Targeted binding of nucleocapsid protein transforms the folding landscape of HIV-1 TAR RNA

Micah J. McCauley; Ioulia Rouzina; Kelly A. Manthei; Robert J. Gorelick; Karin Musier-Forsyth; Mark C. Williams

doi:10.1073/pnas.1510100112

Edited by Steven M. Block, Stanford University, Stanford, CA, and approved September 17, 2015 (received for review May 22, 2015)

Significance

The nucleocapsid protein (NC) of the human immunodeficiency virus type 1 (HIV-1) is critical for HIV-1 replication. NC is required for reverse transcription, in which the viral single-stranded RNA genome is converted into double-stranded DNA for replication in the cell. One critical step of reverse transcription that requires NC involves transient destabilization of the transactivation response (TAR) RNA hairpin into a stable double-stranded RNA–DNA hybrid structure. It is not clear how NC is able to destabilize TAR RNA without destabilizing the resulting double-stranded structure. This work shows for the first time to our knowledge that NC binding to specific defects and sequence contexts in TAR RNA dramatically alters the unfolding landscape, preferentially destabilizing TAR RNA relative to other structures to facilitate reverse transcription.

Abstract

Retroviral nucleocapsid (NC) proteins are nucleic acid chaperones that play a key role in the viral life cycle. During reverse transcription, HIV-1 NC facilitates the rearrangement of nucleic acid secondary structure, allowing the transactivation response (TAR) RNA hairpin to be transiently destabilized and annealed to a cDNA hairpin. It is not clear how NC specifically destabilizes TAR RNA but does not strongly destabilize the resulting annealed RNA–DNA hybrid structure, which must be formed for reverse transcription to continue. By combining single-molecule optical tweezers measurements with a quantitative mfold-based model, we characterize the equilibrium TAR stability and unfolding barrier for TAR RNA. Experiments show that adding NC lowers the transition state barrier height while also dramatically shifting the barrier location. Incorporating TAR destabilization by NC into the mfold-based model reveals that a subset of preferential protein binding sites is responsible for the observed changes in the unfolding landscape, including the unusual shift in the transition state. We measure the destabilization induced at these NC binding sites and find that NC preferentially targets TAR RNA by binding to specific sequence contexts that are not present on the final annealed RNA–DNA hybrid structure. Thus, specific binding alters the entire RNA unfolding landscape, resulting in the dramatic destabilization of this specific structure that is required for reverse transcription.

The transactivation response (TAR) RNA hairpin is a 59-nt sequence in the long-terminal repeat (LTR) of the HIV-1 genome that forms a 24-bp hairpin (Fig. 1A) (1). This structure is essential in promoting viral transactivator protein (Tat)-mediated transcription. The protein–RNA complex further enhances LTR promoter activity (2). The highly stable TAR hairpin structure that stimulates viral RNA transcription becomes a liability during the early stage of a new infection, as TAR hairpins inhibit the minus-strand transfer step required for reverse transcription (1). To alleviate this inhibition, successful reverse transcription requires a key viral chaperone, the nucleocapsid (NC) protein. In vitro experiments have shown a 3,000-fold stimulation of the rate-limiting step of minus-strand transfer in the presence of NC (3), as NC is required to destabilize TAR RNA and the complementary repeat TAR DNA hairpin to allow subsequent strand annealing (1).

Fig. 1.

