Differing Roles of the N- and C-terminal Zinc Fingers in Human Immunodeficiency Virus Nucleocapsid Protein-enhanced Nucleic Acid Annealing*

The replication process of human immunodeficiency virus requires a number of nucleic acid annealing steps facilitated by the hybridization and helix-destabilizing activities of human immunodeficiency virus nucleocapsid (NC) protein. NC contains two CCHC zinc finger motifs numbered 1 and 2 from the N terminus. The amino acids surrounding the CCHC residues differ between the two zinc fingers. Assays were preformed to investigate the activities of the fingers by determining the effect of mutant and wild-type proteins on annealing of 42-nucleotide RNA and DNA complements. The mutants 1.1 NC and 2.2 NC had duplications of the N- and C-terminal zinc fingers in positions 1 and 2. The mutant 2.1 NC had the native zinc fingers with their positions switched. Annealing assays were completed with unstructured and highly structured oligonucleotide com-plements. 2.2 NC had a near wild-type level of annealing of unstructured nucleic acids, whereas it was com-pletely unable to stimulate annealing of highly structured nucleic acids. In contrast, 1.1 NC was able to stimulate annealing of both unstructured and structured substrates, but to a lesser degree than the wild-type protein. Results suggest that finger 1 has a greater role in unfolding of strong secondary structures, whereas finger 2 serves an accessory role that leads to a further increase in the

Human immunodeficiency virus (HIV) 1 is a member of the family Retroviridae and contains a dimeric single-stranded RNA genome. Within the virion, the genome is coated by nucleocapsid (NC) protein (1,2). NC is a basic protein containing 55 amino acids with 15 arginine and lysine residues. This protein contains two CX 2 CX 4 HX 4 C (where X is a variable amino acid) zinc finger motifs that each coordinate a zinc ion (3)(4)(5)(6)(7)(8). All known orthoretroviruses contain one or two zinc fingers in their NC proteins (9). A small basic linker connects the zinc fingers of HIV-1 NC. NMR spectroscopy has shown that the N-and C-terminal tails, as well as the linker region, have a flexible structure (4, 10 -13). HIV-1 NC is encoded in the gag region of the genome and is a product of proteolytic processing of the Gag precursor protein (14).
HIV-1 NC most often binds to nucleic acids nonspecifically, with an occluded binding site of about seven nucleotides (15)(16)(17). This protein coats the genome within the virion, with the ability to protect the genome against nucleases (17,18). Although it binds the complete genome nonspecifically, in vitro studies have shown an increased affinity for TG and UG repeat sequences in DNA and RNA, respectively (19,20). Also, NMR studies have shown specific interactions between NC residues with stem-loops 2 and 3 of the viral packaging signal (⌿) (21)(22)(23)(24). Upon binding to these stem-loops, the N-terminal basic domain of NC forms a 3 10 -helix that interacts with each stem, whereas the C-terminal zinc finger associates with the loop region of the stem-loop (21,23).
HIV-1 NC involvement has been implicated in many processes in the viral life cycle. Recent studies have shown that NC may play a role in integration of the proviral DNA (18,(25)(26)(27). Also, NC sequences in the Gag precursor are a necessity for packaging of the RNA genome (28 -30) and the maturation of the genomic RNA dimer in the virion core (31,32). In vivo studies have specifically linked packaging activity to the zinc finger regions of the Gag (NC) precursor (33,34). In vitro studies have indicated that NC may have important roles in reverse transcription as well. HIV-1 NC has been shown to enhance annealing of the tRNA primer to the primer-binding site to initiate reverse transcription (35)(36)(37)(38); to stimulate minus, plus, and internal transfer (39 -45); and to have an effect on the processivity of reverse transcriptase in DNA synthesis (46 -48).
