Molecular Mechanisms by Which Human Immunodeficiency Virus Type 1 Integrase Stimulates the Early Steps of Reverse Transcription

Abstract

Reverse transcriptase (RT) and integrase (IN) are two essential enzymes that play a critical role in synthesis and integration of the retroviral cDNA, respectively. For human immunodeficiency virus type 1 (HIV-1), RT and IN physically interact and certain mutations and deletions of IN result in viruses defective in early steps of reverse transcription. However, the mechanism by which IN affects reverse transcription is not understood. We used a cell-free reverse transcription assay with different primers and compositions of deoxynucleoside triphosphates to differentially monitor the effect of IN on the initiation and elongation modes of reverse transcription. During the initiation mode, addition of IN stimulated RT-catalyzed reverse transcription by fourfold. The stimulation was specific to IN and could not be detected when the full-length IN was replaced with truncated IN derivatives. The IN-stimulated initiation was also restricted to the template-primer complex formed using tRNA₃^Lys or short RNA oligonucleotides as the primer and not those formed using DNA oligonucleotides as the primer. Addition of IN also produced a threefold stimulation during the elongation mode, which was not primer dependent. The stimulation of both initiation and elongation by IN was retained in the presence of an RT trap. Furthermore, IN had no effect on steps at or before template-primer annealing, including packaging of viral genomic RNA and tRNA₃^Lys. Taken together, our results showed that IN acts at early steps of reverse transcription by increasing the processivity of RT and suppressing the formation of the pause products.

The retroviral integrase (IN) catalyzes integration of a double-stranded DNA copy of the viral RNA genome into the host chromosome and is essential for viral replication (for a review, see reference 5). Human immunodeficiency virus type 1 (HIV-1) IN is initially expressed and incorporated into the virus particle as a part of a Gag-Pol precursor polyprotein. Subsequent processing events by viral protease cleave the polyprotein into individual Gag and Pol components, including reverse transcriptase (RT), a heterodimer having two subunits, of 51 and 66 kDa, and the 32-kDa IN (19). IN comprises three discrete functional domains that are conserved among retroviruses. For HIV-1, the N-terminal domain contains a highly conserved zinc-binding HHCC motif. Coordination of zinc by these His and Cys residues stabilizes a helix-turn-helix structure and promotes multimerization of the full-length IN (7, 33, 53). Mutations in this motif affect the 3′-end processing and strand transfer of IN in vitro (32, 50, 53) and abolish viral infectivity in vivo (10, 40, 52). The central core domain contains the invariant catalytic DD(35)E motif that is conserved in all retroviral INs and retrotransposases (45). Mutations in this motif are detrimental to the enzymatic activity of IN and prohibit provirus formation within infected cells (10, 28, 38). The C-terminal domain, the least conserved among retroviral INs, possesses a nonspecific DNA-binding property (9, 11, 12).

RT catalyzes reverse transcription of the single-stranded viral RNA genome into double-stranded cDNA. In the presence of a template, primer, and deoxynucleoside triphosphates (dNTPs), RT can accomplish the conversion of RNA to DNA in vitro without any other added factors. However, the complete synthesis of the cDNA during viral replication is more complex, and several viral and cellular factors can affect reverse transcription and influence the outcome of viral cDNA synthesis. For HIV-1, these factors include the viral proteins Tat, Nef, Vif, Vpr, matrix, and nucleocapsid (NC), as well as various cis-acting elements within the viral RNA-tRNA complex (1, 8, 17, 26, 30, 34, 46). Many of these factors are known to function at various levels of the virus life cycle, rendering the precise mechanism by which a particular factor influences reverse transcription unclear: the factor may affect reverse transcription directly or perhaps only indirectly through other steps during viral replication.

Studies on HIV-1 replication in cell culture show that certain mutations of IN have pleiotropic effects during HIV-1 replication (10, 37, 40, 49, 51). Besides integration, these include early steps in the viral life cycle, such as reverse transcription (31, 39, 49, 52) and nuclear import (13, 49), as well as the late steps postintegration, such as polyprotein processing, assembly, and maturation (6, 44, 47). Since IN is synthesized and packaged into the immature virion as part of the Gag-Pol polyprotein, many of the alterations in virion morphology and defects at the late steps of replication may be caused by the effect of IN mutations on the Gag-Pol precursor protein (6, 44). Certain IN mutations, however, specifically impair reverse transcription with no apparent effects on the Gag-Pol polyprotein and other steps in the life cycle (31, 39, 52, 54). The mechanisms by which IN or its mutants exert the multiple effects are presently poorly understood.

The reverse transcription defect of several IN-deleted HIV-1 strains can be partially rescued when wild-type (WT) IN is incorporated into virions in trans as a Vpr-IN fusion protein (36, 52; C. W. Dobard and S. A. Chow, unpublished data). The IN-negative viruses, as well as viruses containing the C130S substitution in IN, fail to synthesize the early reverse transcription product, the minus-strand strong-stop DNA (−sssDNA), suggesting that the defect is before or during the early stage of reverse transcription (36, 52, 54). In vitro, HIV-1 RT and IN physically interact and the interacting domain in IN is mapped to the C-terminal domain (18, 54). The biological relevance of this IN-RT interaction, however, is not known. To better understand the mechanism whereby IN can affect this process, we employed a cell-free reverse transcription assay that uses the HIV-1 RNA sequence corresponding to the 5′ end of the genome as the template, tRNA₃^Lys as the primer, recombinant RT, and IN to monitor the early steps of reverse transcription. The early events of reverse transcription include two distinct modes: initiation and elongation (29, 30). Initiation has a primer-specific, distributive mode of polymerization and corresponds to the incorporation of the first 5 nucleotides. Initiation is followed by elongation, which has a nonspecific, processive mode of DNA synthesis. The two modes can be differentially monitored in vitro by selecting different primers and compositions of dNTPs. Our analysis revealed that HIV-1 IN can specifically stimulate both the initiation and elongation modes of viral DNA synthesis in vitro. Furthermore, our results demonstrate that IN augments the processivity of RT during the early events of reverse transcription. From these results, we showed that IN directly contributes to the overall efficiency of reverse transcription, possibly through a functional interaction with RT.

