Introduction

Top Summary Introduction Materials & Methods Results Discussion References

In the last several years, considerable effort has been devoted to the study of L-nucleoside analogs as anti-HIV and anti-HBV agents. Among them, 3TC and its 5-fluoro analog and -L-2',3'-dideoxycytidine and its 5-fluoro derivative, -L-FddC, were demonstrated to have potent activity against both HIV and HBV (Belleau et al., 1989; Doong et al., 1991; Lin et al., 1994a, 1994b). Recently, a novel L-nucleoside analog with an unsaturated ribose residue, -L-2',3'-dideoxy-2',3'-didehydrocytidine and its 5-fluoro derivative, -L-Fd4C, have been shown to possess even more potent anti-HIV and anti-HBV activity in cell culture than the compounds mentioned above (Lin et al., 1996). It was suggested that the termination of viral DNA synthesis catalyzed by HIV RT or HBV DNA polymerase might be the major mechanism of the antiviral action. -L-Fd4C was also reported to be more toxic than 3TC or -L-FddC against the replication of various cell lines in culture (Lin et al., 1996; Faray et al., 1997). The underlying mechanism of the growth inhibitory activity by -L-Fd4C may be attributed to its capacity to inhibit nuclear DNA synthesis. It is noteworthy that -L-Fd4C is the first L-cytidine derivative with a 2',3'-unsaturated sugar moiety discovered to have potent activity against HIV and HBV in cell culture. The X-ray analysis of -D-2',3'-dideoxy-2',3'-didehydro-nucleosides and the interaction of their 5'-triphosphates with HIV RT and some human DNA polymerases have been studied (Dyatkina et al., 1987; St. Clair et al., 1987; Birnbaum et al., 1989; Gurskaya et al., 1991; Harte et al., 1991; Van Roey and Chu, 1992). No information on X-ray analysis or on the molecular interaction of -L-Fd4CTP with HIV RT and DNA polymerases has been reported previously.

In this work, we employed a DNA primer extension technique and sequencing gel analysis to evaluate the kinetic interactions of -L-Fd4CTP with HIV-1 RT and human DNA polymerases , , , and . In this article, we also describe a novel procedure for the synthesis of -L-Fd4CTP.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

dNTPs and ddCTP were purchased from Boehringer Mannheim (Indianapolis, IN). -L-FddC was a kind gift of the late Dr. T.S. Lin. -L-FddCTP was synthesized from -L-FddC in this laboratory. The synthesis of -L-Fd4CTP is described below. The M13mp19(+) strand DNA was isolated as described previously (Sambrook et al., 1989). The 22-base primer 5'-d(GTAAACGACGGCCGTGAATT-3') was synthesized on an Applied Biosynthesis 380A DNA synthesizer at the Yale Oligonucleotide Synthesis Facility. The primer oligonucleotide was labeled at the 5'-position with T4 polynucleotide kinase using 3000 Ci/mM [-³²P]ATP (Amersham, Arlington Heights, IL) annealed to M13mp19 phage DNA as previously described (Sambrook et al., 1989). The complex was purified on a Sephadex G-50 column, and used as a substrate for elongation reaction.

Human pol , pol , and pol  were purified from chronic lymphocytic leukemia cells obtained from patients by leukophoresis. The procedure for purification and the characterization of human DNA polymerases were previously described (Kukhanova et al., 1995). Pol  from human placenta was a kind gift of Dr. D. Mozzherin (Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia). HIV RT was a kind gift of Dr. K. Anderson (Pharmacology Department, Yale University).

Synthesis of -L-Fd4CTP. -L-Fd4C used for the synthesis of -L-Fd4CTP was prepared as described previously (Chen et al., 1997). To the amidine-protected -L-Fd4C (200 mg, 0.709 mmol) in 10 ml CH₂C1₂ and 3Å sieves was added 2-cyanoethyl diisopropyl chlorophosphoramidine (336 mg, 1.42 mmol) followed by diisopropylethylamine (336 mg, 2.8 mmol). The reaction mixture was stirred at room temperature for 2 hr Tributylammonium pyrophosphate (514 mg, 1.4 mmol) and 0.5 M tetrazole (2.9 ml, 1.4 mmol) were then added to the above mixture and stirred for 2 days. At this point m-chloroperbenzoic acid (612 mg, 3.5 mmol) in 5 ml CH₂C1₂ was added and the reaction was stirred for another 1 day. Finally, 2 M ammonia in methanol (0.37 ml, 2.36 mmol) was added to remove the protective groups. After removal of solvents, the crude product was dissolved in 0.2 ml deionized H₂O and applied to a 20 cm DEAE-cellulose column. The product was eluted with a linear gradient of 0 to 0.4 M triethylammonium bicarbonate in deionized H₂O. The fractions were collected and purified by ion exchange high performance liquid chromatography using a Whatman Partisil-SAX column to provide pure -L-Fd4CTP. [¹H]NMR analysis (300 MHz; CD₃OD) yielded these results:  7.96 (d, J = 9 Hz, 1H), 7.85 (m, 1H), 7.38 (m, 1H), 6.95 (s, 1H), 6.53 (d, J = 7.5 Hz, 1H), 5.89 (d, J = 7.5 Hz, 1H), 4.94 (m, 1H), 4.21 (m, 2H). Fast-atom bombardment-mass spectrometry was calculated for C₉H₁₃O₁₂N₃P₃F 467; found, 465.

