AMP is not a detectable substrate.
Inhibitors of IAG-Nucleoside Hydrolase-Both ribose and hypoxanthine exhibit slope-linear competitive patterns using pnitrophenylriboside as substrate (Table II). The K is is 1.35 and 19 mM for hypoxanthine and ribose, respectively. In the presence of 150 mM ribose and 612 M p-nitrophenylriboside, the K is for hypoxanthine increased to 5.6 mM. 3-Deazaadenosine, formycin B, and 7-deazaadenosine (tubercidin) are competitive inhibitors of IAG-nucleoside hydrolase with a K is of 1.76, 13, and 59 M, respectively.
Effect of pH on the Kinetic Constants of IAG-Nucleoside Hydrolase-IAG-nucleoside hydrolase is stable from approximately pH 5 to 10 during the assay. The V max decreases between pH 5.2 to 9.7 with a plateau near pH 7.5 using inosine as the substrate (Fig. 3). The data are consistent with a requirement for two protonated groups essential for catalysis near pH 8.8 and a third protonated group, pK a near 6.5, which increases the catalytic rate 10-fold. There is no effect of pH on V max /K m between pH 5.2 and 8.2. However, the V max /K m decreases as the pH increases from pH 8.2 to 9.4, indicating two groups on the free enzyme involved in the formation of the catalytically competent Michaelis complex, pK a 8.6. The K m decreases slightly from 150 to 25 M between pH 5 and 7, with the K m remaining unchanged above pH 7. The effects of pH on V max with adenosine as substrate also indicate two protonated groups essential for catalysis at pH 8.8, and one protonated group which increases the catalytic rate at pH 6.7.

DISCUSSION
Kinetic Mechanism-The presence of two slope-linear competitive-product inhibition patterns establishes that the kinetic mechanism for IAG-nucleoside hydrolase, of the bloodstream form of T. brucei brucei, is rapid equilibrium random. Both the inhibition constants and the K m for inosine are dissociation constants in a rapid equilibrium system. Hypoxanthine binds over 14 times tighter than ribose, establishing the preferred release of ribose. Both ribose and hypoxanthine bind tighter to the free enzyme than the enzyme substrate ternary complex (Fig. 4). The rapid equilibrium condition holds only when substrate and product release are significantly faster than the catalytic step. The k cat/ K m value of 1.9 ϫ 10 6 M Ϫ1 s Ϫ1 for inosine is approximately 2-3 orders of magnitude less efficient than a diffusion limited enzyme (Fersht, 1985). These data support a rate-limiting catalytic step at pH 7.2. The k cat in the reverse direction was calculated using the values for K eq , k cat in the forward direction, and the kinetic constants for the enzyme substrate complexes utilizing the Haldane relationship. The Haldane equation is given below.
where K hx,r is the dissociation constant for the ribose from the enzyme-hypoxanthine-ribose complex. The value for the k cat,forward is 34 s Ϫ1 compared with 2 s Ϫ1 for k cat,reverse (Fig. 4). The value used for the K eq was 106 . Substrate and Inhibitor Specificity of IAG-Nucleoside Hydrolase-The relatively tight binding of inosine depends on the relationship of both the ribose and hypoxanthine moieties, since both free ribose and hypoxanthine bind weakly to the enzyme. The product of the K i values for hypoxanthine and ribose is 25 M, close to the observed K m for inosine of 18 M. This relationship suggests that the enzymatic substrate recognition is the sum of the interactions of the groups on inosine. 13 Ϯ 1 7-Deazaadenosine (Tubercidin) b 59 Ϯ 5 a The K m or K i(app) value was calculated from the ratio of Ϫinhibitor/ϩinhibitor initial rates as described under "Results." At least three concentrations of inhibitor was used to determine K i(app) b The inhibition constant K is was obtained from experiments using at least five substrate concentrations at each of three inhibitor concentrations. The inhibition constant was obtained by fitting the data to the equation for competitive inhibition using the computer programs of Cleland (1979).