MATERIALS AND METHODS

Proteins and reagents.

WT IN was expressed and purified under nondenaturing conditions using a protocol similar to that described previously (15). N-terminally truncated (amino acid residues 51 to 288) and C-terminally truncated (residues 1 to 212) HIV-1 IN, the core domain of HIV-1 IN (residues 51 to 212), and HIV-1 IN containing the F185H/C280S substitutions (IN-FC) were obtained from Robert Craigie at NIH, whereas the C-terminal domain (residues 213 to 288) of HIV-1 IN was from Alan Engelman at the Dana-Farber Cancer Institute. HIV-1 NC was obtained through the AIDS Research and Reference Program from Louis Henderson. T4 gene 32 protein (T4 g32) was purchased from New England BioLabs. The expression plasmid for the HIV-1 RNA template, pHIV-PBS(+), was obtained from Mark Wainberg at McGill University. The HIV-1 RT heterodimer (p66/p51) was kindly provided by Stuart Le Grice at National Cancer Institute. DNA (D18) and RNA (R18) primers complementary to the HIV-1 primer-binding site (PBS) were purchased from Integrated DNA Technologies, Inc. (IDT; Coralville, IA).

RNA template and primers.

The RNA template, consisting of a 957-nucleotide (nt) HIV-1 RNA sequence spanning the 5′ end of Gag, PBS, U5, and the R region of the long terminal repeat, was transcribed in vitro from linearized pHIV-PBS(+) (3) by using the T7-MEGA short script in vitro transcription kit (Ambion, Austin, TX). The RNA transcript of the appropriate size was gel purified using an RNaid kit according to the manufacturer's instructions (Qbiogene). High-pressure liquid chromatography-purified natural tRNA₃^Lys from human placenta, which is 85 nt, was purchased from BIO S&T (Montreal, Canada). In some experiments, the natural tRNA₃^Lys and DNA primers were 5′ end labeled with [γ-³²P]ATP using T4 polynucleotide kinase.

Reverse transcription assays.

Cell-free reverse transcription reactions were carried out in a final volume of 20 μl RT reaction buffer consisting of 50 mM Tris-HCl (pH 7.8), 50 mM NaCl, 6 mM MgCl₂, and 1 mM Tris(2-carboxyethyl)phosphine-HCl. The natural tRNA₃^Lys, RNA primer, or DNA primer was heat annealed to the RNA template. Annealing reaction mixtures containing 0.5 pmol primer and 0.5 pmol template in a volume of 2 μl were first denatured at 95°C for 5 min and then gradually cooled (70°C to 37°C) to allow for annealing and renaturation of the RNA secondary structure. In some experiments, instead of heat annealing, the placement of tRNA₃^Lys onto the RNA template was carried out by adding 1.5 μM IN, NC, or bovine serum albumin (BSA) into the reaction mixture and incubating at 37°C for 1 h. For RT assays, unless indicated otherwise, RT (20 nM, calculated as p66/p51 heterodimers) was incubated at 37°C for 10 min in the presence or absence of various concentrations of IN. The heat-annealed primer-template complexes (0.5 pmol) were added to the RT-containing reaction mixture, and the mixture was incubated at 37°C for an additional 5 min.

To measure the initiation mode of reverse transcription, we employed a 5-nt primer extension assay as described previously (24, 35). The initiation was started by adding 1 μM final concentrations of dCTP, dTTP, and dGTP and 10 μCi of [α-³²P]dCTP (6,000 Ci/mmol) to the reaction mixture in a final volume of 20 μl. In the presence of [α-³²P]dCTP and the absence of dATP, primers that undergo polymerization are internally labeled but not extended beyond the +5 position. To monitor the elongation mode of reverse transcription, reactions were carried out as described above except that 0.2 mM dNTPs and 10 μCi of [α-³²P]dCTP were added to the reaction mixture. Both initiation and elongation reactions were allowed to proceed at 37°C for 30 min and stopped by the addition of loading dye containing 95% formamide and 20 mM EDTA. Products were separated on 8% polyacrylamide-7 M urea gels and visualized by autoradiography.

Annealing of tRNA₃^Lys to the RNA template.

5′-end-labeled tRNA₃^Lys (25 nM) was incubated at 37°C for 1 h with the RNA template (25 nM) and 1.5 μM of either IN, NC, or BSA in a buffer containing 50 mM Tris-HCl (pH 7.8), 100 mM NaCl, and 1 mM Tris(2-carboxyethyl)phosphine-HCl. Samples were then treated with proteinase K (250 μg/ml) at 37°C for 30 min, and RNA was extracted with phenol-chloroform. After ethanol precipitation, the RNA pellet was dissolved in a loading buffer containing 25 mM Tris-HCl (pH 7.8), 10 mM EDTA, 1% sodium dodecyl sulfate, 12.5% glycerol, and 0.005% bromophenol blue and separated on an 8% nondenaturing polyacrylamide gel at 4°C. The gel was then dried, and radioactive bands were visualized with autoradiography.

