Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Considerable progress has been made in the treatment of patients infected with HIV, the causative agent of AIDS. Combination regimens that include potent reverse transcriptase and protease inhibitors are in clinical practice (Hammer et al., 1997). Nevertheless, therapy failure caused by the emergence of strains that are resistant to currently used drugs is common (Schinazi et al., 1996). Therefore, it is essential to find drugs targeted at alternative steps of the viral replication cycle.

The integrase of HIV is an attractive target for selective antiviral therapy because there is no known functional homolog in human cells (Thomas et al., 1996). The integration of retrotranscribed viral DNA into a host chromosome is an essential step in the replication cycle of retroviruses. The only viral enzyme required for efficient integration of retroviral cDNA is integrase (IN), a protein of 32 kDa encoded by the 3'-end region of the pol gene (Brown, 1997).

A vast series of compounds have been reported to inhibit the integrase activity in oligonucleotide assays; until recently, however, for none of these compounds has it been shown unambiguously that inhibition of the viral replication is caused by interference with the integration step (Neamati et al., 1997; Pommier et al., 1997). One promising compound is AR177 (Zintevir), a guanosine-quartet forming 17-mer composed of deoxyguanosine and thymidine (Rando et al., 1995). Zintevir is a potent inhibitor of HIV-1 integrase activity; nevertheless, we demonstrated that the compound owes its antiviral activity to an interaction with the viral envelope glycoprotein gp120 and not with integrase (Cherepanov et al., 1997; Esté et al., 1998). Recently, the antiviral effect of diketo acid inhibitors was described as being caused by the inhibition of the strand-transfer reaction of HIV-1 integrase (Hazuda et al., 2000).

Chicoric acid (CA; dicaffeoyltartaric acid) is a natural compound obtained from plants (Williams et al., 1996; Bonn, 1998). The chemical structure shows two carboxylic acid functionalities in the tartaric acid part of the molecule that are negatively charged at physiological pH and four weaker phenolic acid functions on the phenyl groups of the caffeic acid moieties (Fig. 1).

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Chemical structures of L-CA, D-CA, L-CATA, and D-CATA.

Dicaffeoylquinic acid (DCQA) derivatives and L-CA were shown to inhibit HIV-1 replication in cell culture at micromolar concentrations. This activity was attributed to the inhibition of integrase as observed in biochemical assays (Robinson et al., 1996b; McDougall et al., 1998), although it was suggested previously that 4,5-DCQA can inhibit HIV-replication in cell based systems, presumably by inhibition of the binding of gp120 to its cellular CD4 receptor (Mahmood et al., 1993). Molecular modeling suggested an interaction of DCQA with the core domain of HIV-1 integrase (Robinson et al., 1996a). Analogs of CA with improved activity against HIV integrase were synthesized (King et al., 1999). Inhibition of HIV integrase by DCQAs was shown to be irreversible (Zhu et al., 1999). An HIV-1 strain resistant to L-CA was selected and the observed resistance was attributed to a glycine-to-serine substitution at position 140 of the integrase (King and Robinson, 1998). CA was also found to improve the in vitro anti-HIV-1 effect of zidovudine and the protease inhibitor AG1350 (Robinson, 1998). We have recently confirmed the antiviral activity of both the D- and L-enantiomers of CA and their tetra-acetyl derivatives and their inhibition of integrase activity in an enzymatic assay (Lin et al., 1999).

