Comparative analysis of the X-ray structures of HIV-1 and HIV-2 proteases in complex with CGP 53820, a novel pseudosymmetric inhibitor.

BACKGROUND: The human immunodeficiency virus (HIV) is the causative agent of acquired immunodeficiency syndrome (AIDS). Two subtypes of the virus, HIV-1 and HIV-2, have been characterized. The protease enzymes from these two subtypes, which are aspartic acid proteases and have been found to be essential for maturation of the infectious particle, share about 50% sequence identity. Differences in substrate and inhibitor binding between these enzymes have been previously reported. RESULTS: We report the X-ray crystal structures of both HIV-1 and HIV-2 proteases each in complex with the pseudosymmetric inhibitor, CGP 53820, to 2.2 A and 2.3 A, respectively. In both structures, the entire enzyme and inhibitor could be located. The structures confirmed earlier modeling studies. Differences between the CGP 53820 inhibitory binding constants for the two enzymes could be correlated with structural differences. CONCLUSIONS: Minor sequence changes in subsites at the active site can explain some of the observed differences in substrate and inhibitor binding between the two enzymes. The information gained from this investigation may help in the design of equipotent HIV-1/HIV-2 protease inhibitors.

Structural and functional studies in vitro on the p6 protein from the HIV-1 gag open reading frame.

Protein p6 from HIV-1 gag open reading frame is reported to affect both the final phase of assembly of the viral particle and the early stage of the gag polyprotein maturation in vitro. Two separate hypotheses have been proposed, on only one of these reported effects. We think that both observations may be eventually explained if p6 protein strongly inhibits the HIV-1 proteinase. Protein p6 was synthesised by solid-phase peptide synthesis. Several methods of folding the p6 protein were tested, each resulting in the random structure according to both CD and 1D proton NMR spectra. A uniformly high exposure of NH protons to the solution was confirmed by temperature-dependent NMR spectra and isotope exchange experiments. Thus the p6 protein does not have any rigid conformation in solution. A rigid structure is not formed after further cleavage by HIV-1 proteinase as neither the protein nor its fragments are cleaved by this proteinase. In addition, the p6 protein itself does not act as inhibitor of HIV-1 proteinase. This excludes a direct role of p6 protein and supports the hypothesis that p6 is involved in forming the appropriate structure of gag polyprotein precursor. The role of slowly cleaved tight gag-proteinase in the final stage of maturation may be to slow down maturation of the precursor polyproteins prior to their transport to final location in the membrane.

Characterization of the proteinase inhibitor IIA from bull seminal plasma by 1H nuclear magnetic resonance. Stability, amide proton exchange and mobility of aromatic residues.

The isoinhibitor IIA from bull seminal plasma was investigated in aqueous solution by 1H nuclear magnetic resonance (n.m.r.). The analysis of the 1H n.m.r. data was based on individual resonance assignments, which are described in the following paper. Large conformation-dependent chemical shifts for aliphatic amino acid side-chains, numerous slowly exchanging amide protons and unusual pH titrations of two aromatic residues show that this protein forms a compact, globular conformation. This form of the protein is stable between pH 4 and 12 at 25 degrees C, and between 5 and 50 degrees C at pH 4.9. At temperatures above 50 degrees C there is evidence for an equilibrium between several different conformations, with the rate of exchange between the different species being in the intermediate range on the n.m.r. time-scale. Preliminary data are presented for the individual exchange rates of 18 backbone amide protons. Among the four aromatic rings, Phe10, Phe38 and Tyr16 undergo rapid 180 flips over the entire temperature range, whereas for Tyr32 a temperature-dependent transition from low-frequency to high-frequency flipping motions was observed.

Preliminary structural comparison of the proteinase isoinhibitors IIA and IIB from bull seminal plasma based on individual assignments of the 1H nuclear magnetic resonance spectra by two-dimensional nuclear magnetic resonance at 500 MHz.

By combined use of amino acid analysis, chemical sequence determination for the N-terminal decapeptide and two-dimensional 1H nuclear magnetic resonance at 500 MHz the amino acid sequence of bull seminal inhibitor IIB was found to coincide with that of the isoinhibitor IIA, except that the N-terminal tripeptide Pyrl-Gly2-Ala3- in HA is replaced by the dipeptide H-Leu2-Phe3- in IIB. Nearly complete, individual proton assignments were obtained for the isoinhibitor IIB, and comparison with the previously obtained corresponding nuclear magnetic resonance data for the isoinhibitor IIA showed that the two proteins must adopt closely similar secondary and tertiary structures in aqueous solution. The individual resonance assignments provide a basis for future, more detailed investigations of the influence of the local primary structure differences on the protein conformation.

