HIV-1 Nucleocapsid Protein Switches the Pathway of TAR RNA/DNA Annealing from Loop-Loop “Kissing” to “Zipper”

Summary

The chaperone activity of human immunodeficiency virus type 1 (HIV-1) nucleocapsid protein (NC) facilitates multiple nucleic acid rearrangements that are critical for reverse transcription of the single-stranded RNA genome into double-stranded DNA. Annealing of the trans-activation response element (TAR) RNA hairpin to a complementary TAR DNA hairpin is an essential step in the minus-strand transfer step of reverse transcription. Previously, we used truncated 27-nucleotide (nt) mini-TAR RNA and DNA constructs to investigate this annealing reaction pathway in the presence and absence of HIV-1 NC. In this work, full-length 59-nt TAR RNA and TAR DNA constructs were used to systematically study TAR hairpin annealing kinetics. In the absence of NC, full-length TAR hairpin annealing is ∼10-fold slower than mini-TAR annealing. Similar to mini-TAR annealing, the reaction pathway for TAR in the absence of NC involves the fast formation of an unstable “kissing” loop intermediate, followed by a slower conversion to an extended duplex. NC facilitates the annealing of TAR by ∼10⁵-fold by stabilizing the bimolecular intermediate (∼10⁴-fold) and promoting the subsequent exchange reaction (∼10-fold). In contrast to the mini-TAR annealing pathway, wherein NC-mediated annealing can initiate through both loop-loop kissing and a distinct “zipper” pathway involving nucleation at the 3′/5′ terminal ends, full-length TAR hairpin annealing switches predominantly to the zipper pathway in the presence of saturated NC.

Introduction

Human immunodeficiency virus type-1 (HIV-1) is a retrovirus that contains its genomic information in the form of two identical single-stranded (ss) RNA molecules. Conversion of the RNA genome into double-stranded viral DNA is an essential step of the retroviral life cycle. Reverse transcription involves multiple steps, including several nucleic acid rearrangements catalyzed by the HIV-1 nucleocapsid protein (NC). NC has been shown to facilitate annealing of the tRNA primer onto the primer binding site ^1-4, annealing of complementary RNA/DNA repeat regions in minus-strand transfer ^5-11, and annealing of complementary DNA sequences in plus-strand transfer ^{7; 12-14}. HIV-1 NC is a short, basic, nucleic acid binding protein, containing two zinc finger domains, each of which has an invariant CCHC metal-ion binding motif ^15-18. The mature protein is produced by proteolytic cleavage of the Gag precursor and is found in the interior of the virus particle, where it is tightly associated with genomic RNA ¹⁹.

NC has multiple functions during the viral replication cycle. In addition to playing essential roles in reverse transcription, NC or the NC domain of Gag have been shown to facilitate genomic RNA packaging, virus assembly, and integration ^20-24. Many of these functions rely on the nucleic acid chaperone activity of NC, which promotes nucleic acid rearrangements leading to thermodynamically more stable structures ^22-26. The chaperone function of NC relies on two main activities, nucleic acid (NA) aggregation ^27-30 and duplex destabilization ^31-38, associated with the cationic N-terminal domain and zinc fingers, respectively.

It has recently been shown that synthesis of long products during reverse transcription in vitro is facilitated under conditions of NA aggregation ³⁹. In addition, DNA synthesis at low RT concentrations became more processive under aggregating conditions, and strand displacement synthesis of the last 99 nucleotides (nt) resulting in production of the DNA flap was shown to be more efficient. Sedimentation analysis and gel-shift annealing studies performed under the same solution conditions as a function of NC concentration, showed that NA aggregation and annealing activities closely parallel each other ³⁰.

Annealing of the trans-activation response element (TAR) DNA hairpin to a complementary TAR RNA hairpin, an essential step in minus-strand transfer, was shown to be accelerated by NC as much as 3000-fold by NC ¹¹. Previous studies used model 27-nt mini-TAR RNA and DNA hairpins derived from the top hairpin loop region of TAR to elucidate the mechanism of NC-mediated TAR RNA/DNA annealing ³⁰. This study showed that the complementary mini-TAR RNA and DNA hairpins anneal via a two-step pathway:

$$R+D\underset{k_{-1}}{\overset{k_{1}D}{\rightleftarrows }}RD\ast \underset{k_{-2}}{\overset{k_2}{\rightleftarrows }}RD$$ (1)

