Zinc Finger Structures in the Human Immunodeficiency Virus Type 1 Nucleocapsid Protein Facilitate Efficient Minus- and Plus-Strand Transfer

Materials.

RNA and DNA-RNA oligonucleotides were purchased from Oligos, Etc. (Wilsonville, Oreg.). DNA oligonucleotides for strand transfer assays were obtained from Lofstrand (Gaithersburg, Md.) or from Oligos, Etc. DNA oligonucleotides for site-directed mutagenesis were obtained from either Operon Technologies, Inc. (Alameda, Calif.) or Life Technologies (Rockville, Md.) An RNase H-minus HIV-1 RT having a point mutation changing the active site residue Glu478→Gln (63) was the generous gift of Stuart Le Grice (HIV Drug Resistance Program, NCI-Frederick Cancer Research and Development Center, Frederick, Md.). The MacVector program was purchased from the Oxford Molecular Company (Oxford, England).

Preparation of wild-type and SSHS NC proteins. (i) Wild-Type NC.

Recombinant wild-type NC (55-amino-acid form) was expressed and purified as described previously (78).

(ii) SSHS NC.

The Cys15→Ser Cys18→Ser Cys28→Ser Cys36→Ser Cys39→Ser Cys49→Ser (SSHS/SSHS) mutant, in the context of the pNL4-3 sequence (1; GenBank accession no. M19921) was prepared as follows. A subclone (pRB581, SSHC/SSHC) was prepared by ligating the 503-bp SpeI-ApaI fragment (containing the Cys15→Ser Cys18→Ser mutations) and the 419-bp ApaI-BclI fragment (containing the Cys36→Ser Cys39→Ser mutations) into the SpeI and SalI restriction sites of the pBluescript KS(+) phagemid vector (Stratagene, La Jolla, Calif.). The inserted fragments were derived from the Cys15→Ser Cys18→Ser and Cys36→Ser Cys39→Ser full-length proviral clones that were reported previously (34). pRB581 was mutated in two successive steps using the Quick Change site-directed mutagenesis system (Stratagene) changing Cys28→Ser and then Cys49→Ser. The Cys28→Ser mutation was performed using oligonucleotides 2549 (5′ACA TAG CCA AAA ATT CCA GGG CAC CTA G-3′; the 5′ end corresponds to nt 1988 in the sequence of pNL4-3 [1]) and Z3414D08 (5′-CTA GGT GCC CTG GAA TTT TTG GCT ATG T-3′; the 5′ end corresponds to nt 2015 of the complementary sequence of the pNL4-3 [1]). The Cys49→Ser mutation was introduced using oligonucleotides Z3414D06 (5′-CCA AAT GAA AGA TTC TAC TGA GAG ACA GG-3′; the 5′ end corresponds to nt 2052 in the sequence of the pNL4-3 [1]) and Z3414D07 (5′-CCT GTC TCT CAG TAG AAT CTT TCA TTT GG-3′; the 5′ end corresponds to nt 2080 of the complementary sequence of the pNL4-3 [1]). The resulting plasmid subclone, termed pDB589, containing the SSHS/SSHS mutation was used to generate the recombinant NC expression plasmid as follows. The gene coding for the mutant NC protein (55-amino-acid form) was amplified by PCR from pDB589 using the sense primer 4658-333 and the antisense primer 4658-356; the PCR product was cloned into the pET32a vector (Novagen, Madison, Wis.), and the NC protein was isolated and purified as described previously (10). All mutations were verified using the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kit and an ABI 373 automated sequencer with the upgraded Big Dye filter wheel (PE Applied Biosystems, Foster City, Calif.).

Construction of the full-length HIV-1 clone containing the SSHS/SSHS mutation in NC.

For preparation of the full length HIV-1 proviral clone containing coding regions for the SSHS/SSHS NC zinc fingers, pDB589 (see above) was cut with the restriction endonucleases SpeI and SalI. The 4,278-bp SpeI/SalI fragment containing the sequences coding for the SSHS/SSHS mutation in the zinc fingers was then subcloned into the homologous sites of pNL4-3 (1) to create the plasmid pDB653.

Analysis of HIV-1 virions containing the SSHS/SSHS mutation in the NC domain of Gag.

