Inactivation of HIV-1 nucleocapsid protein P7 by pyridinioalkanoyl thioesters. Characterization of reaction products and proposed mechanism of action.

The synthesis and antiviral properties of pyridinioalkanoyl thioester (PATE) compounds that target nucleocapsid p7 protein (NCp7) of the human immunodeficiency virus type 1 (HIV-1) have been described previously (Turpin, J. A., Song, Y., Inman, J. K., Huang, M., Wallqvist, A., Maynard, A., Covell, D. G., Rice, W. G., and Appella, E. (1999) J. Med. Chem. 42, 67-86). In the present study, fluorescence and electrospray ionization-mass spectrometry were employed to determine the mechanism of modification of NCp7 by two lead compounds, N-[2-(5-pyridiniovaleroylthio)benzoyl]sulfacetamide bromide and N-[2-(5-pyridiniovaleroylthio)benzoyl]-4-(4-nitrophenylsulfonyl )anili ne bromide (compounds 45 and 47, respectively). Although both compounds exhibit antiviral activity in cell-based assays, we failed to detect appreciable ejection of zinc from NCp7 under conditions in which previously described NCp7-active disulfides readily eject zinc. However, upon "activation" by Ag(+), compound 45 reacted with NCp7 resulting in the zinc ejection from both zinc fingers. The reaction followed a two-step mechanism in which zinc was ejected from the carboxyl-terminal zinc finger faster than from the amino-terminal zinc finger. Both compounds covalently modified the protein with pyridinioalkanoyl groups. Compound 45 modified cysteines 36 and 49 of the carboxyl-terminal zinc finger. The results obtained herein demonstrate that PATE compounds can be constructed that selectively target only one of the two zinc fingers of NCp7, thus providing an impetus to pursue development of highly selective zinc finger inhibitors.

Development of drug-resistant HIV 1 strains in response to therapy with inhibitors of the viral reverse transcriptase (1)(2)(3) and protease enzymes (4,5) has necessitated the search for novel antiretroviral agents that are directed against new molecular targets. The involvement of HIV-1 NCp7 zinc fingers in multiple phases of the HIV-1 replication cycle and their mutationally non-permissive nature has provided incentives for choosing this protein as a target for antiretroviral therapy. Moreover, mutations or modifications of either the conserved zinc chelating or non-chelating residues have resulted in loss of NCp7-mediated activities, including rendering the HIV noninfectious (6 -8). These observations gave impetus to explore several types of organic compounds that selectively target NCp7 protein. First among them being 3-nitrosobenzamide (9) followed by a series of 2,2Ј-dithiobis(benzamide) disulfides (DIBA) (10) and azodicarbonamide (11) that inhibited a wide range of HIV-1 isolates. Even though 2,2Ј-dithiobis(benzamide) disulfides represent a new class of highly specific antiretroviral agents, the disulfide bond is susceptible to reduction in vivo, resulting in the loss of antiviral activity. To circumvent this problem, we synthesized novel pyridinioalkanoyl thioester (PATE) derivatives (12). Of various such compounds synthesized, two of them, compounds 45 and 47 ( Fig. 1), showed superior antiviral activity. Both compounds were specific for NCp7, demonstrated antiviral activity in the presence of reduced glutathione, and showed minimal cytotoxicity. A close examination of their antiviral activities suggested that they act at different stages of viral replication. Compound 47 penetrated tumor necrosis factor-␣ induced U1 cells where it initiated Gag precursor cross-linking and inhibition of precursor processing. However, this compound was unable to initiate crosslinking of NCp7 protein within cell-free virions. In contrast, compound 45 did not inhibit the Gag precursor processing, and its anti-viral activity was ascribed to its ability to produce extensive cross-linking of NCp7 protein in cell-free virions (12).
The purpose of the current study was to determine the mechanism of action of PATEs and to determine the sites of modification on the NCp7 protein. For this purpose, a fluorescencebased assay using a zinc-specific fluorophore, Newport Green (NPG), and mass spectrometry were used to follow the kinetics of zinc ejection and to determine the site(s) of covalent modification on the target protein.

