Simultaneous purification and comparative characterization of six serine hydrolases from rat liver microsomes

Rat liver microsomes contain many serine hydrolases, which can be demonstrated in electropherograms with carboxylesterase stain and with an active-site-directed radioactive organophosphate. Five of the most prominent of these enzymes plus dipeptidyl aminopeptidase IV, a microsomal serine hydrolase without activity against simple esters, have been highly purified with a simultaneous procedure after solubilization with saponin. The five carboxylesterases belong to at least three groups of chemically different proteins. Terminal amino acids, amino acid composition, and substrate specificity are different, while the subunit molecular weight of all esterases is very similar (about 60,000). All purified carboxylesterases have monooleylglycerol-cleaving capacity. The subunit weight (84,000) and the N-terminal amino acid (serine) of the peptidase differ from those of all isolated carboxylesterases. The data are correlated to other reports on individual serine hydrolases from rat liver.     

Effects of clofibric acid and tiadenol on cytosolic long-chain acyl-CoA hydrolase and peroxisomal beta-oxidation in liver and extrahepatic tissues of rats

The effects of two peroxisome proliferators, p-chlorophenoxyisobutyric acid (clofibric acid) and 2,2'-(decamethylenedithio)diethanol (tiadenol), on cytosolic long-chain acyl-CoA hydrolase and peroxisomal beta-oxidation were studied in several organs of rat. Among organs of control rats, the brain had the highest activity of long-chain acyl-CoA hydrolase, followed by testis, and a low activity was found in other tissues. Administration of the peroxisome proliferators caused a marked increase in activity of long-chain acyl-CoA hydrolase in both liver and intestinal mucosa and a slight increase in the activity in kidney, but little affected acyl-CoA hydrolase activity in either brain, testis, heart, spleen and skeletal muscle. In accordance with the change in the activity of acyl-CoA hydrolase, the activity of peroxisomal beta-oxidation was markedly increased in liver, intestinal mucosa and kidney, and a slight increase was found in brain and testis, whereas peroxisome proliferators little affected the activity in other organs tested. Gel filtration of cytosol from intestinal mucosa showed that clofibric acid caused an appearance of a new peak in intestinal mucosa. Although cytosol of liver, intestinal mucosa, brain and testis contained two 4-nitrophenyl acetate esterases with different molecular weights (about 105,000 and about 55,000), these esterases are different from cytosolic long-chain acyl-CoA hydrolases of these four organs in respect of molecular weight. The administration of clofibric acid little affected cytosolic 4-nitrophenyl acetate esterases. Comparative studies on cytosolic long-chain acyl-CoA hydrolases from these four organs showed that liver hydrolase I (molecular weight of about 80,000) had properties similar to those of brain and testis enzymes. On the other hand, intestinal mucosa enzyme was different from either hepatic hydrolase I or II (molecular weight of about 40,000). The results from the present study suggest that inductions of peroxisomal beta-oxidation and cytosolic long-chain acyl-CoA hydrolases are essential responses of rats to peroxisome proliferators not only in liver but also in intestinal mucosa and that induced hydrolases are not attributable to non-specific esterases.     

Long-chain acyl-CoA hydrolase in the brain

Summary. Long-chain acyl-CoA hydrolases are a group of enzymes that cleave acyl-CoAs into fatty acids and coenzyme A (CoA-SH). Because acyl-CoAs participate in numerous reactions encompassing lipid synthesis, energy metabolism and regulation, modulating intracellular levels of acyl-CoAs would affect cellular functions. Therefore, acyl-CoA synthetases have been intensively studied. In contrast, acyl-CoA hydrolases have been less investigated, especially in the brain despite the fact that its long-chain acyl-CoA hydrolyzing activity is much higher than that in any other organ in the body. However, recent studies have dissected the multiplicity of this class of enzymes on a genomic basis, and have allowed us to discuss their function. Here, we describe a cytosolic long-chain acyl-CoA hydrolase (referred to as BACH) that is constitutively expressed in the brain, comparing it with other acyl-CoA hydrolases found in peripheral organs that have a role in fatty acid oxidation.     

Triglyceride, diglyceride, monoglyceride, and cholesterol ester hydrolases in chicken adipose tissue activated by adenosine 3':5'-Monophosphate-dependent protein kinase. Chromatographic resolution and immunochemical differentiation from lipoprotein lipase.