Here we investigated in detail the mechanism underlying the antiviral effect of CA derivatives in cell culture. We have selected HIV-1 strains partially resistant to the drugs. Sequencing of these strains revealed mutations in the gene encoding gp120 but not in the integrase gene. Recombination of the gp120 gene of the L-CA and D-CA resistant strain in a NL4.3 wild-type strain fully rescued the phenotypic resistance toward L-CA and D-CA. These results demonstrate that the phenotypic resistance to L-CA and D-CA can be completely explained by mutations in gp120. We conclude that the primary antiviral target of CA and its tetra-acetate esters is the envelope glycoprotein gp120 of HIV.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Compounds. L-CA, D-CA, and their tetra-acetate esters (L-CATA and D-CATA) were synthesized according to the method described by Zhao and Burke (1998). Dextran sulfate (DS) (MW 5000) was purchased from Sigma (St. Louis, MO, USA). The bicyclam derivatives AMD3100 and AMD2763 were provided by AnorMED (Langley, BC, Canada) and were synthesized as described previously (Bridger et al., 1995). 3'-Azido-3'-deoxythymidine (AZT) was synthesized according to the method described by Horwitz et al. (1964). The oligonucleotide AR177 was provided by Aronex Pharmaceutical (Woodlands, TX). Ritonavir was kindly provided by Abbott Laboratories (Abbott Park, IL).

Viruses, Cells, Antiviral Activity Assays, and Cytotoxicity Assays. Anti-HIV activity and cytotoxicity measurements in MT-4 cells (Harada et al., 1985) were based on viability of cells that had been infected or not infected with HIV in the presence of various concentrations of the test compounds. After the MT-4 cells had been allowed to proliferate for 5 days, the number of viable cells was quantified by a tetrazolium-based colorimetric method (MT-4/MTT assay or 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium method) as described by Pauwels et al. (1988).

Selection of L-CA- and D-CA Resistant HIV-1 NL4.3. L-CA- and D-CA-resistant HIV-1 strains were obtained after sequential passaging of HIV-1 NL4.3 virus in the presence of increasing concentrations of L-CA or D-CA, respectively, in MT-4 cells. At the start of the selection, NL4.3 virus was inoculated on MT-4 cells in the presence of 5 µM L-CA or D-CA. When the cytopathic effect (CPE) was observed, cell culture supernatant was used as inoculum to infect new MT-4 cells at the same concentration of compound. The second time CPE was observed, the concentration of L-CA or D-CA was increased two times. After serial passaging, we were able to culture resistant virus in the presence of 63.2 µM L-CA and 84.3 µM D-CA, respectively.

Time of Addition Experiment. (Pauwels et al., 1990) MT-4 cells were infected with HIV-1 (III_B) at a multiplicity of infection of 0.5, and the test compounds were added at different times after infection. Viral p24 Ag production was determined at 31 h after infection by the p24 Ag enzyme-linked immunosorbent assay (NEN).

Virus Adsorption Assay. In this assay, the inhibitory effects of L- and D-CA, DS, and the bicyclam AMD3100 on virus adsorption to MT-4 cells were measured. We evaluated the binding capacity of NL4.3 wild-type strain in parallel with the L-CA- and D-CA-resistant NL4.3 strains. Therefore, MT-4 cells (5 × 10⁵ cells per tube) were incubated with the respective HIV strains (corresponding to 100 ng of p24) in the presence or absence of serial dilutions of the test compounds. After a 2-h incubation at 37°C, the cells were washed extensively with PBS to remove the unadsorbed virus particles. Then the cells were lysed with PBS containing 0.5% of tergitol Nonidet P-40 (Sigma, St. Louis, MO). The amount of p24 Ag was determined by the p24 Ag Elisa.