The binding specificity of the ribose moiety involves the 3Ј-and 5Ј-hydroxyl groups of the ribose ring. The comparison of V max and K m values for adenosine to those from 2Ј-, 3Ј-, or 5Јdeoxyadenosine demonstrates that the ribose hydroxyl groups are important for binding and/or catalysis. The 2Ј-OH group of adenosine has a greater effect on V max (1000-fold) than on K m (2-fold), indicating that it is important in the stabilization of the enzymatic transition state rather than the formation of the Michaelis complex. In contrast, the 3Ј-OH group is important both in the formation of the Michaelis complex (40-fold increased K m ) and also the stabilization of the transition state, as indicated by the 1000-fold decrease in V max . The 5Ј-OH contributes more to the formation of the Michaelis complex than catalysis since the K m increases 600-fold, whereas V max decreases only 25-fold. The binding specificity of the purine or pyrimidine-leaving group is much less stringent. The K m values differ only approximately 5-fold between the purines and pyrimidine, ranging from 15 to 106 M. Other purine and pyrimidine analogs also bind with K m values in the 20 M range. Purine nucleoside, which is devoid of amino and/or hy-droxyl groups attached to the heterocyclic ring, binds as tightly to IAG-nucleoside hydrolase as the other purines. Only p-nitrophenylriboside, which lacks the heterocyclic ring structure, has a significantly higher K m .
The naturally occurring purine nucleosides are substrates of IAG-nucleoside hydrolase with k cat /K m values around 10 6 M Ϫ1 s Ϫ1 . The pyrimidines are much poorer substrates with k cat /K m values around 10 3 M Ϫ1 s Ϫ1 . The activation and binding energy difference (⌬⌬G ab ) between two substrates can be compared using the following equation, where R is the gas constant (1.99 cal/T-mol) and T is the absolute temperature, 303 K. ⌬⌬G ab ranges from 3 to 6 kcal/ mol for k cat /K m between purines and pyrimidines. Substituting K m values for k cat /K m indicates only about 1 kcal/mol binding energy difference between purines and pyrimidines. Therefore, approximately 3 to 5 kcal/mol of binding energy, about the energy of a hydrogen bond, is used to lower the energy of activation of the purines, compared to the pyrimidines (Fersht, 1985). These data support the hypothesis that protonation of the leaving group is an essential component in catalysis of purines and that IAG-nucleoside hydrolase is able to protonate the purine-leaving group more efficiently than the pyrimidineleaving group.
Ionizable Groups in Catalysis-The pH profile for hydrolysis of inosine indicates that three protonated groups contribute to the full catalytic capacity of IAG-nucleoside hydrolase. The presence of a plateau region around pH 7.5 indicates that two forms of the enzyme are catalytically active. One group with a pK a near 6.5 and two groups with pK a values near 8.8 are required for the most efficient catalysis. Loss of a proton from the most acidic group lowers the V max approximately 10-fold. However, the loss of the two protons on the amino acids, pK a 8.8, completely inactivates IAG-nucleoside hydrolase. The V max /K m pH profile shows no change in catalytic efficiency between pH 5 and 8. However, above pH 8 the catalytic efficiency decreases significantly due to two ionizable groups on the free enzyme, pK a 8.6. The observed K m is lowered from 150 M at pH 5.2 to a constant value of 25 M above pH 7. There are two possible explanations for this observation: 1) there is one group on the enzyme with a pK a of 6.0 which aids in the formation of the Michaelis complex, since there are no titratable groups on the substrate over this pH range; or 2) catalysis is no longer rate-limiting, thus changing K m from a thermodynamic binding constant (K s ) to a kinetic constant. However, between pH 7 and 10 these data are consistent with inosine binding in rapid equilibrium with the enzyme since there is no FIG. 3. Effect of pH on the kinetic constants for IAG-nucleoside hydrolase. The upper panel demonstrates, using inosine as substrate, the decrease in V max , with a plateau, as the pH increases, V max /K m remaining constant until pH 8 then decreasing with increasing pH, and the K m changing from 150 to a plateau of 25 M at pH 7. The data were fitted to the appropriate equation (see "Materials and Methods") using the Mardquat algorithm. The ordinate scale for V max is micromoles min Ϫ1 mg Ϫ1 (S.A.), V max /K m is S.A./mM, and for K m is mM. The experimental points are from the best fits of the data to the Michaelis-Menten equation with a minimum of 5 different substrate concentrations. The lower panel shows that the effect of pH on the V max profile with adenosine as substrate is experimentally indistinguishable to the inosine profile.