Hormone-sensitive lipase and cholesterol ester hydrolase of chicken adipose tissue were markedly activated by adenosine 3':5'-monophosphate (cAMP)-dependent protein kinase (on the average, 235 to 275%; occasionally as much as 1000%). Diglyceride and monoglyceride hydrolases were also activated, but to a lesser extent (60 to 87%). The activation of all four hydrolases was inhibited by protein kinase inhibitor and reversed by the addition of exogenous protein kinase. Following activation by cAMP-dependent protein kinase, all four hydrolases were deactivated in a Mg2+-dependent reaction and then reactivated to or near initial levels on incubation with cAMP and Mg2+-ATP. The reversible deactivation is assumed to reflect activity of one or more protein phosphatases. The maximum activation obtainable for the four hydrolases decreased when the tissue had been previously exposed to glucagon, indicating that the glucagon-induced activation was probably similar to or identical with the activation demonstrated in cell-free preparations. The pH optima for the four hydrolase activities were similar (7.13 to 7.38). Although the absolute activities and relative degrees of kinase activation differed according to the particular emulsified substrates used, the results do not rule out the possibility that all four hydrolase activities are referable to a single hormone-sensitive hydrolase. Hormone-sensitive acyl hydrolases were separated from lipoprotein lipase by heparin-Sepharose affinity chromatography. Lipoprotein lipase was active against triolein, diolein, and monoolein, but not cholesterol oleate. Incubation of lipoprotein lipase with exogenous protein kinase, cAMP, and Mg2+ATP had no effect on any of the three hydrolase activities. Lipoprotein lipase was further purified to homogeneity and used to prepare antiserum in rabbits. The immunoglobin G fraction from these antisera completely inhibited lipoprotein lipase eluted from heparin-Sepharose columns. However, the hormone-sensitive hydrolase activities (not retained on heparin-Sepharose affinity chromatography) were not inhibited by anti-lipoprotein lipase immunoglobin G, and anti-lopoprotein lipase immunoglobin G did not affect the activation process in crude fractions. Thus, hormone-sensitive lipase and lipoprotein lipase, functionally distinct enzymes, have been physically resolved and immunochemically distinguished. Apparently lipoprotein lipase activity is not regulated, at least directly, by cAMP-dependent protein kinase.     

Purification and characterization of three parathion hydrolases from gram-negative bacterial strains

Three unique parathion hydrolases were purified from gram-negative bacterial isolates and characterized. All three purified enzymes had roughly comparable affinities for ethyl parathion and had broad temperature optima at ca. 40 degrees C. The membrane-bound hydrolase of Flavobacterium sp. strain ATCC 27551 was composed of a single subunit of approximately 35,000 daltons (Da) and was inhibited by sulfhydryl reagents such as dithiothreitol (DTT) and by metal salts such as CuCl2. The cytosolic hydrolase of strain B-1 was composed of a single subunit of approximately 43,000 Da and was stimulated by DTT and inhibited by CuCl2. The membrane-bound hydrolase of strain SC was composed of four identical subunits of 67,000 Da and was inhibited by DTT and stimulated by CuCl2. The substrate ranges of the three enzymes also differed, as evidenced by their relative affinities for parathion and the related organophosphate insecticide O-ethyl-O-4-nitrophenyl phenylphosphonothioate (EPN). The B-1 hydrolase displayed equal affinity for both compounds, the Flavobacterium enzyme showed twofold-lower affinity for EPN than for parathion, and the SC hydrolase displayed no activity toward EPN. The range in characteristics of these three enzymes can be exploited in different waste disposal strategies.     

Purification and characterization of three parathion hydrolases from gram-negative bacterial strains.

Three unique parathion hydrolases were purified from gram-negative bacterial isolates and characterized. All three purified enzymes had roughly comparable affinities for ethyl parathion and had broad temperature optima at ca. 40 degrees C. The membrane-bound hydrolase of Flavobacterium sp. strain ATCC 27551 was composed of a single subunit of approximately 35,000 daltons (Da) and was inhibited by sulfhydryl reagents such as dithiothreitol (DTT) and by metal salts such as CuCl2. The cytosolic hydrolase of strain B-1 was composed of a single subunit of approximately 43,000 Da and was stimulated by DTT and inhibited by CuCl2. The membrane-bound hydrolase of strain SC was composed of four identical subunits of 67,000 Da and was inhibited by DTT and stimulated by CuCl2. The substrate ranges of the three enzymes also differed, as evidenced by their relative affinities for parathion and the related organophosphate insecticide O-ethyl-O-4-nitrophenyl phenylphosphonothioate (EPN). The B-1 hydrolase displayed equal affinity for both compounds, the Flavobacterium enzyme showed twofold-lower affinity for EPN than for parathion, and the SC hydrolase displayed no activity toward EPN. The range in characteristics of these three enzymes can be exploited in different waste disposal strategies.     