HIV-1 NC has also been shown to possess chaperone activity (1, 18, 49 -53), meaning it can aid in the unfolding of nucleic acid structures to enhance the annealing of more thermodynamically favorable structures (containing more base pairs) (54). This chaperone activity aids in tRNA/primer-binding site annealing, annealing of RNA and DNA in strand transfer, and genome dimerization and maturation. It is not clear if there are particular regions of NC that are responsible for chaperone activity or various components (helix destabilization (unwinding) or hybridization of complements) of this chaperone activity. A number of investigators have completed annealing assays with mutant NC proteins to determine what residues are necessary for NC chaperone capability. Studies examining the effect of Moloney murine leukemia virus (MMLV) NC mutants on the annealing of tRNA Pro to MMLV RNA, as well as the dimerization of the genomic RNA segments, showed that the basic regions of NC are necessary for enhancement (55,56). These studies indicated that the single zinc finger of MMLV NC could be removed and that chaperone activity was retained.
Similar experiments were completed with genomic RNA sequences and NC derived from HIV. RNA dimerization and tRNA 3 Lys binding were observed in the presence of HIV-1 NC mutants containing only amino acid sequences exterior to the zinc fingers (49,(57)(58)(59). Mutants with zinc fingers replaced by glycine-glycine linkers were able to enhance annealing, although the concentration of these proteins was often higher than used with wild-type NC. In vitro experiments have also been completed with ribozymes that will cleave RNA sequences upon annealing to them (60,61). Enhanced RNA cleavage was observed in the presence of wild-type NC as well as mutants containing the conserved basic residues with deleted zinc fingers (60). These reports indicate that the basic residues of NC are an absolute requirement for chaperone activity (56,59,60,62).
Other annealing experiments have also been completed with NC proteins that retained the zinc fingers, but that contained mutated residues within these regions. Guo et al. (63) completed annealing assays using sequences from the trans-activation response region of the HIV-1 genome. They used three types of proteins with mutations that alter either the structure of the zinc finger or the residues within the zinc finger that are not involved in zinc binding. Optimal annealing activity was observed only when the N-terminal zinc finger was not mutated. The effects of mutations to the C-terminal zinc finger were not as pronounced, although these results varied with the type of mutant constructed. The results are supported by experiments that tested the effect of altering the zinc finger structure on tRNA 3 Lys annealing to the primer-binding site (64). In this case, the zinc fingers were shown to be necessary for the annealing reaction. Also, annealing assays completed with the nucleic acid sequences from the regions of (Ϫ)-and (ϩ)-strand transfer in the presence of NC mutants unable to bind zinc showed that the requirement for the zinc finger structure differed depending on the nucleic acid sequence being used (65). Last, single DNA molecule stretching experiments have been completed with NC to examine its ability to destabilize the helix and to aid in transition to a coil structure. HIV-1 NC mutants were used in these experiments, and it was observed that the N-terminal finger of HIV-1 NC must be in the native position for optimal chaperone activity (52,66).
The study presented here investigates how NC-enhanced nucleic acid annealing differs for unstructured and structured nucleic acids. NC was shown to increase the annealing of both types of nucleic acids, indicating that it stimulates both the unfolding of structured nucleic acids and the direct hybridization of complements. NC zinc finger mutants were also used in annealing assays, which provided additional insight into the potential role of zinc fingers in annealing. The N-terminal zinc finger was shown to be necessary for the annealing of sequences with a high degree of secondary structure, whereas the C-terminal zinc finger was shown to be required for maximal annealing activity. Overall, the results suggest that the Nterminal finger is more important in unwinding secondary structures, whereas the C-terminal finger plays an accessory role.

EXPERIMENTAL PROCEDURES
Materials-RNA oligonucleotides were purchased from Dharmicon Research, Inc. DNA oligonucleotides were purchased from Integrated DNA Technologies and Invitrogen. T4 polynucleotide kinase, Klenow polymerase, and T4 RNA ligase were obtained from New England Biolabs Inc. SP6 polymerase, alkaline phosphatase, NTPs, dNTPs, and RNase-free DNase I were obtained from Roche Applied Science. Radiolabeled compounds were purchased from PerkinElmer Life Sciences. All other chemicals were from Fisher or Sigma.
