The cyclin-dependent kinases cdk2 and cdk5 act by a random, anticooperative kinetic mechanism.

cdk2.cyclin E and cdk5.p25 are two members of the cyclin-dependent kinase family that are potential therapeutic targets for oncology and Alzheimer's disease, respectively. In this study we have investigated the mechanism for these enzymes. Kinases catalyze the transfer of phosphate from ATP to a protein acceptor, thus utilizing two substrates, ATP and the target protein. For a two-substrate reaction, possible kinetic mechanisms include: ping-pong, sequential random, or sequential ordered. To determine the kinetic mechanism of cdk2.GST-cyclin E and cdk5.GST-p25, kinase activity was measured in experiments in which concentrations of peptide and ATP substrates were varied in the presence of dead-end inhibitors. A peptide identical to the peptide substrate, but with a substitution of valine for the phosphoacceptor threonine, competed with substrate with a K(i) value of 0.6 mm. An aminopyrimidine, PNU 112455A, was identified in a screen for inhibitors of cdk2. Nonlinear least squares and Lineweaver-Burk analyses demonstrated that the inhibitor PNU 112455A was competitive with ATP with a K(i) value of 2 microm. In addition, a co-crystal of PNU 112455A with cdk2 showed that the inhibitor binds in the ATP binding pocket of the enzyme. Analysis of the inhibitor data demonstrated that both kinases use a sequential random mechanism, in which either ATP or peptide may bind first to the enzyme active site. For both kinases, the binding of the second substrate was shown to be anticooperative, in that the binding of the first substrate decreases the affinity of the second substrate. For cdk2.GST-cyclin E the kinetic parameters were determined to be K(m, ATP) = 3.6 +/- 1.0 microm, K(m, peptide) = 4.6 +/- 1.4 microm, and the anticooperativity factor, alpha = 130 +/- 44. For cdk5.GST-p25, the K(m, ATP) = 3.2 +/- 0.7 microm, K(m, peptide) = 1.6 +/- 0.3 microm, and alpha = 7.2 +/- 1.8.

A fluorescence resonance energy transfer method for measuring the binding of inhibitors to stromelysin.

A sensitive fluorescence resonance energy transfer method was developed for the direct measurement of the dissociation constants of stromelysin inhibitors. The method is applied to the thiadiazole class of stromelysin inhibitors and it takes advantage of the fact that, upon binding to the active site of enzyme, the thiadiazole ring, with its absorbance centered at 320 nm, is able to quench the fluorescence of the tryptophan residues surrounding the catalytic site. The changes in fluorescence are proportional to the occupancy of the active site: Analysis of the fluorescence versus inhibitor concentration data yields dissociation constants that are in agreement with the corresponding competitive inhibitory constants measured by a catalytic rate assay. The affinity of nonthiadiazole inhibitors of stromelysin-such as hydroxamic acids and others-can be determined from the concentration-dependent displacement of a thiadiazole of known affinity. Using this displacement method, we determined the affinities of a number of structurally diverse inhibitors toward stromelysin. Since the three tryptophan residues located in the vicinity of the active site of stromelysin are conserved in gelatinase and collagenase, the method should also be applicable to inhibitors of other matrix metalloproteinases.

Dynamics of stromelysin/inhibitor interactions studied by 15N NMR relaxation measurements: comparison of ligand binding to the S1-S3 and S'1-S'3 subsites.