Medium-chain fatty acid synthesis by goat mammary-gland fatty acid synthetase. The effect of limited proteolysis.

Fatty acid synthetase from goat mammary gland was subjected to limited proteolysis by trypsin and elastase. Both proteolytic enzymes selectively cleaved the chain-terminating thioester hydrolase component from the enzyme complex, leaving all other partial activities intact in the core peptides. Trypsin, but not elastase, caused extensive degradation of the released thioester hydrolase. The released thioester hydrolase could be purified to homogeneity by gel filtration. The molecular weight was estimated as 29 000 and the enzyme showed only significant hydrolytic activity toward long-chain acyl-CoA esters. The core peptides retained the ability to synthesize medium-chain acyl-CoA esters in the presence of 2,6-di-O-methyl-alpha-cyclodextrin. The results conclusively show that the terminating thioester hydrolase of goat mammary-gland fatty acid synthetase is not involved in termination of medium-chain-length fatty acid synthesis by this enzyme.     

Human liver cytosolic epoxide hydrolases

Human liver epoxide hydrolases were characterized by several criteria and a cytosolic cis-stilbene oxide hydrolase (cEHCSO) was purified to apparent homogeneity. Styrene oxide and five phenylmethyloxiranes were tested as substrates for human liver epoxide hydrolases. With microsomes activity was highest with trans-2-methylstyrene oxide, followed by styrene 7,8-oxide, cis-2-methylstyrene oxide, cis-1,2-dimethylstyrene oxide, trans-1,2-dimethylstyrene oxide and 2,2-dimethylstyrene oxide. With cytosol the same order was obtained for the first three substrates, whereas activity with 2,2-dimethylstyrene oxide was higher than with cis-1,2-dimethylstyrene oxide and no hydrolysis occurred with trans-1,2-dimethylstyrene oxide. Generally, activities were lower with cytosol than with microsomes. The isoelectric point for both microsomal styrene 7,8-oxide and cis-stilbene oxide hydrolyzing activity was 7.0, whereas cEHCSO had an isoelectric point of 9.2 and cytosolic trans-stilbene oxide hydrolase (cEHTSO) of 5.7. The cytosolic epoxide hydrolases could be separated by anion-exchange chromatography and gel filtration. The latter technique revealed a higher molecular mass for cEHCSO than for cEHTSO. Both cytosolic epoxide hydrolases showed higher activities at pH 7.4 than at pH 9.0, whereas the opposite was true for microsomal epoxide hydrolase. The effects of ethanol, methanol, tetrahydrofuran, acetonitrile, acetone and dimethylsulfoxide on microsomal epoxide hydrolase depended on the substrate tested, whereas both cytosolic enzymes were not at all, or only slightly, affected by these solvents. Effects of different enzyme modulators on microsomal epoxide hydrolase also depended on the substrates used. Trichloropropene oxide and styrene 7,8-oxide strongly inhibited cEHCSO whereas cEHTSO was moderately affected by these compounds. Immunochemical investigations revealed a close relationship between cEHCSO and rat liver microsomal, but not cytosolic, epoxide hydrolase. Interestingly, cEHTSO has no immunological relationship to rat microsomal, nor to rat cytosolic epoxide hydrolase. cEHTSO from human liver differed also from its counterpart in the rat in that it was only moderately affected by tetrahydrofuran, acetonitrile and trichloropropene oxide. Five steps were necessary to purify cEHCSO. The enzyme has a molecular mass (49 kDa) identical to that of rat liver microsomal epoxide hydrolase.     

Comparison of some physicochemical and kinetic properties of S-adenosylhomocysteine hydrolase from bovine liver, bovine adrenal cortex and mouse liver