Preparation of Wild-type and Mutant NC Proteins-HIV-1 NC from the HIV-1 AIDS-associated retrovirus strain was prepared as described (15). Wild-type and mutant NC proteins from the HIV-1 NL4-3 strain were prepared as explained previously (26). There are four differing amino acids between the NC proteins derived from the two HIV strains. NC from the AIDS-associated retrovirus strain contains two arginine residues in the C-terminal zinc finger that are substituted with lysine residues in NC from the NL4-3 strain. Aliquots of HIV-1 NC were prepared and stored in 50 mM Tris-HCl (pH 7.5), 10% glycerol, and 5 mM 2-mercaptoethanol at Ϫ80°C. Fresh aliquots were used for each experiment.
Preparation of Oligonucleotides-RNA oligonucleotide 0.0rna was purchased from Dharmicon Research, Inc., and DNA oligonucleotide 0.0dna was from Invitrogen. The most highly structured pair of oligonucleotides, 21.7rna and 21.7dna, were purchased from Integrated DNA Technologies. The RNA contained a 5Ј-fluorescein-6-carboxamidohexyl (FAM) end label, whereas the DNA contained a 4-[[(4-dimethylamino)phenyl]-azo]benzenesulfonicamino (DABCYL) group. Oligonucleotides 7.5rna and 16.3rna were transcribed from DNA oligonucleotide pairs. One DNA strand of each pair was 61 nucleotides long and contained the sequence for the SP6 promoter at the 5Ј-end, followed by the DNA sequence corresponding to the desired RNA. Twenty pmol of the 61-mer was combined with 40 pmol of a complementary 42-mer DNA in hybridization buffer (50 mM Tris-HCl (pH 8.0), 1 mM dithiothreitol, and 80 mM KCl). The hybrid reaction was heated to 65°C for 5 min and then cooled slowly to room temperature. The 5Ј-overhang was filled in using Klenow polymerase at 37°C. The hybrid was incubated for 1 h with Klenow polymerase and 200 M dNTPs. The double-stranded DNA product was extracted with phenol/chloroform/ isoamyl alcohol (25:24:1) and precipitated with ethanol. SP6 polymerase was then used to transcribe the 42-nucleotide RNA product. DNA was digested with DNase I for 10 min at 37°C. The reaction was stopped by adding an equal volume of 2ϫ formamide dye. The RNA was heated to 90°C for 3 min and then subjected to gel electrophoresis on an RNase-free 10% denaturing polyacrylamide gel. RNA was excised, eluted in formamide elution buffer (80% formamide, 400 mM NaCl, 1 mM EDTA, and 40 mM Tris-HCl (pH 7.0)), and precipitated with ethanol. The 5Ј-end of the RNA was dephosphorylated using calf intestinal alkaline phosphatase. Seventy-five pmol of RNA was incubated with calf intestinal alkaline phosphatase for 1 h at 37°C. The dephosphorylated RNA was extracted with phenol/chloroform/isoamyl alcohol (25: 24:1) and precipitated with ethanol. All of the aforementioned oligonucleotides were resuspended in 30 l of water and quantitated spectrophotometrically.
RNA End Labeling-Oligonucleotides 0.0rna, 7.5rna, and 16.3rna were labeled at the 5Ј-end with [␥-32 P]ATP. Fifty pmol of dephosphorylated RNA was end-labeled using T4 polynucleotide kinase. The 5Јend of the most highly structured sequence, 21.7rna, was connected to a FAM molecule. Therefore, this RNA was 3Ј-end-labeled with [5Ј-32 P]cytidine 3Ј,5Ј-bisphosphate as described. 2 All labeled RNAs were purified on a 10% polyacrylamide gel, eluted, and precipitated as described above. RNAs were resuspended in 70 l of TE buffer (10 mM Tris-HCl (pH 8) and 1 mM EDTA (pH 8)) and quantitated spectrophotometrically.