The pNL4-3 and pDB653 plasmid constructs were transfected into 293T cells as previously described (29, 32). Virus-containing supernatant fluids were clarified by low-speed centrifugation. Samples were taken for RT assays and activities were determined as described earlier (32). The procedures for Western and Northern blot analyses are given in reference 28. Infectivity was determined using either CD4-producing HeLa cells containing a long terminal repeat (LTR)–β-galactosidase construct (HCLZ) (30) or AA2-clone 5 cells (32). Viral DNA was isolated from AA2-clone 5 cells infected with the mutant and wild-type viruses as described earlier (3). PCR analysis of 2-LTR-circularized viral DNA species was performed as described previously (25, 32).

Analysis of complete primer removal.

The assay system models the step in which RNase H cleavage has already removed one base from the 3′ terminus of tRNA₃^Lys, leaving a 3′ rA attached to the 5′ end of minus-strand DNA (11) and measures formation of the plus-strand 80-nt transfer DNA product (77). Assay conditions, separation of the DNA products by gel electrophoresis, and PhosphorImager analysis of the gel data are described in detail by Wu et al. (77). Note that the chimeric DNA-RNA oligonucleotide and the 17-nt RNA were added in fivefold excess of the amount of (+) SSDNA to ensure complete annealing of (+) SSDNA to these two reactants (77); wild-type or SSHS NC was then added to reaction mixtures prior to addition of wild-type or RNase H-minus (63) RT (77). Reactions were initiated by addition of MgCl₂ and deoxyribonucleotide triphosphates (77). Both the plus- and minus-strand 80-bp DNA was synthesized; however, only the plus-strand was detected, since 5′ ³²P-labeled (+) SSDNA was used for the reaction.

Annealing reactions.

For plus-strand annealing, the reactants were 5′ ³²P-labeled (+) SSDNA (50 nt) and minus-strand acceptor DNA (48 nt). Annealing of the complementary 18-nt PBS sequences in each of the reactants was performed in the presence of NC, as described in reference 77 for annealing alone. Unannealed labeled (+) SSDNA and the labeled DNA duplex were separated by electrophoresis in 6% polyacrylamide gels, and the amounts of the unannealed and annealed DNAs were quantified by using a PhosphorImager (Molecular Dynamics) and ImageQuant software. Minus-strand annealing reaction mixtures containing 5′ ³²P-labeled (−) SSDNA (131 nt) and acceptor RNA (148 nt) were incubated with NC and analyzed according to procedures given in Guo et al. (37), except that 7.5% polyacrylamide gels rather than agarose gels were used to separate unannealed labeled (−) SSDNA and the labeled RNA-DNA hybrid.

Complete minus-strand transfer system.

The components present in reaction mixtures, incubation conditions, and analysis of DNA products are given in Guo et al. (36). Note that the short DNA oligonucleotide, which primes synthesis of (−) SSDNA in this system, was labeled with ³²P at its 5′ end, as described in reference 38.

SSHS mutant NC protein.

NC plays an essential role in HIV-1 minus-strand (2, 8, 17, 18, 20, 36, 42, 58, 61, 79) and plus-strand (5, 77) DNA transfer. To investigate the question of NC zinc finger function in these reactions, we constructed a clone expressing an NC protein in which the three cysteines in each zinc finger were changed to serine, i.e., the six S atoms in the zinc coordination sites in NC were replaced by six O atoms (see Materials and Methods). This is the smallest change that can be made in NC that will essentially destroy the metal ion binding sites, while still allowing retention of neighboring residues such as certain basic and aromatic amino acids, which are also important for HIV-1 NC nucleic acid binding (13, 15, 64, 74, 76). This mutation also prevents disulfide bond formation, which can occur when less than six cysteine residues are modified (12).

Effect of wild-type and SSHS NC proteins on complete removal of the tRNA₃^Lys primer from minus-strand donor DNA.

In previous work with a reconstituted HIV-1 plus-strand transfer system, we demonstrated that synthesis of the 80-bp strand transfer product is dependent on the removal of 9 nt from the 3′ end of tRNA₃^Lys: (i) initial RNase H cleavage of the 3′ rA and (ii) the removal of an additional 8 nt by secondary RNase H cleavage and/or the nucleic acid chaperone activity of NC (77). To investigate the effect of replacing the cysteine residues in the two zinc fingers of NC on complete primer removal, we used a model system that mimics the step in which RNase H has already cleaved the 3′ rA (references 11 and 77 and references therein) from the 18 nt sequence at the 3′ end of tRNA₃^Lys (Fig. 1A) (77). In this system, the readout is synthesis of the 80-bp transfer product: primer removal is followed by annealing of (+) SSDNA to acceptor DNA (Fig. 1A) and subsequent elongation of plus- and minus-strand DNA. As described in Materials and Methods, 5′ ³²P-labeled (+) SSDNA was annealed to the chimeric DNA-RNA oligonucleotide and the 17-nt RNA; wild-type or mutant NC was then added, followed by the addition of either wild-type or RNase H-minus (63) RT. The amount of labeled plus-strand 80-nt transfer product was quantified by PhosphorImager analysis of gel data. [Since only (+) SSDNA is labeled in this experiment, the minus-strand 80-nt transfer product is not detected.]

