Structural and functional characterization of Streptomyces plicatus beta-N-acetylhexosaminidase by comparative molecular modeling and site-directed mutagenesis

We have sequenced the Streptomyces plicatus beta-N-acetylhexosaminidase (SpHex) gene and identified the encoded protein as a member of family 20 glycosyl hydrolases. This family includes human beta-N-acetylhexosaminidases whose deficiency results in various forms of GM2 gangliosidosis. Based upon the x-ray structure of Serratia marcescens chitobiase (SmChb), we generated a three-dimensional model of SpHex by comparative molecular modeling. The overall structure of the enzyme is very similar to homology modeling-derived structures of human beta-N-acetylhexosaminidases, with differences being confined mainly to loop regions. From previous studies of the human enzymes, sequence alignments of family 20 enzymes, and analysis of the SmChb x-ray structure, we selected and mutated putative SpHex active site residues. Arg162 --> His mutation increased Km 40-fold and reduced Vmax 5-fold, providing the first biochemical evidence for this conserved Arg residue (Arg178 in human beta-N-acetylhexosaminidase A (HexA) and Arg349 in SmChb) as a substrate-binding residue in a family 20 enzyme, a finding consistent with our three-dimensional model of SpHex. Glu314 --> Gln reduced Vmax 296-fold, reduced Km 7-fold, and altered the pH profile, consistent with it being the catalytic acid residue as suggested by our model and other studies. Asp246 --> Asn reduced Vmax 2-fold and increased Km only 1.2-fold, suggesting that Asp246 may play a lesser role in the catalytic mechanism of this enzyme. Taken together with the x-ray structure of SmChb, these studies suggest a common catalytic mechanism for family 20 glycosyl hydrolases.     

The contribution of noncatalytic phosphate-binding subsites to the mechanism of bovine pancreatic ribonuclease A

The enzymatic catalysis of polymeric substrates such as proteins, polysaccharides or nucleic acids requires precise alignment between the enzyme and the substrate regions flanking the region occupying the active site. In the case of ribonucleases, enzyme-substrate binding may be directed by electrostatic interactions between the phosphate groups of the RNA molecule and basic amino acid residues on the enzyme. Specific interactions between the nitrogenated bases and particular amino acids in the active site or adjacent positions may also take place. The substrate-binding subsites of ribonuclease A have been characterized by structural and kinetic studies. In addition to the active site (p1), the role of other noncatalytic phosphate-binding subsites in the correct alignment of the polymeric substrate has been proposed. p2 and p0 have been described as phosphate-binding subsites that bind the phosphate group adjacent to the 3' side and 5' side, respectively, of the phosphate in the active site. In both cases, basic amino acids (Lys-7 and Arg-10 in p2, and Lys-66 in p0) are involved in binding. However, these binding sites play different roles in the catalytic process of ribonuclease A. The electrostatic interactions in p2 are important both in catalysis and in the endonuclease activity of the enzyme, whilst the p0 electrostatic interaction contributes only to binding of the RNA.     

Three-dimensional structure of phosphoenolpyruvate carboxylase: a proposed mechanism for allosteric inhibition

The crystal structure of phosphoenolpyruvate carboxylase (PEPC; EC 4. 1.1.31) has been determined by x-ray diffraction methods at 2.8-A resolution by using Escherichia coli PEPC complexed with L-aspartate, an allosteric inhibitor of all known PEPCs. The four subunits are arranged in a "dimer-of-dimers" form with respect to subunit contact, resulting in an overall square arrangement. The contents of alpha-helices and beta-strands are 65% and 5%, respectively. All of the eight beta-strands, which are widely dispersed in the primary structure, participate in the formation of a single beta-barrel. Replacement of a conserved Arg residue (Arg-438) in this linkage with Cys increased the tendency of the enzyme to dissociate into dimers. The location of the catalytic site is likely to be near the C-terminal side of the beta-barrel. The binding site for L-aspartate is located about 20 A away from the catalytic site, and four residues (Lys-773, Arg-832, Arg-587, and Asn-881) are involved in effector binding. The participation of Arg-587 is unexpected, because it is known to be catalytically essential. Because this residue is in a highly conserved glycine-rich loop, which is characteristic of PEPC, L-aspartate seemingly causes inhibition by removing this glycine-rich loop from the catalytic site. There is another mobile loop from Lys-702 to Gly-708 that is missing in the crystal structure. The importance of this loop in catalytic activity was also shown. Thus, the crystal-structure determination of PEPC revealed two mobile loops bearing the enzymatic functions and accompanying allosteric inhibition by L-aspartate.     