RNA/DNA Annealing Assay-32 P-5Ј-end-labeled RNA and the complementary DNA were diluted in TE buffer, separately heated to 90°C for 3 min, and then transferred to ice for 5 min. The RNA (5 nM final oligonucleotide concentration and 0.21 M nucleotide concentration) was then preincubated at 37°C in the presence or absence of mutant or wild-type HIV-1 NC (2 M final concentration) for 2 min. Complementary DNA (10 nM final concentration and 0.42 M nucleotide concentration) was separately preincubated at 37°C for 2 min in the presence or absence of HIV-1 NC (2 M final concentration). To start the reactions, 17 l of DNA solution was added to 90 l of reaction mixture containing the RNA. Final reaction concentrations were 50 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 1 mM dithiothreitol, 6 mM MgCl 2 , 80 mM KCl, and 100 M ZnCl 2 . Aliquots of 15 l were removed at specific time points (as indicated) and added to 7.5 l of stop solution (0.25% bromphenol blue, 20% glycerol, 20 mM EDTA (pH 8), 0.2% SDS, and 0.4 mg/ml yeast tRNA) (51). All reactions were incubated in stop solution at 37°C for 1 min before being transferred to ice. Reactions were then subjected to electrophoresis on 12 or 15% native polyacrylamide gels. Gels were dried and subjected to autoradiography (67) or phosphoimager analysis using a Bio-Rad GS-525 phosphoimager. Percent annealing was determined by dividing the amount of annealed product (A) in each lane by the total RNA (annealed and single-stranded (S)) in each lane and multiplying by 100 Annealing Detected by Fluorescence Resonance Energy Transfer-As was previously noted, 21.7rna was purchased with a FAM molecule on its 5Ј-end. The complementary DNA was purchased with a 3Ј-DABCYL group. Annealing assays were completed at 25°C using a Fluoromax-2 spectrofluorometer (Jobin Yvon Instruments S. A., Inc.). The RNA and DNA were separately incubated in the presence or absence of wild-type NC, 1.1 NC, or 2.2 NC. The reactions were started by mixing 10.5 l of the DNA/NC solution and 59.5 l of the RNA/NC solution. The final concentrations of the RNA and DNA were 5 and 10 nM, respectively (0.63 M total nucleotide concentration). NC was used at a final concentration of 2 M, and the final concentrations of the reagents in the buffer were as follows: 50 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 1 mM dithiothreitol, 6 mM MgCl 2 , 80 mM KCl, and 100 M ZnCl 2. The excitation wavelength was 494 nm with a bandwidth of 1 nm. The emission bandwidth was 5 nm, and the spectrum was observed from 508 to 570 nm. The emission spectrum was taken at 0, 1, 2, 4, 8, and 16 min. The intensity of the emission peaked at 517 nm. The intensity ratio (I r ) was determined by dividing the peak intensity at a given time (I t ) by the peak intensity at time 0 (I 0 ) (I r ϭ I t /I 0 ). Three experiments were completed for each type of NC, and an average intensity ratio was determined and plotted versus time.

RESULTS
Structure of Substrate RNAs-The RNAs used in these experiments each contained 42 nucleotides. The structures are shown in Fig. 1. These structures were determined using both RNAdraw and mfold (68,69). These structures were confirmed using RNase A and RNase T1 mapping (data not shown). The Gibbs free energy for unfolding is shown next to the structure. Nucleic acids were designed such that they would contain the same number of nucleotides, but have an increasing strength of secondary structures. Structures were chosen that did not have a high degree of GU (or GT in the complementary DNA) repeats due to reports indicating that NC may preferentially bind to these sequences (19,20).