Probing the interaction of the TAR RNA hairpin and NC. (A) The 59-nt sequence and predicted secondary structure of HIV-1 TAR RNA include 24 bp. (B) NC consists of two zinc fingers and a basic N terminus. Basic residues are shown in blue and acidic residues in red, and black denotes zinc-coordinating amino acids. Aromatic residues Phe and Trp are marked. (C) An optical tweezers experiment tethers a single RNA hairpin between two beads through long DNA handles. The micropipette is translated to increase the tension. (D and E) Control constructs excluding the hairpin show an elastic response typical of DNA for three cycles of extension/release (solid/dotted green lines). Experimental constructs incorporating TAR RNA hairpins reveal the same elastic response as the DNA handles until interrupted by sudden hairpin opening at ∼12 pN (solid/dotted blue lines for TAR and solid/dotted red lines for TAR with NC). Thick solid lines are fits to polymer elasticity models from SI Materials and Methods, section 1. Unfolding is characterized by a measured force (F_op) and length increase (Δx_op), and the hairpin closing force (F_cl) is identified where the release and extension data overlap. The shaded region represents the net work done by the instrument to open the hairpin (W_op) and represents the energy required to extend the handles as the hairpin remains folded (ΔW_d) minus the energy required to extend the handles and the unfolded hairpin construct over the same extension range (ΔW_d+r). (F) Histograms of measured opening lengths for pulling rates shown in pN/s. (G) At each pulling rate the average length (solid symbols) shows a variability, which disappears when corrected for polymer elasticity (open symbols), giving the number of bases unfolded: n = 47.8 ± 1.3 for the TAR hairpin and N_NC = 48.4 ± 0.5 in the presence of NC.

HIV-1 NC is only 55 aa long, consisting of two highly conserved CCHC zinc fingers and a basic N terminus (1) (Fig. 1B). The multiple roles of NC during reverse transcription all use the same “chaperone” activity (1), which describes HIV-1 NC’s ability to facilitate the rearrangement of nucleic acids into the most stable structures, with the lowest free energy (1). This chaperone activity is characterized by nucleic acid aggregation, duplex destabilization, and rapid kinetics of protein–nucleic acid interactions (3, 4). Aromatic residues in each zinc finger stack with single-stranded nucleic acid bases, resulting in preferential binding to single-stranded nucleic acids (5). Recent studies suggest that preferred sites for NC-induced destabilization involve guanine-containing base pairs at the boundary with defects in the duplex, which include mismatches, loops, and bubbles (6 ⇓–8). However, the contribution of this localized RNA structure destabilization by NC to the facilitation of RNA unfolding and refolding as well as the magnitude of the destabilization remains unclear.

In this work we use single-molecule optical tweezers (OT) to force-unfold the TAR RNA hairpin and characterize the energy landscape of hairpin unfolding. We show that NC dramatically alters the TAR unfolding transition state position and energy. Our results quantify the destabilization induced by specific NC binding to a limited number of paired guanine bases located at the boundaries of TAR stem defects. Such binding effectively creates larger loops in the already interrupted TAR RNA secondary structure, shifting the transition state position and significantly increasing the spontaneous RNA opening probability. In addition to quantifying NC-induced changes to the TAR unfolding landscape, this work provides a previously unidentified case study of a protein that specifically destabilizes particular elements of nucleic acid secondary structure during the unfolding process.

Results

TAR Hairpin Unfolding Is a Two-State Process.

To quantify the effect of NC on TAR RNA hairpins, the 59-nt hairpin (Fig. 1A) was ligated to long DNA handles and tethered in a dual-beam optical tweezers apparatus in both the absence and presence of NC (Fig. 1 B and C). For constructs where the TAR hairpin was omitted, cycles of extension and release exhibit the smooth curvature characteristic of DNA handle elasticity (Fig. 1 D and E) over a fixed pulling rate. When TAR hairpins are incorporated, extension data show discrete length increases due to force-induced hairpin unfolding. The onset of hairpin opening and the completion of hairpin refolding are determined for each cycle (the onset of refolding at very low forces could not be reliably determined).

Unfolding is characterized by a single length increase of ∼22 nm for nearly all cycles (Fig. 1F). Constant force experiments have shown weak intermediate states for shortened TAR hairpins, but these are not observable in our rapid force-ramp experiments (9). Correcting for the force-dependent elasticity of the stretched construct, as described in SI Materials and Methods, section 1, and for the finite width of the folded stem of 2 nm, as seen for other RNA hairpins (10), gives a corrected unfolded hairpin length, N. The average number of bases unfolded is 47.8 ± 1.3 bases for TAR alone and 48.4 ± 0.5 bases in the presence of NC (this length is compared with the full hairpin length below). In our experiments, TAR refolding during release, especially at higher force-ramp rates, is much more variable than unfolding during stretching and may not occur as a two-state process (11).

NC Reduces the Equilibrium Free Energy of TAR RNA Hairpin Unfolding.