Measuring RT processivity.

The effect of IN on the processivity of RT during reverse transcription was assessed in the presence of a trap that effectively sequestered any free RT dissociated from the labeled primer-template complex during the initiation and elongation assays described above. The RT trap was prepared by annealing the 957-nt HIV-1 RNA template with a 23-mer RNA primer (Integrated DNA Technologies, Inc.) that is complementary to the 18-nt PBS plus the 5 nucleotides 5′ of the PBS. Reaction mixtures containing RT (20 nM) with or without 10-fold molar excess IN were incubated with template-primer complexes (25 nM) at 37°C for 10 min. The initiation mode of reverse transcription was started by adding 1 μM final concentrations of dCTP, dTTP, and dGTP; 10 μCi of [α-³²P]dCTP; and 125 nM RT trap. For elongation, the reaction was carried out in the presence of 0.2 mM dNTPs, 10 μCi of [α-³²P]dCTP, and 125 nM RT trap. All reactions were allowed to proceed at 37°C for 30 min and stopped by the addition of loading dye containing 95% formamide and 20 mM EDTA. Products were separated on 8% polyacrylamide-7 M urea gels and visualized by autoradiography.

RESULTS

IN stimulates the initiation mode of the early reverse transcription catalyzed by RT.

In vitro studies on HIV-1 viral cDNA synthesis reveal that the early events of reverse transcription occur in a two-step process: initiation followed by elongation (29). The initiation step of reverse transcription can be recapitulated in vitro using a cell-free assay mixture that consists of HIV-1 viral RNA template, human tRNA₃^Lys, RT (p66/p51), and selective compositions of dNTPs (35). The HIV-1 RNA template contains a 5′-UAGCAG sequence immediately upstream of the PBS (Fig. 1A). In the presence of [α-³²P]dCTP, dTTP, and dGTP and in the absence of dATP, the tRNA₃^Lys primers that undergo polymerization will not extend beyond the +5 position. Unlike the +5 product, the +3 and +1 products are abortive transcripts and resulted from dissociation of RT from the primer-template complex at the two known pause sites (35). In the presence of 25 nM primer-template complexes, initiation products were synthesized at RT concentrations as low as 5 nM (Fig. 1B, lane 2). However, at this concentration, the formation of the +5 product was less than 1% of the total initiation products. In contrast, at high concentrations of RT (55 nM and above), a majority of the initiation products (≥70%) were extended to the +5 position (Fig. 1B, lanes 4 to 6). Therefore, under the described reaction conditions, the linear range of RT for the initiation assay was between 5 and 55 nM.

FIG. 1.

Effects of RT concentration on the initiation mode of reverse transcription. (A) Schematic representation of the in vitro extension assay used to study the initiation mode of reverse transcription. Human tRNA₃^Lys primers (thin lines with looped 5′ ends; 25 nM) were heat annealed to the PBS of the 957-nt HIV-1 RNA template (thick line; 25 nM), and the template-primer complex was then incubated with purified RT heterodimers. The initiation was started by adding 1 μM dGTP, dTTP, and [α-³²P]dCTP (asterisks) and allowed to proceed at 37°C for 30 min. The extension products labeled +1, +3, and +5 represent the corresponding numbers of nucleotides added to the 3′ end of the primer. (B) Formation of initiation products as a function of RT concentration. The initiation mode of reverse transcription was assayed under conditions described above in the presence of 0 to 540 nM RT. The upper panel shows the separation of the ³²P-labeled products on a sodium dodecyl sulfate-polyacrylamide gel, with the identities of the extended products indicated to the right. The lower panel shows a bar graph of the formation of the +5 product expressed as a percentage of the total extension product. Values are means ± standard errors of the means of three independent experiments.

To investigate the effect of IN on the initiation mode, we carried out the reaction using 20 nM of RT, which was within the linear range of the assay, in the presence or absence of various concentrations of IN (Fig. 2A). In three independent experiments, IN acted in a concentration-dependent manner to increase the percentage of +5 initiation product compared to RT alone (Fig. 2A). At a fivefold molar excess of IN over RT, the +5 product represented approximately 50% of the total initiation products synthesized (Fig. 2A, lane 3). At a 10-fold molar excess of IN, the synthesis of the +5 product reached saturation at 80%, which was 4-fold higher than that in the presence of RT alone (Fig. 2A, lane 4 versus 1). Unlike synthesis of the +5 product, synthesis of the +1 and +3 pause products decreased and remained unchanged, respectively (Fig. 2A, lanes 2 to 5). If we assume that active IN is a tetramer, then the ratio coinciding with maximum stimulation would be approximately 1 RT heterodimer to 2.5 IN tetramers.

FIG. 2.