S-Adenosyl-L-homocysteine hydrolase (EC 3.3.1.1) was purified to apparent homogeneity from bovine liver, bovine adrenal cortex and mouse liver. All enzymes were tetramers, composed of two types of subunit present in the proportion 1:1, as judged by SDS-polyacrylamide gel electrophoresis. The partition coefficient was exactly the same for these enzymes on high-performance gel permeation chromatography, and they co-sedimented in density gradients, suggesting the same molecular size and form of S-adenosylhomocysteine hydrolase from these sources. The bovine enzymes differed from the mouse liver enzyme with respect to isoelectric point (pI = 5.35, versus pI = 5.7), affinity for DEAE-cellulose, and migration of subunits on SDS-polyacrylamide gel electrophoresis with SDS from some commercial sources. The enzymes were not substrates for cAMP-dependent protein kinase. The apparent Km values for adenosine (0.2 microM) and S-adenosylhomocysteine (0.75 microM) were the same for all three enzymes. The ratio between Vmax for the synthesis and hydrolysis of S-adenosylhomocysteine was about 4 for the mouse liver enzyme, and about 6 for the bovine enzymes. It is concluded that only subtle kinetic and physicochemical differences exist between S-adenosylhomocysteine hydrolase from these bovine and mouse tissues. This suggests that differences in experimental procedures rather than species- and organ-differences of S-adenosylhomocysteine hydrolase are responsible for the variability in kinetic and physicochemical parameters reported for the mammalian hydrolase.     

Hormone-sensitive lipase of adipose tissue.

Some physiologic aspects of the mobilization and fate of free fatty acids are reviewed. The molecular mechanism of the activation of hormone-sensitive lipase in adipose tissue is then discussed. Recent evidence established that hormone-sensitive lipase, concerned with fat mobilization, is both functionally and immunochemically distinct from lipoprotein lipase, concerned with uptake of plasma triglycerides. Lipoprotein lipase activity is not altered by cyclic AMP-dependent protein kinase. The latter enzyme enhances not only triglyceride hydrolase but also monoglyceride, diglyceride and cholesterol ester hydrolase activities in chicken adipose tissue. Finally, it is shown that the activation of all four acyl hydrolases is reversible, the deactivation being magnesium-dependent. Protein phosphatase fractions from heart and liver active against phosphorylase a can reversibly deactivate adipose tissue hormone-sensitive lipase, implying a low degree of substrate specificity for lipase phosphatase.     

The Nudix hydrolases of Deinococcus radiodurans

All 21 of the Nudix hydrolase genes from the radiation-resistant organism Deinococcus radiodurans have been cloned into vectors under the control of T7 promoters and expressed as soluble proteins in Escherichia coli. Their sizes range from 9.8 kDa (91 amino acids) to 59 kDa (548 amino acids). Two novel proteins were identified, each with two Nudix boxes in its primary structure, unique among all other known Nudix hydrolases. Extracts of each of the expressed proteins were assayed by a generalized procedure that measures the hydrolysis of nucleoside diphosphate derivatives, and several enzymatic activities were tentatively identified. In addition to representatives of known Nudix hydrolase subfamilies active on ADP-ribose, NADH, dinucleoside polyphosphates or (deoxy)nucleoside triphosphates, two new enzymes, a UDP-glucose pyrophosphatase and a CoA pyrophosphatase, were identified.     

Specificity of two different purified acylcarnitine hydrolases from rat liver, their identity with other carboxylesterases, and their possible function

One of the previously described five purified monoglyceride-cleaving carboxylesterases from rat liver microsomes proved to be a carnitine ester hydrolase. This esterase, with an isoelectric point of 5.2, is most active with medium-chain acyl-L-carnitines (C12-C14). The esterase is also remarkably active with 1,3-diglycerides, especially 1,3-dioctanoylglycerol, that are hydrolyzed faster than the corresponding 1-monoglycerides and triglycerides. Only one of the other four purified carboxylesterases has moderate acylcarnitine-hydrolyzing activity. An altered procedure for the separation of the two microsomal acylcarnitine-cleaving enzymes is described. Both enzymes hydrolyze carnitine esters optimally at pH 8 and both are inactive with acetylcarnitine, palmitoyl-CoA, and butyrylthiocholine. The possible natural functions of the hydrolases are discussed. Besides their detoxifying action on natural membrane-lysing detergents (like carnitine esters and lysophospholipids), these enzymes could be involved in the transport of carnitine out of the liver.     

Acid hydrolases in early and late endosome fractions from rat liver.