Genotypic Analysis of gp120 Encoding Sequences. MT-4 cells were infected with the HIV-1(NL4.3) strains in vitro selected in the presence of L-CA and D-CA. DNA extraction of proviral DNA was performed using the QIAamp blood Kit (Qiagen, Westburg, Leusden, The Netherlands). A 2105-nucleotide base pair fragment (codons 1 to 445) of gp120 was amplified in a nested polymerase chain reaction (PCR) using Expand High Fidelity PCR system (Boehringer Mannheim, Roche, Germany), which is composed of an enzyme mix containing thermo-stable Taq DNA and Pwo DNA polymerase with 3'-5' exonuclease proofreading capacity. The outer PCR reaction was performed on a Perkin Elmer Gene Amp PCR system 9600 and the inner PCR reaction on a Biometra Trioblock (Westburg) using the primers AV310 (5'-AGC AGG ACA TAA T/CAA GGT AGG-3' corresponding to position 5447-5467 of NL4.3) and AV311 (5' GGA GAA GTG AAT TAT ATA AG/AT ATA AAG TAG-3' corresponding to position 7630-7659 of NL4.3), followed by the primers AV312 (5'-AGA A/GGA C/TAG ATG GAA CAA GCC CCA G-3' corresponding to position 5549-5573 of NL4.3) and AV313 (5'-GAC CTG GAG GAG GAA/G ATA TGA G/AGG A-3' corresponding to position 7605-7629 of NL4.3). The outer cycling conditions were as follows: a first denaturation step of 3 min at 95°C followed by 40 cycles consisting of 45 sec at 95°C, 30 sec at 50°C, and 2 min at 72°C. A final extension was performed at 72°C for 10 min. For the inner cycling, the following conditions were used: after 3 min at 95°C, 30 cycles of 45 sec 95°C, 30 sec 58°C, 2 min 72°C and 10 min 72°C extension.

gp120-Recombination. MT-4 cells were subcultured at a density of 500,000 cells/ml on the day before transfection. Cells were pelleted and resuspended in phosphate-buffered saline at a concentration of 3.1 × 10⁶ cells/ml. For each transfection, 2.5 × 10⁶ cells (0.8 ml) were used. Transfections were performed by electroporation using an EASYJECT (Eurogentec, Seraing, Belgium). Cells were cotransfected with 10 µg of a linearized gp120-deleted NL4.3 clone (V. Fikkert, C. Pannecouque, P. Cherepanov, C. Van Laethem, E. De Clercq, A.M. Vandamme, and M. Witvrouw, in preparation.) and 2 µg of purified and concentrated AV312-AV313 inner PCR product (PCR Purification Kit; Qiagen, Westburg, The Netherlands). The electroporation conditions were 300 µF and 300 V. After 30-min incubation at room temperature, the transfected cell suspension in 5 ml of culture medium was incubated at 37°C in a humidified atmosphere with 5% CO₂. When full CPE was observed in the culture (about 8 days after transfection), cells were pelleted, and the supernatant containing the recombinant virus was stored in 1 ml aliquots at 80°C for subsequent infectivity and drug susceptibility determinations in the MT-4/MTT assay.

HIV-1 Integrase Assays. The recombinant enzyme preparation of wild-type His-tagged integrase the substrate and target DNA were as described previously (Cherepanov et al., 1997; Debyser et al., 2000). The following high-performance liquid chromatography-purified deoxynucleotides were purchased from Amersham-Pharmacia Biotech: INT1, 5'-TGTGGAAAATCTCTAGCAGT; INT2, 5'-ACTGCTAGAGATTTTCCACA; T35, 5'-ACTATACCAGACAATAATTGTCTGGCCTGTACCGT; and SK70, 5'-ACGGTACAGGCCAGACAATTATTGTCTGGTATAGT. The oligonucleotides INT1 and INT2 correspond to the U5 end of the HIV-1 long-terminal repeats. The 3'-processing, overall integration, and strand-transfer assays were slightly modified from published procedures. The final reaction mixture for the 3'-processing assays contained 20 mM HEPES, pH 7.5, 5 mM dithiothreitol, 10 mM MgCl₂, 75 mM NaCl, 5% (v/v) polyethylene glycol 8000, 15% (v/v) dimethyl sulfoxide, 30 nM concentrations of the oligonucleotide substrate, and 230 nM His-tag IN in a volume of 10 µl. Reactions were started by the addition of the enzyme. Inhibitors were incubated shortly with the reaction components before the addition of IN. Reactions were allowed to proceed at 37°C for 7 min and stopped by the addition of formamide dye. In the overall integration assay, the reaction is allowed to proceed for 60 min before the addition of formamide dye.