IN stimulates the initiation mode of reverse transcription catalyzed by RT. (A) Effect of IN on initiation. The initiation reaction was carried out in the presence of 20 nM RT alone (lane 1) or 20 nM RT plus 20 to 400 nM of IN (lanes 2 to 5) as described in Materials and Methods. (B) Effect of IN on +1 product formation. Primers were limited to a 1-nucleotide extension by adding only [α-³²P]dCTP to initiate reverse transcription. The reaction was carried out with 20 nM RT alone (left) or RT plus a 10-fold molar excess of IN (200 nM) and monitored at the indicated times for the synthesis of the +1 product. (C) Specificity of IN stimulation on +5 product formation during initiation. The initiation reaction with 20 nM RT was carried out in the absence of added protein (lane 1) or in the presence of a 10-fold molar excess of IN (lane 2), BSA (lane 3), T4 g32 (lane 4), or NC (lane 5). (D) Mapping of the IN domains required for stimulating the initiation mode of reverse transcription. RT (20 nM) was incubated without (lane 1) or with (lane 2) 200 nM of full-length IN or the truncated derivatives of IN: N-terminally truncated IN (lane 3; residues 50 to 288), C-terminally truncated IN (lane 4; residues 1 to 212), core domain only (lane 5; residues 50 to 212), and C-terminal domain only (lane 6; residues 213 to 288). In all panels, the numbers on the right denote the identities of the extension products.

In the presence of IN, the increase in the +5 product was coupled with a decrease in the +1 product, suggesting that IN stimulated the initiation mode of reverse transcription by suppressing the abortive products without affecting the overall yield of the reaction. To test this, the initiation assay was carried out as described earlier with or without a 10-fold molar excess of IN except that the reaction was started by adding labeled dCTP only. Since dCTP is the only nucleotide available in the reaction, the tRNA is limited to a single base extension. Under such a condition, addition of IN did not increase the formation of the +1 product compared to RT alone (Fig. 2B). The result confirmed that IN stimulates the formation of the +5 initiation product without affecting the overall yield of the initiation reaction. Furthermore, the data suggested that the stimulation of initiation by IN is specific to the polymerization activity of RT subsequent to the addition of the first nucleotide.

To determine the specificity of the stimulation of +5 product formation by HIV-1 IN, we tested the effect of other DNA- and RNA-binding proteins using the initiation assay (Fig. 2C). Addition of BSA, which served as a negative control, did not show stimulation, and the level of the +5 product was similar to that in the reaction with RT alone (Fig. 2C, lane 1 versus 3). HIV-1 NC, an RT- and RNA-binding protein known to promote primer annealing and strand transfer (23, 34), did not change the percentage of +5 product formation (Fig. 2C, lane 5). T4 g32, a single-stranded DNA-binding protein that improves the yield and efficiency of reverse transcription reactions by increasing the processivity or stability of RT (22, 43), increased the total initiation products about twofold (Fig. 2C, lane 4), validating its use as an RT-stimulating protein. However, unlike what was found for IN, the percentage of the +5 product formed remained similar to that for RT alone, suggesting that IN and T4 g32 stimulate RT by different mechanisms. The results demonstrated that the stimulation of the +5 product observed in the initiation assay is specific to IN.

We showed previously that HIV-1 IN physically interacts with RT and that the RT-interacting domain resides within the C-terminal domain of IN (54). To address whether a specific domain of IN is also sufficient for stimulating RT during initiation in vitro, full-length IN and various truncated IN mutants were added at a 10-fold molar excess to RT and tested for their ability to stimulate initiation (Fig. 2D). No significant increase in the percentage of the +5 initiation product was observed with N-terminally truncated IN (Fig. 2D, lane 3), C-terminally truncated IN (Fig. 2D, lane 4), or the RT-interacting C-terminal domain of IN (Fig. 2D, lane 6). The results suggested that the stimulatory effect of IN on the initiation mode requires the full-length intact protein.

Unlike what was found for other truncation derivatives, addition of the IN core domain inhibited RT-catalyzed initiation (Fig. 2D, lane 5). Several peptides derived from the core domain of HIV-1 IN have been shown to bind RT and inhibit the DNA polymerization of RT in vitro (42). The inhibitory mechanism is not clear but is postulated to be a result of steric hindrance that obstructs the formation of RT-DNA complexes. Whether a similar mechanism is involved in inhibiting the RT-catalyzed initiation by the IN core domain is not known.

IN stimulatory effect on the initiation mode of reverse transcription is primer dependent.

Efficient initiation of HIV-1 reverse transcription in vitro is thought to require the three-dimensional structure of the initiation complex, which involves extensive intra- and intermolecular interactions between tRNA₃^Lys and the viral RNA template (20, 21). Therefore, priming efficiency varies depending on the template and primer used for assaying initiation in vitro. To investigate whether different interactions between the primer-template complex contribute to the IN stimulatory effect, initiation was evaluated using an 18-nt RNA or DNA primer complementary to the PBS region of the viral RNA template. Although the linear ranges of RT concentration were similar, the initiation was far less efficient at producing the +5 product when the reaction was primed with the 18-nt RNA than when primed with tRNA₃^Lys (Fig. 3A and 1B). At 20 nM RT, the reaction with the 18-nt RNA primer synthesized 10-fold less +5 product than the reaction with the tRNA₃^Lys primer. However, similar to what was found for tRNA-primed initiation, addition of a 10-fold molar excess of IN to the 18-nt RNA-primed initiation complex produced a fourfold increase in +5 product formation (Fig. 3B). Unlike what was found for initiation complexes primed with tRNA or 18-nt RNA, IN-mediated stimulation was not observed when an 18-nt DNA was used as a primer (Fig. 3C). The reaction was very efficient even when assayed at the earliest time point. Our results suggested that the IN stimulatory effect on the initiation mode requires the use of RNA as a primer.