EXPERIMENTAL PROCEDURES
NCp7 Purification and Reconstitution-Recombinant NCp7 was purified using a pET3A-expressing NCp7 plasmid propagated in Escherichia coli strain BL21(DE3) pLysE. Briefly, the cells were grown at 37°C in the presence of 100 g/ml ampicillin and 34 g/ml chloramphenicol to an absorbance of 0.5 at 600 nm. Protein expression was induced with 1 mM isopropyl-␤-D-thiogalactopyranoside. After 3 h, the cells were harvested by centrifugation (13), lysed with buffer consisting of 50 mM Tris-HCl, pH 8.0, 10% (v/v) glycerol, 0.1 M NaCl, 0.1 mM ZnCl 2 , 5 mM dithiothreitol, 2 mM EDTA, and a protease inhibitor mixture (Roche Molecular Biochemicals). After clearing the debris by centrifugation, the supernatant was acidified with acetic acid (to a final 10% v/v), centrifuged again, and loaded onto a reverse phase C8 HPLC column (Vydac, Hesperia, CA). The chromatogram was developed using a linear gradient of 0 -55% acetonitrile, 0.04% trifluoroacetic acid.
NCp7 eluted at about 25% acetonitrile. The purity and integrity of the protein was confirmed by electrospray ionization mass spectrometry on a Finnigan MAT SSQ 7000 (San Jose, CA) mass analyzer. The apoprotein had a molecular mass of 6369 Ϯ 1 Da (calculated mass is 6369.4 Da). Lyophilized NCp7 was reconstituted with 20 mM sodium phosphate buffer, pH 7.2, containing 10% (v/v) glycerol and 0.1 mM ZnCl 2 . Excess zinc was removed using a Centricon-3 ultrafiltration unit (M r cut-off 3000 Da, Amicon), and the resulting reconstituted NCp7 was used in all the studies.
Peptide Synthesis-Peptides corresponding to amino-(13-30) and carboxyl-(32-52) terminal zinc fingers were synthesized by the solid phase method with Fmoc chemistry using an Applied Biosystems 430A peptide synthesizer (Applied Biosystems, Foster City, CA). Peptides were cleaved from the resin, and side chain protecting groups were removed by incubating in reagent K (trifluoroacetic acid/phenol/thioanisole/H 2 O/EDT, 82.5:5:5:5:2.5) for 3 h at room temperature. The peptides were purified by HPLC on a pH-stable C-8 column (Vydac, Hesperia, CA). The masses of the peptides were confirmed using Finnigan MAT SSQ 7000 mass analyzer.
NCp7 Zinc Ejection Monitored by NPG Fluorescence-The zinc ejection assay buffer consisted of 10% (v/v) glycerol, 20 mM sodium phosphate buffer, pH 7.2. Zinc ejection was monitored by following the increase in the fluorescence of a zinc selective fluorophore, Newport Green (10 M) (Molecular Probes, Eugene OR), in the assay buffer at room temperature. Under these conditions, NPG fluorescence increases linearly with increasing zinc concentration over the range 0.1 to 2 M ZnCl 2 . The effect of zinc concentration on NPG fluorescence was essentially unchanged upon increasing NPG to 20 M, while it decreased 10% upon increasing the phosphate concentration to 30 mM. Lowering the phosphate concentration to 10 mM did not significantly affect NPG fluorescence. Therefore, 10 M NPG and 20 mM phosphate were used as the standard condition. Zinc ejection was initiated by the addition of PATEs in dimethyl sulfoxide (5-80 M final concentration) to NCp7 (1 M) in assay buffer containing 10 M NPG at room temperature (24 Ϯ 1°C). The increase in fluorescence at 540 nm ( ex ϭ 490 nm) was monitored as a function of time using an Aminco Bowman Series 2 Luminescence Spectrometer. In experiments in which silver nitrate (20 M) was used, it was added prior to the addition of the PATEs, and the nonspecific increase in fluorescence was recorded.