We have examined the distribution of the cation-independent mannose 6-phosphate receptor and five acid hydrolases in early and late endosomes and a receptor-recycling fraction isolated from livers of estradiol-treated rats. Enrichment of mannose 6-phosphate receptor mass relative to that of crude liver membranes was comparable in membranes of early and late endosomes but was even greater in membranes of the receptor-recycling fraction. Enrichment of acid hydrolase activities (aryl sulfatase, N-acetyl-beta-glucosaminidase, tartrate-sensitive acid phosphatase, and cholesteryl ester acid hydrolase) and cathepsin D mass was also comparable in early and late endosomes but was considerably lower in the receptor-recycling fraction. The enrichment of two acid hydrolases, acid phosphatase and cholesteryl ester acid hydrolase, in endosomes was severalfold greater than that of the other three examined, about 40% of that found in lysosomes. Acid phosphatase and cholesteryl ester acid hydrolase were partially associated with endosome membranes, whereas cathepsin D was found entirely in the endosome contents. These findings raise the possibility that lysosomal enzymes traverse early endosomes during transport to lysosomes in rat hepatocytes and suggest that the greater enrichment of some acid hydrolases in endosomes is related to their association with endosome membranes. Despite the substantial enrichment of lysosomal enzymes in hepatocytic endosomes, we found that two, cholesteryl ester acid hydrolase and cathepsin D, did not degrade cholesteryl esters and apolipoprotein B-100 of endocytosed low density lipoproteins in vivo, presumably because they are inactive at the pH within endosomes.     

Specificity of purified monoacylglycerol lipase, palmitoyl-CoA hydrolase, palmitoyl-carnitine hydrolase, and nonspecific carboxylesterase from rat liver microsomes

The comparative substrate specificities of five purified serine hydrolases from rat liver microsomes have been investigated, especially their action upon natural lipoids. All enzymes had high carboxylesterase activities with simple aliphatic and aromatic esters and thioesters. The broad pH optima were in the range of pH 6-10. Synthetic amides were less potent substrates. The hydrolytic activities towards palmitoyl-CoA and monoacyl glycerols were generally high, whereas phospholipids and palmitoyl carnitine were cleaved at moderate rates. Acetyl-CoA, acetyl carnitine, and ceramides were not cleaved at all. The closely related hydrolases with the highest isoelectric points (pI 6.2 and 6.4) were most active with palmitoyl-CoA and palmitoyl glycerol. One of these enzymes might also be responsible for the low cholesterol oleate-hydrolyzing capacity of rat liver microsomes. Among the other hydrolases, that with pI 6.0 showed significant activities with simple butyric acid esters, 1-octanoyl glycerol, and octanoylamide. The esterase with pI 5.6 had the relatively highest activities with palmitoyl carnitine and lysophospholipids. The purified enzyme with pI 5.2 showed some features of the esterase pI 5.6, but generally had lower specific activities, except with 4-nitrophenyl acetate. The lipoid substrates competitively inhibited the arylesterase activity of the enzymes. The varying activities of the individual hydrolases were influenced in parallel by a variety of inhibitors, indicating that the purified hydrolases possessed a relatively broad specificity and were not mixtures of more specific enzymes. The nomenclature of the purified hydrolases is discussed.     

On the organ specificity of neutral glycerolester hydrolase of various tissues; an immunological study

Rat liver microsomal glycerol monoester hydrolase (EC 3.1.1.23) has been purified 130 fold. The enzyme has a molecular weight of about 60,000. An antibody raised against this enzyme in rabbit did not inhibit heparin-releasable liver lipase, which hydrolyses long-chain 1- and 2-monoglycerides effectively. This confirms an earlier conclusion, based on results obtained with an antibody raised against the latter enzyme, that the non-releasable and heparin-releasable liver enzymes are different proteins. The antibody against the liver microsomal glycerol monoester hydrolase, however, inhibited also the monoglyceridase activities of acetone powder extracts of rat small intestinal epithelial microsomes and rat epididymal fat pads, suggesting structural similarities between the endoplasmic reticulum hydrolases of various tissues. These findings also apply to pig where an antibody against adipose tissue lipases inhibits the monoglyceridase activities of small intestinal and liver microsomal acetone powder extracts.     

Induction of hepatic long-chain acyl-CoA hydrolase by clofibric acid administration

The induction of hepatic long-chain acyl-CoA hydrolase in the cytosolic fraction by administration of clofibric acid (p-chlorophenoxyisobutyric acid) was compared in rats, mice and guinea-pigs. In rats, two long-chain acyl-CoA hydrolases were induced by the administration of clofibric acid. In mice, only one long-chain acyl-CoA hydrolase was induced, and this hydrolase had properties similar to those of the lower-molecular-weight hydrolase induced in the hepatic cytosol of rats. In hepatic cytosol of guinea-pig, no hydrolase was induced by the administration of clofibric acid.     

Independent temporal expression of two N-substituted aminoacyl-tRNA hydrolases during the development of Artemia salina.