FIG. 3.

Effect of different primers on the initiation mode stimulated by IN. (A) Initiation using an 18-nt RNA (R18) as a primer in the absence of IN. The template-primer complex was formed by annealing 25 nM R18 with the 957-nt RNA template (25 nM), and initiation was conducted in the presence of various concentrations of RT (0 to 165 nM). (B) Effect of IN on the initiation mode using complexes formed with R18 as the primer. The initiation reaction was carried out by adding 20 nM RT and various concentrations of IN (0 to 200 nM) as indicated. (C) Effect of IN on the initiation mode using complexes formed with an 18-nt DNA (D18) as the primer. The initiation reaction was performed in the presence of 20 nM RT, with or without 200 nM IN, and monitored at 1 and 3 min after the start of the reaction. In all panels, the numbers on the right have the same meaning as in Fig. 1.

IN stimulates the elongation mode of reverse transcription.

To assay the elongation mode of reverse transcription, we monitored the yield of a reverse-transcribed product that was fully extended to the 3′ end of the RNA template. Under the described condition and using tRNA as a primer, the fully elongated product is 259 nt and corresponds to the −sssDNA synthesized during HIV-1 replication. Similar to that for the initiation mode, the linear range of RT for producing the fully elongated 259-nt product was between 5 and 55 nM (Fig. 4A). Increasing the RT concentration above 55 nM did not significantly increase the yield of 259-nt product synthesis, suggesting that the reaction had reached saturation.

FIG. 4.

Stimulation by IN of the elongation mode of the RT-catalyzed reverse transcription and the effect of primers on IN stimulation. (A) Effect of RT concentration on the elongation mode of reverse transcription. The 957-nt HIV-1 RNA template (25 nM) was heat annealed with 25 nM human tRNA₃^Lys primer, and the resulting template-primer complex was incubated with various concentrations of RT (0 to 540 nM), 200 μM dNTPs, and 10 μCi of [α-³²P]dCTP. The filled arrowhead indicates the position of the fully extended product, which is 259 nt when tRNA is used as the primer. The numbers to the left are the lengths in nucleotides of the DNA size markers. (B) Effect of IN on the elongation reaction using tRNA₃^Lys as the primer. The elongation was monitored after mixing 20 nM RT in the presence of various concentrations of IN as indicated. The fully extended product was identical to that of panel A. (C) Effect of IN on the elongation reaction using R18 as the primer. (D) Effect of IN on the elongation reaction using D18 as the primer. In panels C and D, the reaction was identical to that of panel B except that either R18 or D18 was used to form the primer-template complex with the 957-nt HIV-1 RNA template. The open arrowheads indicate the positions of the fully extended product, which is 192 nt when R18 or D18 is used as the primer.

In the presence of 20 nM RT, addition of IN stimulated RT-catalyzed elongation in a concentration-dependent manner. The stimulation reached saturation at around 100 nM IN (Fig. 4B). To assess the effect of the primer on the elongation mode of reverse transcription stimulated by IN, RNA templates primed with either 18-nt RNA or 18-nt DNA were tested for the ability to produce the fully elongated 192-nt product in the presence or absence of IN. Regardless of the primer-template complex used, addition of 10-fold molar excess of IN stimulated the synthesis of the 192-nt product compared to RT alone (Fig. 4C and D). A similar stimulation by IN was also observed when the elongation reaction was assayed using a 5′-end labeled DNA primer instead of incorporation of labeled dCTP (data not shown). The results demonstrated that the presence of IN enhances the synthesis of the fully elongated product and that the stimulation of elongation is primer independent.

Since initiation complexes primed by DNA bypass the slow initiation step and proceed immediately into the elongation step (3), the result using the D18 primer provided a better means to quantify the effect of IN on the elongation mode alone. Based on the synthesis of the fully elongated 192-nt product, the maximum stimulation of the elongation step attained in the presence of 10-fold molar excess IN was threefold (Fig. 4D).

IN has no effect on the annealing of the tRNA₃^Lys primer to the RNA template.

Studies comparing primer-template complexes annealed by heat or NC showed that initiation is more efficient when the primer-template complex is annealed by NC (34, 35). This observation is attributed to NC's ability to promote and stabilize integral inter- and intramolecular interactions that are unlikely to be formed between tRNA₃^Lys and the viral template by heat annealing (20, 34). Although IN has never been shown to bind RNA, we could not completely exclude the possibility that IN stimulates early reverse transcription by stabilizing or promoting the formation of a more favorable initiation complex. To assess IN′s effect on primer annealing, we modified the initiation assay by forming the primer-template complex under different conditions: either heat annealing or annealing in the presence of BSA, NC, or IN at 37°C (Fig. 5A). The primer-template complex formed by heat annealing or in the presence of NC was used as a positive control, while the primer-template complex formed by adding BSA was used as a negative control. The yield of the primer-template complex was measured indirectly by determining the formation of the +5 initiation product. Under the described condition, only the reaction with either the heat- or the NC-annealed complex was able to efficiently produce the initiation product. The reaction with the IN-annealed complex, similar to the negative control using BSA, resulted in background levels of initiation products. The result suggested that the IN-mediated stimulation of initiation is not due to an increase in annealing or early priming events and occurs after the formation of the primer-template complex. The data further corroborate the earlier observation (Fig. 2B) that IN has no effect on the formation of +1 initiation product.