Inhibition of HIV RT and human DNA polymerases by -L-Fd4CTP, -L-FddCTP, and ddCTP. The incubation mixture (10 µl) contained a buffer optimal for each enzyme as described previously (Kukhanova et al., 1995), 50 nM complex (M13mp19 phage DNA-22-mer oligonucleotide primer), 2 µM dCTP, 20 µM each of three other dNTPs, 0.5 µCi [-³²P]dCTP, different amounts of dCTP analogs, and 0.5-1 unit of enzyme. One unit of activity was defined as the amount of enzyme that catalyzed the incorporation of 1 nmol of dTMP into activated DNA per hour at 37° for human DNA polymerases, and as the amount of HIV RT that incorporated 1 nmol of dTMP into the poly(rA)oligo(dT) complex. The reaction mixtures were initiated with enzyme and incubated at 37°; the rate of incorporation of [-³²P]dCMP residues into DNA was linear with respect to time. After incubation, 8 µl of reaction mixture was spotted onto DE81 paper. The paper was washed with 0.5 M NH₄HCO₃, dried, and counted as described previously (Kukhanova et al., 1995).

DNA primer extension assays. The reaction mixture (8 µl) for conducting the single nucleotide incorporation into the standing start point of DNA contained the optimal buffer, 25 nM 5'-[³²P]-primer annealed to M13mp19 phage DNA as above, and different amounts of dCTP analogs as shown in Fig. 4. The reactions were initiated by the addition of enzyme, and the incubation time of all experiments was chosen so that the yield of the reaction products was linear as a function of time. The reactions were terminated by adding 5 µl of formamide stop solution, and the reaction products were analyzed by autoradiography after separation on 15% polyacrylamide sequencing gels as described previously (Kukhanova et al., 1995).

Chain-termination sequencing reactions. A modification of the dideoxy-chain termination sequencing procedure (Sanger et al., 1977) was used to assay -L-Fd4CTP for base-specific chain termination. Reaction mixtures (8 µl) contained buffer; 25 nM 5'-[³²P]-primer-template complex as above; 20 µM each dATP, dTTP, and dGTP; 1 µM dCTP; and 1 unit of HIV RT or 2 units of pol . Additionally, the reaction mixtures contained -L-Fd4CTP: 2 µM, 5 µM, and 20 µM, when reactions were catalyzed with pol  and 0.25 µM, 0.5 µM, and 2 µM when catalysis was provided by HIV RT. As a control, DNA sequence analysis was performed using ddNTPs as chain terminators. After incubation at 37° for 30 min, the reaction mixtures were chased for 30 min with an additional 2 µl of a solution containing 250 µM of all four natural dNTPs. The reaction products were analyzed by 12% polyacrylamide gel electrophoresis as described above.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Synthesis of -L-Fd4CTP. Because of the acid instability of the 2',3' double bond in -L-Fd4C, many conventional methods used for the 5'-phosphorylation have failed to provide the desired -L-Fd4CTP (Hoard and Ott, 1965). After many unsuccessful attempts, because of the base-mediated hydrolysis of the neutral tributylammonium (S-acyl-2-thioethyl)-bearing monophosphate prodrug of -L-Fd4C, we finally succeeded in the synthesis of -L-Fd4CTP using the one-pot procedure outlined in Fig. 1. In this case, 2-cyanoethyl diisopropyl chlorophosphoramidite was employed as the 5'-monophosphorylating agent (Beancage, 1993). [¹H]NMR characteristics of the product are described in Materials and Methods. The retention time of -L-Fd4CTP on an anion exchange high performance liquid chromatography column was 19.2 min, which is the same retention time as the [³H]-L-Fd4CTP obtained by enzymatic synthesis.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Outline of one-pot procedure used for successful synthesis of -L-Fd4CTP.