HIV-1 NC-enhanced Annealing of Structured and Unstructured RNAs-Annealing assays were completed to examine how HIV-1 NC affects the annealing of unstructured and struc-tured nucleic acids. Fig. 2 shows the sequence of HIV-1 NC. The 5Ј-end-labeled RNA and its complementary DNA were separately preincubated at 37°C in the presence or absence of NC for 2 min as indicated under "Experimental Procedures." The DNA was then added to the RNA reaction mixture, and time courses were completed. The final concentration of DNA was twice that of RNA. Time courses for oligonucleotides 0.0rna and 7.5rna were run for 4 min. The time was extended to 16 and 32 min for 16.3rna and 21.7rna, respectively. The annealing reactions were subjected to gel electrophoresis, which allowed for the determination of the percent of the RNA that annealed to the complementary DNA. Annealed RNA shifted up in the autoradiogram because the double-stranded nucleic acid migrated slower through the native gel. A reaction containing just the RNA, DNA, and hybridization buffer was heated and slowly cooled and used as a hybrid control, which showed where the annealed RNA and DNA ran in the gel (data not shown). The 2 M concentration of NC used stimulated annealing to the maximal level, as more NC led to no further stimulation (data not shown). Fig. 3 shows autoradiograms of experiments with each of the RNA substrates. It is evident that with each substrate, HIV-1 NC did increase the annealing of RNA to DNA relative to reactions without NC. Note also that for the more strongly structured substrates (16.3rna and 21.7rna), a band was observed above the RNA/DNA hybrid species (marked as Dimer on Fig. 3). This species migrating at this position was observed most prominently in assays with RNA alone and likely represents dimerized RNA. Although less thermodynamically stable than hybrids (⌬G ϭ Ϫ34.7 versus Ϫ57.8 for the 21.7rna dimer and hybrid, respectively) (70), dimers are still quite stable and would be expected to form. The results from experiments with all the substrates were quantified using a phosphoimager, and the data were displayed graphically. The graphs shown in Fig.   FIG. 1. Structures of RNAs used in annealing assays. RNAs were designed to contain the same number of nucleic acids (42), although the sequence composition differed. Folding and free energy calculations were done using RNAdraw and mfold using default conditions and 37°C (68,69). Free energies for unfolding are given in kilocalories/mol. Each structure was named based on the free energies calculated from the program. A, 0.0rna does not have any predicted base pairs. B, 7.5rna contains 14 A-U base pairs and no G-C pairs. C, G-C content is increased in 16.3rna, which has 4 G-C base pairs and 11 A-U base pairs. D, the strongest structure, 21.7rna, contains 15 base pairs, with 7 G-C pairs and 8 A-U pairs. Concentric circles denote the 5Ј-end. 4 display the percent of RNA annealed in the presence or absence of HIV-1 NC for each substrate. Annealing at time 0 was only detected for 0.0rna and 7.5rna. This was probably due to the ease with which these substrates are able to anneal either in the sample buffer or during electrophoresis. The annealing reaction for the 0.0rna structure is shown in Fig. 4A. There was a clear increase in the RNA/DNA annealing even without NC. Over 4 min, annealing increased to 65% (hybrid/ (single-stranded ϩ hybrid) ϫ 100). However, when NC was added, the rate of annealing was accelerated considerably. Both the rate of annealing and the amount of annealed product at 4 min were enhanced. These results demonstrate a role for NC in annealing that is separate from a helix-destabilizing (unwinding) activity since this substrate presumably has no helices that require unwinding.
Shown in Fig. 4B is the reaction with 7.5rna. In this case, stimulation by NC was more evident than with 0.0rna, as reactions in the absence of NC were slower with this substrate. Comparing panel B with panels C and D, it is evident that annealing in the absence of NC was much greater for 7.5rna than for the stronger structures. In reactions with 7.5rna without NC, there was a considerable amount of annealing that occurred even over the shorter times used, indicating that this weak stem-loop does not need substantial unwinding activity to hybridize to the complement. Fig. 4 (C and D) shows the graphical representation of the strand annealing for the strongest structures. For each of these structures, a substantial amount of annealing did not occur in the absence of HIV-1 NC. Annealing assays for 16.3rna and 21.7rna were run over longer periods of time due to slower rates of annealing. In each case, the structure prevented high levels of annealing in the absence of NC. However, the addition of NC allowed the annealing of complementary RNA and DNA, presumably because NC aided in the unfolding of secondary structure. The rate of annealing was much less than was observed with the weaker structures, suggesting that the ratelimiting step for these substrates was destabilization of intraor intermolecular interactions. Although NC was clearly able to stimulate this step, it was still the slow step in the reaction for the highly structured substrates.