Here, the first bimolecular step leads to the formation of the annealing intermediate RD*, with the forward association rate constant k₁, and the intermediate dissociation rate k_-1. The annealing intermediate RD* is subsequently converted into the fully annealed RNA/DNA duplex RD via a second monomolecular step, which is characterized by the strand exchange rate k₂ and reverse rate k_-2. DNA mutagenesis data showed that in the absence of NC, mini-TAR RNA/DNA annealing is nucleated via an extended loop-loop “kissing” interaction ³⁰. The annealing intermediate in this case involves 17 intermolecular base pairs (bp), including nt from the hairpin loop and from the adjacent stem and 3-nt bulge (see Figure 1). In the presence of saturating amounts of NC, an additional annealing pathway was observed, involving nucleation near the 3′/5′ ends ³⁰. The later pathway has also been termed the “zipper” mechanism (Figure 1)⁴⁰. Multiple annealing pathways have also been proposed for annealing of full-length TAR DNA to complementary DNA oligonucleotides based on single molecule studies ⁴⁰. However, a kinetic analysis using full-length TAR RNA/DNA hairpins derived from the HIV-1 MAL isolate and NC(12-55), concluded that NC facilitated this annealing reaction primarily via the zipper pathway ⁴¹. A recent single molecule spectroscopy study using full-length TAR DNA/cDNA hairpins and full-length HIV-1 NC under non-aggregating conditions is consistent with this conclusion ⁴².

Figure 1.

Two pathways of full-length TAR RNA/DNA annealing. Top: predicted secondary structures of full-length TAR RNA (black) and TAR DNA (gray) hairpins. Sequences are derived from the HIV-1 NL4-3 isolate. The TAR RNA and DNA secondary structures were predicted by m-fold analysis⁵¹. Residues mutated in different DNA constructs are circled. Bottom: (left) loop-loop “kissing” pathway, which involves initial formation of an extended “kissing” complex followed by subsequent strand exchange to form the fully annealed duplex; (right) “zipper” pathway involving nucleation through the 3′/5′ termini resulting in formation of a zipper intermediate followed by conversion to the fully annealed duplex.

In this work, we use gel-shift assays to systematically study the annealing kinetics of full-length 59-nt TAR RNA and DNA hairpins in the presence and absence of wild-type HIV-1 NC. The substrates and conditions used here were designed to closely mimic the in vivo process. In particular, the use of full-length NC and gel-based rather than fluorescence-based assays allows reactions to be performed under conditions that promote NA aggregation. In addition, the role of RNA/DNA sequence complementarity and the NC concentration-dependence of the annealing reaction are examined. The data obtained in the absence of NC support an overall slow two-step structure-based annealing mechanism, involving nucleation via a loop-loop kissing interaction followed by conversion to an extended annealed duplex. In contrast, the presence of saturating NC switches the annealing mechanism to the zipper pathway involving intermediate complex formation between the 3′/5′ ends of the hairpin stems.

Results and Discussion

As observed previously for mini-TAR RNA/DNA annealing³⁰, under all conditions examined here the annealing kinetics appears to be bi-exponential, and the percent of the molecules annealed as a function of time, P(t), can be described by the fast and the slow rates, k_f, and k_s, and the fraction of the fast component f according to:

$$P(t)=P_{\infty }\cdot \left(f\cdot (1-e^{-k_{f}t})+(1-f)\cdot (1-e^{-k_{s}t})\right).$$ (2)

Previous studies ³⁰ showed that k_f and k_s correspond to the rate of formation of the intermediate and of the fully annealed duplex, respectively. The fraction of the fast component f is the probability of intermediate formation and P_∞ is the final equilibrium percent of RNA annealed. The following equations describe the relationship between these kinetic parameters and the elementary rates of the two-step process ⁴³:

$$k_f=k_{1}D+k_{-1}\text{and}{k}_s=f\cdot k_2+k_{-2},$$ (3)

where

$$f=\frac{k_{1}D}{k_{1}D+k_{-1}}=\frac{D/K_d^{\ast }}{D/K_d^{\ast }+1}\text{and}{P}_{\infty }=100\cdot f\cdot \frac{k_2}{k_2+k_{-2}}=100\cdot \frac{D/K_d}{D/K_d+1}$$ (4)

Here, $K_d^{\ast }$ and K_d are the equilibrium dissociation constants of the intermediate RD* and the final annealed product RD, respectively. Experimentally, both rates are observed when they are sufficiently different from each other, i.e., k_f » k_s, and when the intermediate complex is reasonably stable, i.e., 0.1 < f < 1. The annealing of the TAR molecules shown in Figure 1 is strongly driven by formation of 14 new bp, such that the TAR hairpins anneal irreversibly both in the presence or absence of NC, i.e., P_∞=100%, and k_-2 is negligible. In addition, under all conditions examined here, the slow rate k_s is proportional to f, in accord with the second relationship in (3).