Encysted embryos of the crustacean Artemia salina contain an enzymatic activity which hydrolyzes N-acetylphenylalanyl-tRNA to N-acetylphenylalanine and tRNA. The enzyme apparently does not hydrolyze other free or N-substituted aminoacyl-tRNAs. The levels of this enzyme do not significantly change during embryonic and early larval development. In contrast, an unspecific hydrolase active on several N-substituted aminoacyl-tRNAs is practically absent in the encysted embryos and during embryogenesis and appears abruptly during larval development. The independent temporal expression of these two hydrolases during Artemia salina differentiation makes this organism siuitable for the study of the physiological role of these enzymes.     

Cytosolic and microsomal epoxide hydrolases are immunologically distinguishable from each other in the rat and mouse

Antibodies raised to homogeneous rat liver microsomal epoxide hydrolase were used to distinguish microsomal epoxide hydrolase from epoxide hydrolase of cytosolic origin in mice and rats. Using double diffusion analysis in agarose gels, we show that anti-rat liver microsomal epoxide hydrolase forms a single precipitin line with solubilized microsomes from rat and mouse liver, but no reaction is seen with the corresponding cytosolic fractions. Rat or mouse microsomal epoxide hydrolase activity (using benzo[a]pyrene 4,5-oxide as substrate) can be completely precipitated out of solubilized preparations by the antibody, which is equipotent against rat and mouse microsomal epoxide hydrolase. No precipitation of cytosolic hydrolase activity (using trans-beta-ethyl styrene oxide as substrate) is seen with any concentration of the antibody tested. Thus, in the case of microsomal epoxide hydrolase, extensive immunological cross-reactivity exists between the two species, rat and mouse. In contrast, no cross-reactivity is detectable between cytosolic and microsomal epoxide hydrolase, even when enzymes from the same species are compared. We conclude that microsomal and cytosolic epoxide hydrolase activities represent distinct and immunologically non-cross-reactive protein species.     

Purification of an epoxide hydrolase from Rhodotorula glutinis

The epoxide hydrolase from Rhodotorula glutinis was isolated and initially characterized. The enzyme was membrane associated and could be solubilized by Triton X-100. Purification yielded an enzyme with sp. act. of 66 mol 1,2-epoxyhexane hydrolyzed min^–1 mg^–1 protein. The enzyme was not completely purified to homogeneity but, nevertheless, a major protein was isolated by SDS-PAGE for subsequential amino acid determination of peptide fragments. From sequence alignments to related enzymes, a high homology towards the active site sequences of other microsomal epoxide hydrolases was found. Molecular mass determinations indicated that the native enzyme exists as a homodimer, with a subunit molecular mass of about 45 kDa. Based upon these, this epoxide hydrolase is structurally related to other microsomal epoxide hydrolases.     

Structure and function of a novel purine specific nucleoside hydrolase from Trypanosoma vivax

The purine salvage pathway of parasitic protozoa is currently considered as a target for drug development because these organisms cannot synthesize purines de novo. Insight into the structure and mechanism of the involved enzymes can aid in the development of potent inhibitors, leading to new curative drugs. Nucleoside hydrolases are key enzymes in the purine salvage pathway of Trypanosomatidae, and they are especially attractive because they have no equivalent in mammalian cells. We cloned, expressed and purified a nucleoside hydrolase from Trypanosoma vivax. The substrate activity profile establishes the enzyme to be a member of the inosine-adenosine-guanosine-preferring nucleoside hydrolases (IAG-NH). We solved the crystal structure of the enzyme at 1.6 A resolution using MAD techniques. The complex of the enzyme with the substrate analogue 3-deaza-adenosine is presented. These are the first structures of an IAG-NH reported in the literature. The T. vivax IAG-NH is a homodimer, with each subunit consisting of ten beta-strands, 12 alpha-helices and three small 3(10)-helices. Six of the eight strands of the central beta-sheet form a motif resembling the Rossmann fold. Superposition of the active sites of this IAG-NH and the inosine-uridine-preferring nucleoside hydrolase (IU-NH) of Crithidia fasciculata shows the molecular basis of the different substrate specificity distinguishing these two classes of nucleoside hydrolases. An "aromatic stacking network" in the active site of the IAG-NH, absent from the IU-NH, imposes the purine specificity. Asp10 is the proposed general base in the reaction mechanism, abstracting a proton from a nucleophilic water molecule. Asp40 (replaced by Asn39 in the IU-NH) is positioned appropriately to act as a general acid and to protonate the purine leaving group. The second general acid, needed for full enzymatic activity, is probably part of a flexible loop located in the vicinity of the active site.