Ras-GRF, the activator of Ras, is expressed preferentially in mature neurons of the central nervous system

Neprilysin is a neutral peptidase that cleaves small peptide substrates on the amino-side of hydrophobic amino acid residues. In the present study, we have used inhibition of non-mutated and mutated enzymes with dipeptide inhibitors and hydrolysis of the substrate [Leu5, Arg6]enkephalin in order to evaluate the contribution of the S2' subsite to substrate and inhibitor binding. Our results suggest that (1) Arg-102 and Asn-542 provide major contributions to the interaction of the enzyme with the P2' residue of the substrate, (2) the S2' subsite is vast and can accommodate bulky side chains, and (3) Arg-102 restricts access to the S2' subsite to some side chains such as arginine.     

Localization of the active site of type II dehydroquinases. Identification of a common arginine-containing motif in the two classes of dehydroquinases

A novel method based on electrospray mass spectrometry (Krell, T., Pitt, A. R., and Coggins, J. R. (1995) FEBS Lett. 360, 93-96) has been used to localize active site residues in the type I and type II dehydroquinases. Both enzymes have essential hyper-reactive arginine residues, and the type II enzymes have an essential tyrosine residue. The essential hyper-reactive Arg-23 of the Streptomyces coelicolor type II enzyme has been replaced by lysine, glutamine, and alanine residues. The mutant enzymes were purified and shown by CD spectroscopy to be structurally similar to the wild-type enzyme. All three mutant enzymes were much less active, for example the kcat of the R23A mutant was 30,000-fold reduced. The mutants all had reduced Km values, indicating stronger substrate binding, which was confirmed by isothermal titration calorimetry experiments. A role for Arg-23 in the stabilization of a carbanion intermediate is proposed. Comparison of the amino acid sequence around the hyper-reactive arginine residues of the two classes of enzymes indicates that there is a conserved structural motif that might reflect a common substrate binding fold at the active center of these two classes of enzyme.     

Role of arginine 439 in substrate binding of 5-aminolevulinate synthase

5-Aminolevulinate synthase (EC 2.3.1.37) catalyzes the first reaction in the heme biosynthetic pathway in nonplant eukaryotes and some prokaryotes. Homology sequence modeling between 5-aminolevulinate synthase and some other alpha-family pyridoxal 5'-phosphate-dependent enzymes indicated that the residue corresponding to the Arg-439 of murine erythroid 5-aminolevulinate synthase is a conserved residue in this family of pyridoxal 5'-phosphate-dependent enzymes. Further, this conserved arginine residue in several enzymes, e.g., aspartate aminotransferase, for which the three-dimensional structure is known, has been shown to interact with the substrate carboxyl group. To test whether Arg-439 is involved in substrate binding in murine erythroid 5-aminolevulinate synthase, Arg-439 and Arg-433 of murine erythroid 5-aminolevulinate synthase were each replaced by Lys and Leu using site-directed mutagenesis. The R439K mutant retained 77% of the wild-type activity; its K(m) values for both substrates increased 9-13-fold, while the activity of R433K increased 2-fold and the K(m) values for both substrates remained unchanged. R439L had no measurable activity as determined using a standard 5-aminolevulinate synthase enzyme-coupled activity assay. In contrast, the kinetic parameters for R433L were comparable to those of the wild-type. Dissociation constants (Kd) for glycine increased 5-fold for R439K and at least 30-fold for R439L, while Kd values for glycine for both R433K and R433L mutants were similar to those of the wild-type. However, there was not much difference in methylamine binding among the mutants and the wild-type, excepting of a 10-fold increase in K(d)methylamine for R439L. R439K proved much less thermostable than the wild-type enzyme, with the thermotransition temperature, T1/2, determined to be 8.3 degrees C lower than that of the wild-type enzyme. In addition, in vivo complementation analysis demonstrated that in the active site of murine erythroid 5-aminolevulinate synthase, R439 is contributed from the same subunit as K313 (which is involved in the Schiff base linkage of the pyridoxal 5'-phosphate cofactor) and D279 (which interacts electrostatically with the ring nitrogen of the cofactor), while another subunit provides R149. Taken together, these findings suggest that Arg-439 plays an important role in substrate binding of murine erythroid 5-aminolevulinate synthase.     