Mass Spectrometry-NCp7 (1-2 M) treated with the PATEs was separated on a reverse phase C-18 column using a linear gradient (5-50%) of solvent B (90% acetonitrile containing 0.04% trifluoroacetic acid). In most cases, the modified protein eluted as two poorly separated peaks. Each peak was collected separately, lyophilized, and redissolved in 50% methanol containing 1% acetic acid for mass analysis on a Mass Analyzer SSQ 7000 (Finnigan).
For ESI-MS studies, the above peaks were redissolved in a minimal volume of 20 mM Tris buffer, pH 7.8, containing 5 mM CaCl 2 and 2 mM dithiothreitol. Clostripain (Promega) was added (1:20) and incubated for 3 h at 37°C. Mass spectrometric data for the clostripain peptides were obtained from collision-induced dissociation (CID) spectra using a Finnigan-MAT LCQ ion trap instrument equipped with electrospray interface (ESI) after introduction via a polyimide-coated fused silica microcapillary reverse phase-HPLC system (14). The mass spectrometer was set for analyzing the positive ions and was operated on either double or triple play mode in which the instrument was set up to automatically acquire: 1) a full scan, 2) a ZoomScan of the (MϩnH) n ϩ ion above a preset threshold, and 3) a tandem MS/MS spectrum (relative collision energy ϭ 30%) from that ion. ZoomScan was not monitored during the double play mode. The observed masses resulting from CID were compared with the predicted pattern of NCp7 fragmentation generated by the program Protein Prospector (University of California, San Francisco, CA).

RESULTS
Zinc Ejection from NCp7 by the Pyridinioalkanoyl Thioesters-Since the PATEs analyzed in this study had a very high absorbance at 280 nm, the commonly used method of monitoring the rate of zinc ejection by following the decrease in tryptophan fluorescence (excitation at 280 nm) was not feasible (12,15). Instead, a method of monitoring zinc ejection directly using NPG was adopted. NPG exhibits an increase in fluorescence upon binding to Zn 2ϩ . Upon the addition of NCp7, there was a small but definite increase in the NPG fluorescence, which remained constant over the monitored time course. Addition of either compound 45 or 47 (5-80 M) (N-[2-(5-pyridiniovaleroylthio)benzoyl]sulfacetamide bromide and N-[2-(5-pyridiniovaleroylthio)benzoyl]-4-(4-nitrophenylsulfonyl)aniline bromide, compounds 45 and 47, respectively) did not result in any apparent release of zinc from NCp7. Even after prolonged incubation (18 h) in the presence of these compounds, zinc remained bound to NCp7 as demonstrated by the subsequent rapid ejection of zinc upon addition of dithiane 1,1-dioxide (16), a previously established NCp7 inhibitor (Fig. 2). This result suggested that the PATEs alone may be considerably less reactive than the disulfides and may require "activation" in vitro. Thioesters are selectively activated by silver ions, a method routinely employed in large segment condensation in peptide synthesis (17,18). Therefore, we evaluated the effect of silver and other metal ions to "activate" the PATEs. Zinc ejection was initiated by the addition of compound 45 (5-80 M) to NCp7 (1 M) in zinc ejection assay buffer containing 10 M NPG and 20 M AgNO 3 . The zinc ejection was rapid and appeared to be complete by 5 min. Silver ions alone were unable to eject zinc during the time course employed. Interestingly, no zinc ejection was measurable with compound 47 even after a 1-h incubation, indicating a much slower process than with compound 45.
Kinetics of Zinc Ejection by Compound 45-Time-dependent zinc ejection was monitored at different concentrations of compound 45. Representative curves for three concentrations are shown in Fig. 3A. The data fit well to a double exponential curve with well separated time constants. The dependence of the observed rate constants on the concentration of compound 45 is shown in Fig. 3B. The observed rate constant for the faster phase, k f,obs , increases with increasing concentration of compound 45. The observed rate constant for the slower phase is independent of the concentration of compound 45, within experimental error, and can be expressed as k s,obs ϭ 4.1 Ϯ 1.2 ϫ 10 Ϫ3 s Ϫ1 .