Creation of the G140S IN Mutant Using Site-Directed Mutagenesis. The G140S mutant was generated by using QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Two oligonucleotides, G140S+ (26-mer, 5'-TGTAGGGAATGCTAAATTCCTGCTTG-3'), and G140S (28-mer, 5'-GATCAAGCAGGAATTTAGCATTCCCTAC-3') embedding the 140 mutation were designed and purchased from Bioserve (Laurel, MD). PCR performed with these two primers and pET-15b-IN1-288/F185K/C280S plasmid (Jenkins et al., 1996) as starting template, produced the pET-15b-IN1-288/F185K/C280S/G140S plasmid containing the desired mutation. After transformation of Escherichia coli strain XL1-Blue by this plasmid, five colonies were sequenced and all of them contained the G140S mutation.

Expression and Purification of F185K/C280S/G140S Integrase from E. coli. pET-15b-IN1-288/F185K/C280S/G140S plasmid was expressed in E. coli strain BL21 as described previously (Jenkins et al., 1996) with the following modifications. Cells were grown in 1000 ml of LB medium (Digene, Beltsville, MD) containing 50 µg/ml of ampicillin until absorbance reached 0.8 at 600 nm. Protein expression was induced for 3 h by addition of 0.4 mM isopropyl--D-thiogalactopyranoside, and cells were harvested and resuspended in lysis buffer containing 25 mM HEPES, pH 7.5, 1 M NaCl, 5 mM imidazole, 2 mM -mercaptoethanol, and 0.3 mg/ml lysozyme. After 30 min on ice and sonication, lysed cells were centrifuged for 20 min at 30,000g and the supernatant was applied onto a nickel-Sepharose column. Integrase retained on the column was washed with a buffer containing 25 mM HEPES, pH 7.5, 0.5 M NaCl, 2 mM -mercaptoethanol, and an increasing imidazole concentration from 20 to 250 mM. The protein was then eluted with the same buffer containing 750 mM imidazole and dialysed overnight against 25 mM HEPES, pH 7.5, 1 M NaCl, 2 mM, 2 mM EDTA, 10 mM dithiothreitol, 2 mM -mercaptoethanol, 100 mM imidazole, and 10% glycerol.

IN Assays Using the F185K/C280S Integrase and its G140S Mutant. These methods were performed essentially as described previously (Mazumder et al., 1999).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Antiviral Activity Spectrum. The two enantiomers of CA (L-CA and D-CA) and their respective tetra-acetate esters (L-CATA and D-CATA) (Fig. 1) proved active against different strains of HIV-1 (III_B, NL4.3) and against HIV-2 (ROD) with EC₅₀ values ranging from 1.7 to 70.6 µM, whereas SIV (MAC251) was not sensitive to their inhibitory effects at subtoxic concentrations in MT-4 cells (Table 1). L-CA and D-CA were toxic for the cells at 115 and 111 µM, respectively, resulting in a selectivity index of 22 for L-CA and 65 for D-CA against the replication of HIV-1(III_B) in MT-4 cells. L- and D-CATA showed a selectivity index of 18 and 24, respectively. The CA derivatives showed reduced activity against HIV-1 strains that were resistant toward DS 5000, the G-quartet AR177, and the bicyclams AMD2763 or AMD3100 (Table 2). These compounds are known to interfere with viral entry and the resistance of viral strains against the inhibitory effects of these compounds is caused by mutations in gp120 (De Vreese et al., 1996; Esté et al., 1997; Esté et al., 1998). The bicyclams are specific CXCR4 antagonists and do not interact directly with gp120 (Schols et al., 1997). The cross-resistance observed with the CA derivatives against viruses known to be resistant to gp120-targeted compounds suggests an interaction of the CA analogs targeted at the envelope glycoprotein gp120.