During the force-extension cycle of an OT experiment, the measured work of unfolding (W_op) consists of the difference between the integrated work required to extend the folded construct (ΔW_d) minus the work required to extend an unfolded hairpin construct (SI Materials and Methods, section 1) over the same extension range (ΔW_d+r) (Fig. 1D): W_op = ΔW_d – ΔW_d+r (12). An alternate approach considers only the net work performed by the instrument across the unfolding transition as the sum of the free energy of base pair opening and the entropy required to extend the open RNA hairpin and change the DNA handle extension (13). Both approaches yielded the same result within uncertainty. Similar calculations determine the work obtained through folding, where error introduced by the presence of intermediately folded states should be less than the uncertainty of these experiments. Probability distributions of measured unfolding/folding energies, P_op(W) and P_cl(W), are shown for varying pulling rates (Fig. 2 A and B), and the measured work clearly varies with pulling rate. According to the Crooks fluctuation theorem, the intersection of the opening and closing distributions of measured work reflects the free energy of unfolding (W = ΔG_o) (14). This result is verified to be independent of the pulling rate, within uncertainty (Fig. 2C). Averaging over all pulling rates gives a final result of ΔG_o = 43.3 ± 0.9 k_BT for TAR unfolding in the absence of NC (matching the equilibrium result above) and ΔG_o, _NC = 28.3 ± 0.9 k_BT in the presence of NC. Alternative methods of deducing the free-energy change (15, 16) are shown in SI Materials and Methods, section 2 (Table S1 and Fig. S1), although the key result, the destabilization of the hairpin upon the addition of NC, is the same for all approaches within uncertainty.

Fig. 2.

Equilibrium energies of hairpin opening. (A) Normalized probability densities of measured work during unfolding [P_op(W), blue] and closing [P_cl(W), cyan] for the TAR RNA hairpin over pulling rates of 30 pN/s, 10 pN/s, and 0.8 pN/s (n = 250 opening events). Solid lines are fits to Gaussian distributions to guide the eye. Distributions cross at the equilibrium work (W = ΔG_o), marked by solid circles. (B) Distributions of work done during unfolding (red) and folding (pink) in the presence of NC (n = 162 cycles). (C) TAR (blue) and TAR with NC (red) unfolding free energies for distributions shown in A and B. Averaged over all rates, the measured free energy and SE are ΔG_o = 43.3 ± 0.9 k_BT in the absence and ΔG_o,NC = 28.3 ± 0.9 k_BT in the presence of NC. Details of the free-energy measurement are discussed in SI Materials and Methods, section 2.

View this table:

Table S1.

Free-energy measurements of TAR RNA hairpin unfolding

Fig. S1.

Bennett’s analysis of TAR RNA hairpin unfolding. Plots of Eq. S5 at pulling rates of 30 pN/s, 10 pN/s, and 0.8 pN/s for TAR hairpins alone (blue) and in the presence of 50-nM NC (red). The intersection of z_cl(x) − z_op(x) with the line z = x/k_BT gives the estimate of the energy of unfolding at that rate, and these are shown in Table S1. Averaged, these values give ΔG_o = 44.2 ± 1.6 k_BT for TAR hairpins and ΔG_o,NC = 30.3 ± 0.7 k_BT after NC is added.

NC Shifts the Transition State of Hairpin Unfolding.

Force probability distributions P(F_op) are fitted across all pulling rates to the dynamic force spectroscopy model of Dudko et al. (17), which is described in SI Materials and Methods, section 3 (Fig. 3A and Figs. S2 and S3). Fits to this model determine rates of hairpin opening in the absence of force (kopo) and the distance to and height of the transition barrier (xop† and ΔGop†). TAR RNA alone exhibits xop† = 9.9 ± 1.1 nm, corresponding to roughly half of the hairpin opening length, whereas ΔGop† = 27.0 ± 2.2 k_BT and kopo = (8 ± 5) × 10⁻⁹ s⁻¹. Surprisingly, at the highest pulling rate, addition of NC increases the observed hairpin unfolding force, suggesting initially that NC might stabilize the hairpin (Fig. 3B). However, these effects are associated with a shorter distance from the folded to the transition state, xop,NC† = 4.8 ± 0.6 nm. In other words, the opening transition rate is less facilitated by force due to the shorter TAR RNA elongation required to reach the transition state in the presence of NC protein. Furthermore, ΔGop,NC† = 14.3 ± 1.3 k_BT in the presence of NC is reduced to almost half of its value without NC. Finally, the rate of hairpin opening in the absence of force increases ∼10,000-fold in the presence of NC to kop,NCo = (1.2 ± 0.8) × 10⁻⁴ s⁻¹. The short transition state distance for TAR is unusual for a long hairpin, and the still shorter distance in the presence of NC is striking.