Assignment of the 1H nuclear magnetic resonance spectrum of the proteinase inhibitor IIA from bull seminal plasma by two-dimensional nuclear magnetic resonance at 500 MHz.

The assignment of the 1H nuclear magnetic resonance (n.m.r.) spectrum of the protease inhibitor IIA from bull seminal plasma is described and documented. The assignments are based entirely on the amino acid sequence and on two-dimensional n.m.r. experiments at 500 MHz. Individual assignments were obtained at 18 degrees C and 45 degrees C for the backbone protons of all 57 amino acid residues, with the single exception of the N-terminal pyroglutamate amide proton. The amino acid side-chain resonance assignments are complete, with the exception of 17 long side-chains, i.e. Pro13, Met43 and all the Glu, Gln, Lys and Arg, where only one or two resonances of C beta H2 and in some cases C gamma H2 could be identified. The sequential assignments showed that the order of the two C-terminal residues in the previously established primary structure had to be changed; this was then confirmed by chemical methods. The chemical shifts for the assigned resonances at 18 degrees C and 45 degrees C are listed for an aqueous solution at pH 4.9. A preliminary characterization of the polypeptide secondary structure was obtained from the observed patterns of sequential connectivities.

Crystal structures of dialkylglycine decarboxylase inhibitor complexes.

The crystal structures of four inhibitor complexes of dialkylglycine decarboxylase are reported. The enzyme does not undergo a domain closure, as does aspartate aminotransferase, upon inhibitor binding. Two active-site conformations have been observed in previous structures that differ in alkali metal ion content, and two active-site conformations have been shown to coexist in solution when a single type of metal ion is present. There is no indication of coexisting conformers in the structures reported here or in the previously reported structures, and the observed conformation is that expected based on the presence of potassium in the enzyme. Thus, although two active-site conformations coexist in solution, a single conformation, corresponding to the more active enzyme, predominates in the crystal. The structure of 1-aminocyclopropane-1-carboxylate bound in the active site shows the aldimine double bond to the pyridoxal phosphate cofactor to be fully out of the plane of the coenzyme ring, whereas the Calpha-CO2(-) bond lies close to it. This provides an explanation for the observed lack of decarboxylation reactivity with this amino acid. The carboxylate groups of both 1-aminocyclopropane-1-carboxylate and 5'-phosphopyridoxyl-2-methylalanine interact with Ser215 and Arg406 as previously proposed. This demonstrates structurally that alternative binding modes, which constitute substrate inhibition, occur in the decarboxylation half-reaction. The structures of d and l-cycloserine bound to the active-site show that the l-isomer is deprotonated at C(alpha), presumably by Lys272, while the d-isomer is not. This difference explains the approximately 3000-fold greater potency of the l versus the d-isomer as a competitive inhibitor of dialkylglycine decarboxylase.

Substrates and inhibitors of human T-cell leukemia virus type 1 (HTLV-1) proteinase.

Human T-cell leukemia virus type 1 (HTLV-1) proteinase, a 125 residue polypeptide, was chemically synthesized using the solid phase method. The crude product was purified, renaturated and proteolytic activity was tested using oligopeptide substrates derived from processing sites of various retroviral polyproteins. Cleavage of the oligopeptide substrates together with an initial study using a series of HIV-1 and MAV (myeloblastosis associated virus) proteinase inhibitors suggest that the substrate specificity of HTLV-1 proteinase is very close to that of BLV (bovine leukemia virus) proteinase and distinct from that of both HIV-1 and MAV proteinases.

Peptide inhibitors of HIV-1 and HIV-2 proteases: a comparative study.

HIV-1 and HIV-2 proteases (PR) which play the key role in the formation of infectious viral particles offer a target for inhibitors that could block the maturation step. Inhibitors o HIV-1 PR exhibit mostly 1-2 orders of magnitude weaker affinity for HIV-2 PR. The subsite specificity study of the HIV-1 and HIV-2 proteases performed with inhibitors varying in the type of nonhydrolysable bonds and amino acid residues in the P1, P1'and P2'positions has led us to the design of inhibitors with 2S,4S and 2R,4S stereomeres of the hydroxyethylene isostere and Glu or Gln in the P2'positions. These compounds inhibit HIV-1 and HIV-2 proteases in vitro in subnanomolar concentrations and exhibit the activity in tissue culture.

Structure of a protein G helix variant suggests the importance of helix propensity and helix dipole interactions in protein design.

Six helix surface positions of protein G (Gbeta1) were redesigned using a computational protein design algorithm, resulting in the five fold mutant Gbeta1m2. Gbeta1m2 is well folded with a circular dichroism spectrum nearly identical to that of Gbeta1, and a melting temperature of 91 degrees C, approximately 6 degrees C higher than that of Gbeta1. The crystal structure of Gbeta1m2 was solved to 2.0 A resolution by molecular replacement. The absence of hydrogen bond or salt bridge interactions between the designed residues in Gbeta1m2 suggests that the increased stability of Gbeta1m2 is due to increased helix propensity and more favorable helix dipole interactions.