This report describes the backbone amide dynamics of the uniformly 15N labeled catalytic domain of human stromelysin complexed to PNU-99533, a hydroxamate-containing ligand that binds to the S'1-S'3 region (right side) of the stromelysin active site, and to PNU-107859 and PNU-142372, both thiadiazole-containing ligands that bind to the S1-S3 region (left side) of the stromelysin active site. 15N R1, R2 and NOE NMR relaxation measurements were recorded and analyzed for each complex. Different dynamic behaviors were observed for stromelysin complexed to the two classes of ligands, indicating that it may be possible to use protein dynamics to distinguish between different binding orientations. In the absence of bound ligand at the S1-S3 subsites, the S1-S3 residues were found to be relatively rigid. In contrast, the S'1-S'3 subsites were found to be flexible in the absence of interactions with ligand. The relative rigidness of the S1-S3 subsites may be responsible for MMP binding specificity by discriminating between ligands of different shapes. By contrast, the inherent flexibility of the S'1-S'3 subsites allows structural rearrangement to accommodate a broad range of incoming substrates or inhibitors. Similarities and differences in dynamics observed for each complex provide insights into the interactions responsible for protein-ligand recognition. The relevance of protein dynamics to structure-based drug design is discussed.

Thermodynamic and circular dichroism studies differentiate inhibitor interactions with the stromelysin S(1)-S(3) and S(')(1)-S(')(3) subsites.

Interactions of stromelysin with a series of inhibitors representative of three chemical templates with distinct binding modes were examined. Unfolding temperatures for inhibitor complexes were 10 degrees C to 15 degrees C greater than for apo stromelysin. Minor changes in ellipticity in the far-UV CD spectra of complexes indicated that ligand-induced conformational changes were localized to the binding site and did not involve gross changes in protein folding. Isothermal titrating calorimetry of thiadiazole-containing inhibitors, which bind in the S(1)-S(3) subsites of stromelysin, indicated that the binding interaction was exothermic and only slightly favorable entropically. Near-UV CD spectra showed large positive ellipticity increases from 250 to 300 nm, consistent with an interaction between the benzene ring of the inhibitor and stromelysin residues Tyr155 and Tyr168. Interactions between stromelysin and amide-hydroxamate ligands, which bind in the S(')(1)-S(')(3) subsites, were found to be both enthalpically and entropically driven. Binding of this class of ligands resulted in modest negative ellipticity changes at 260-285 nm and positive increases at 292 nm. Stromelysin complexed to a lactam-hydroxamate inhibitor with structure extending into both the S(1)-S(3) and S(')(1)-S(')(3) subsites showed increased ellipticity at 245 nm and negative changes at 260-285 and 295 nm.

Synthesis of a series of stromelysin-selective thiadiazole urea matrix metalloproteinase inhibitors.

The synthesis and enzyme inhibition data for a series of thiadiazole urea matrix metalloproteinase (MMP) inhibitors are described. A broad screening effort was utilized to identify several thiadiazoles which were weak inhibitors of stromelysin. Optimization of the thiadiazole leads to include an alpha-amino acid side chain with variable terminal amide substituents provided a series of ureas which were moderately effective stromelysin inhibitors, with Ki's between 0.3 and 1.0 microM. The most effective analogues utilized an L-phenylalanine as the amino acid component. In particular, unsubstituted 46 had a Ki of 710 nM, while the p-fluoro analogue 52 displayed increased potency (100 nM). Stromelysin inhibition was further improved using a pentafluorophenylalanine substituent which resulted in 70, a 14 nM inhibitor. While gelatinase inhibition was generally poor, the use of 1-(2-pyridyl)piperazine as the amide component usually provided for enhanced activity, with 71 inhibiting gelatinase with a Ki of 770 nM. The combination of this heterocycle with a p-fluorophenylalanine substituent provided the only analogue, 69, with collagenase activity (13 microM). The SAR for analogues described within this series can be rationalized through consideration of the X-ray structure recently attained for70 complexed to stromelysin. Uniquely, this structure showed the inhibitor to be completely orientated on the left side of the enzyme cleft. These results suggest that thiadiazole urea heterocycles which incorporate a substituted phenylalanine can provide selective inhibitors of stromelysin. Careful selection of the amide substituent can also provide for analogues with modest gelatinase inhibition.

The constituent tryptophans and bisANS as fluorescent probes of the active site and of a secondary binding site of stromelysin-1 (MMP-3).