FIG. 4. Kinetic mechanism for IAG-nucleoside hydrolase. The numbers near the arrows for binding and release of substrates and products are dissociation constants for each complex. The turnover number, k cat , in the forward direction was determined from V max , and the k cat in the reverse direction was determined using K eq and the Haldane relationship. E represents free enzyme. The dissociation constant for ribose interacting with the E⅐hypoxanthine complex was estimated from the three experimentally determined dissociation constants and the expression for the thermodynamic box of product release. change in K m over this pH range and the kinetic mechanism of the enzyme is rapid equilibrium with inosine as substrate at pH 7.2 (Tipton and Dixon, 1979).
Protonation of the Leaving Group-Since proton donation to the leaving group has been implicated in both chemical and enzymatic depurinations the most likely positions are N-1, N-7, N-3, and the oxo or amino group on C-6 of the purine (Garrett and Mehta, 1972;Mentch et al., 1987). The effects of pH on the V max profile with adenosine as substrate are experimentally indistinguishable from those with inosine (Fig. 3). The V max for adenosine and purine riboside decreased to 50 and 2% of inosine, respectively, while the K m for these substrates are experimentally the same. These data indicate that the proton donation to the leaving group, which is essential for catalysis, is not likely to involve the purine's N-1 or the oxo or amino group of C-6 since the electronic configuration of these functional groups are very distinct . Both 3-deazaadenosine and 7-deazaadenosine (tubercidin) bind well to the enzyme, but are not substrates, demonstrating the essential nature of N-3 and N-7 in catalysis. The observation that formycin B binds tighter than tubercidin is consistent with the hypothesis that N-7 of the purine ring acts as a proton acceptor in the transition state.
Comparison of Nucleoside Hydrolases from T. brucei brucei and C. fasciculata-The kinetic constants and molecular weights for the three well characterized nucleoside hydrolases are seen in Table III. The IAG-nucleoside hydrolase is most similar in substrate and inhibitor specificity to the GI-nucleoside hydrolase. However, the differences in molecular weight and the ability of the IAG-nucleoside hydrolase to utilize adenosine as a substrate clearly distinguishes the enzyme. The IAGand GI-nucleoside hydrolases differ in subunit aggregation, stability to heat, inhibitor specificity, and in the dependence of the kinetic constants on pH when compared to the IU-nucleoside hydrolase Estupiñ á n and Schramm, 1994). Both IAG-and GI-nucleoside hydrolase are inhibited by formycin B and tubercidin in the M range, whereas these modulators inhibit the IU-nucleoside enzyme in the millimolar range. Finally, the purine-specific IAG-and GI-nucleoside hydrolases have k cat /K m values approximately 2 orders of magnitude greater than the IU-nucleoside hydrolase and are thus superior biological catalysts.