The kinetic data were fit to a two-step mechanism, in which the first step consisted of a reversible, bimolecular reaction between the protein and the PATE compound, followed by covalent modification of the protein and ejection of 1 eq of zinc. In the second step, a conformational change or rearrangement in the protein is rate-limiting and results in the release of a second equivalent of zinc, independent of the concentration of compound 45. This mechanism can be expressed as, The large separation in the time constants of the initial fit to the zinc release data suggested that the two phases were largely distinct and could be considered separately. Applying a standard steady-state assumption for the intermediate I results in, is the concentration of compound 45. The inverse of the observed rate constant is indeed linear with the inverse of the concentration of compound 45 for the five highest concentrations and exhibits a distinctly non-zero intercept (not shown). From this analysis, k 2 ϭ 0.125 s Ϫ1 and k 1 /(k Ϫ1 ϩ k 2 ) ϭ 2.2 ϫ 10 4 M Ϫ1 for the faster phase of zinc ejection. The expected dependence of the rate constant k ss on the concentration of compound 45 is shown in Fig. 3B as a solid line and gives a good fit to the observed rate constants for the faster phase. The observed rate constants are somewhat faster than expected for the two lowest concentrations of compound 45, the sort of deviation expected when pseudo-first order conditions are not strictly met. The value of k 3 is determined from the average value of k s,obs and is shown in Fig. 3B as a dashed line, indicating the lack of dependence of the observed rate constant on the concentration of compound 45. Application of the kinetic model described by Equations 1 and 2 without assuming a steady state did not lead to a significantly improved fit to the data. The overall yield of the reaction, as indicated by the plateau value of the relative fluorescence shown in Fig. 3A, is distinctly lower at the lowest concentration of compound 45 and increases with increasing concentration to a limiting value. The kinetic model described by Equations 1 and 2 contains sequential irreversible reactions preceded by a reversible binding step. For reactions containing limiting amounts of NCp7, this model predicts that the eventual extent of the reaction should be the same, regardless of the concentration of compound 45, although the rate of the reaction is dependent on compound 45 concentration. The low amount of relative fluorescence initially established in a reaction containing 5 M compound 45 increases rapidly to the limiting plateau value upon subsequent addition of compound 45 to a final concentration of 60 M (data not shown). This result indicates that the limiting plateau values obtained at low compound concentration do not result from irreversible adsorption of the protein or precipitation of silver ion. The lower plateau values obtained at low compound concentration may be the consequence of the requirement for activation of the compound, which for unknown reasons is less efficient at low concentrations. We note, however, that similarly reduced plateau values have been reported for inactivation of NCp7 by lower concentrations of the unrelated disulfide benzamides, which do not require activation (15).
To identify which zinc finger of NCp7 reacted rapidly with compound 45, peptides corresponding to the amino-and carboxyl-terminal zinc fingers (residues 13-30 and 32-52, respec- tively) were synthesized. Previously, it has been shown that the isolated zinc fingers of NCp7 bind zinc with affinities similar to that of the whole protein (19). The zinc ejection from both fingers by compound 45 was monitored as described for the whole protein. Interestingly, compound 45 was unable to eject zinc from the amino-terminal finger (Fig. 4A). The fact that the amino-terminal zinc finger was indeed coordinated with zinc was demonstrated by addition of dithiane 1,1-dioxide, which ejected zinc as expected. Compound 45 caused ejection of zinc from the carboxyl-terminal zinc finger as shown in Fig. 4B. The data are well described as a single exponential curve.