Figure 1B shows the results obtained from reactions with wild-type NC and mutant or wild-type RT (upper and lower panels, respectively). In accord with our earlier findings (77), the presence of wild-type NC made it possible to synthesize the strand transfer product in the absence of RNase H activity (upper panel). (In the absence of NC and RNase H activity, strand transfer did not occur [Fig. 1B, upper panel; references 5, 65, and 77]). However, if both wild-type NC and wild-type RT were added to the reaction, in effect mimicking in vivo conditions, the amount of strand transfer was stimulated by severalfold (lower panel).

A rather different picture emerged from analysis of reactions containing SSHS NC and mutant or wild-type RT (Fig. 1C, upper and lower panels, respectively). In the presence of mutant NC, the amount of transfer product made was independent of the ability of RT to catalyze RNase H cleavage. Moreover, the data in Fig. 1C were virtually identical to the results obtained under conditions where wild-type NC was present together with RNase H-minus RT (Fig. 1B, upper panel).

Taken together, the data of Fig. 1 suggest that to maximize complete removal of the tRNA primer, both the presence of the zinc finger structures of NC and the RNase H activity of RT are required.

Effect of wild-type and SSHS NC proteins on the annealing reactions in plus- and minus-strand transfer.

We have previously shown that NC stimulates annealing of the complementary PBS sequences in (+) SSDNA and minus-strand acceptor DNA during HIV-1 plus-strand transfer (77). In principle, it is possible that the effect of the zinc fingers observed in the primer removal experiment (Fig. 1) actually reflects an enhancing effect on annealing and not on primer removal per se. (In our system, elongation is not affected by NC [77].) It was therefore of interest to examine the effect of the SSHS mutation on this annealing reaction. The amount of 5′ ³²P-labeled (+) SSDNA hybridized to minus-strand acceptor DNA was determined as described in Materials and Methods; the NC concentration was 0.28 μM. Note that the short dashed line at the bottom of Fig. 2A-1 represents annealing in the absence of NC (77).

Both wild-type and mutant NCs had a marked stimulatory effect on the rate and extent of annealing, and the annealing curves were almost superimposable (Fig. 2A-1). Replotting the data in Fig. 2A-1 as semilogarithmic plots (Fig. 2A-2) indicated that the plots show curvature, as previously reported for this system (77) and as is typical for most annealing reactions (reference 21 and references therein). In this experiment, the half-life (t_1/2) values for the overall reaction were calculated to be 8.2 and 10 min for the reactions with wild-type and SSHS NC proteins, respectively. Calculation of the initial rates gave even closer t_½ values: 6.2 min for wild-type NC and 6.8 min for SSHS NC (data not shown). At a lower NC concentration (0.14 μM), the mutant NC was still as active as the wild-type protein (data not shown).

The data presented in Fig. 2A-1 and A-2 demonstrate that maintaining the CCHC motif in both zinc fingers of HIV-1 NC is not a requirement for promoting the annealing of the complementary PBS sequences in the plus-strand transfer intermediates. Moreover, these findings lead to the conclusion that the results in Fig. 1 reflect a direct effect of the zinc finger structures on the primer removal step.

A similar experiment was performed to determine the effect of the cysteine-to-serine substitution on the annealing reaction that occurs during HIV-1 minus-strand transfer, i.e., annealing of the complementary R sequences in (−) SSDNA and the 3′ end of viral RNA. (Note that in our constructs, R contains the entire TAR sequence [36].) The amount of 5′ ³²P-labeled (−) SSDNA hybridized to plus-strand acceptor RNA was determined using the conditions described in Materials and Methods. For comparison, data obtained for annealing in reactions without NC are represented in Fig. 2B-1 by a short dashed line (37).