Chemical and Kinetic Mechanism of the Three Nucleoside
Hydrolases-The kinetic mechanism is rapid equilibrium random for both IAG-nucleoside hydrolase and the IU-nucleoside hydrolase. Ribose is released first from the IAG-nucleoside hydrolase with hypoxanthine being released first from the IUnucleoside hydrolase. The chemical mechanism of the IAG-, GI-, and IU-nucleoside hydrolases involves the formation of an oxycarbonium ion in the transition state, characteristic of an S n 1-like mechanism, as determine by kinetic isotope effects . 2 However, the magnitude of the oxycarbonium ion formation and/or the reaction pathway by which the transition state is reached differs for each of the nucleoside hydrolases. The synthetic substrate p-nitrophenylriboside has been used to determine the reaction pathway by which Nribohydrolases reach their oxycarbonium-ion transition state. N-Ribohydrolases which utilize protonation of the purine base to achieve the oxycarbonium-ion transition state have a high specificity for the aglycone and poor reaction rates with pnitrophenylriboside, whereas N-ribohydrolases which utilize activation of the ribose ring have much lower aglycone specificity and high reaction rates with p-nitrophenylriboside (Mazzella et al., 1996).
The IU-nucleoside hydrolase has a much broader aglycone specificity than the IAG-or GI-nucleoside hydrolase, as seen in Table III. Whereas the IU-nucleoside hydrolase possesses the ability to cleave both purines and pyrimidines, the IAG-and GI-nucleoside hydrolases have a very strong preference for purines. The role of the 5Ј-methoxy group in catalysis differs between the hydrolases since 5Ј-deoxyadenosine is a substrate 2 D. W. Parkin and V. L. Schramm, unpublished results.

TABLE III
Comparison of kinetic constants, native molecular weight, and subunits for IAG-nucleoside hydrolase from T. brucei brucei and the IU and GI-nucleoside hydrolase from C. fasciculata for the IAG-nucleoside hydrolase but not a substrate for the IUor GI-nucleoside hydrolases Estupiñ á n and Schramm, 1994). The pH profiles of the kinetic constants also establish fundamentally different catalytic mechanisms between the IAGand GI-nucleoside hydrolase and the IU-nucleoside hydrolase. The pH profile studies show that protonation of the leaving group is essential in catalysis for the IAG-and GI-nucleoside hydrolases. The IAG-and GI-nucleoside hydrolases demonstrate an absolute requirement for two groups with pK a values around 9 to be protonated for catalytic activity. Protonation of a third group, pK a 6.5, on the IAG-nucleoside hydrolase increases the rate of catalysis but is not essential for activity. A similar pH profile pattern for V max was obtained for the GInucleoside hydrolase (Estupiñ á n and ). In contrast, the IU-nucleoside hydrolase requires one group to be unprotonated, pK a 7.1, and only one group to be protonated, pK a 9.1.
The above data support the hypothesis that the chemical mechanism of the IAG-and GI-nucleoside hydrolases is different from the IU-nucleoside hydrolase. These observations are consistent with the hypothesis that the oxycarbonium-ion transition state for the IAG-and GI-nucleoside hydrolases is reached via protonation of the purine ring, whereas the oxycarbonium-ion transition state for IU-nucleoside hydrolase is reached via the enzyme activating the ribosyl moiety.
Conclusion-The IAG-nucleoside hydrolase of T. brucei brucei is the first purine salvage enzyme to be purified and characterized from an African trypanosome. The enzyme exhibits high catalytic efficiency, specificity for all purines, and is an abundant protein in the bloodstream form of T. brucei brucei, giving it properties consistent with an active role in the purine salvage pathway under physiological conditions. The relatively prominent role of the purine-specific enzyme has been established in C. fasciculata with over 80% of the inosine in C. fasciculata being metabolized via GI-nucleoside hydrolase (Estupiñ á n and . Kinetic isotope effects and subsequent transition state analysis of IU-nucleoside hydrolase has resulted in the design of several new transition state inhibitors Boutellier et al., 1994). The determination of the K is of these transition state inhibitors for both IU-nucleoside hydrolase and IAG-nucleoside hydrolase, and kinetic isotope effect studies for IAG-nucleoside hydrolase, are presently being conducted. The specificity, distribution, and mechanism of this family of unique salvage enzymes may eventually permit the targeting of the purine salvage pathways for novel antibiotic design. Using the above data and the subsequent transition state analysis of the kinetic isotope effects, novel transition state inhibitors could be designed for the IAGnucleoside hydrolase as potential antiparasitic agents.