Inhibition of HIV RT and human DNA polymerases by -L-Fd4CTP, -L-FddCTP, and ddCTP. The structure of -L-Fd4CTP is shown in Fig. 2. The inhibition pattern of HIV RT and human DNA polymerases by -L-Fd4CTP was compared with that by ddCTP and -L-FddCTP with the use of the complex of a 22-mer oligonucleotide primer annealed to M13mp19 phage DNA (Fig. 3). -L-Fd4CTP was about 4-fold more potent as an inhibitor than ddCTP for HIV RT, and 8-10-fold more effective an inhibitor than -L-FddCTP. -L-Fd4CTP effectively inhibited pol , but in contrast to HIV RT, the efficiency of inhibition was about one order of magnitude lower than that of ddCTP. The inhibitory effect of -L-Fd4CTP on pol  was twice that of ddCTP. The concentrations of -L-Fd4CTP that inhibited the yield of products by 50% were 0.20 ± 0.05 µM, 1.8 ± 0.5 µM, and 4 ± 1 µM for HIV RT, pol , and pol , respectively. By comparison, the K_i values of 5'-triphosphate of 3TC have been reported as 1 µM, 0.01 µM, and 1.2 µM for HIV RT, pol , and pol , respectively (Chang et al., 1992). -L-Fd4CTP at a concentration as high as 40 µM had no inhibitory effect on pol , but inhibited pol  by 10-20% at 20 µM. The effects of ddCTP and -L-FddCTP on pol  was in accordance with those previously described (Copeland et al., 1992; Kukhanova et al., 1995).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Structure of -L-Fd4CTP.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Inhibition of M13mp19 phage DNA synthesis catalyzed by HIV RT or human pol , pol , or pol  with A, -L-Fd4CTP; B, -L-FddCTP; and C, ddCTP. The assay conditions are described in Materials and Methods. The incorporation of [-32P]dCTP into primer-template without inhibitors was taken as 100%. Points, average of three experiments, with duplicates performed in each experiment.

Kinetic parameters of the incorporation of -L-Fd4CTP into the standing start point of DNA chain. The -L-FddCTP and -L-Fd4CTP were evaluated for their incorporation into the standing start point of DNA by HIV RT and human DNA polymerases, in a system containing M13mp19 phage DNA annealed with 5'-[³²P]-22-mer oligonucleotide primer. Although a large body of experimental data exists on steady state kinetic constants of HIV RT and human DNA polymerases with respect to ddCTP and -L-FddCTP, these data were generated not only under different conditions, but more importantly, with different template-primers (St. Clair et al., 1987; Ono et al., 1989; Eriksson et al., 1995; Kukhanova et al., 1995; Wilson et al., 1996; Ueno and Mitsuya., 1997). Steady state kinetic constants are known to be greatly influenced by the template complex and salt concentration (Beard and Wilson, 1993). We therefore compared the K_m and V_max values of HIV RT and human DNA polymerases with respect to ddCTP, -L-FddCTP, and -L-Fd4CTP using the same template and a buffer optimal for each enzyme. Fig. 4 illustrates the dose dependence of the incorporation of -L-Fd4CTP into the standing start point of 5'-[³²P]-primer by HIV RT, pol , pol , pol , and pol . The bands corresponding to the 23-mer primer were scanned with a densitometer, and the K_m and relative V_max values were estimated in the concentration ranges shown in Fig. 4. Similar methods were applied for estimation of the K_m and V_max values of ddCTP and -L-FddCTP. The elongation of the primer with -L-Fd4CTP by HIV RT and DNA polymerases in the standing start point was saturated at 0.25 µM-0.5 µM for HIV RT, 1 µM-2 µM for pol , and 5 µM for pol . Table 1 shows the K_m and V_max values for all analogs, and the ratio of the V_max to K_m reflects the efficiency of the analogs' incorporation into the standing start point. The behavior of HIV RT and pol  in respect to all three analogs was similar. Both HIV RT and pol  were able to use all three analogs, but -L-Fd4CTP was the most effective substrate for HIV RT. Its incorporation efficiency was 4-fold higher than that of ddCTP, which had been the most potent nucleotide inhibitor of HIV RT discovered previously. The -L-Fd4CTP is also a good substrate for pol  and pol , but in contrast to HIV RT, its incorporation into DNA by pol  was 7-8-fold lower than that of ddCTP, and about 3-fold lower than by pol . Pol  could use -L-Fd4CTP as a substrate, but only at high concentration (>40 µM) (Fig. 4). As a consequence, a direct estimation of the K_m value could not be achieved for this enzyme. It should be mentioned that ddCTP or -L-FddCTP up to a concentration of 100 µM could not be incorporated into DNA by pol  (data not shown). These data excluded possible contamination of pol  samples with pol  or pol , and correlated with findings reported previously (Copeland et al., 1992). All steady state kinetic values presented here must be regarded as apparent because a saturating primer-template concentration was not achieved in these experiments. Another complicating factor in estimation of the K_m values for pol  was the 3'->5' exonuclease activity of this enzyme, which could degrade primers terminated by different analogs at different rates. To overcome these problems, we determined the relative efficiency of HIV RT or DNA polymerases in incorporating different dCTP analogs. Comparison was made mainly with ddCTP, which is the most effective inhibitor of HIV RT and pol . Inhibition of pol  subsequently leads to delayed cellular toxicity (Chen and Cheng, 1989).

