Sensitive, soluble chromogenic substrates for HIV-1 proteinase.

By replacement of the P1' residue in a capsid/nucleocapsid cleavage site mimic with 4-NO2-phenylalanine (Nph), an excellent chromogenic substrate, Lys-Ala-Arg-Val-Leu*Nph-Glu-Ala-Met, for HIV-1 proteinase (kappa cat = 20 s-1, Km = 22 microM) has been prepared. Substitution of the Leu residue in P1 with norleucine, Met, Phe, or Tyr had minimal effects on the kinetic parameters (kappa cat and kappa cat/Km) determined at different pH values, whereas peptides containing Ile or Val in P1 were hydrolyzed extremely slowly. The spectrophotometric assay has been used to characterize the proteinase further with respect to pH dependence, ionic strength dependence, and the effect of competitive inhibitors of various types.

or by spectrophotometric monitoring of the decrease in absorbance at 300 nm resulting from hydrolysis of the scissile peptide bond (8). In some assays, rates were determined by averaging the absorbance decrease over the range of 284-324 nm using a Hewlett-Packard diode array spectrophotometer.

AND DISCUSSION
Previously published studies on HIV PR have concentrated largely on substrates that are based on the MA/CA junction in the gag polyprotein.
To explore other determinants for proteinase recognition, this work has focused on the -Leu*Alacleavage site at one of the CA/NC junctions.
Cleavage of Peptide 1 (Table I)   Samples of individual peptides were incubated with HIV-1 proteinase (between 25 and 100 ng) at 30 "C in 100 mM sodium acetate buffer, pH 4.7 containing 4 mM EDTA, 5 mM mercaptoethanol, and sufficient NaCl to give a final ionic strength of 1 M. Aliquots were removed at appropriate times and the reaction was stopped by the addition of 5% (v/v) trifluoroacetic acid for analysis by reverse-phase HPLC (for Peptides 1, 2, and 9-11). Alternatively, with Peptides 2-10, the cleavage reaction was monitored directly by following the decrease in absorbance at 300 nm (see the legends to Tables II, III,  and IV).
In previous reports with archetypal aspartic proteinases, it has been demonstrated that when a Nph residue occupies the P1' position, cleavage results in a shift of the absorbance maximum (6,8). Thus, the cleavage of Peptide 2 was also monitored by spectroscopy at 300 nm. The values for the initial rate of hydrolysis derived from the changes in absorbance also displayed Michaelis-Menten kinetics. At pH 4.7 and an ionic strength of 1 M, a K,,, of 22 PM, and kCat of 20 s-' were determined, in excellent agreement with the values given above from the HPLC method. The K, value is lower than any reported previously for a synthetic substrate interacting with HIV-l PR. The k,,, is comparable to that reported for the hydrolysis of a longer -Phe*Pro-containing peptide (7) and is in the range of values determined previously for catalysis by mammalian aspartic proteinases (6,8). These mean values were derived from six independent experiments.
However, when a different preparation of Peptide 2 was used as substrate, a mean value of 7 fiM was derived for the K,,, from three replicate estimations ( kCat was consistent between preparations).
This variation is most likely due to the well documented susceptibility of methionine residues in synthetic peptides to oxidation (9). Consequently, having demonstrated the value of a Nph-containing peptide that otherwise duplicates the sequence of residues at the CA/NC junction, the COOH-terminal Met residue in Peptide 2 was replaced by the more stable norleucine isostere. This was incorporated either as the -NH, or -Gly-NH2 derivatives in the series of synthetic substrates (Peptides 3-10) in which the residue in the Pi position was varied systematically to characterize further the binding site of HIV-l PR. The values obtained for k,., (Table II)   Ile -0 10 Val -0 all of the others terminated with -Nle-NH2 or -Nle-Gly-NH, (see Table I  kinetic parameters was carried out for the hydrolysis of Peptides 2 and 6 ( Fig. 1). The data points for both substrates fit well to a single line. In the upper panel (Fig. l), a decrease in the parameter log kcat/Km defines an apparent pK, of 5.7. This dissociation can arise from either the free enzyme or the free substrate. Peptides 2 and 6 both have a Glu residue in the PZ' position in a predominantly hydrophobic sequence. However, a similar pH dependence was observed for the hydrolysis of an unrelated peptide substrate that did not contain an acidic residue but which also included the sequence -Arg-Val-Tyr*Nph-Val-around the scissile peptide bond (data not shown). On this preliminary basis, then, it would appear that the dissociation may arise from a group in the enzyme, e.g. AspZ9 or Asp3", known to be in the vicinity of the substrate binding cleft (11). A decrease in binding affinity at higher pH has been observed previously with inhibitors and other substrates (7). A more detailed rationalization should be obtained by synthesis of further analogs of Peptide 6 containing Gln or other amino acids in the PZ' position together with mutagenesis of the HIV PR.