According to eqs. (3) and (4), the fitted parameters k_f, k_s, and f can be used to estimate the elementary rates of the two-step annealing process using the following relationships:

$$k_1=k_f\cdot \frac{f}{D},k_{-1}=k_f\cdot (1-f),k_2=k_s/f,K_d^{\ast }=\frac{k_{-1}}{k_1}=\frac{1-f}{f}\cdot D$$ (5)

Kinetics of full-length TAR RNA/DNA annealing in the absence of NC

Dependence of full-length TAR RNA/DNA annealing kinetics on the concentration of TAR DNA

Figure 1 shows the full-length 59-nt TAR RNA and TAR DNA substrates used in this work. Gel-shift assays with [³²P]-labeled TAR RNA and unlabeled TAR DNA were used to quantitatively monitor complex formation as a function of time. Typical annealing time courses are shown in Figure 2(a). The annealing reactions were performed using several TAR DNA concentrations, with DNA in large excess over TAR RNA. In the absence of NC, the annealing reaction is extremely slow and irreversible. Under these pseudo-first order conditions, the reaction rate increases with increasing DNA concentration, which is consistent with a bimolecular association. To quantitatively analyze the annealing kinetics, the annealing time courses in Figure 2(b) were fit to the general two-exponential expression (2) to obtain k_s, k_f, and f. P_∞ was set to 100% since the reaction is irreversible.

Figure 2.

Time course of TAR RNA/DNA hairpin annealing in the absence of NC. (a) Typical gel-shift analysis performed with 50 nM TAR RNA and 1 μM TAR DNA as a function of time. The numbers at the top of each lane correspond to minutes. Bound (B) and free (F) RNA bands are labeled on the left. (b) Time course showing the percent of TAR RNA (15 nM) annealed to variable concentrations of TAR DNA (as indicated on each curve) at 37 °C. Lines represent double-exponential fits of the data to equation (2), with the equilibrium percent annealed fixed at 100%. The insert shows the dominant slow annealing rate, k_s obtained from the fits of the data shown in (b), as a function of TAR DNA concentration, D. The line is a linear fit of the data with a slope corresponding to k_eff=10±5 M^-1s^-1. (c) Temperature dependence of TAR RNA/DNA annealing. Time courses for TAR RNA (15 nM) annealing to TAR DNA (1.5 μM) were measured in the absence of NC in 20 mM Na⁺ at the indicated temperatures.

The fraction of the fast component, f, is very small (<0.1), but increases with increasing DNA concentration, D. The D dependence of the dominant slow annealing rate, k_s, is presented in the insert in Figure 2(b). As expected, k_s increases proportionally to D, and the slope of k_s versus D yields an apparent bimolecular association rate constant k_eff of 10(±5) M^-1 s^-1. The small probability of intermediate formation (i.e., f«1) and the slow strand exchange rate both contribute to the extremely slow overall rate of annealing observed in the absence of NC. As shown in Figure 2(c), increasing temperature leads to a moderate rate enhancement. Vant Hoff analysis of these data yields an enthalpy for the rate-limiting step of ΔH=27±10 kcal/mol (data not shown). This value is comparable to the ΔH value determined for mini-TAR annealing (ΔH=15±5 kcal/mol) ³⁰. This result is consistent with the hypothesis that mini-TAR and TAR anneal via the same pathway in the absence of NC. Interestingly, TAR annealing is ∼10-fold slower than mini-TAR annealing in the absence of protein chaperone ³⁰. We hypothesize that this is due to the larger entropy loss upon intermediate complex formation in the case of the longer hairpins ⁴⁴. Alternatively, slower annealing of longer hairpins may be due to a greater electrostatic repulsion. However, an analysis of the salt-dependence of the TAR annealing kinetics does not support the latter explanation ⁴⁵.

DNA mutational analysis

To determine the pathway of full-length TAR RNA/DNA annealing, kinetic studies were performed using various mutant TAR DNA hairpins. To allow for more accurate quantification of extremely slow annealing kinetics observed in the absence of NC, assays were performed in the presence of 100 mM Mg²⁺, which is known to facilitate the annealing reaction without altering the annealing pathway ⁴⁵. The importance of loop-loop complementarity was examined by studying the annealing of TAR RNA to a double (C27A/C28A) loop mutant DNA hairpin (Figure 1, right). Indeed, the C27A/C28A loop mutant displayed dramatically slower annealing compared to WT DNA (Figure 3). In contrast, when the two G:C base pairs at the bottom of the stem are inverted (B-stem mutant, see Figure 1), the rate of annealing is unaffected (Figure 3). These results suggest that loop-loop complementarity is critical for full-length TAR RNA/DNA annealing, while stem end complementarity is not. An additional TAR DNA variant was constructed containing a 3-nt insertion that results in closure of the 3-nt single-stranded bulge located near the top of the hairpin (T-zip, see Figure 1). This change creates 12 continuous bp adjacent to the hairpin loop and results in a dramatic reduction in the annealing rate (Figure 3).