The active site of the catalytic domain of stromelysin-1 (matrix metalloproteinase-3, MMP-3) was probed by fluorescence quenching, lifetime, and polarization of its three intrinsic tryptophans and by the environmentally sensitive fluorescent reporter molecule bisANS. Wavelength-dependent acrylamide quenching identified three distinct emitting tryptophan species, only one of which changes its emission and fluorescence lifetime upon binding of the competitive inhibitor Batimastat. Significant changes in the tryptophan fluorescence polarization occur upon binding by any of the three hydroxamate inhibitors Batimastat, CAS108383-58-0, and Celltech CT1418, all of which bind in the P2'-P3' region of the active site. In contrast, the inhibitor CGS27023A, which is thought to bind in the P1-P1' region, does not induce any change in tryptophan fluorescence polarization. The use of the fluorescent probe bisANS revealed the existence of an auxiliary binding site extrinsic to the catalytic cleft. BisANS acts as a competitive inhibitor of stromelysin with a dissociation constant of Ki=22 microM. In addition to this binding to the active site, it also binds to the auxiliary site with a dissociation constant of 3.40+/-0.17 microM. The auxiliary site is open, hydrophobic, and near the fluorescing tryptophans. The binding of bisANS to the auxiliary site is greatly enhanced by Batimastat, but not by the other competitive inhibitors tested.

Solution structures of stromelysin complexed to thiadiazole inhibitors.

Unregulated or overexpressed matrix metalloproteinases (MMPs), including stromelysin, collagenase, and gelatinase. have been implicated in several pathological conditions including arthritis and cancer. Small-molecule MMP inhibitors may have therapeutic value in the treatment of these diseases. In this regard, the solution structures of two stromelysin/ inhibitor complexes have been investigated using 1H, 13C, and 15N NMR spectroscopy. Both-inhibitors are members of a novel class of matrix metalloproteinase inhibitor that contain a thiadiazole group and that interact with stromelysin in a manner distinct from other classes of inhibitors. The inhibitors coordinate the catalytic zinc atom through their exocyclic sulfur atom, with the remainder of the ligand extending into the S1-S3 side of the active site. The binding of inhibitor containing a protonated or fluorinated aromatic ring was investigated using 1H and 19F NMR spectroscopy. The fluorinated ring was found to have a reduced ring-flip rate compared to the protonated version. A strong, coplanar interaction between the fluorinated ring of the inhibitor and the aromatic ring of Tyr155 is proposed to account for the reduced ring-flip rate and for the increase in binding affinity observed for the fluorinated inhibitor compared to the protonated inhibitor. Binding interactions observed for the thiadiazole class of ligands have implications for the design of matrix metalloproteinase inhibitors.

Structural characterizations of nonpeptidic thiadiazole inhibitors of matrix metalloproteinases reveal the basis for stromelysin selectivity.

The binding of two 5-substituted-1,3,4-thiadiazole-2-thione inhibitors to the matrix metalloproteinase stromelysin (MMP-3) have been characterized by protein crystallography. Both inhibitors coordinate to the catalytic zinc cation via an exocyclic sulfur and lay in an unusual position across the unprimed (P1-P3) side of the proteinase active site. Nitrogen atoms in the thiadiazole moiety make specific hydrogen bond interactions with enzyme structural elements that are conserved across all enzymes in the matrix metalloproteinase class. Strong hydrophobic interactions between the inhibitors and the side chain of tyrosine-155 appear to be responsible for the very high selectivity of these inhibitors for stromelysin. In these enzyme/inhibitor complexes, the S1' enzyme subsite is unoccupied. A conformational rearrangement of the catalytic domain occurs that reveals an inherent flexibility of the substrate binding region leading to speculation about a possible mechanism for modulation of stromelysin activity and selectivity.