The observed rate constant for zinc ejection from the isolated carboxyl-terminal zinc finger peptide increases with the concentration of compound 45 over the range 5 to 60 M, with no further increase at 80 M, as shown in Fig. 4C. Following the same course of analysis as was used for the intact protein, a steady state assumption was applied. The double reciprocal plot for the five highest concentrations is linear, suggesting that a steady state analysis provide a good description at the higher concentrations (not shown). The value of k 2 ϭ 0.95 s Ϫ1 is obtained from the intercept and the value of k 1 /(k Ϫ1 ϩ k 2 ) ϭ 5.6 ϫ 10 4 M Ϫ1 from the ratio of the slope and the intercept. The dependence of the steady state rate constant on the concentration of compound 45 is shown in Fig. 4C as a solid line and provides a good fit to the observed rate constants. Furthermore, the qualitative agreement of the parameters obtained for intact NCp7 and the isolated carboxyl-terminal zinc finger peptide indicates that the faster phase of zinc release from the intact protein can be ascribed to the carboxyl-terminal zinc finger.
Covalent Modification of NCp7 with PATEs-To determine whether the zinc ejection is accompanied by covalent modification of the protein, the products of the reaction of NCp7 (2 M) and a PATE (40 M) in the presence of AgNO 3 (10 M) were separated on reverse phase C18 column using a water/acetonitrile gradient system. The protein eluted in 2 fractions, which were collected separately, and the mass of the protein products was determined on an SSQ 7000 (Finnigan) mass analyzer. The results are tabulated in Table I. In the absence of any modification, NCp7 had an apparent molecular mass of 6369 Ϯ 1 Da, corresponding to the apoprotein (zinc is lost from the protein under conditions used to separate species by HPLC). When silver ion was added to NCp7, the hydrophobicity of NCp7 decreased, as indicated by its earlier elution on a reverse phase C18 column, and the molecular mass increased to 6906 Ϯ 1 Da, corresponding to apo-NCp7 with 5 bound silver ions. Since there are 4 dicarboxylic amino acids (three Glu and one Asp) plus the carboxyl terminus of the protein, it is expected that the silver ions form a salt with NCp7. Upon addition of a PATE, a new, slightly more hydrophobic peak was observed containing a mixture of modified peptides with molecular masses of 7066 Ϯ 1 Da and 7026 Ϯ 1 Da corresponding to apo-NCp7 with 5 silver atoms plus one or two pyridinioalkanoyl groups (⌬m ϭ 162 Ϯ 1 Da per group), respectively. Interestingly, although no appreciable zinc ejection was observed with compound 47 even after a 1-h incubation, a minor amount of NCp7 modified with one or two pyridinioalkanoyl groups was detected (Table I).
Other metals were tested for their ability to promote zinc ejection and covalent modification of NCp7 by compound 45. Only Fe 2ϩ (40 M) was able to do so after 4 h of incubation, while Fe 3ϩ , Mg 2ϩ , and Ca 2ϩ showed no effect. Ferrous ions, unlike silver, did not form a stable complex with NCp7 that was detected by MS and the resulting modified protein had a mass of 6531 Ϯ 1 Da, corresponding to the addition of a single pyridinioalkanoyl group. In addition, the yield of the modified protein was much lower than that obtained when silver ions were used as the activator.
Identification of the Modified Residues by Tandem Mass Spectrometry-The NCp7 product produced by reaction with compound 45 was subjected to proteolytic digestion with clostripain, and the digest was analyzed by LC/MS/MS. Clostripain, which cleaves on the carboxyl side of arginine, is expected to cleave the protein into three large and three small peptides. Table II lists the ions and sequence positions obtained. The mass of the fragment corresponding to the carboxyl-terminal zinc finger, amino acids 32-52, indicated addition of two pyridinioalkanoyl groups per fragment (⌬m ϭ 162 Ϯ 1 per group). Fragments corresponding to the amino-terminal zinc finger did not exhibit evidence of modification.