Time (Site) of Intervention. A time-of-addition experiment was performed to pinpoint the possible step(s) in the replicative cycle of HIV-1 that is (are) inhibited by D- and L-CA. HIV-1 replication, as measured by p24 Ag production, could only be inhibited by DS if added at the time of infection. Addition of the nucleoside reverse transcriptase inhibitor AZT could be delayed up to 4 h after infection without loss of inhibitory activity. Addition of L- and D-CA at different concentrations could not be delayed, like DS, more than 1 h after infection as is shown in Fig. 2. These results indicate that CA and its analogs interact with an early stage (binding/fusion) of the viral replicative cycle.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Time-of-addition experiment. MT-4 cells were infected with HIV-1(IIIB) at a multiplicity of infection of 0.5, and the test compounds were added at different times after infection. Viral p24 Ag production was determined at 31 h after infection and is expressed as the log10 of the p24 Ag content in pg/ml. , control; , DS (20 µM); , AZT (1.9 µM); ×, Ritonavir (2.8 µM); , L-CA (50 µM).

Selection of HIV-1 (NL4.3) Mutant Strains. HIV-1 strains resistant to L- and D-CA were generated in MT-4 cells by passaging the virus in the presence of increasing concentrations of L- and D-CA, respectively. An HIV-1(NL4.3) selected strain, NL4.3 L-CA^res was able to grow in the presence of 63.2 µM L-CA after 23 passages. NL4.3 L-CA^res proved to be 5- to 10-fold resistant to L-CA (and its derivatives) and 26-fold resistant to DS (Table 3). NL4.3 D-CA^res was obtained after 26 passages in the presence of increasing concentrations of D-CA (final concentration was 84.3 µM). NL4.3 D-CA^res was 5-to 12-fold resistant to D-CA (and its derivatives) and 16-fold resistant to DS (Table 3).

Virus Adsorption Assay. To confirm that the anti-HIV activity of L-CA and D-CA is attributed to the inhibition of virus binding or fusion to the cells, a virus adsorption assay was performed. A known adsorption inhibitor DS and a fusion inhibitor AMD3100 were included as control specimens. Cells were infected with wild-type, L-CA- or D-CA-resistant NL4.3 strain. DS, L-CA, and D-CA inhibited the binding of wild-type virus to the cells with an IC₅₀ of 0.01 µM, 14.3 µM, and 2.1 µM, respectively. As expected, the CXCR4 antagonist AMD3100 (Schols et al., 1997) did not inhibit virus adsorption (Table 5). The inhibitors DS, L-CA, and D-CA lost their inhibitory effect on the cell binding if the L-CA- and D-CA-resistant NL4.3 strains were used. These results indicate that the mechanism of anti-HIV action of L-CA and D-CA is based on inhibition of adsorption. Compounds had no effect on the binding of mAb to CD4 or CXCR4 (data not shown).

gp120-Recombination. To verify that the mutations in gp120 were sufficient to explain the resistant phenotype, recombinant viruses were constructed that contain the gp120 gene of the strains selected in vitro for resistance against L-CA and D-CA. These recombinant strains are referred to as L-CA^resRV and D-CA^resRV. The resistant profiles of the recombinant viruses were similar to the in vitro selected strains: both L- and D-CA^resRV were less susceptible to inhibition by DS, L-CA, and D-CA compared with the NL4.3RV strain in which the wild-type gp120 gene is recombined (Table 6). No cross-resistance was observed toward the fusion inhibitor AMD3100.