Figure 3.

Mutational analysis of TAR RNA/DNA annealing in the absence of NC. Percent TAR RNA (15 nM) annealed to various TAR DNA molecules (500 nM) as a function of time. The TAR DNA variants used are indicated by each curve (WT, Loop, T-zip or B-stem, see Figure 1). Annealing was performed at 37 °C in the presence of 100 mM Mg²⁺. Lines represent double-exponential fits of the data to equation (2), with the equilibrium percent annealed fixed at 100%.

For all DNA mutants, the annealing kinetics measured in the presence of 100 mM Mg²⁺ was clearly biphasic. The k_f, k_s and f parameters were determined by fitting the experimental annealing time courses shown in Figure 3 to eq. (2). The elementary reaction rates determined from these fits using eqs. (5) are plotted in Figure 4. As shown in Figure 4(a), the intermediate association rate, k₁, decreased by ∼7-fold when mutations were present in the top portion of the hairpin (i.e., Loop or T-zip). In contrast, the B-stem mutation had no effect on this rate. These results are consistent with a mechanism of TAR RNA/DNA annealing involving an extended loop-loop kissing intermediate (see Figure 1, lower left). As expected based on this annealing pathway, the B-stem mutation does not affect k₁.

Figure 4.

Elementary reaction rates and intermediate dissociation constant for annealing of TAR RNA to various TAR DNA constructs in the absence of NC. The elementary rates (k₁, panel a; k_-1, panel c; k₂, panel d) and intermediate dissociation constant (panel b) for the two-step annealing process were obtained by analysis of annealing time courses in Figure 3 using equations (2-5).

In agreement with these results, the intermediate dissociation rate, k_-1, was not significantly affected by T-zip or B-stem mutations, but increased ∼5-fold in the presence of the Loop mutation (Figure 4(c)). As a result of these changes in the on and off rates for the formation of the reaction intermediate, the intermediate dissociation constant, K_d*, increased ∼30-fold and ∼10-fold in the presence of Loop and T-zip mutations, respectively (Figure 4(b)). Interestingly, the conversion rate, k₂, is similar (i.e., within 2-fold of WT) for all of the TAR DNA constructs (Figure 4(d)). The small value of this rate, k₂,=0.03±0.02 min^-1, suggests that strand exchange is rate-limited by the cooperative melting of a large portion of one or both stems, followed by much faster strand exchange. This hypothesis is consistent with the observation that k₂ is essentially the same for mini-TAR and TAR hairpins ⁴⁵, i.e., that the strand exchange rate does not depend on hairpin stem length. Taken together, these data are consistent with an annealing mechanism in the absence of NC that involves nucleation through an extended loop-loop kissing interaction followed by slower conversion to a fully extended annealed duplex.

Kinetics of full-length TAR RNA/DNA annealing in the presence of NC

NC greatly facilitates full-length TAR RNA/DNA annealing. The time course of annealing in the presence of saturating amounts of NC (4:1 nt:NC ratio in low salt buffer, i.e., 20 mM Na⁺ and 0.2 mM Mg²⁺) is well described by a single-exponential expression (Figure 5). The final yield of annealed product in the reactions shown in Figure 5b (D = 30-90 nM) is close to 100%. Fitting the annealing time courses to a single-exponential expression yields an annealing rate of 4±2 min^-1. Although under these conditions, the annealing rate appears to be independent of D (due to the time resolution of the assay), experiments performed at lower D (10-20 nM) show a reduced rate consistent with a bimolecular annealing mechanism (see inset in Figure 5(b)). The equilibrium percent of annealed product also decreases at lower D, suggesting that the annealing reaction is reversible under these conditions, as observed previously³⁰. The annealing rate measured for D = 30-60 nM is ∼10-fold faster than the rate of NC-facilitated mini-TAR RNA/DNA annealing at similar D³⁰. Given that TAR anneals ∼10-fold slower in the absence of NC, the overall NC-facilitated rate enhancement is ∼100-fold greater for full-length TAR annealing compared to mini-TAR.

Figure 5.

Time course of TAR RNA/DNA hairpin annealing in the presence of saturating amounts of NC. (a) Typical gel-shift analysis performed with 15 nM TAR RNA and 45 nM TAR DNA. The numbers at the top of each lane correspond to minutes. The positions of bound (B) and free (F) RNA bands are indicated on the left. (b) Time dependence of the percent of TAR RNA (15 nM) annealed to variable concentration of TAR DNA (as indicated by each curve) at 37 °C and a 4:1 nt:NC ratio. The 10-20 nM DNA curves are shown as an inset. (c) TAR RNA/DNA annealing time courses (15 nM TAR RNA+ 45 nM TAR DNA) obtained at different temperatures, as indicated in the Figure. All lines are single-exponential fits of the data.