Fig. 3.

Kinetics and thermodynamics of TAR unfolding and folding. (A) Normalized TAR RNA opening probabilities P_op(F) recast as points for pulling rates of 30 pN/s, 10 pN/s, and 0.8 pN/s (dark blue, blue, and cyan) globally fitted to Eq. S7 (solid lines) described in SI Materials and Methods, section 3, and including residuals. Standard histogram bin uncertainties are omitted for clarity, and fits shown are for shape factor ν = 0.5, as discussed in the text. Averaged fitted parameters (with SE) were found: Δxop† = 9.9 ± 1.1 nm, ΔGop† = 27.0 ± 2.2 k_BT, and kopo = (8 ± 5) × 10⁻⁹ s⁻¹ for TAR with n = 250 and χυ2 ∼ 1 for all fits. (B) In the presence of 50 nM NC, global fits over 30 pN/s, 10 pN/s, and 0.8 pN/s (dark red, red, and pink) yield Δxop,NC† = 4.8 ± 0.6 nm, ΔGop,NC† = 14.3 ± 1.3 k_BT, and kop,NCo = (1.2 ± 0.8) × 10⁻⁴ s⁻¹, where n = 162 and χυ2 ∼ 0.8 for all fits. Histogram bin widths were scaled for direct comparisons between A and B. (C) TAR RNA hairpin opening rate as a function of unzipping force, k_op(F), omitting/including NC, calculated according to Eqs. S9 and S10 in SI Materials and Methods, section 4. Solid lines represent the fit to P_op(F), and dotted lines are direct fits of k_op(F) to Eq. S6. The full list of fitted values is compared in Table S2.

Fig. S2.

Stretching force rate dependence of P_op(F) and P_cl(F). (A) Normalized probability of opening/closing (blue/cyan) vs. force for TAR RNA (n = 250) and (B) distributions of opening/closing (red/pink) for TAR in the presence of 50-nM NC (n = 162). Opening forces are fitted in the main text and shown in Fig. 3.

Fig. S3.

Fitting minima for opening force data, P_op(F). Values of χ² for the fits of Eq. S7 to the data of P_op(F), for the shape factor v = ½. Variations about the minima for (A) TAR RNA and (B) TAR with NC, for each free parameter, with the other two held fixed. Minimized values are shown in Table S2, along with fitted uncertainties determined from the parameter values that increase χ² by 8, which represents the 95% confidence interval for three fitting parameters.

Complementary analysis of the kinetics based on the cumulative probability of unfolding yielded the unfolding rate as a function of force, k_op(F) (18). This is reported in the presence and absence of NC, according to SI Materials and Methods, section 4. Force-dependent opening rates for the three pulling rates agree well with each other (Fig. 3C) and can be universally fitted to yield transition state parameters similar to those discussed above within uncertainty (Table S2). The kinetics of hairpin closing, however, do not agree across the pulling rates, likely due to the presence of intermediates (Fig. S4) (19). The main effect of NC is to lower the transition state energy and move it closer to the closed hairpin state, a result well outside the uncertainty of these fits.

View this table:

Table S2.

Transition state TAR RNA hairpin unfolding properties

Fig. S4.