Crystal structures of a marginally active thymidylate synthase mutant, Arg 126-->Glu.

Thymidylate synthase (TS) is a long-standing target for anticancer drugs and is of interest for its rich mechanistic features. The enzyme catalyzes the conversion of dUMP to dTMP using the co-enzyme methylenetetrahydrofolate, and is perhaps the best studied of enzymes that catalyze carbon-carbon bond formation. Arg 126 is found in all TSs but forms only 1 of 13 hydrogen bonds to dUMP during catalysis, and just one of seven to the phosphate group alone. Despite this, when Arg 126 of TS from Escherichia coli was changed to glutamate (R126E), the resulting protein had kcat reduced 2000-fold and Km reduced 600-fold. The crystal structure of R126E was determined under two conditions--in the absence of bound ligand (2.4 A resolution), and with dUMP and the antifolate CB3717 (2.2 A resolution). The first crystals, which did not contain dUMP despite its presence in the crystallization drop, displayed Glu 126 in a position to sterically and electrostatically interfere with binding of the dUMP phosphate. The second crystals contained both dUMP and CB3717 in the active site, but Glu 126 formed three hydrogen bonds to nearby residues (two through water) and was in a position that partially overlapped with the normal phosphate binding site, resulting in a approximately 1 A shift in the phosphate group. Interestingly, the protein displayed the typical ligand-induced conformational change, and the covalent bond to Cys 146 was present in one of the protein's two active sites.

Subsite specificity of the proteinase from myeloblastosis associated virus.

The subsite requirements of the aspartic proteinase from the myeloblastosis-associated virus (MAV) for the cleavage of peptide substrates were studied with a series of synthetic peptides of general structure Ala-Thr-P4-P3-P2-P1*Nph-Val-Arg-Lys-Ala. The residues in positions P4, P3, P2 and P1 were varied and the kinetic parameters for the cleavage of substrates in 2.0 M NaCl were spectrophotometrically determined at pH 6.0 and 37 degrees C. The acceptance of amino acid residues in particular subsites is similar to that observed with the human immunodeficiency virus type 1 (HIV-1) proteinase in our earlier studies on the same substrate series: hydrophobic or aromatic residues are preferable in P1 position, a broad variety of residues are acceptable in P3 whereas the residues occupying P2 plays the decisive role in the substrate cleavage as evidenced by its dramatic influence on both kcat and Km values. The most remarkable difference between the two enzymes was found in P3 and P4 subsites. In P3, the introduction of negatively charged glutamate increases the substrate binding by the MAV proteinase 12-fold and decreases binding by the HIV-1 proteinase. In P4, Pro in this series is a favourable residue for the MAV proteinase and is strongly inacceptable for HIV-1 the proteinase. The pH profile of the cleavage was studied with a chromogenic substrate and differences between HIV-1 and MAV proteinases are discussed.

Processed enzymatically active protease (p15gag) of avian retrovirus obtained in an E. coli system expressing a recombinant precursor (Pr25lac-delta gag).

Processing proteases of avian and mammalian retroviruses cut the polyprotein precursors encoded by the retroviral genes into mature functional proteins. Retroviral processing proteases are still a rather poorly characterized group as to their relation to other proteases, specificity, and mechanism of enzymatic action. In avian retroviruses the generation of the processing protease itself comprises a processing cleavage event - the protease p15gag is cut off the carboxy-terminus of a gag polyprotein precursor, Pr76gag. We report here that direct and efficient production of the avian retrovirus processing protease p15gag (required for structure-function studies and rational design of inhibitors) was obtained in an E. coli system, where massive expression of a size-reduced, recombinant precursor (Pr25lac-delta gag) was accompanied by its structurally accurate processing.

Reduced-bond tight-binding inhibitors of HIV-1 protease. Fine tuning of the enzyme subsite specificity.

Truncation of a peptide substrate in the N-terminus and replacement of its scissile amide bond with a non-cleavable reduced bond results in a potent inhibitor of HIV-1 protease. A series of such inhibitors has been synthesized, and S2-S3' subsites of the protease binding cleft mapped. The S2 pocket requires bulky Boc or PIV groups, large aromatic Phe residues are preferred in P1 and P1' and Glu in P2'. The S3' pocket prefers Phe over small Ala or Val. Introduction of a Glu residue into the P2' position yields a tight-binding inhibitor of HIV-1 protease, Boc-Phe-[CH2-NH]-Phe-Glu-Phe-OMe, with a subnanomolar inhibition constant. The relevant peptide derived from the same amino acid sequence binds to the protease with a Ki of 110 nM, thus still demonstrating a good fit of the amino acid residues into the protease binding pockets and also the importance of the flexibility of P1-P1' linkage for proper binding. A new type of peptide bond mimetic, N-hydroxylamine -CH2-N(OH)-, has been synthesized. Binding of hydroxylamino inhibitor of HIV-1 protease is further improved with respect to reduced-bond inhibitor.