A second apparent proton dissociation is observed in the upperpanel of Fig. 1 at a pH of approximately 7.3. The change in molar extinction coefficient at 300 nm upon cleavage of Nph-containing peptides is too small at pH values of 7 or above to permit accurate measurement of rates of hydrolysis. Thus, the data above pH 7 (Fig. 1) were obtained by the HPLC method of analysis and hence are of lower precision. Once again, this ionization may arise from a side chain in the enzyme or from dissociation of a proton from a basic group in the NHz-terminal segment of the substrate (e.g. the primary amino group at the NH, terminus of both substrates) which would remove a positive charge and facilitate binding to the hydrophobic cleft. Synthesis of further peptides, acylated at the NH, terminus and/or with replacement of the Lys and Arg residues will resolve this issue. and Tris (pH 7.6-9.0). Reactions at 37°C were initiated by the addition of HIV-1 proteinase (into a final volume of 800 ~1) and monitored by the decrease in absorbance at 300 nm. Values for k,., and K,,, (expressed in millimolar) were derived from plots of initial velocity, thus determined spectrophotometrically, against substrate concentration.
Values at pH 7.0, 7.6,8.2 and 9.0 were also determined by the less precise HPLC method (7). To reflect this, a dashed line is used at these higher pH values. 0, Lys-Ala-Arg-Val-Leu*Nph-Glu-Ala-Met.
Cl, Lys-Ala-Arg-Val-Nle*Nph-Glu-Ala-Nle-NH*. of an inflection in this plot is in contrast to many previous investigations of archetypal aspartic proteinases (e.g. see Ref. 6) where the apparent loss of a proton (presumably from the catalytic aspartic acid residues) occurs at pH values between 5 and 6. Even with the renin branch of the aspartic proteinase family (12), activity declines rapidly at neutral pH and above. Thus, HIV PR would appear to be able to provide efficient catalysis at the pH values likely to exist within virally infected cells. The acid pH "optimum" for HIV-l PR described previously would appear to be a reflection of the inability of substrates to interact strongly with this enzyme rather than a reduction in catalytic potency at higher pH values.
Previous investigations on other retroviral proteinases have shown that high ionic strength facilitates active site interaction with oligopeptide substrates (13). The effect of ionic strength on the activity of HIV PR was thus examined using the spectrophotometric assay. As the ionic strength was raised from 0. 5  To emphasize the advantages of the spectrophotometric assay, inhibition constants were determined for several inhibitors (ranging in potency) of HIV PR.
The first two of these compounds were weak competitive inhibitors at pH 4.7 (KJ = 440 and 150 nM, respectively). The third was essentially ineffective at concentrations up to 19 PM.
It has been shown that -Tyr*Nph-, -Met*Nph-, or -Leu*-Nph-(as well as -Nle*Nph-or -Phe*Nph-) can all be substituted satisfactorily for the naturally occurring -Leu*Alascissile bond in a peptide that otherwise duplicates one of the CA/NC junctions (for terminology, see Ref. 14). In contrast, insertion of these three dipeptide pairs into sequences corresponding to the MA/CA, the other CA/NC, and (a derivative of) the RT/IN junctions in the HIV-1 polyprotein, to give, respectively, Arg-Ser-Gln-Asn-Tyr*Nph-Ile-Val-Gln, Asn-Thr-Ala-Thr-Ile-Met'Nph-Gln-Arg-Gln-Arg-Gly-Asn, and Arg-Arg-Gln-Val-Leu*Nph-Leu-Glu-Lys-Arg generated peptides which exhibited negligible rates of hydrolysis by HIV-l PR at pH 4.7. Therefore, not only the relative size/nature of the residues contributing the scissile peptide bond but also the identities of those residues in the flanking positions have a considerable influence on the conformation that can be adopted by the substrate within the active site cleft.
Results in this communication establish a simple, rapid method to quantitatively evaluate active site interactions in HIV-l PR and, potentially, in mutant forms of this enzyme. Utilizing synthetic peptide chemistry, further investigations are now in progress to attempt to capitalize on the newlyavailable three-dimensional structure of an inhibited-HIV-l PR complex (11) to detail the structural demands of these other subsites within the active site cleft of this essential viral enzyme.