Integrase Assays. Integrase catalyzes two subsequent reactions: 3'-processing and DNA strand transfer (Fig. 3A). It can also catalyze a reversal of the integration reaction, denoted the disintegration reaction (Fig. 3B). We have previously established enzymatic assays that can evaluate potential inhibition of both the 3-processing or strand-transfer reaction independently from each other (Cherepanov et al., 1997). All molecules tested inhibited the 3'-processing reaction of HIV-1 IN at concentrations that were in accordance with the reported values (Table 7). The compounds were inactive in the DNA strand-transfer assay (Table 7). The IC₅₀ values obtained for the overall integration reaction (3'-processing + DNA strand transfer) were similar to the values obtained for the 3'-processing reaction.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   HIV-1 integrase catalytic assays. A, a 21-mer blunt-end oligonucleotide corresponding to the U5 end of the HIV-1 proviral DNA, 5' end-labeled with 32P, reacts with purified HIV-1 integrase. The initial step involves nucleolytic cleavage of two bases from the 3' end, resulting in a 19-mer oligonucleotide. The second step, 3'-end joining or strand transfer, involves joining of this recessed 3' end to the 5' end of an integrase-induced break in another identical oligonucleotide, which serves as the target DNA. B, HIV-1 integrase disintegration assay. The substrate oligonucleotide mimics a strand transfer step product (i.e., a Y oligonucleotide containing a 15-mer oligonucleotide 5'-end-labeled with 32P). HIV-1 integrase-mediated disintegration generates a 30-mer oligonucleotide.

Sensitivity of the G140S IN Mutant toward the Compounds. An HIV strain resistant to L-CA was reported to contain the G140S substitution in the integrase gene (King and Robinson, 1998). We now have constructed and purified recombinant HIV-1 IN carrying this mutation. The enzymatic activity of the mutant enzyme was somewhat lower than that of the wild-type enzyme, especially with respect to the DNA strand transfer activity (data not shown). The wild-type integrase used in this assay carried F185K/C280S mutations to increase the solubility (Jenkins et al., 1996). The G140S mutation was introduced in this soluble mutant. All CA derivatives tested inhibited the 3'-processing activity and the disintegration activity of HIV-IN and the G140S mutant to the same extent (Table 8). The compounds also inhibited IN-DNA cross-linking within a similar range for both enzymes (data not shown). Taken together, these data show that the CAs without any discrimination can inhibit wild-type IN and the G140S mutant. This indicates that the G140S mutation in the IN gene described earlier cannot be hold responsible for the resistant phenotype of the selected virus.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown that the anti-HIV activity of CA and its derivatives (L-CA, D-CA, L-CATA, and D-CATA) in cell culture is caused by interference of the compounds with viral entry, much alike polyanionic molecules such as dextran sulfate (DS) and the G-quartet AR177 (Zintevir). The structure of CA contains several negatively charged residues, which is a point of similarity with other polyanionic compounds. However, CAs have only a limited number of negative charges in contrast with the other described polyanionic anti-HIV compounds.

We have shown that CA and CATA inhibit the virus replication of different HIV-1 (III_B, NL4.3) strains and HIV-2 (ROD) strain with EC₅₀ values ranging from 1.7 to 70.6 µM. The cytotoxicity (CC₅₀) for MT-4 cells is in the range of 74 to 115 µM. The loss of antiviral activity of CA and CATA against HIV-1 NL4.3 strains that are resistant to inhibitors of viral entry (DS, AR177, AMD2763, AMD3100) and that are known to contain mutations in the viral envelope gp120 suggested to us that the antiviral activity of CA may involve an interaction with gp120.

In a time-of-addition experiment, the addition of inhibitors can be postponed after infection as long as the replication step with which the inhibitor interferes has not taken place. Like DS, CA had to be present at the time of virus adsorption to inhibit HIV replication, confirming the interaction with an early process in the viral replication cycle. For a "true" integrase inhibitor, one should be able to postpone addition of the compound for more than 6 h after infection.

In a direct virus-cell binding assay, binding of wild-type HIV but not of L-CA- or D-CA-resistant virus was clearly inhibited by DS, L-CA, and D-CA, whereas the bicyclam AMD3100, a CXCR4 antagonist, was inactive in the binding assay. These data provide direct proof that the CA derivatives block virus adsorption.