FIG. 5.

IN has no effect on the annealing of the tRNA primer to the HIV-1 RNA template. (A) Efficiency of primer annealing measured by the initiation assay. RNA templates (25 nM) and tRNA₃^Lys primers (25 nM) were mixed and subjected to heat annealing (HA) or incubation with 1.5 μM of BSA, NC, or IN at 37°C for 1 h. The efficiency of annealing was assessed by measuring the synthesis of initiation products after the addition of 0.15 μM RT and 1 μM dGTP, dTTP, and [α-³²P]dCTP. The reaction was allowed to proceed for 1, 3, or 30 min. The identities of the initiation products are indicated by the numbers to the right and have the same meaning as in Fig. 1. (B) Formation of the template-primer complex. ³²P-labeled tRNA₃^Lys primers and RNA templates were annealed by heat treatment (lane 2) or incubated without added protein (lane 1) or with 30 pmol of BSA (lane 3), IN (lane 4), or NC (lane 5) at 37°C for 1 h. Primer-template complexes were recovered, resuspended in loading buffer, and separated on nondenaturing 8% polyacrylamide gels at 4°C. Free tRNA₃^Lys primer (*P) and annealed RNA-tRNA₃^Lys complexes (T/*P) are as indicated on the right. (C) Initiation using template-primer complexes isolated from cell-free virus particles. RNA was isolated from wild-type (NL4-3) or IN-negative virions (NLΔIN), and equal amounts (500 ng) of total nucleic acids were used as the source of primer-template complexes in the initiation assay described in Materials and Methods. Heat-annealed RNA template-tRNA₃^Lys complexes were subjected to the same viral RNA extraction condition and served as a positive control (HA). The initiation reaction was carried out in the presence of 1 μM dGTP, dTTP, and [α-³²P]dCTP. Because of the small quantity of the virion-derived primer-template complex, the concentration of RT used was increased to 165 nM. The numbers on the right have the same meaning as in panel A.

We also evaluated IN′s involvement in primer annealing by assessing directly its ability to anneal tRNA₃^Lys to the viral RNA template using a nondenaturing gel shift assay (Fig. 5B). The viral RNA template and labeled tRNA₃^Lys were incubated with IN, NC, or BSA. After incubation, protein was removed by proteinase K treatment, and the RNA was separated by electrophoresis through a nondenaturing 8% polyacrylamide gel. As expected, addition of NC promoted the formation of the tRNA-template complex to a similar extent as the heat-annealed control (Fig. 5B, lanes 2 and 5), whereas no complex was detected in the presence of BSA (Fig. 5B, lane 3). Addition of IN also did not result in the annealing of the labeled tRNA₃^Lys primer to the viral RNA template (Fig. 5B, lane 4).

To further substantiate that IN was not involved at the annealing step, we isolated primer-template complexes from WT NL4-3 and IN-negative (NLΔIN) virions and determined their ability to undergo primer extension. Previous studies have shown that RNAs extracted from viruses defective in RNA packaging or annealing are also incapable of undergoing primer extension (4, 16). WT HIV-1 or viruses lacking IN were lysed, and total viral RNA was extracted. The total amounts of viral nucleic acids, as determined by measuring optical density, for WT and NLΔIN virions were similar (data not shown). Equal amounts of total nucleic acids isolated from WT or NLΔIN virions were then used in place of the heat-annealed template-primer complex in the initiation assay described previously. As a control for RNA isolation, heat-annealed template-primer complexes were extracted using the identical procedure. We found that the primer-template complex isolated from NLΔIN virions, similar to that from WT virions, was competent for primer extension (Fig. 5C). The result obtained from infectious virions, therefore, corroborated that from the in vitro assays in showing that IN is not essential for primer-template annealing. The IN-mediated stimulation of initiation occurs subsequent to the annealing of the tRNA primer to the template. Since the formation of the primer-template complex requires the incorporation of template RNA into the virion, the result also suggested that IN is not necessary for RNA packaging.

IN enhances the processivity of RT during both initiation and elongation modes of reverse transcription.

Reverse transcription catalyzed by recombinant HIV-1 RT is inefficient in vitro (14, 25). This is partly due to RT's poor processivity, which can be measured by determining the length of primer extension occurring from the time RT binds and initiates synthesis until it dissociates from the primer-template. In this study, we measured the processivity of RT by comparing the distributions of abortive and full-length products synthesized from a single round of extension. We showed earlier that a 10-fold molar excess of IN stimulated the synthesis of initiation and elongation products during early reverse transcription (Fig. 2A and 4B). Since IN has no effect on the formation of the primer-template complex, we hypothesized that the increase in the early reverse transcription products is due to an increase in processivity, brought forth by IN′s ability to stabilize the association of RT to the primer-template complex. To test the effect of IN on RT processivity, we performed initiation and elongation reactions in the presence of a trap sequestering any free RT that dissociates from the primer-template complex during polymerization. We chose the 957-nt HIV-1 RNA template annealed with a 23-nt RNA primer as a trap because, unlike homopolymeric DNA or heparin, such a construct did not bind IN and only negligible amounts of initiation and elongation products were detected if RT was added after the introduction of 125 nM RT trap (Fig. 6A and B, lane 1). In the control initiation reaction without the trap, as shown earlier in Fig. 2, addition of IN resulted in an increase of the +5 initiation product (Fig. 6A, lane 3 versus 2). In the presence of the trap and without IN, the level of the +5 product decreased while the level of the +1 product increased (Fig. 6A, lane 4 versus 2). This observation is consistent with the low processivity of RT and its propensity to dissociate from the primer-template at pause sites (27). However, the formation of the +5 product could be restored in the presence of the trap when the initiation reaction was carried out in the presence of IN (Fig. 6A, lane 5 versus 4).