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Gel analysis of the capacity of HIV RT, pol , pol , pol , and pol  to incorporate -L-Fd4CTP into the standing start point of 5'-[32P]-22 mer oligonucleotide primer annealed with M13mp19 phage DNA. Assays were performed as described in Materials and Methods. The concentrations of -L-Fd4CTP used for each enzyme are shown. Arrows indicate the incorporation of dCTP (0.5 µM). Bands under position of primer, 3'->5' exonuclease activity of pol  and pol .

View this table: [in this window] [in a new window]   TABLE 1 Apparent kinetic constants of interaction of ddCTP, -L-FddCTP, and -L-Fd4CTP with standing start point of primer-template Values and standard deviations are from one experiment repeated at least three times with very close results. The Km and Vmax were calculated as described under Materials and Methods.

Chain-termination of DNA synthesis by -L-Fd4CTP. The competition of -L-Fd4CTP with dCTP for incorporation into the growing DNA strand by HIV-1 or pol  was assayed using the modified dideoxy-chain termination sequencing procedure described previously (Sanger et al., 1977). The same primer-template as above was used in these experiments (Fig. 5). In the absence of -L-Fd4CTP and in the presence of all four natural dNTPs, the synthesis proceeded up to 70 nucleotides (Fig. 5A, lane 8). It should be noted that some natural stop sites at positions C28 and G38-40 were observed when the reaction mixture contained all four natural dNTPs. Fig. 5, lanes 1-4, shows the sequence of primer extension using ddNTPs as chain-terminators and pol  or HIV RT (Fig. 5, A and B). As the concentrations of -L-Fd4CTP were increased from 0.25 µM to 2 µM for HIV RT, and from 2 µM to 20 µM for pol , more and more DNA fragments terminated at C-sites were observed (C23, C27, C29, C34-36. C43-44). These observations indicate that -L-Fd4CMP residues were incorporated into the DNA chain in place of dCMP. It is noteworthy that the concentration of -L-Fd4CTP that produced the DNA sequencing pattern with HIV RT was 8-10-fold less than that observed with pol . This result shows that -L-Fd4CTP can compete with dCTP for incorporation when HIV RT or pol  are used as the enzymes. This observation is supported by the data presented in Table 1 and Figs. 3 and 4.

View larger version (83K): [in this window] [in a new window]   Fig. 5.   Gel analysis of the chain terminating sequencing reaction with -L-Fd4CTP with the use of A, pol , or B, HIV RT. The letters on both sides of the figure indicate the sequence of the growing DNA chain after the primer. A, lane 8, shows the DNA synthesis catalyzed by pol  without chain terminators. A and B, lanes 1-4, DNA sequence analysis using ddNTPs as chain terminators and A, pol , or B, HIV RT, respectively. A and B, lanes 5-7, DNA sequencing analysis using -L-Fd4CTP as the chain terminator at, in A, 2 µM, 5 µM, and 20 µM, respectively, and, in B, 0.25 µM, 0.5 µM, and 2 µM, respectively. Reaction conditions are described in Materials and Methods.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Considerable effort has been directed toward the search for a novel nucleoside structure among L-nucleoside analogs for use as an anti-HIV and anti-HBV agent. Recently, a novel 2',3'-unsaturated L-nucleoside derivative, -L-Fd4C, was found to be the most active agent against HBV, and also to exhibit significant anti-HIV activity in cell culture (Lin et al., 1996; Faray et al., 1997). The cell growth inhibitory activity of this compound against CEM cell lines was about the same as that of ddC and was 10-fold higher than that of -L-FddC. Its ability to inhibit mitochondrial DNA synthesis was much less than that of ddC (Lin et al., 1996). The mechanism of its antiviral and cell growth inhibition activities may be due to the effect of its 5'-triphosphate metabolite on viral and human DNA polymerases. Recently, the formation of 5'-triphosphate metabolites in cells treated with -L-Fd4C has been demonstrated (Dutschman et al., 1998). The efficiency of formation of the phosphorylated metabolites of -L-Fd4C was higher than that of -L-FddC and 3TC in cell culture. Whether the more potent anti-HIV action of -L-Fd4C relative to that of -L-FddC is caused by the different affinity of their 5'-triphosphate metabolites for HIV RT is not clear.