Conserved residues are functionally distinct within transketolases of different species

Most of the amino acid residues which interact with thiamine pyrophosphate are highly conserved among enzymes which use this cofactor. The possible roles of several such residues in cofactor binding, catalysis, and/or substrate binding were examined for human transketolase. Mutations in H110 resulted in dramatic reductions to 2% or less of the normal activity. No alterations were found in the K(m)app's for the cofactor or for the donor and acceptor substrates. Alterations in Q428 resulted in a less severe loss of activity and also no changes in the K(m)app's. On the basis of the results, H110, an invariant residue, is proposed to function as a base which abstracts a proton from the protonated 4'-iminopyrimidine ring. The deprotonated 4'-imino moiety is required for generation of the C2-thiazolium carbanion which attacks the donor substrate. Interestingly, the function in the human enzyme of this invariant histidine is distinct from its role in yeast transketolase in which it aids in binding donor substrate and in subsequent catalytic events. Q428 is suggested to play a supportive role by stabilizing and orientating a water molecule which mediates the interaction between the 4'-amino group and H110. In other TPP-utilizing enzymes, the equivalent residue of Q428 is a histidine and is thought to deprotonate the 4'-amino group.     

Deletion of a highly motional residue affects formation of the Michaelis complex for Escherichia coli dihydrofolate reductase

Analysis of the dihydrofolate reductase (DHFR) complex with folate by two-dimensional heteronuclear (1H-15N) nuclear magnetic relaxation revealed that isolated residues exhibit diverse backbone fluctuations on the nanosecond to picosecond time scale [Epstein, D. M., Benkovic, S. J., and Wright, P. E. (1995) Biochemistry 34, 11037-11048]. These dynamical features may be significant in forming the Michaelis complex. Of these residues, glycine 121 displays large-amplitude backbone motions on the nanosecond time scale. This amino acid, strictly conserved for prokaryotic DHFRs, is located at the center of the betaF-betaG loop. To investigate the catalytic importance of this residue, we report the effects of Gly121 deletion and glycine insertion into the modified betaF-betaG loop. Relative to wild type, deletion of Gly121 dramatically decreases the rate of hydride transfer 550-fold and the strength of cofactor binding 20-fold for NADPH and 7-fold for NADP+. Furthermore, DeltaG121 DHFR requires conformational changes dependent on the initial binary complex to attain the Michaelis complex poised for hydride transfer. Surprisingly, the insertion mutants displayed a significant decrease in both substrate and cofactor binding. The introduction of glycine into the modified betaF-betaG loop, however, generally eliminated conformational changes required by DeltaG121 DHFR to attain the Michaelis complex. Taken together, these results suggest that the catalytic role for the betaF-betaG loop includes formation of liganded complexes and proper orientation of substrate and cofactor. Through a transient interaction with the Met20 loop, alterations of the betaF-betaG loop can orchestrate proximal and distal effects on binding and catalysis that implicate a variety of enzyme conformations participating in the catalytic cycle.     

Structure and mechanism of action and inhibition of steroid dehydrogenase enzymes involved in hypertension

Members of the NADPH-dependent short chain dehydrogenase/reductase (SDR) family control blood pressure, fertility, and natural and neoplastic growth. Despite the fact that only one amino acid residue is strictly conserved in the 100 known members of the family, all appear to have a dinucleotide-binding Rossmann fold and homologous catalytic residues including the conserved tyrosine. Variation in the binding pocket creates specificity for steroids, prostaglandins, sugars and alcohols. The critically important tyrosine appears to maintain a fixed position relative to the scaffolding of the Rossmann fold and the cofactor position, while the substrate-binding pocket alters in such a way that the dehydrogenation/reduction reaction site is brought into bonding distance of the tyrosine hydroxyl group. Licorice induces high blood pressure by inhibiting an SDR in the kidney. The crystal structure of the complex of 3alpha,20beta-hydroxysteroid dehydrogenase and carbenoxolone reveals the mechanism of enzyme inhibition by licorice. The most potent dehydrogenase enzyme inhibitors are those that displace substrate and cofactor and form strong hydrogen bonds to one or more amino acid residues involved in catalysis.     