Increasing temperature facilitates TAR annealing further (Figure 5(c)). The Van't Hoff analysis of the temperature dependence of the dominant fast rate (data not shown) yields an enthalpy for the rate-limiting step of 28±8 kcal/mol, consistent with extensive melting occurring prior to complex formation. This enthalpy value is ∼2-fold higher than previously reported for mini-TAR annealing ³⁰. This result implies that in the presence of saturating amounts of NC, annealing of the longer TAR hairpins involves pre-melting of the more stable region than in the case of mini-TAR hairpin annealing.

DNA mutational analysis

We next used the mutant TAR DNA constructs shown in Figure 1 (right) to investigate the annealing pathway of full-length TAR in the presence of saturating amounts of NC. Under these conditions, annealing in the presence of the Loop or T-zip variants is very fast and similar to that of the WT DNA (Figure 6). In contrast, the annealing rate with the B-stem variant is significantly reduced. The fast component dominates annealing (i.e. f∼1), thus making determination of k_-1 and k₂ rates less reliable than in the absence of NC. The effects of TAR DNA mutations on the elementary annealing rates are summarized in Figure 7. Relative to the reactions performed in the absence of NC, saturating amounts of NC resulted in 100-fold faster k₁ for WT, Loop, as well as T-zip mutant TAR DNA. In contrast, k₁ was ∼6-fold slower for the B-stem mutant (Figure 7(a)). The intermediate complex dissociation rate, k_-1, was similar for all the TAR DNA constructs, with a slightly faster dissociation measured for the B-stem variant (Figure 7(c)). In part as a result of this, but largely due to the differences in k₁, the reaction intermediate was an order of magnitude more stable for the WT, Loop, and T-zip TAR constructs relative to the B-stem variant (Figure 7(b)). These results are consistent with an annealing pathway that involves nucleation via the stem ends. The B-stem mutation introduces non-complementarity between DNA and RNA in this region, thereby making intermediate association slower, and dissociation faster at this nucleation site. The strand exchange rate, k₂, is similar for WT, Loop mutant, and B-stem mutant. The ∼10-fold slower exchange rate determined for the T-zip mutant may be due to stabilization of the stem by introduction of 3 additional bp (Figure 7(d)). Taken together, these data are consistent with WT TAR RNA/DNA annealing via the zipper pathway in the presence of saturating amounts of NC.

Figure 6.

Mutational analysis of full-length TAR RNA/DNA annealing in the presence of saturating amounts of NC. Percent TAR RNA (15 nM) annealed to various TAR DNA constructs (45 nM), as indicated on each curve (WT, Loop, T-zip or B-stem, see Figure 1), as a function of time at 37 °C in the presence of a 4:1 nt:NC ratio. Lines represent single-exponential fits of the data.

Figure 7.

Comparison of elementary rates and intermediate dissociation constant of WT and mutant TAR DNA constructs in the presence of saturating concentrations of NC. Elementary rates (k₁, panel a; k_-1, panel c; k₂, panel d) and intermediate dissociation constants (panel b) were obtained by using equations (2-5) to analyze the annealing time courses presented in Figure 6 for TAR RNA (15 nM) annealed to various TAR DNA constructs (45 nM) at 37 °C and a 4:1 nt:NC ratio.

Effect of varying NC concentration on WT and mutant TAR DNA annealing

To examine the NC-induced TAR RNA/DNA annealing pathway switch in more detail, the annealing of 15 nM TAR RNA to 150 nM WT or mutant TAR DNA was measured in the presence of varying concentrations of NC (Figure 8). The latter is presented as fractional NA saturation with NC, Θ_NC. Under these low salt conditions (20 mM Na⁺ and 0.2 mM Mg²⁺) all added NC is bound to NA, and Θ_NC can be calculated as

Figure 8.

Annealing of WT and mutant TAR DNA to TAR RNA in the presence of variable amounts of NC. TAR RNA (15 nM) was annealed to 150 nM of WT TAR DNA (a), Loop mutant TAR DNA (b), or B-stem mutant TAR DNA (c) as a function of time in the presence of various concentrations of NC in low salt (20 mM Na⁺, 0.2 mM Mg²⁺). Fractional NA saturation with NC, Θ_NC, calculated according to eq. (6), is indicated for each curve. Lines represent double-exponential fits of the data.