FIG. 6.

IN enhances the processivity of RT. (A) Effect of IN on initiation in the presence of an RT trap. Heat-annealed RNA template-tRNA₃^Lys complexes were mixed with 20 nM RT alone (lanes 2 and 4) or 20 nM RT plus 200 nM IN (lanes 3 and 5) at 37°C for 10 min, and the initiation reaction was started by adding 1 μM dGTP, dTTP, and [α-³²P]dCTP. The processivity of RT was assessed by measuring the synthesis of the fully extended initiation product (+5) in the absence (lanes 1 and 2) or presence (lanes 3 and 4) of 125 nM RT trap. The numbers to the right have the same meaning as in Fig. 1. (B) Effect of IN on elongation in the presence of an RT trap. The reaction condition was identical to that of panel A except that 200 μM dNTPs were added to start the reaction. The fully extended 259-nt elongation product and the lengths in nucleotides of the DNA size markers are indicated to the right. In both panels, lane 1 represents the control for the trap assay, in which the RT trap was added to the reaction mixture before RT.

Similarly, the synthesis of the fully elongated, 259-nt product and intermediate abortive products during elongation was greatly diminished in the presence of the trap and without IN during elongation (Fig. 6B, lane 4 versus 2), and addition of IN increased concomitantly the yield of the fully elongated product and the intermediate abortive products in the presence of the trap (Fig. 6B, lane 5 versus 4). Thus, IN stimulated both the initiation and elongation modes of reverse transcription by enhancing the processivity of RT.

An RT-noninteracting IN mutant has no stimulatory effect on initiation and elongation.

WT IN specifically interacts with RT (18, 54). If IN stimulates the two modes of early reverse transcription by interacting with RT, we expect that IN mutants that fail to interact with RT would have no effect on RT-catalyzed initiation and elongation. Using the far-Western blotting assay, we showed that an HIV-1 IN mutant, IN-FC, which contains Phe-to-His and Cys-to-Ser substitutions at residues 185 and 280, respectively (F185H/C280S), was much reduced in its ability to bind RT compared with WT IN or NC (Fig. 7A). As expected, when WT IN was replaced by IN-FC in the reverse transcription assays, little or no stimulation was observed in both initiation (Fig. 7B, lane 3) and elongation reactions (Fig. 7C, lane 3).

FIG. 7.

The RT noninteracting IN mutant (IN-FC) does not stimulate RT-catalyzed reverse transcription. (A) IN-FC does not bind RT. Two microliters containing 20 pmol of WT IN, NC, or IN-FC was spotted onto a nitrocellulose membrane. The dried membrane was subjected to UV cross-linking for 30 s, then blocked in HBB buffer (25 mM HEPES [pH 7.8], 10 mM ZnCl₂, 5 mM MgCl₂, and 25 mM NaCl) plus 5% milk at ambient temperature for 4 h. After being washed three times with HBB buffer, the membrane was incubated overnight at 4°C with shaking in the presence of 0.1 μM RT in HBB buffer, 1% milk, and 0.05% NP-40 (bottom). As a negative control, an identical membrane was incubated under the same conditions without RT (top panel). The membrane was then washed three times with HBB buffer and probed with the mouse anti-RT monoclonal antibody at 1:500 dilution at 37°C for 4 h. A chemiluminescence substrate kit (SuperSignal; Pierce) was used for detection according to the manufacturer's instructions. (B) Effect of IN-FC on initiation. Heat-annealed RNA template-tRNA₃^Lys complexes were mixed with 20 nM RT either alone (lane 1) or in the presence of 200 nM IN (lane 2) or 200 nM IN-FC (lane 3) at 37°C for 10 min, and the initiation reaction was started by adding 1 μM dGTP, dTTP, and [α-³²P]dCTP. The numbers to the right have the same meaning as in Fig. 1. (C) Effect of IN-FC on elongation. The reaction condition was identical to that of panel B except that 200 μM dNTPs were added to start the reaction. The fully extended 259-nt elongation product and the lengths in nucleotides of the DNA size markers are indicated to the right of the panel.

DISCUSSION

During reverse transcription, the complete synthesis of the viral cDNA involves the coordination of many intra- and intermolecular interactions and rearrangements within the reverse transcription complex. A detailed understanding of the molecular mechanisms that govern the formation and execution of these complexes remains obscure. As a result, distinguishing between factors that specifically affect reverse transcription and those that are involved at steps preceding reverse transcription, such as RNA packaging, primer annealing, and uncoating, remains a difficult task. IN has been shown to be involved during HIV-1 reverse transcription, but the mechanism of this involvement is currently unknown (52, 54). Although RT itself is sufficient to catalyze reverse transcription in vitro, our results show that, under physiologically relevant conditions, IN is a bona fide factor stimulating early events of reverse transcription. In vitro analysis indicates that the initiation and elongation modes of reverse transcription are stimulated by IN in a concentration-dependent manner. The stimulation of initiation, especially the formation of the +5 initiation product, is specific to IN and not shared by other known RT- and RNA-binding proteins tested in this study. In addition, full-length IN is required for stimulation, which is in agreement with in vivo studies revealing that full-length IN is required for intracellular reverse transcription (36, 52).