Effects of substitutions of the conserved histidine residues in human gamma-glutamyl transpeptidase

gamma-Glutamyl transpeptidase possesses two histidine residues at positions 383 and 505 which are conserved in all mammalian and bacterial species. In order to elucidate the functions of these residues, we prepared mutants in which these residues were replaced by Ala. Kinetic analysis of the hydrolysis of L-gamma-glutamyl-p-nitroanilide indicated that substitution at His-383 decreased the Vmax value to 14% of that of the wild type, but had no effect on Vmax/K(m). In reactions involving glycylglycine as the acceptor substrate, the Vmax value of this mutant decreased to 38% with little alteration of Vmax/K(m) for L-gamma-glutamyl-p-nitroanilide as a gamma-glutamyl donor, but with a significant reduction of Vmax/K(m) for the acceptor. These results show that this substitution causes impairment of the step in which the free enzyme is regenerated from the gamma-glutamyl enzyme by water or an acceptor substrate. On the other hand, replacement of His-505 resulted in a decrease of the Vmax value for transpeptidation to about 10% of that of the wild type despite no substantial effect on the Vmax value for the hydrolysis reaction. However, this substitution did not affect Vmax/K(m) for the acceptor on transpeptidation. Thus, the formation of a non-productive enzyme-substrate complex with the acceptor substrate would decrease the Vmax value on transpeptidation. These results suggest that His-383 plays an important catalytic role in facilitating the degradation of the gamma-glutamyl-enzyme through hydrolysis or transfer of the gamma-glutamyl moiety to an acceptor. It was also shown that His-505 is important in the formation of a complex of the gamma-glutamyl enzyme with the acceptor substrate even though it plays no critical role in the catalysis. Although the pH-dependence profile and the van't Hoff plot for the ionic group responsible for enzyme activity were consistent with the requirement of a histidine residue, neither of the conserved histidines could be assigned as such an ionic group. This suggests that another histidine residue(s) might play an essential role in the enzyme function.     

Evidence for essential arginine residues at the active sites of maize branching enzymes

Alignment of 23 branching enzyme (BE) amino acid sequences from various species showed conservation of two arginine residues. Phenylglyoxal (PGO) was used to investigate the involvement of arginine residues of maize BEI and BEII in catalysis. BE was significantly inactivated by PGO in triethanolamine buffer at pH 8.5. The inactivation followed a time- and concentration-dependent manner and showed pseudo first-order kinetics. Slopes of 0.73 (BEI) and 1.05 (BEII) were obtained from double log plots of the observed rates of inactivation against the concentrations of PGO, suggesting that loss of BE activity results from as few as one arginine residue modified by PGO. BE inactivation was positively correlated with [14C]PGO incorporation into BE protein and was considerably protected by amylose and/or amylopectin, suggesting that the modified arginine residue may be involved in substrate binding or located near the substrate-binding sites of maize branching enzymes I and II.     

Relationship between conserved consensus site residues and the productive conformation for the TPQ cofactor in a copper-containing amine oxidase from yeast

A highly conserved asparagine residue is contained in the consensus site sequences of all known copper-containing amine oxidases (CAOs). On the basis of published crystallographic structures, the asparagine is found to reside proximal to the active site redox cofactor, 2,4,5-trihydroxyphenylalanine quinone (TPQ). In this study, the conserved asparagine was changed to an alanine in a CAO from Hansenula polymorpha expressed in Saccharomyces cerevisiae, and the mutant's catalytic properties were characterized using steady-state kinetics and resonance Raman spectroscopy. Several lines of evidence point to TPQ exisiting in an nonproductive orientation in the mutant, including reductions in several steady-state parameters and an accumulation of an inactive product Schiff base complex when the enzyme is incubated with methylamine as the substrate. This product Schiff base complex was previously found to form following mutation of another conserved consensus site residue, a glutamate (or aspartate) at the C + 1 position from TPQ [Cai, D., Dove, J., Nakamura, N., Sanders-Loehr, J., and Klinman, J. P. (1997) Biochemistry 36, 11472-11478]. The results suggest that these two residues are crucial in maintaining the balance of cofactor mobility versus rigidity expected to be necessary during the dual processes of biogenesis and catalysis, respectively, that all CAOs must accomplish. In addition, a previously unidentified structural linkage between these two highly conserved residues is proposed which spans both subunits of the dimeric CAOs, and may have implications for intersubunit communication.     