Stretching force dependence of rates k_op(F) and k_cl(F). (A) k_op(F) (solid symbols) and k_cl(F) (open symbols) for TAR RNA pulled at rates of 30 pN/s, 10 pN/s, and 0.8 pN/s (dark blue, blue, and cyan), calculated from the data of Fig. S2, using Eqs. S9 and S10. (B) k_op(F) (solid symbols) and k_cl(F) (open symbols) for TAR RNA with 50-nM NC pulled at rates of 30 pN/s, 10 pN/s, and 0.8 pN/s (dark red, red, and pink). In both cases the opening rate is pulling rate independent while the closing rate depends on the pulling rate, reflecting multistate TAR RNA closing kinetics. Fits of Eq. S3 to the opening data are discussed in the text.

mfold Quantifies Destabilization of TAR Due to Specific NC Binding.

mfold provides a theoretical estimate of the overall energy of hairpin folding and the energy per base pair for a given sequence (20). The free-energy profile for opening n nucleotides of TAR RNA at zero force was calculated as a sum of unzipping free energies for the corresponding elements of the TAR structure, using mfold energies per base pair, G(n_i, F = 0) (Fig. 4A). Calculations were performed at standard conditions approximately equivalent to our experimental conditions, as discussed in SI Materials and Methods, section 6. In addition to the hairpin unfolding energy, further insight can be derived by subtracting the mechanical work F_1/2⋅x done by the applied force, to find the landscape, G(x_i, F_1/2), where F_1/2 is the critical force at which folded and unfolded states have equal free energy (Fig. 4B) (10, 21). For the full TAR RNA sequence, F_1/2 = 10.1 ± 0.1 pN, after correcting for single-stranded RNA (ssRNA) elasticity during unfolding and the 2-nm width of the hairpin stem.

Fig. 4.

Transition state predictions for the TAR hairpin in the absence and presence of NC. (A) Theoretical energy profiles determined from mfold. Horizontal and vertical lines indicate Δx_op and ΔG_o for TAR RNA (blue), for the lowest energy state seen in these experiments, as discussed in the text. (B) Free-energy profiles at F_1/2, where folded and unfolded state free energies are equal, with corrections for ssRNA elasticity and added potentials for open and closed states and probability distributions for the two states (green) (SI Materials and Methods, section 5). Horizontal and vertical lines indicate theoretical estimates of xop† and ΔG1/2† for TAR RNA (blue): F_1/2 = 10.1 ± 0.1 pN, xop† = 11.3 ± 0.9 nm, and ΔGop† = 30.7 ± 2.0 k_BT. (C and D) Predicted destabilization of TAR in the presence of NC (for four bound NCs). The resulting landscapes (red) give F_1/2,NC = 7.9 ± 0.1 pN, xop,NC† = 4.7 ± 0.9 nm, and ΔGop,NC† = 14.8 ± 1.2 k_BT. (E) Potential NC binding sites located at defect-adjacent G-containing base pairs are circled in red for four test cases. Sites marked with solid circles were uniformly destabilized by a total δG_o = 16 k_BT. (F) Values of the transition state location and height calculated from mfold for each case shown in E. A 2D Z-test of the data with the various models gives a probability of 0.74 for the four-site model and 0.08 for the five-site model, whereas the three- and six-site models each have probability of less than 0.001.

Our experimentally determined value of the unfolding energy for TAR RNA alone, 44 k_BT, is significantly lower than the value calculated in mfold, 56 k_BT. Similarly, only 48 bases of an expected length of 59 bases were observed to open during experimental unfolding studies, as noted above. A consistent explanation for both observations is found in the calculated TAR RNA G(x_i, F_1/2) profile, which shows a deep free-energy minimum at an extension of ∼5 nm corresponding to the lowest C bulge (Fig. 4B). We also show the calculated occupancy probability, smoothed to represent the system elasticity. The transition is predicted to be primarily between a partially frayed state and the unfolded state, as seen for other hairpins (10, 21). The energy landscape suggests a separate unfolding step at low extension, which is not observed, although the landscape near the fully folded state may not be precisely represented by the model. Theoretical values of the unfolded and transition states are determined to be ΔG_o = 41.5 ± 1.5 k_BT, xop† = 11.3 ± 0.9 nm, and ΔGop† = 30.7 ± 2.0 k_BT, where the uncertainties are primarily due to uncertainty in the location of the frayed state. The unfolding length of Δx = 21.8 ± 0.5 nm, determined at F_1/2, is in good agreement with the unfolding length at the transition force measured in the OT experiments.