In the presence of wild-type NC, annealing proceeded much faster than it did with SSHS NC: the wild-type reaction was essentially complete by ∼5 min, whereas, it took ∼40 min for SSHS NC-catalyzed annealing to reach a plateau level. However, the extent of annealing was approximately the same for both proteins (Fig. 2B-1). In this case, the semilogarithmic plots were linear (Fig. 2B-2; references 37 and 79), reflecting pseudo first-order kinetics (79). The t_½ values for the wild-type and SSHS reactions, calculated from the plots shown in Fig. 2B-2, were 1.1 and 8.9 min, respectively. This indicates that wild-type NC catalyzed annealing at a rate 8-fold greater than that observed in the reaction with SSHS NC.

Thus, in contrast to the data for plus-strand annealing (Fig. 2A-1 and A-2), the results of Fig. 2B-1 and B-2 demonstrate that the presence of the two CCHC motifs has a major impact on annealing in minus-strand transfer.

Requirement for zinc finger structures to inhibit self-priming in the complete minus-strand transfer system.

The observation that the zinc finger structures facilitate annealing in minus-strand transfer (Fig. 2B) led us to investigate whether zinc coordination plays a role in the ability of NC to reduce self-priming from (−) SSDNA (8, 36, 44). To approach this question, we tested the effects of wild-type and SSHS NC proteins on minus-strand transfer in the complete HIV-1 reconstituted system (36).

Figure 3A illustrates gel analysis of reactions containing increasing concentrations of wild-type and mutant NC proteins. In agreement with previous findings (36), inspection of the gel shows that synthesis of the strand transfer product (T) was greatly stimulated by wild-type NC (lanes 2 to 6) compared to the amount made in the absence of NC (lane 1). In addition, wild-type NC significantly reduced synthesis of SP DNAs. By contrast, SSHS NC promoted low levels of strand transfer and was unable to effectively inhibit self-priming (lanes 7 to 11).

Quantification of the gel data by PhosphorImager analysis confirmed these observations. Figure 3B shows that the presence of wild-type NC dramatically increased synthesis of the strand transfer product in a dose-dependent manner (∼32-fold with saturating levels of NC); synthesis reached a plateau at ∼1.6 μM NC (1.75 nt/NC). However, in the presence of SSHS NC, the increase in synthesis of the strand transfer product was ∼6-fold at the highest NC concentration tested (3.2 μM, 0.88 nt/NC). In the linear portion of the curves, e.g., at concentrations of 0.4 and 0.8 μM, stimulation of strand transfer was ∼8-fold greater with wild-type NC. The reactions were also analyzed for the synthesis of SP products (Fig. 3C). As expected (36), increasing concentrations of wild-type NC greatly reduced the amounts of SP DNAs that were made. The SSHS NC protein had almost no effect on self-priming and, in repeated experiments, only a slight reduction in synthesis of SP products was observed.

The results presented above demonstrate that zinc coordination is required for two crucial aspects of NC activity in minus-strand transfer: (i) to promote efficient annealing (Fig. 2B) and (ii) to inhibit self-priming (Fig. 3A and C). Thus, a mutant such as SSHS NC, lacking zinc finger structures, has only a small stimulatory effect on the overall minus-strand transfer reaction (Fig. 3A and B).

Analysis of HIV-1 virions with the SSHS mutation in both zinc fingers.

The results from the in vitro experiments described above suggested that virions with the SSHS/SSHS mutation in the NC domain of Gag would be defective. To investigate this possibility, virus particles were harvested and characterized in a number of different assays. Western blot analysis showed that the protein band patterns of the mutant virus were indistinguishable from those of the wild-type. In both cases, CA protein (p24^CA) was clearly visible, indicating that the viral protease was expressed and could catalyze normal cleavage of the Gag precursor (data not shown). Additional results are summarized in Table 1. The ratios of p24^CA to RT activity were similar for the mutant and wild-type viruses. The full-length genomic RNA content of the mutant, determined by Northern blot analysis, was <10% of the wild-type virus, a result comparable to findings we reported for similar mutants of HIV-1 (34). In addition, the mutant was replication defective in assays utilizing HCLZ (30) and AA2-clone 5 cells, which measure infectivity in single- and multiple-rounds of replication, respectively. Analysis of 2-LTR-circularized viral DNA species showed that the expected 170-bp PCR product (25, 32) was easily detected in DNA from cells infected with the wild-type virus but was not observed in DNA obtained from cells infected with the mutant.

These results demonstrate that SSHS/SSHS mutant virions are noninfectious and have defects in synthesis of 2-LTR circular viral DNA as well as in viral RNA packaging.