$${\Theta }_{\text{NC}}=(\text{NC:nt})\times 6,$$ (6)

where NC:nt is the protein to nt concentration ratio and assuming that 1 NC binds to 6±1 nt at saturation ^{11; 22-24; 36; 46-49}. As expected, at saturation (i.e., Θ_NC ≥ 1), the annealing in the presence of the Loop mutant DNA (Figure 8(b)) is very fast and similar to that of WT (Figure 8(a)), whereas annealing with the B-stem variant is slightly suppressed (Figure 8(c)). Further decrease in Θ_NC (Θ_NC = 0.75) leads to a significant reduction of the annealing rate for both B-stem and the Loop mutants compared to WT DNA, with a more pronounced effect on the latter. These results suggest that under slightly sub-saturating concentrations of NC, full-length TAR RNA/DNA annealing occurs with comparable efficiency via both kissing and zipper pathways, with annealing via the stem ends slightly favored. This is consistent with the preferential annealing of WT TAR RNA/DNA via the zipper pathway under saturated NC binding conditions.

The dissociation constant for the annealing intermediate, as well as elementary rates for the two-step reaction, determined based on the data shown in Figure 8, are presented in Figure 9. The elementary rates could only be determined for Θ_NC ≥ 0.375, since at lower Θ_NC the annealing is extremely slow and has a negligible fast component. In addition, k₂ can not be determined at a Θ_NC > 0.75, since the annealing is dominated by the fast component. Figure 9(a) shows that the B-stem mutant has the slowest association rates over the range of 0.375 ≤ Θ_NC ≤ 1.5. The Loop mutant has a slightly faster k₁ than the B-stem mutant but remains slower than the WT. While k₁ is very fast for all the TAR DNA constructs tested at high Θ_NC, WT TAR DNA requires less NC to reach its maximum k₁. These results are consistent with the existence of multiple annealing pathways, with a greater contribution from the zipper pathway when Θ_NC ≥ 0.4.

Figure 9.

Comparison of elementary rates (panels a, c, and d) and intermediate dissociation constants (panel b) for annealing of WT and mutant TAR DNA hairpins to TAR RNA in the presence of variable amounts of NC. Fractional NA saturation with NC, Θ_NC, was calculated according to eq. (6). Parameters were obtained using equations (2-5) for the analysis of the annealing time courses presented in Figure 8.

There is no significant difference in the intermediate dissociation rate, k_-1, between WT TAR DNA and the two mutant constructs (Figure 9(c)), suggesting that dissociation of the intermediate formed with either the Loop or B-stem variants is comparable in the range of Θ_NC studied here. The same conclusion holds for k₂, which also appears to be independent of Θ_NC (Figure 9(d)). Based on the known duplex destabilizing ability of NC, one would expect k₂ to increase with increasing NC concentration. However, we hypothesize that this dependence is most likely obscured by the NC-induced annealing pathway switch from the extended loop-loop kissing pathway to the zipper pathway at Θ_NC ∼0.4. This switch is not abrupt; both pathways contribute to annealing, with the zipper contribution taking over completely upon NC saturation.

Role of NC-induced nucleic acid aggregation

The major effect of increasing NC concentration is to enhance the bimolecular association step of annealing (see Figure 9(a)). This, in turn, results in strong stabilization of the annealing intermediate (Figure 9(b)). To examine the role of NC-induced NA aggregation in bimolecular intermediate stabilization, a sedimentation assay was performed to measure the fraction of aggregated full-length TAR RNA molecules, f_a, as a function of fractional NC binding under conditions similar to those of the annealing assays. Figure 10 shows that the amount of NC required to stabilize the reaction intermediate (continuous line, f(Θ_NC)) is virtually equivalent to the amount needed to induce NA aggregation (dotted line, f_a(Θ_NC)). These data are consistent with the results obtained for mini-TAR³⁰, and provide further support for the notion that NC-induced NA aggregation constitutes a major component of NC's chaperone function.

Figure 10.

Fraction of fast annealing component (f) and aggregated fraction of TAR RNA (f_a) as a function of fractional NA saturation with NC. f was obtained from the analysis of the data presented in Figure 8 using eqs. (2-5). f_a was obtained by sedimentation assays as described in the Materials and Methods carried out under the same solution conditions as the annealing experiments.

Summary and Conclusions

In this work, the mechanism of 59-nt TAR RNA/DNA hairpin annealing was examined as a function of NA saturation with HIV-1 NC. Under all solution conditions tested, TAR hairpins anneal via the two-step pathway described by eq. (1), involving formation of a bimolecular annealing intermediate in pre-equilibrium to the subsequent strand exchange step resulting in the fully annealed RNA/DNA duplex. Interestingly, the nature of the annealing intermediate, i.e., the annealing pathway, changes as more NC is added. DNA mutational analysis showed that in the absence of NC or at low NC binding levels, the TAR RNA/DNA annealing proceeds via an extended loop-loop kissing intermediate (see Figure 1, lower left). In contrast, when levels of NC binding approach saturation, TAR RNA/DNA anneal through the stem ends (see Figure 1, lower right). Both pathways contribute to annealing at intermediate levels of NC binding (see Figure 9 (a) and (b)).