Previous studies have shown a preference for NC binding to G residues located adjacent to local defects, such as G⋅U wobble pairs, bulges, mismatches, and loops (6, 7). All such potential sites in TAR RNA are circled in Fig. 4E. In our model, we reduced the free energy of these potential sites by fixed amounts, δG_NC(n_i), and observed the resulting free-energy profiles, G_NC(n_i, F = 0) and G_NC(x_i, F_1/2). To match the experimental difference measured in OT experiments, the total destabilization was required to match that observed in experiments within uncertainty. Next, we looked for the combination of NC-induced destabilization that would recover the measured properties of the transition state, especially the opening distance to the transition, xop†. Spreading evenly the net destabilization over all seven potential sites, C⁷⋅G⁵⁴, C⁹⋅G⁵², G¹²⋅C⁴⁹, G¹⁶⋅C⁴⁵, C¹⁸⋅G⁴⁴, G²⁶⋅C³⁹, and C²⁹⋅G³⁶, did not lead to the correct position of the transition state. Extensive variation of δG_NC(n_i) shows the loop-adjacent site C²⁹⋅G³⁶ is not destabilized by NC. Presented in Fig. 4 are selected mfold models with six, five, four, or three sites each destabilized by equal amounts (solid lines denote destabilized sites). Although it is possible that different sites may show varying amounts of destabilization, these details cannot be resolved in our experiments. The energy landscape, G_NC(n_i, F = 0) and G_NC(x_i, F_1/2), is determined for each case, and the example of four NC binding sites is shown in Fig. 4 C and D. For the models in Fig. 4E, Fig. 4F illustrates the results for the calculated transition state parameters ΔGop† and xop† from this analysis. The four-binding-site model is the best match to the measured results of the OT experiments within uncertainty, giving the transition state parameters, xop, NC† = 4.7 ± 0.9 nm and ΔGop,NC† = 14.8 ± 1.2 k_BT, and the equilibrium TAR unfolding free energy in the presence of NC, ΔG_{o, NC} = 25.6 ± 1.5 k_BT. We find a new transition force, F_1/2, _NC = 7.9 ± 0.1 pN, and unfolding length Δx_NC = 20.8 ± 0.5 nm. Comparisons with experiment are made in Fig. 5 and Table S2 and are discussed below. Although there is a small probability associated with the five-NC binding-sites state model (Fig. 4), the following discussion is not altered by this possibility. Finally, this close match to experiment was found only for this very specific combination of δG_NC(n_i) values, in which no destabilization occurs in the part of the TAR stem below this new transition state. Thus, NC destabilizes the TAR RNA hairpin by targeting specific sites on that part of the stem, further reducing the distance from the closed state to the transition state.

Fig. 5.

Specific binding of NC transforms the energy landscape of the TAR RNA hairpin. Shown is a summary of experimental and computational results for TAR RNA hairpin force-unfolding, combining the overall opening length and unfolding free energy with the transition state location into a unified free-energy landscape. Theory and experiment are compared in the absence (cyan/blue) and the presence (pink/red) of 50 nM NC and presented at a common external force of F_1/2,NC = 7.7 pN. Lines are interpolations to guide the eye. Transition state distances and energies are shown relative to a folded state where the bottom part of the stem is frayed as described in the text. The zero-force TAR and TAR + NC data are offset by a qualitative preference of NC for ssRNA (SI Materials and Methods, section 9). Diamonds locate the transition state for TAR RNA and for TAR in the presence of NC (ovals highlight binding sites). NC dramatically enhances the force-free rate of TAR unzipping by specifically destabilizing the interrupted upper part of the TAR stem, thereby shortening the region at the bottom of the TAR hairpin that has to open before reaching the transition state.

SI Materials and Methods

1) Models of Polymer Elasticity.

The lengths of the DNA handles and the unfolded TAR RNA are force dependent. Because unfolding occurs over a range of forces, this elasticity must be modeled to effectively compare the measured lengths (Fig. S1). The DNA construct and the DNA handles of the TAR construct (Fig. 1C) may be characterized by the worm-like chain model of polymer elasticity, where the observed length per base pair of the polymer is dependent on the applied force b(F):b(F)=Bds[1−12(kBTPdsF)1/2+FSds].[S1]DNA is described by an overall contour length (B_ds = 0.340 ± 0.001 nm/bp), a persistence length (P_ds = 30 ± 4 nm), and an elastic modulus (S_ds = 1,200 ± 200 pN). Fig. 1 D and E (in the main text) shows fits to the DNA construct used as a control in these experiments (32).