Mutant NC Proteins Display Differing Effects on Annealing, Depending on the Type of Nucleic Acid Structure-We next wanted to investigate what roles the zinc fingers of NC have in nucleic acid annealing. Strand annealing assays were completed with three mutant NC proteins and wild-type NC derived from the NL4-3 strain of HIV. Wild-type NC from the NL4-3 strain had similar annealing activity compared with wild-type NC from the AIDS-associated retrovirus strain (compare Figs. 4 and 6). As was previously noted, the N-and C-terminal zinc fingers of NC are not biologically equivalent (28). It was therefore possible that the fingers may have differential activities with respect to the hybridization (stimulation of hybrid formation in the absence of structure) versus unwinding (helix destabilization) activities of NC. Three NC mutants, 1.1 NC, 2.2 NC and 2.1 NC, were analyzed. The first mutant, 1.1 NC, contains two N-terminal zinc fingers, the first in its native position and the second in the position of the native C-terminal zinc finger. Two copies of the C-terminal zinc finger are present in 2.2 NC, and 2.1 NC is a finger switch mutant in which the positions of the zinc fingers are switched. Annealing assays were completed for each structure in the presence of the mutant NC proteins. An autoradiogram for 0.0rna and 16.3rna is shown in Fig. 5. It is evident that the mutants had differential effects on the annealing of the different types of structures. Graphical representations of annealing assays with mutants for each substrate are shown in Fig. 6.
Results of annealing assays completed with 0.0rna showed that all mutants enhanced annealing of this substrate. Because of the relatively high background with this structure, it is difficult to evaluate potential quantitative differences between the mutant and wild-type NC proteins with this substrate. All of the mutants appeared to display annealing activity comparable to that of the wild-type protein. Interestingly, Fig. 5 (upper panels) shows that the annealed products with 1.1 NC did not form a discrete band, but instead shifted up to form two smeared bands. These products were both quantitated as an- nealed product. These results indicate that all mutants contain hybridization activity that enhances the annealing of unstructured sequences to a similar extent.
In contrast to the results with 0.0rna, the mutants showed differential activity with the 7.5rna substrate. The annealing assays completed in the presence and absence of HIV-1 NC with 7.5rna showed that hybridization will occur without NC, but at a much reduced rate compared with reactions with wild-type NC (Fig. 6B). Reactions completed in the presence of each mutant showed an increase in the rate of annealing compared with reactions in the absence of NC. However, 2.2 NC had a much lower ability to enhance the annealing with 7.5rna compared with 0.0rna. With this substrate, both 1.1 NC and 2.1 NC were considerably better than 2.2 NC, whereas they were only slightly below wild-type NC. The results suggest that the first zinc finger may be more important in unwinding nucleic acid strands, even for relatively weak stem-loop structures. This point was further investigated using the stronger structures as described below.
Consistent with the trend observed with 7.5rna, the ability of 2.2 NC to enhance hybrid formation decreased with an increase in the strength of the nucleic acid structure. This mutant stimulated hybridization to the least extent with each of the structured RNAs. In fact, 2.2 NC showed essentially no stimulation with 21.7rna. In contrast, 1.1 NC and 2.1 NC retained some activity with all the structures, with 1.1 NC in general increasing the annealing rate slightly more than 2.1 NC. The results further support the hypothesis that finger 1 is more important for unwinding. Also, the results with the switch mutant 2.1 NC indicate that the location of the fingers is important. If this were not the case, then 2.1 NC should have had wild-type levels of activity since it possesses both fingers. This mutant did seem to have wild-type levels of annealing activity as judged from assays with 0.0rna, but unwinding was reduced.