A scintillation proximity assay for UDP-GalNAc:polypeptide, N-acetylgalactosaminyltransferase.

A rapid and simple method for quantitating the reaction product of UDP-GalNAc:polypeptide, N-acetylgalactosaminyltransferase (GalNAc-transferase) by scintillation proximity assay (SPA) was developed. The assay quantitates the radioactivity incorporated from 3H-labeled UDP-GalNAc into a biotin-labeled acceptor peptide, as measured after adsorption of the acceptor peptide to avidin-coated SPA beads. The acceptor peptide, PPASTSAPG (Elhammer et al. (1993) J. Biol. Chem. 268, 10029-10038) was conjugated to biotin using a di-beta-alanine spacer arm. The conjugated peptide reacted readily with the enzyme and it had an apparent Km comparable to that of the parent peptide. Using a reaction mixture consisting of 4 mg of SPA beads, 17 microM acceptor, 0.5 microM nucleotide sugar, and 7.5 U/ml enzyme, the time dependence of product formation obeyed Michaelis-Menten-type kinetics throughout the full course of the reaction-until exhaustion of the donor substrate-and the beginning portion of the reaction was sufficiently linear for calculating accurate initial rates. Analysis of the time dependency yielded an apparent Km of 0.38 +/- 0.12 microM for UDP-GalNAc. The assay is conveniently carried out in 96-well microtiter plates; it is ideally suited for assaying large numbers of samples and for screening large collections of chemicals for competitive inhibitors.

A heme-binding protein from hemolymph and oocytes of the blood-sucking insect, Rhodnius prolixus. Isolation and characterization.

A heme-binding protein has been isolated and characterized from both the hemolymph and oocytes of the blood-sucking insect, Rhodnius prolixus. The protein from both sources is identical in most aspects studied. The Rhodnius heme-binding protein (RHBP) is composed of a single 15-kDa polypeptide chain coiled in a highly alpha-helical structure which binds non-covalently one heme/polypeptide chain. This RHBP is not produced by limited degradation of hemoglobin from the vertebrate host, since specific polyclonal antibodies against it do not cross-react with rabbit hemoglobin, and since it differs from hemoglobin in having a distinct amino-acid composition and NH2-terminal sequence. The spectrum of the dithionite-reduced protein has peaks at 426, 530, and 559 nm and resembles that of a b-type cytochrome. RHBP from hemolymph is not saturated with heme and promptly binds heme added to the solution. The oocyte protein, on the other hand, is fully saturated and is not capable of binding additional heme.

A vector projection method for predicting the specificity of GalNAc-transferase.

The specificity of UDP-GalNAc:polypeptide N-acetylgalactosaminytransferase (GalNAc-transferase) is consistent with the existence of an extended site composed of nine subsites, denoted by P4, P3, P2, P1, P0, P1', P2', P3', P4', where the acceptor at P0 is being either Ser or Thr. To predict whether a peptide will react with the enzyme to form a Ser- or Thr-conjugated glycopeptide, a vector projection method is proposed which uses a training set of amino acid sequences surrounding 90 Ser and 106 Thr O-glycosylation sites extracted from the National Biomedical Research Foundation Protein Database. The model postulates independent interactions of the 9 amino acid moieties with their respective binding sites. The high ratio of correct predictions vs. total predictions for the data in both the training and the testing sets indicates that the method is self-consistent and efficient. It provides a rapid means for predicting O-glycosylation and designing effective inhibitors of GalNAc-transferase.

The tetracycline analogs minocycline and doxycycline inhibit angiogenesis in vitro by a non-metalloproteinase-dependent mechanism.