The unfolded RNA hairpin is described by the freely jointed chain model, a polymer model appropriate for single-stranded nucleic acids, where the measured length of the polymer depends on the applied force b(F):b(F)=Bss[coth(2PssFkBT)1/2−12kBTPssF].[1+FSss].[S2]The unfolded TAR RNA hairpin is characterized by an overall contour length (B_ss = 0.570 ± 0.001 nm/bp), a persistence length (P_ss = 0.8 ± 0.2 nm), and an elastic modulus (S_ss = 800 ± 200 pN). This model could not be fitted independently to force-extension data of the TAR RNA hairpin, as the handles were always present. Eq. S2 has been used successfully to characterize long chains of single-stranded DNA (32). Fig. 1 D and E shows a model that adds the extension of the dsDNA handles and the unfolded TAR RNA hairpin to determine the total construct extension, which is fitted to the measured force-extension curves (FECs), using only the length of the unfolded hairpin as a variable. In practice, this is equivalent to measuring the length change from F_op(dsDNA) to F_op(dsDNA + ssRNA), as shown in Fig. 1D. Elasticity corrections to the unfolded length are shown in Fig. 1F.

2) Comparisons of Unfolding Free-Energy Measurements.

There are several ways to extract the unfolding free energy, ΔG_o, from the work distributions of Fig. 2. The case when the molecular extension and release curves overlap suggests an almost equilibrium molecular unfolding and refolding, which in the case of TAR RNA alone is approached at our slowest force ramp rate of 0.8 pN/s (Fig. 2A). At this pulling rate the average opening and closing forces, F_1/2 = 10.4 pN, are the same, indicating that the hairpin is equally likely to be unfolded or folded. Integrating the measured work at this rate (W_d – W_d+r from Fig. 1) gives the free energy of unfolding, ΔG_o = 43.5 ± 0.5 k_BT. For TAR in the presence of NC, we may estimate F_1/2 = 7.7 pN and ΔG_o,NC = 28.3 ± 0.5 k_BT.

In general, the work measured during extension and release will give probability distributions P_op(W) and P_cl(W) that will vary with the pulling rate (r), as seen explicitly in Fig. 2. It has been shown that the difference between these distributions is closely related to the difference between the measured work and the free energy of the transition (ΔG_o) (14):Pop(W)Pcl(W)=e(W−ΔGo)/kBT.[S3]According to this relation, the value of the work where the distributions are equal should correspond to P_op(W) = P_cl(W) and thus W = ΔG_o, as first shown by Crooks (14). These crossing points are marked in Fig. 2, where Gaussian distributions have been drawn to guide the eye. Averaged over all pulling rates, ΔG_o = 43.3 ± 0.9 k_BT and ΔG_o,NC = 28.3 ± 0.9 k_BT are close to the integrated equilibrium values found above, and all values are compared in Table S1.

An alternate method integrates across the full distribution of the data, to give the free energy as a function of sum of the measured work according to Jarzynski’s equality (15):ΔGo=−kBT⁡ln〈exp(−WkBT)〉.[S4]Although using the full unfolding distribution is advantageous, this technique is sensitive to the existence of statistical outliers that bias the average toward arbitrarily low estimates of the free energy. This is particularly problematic at low pulling rates, where low measured forces are corrupted by drifts in the baseline that are more significant at slower speeds. Values may be reliably determined only at the fastest pulling rates (Table S1), ΔG_o = 44.4 ± 0.6 k_BT and ΔG_o,NC = 29.5 ± 1.0 k_BT.