This work has demonstrated that -L-Fd4CTP is a more potent inhibitor of HIV RT than -L-FddCTP is. The inhibitory potency is correlated with the efficiency of -L-Fd4CTP and -L-FddCTP as a substrate for HIV RT (Fig. 3, Table 1). -L-Fd4CTP is a good substrate for pol  and pol , but it is a better substrate for HIV RT. The K_i value of -L-Fd4CTP has been reported as 5-fold less than that of 5'-triphosphate of 3TC (Chang et al., 1992). The high affinity of -L-Fd4CTP for HIV RT may be caused by a preferable recognition of the2',3'-dideoxy-2',3'-didehydro structure of the ribose residue compared with that of the 2',3'-dideoxy structure by HIV RT.

The X-ray analysis of -D-2',3'-dideoxy-2',3'-didehydro-nucleosides showed that the 2',3' double bond limits the conformational flexibility of the sugar ring. In contrast to the deoxyribose ring, the didehydrofurane ring is nearly planar with O(4') being slightly above the plane of the other four atoms (endo-configuration) (Birnbaum et al., 1989; Harte et al., 1991; Van Roey and Chu, 1992). It has been suggested that such a planar conformation of the ribose residue may represent a transition state of the substrates of HIV RT and some cellular kinases (Harte et al., 1991; Krayevsky and Watanabe, 1993). It has been shown that natural dNTPs in a complex with Escherichia coli DNA polymerase I undergo conformational rearrangement resulting in significant flattening of the deoxyribose residue (Ferrin and Mildvan, 1985, 1986). It is logical to predict that the conformation of the ribose residue in L- and D-enantiomers of unsaturated nucleosides is similar, and flattened conformation might mimic a transition state of -L-Fd4CTP during enzymatic reaction and facilitate its binding to HIV RT. This hypothesis is con-sistent with the observations made by others, that both -L- and -D-enantiomers of d4NTP were more effective inhibitors of HIV RT than were -L-ddNTP and -D-ddNTP, respectively (Ono et al., 1989; Wilson et al., 1996). Our results do not prove that a flattened conformation of the ribose residue of nucleosides is the only requirement for their potent anti-HIV activity, but our findings do show that a flattened conformation of the ribose residue has a significant effect on the interaction of their 5'-triphosphates with HIV RT.

A high activity of unsaturated nucleoside analogs in cell culture indicates also that these compounds are substrates for all steps of phosphorylation. Whatever the mechanism, these results may be important in the design of anti-HIV compounds. Although pol  showed behavior similar to that of HIV RT toward some dNTP analogs (Eriksson et al., 1995), -L-Fd4CTP was a much less potent substrate for pol  than for HIV RT. It should be noted that compared with ddC, both -L-FddC and -L-Fd4C are much poorer inhibitors of mitochondrial DNA synthesis in cell culture (Lin et al., 1996; Dutschman et al., submitted for publication). The lack of potent inhibition of mitochondrial DNA synthesis in cell culture by -L-Fd4C could be caused by either less efficiency of the interaction of its 5'-triphosphate metabolites with pol , or an insufficient amount of the 5'-triphosphate metabolite in the mitochondria. This question is under investigation. -L-Fd4CTP is not a substrate for pol , and is a very poor substrate for pol . Its incorporation into DNA was detected only at a high concentration of -L-Fd4CTP in the primer extension assay (Fig. 4). The ability of -L-Fd4CTP to be a substrate for pol  could account for its inhibition of cell growth. Other possible mechanisms of the growth inhibitory action of -L-Fd4C, unrelated to interaction of its 5'-triphosphate with pol , are under current investigation.