The tetracycline analogs minocycline and doxycycline are inhibitors of metalloproteinases (MMPs) and have been shown to inhibit angiogenesis in vivo. To further study the mechanism of action of these compounds we tested them in an in vitro model of angiogenesis: aortic sprouting in fibrin gels. Angiogenesis was quantitated in this system by a unique application of planar morphometry. Both compounds were found to potently inhibit angiogenesis in this model. To further characterize the activity of these compounds against MMPs, we determined the IC50S of both compounds against representatives of three classes of metalloproteinases: fibroblast collagenase, stromelysin, and gelatinase A. Doxycycline was found to inhibit collagenase, gelatinase A and stromelysin with IC50S of 452 microM, 56 microM and 32 microM, respectively. Minocycline was found to inhibit only stromelysin in the micromolar range with an IC50 of 290 microM. Since these results suggest that these compounds may not have been inhibiting in vitro angiogenesis by an MMP-dependent mechanism, we decided to test the effects of the potent MMP inhibitor BB-94. This compound failed to inhibit aortic sprouting in fibrin gels, thus strongly suggesting that both doxycycline and minocycline act by an MMP-independent mechanism. These results have implications for the mechanism of action of tetracycline analogs, particularly where they are being considered for the treatment of disorders of extracellular matrix degradation including periodontal disease, arthritis, and tumor angiogenesis.

Purification of a recombinant human respiratory syncytial virus chimeric glycoprotein using reversed-phase chromatography and protein refolding in guanidine hydrochloride.

FG glycoprotein is a recombinant chimeric protein consisting of the extracellular portions of human respiratory syncytial virus (RSV) F and G glycoproteins. In theory, highly purified FG glycoprotein may be effective as a RSV vaccine. Recombinant FG glycoprotein was expressed using the baculovirus/insect cell system. FG glycoprotein was isolated from cell culture supernatants using S Sepharose ion-exchange chromatography, Cu(2+)-immobilized metal affinity chromatography, preparative reversed-phase high-performance liquid chromatography, denaturation with 6 M guanidine hydrochloride, and protein refolding in Tween 80 detergent. The purified FG glycoprotein was concentrated on a S Sepharose column and exchanged into an appropriate buffer for vaccine formulation. Five batches of FG glycoprotein with protein purity of 92-99% were produced using this purification process. FG glycoprotein produced using reversed-phase chromatography and protein refolding was compared with nondenatured FG glycoprotein using a panel of 14 monoclonal antibodies directed against conformational and linear epitopes on RSV F and G glycoproteins. The results of these studies indicated that refolded FG glycoprotein had the same three-dimensional structure as nondenatured FG glycoprotein.

Interstitial collagenase is required for angiogenesis in vitro.

Human umbilical vein endothelial cells (HUVECs) invade collagen gels and establish vascular-like structures within the gel following stimulation with phorbol esters. This process was quantitated by measuring release of radioactivity from gels composed of [3H]collagen. Collagen was steadily degraded over the period of several weeks by phorbol ester-treated cells while little collagenolysis by cells not receiving phorbol ester was noted. Examination of matrix metalloproteinases (MMPs) secreted by HUVECs revealed a prominent induction of interstitial collagenase. Production of the mature forms of gelatinase A was also stimulated, as was the secretion of gelatinase B. Stromelysin was not detected. Two inhibitors of MMPs, the naturally occurring tissue inhibitor of metalloproteinases (TIMP; 10 micrograms/ml) and the synthetic, peptide inhibitor BB-94 (1 microM) were both effective at blocking HUVEC-mediated collagen degradation. Morphological examination of control, PMA-treated HUVECs, as well as PMA-treated HUVECs receiving TIMP or BB-94, revealed that MMP inhibition resulted in a block to invasion and tubule formation within the collagen gels. Similar results for MMP expression and inhibition of tubule formation in vitro were obtained with human dermal microvascular endothelial cells. Examination of collagen proteolytic fragments revealed that both BB-94 and TIMP blocked cleavage of the alpha 1 and alpha 2 chains of type I collagen and the appearance of tropocollagen fragments A and B, demonstrating that the inhibitors were acting directly upon interstitial collagenase. Our results demonstrate that interstitial collagenase is required for angiogenesis in vitro.