The conserved asparagine and arginine are essential for catalysis of mammalian adenylyl cyclase

Mammalian adenylyl cyclases have two homologous cytoplasmic domains (C1 and C2), and both domains are required for the high enzymatic activity. Mutational and genetic analyses of type I and soluble adenylyl cyclases suggest that the C2 domain is catalytically active and the C1 domain is not; the role of the C1 domain is to promote the catalytic activity of the C2 domain. Two amino acid residues, Asn-1025 and Arg-1029 of type II adenylyl cyclase, are conserved among the C2 domains, but not among the C1 domains, of adenylyl cyclases with 12 putative transmembrane helices. Mutations at each amino acid residue alone result in a 30-100-fold reduction in Kcat of adenylyl cyclase. However, the same mutations do not affect the Km for ATP, the half-maximal concentration (EC50) for the C2 domain of type II adenylyl cyclase to associate with the C1 domain of type I adenylyl cyclase and achieve maximal enzyme activity, or the EC50 for forskolin to maximally activate enzyme activity with or without Gsalpha. This indicates that the mutations at these two residues do not cause gross structural alteration. Thus, these two conserved amino acid residues appear to be crucial for catalysis, and their absence from the C1 domains may account for its lack of catalytic activity. Mutations at both amino acid residues together result in a 3,000-fold reduction in Kcat of adenylyl cyclase, suggesting that these two residues have additive effects in catalysis. A second site suppressor of the Asn-1025 to Ser mutant protein has been isolated. This suppressor has 17-fold higher activity than the mutant and has a Pro-1015 to Ser mutation.     

Molecular basis of a hereditary type I protein S deficiency caused by a substitution of Cys for Arg474

The molecular basis for a hereditary type I protein S (PS) deficiency was investigated. DNA sequence analysis in the proband showed a novel missense mutation substituting Cys (TGT) for Arg474 (CGT) that is a highly conserved amino acid residue among the related proteins. This missense mutation cosegregated with the type I PS deficiency in this family. Transient expression studies showed that the secretion of the recombinant Cys-mutant PS was markedly decreased compared with that of the recombinant wild-type PS, reproducing the observed phenotype of type I deficiency. Stable expression and pulse-chase experiments demonstrated an intracellular degradation and an impaired secretion of the recombinant Cys-mutant PS. Furthermore, the substitution of Arg474 by Ala or Glu, but not by Lys, markedly reduced the secretion of the recombinant PS mutants in transient expression studies, suggesting that a positively charged basic amino acid might be needed at residue 474 and might play a key role in the protein structure and conformation of the sex hormone binding globulin-homology domain of the PS molecule. We postulate that the loss of the highly conserved Arg474 might be responsible for the type I PS deficiency inherited in this family.     

Trypanosoma brucei gamma-glutamylcysteine synthetase. Characterization of the kinetic mechanism and the role of Cys-319 in cystamine inactivation

The parasitic protozoan Trypanosoma brucei utilizes a conjugate of glutathione and spermidine, termed trypanothione, in place of glutathione to maintain cellular redox balance. The first committed step in the biosynthesis of glutathione and thereby trypanothione, is catalyzed by gamma-glutamylcysteine synthetase (gamma-GCS). We have determined the kinetic mechanism for T. brucei gamma-GCS. The kinetics are best described by a rapid equilibrium random ter-reactant mechanism, in which the model derived Kd values for the binding of L-Glu, L-alpha-aminobutyrate, and ATP to free enzyme are 2.6, 5.1, and 1.4 mM, respectively. However, significant dependences exist between the binding of some of the substrate pairs. The binding of either ATP or L-Glu to the enzyme increases the binding affinity of the other by 18-fold, whereas the binding of L-Glu or L-alpha-aminobutyrate decreases the binding affinity of the other by 6-fold. Similarly to the mammalian enzyme, cystamine is a time-dependent, irreversible inhibitor of T. brucei gamma-GCS. It has been suggested by several studies that cystamine labels an active site Cys residue essential for catalysis. Among the enzymes reported to be inactivated by cystamine, only one Cys residue is invariant (Cys-319 in T. brucei gamma-GCS). Mutation of Cys-319 to Ala in T. brucei gamma-GCS renders the enzyme insensitive to cystamine inactivation without significantly affecting the enzyme's catalytic efficiency, kinetic mechanism, or substrate affinities. These studies suggest that cystamine inactivates the enzyme by blocking substrate access to the active site and not by labeling an essential active site residue.     