RNA/DNA Annealing Detected by Fluorescence Resonance Energy Transfer-To confirm the above results by a second technique, fluorescence quench experiments were also completed using 21.7rna and 21.7dna. The RNA was synthesized with a FAM molecule on its 5Ј-end. This molecule fluoresces when excited by light at 494 nm. However, if this molecule comes within close proximity to a DABCYL molecule, which is a dark quencher, the fluorescence will be quenched. 21.7dna in these assays contained a 3Ј-DABCYL molecule. Annealing of the RNA and DNA would therefore lead to quenching of the FAM fluorescence. Fluorescence decay experiments were completed in the absence of NC and in the presence of wild-type NC, 1.1 NC, and 2.2 NC. A graph showing fluorescence decay is shown in Fig. 7. Wild-type NC showed a large and relatively rapid decrease in fluorescence, whereas no major decrease in intensity was shown without NC. This shows that the addition of NC increases the amount of FAM molecules that are quenched by coming into close proximity to the DABCYL molecule, more specifically by the annealing of the RNA and DNA. The mutants used, 1.1 NC and 2.2 NC, showed very different intensity ratio profiles. The intensity ratio with 2.2 NC was similar to that observed without NC, whereas reactions with 1.1 NC decreased less than wild-type NC, but considerably more than reactions with 2.2 NC. These data clearly support the findings from experiments performed in the gel-based annealing assay. DISCUSSION HIV-1 NC has been shown to contain chaperone activity that can aid in the many nucleic acid annealing steps that occur during the viral life cycle. This protein contains two zinc fingers, which have been studied extensively. The CCHC motifs cannot be substituted for one another in strand transfer assays that mimic retrovirus recombination, and mutants with alterations to the zinc fingers have been shown to reduce or eliminate viral infectivity (28,63). However, the role of the individual zinc fingers in nucleic acid annealing is unclear. In this report, we show that the hybridization and unwinding components of NC annealing activity are proportioned unequally in the two fingers, with the latter being more prevalent in finger 1. Finger 2, although apparently possessing little unwinding activity, clearly enhances the overall activity of NC since both fingers in the appropriate context are required for full activity (Fig. 6).
The annealing assays presented here with wild-type HIV-1 NC showed that NC enhanced the formation of hybrids between complementary RNAs and DNAs that contained varying degrees of secondary structure. Structured sequences require unfolding to be available for annealing to their complements. In the presence of HIV-1 NC, the rate of annealing was dramatically increased for all the substrates that were tested. NC enhanced even the 0.0rna substrate, which had no predicted structure (Figs. 3 and 4). This result strongly supports a role for NC in enhancing hybridization of nucleic acids irrespective of secondary structure. NC also stimulated annealing using highly structured substrates (Figs. 3 and 4). With 16.3rna and 21.7rna, destabilization of the strong secondary structures limited the rate of annealing, and NC unwinding activity was required for faster annealing. This suggests that NC possesses two related activities, hybridization and helix-destabilizing activities. It is possible that there is only one activity that stimulates hybridization and then unwinds secondary structure through strand invasion (71). However, others have clearly shown that NC destabilizes secondary structure in tRNAs and other nucleic acids in the absence of a complementary strand (37,72). This provides clear evidence for unwinding activity, whereas the enhanced hybridization in the absence of structure shown here supports a separate annealing activity. In addition, in a recent study that used fluorescence quenching experiments to examine strand exchange of short oligonucleotides, it was shown that HIV-1 NC highly enhanced the rate of annealing of complementary strands while moderately enhancing unwinding of the helix DNA (53). Again, this argues for two related but distinguishable activities.