The specificity of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase as inferred from a database of in vivo substrates and from the in vitro glycosylation of proteins and peptides.

The acceptor substrate specificity of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase (GalNAc-transferase) was inferred from the amino acid sequences surrounding 196 O-glycosylation sites extracted from the National Biomedical Research Foundation Protein Database. When analyzed according to the cumulative enzyme specificity model (Poorman, R.A., Tomasselli, A.G., Heinrikson, R.L., and Kézdy, F.J. (1991) J. Biol. Chem. 266, 14554-14561) these data were found to be consistent with an enzymatic active site which interacts with an 8-amino-acid long segment of the substrate, spanning 3 amino acid residues preceding and 4 amino acid residues following the reactive serine or threonine. The model postulates independent interactions of the 8 amino acid moieties with their respective binding sites, designated as subsites P3 through P0 and P1' to P4'. High selectivity is expressed at all subsites toward serine, threonine, and proline. The inferred specificity was confirmed by in vitro bovine colostrum GalNAc-transferase-catalyzed glycosylation of unglycosylated proteins containing predicted sites for O-glycosylation and synthetic peptides designed to be GalNAc acceptors. In synthetic peptides the bovine colostrum GalNAc-transferase glycosylates threonine about 35 times faster than serine. Our results suggest that the specificity of the enzyme is not dependent on any particular secondary structure of the substrate but, rather, it is determined by the amino acids in the acceptor peptide segment as well as by the accessibility of this segment. It also appears likely that bovine colostrum GalNAc-transferase is able to catalyze in vivo the glycosylation of both threonine and serine residues.

Inhibition of acetaminophen and lorazepam glucuronidation in vitro by probenecid.

The effect of probenecid on glucuronidation of acetaminophen and lorazepam in hepatic microsomes from various species was studied to see if in vitro results were consistent with previous in vivo observations. Mouse, rat, and human microsomes were incubated with acetaminophen and probenecid while monkey microsomes were incubated with lorazepam and probenecid. Glucuronidation rates in all species varied with substrate, protein, and detergent concentrations. Mice exhibited faster rates of glucuronidation than rats or humans. All species showed inhibition of glucuronidation of acetaminophen or lorazepam when probenecid was added. Analysis suggested competitive inhibition. Thus, in vitro studies support in vivo results and confirm that the inhibition takes place at the hepatic level.

Probing the structure-function relationship of nerve growth factor.

We compared the receptor binding, antigenicity, biological activation, and cell-mediated proteolytic degradation properties of mouse nerve growth factor (mNGF) and human NGF (hNGF). The affinity of hNGF toward human NGF-receptor is greater than that of mNGF, but the affinity of mNGF toward rat NGF-receptor is greater than that of hNGF. Thus, the specificity of the interaction between NGF and its receptor resides both on the NGF and on its receptor. Using a group of anti-NGF monoclonal antibodies that competitively inhibit the binding of NGF to receptor, sites differing between mNGF and hNGF were detected. Together, these results indicate that the sites on hNGF and mNGF, responsible for binding to NGF-receptor, are similar but not identical. In comparing the relative abilities of mNGF and hNGF to stimulate a biological response in PC12 cells, we observed that mNGF was better at stimulating neurite outgrowth than was hNGF, consistent with the differences observed for receptor binding affinity. However, the ED50 for biological activation is approximately 100-fold lower than the Kd for receptor occupancy, and, thus, the dose-response curve is not consistent with a simple activation proportional to receptor occupancy. The data are consistent with a model requiring a low-level threshold occupancy of NGF-receptor (Kd = 10(-9) M) in order to stimulate full biological activity. Finally, we observed the degradation of NGF by PC12 cells. We found that the NGF molecule is significantly degraded via a receptor-mediated uptake mechanism. Together, the data provide insight into regions of the NGF molecule involved in contacts with the receptor leading to formation of the NGF:NGF-receptor complex. Additionally, they establish the link between occupancy of receptor and biological activation and the requirement for receptor-mediated uptake in order to degrade NGF proteolytically in cultured PC12 cells.