Glutamate-459 is important for Escherichia coli branching enzyme activity

The branching enzyme belongs to the amylolytic family, a group of enzymes that cleave and/or transfer chains of glucan. The amylolytic enzymes are homologous and all contain four conserved regions, proposed to contain the active site. By primary structure analysis, a conserved position unique to branching enzymes has been identified. This residue, which is either Asp or Glu, depending on the species, is located immediately after the putative catalytic Glu-458 (Escherichia coli numbering). Branching enzymes differ from other amylolytic enzymes in having this acid pair, and we asked if this motif could be essential for branching enzyme action. We used site-directed mutagenesis of the Glu-459 residue in the E. coli branching enzyme in order to determine the significance of the conserved Asp/Glu in branching enzymes. A substitution of Glu-459 to Asp resulted in increased specific activity compared to wild-type, suggesting that the mutation had created a more efficient enzyme. Changing Glu-459 to Ala, Lys, or Gln lowered the specific activities and altered the preferred substrate from amylose to amylopectin.     

A single cleavage assay for T5 5'-->3' exonuclease: determination of the catalytic parameters forwild-type and mutant proteins

Bacteriophage T5 5'-->3' exonuclease is a member of a family of sequence related 5'-nucleases which play an essential role in DNA replication. The 5'-nucleases have both exonucleolytic and structure-specific endo-nucleolytic DNA cleavage activity and are conserved in organisms as diverse as bacteriophage and mammals. Here, we report the development of a structure-specific single cleavage assay for this enzyme which uses a 5'-overhanging hairpin substrate. The products of DNA hydrolysis are characterised by mass spectrometry. The steady-state catalytic parameters of the enzyme are reported and it is concluded that T5 5'-->3' exonuclease accelerates the cleavage of a specific phosphodiester bond by a factor of at least 10(15). The catalytic assay has been extended to three mutants of T5 5'-->3' exonuclease, K83A, K196A and K215A. Mutation of any of these three lysine residues to alanine is detrimental to catalytic efficiency. All three lysines contribute to ground state binding of the substrate. In addition, K83 plays a significant role in the chemical reaction catalysed by this enzyme. Possible roles for mutated lysine residues are discussed.     

Contribution of arginine-82 and arginine-86 to catalysis of RNases from Bacillus intermedius (binase)

To elucidate the functional role of Arg82 and Arg86 in the enzyme activity of binase, the extracellular ribonuclease of Bacillus intermedius, we used site-directed mutagenesis. On cleavage of various substrates the catalytic activity of binase mutant Arg86 Ala is 2.7 x 10(3) - 7.7 x 10(3) times less than that of the native enzyme. The decrease in activity is determined preferentially by the decrease in the molecular rate constant kcat with a relatively small change of enzyme-substrate affinity, characterized by Km. This is the expected result if Arg86 acts to lower the energy of a transition state of the reaction. The replacement of Arg82 by Ala causes a 5-19-fold activity decrease, depending on the substrate. We propose that this residue does not have a direct catalytic function in the molecular mechanism of the binase action and that the activity decrease of binase on the replacement of Arg82 by alanine is mediated by the effect of Arg82 on the pK of catalytic residues.     

Structural mapping of the active site specificity determinants of human tissue-type plasminogen activator. Implications for the design of low molecular weight substrates and inhibitors

The recent structure determination of the catalytic domain of tissue-type plasminogen activator (tPA) suggested residue Arg174 could play a role in P3/P4 substrate specificity. Six synthetic chromogenic tPA substrates of the type R-Xaa-Gly-Arg-p-nitroanilide, in which R is an N-terminal protection group, were synthesized to test this property. Although changing the residue Xaa (in its L or D form) at position P3 from the hydrophobic Phe to an acidic residue, Asp or Glu, gave no improvement in catalytic efficiency, comparative analysis of the substrates indicated a preference for an acidic substituent occupying the S3 site when the S4 site contains a hydrophobic or basic moiety. The 2.9 A structure determination of the catalytic domain of human tPA in complex with the bis-benzamidine inhibitor 2, 7-bis-(4-amidinobenzylidene)-cycloheptan-1-one reveals a three-site interaction, salt bridge formation of the proximal amidino group of the inhibitor with Asp189 in the primary specificity pocket, extensive hydrophobic surface burial, and a weak electrostatic interaction between the distal amidino group of the inhibitor and two carbonyl oxygens of the protein. The latter position was previously occupied by the guanidino group of Arg174, which swings out to form the western edge of the S3 pocket. These data suggest that the side chain of Arg174 is flexible, and does not play a major role in the S4 specificity of tPA. On the other hand, this residue would modulate S3 specificity, and may be exploited to fine tune the specificity and selectivity of tPA substrates and inhibitors.