To identify the molecular target(s) of CA, HIV strains resistant to the compounds were developed in cell culture. We have previously shown that HIV-1 strains that are resistant to inhibitors targeted at the binding/fusion step of replication can emerge after sequential passaging of the virions in cell culture in the presence of increasing concentrations of the compounds. The resistance to L-CA developed after 23 passages of HIV-1 NL4.3 in MT-4 cells in the presence of increasing concentrations of the compound. The resistant strain was able to replicate in the presence of L-CA at concentrations up to 63.2 µM. DNA sequence analysis of the gp120-encoding region showed the emergence of mutations in the L-CA-resistant strain that were not present in the wild-type strain. The mutation found in the integrase gene (C56Y) has been described as a polymorphism (Pommier et al., 1997). To definitively rule out the role of the integrase gene in the resistant phenotype, the gp120 gene of each resistant virus was recombined in a wild-type NL4.3 virus, deleted for gp120. The recombinant viruses were capable of infecting MT-4 cells. The resistance profile of the recombined virus toward the compounds L-CA, D-CA, DS, and AMD3100 is essentially the same as the resistant strain selected in vitro. It can thus be concluded that the resistance phenotype can be completely rescued by the recombination of the gp120 gene.

The genotypic analysis of the gp120 coding region reveals a typical mutation in the V3 loop, N295D, that has already been described for DS and SOF-1 (Esté et al., 1997; Schols et al., 1998). In addition, four novel mutations (S160N, T372N, T378N, and E379Q) were selected in the presence of L-CA, whereas in the presence of D-CA, the mutations S160N and E379K arose in gp120. Further investigations regarding the mechanism of interference by different compounds with gp120 will be described elsewhere (V. Fikkert, C. Pannecouque, P. Cherepanov, C. Van Laethem, E. De Clercq, A.M. Vandamme, and M. Witvrouw, in preparation.).

Theoretically, it cannot be excluded that further passage of CA-resistant HIV strains in further-increasing concentrations of L-CA might eventually lead to the emergence of mutations in the integrase gene. However, it should be pointed out that selection of resistance occurred at concentrations of L-CA that were already toxic for the host cell. Furthermore, the instability of the ester function linking the tartaric and the caffeic acid moieties confounds achieving high drug levels in cell culture. Synthesis of congeners with esters being replaced by amide functions could lead to more stable and potentially more active compounds.

The CA derivatives inhibited the enzymatic activities of both wild-type integrase and G140S mutant enzyme to the same extent. The G140S mutation was found after passaging an HIV strain in the presence of L-CA (King and Robinson, 1998). Our results question the role of this mutation in the observed resistance to CA. In light of our results, it would be interesting to sequence the gp120 glycoprotein of the reported (King and Robinson, 1998) virus strain. However, the G140S mutated integrase carries F185K/C280S mutations to increase solubility (Jenkins et al., 1996). It might be possible that these additional mutations affect the resistance of the G140S enzyme to CA.

The CA derivatives L-CA, L-CATA, and their enantiomers act primarily on the binding of viral gp120 with the cells. The precise mechanism of interaction warrants further investigation. In the case of integrase inhibitors, proof of inhibition of the integration step in cell culture is rather difficult to achieve (Hazuda et al., 2000), emphasizing the need for developing a good cellular integration system. The development of a human cell line stably expressing HIV-1 integrase at high expression levels will be helpful in the development of a cellular integration assay (Cherepanov et al., 2000). Assays with purified preintegration complexes that more closely resemble the actual integration step in the infected cell should also contribute to this end. In fact, L-CA was found to be inactive in an HIV-1 preintegration complex assay in a microtiter plate format (Hansen et al., 1999).

The observation of the inhibitory effect of CA on HIV entry does not necessarily compromise the future of L-CA and its analogs as anti-HIV-1 drugs. On the contrary, because of renewed interest in viral entry as a potential antiviral target and because binding and fusion of HIV to the cell are mediated in part by the gp120 molecule, these inhibitors of gp120 ought to be further developed. CA derivatives clearly belong to a new class of HIV entry inhibitors, as indicated by the emergence of novel resistance mutations in the viral gp120 glycoprotein on repeated exposure to the compound.