Zinc inhibition of renin and the protease from human immunodeficiency virus type 1.

We report here for the first time that Zn2+ is an effective inhibitor of renin and the protease from HIV-1, two aspartyl proteinases of considerable physiological importance. Inhibition of renin is noncompetitive and is accompanied by binding of 1 mol of Zn2+/mol of enzyme. Depending on the substrate, inhibition of the HIV protease by Zn2+ can be either competitive or noncompetitive, but in neither case is loss of activity due to disruption of the protease dimer. Inhibition of both enzymes is first order with respect to Zn2+ and is rapidly reversed by addition of EDTA. Ki values are strongly pH dependent and optimal in the range of 20 microM at or above pH 7. All of the data in hand suggest that the inhibitory effect of Zn2+ is a consequence of its binding at, or near, the active-site carboxyl groups of these aspartyl proteinases. This inhibition of the viral enzyme may help to explain some of the beneficial effects seen in AIDS patients who have received Zn2+ therapy.

Dissociative inhibition of dimeric enzymes. Kinetic characterization of the inhibition of HIV-1 protease by its COOH-terminal tetrapeptide.

Human immunodeficiency virus 1 (HIV-1) protease is an aspartyl protease composed of two identical protomers linked by a four-stranded antiparallel beta-sheet consisting of the NH2- and COOH-terminal segments (Weber, I.T. (1990) J. Biol. Chem. 265, 10492-10496). Kinetic analysis of the HIV-1 protease-catalyzed hydrolysis of a fluorogenic substrate demonstrates that the enzyme is an obligatory dimer. At pH = 5.0, 0.1 M sodium acetate, 1 M NaCl, 1 mM EDTA buffer, 37 degrees C, the equilibrium dissociation constant, Kd = 3.6 +/- 1.9 nM. We found that the tetrapeptide Ac-Thr-Leu-Asn-Phe-COOH, corresponding to the COOH-terminal segment of the enzyme, is an excellent inhibitor of the enzyme. Kinetic analysis shows that the inhibitor binds to the inactive protomers and prevents their association into the active dimer (dissociative inhibition). The dissociative nature of this inhibition is consistent with the results obtained from sedimentation equilibrium experiments in which the apparent molecular weight of the enzyme was observed to be 20,800 +/- 1,500 and 12,100 +/- 300, in the absence and presence of the COOH-terminal tetrapeptide, respectively. The dissociation constant of the protomer-inhibitor complex is Ki = 45.1 +/- 1.8 microM. This is the first kinetic analysis and direct experimental demonstration of noncovalent dissociative inhibition.

A cumulative specificity model for proteases from human immunodeficiency virus types 1 and 2, inferred from statistical analysis of an extended substrate data base.

Statistical analysis of an expanded data base of regions in viral polyproteins and in non-viral proteins that are sensitive to hydrolysis by the protease from human immunodeficiency virus (HIV) type 1 has generated a model which characterizes the substrate specificity of this retroviral enzyme. The model leads to an algorithm for predicting protease-susceptible sites from primary structure. Amino acids in each of the sites from P4 to P4' are tabulated for 40 protein substrates, and the frequency of occurrence for each residue is compared to the natural abundance of that amino acid in a selected data set of globular proteins. The results suggest that the highest stringency for particular amino acid residues is at the P2, P1, and P2' positions of the substrate. The broad specificity of the HIV-1 protease appears to be a consequence of its being able to bind productively substrates in which interactions with only a few Pi or Pi' side-chains need be optimized. The analysis, extended to 22 protein segments cleaved by the HIV-2 protease, delineates marked differences in specificity from that of the HIV-1 enzyme.