High-level expression of enzymatically active bovine leukemia virus proteinase in E. coli.

An E. coli plasmid expressing efficiently an artificial precursor of bovine leukemia virus (BLV) proteinase under transcriptional control of the phage T7 promoter was constructed. The expression product accumulates in the induced E. coli cells in the form of insoluble cytoplasmic inclusions. Solubilization of the inclusions and a refolding step yield almost pure and completely self-processed proteinase. Purification to homogeneity was achieved by ion-exchange chromatography and reverse-phase HPLC. On a preparative scale, a high yield of enzymatically active proteinase was obtained. An initial study using a series of synthetic peptide substrates shows a distinct substrate specificity of BLV proteinase.

Sub-site preferences of the aspartic proteinase from the human immunodeficiency virus, HIV-1.

A series of synthetic, chromogenic substrates for HIV-1 proteinase with the general structure Ala-Thr-His-Xaa-Yaa-Zaa*Nph-Val-Arg-Lys-Ala was synthesised with a variety of residues introduced into the Xaa, Yaa and Zaa positions. Kinetics parameters for hydrolysis of each peptide by HIV-1 proteinase at pH 4.7, 37 degrees C and u = 1.0 M were measured spectrophotometrically and/or by reverse phase FPLC. A variety of residues was found to be acceptable in the P3 position whilst hydrophobic/aromatic residues were preferable in P1. The nature of the residue occupying the P2 position had a strong influence on kcat (with little effect on Km); beta-branched residues Val or Ile in this position resulted in considerably faster peptide hydrolysis than when e.g. the Leu-containing analogue was present in P2.

A spectroscopic study of the hemin-human-alpha-fetoprotein system.

The binding of hemin to human alpha-fetoprotein has been estimated by means of fluorescence and spectrophotometric titration. Spectrophotometric titration discloses one strong binding site for hemin with an association constant of 1.5 X 10(7) M-1. The binding causes a shift of the absorption maximum to a higher wavelength and a rise in the molar absorption coefficient. Fluorescence reveals that the binding of hemin to human AFP quenches the protein fluorescence, which changes in character from a tryptophan type to a tyrosine type. As postulated by our results, the binding of hemin to human AFP is similar to the binding of hemin to HSA.

Fluorescence spectroscopy, fluorescence polarization, and circular dichroism in studies on pH-dependent changes in the alpha-fetoprotein molecule.

Conformational transitions of human alpha-fetoprotein (AFP) in the pH range from 2 to 12 were studied by fluorescence spectroscopy, fluorescence polarization, and circular dichroism in order to compare the molecular architecture of AFP with that of human serum albumin (HSA). In a previous paper we have found that AFP has a remarkably hydrophilic exposed molecular surface at neutral pH and possesses extensive hydrophobic binding sites located in crevices. Conformational changes of AFP occur in the acid and alkaline pH regions; extensive, hydrophobic areas of AFP are exposed by both acid and alkaline transitions. Similar results were obtained using optical methods. Both fluorescence and CD measurements disclosed conformational transitions of AFP induced by acidic or alkaline pH. Changes of the fluorescence intensity of AFP and HSA were compared with the model substances, that is, with lysozyme and Gly-Trp dipeptide. Again AFP and HSA show similar features, especially at the neutral pH and during alkaline transition. In the acid pH region, decrease of the AFP emission fluorescence intensity was greater than that of HSA and evidence for some irreversible conformational changes of AFP was obtained. CD spectra of both proteins also show a very similar pattern. The changes of molar ellipticity with pH for HSA are very much like those found for AFP. We have estimated high alpha-helix content--67% at pH 7.6 and 47% at pH 2.11. These figures are very close to those given for bovine albumin and rat alpha-fetoprotein. These findings provide additional support for our former findings that the molecular architecture of human alpha-fetoprotein has features similar to those of human serum albumin.

Biological tolerance of copolymers of 2-hydroxyethyl methacrylate-vinylpyridine.

The healing-in of implants of terpolymer gels having different chemical composition was studied in animals. Histological analysis showed that all polymers under investigation are healed-in in a similar way by encapsulation with collagen fibrous tissue. Neither cancerogenic effect nor calcification were observed within the interval studied.