We hypothesize that the existence of two competing pathways for annealing of the TAR RNA/DNA hairpins is due to the comparable stabilities of the two 4-bp helical segments in TAR RNA (see Figure 1, top left). In particular, based on theoretical predictions, the 4-bp helix adjacent to the hairpin loop is ∼1 kcal/mol more stable than the 4-bp helix proximal to the 5′/3′ termini ^{50; 51}. While melting of the more stable helix is less probable, it allows for formation of a 17-bp extended loop-loop intermediate (Figure 1, lower left). In contrast, melting of the terminal 4-bp helix leads to formation of an RNA/DNA intermediate containing only 9 bp. Whereas the former pathway is plausible in the absence of NC, in the presence of saturated NC, which stabilizes both intermediates, the annealing switches to the pathway that requires the least structure pre-melting, i.e. the zipper pathway.

Previous studies of the annealing of mini-TAR RNA/DNA were also consistent with two annealing pathways, with the loop pathway dominating in the absence of NC ³⁰. However, in case of the shorter hairpin, which only contains two helical regions of similar stability, the stem pathway does not dominate even at saturated NC ³⁰.

A recent study examined the annealing of small complementary DNA hairpins, PBS(+) and PBS(-), representing the rate-limiting step of plus-strand transfer during reverse transcription ⁵². A truncated version of NC, NC(12-55), which lacks the N-terminal basic domain and is therefore a poor aggregating agent, was also able to induce a pathway switch when compared to the reaction performed in the absence of NC ⁵². Another study from the same group examined the effect of NC(12-55) on the annealing of TAR DNA to its DNA complement cTAR ⁴¹. As in the PBS annealing study⁵², NC(12-55) stimulated the TAR annealing reaction by ∼100-fold. This is lower than the ∼10³–fold and ∼10⁵–fold stimulation observed for mini-TAR ³⁰ and TAR annealing (this work), respectively, using full-length NC(1-55). Indeed, we have shown that the duplex destabilizing and minor aggregation activities of NC(12-55) each contribute ∼10-fold to the overall rate enhancement (M.-N. Vo, M. Mitra, I. Rouzina, and K. Musier-Forsyth, manuscript in preparation). Taken together, these results support a major role for NC-induced nucleic acid aggregation in its chaperone activity.

NC modulates the annealing pathway of structured complementary NA via both of its major activities, NA aggregation and destabilization ^{4; 26; 30}. In vivo these diverse capabilities of NC may ensure efficient obligatory and random strand transfer events during reverse transcription ^{53; 54}. The major effect of NC on the efficiency of strand transfers may also explain the increased production of full-length RNA transcripts during in vitro reverse transcription ³⁹, as well as improved strand displacement synthesis of the central DNA flap ⁵⁵ and greater reverse transcriptase processivity ⁵⁵, which are both observed under conditions that also promote NC-induced NA aggregation ^{39; 55}.

Materials and Methods

Protein, RNA and DNA preparation

The NC protein used in this work was prepared by solid-phase synthesis as described ⁴⁰, and its purity was estimated to be > 95% by SDS-polyacrylamide gel electrophoresis. The concentration of NC was determined by measuring its absorbance at 280 nm and using ε₂₈₀ = 6050 mol^-1 cm^-1.

All DNA oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA). Full-length TAR RNA was generated by in vitro transcription using a PCR-amplified template encoding the 59-nt TAR sequence downstream of a T7 RNA polymerase promoter. The T7 promoter was appended during the PCR reaction and the sequences of DNA template and primers for the PCR reaction were: TAR DNA template, 5′-GGGTCTCTCTGGTTAGACCAGATCTGAGCCTGGGAGCTCTCTGGCTAACTAGGGAACCC-3′ and its complementary sequence 5′-GGGTTCCCTAGTTAGCCAGAGAGCTCCCAGGCTCAGATCTGGTCTAACCAGAGAGACCC-3′; T7-TAR primer 1, 5′-CTGTAATACGACTCACTATAGGGTCTCTCTGGTTAGACCAG-3′ and TAR primer 2, 5′-GGGTTCCCTAGTTAGCCAGAGAGC-3′. PCR reactions were performed with PfuTurbo DNA polymerase (Stratagene, La Jolla, CA) according to the enzyme manufacturer's protocol using the provided buffer. Briefly, each reaction included 30 cycles of denaturation, annealing and extension at temperatures of 94 °C for 45 sec, 50 °C for 45 sec and 72 °C for 2 min, respectively, followed by one cycle of extension at 72 °C for 10 min. PCR-generated DNA template was extracted with phenol:chloroform:isoamyl alcohol (25:24:1, by volume) and precipitated with ethanol prior to use. The purified DNA template (5 μg per 100 μL reaction) was in vitro transcribed according to standard procedures, followed by DNase treatment to remove the DNA template. All RNA and DNA oligonucleotides were gel purified on 12% (w/v) denaturing polyacrylamide gels, dissolved in diethyl pyrocarbonate-treated water, and stored at −20 °C Full-length TAR RNA was internally [³²P]-labeled via in vitro transcription under similar conditions, except that the GTP concentration was lowered from 4 mM to 1 mM and 17 mCi/ml [α-³²P]-GTP was added to the reaction.