The exact site of the modification was determined by CID of the modified fragment. As shown in Fig. 5A, the CID spectrum of the ϩ6 ion with a parent mass of 2675.82 Da contains a series of b ions that exhibit evidence of modification as of ions, exhibiting a mass shift of 268 Da mass units, is consistent with the peptide being modified at Cys 49 by one pyridinioalkanoyl group and one silver atom at Glu 51 (Fig. 5, panel B). Two other intense ions (m/z of 656.9 and 676.7) correspond to doubly charged internal fragments of modified b 10 and y 10 -NH 3 , respectively. Collectively, these data demonstrate the modification of NCp7 by two pyridinioalkanoyl groups, one at each of Cys 36 and Cys 49 . The covalent modification of the zinc finger peptides by compound 45 was confirmed by HPLC followed by mass analysis. The observed masses of the apopeptides of 1976.94 and 2352.8 Da agree well with the calculated values of 1977.5 (aminoterminal peptide) and 2352.04 Da (carboxyl-terminal peptide) in the absence of silver. As expected, the observed masses shift to 2191.3 and 2672 Da in the presence of silver corresponding to the addition of two and three silver atoms to the amino-and carboxyl-terminal peptides, respectively. Upon addition of compound 45 (40 M), the mass of carboxyl-terminal peptide increased, corresponding to various modifications listed in Table  III. In accord with results presented above, the amino-terminal zinc finger peptide did not show evidence of any covalent modification. The modified peptides were analyzed by CID, yielding a spectrum dominated by the ϩ3 and ϩ4 charged internal peptides (data not shown). As with the carboxyl-terminal zinc finger fragment from the intact protein, we observed b series ions indicating modification of Cys 36 and y series ions indicating modification of Cys 49 with pyridinoalkanoyl groups.

DISCUSSION
Understanding the detailed mechanism of action of an antiviral agent and identifying the possible site(s) of interaction on the target protein is important in developing highly optimized and specific drugs. In this study, the mechanism of NCp7 inhibition by two novel PATEs, compounds 45 and 47, was probed with fluorescence-based assays and mass spectrometric techniques. Although both compounds were very active in cell and target based assays in inhibiting HIV-1, and preliminary studies indicated that these compounds acted through modification of NCp7 zinc fingers (12), they failed to show any appreciable zinc ejection in vitro from purified, recombinant NCp7 or its synthetic peptides under the standard zinc ejection assay conditions. However, zinc ejection was evident upon "activation" with silver ions. Surprisingly, the zinc ejection was observed only with compound 45 and not with compound 47. A detailed analysis of the time course of zinc ejection produced by compound 45, in the presence of silver, suggested the formation of a loosely bound intermediate prior to zinc ejection. Zinc ejection followed a double exponential curve that can best be interpreted as the ejection of carboxyl-terminal finger Zn 2ϩ at a rate 15-fold greater than ejection of amino-terminal Zn 2ϩ (Fig. 3). Further proof of the specificity of interaction and sequential release of Zn 2ϩ from two fingers was obtained with synthetic peptides corresponding to amino-and carboxyl-terminal zinc fingers (Fig. 4). Interestingly, the complete lack of reactivity of the isolated amino-terminal peptide provides further support that the release of the second equivalent of zinc from intact NCp7 results from the collapse of the protein structure and not a direct attack of the PATE on the amino-terminal zinc finger. In accord with this observation, mass analysis of compound 45-treated NCp7 as well as the zinc finger peptides revealed that both modifications were present exclusively on the carboxyl-terminal zinc finger. Sequence data indicated that the peptide corresponding to carboxyl-terminal zinc finger (32-52) was modified at Cys 36 and Cys 49 as b and y ions with a mass shift of 162 Ϯ 1 mass units were found. Various multiply charged internal fragments corresponding to these modified cysteines were also observed.