As discussed in the Introduction, there have been many studies in which NC mutants were used to determine which amino acid residues are important for nucleic acid annealing. Experiments have shown that the binding of tRNA 3 Lys to the primer-binding site and the dimerization of sequences containing the ⌿ site require only peptides outside of the zinc fingers, including the basic backbone residues (55,59). It is known that the basic residues of NC are vital for RNA binding in vitro (73). Therefore, the decreased amount of annealing in the presence of mutants lacking basic backbone residues could be primarily due to a lack of NC binding to the RNA. Lapadat-Tapolsky et al. (49) completed annealing experiments with oligonucleotides containing sequences from the R region of the HIV-1 genome. Their report indicated that it is the backbone residues that are important for annealing rather than the zinc fingers. In agreement with this, binding assays completed with NC mutants containing only the zinc finger residues showed a highly reduced binding affinity compared with the wild-type protein (53). Therefore, there is a clear role for the basic amino acids in the backbone of NC in annealing; however, these reports do not preclude involvement of the zinc fingers in annealing. Studies conducted with NC mutants containing the zinc fingers but with alterations that affect the structure or sequence of the fingers show clear effects on the annealing activity of the protein. Single molecule stretching experiments completed with 2.1 NC and 1.1 NC showed that it was important to have the N-terminal zinc finger in its native position to enhance the helix-coil transition (66). These findings are in agreement with the work presented here. We have shown that the N-terminal zinc finger must be present for annealing of the strongest nucleic acid structures, and the highest activity was observed when this finger was in its native position (1.1 NC or wild-type NC) (Fig. 6D). This is also in accordance with annealing assays completed with the trans-activation response region. This region of the genome is highly structured, and annealing transactivation response RNA to complementary DNA is substantial only in mutants containing the N-terminal zinc finger in the position observed in wild-type NC (63). It is important to note that of the viruses containing the NC mutations used in this study, only 1.1 NC was infectious, although there was a reduction in infectivity compared with wild-type NC (28,29,63). Taken together, these results indicate that the N-terminal zinc finger of HIV-1 NC is important for the unfolding of strong secondary nucleic acid structure and double-stranded DNA unwinding. It is clear from the infection experiments that the role of finger 2 is nonessential or can be partly compensated by finger 1 and the protein backbone. However, our experiments indicate that finger 2 enhances overall NC activity, and the reduction in (rather than complete loss of) infectivity of mutants lacking this finger also suggests an enhancement role. Finger 2 of NC has not previously been shown to be important for annealing, although annealing assays have not been conducted with unstructured RNAs. The results we obtained indicate that finger 2 may enhance the hybridization of unfolded RNAs, although it cannot enhance the unfolding of strongly structured RNAs. Therefore, it is possible that this zinc finger is involved in enhancing the collisions and binding of complementary nucleic acids. It should be noted that it is also possible that this effect is due solely or at least mostly to the basic backbone residues exterior to the zinc fingers in 2.2 NC. Other NC mutants will have to be examined to more clearly determine the role of finger 2.
Whereas HIV-1 NC contains two zinc fingers, MMLV NC has only one. For strong stop minus-strand DNA transfer to occur in MMLV replication, the only structure that must unfold is much weaker than the HIV-1 trans-activation response hairpin required for minus-single-stranded DNA transfer in HIV (74). Also, Williams et al. (66) have pointed out that retroviruses that contain long repeat regions with numerous hairpin structures have NC proteins with two zinc fingers, whereas those with short repeat regions and minimal structure have NC proteins with only one zinc finger (75,76). Consistent with this is the possibility that a second finger in HIV is required to enhance the unwinding activity of NC.
It is not clear what causes the apparent differences between the activities of the two zinc fingers. There are five amino acid residues that differ between the N-and C-terminal zinc fingers of HIV (NL4-3 strain). These include from finger 1 to finger 2: phenylalanine to tryptophan, asparagine to lysine, isoleucine to glutamine, alanine to methionine, and asparagine to aspartate. Nucleic acid annealing would be favored by a minimization of the electrostatic repulsion between the acidic phosphate groups in complementary strands. HIV-1 NC is a highly basic molecule with a net charge of ϩ13 and would be capable of minimizing that repulsion. The apparent unwinding activity in the N-terminal zinc finger could be related to the presence of more hydrophobic residues in this finger. These residues may play a role in disrupting hydrophobic base stacking interactions by associating with the hydrophobic rings of the bases. We are currently designing mutants in which the N-terminal zinc finger amino acids will be incrementally replaced with those found in the C-terminal zinc finger. These mutants will be used in annealing assays with structured RNA sequences to determine which residues are important for the nucleic acid unfolding capability of the N-terminal zinc finger.
The results presented here indicate that the C-terminal zinc finger of NC plays an accessory role that enhances overall NC chaperone activity, whereas the N-terminal zinc finger is more pivotal and has a clear role in unfolding of secondary structure. Understanding the methods by which NC displays chaperone activity and what roles the zinc fingers have in this activity could be very useful to the design of therapeutic agents that could inhibit the function of NC throughout the viral life cycle.