The concentrations of the RNA and DNA oligonucleotides were determined by measuring their absorbances at 260 nm and using the following extinction coefficients: TAR RNA (59-mer), 5.34 × 10⁵ M^-1 cm^-1; TAR DNA (59-mer), 5.65 × 10⁵ M^-1 cm^-1; C27A/C28A TAR DNA (59-mer), 5.74 × 10⁵ M^-1 cm^-1; B-stem TAR DNA (59-mer), 5.65 × 10⁵ M^-1 cm^-1; T-zip TAR DNA (62-mer), 5.80 × 10⁵ M^-1 cm^-1.

Prior to use, all oligonucleotides were refolded in 25 mM Hepes (pH 7.5) and 100 mM NaCl at a concentration 100× greater than that used in the annealing reactions. This was accomplished by incubation at 80 °C for two min, cooling to 60 °C for two min, followed by addition of MgCl₂ to a final concentration of 10 mM and placement on ice.

Annealing assays

For annealing assays, refolded [³²P]-labeled full-length TAR RNA was combined with refolded unlabeled complementary WT or mutant TAR DNA constructs in a solution containing 20 mM Hepes (pH 7.5), 20 mM NaCl and 0.2 mM MgCl₂ (this MgCl₂ concentration comes from the refolding buffer) at 37 °C, except when otherwise indicated.

In the absence of NC, DNA concentration-dependence studies were performed with 50 nM RNA and various concentrations of DNA, as indicated in the figure legends. Reactions were initiated by adding DNA to RNA, followed by incubation in the reaction buffer for the indicated time. Reactions were quenched by placing solutions on ice followed by addition of glycerol to 5% final concentration. Samples were loaded onto 12% SDS polyacrylamide gels (375 mM Tris-HCl, pH 8.8, 0.1% SDS; 19:1 acrylamide/bisacrylamide (w/v)) run at 25 °C in Tris-glycine (25 mM Tris, 250 mM glycine, pH 8.3) running buffer.

In the presence of NC, studies of the DNA concentration-dependence of TAR RNA/DNA annealing were performed by mixing 15 nM RNA with various concentrations of DNA prior to the addition of NC to achieve a 4:1 nt:NC ratio. Thus, final NC concentrations were 0.664 μM, 0.885 μM, 1.11 μM and 1.55 μM in the presence of 30 nM, 45 nM, 60 nM and 90 nM TAR DNA, respectively. Annealing reactions were quenched at the indicated times by incubation with 1% (w/v) SDS and immediate placement on ice for 5 min. Samples were subjected to phenol/chloroform-extraction (2×), followed by addition of glycerol to 5% final concentration and separation on 12% SDS-polyacrylamide gels as described earlier. Gels were visualized using a Bio-Rad Molecular Imager FX and quantified with Bio-Rad Quantity One Software. NC concentration-dependence studies were carried out similarly with 15 nM RNA and 150 nM DNA in the presence of various concentrations of NC, as indicated in the figure legends.

Sedimentation/aggregation assays

Refolded [³²P]-labeled full-length TAR RNA (15 nM) was combined with TAR DNA (45 nM) in a solution containing 20 mM Hepes, pH 7.5, 20 mM NaCl and 0.2 mM MgCl₂. Upon addition of NC to final concentrations of 0.1, 0.2, 0.5, 1.0, or 2.0 μM, reactions (40 μL) were incubated at 37 °C for 30 min. At the end of the incubation period, solutions were centrifuged at 12,000 rpm in a microcentrifuge for 20 min. Supernatant (5 μL) was collected and analyzed by scintillation counting. The percent radioactivity remaining in the supernatant relative to the RNA-only sample (set to 100%) was plotted as a function of NC concentration.

Abbreviations

TAR

trans-activation response element

NC

nucleocapsid protein

bp

base pair

NA

nucleic acid

nt

nucleotide(s)

ss

single-stranded

ds

double-stranded

WT

wild-type