It has been previously demonstrated that 2,2Ј-dithiopyridine (20), disulfide benzamides (15), and 2,2Ј-dithiobis(benzamide) (21), all disulfide-based inhibitors, follow a similar, sequential   zinc ejection pathway. By comparing the rates of zinc ejection using Trp 37 fluorescence (an indicator of carboxyl-terminal zinc ejection) and zinc-specific fluorophores (indicator of zinc ejection from the whole protein), it has been postulated that carboxyl-terminal zinc is released prior to amino-terminal zinc. Our results provide direct evidence of this differential zinc ejection. The greater reactivity of the carboxyl-terminal zinc finger toward attack by a variety of electrophiles has been attributed to a combination of steric factors and differences in nucleophilicity of the cysteines (22)(23)(24). Molecular modeling studies based on density-functional theory suggested that the thiolate of Cys 49 is the most electron rich and the most reactive toward highly polarizable electrophiles (23). Molecular modeling studies show that the reactive sites of the carboxyl-terminal zinc finger lie in a more contiguous reactive surface compared with the amino-terminal finger (24). Docking of some antiviral agents, 2,2Ј-dithiobis(benzamide) disulfides-1 in particular, revealed that the structure of the carboxyl-terminal finger allows a significantly closer approach of the ligands to the cysteine thiolates, compared with the amino-terminal finger (24).
As mentioned earlier, zinc ejection was observed only with compound 45. However, mass spectrometry, which is more sensitive than the zinc ejection assay, revealed that compound 47 also modified NCp7 to a limited extent (Table I), resulting in addition of one or two pyridinioalkanoyl groups. The low reactivity of compound 47 toward NCp7 was suggested previously by its poor cross-linking of cell-free virions (12). However, compound 47 was active in initiating Gag precursor cross-linking (12). Since the structural context of the zinc fingers in NCp7 differs significantly from that of the zinc fingers in Pr55gag and Pr160 gag-pol, the observed differences in reactivity of the two compounds probably reflects the influence of local protein conformation. Further experiments are needed to define the interaction of compound 47 with either NCp7 or the Pr55gag precursor.
Based on the results presented, we propose the following mechanism for the action of compound 45. As depicted schematically in Fig. 6, the reaction between compound 45 and NCp7 is initiated by the formation of a loosely bound complex followed by a relatively slow nucleophilic attack from the sulfur atom of the "reactive" Cys 49 toward the carbonyl carbon of the activated PATE to form a new thioester bond in a transacylation reaction. Loss of cysteine coordination results in the release of zinc from the carboxyl-terminal finger facilitating a rapid attack by another molecule of compound 45 at Cys 36 . This process is followed by the slow structural collapse of NCp7, resulting in release of zinc from the amino-terminal finger, a process that is independent of concentration of compound 45.
This model includes several notable features. First, it is interesting to note that only two molecules of the antiviral compound are required to completely inactivate one molecule of NCp7 as opposed to three molecules of disulfide based inhibitors (15,20,21) due to the lack of reactivity toward the amino-terminal zinc finger. Second, both the need for activation and the initial formation of a noncovalent complex preceding transacylation are indicative of the lower reactivity of thioester compounds compared with disulfide agents. However, it should be noted that the NCp7 used in this study is a purified protein and is devoid of viral RNA normally associated with it. Under more native, in vivo conditions, the initial recognition step may be enhanced and additionally may result in activation of the PATE through well positioned amino acid side chains of NCp7 with consequent acceleration of the transacylation reaction without the requirement for an external activator.
Finally, a major concern with the utilization of NCp7 zinc fingers as anti-HIV targets has been the issue of selectivity of compounds for zinc fingers of target proteins. The PATEs were designed with this issue in mind, such that the compounds were selected for minimal chemical reactivity while maintaining anti-HIV activity (12). In this paper we have identified a PATE that reacts with only one of the two highly similar zinc finger domains of the NCp7 protein. These findings clearly illustrate that zinc fingers exhibit differential susceptibility to chemical entities. The exact susceptibility of various zinc fingers may be due to slight differences in chemical potentials of the sulfur atoms, their solvent accessibility, atomic surface geometric and hydrophilic/hydrophobic characteristics, and other factors (12,23). Nevertheless, the proof now exists for selective targeting of zinc fingers, and efforts must be undertaken to further exploit those differences.