Our study showed that IN stimulates the early events of reverse transcription by increasing RT processivity, possibly by stabilizing RT during polymerization through protein-protein interactions. This model is supported by the observation that, in the presence of an RT trap that restricts RT to a single cycle of polymerization, addition of IN increases initiation and elongation product formation more than 5- and 20-fold, respectively, compared to RT alone. In the absence of the RT trap, IN increases the synthesis of the +5 initiation product while concomitantly decreasing the +1 initiation product, suggesting that IN functions by suppressing the abortive products without affecting the overall yield of the reaction. We think that it is unlikely that IN stimulates polymerization at the primer-binding site by functioning as a processivity clamp, since HIV-1 RT already binds persistently to the tRNA-RNA initiation complex even in the absence of IN. In addition, any increase in primer-template binding affinity would be expected to have quantifiable effects on the formation of the +1 initiation product, which is not observed. Another possible mechanism is that IN stimulates reverse transcription by facilitating the proper annealing of unfinished or distorted tRNA-RNA complexes, which are relatively unstable structures and would otherwise impede polymerization. The observations that IN fails to promote annealing of 5′-end-labeled tRNA₃^Lys to a cRNA molecule and does not induce tRNA primer extension in the absence of a heat-annealed initiation complex and that viral RNA-tRNA complexes extracted from virions lacking IN can carry out primer extension argue against a primer-annealing mechanism.

The activity of RT during the initiation mode is increased by IN when an RNA-primed but not a DNA-primed initiation complex is used in place of the endogenous tRNA₃^Lys-RNA complex. Previous reports showed that DNA-primed initiation complexes bypass the slow initiation step and proceed immediately into the elongation mode (3). This phenomenon occurs because RT dissociates less frequently from the heteroduplex primer-template (DNA-RNA), thus leading to a decrease in abortive synthesis at various pausing sites (2). Since our data support the finding that IN stimulates RT-catalyzed reverse transcription by enhancing processivity, it is not surprising that IN has no effect during initiation when a DNA-primed complex is used. These results further support the in vivo data showing that IN is required at the early step of reverse transcription, during the RNA-dependent polymerization to form the −sssDNA (31, 52, 54).

The effects of IN on RT activities have been examined previously using various in vitro reverse transcription assays, and different results have been reported (18, 41, 48). One study showed that IN inhibits the DNA-dependent but not RNA-dependent polymerase activity of HIV-1 RT (48), whereas two other studies reported that both the RNA- and DNA-dependent polymerase activities of RT and its processivity are unaffected by the presence of IN (18, 41). The discrepancy of the IN effect on RT may be attributed to the specific step of reverse transcription that is monitored or the manner by which the in vitro reverse transcription assay is carried out. For instance, some earlier studies used poly(rA)/oligo(dT) or poly(dA)/oligo(dT) as the RT primer-template complex to evaluate the effect of IN on extension (41, 48). We showed here that the stimulation of reverse transcription by IN requires an RNA primer, which could explain the lack of stimulation in the previous reports. In addition to physiologically relevant HIV-1 RNA substrates, the concentration of RT used is critical, as saturating levels of RT can mask IN′s stimulatory effect. Therefore, the inability to detect stimulatory effects by IN is possibly attributed to a combination of carrying out the reactions using saturating amounts of RT together with DNA-primed initiation complexes.

Many IN mutants are defective in reverse transcription (10, 37, 40, 49, 51). Although some of these mutations may have secondary effects, thereby affecting reverse transcription indirectly through other steps during viral replication (6, 44), certain IN mutations appear to specifically impair reverse transcription (31, 39, 52, 54). The mutations that purportedly specifically affect reverse transcription are found in different domains of HIV-1 IN, such as replacement of His at position 12 or 16 with Ala, replacement of Cys at position 130 with Ser, replacement of Phe at position 185 with Ala, and deletion of the last 22 residues in the C terminus (31, 39, 52, 54). The common phenotype exhibited by the wide array of different mutations scattered throughout the IN molecule raises the possibility that IN may affect reverse transcription not by a single mechanism but through multiple mechanisms. The direct stimulation by IN of RT-catalyzed reverse transcription suggests that the RT-IN interaction is functional and biologically relevant. Disruption of this RT-IN interaction may form the mechanistic basis for the reverse transcription defect produced by some IN mutations, as exemplified by the IN mutant containing the F185K/C280S substitutions. It is important to note that an HIV-1 mutant containing a F185A substitution in IN is replication defective and is impaired in early reverse transcription (52).

RT and IN are the viral enzymes responsible for catalyzing the essential steps of reverse transcription and integration, respectively, during the early stage of the retroviral life cycle. This study shows that HIV-1 IN can specifically stimulate both the initiation and elongation modes of viral DNA synthesis in vitro by increasing the overall processivity of RT. Both HIV-1 RT and IN are part of a large nucleoprotein complex that is integral for all the functions and activities of the virus subsequent to attachment to and entry into an infected cell. The physical and functional interactions between RT and IN may serve to ensure that the virion-derived nucleoprotein complex is assembled properly and can perform efficiently. Further investigations on the interaction between these two key retroviral enzymes may lead to a better understanding of the biological significance of the RT-IN complex formation during the early stages of HIV-1 replication.