A final method is especially useful when the distributions have little or even no overlap. The acceptance ratio developed by Bennett shows that the best estimate of the free energy may be found from the nonequilibrium relation of Eq. S3 by multiplying both sides by a specific function f(x) (13, 16). The statistical error is minimized across all rates for the solution:zcl(x)−zop(x)=ΔGokBT;zop(x)=ln〈fx(W)exp(−WkBT)〉op, zcl(x)=ln〈fx(W)〉cl,fx(W)=exp(x2kBT)1+exp(x−WkBT).[S5]To calculate z_op(x) and z_cl(x), measured values of W are averaged across the opening and closing transitions as indicated for each value of x. A plot of the difference z_op(x) – z_cl(x), shown in Fig. 1, will cross the line z = x/k_BT at ΔG_o independent of the pulling rate. Values are shown in Table S1, and the averages are close to the values above: ΔG_o = 44.2 ± 1.6 k_BT and ΔG_o,NC = 30.3 ± 0.7 k_BT.

The summaries of the measured energies in Table S1 show good agreement across all of the techniques. Comparisons between the methods have been made for other hairpins and have quantified the dissipated work in nonequilibrium experiments. Large values of the dissipated work may introduce a systematic bias in the data (13). In this work, the combination of long handles and a soft trap (discussed below) leads to relatively small values of the dissipated work. Crucially, the change in the hairpin stability upon addition of NC is the same within uncertainty for all techniques.

3) Opening Force Distribution Analysis to Obtain Transition State Parameters.

Dudko et al. (17) calculate the average escape time, which is the reciprocal of the average opening rate, of the molecule from its potential well U₀(x) via diffusion in the presence of an unfolding force, to obtain explicit expressions for both P_op(F) and k_op(F) for specific shapes of U_o(x),kop(F)=kopo(1−xop†FΔGop†v)1/v−1⁡exp{ΔGop†kBT[1−xop†FΔGop†v]1/v},[S6]Pop(F)=kop(F)rexp{kopoxop†r−kop(F)xop†r(1−xop†FΔGop†v)1−1/v}.[S7]Here ΔGop† is the height of the opening transition barrier located at a distance xop† from the closed state, with a zero-force rate of thermal unzipping kopo, and ν is a zero-force energy shape factor, which is taken to be either 2/3 or 1/2 for a linear-cubic and cusp free-energy surface, respectively (17). These are the only two shapes of the zero-force opening potential, U_o(x), for which the analytical expressions for k_op(F) and P_op(F) (Eqs. S6 and S7) are available (17). The actual opening free-energy profile for nucleic acid hairpins, and especially for “patterned” or “interrupted” hairpins, like TAR, is not well approximated by either of these two shapes. The positions of the transition barrier in the absence and presence of the force are very different from each other (Fig. 5) and are both determined by the details of the hairpin structure. Therefore, although elongation to the transition state, xop†, is well defined and can be obtained by fitting the experimental k_op(F) and P_op(F) dependencies to [S6] and [S7] irrespective of U_o(x), the fitted values of ΔGop† and kopo will vary somewhat with the choice of U_o(x). The typical practice is to fit P_op(F) (Fig. S2) to [S7] for both values of ν and report the average value, as we do in Table S2 (12). Importantly, adding NC produces equivalent changes to these parameters within uncertainty for both model potentials, providing a reliable determination of the effects of NC on TAR unfolding despite the limitations of the method. Instrumental noise could bias the experiment toward short values of the transition state distance (and corrupt the overall free-energy estimate), so comparing forces across several pulling rates is crucial to validate this result.

In practice, using Eqs. S6 and S7 to fit the experimental k_op(F) and P_op(F) dependencies is difficult, as many combinations of plausible initial guesses for the fitting parameter values lead to either no fit or unreasonable parameter values. To test our results from the fits to Eqs. S6 and S7, we can use a simple explicit relationship between the maximum values of P_op(F), Fopmax and Popmax, and the fitting parameters xop† and ΔGop†, in the formΔGop†kBT=Fopmaxxop†kBT[xop†xop†−PopmaxkBT⋅e(1−v)].[S8]Thus, instead of fitting the complete distributions to Eqs. S6 and S7 we can first use our initial estimates of parameters Fopmax and Popmax to calculate xop† and ΔGop† according to Eq. S8. Although Eq. S8 cannot by itself define both ΔGop† and xop†, choosing one of these parameters yields the other. The choice of xop† that yields the most consistent values of ΔGop† for all force ramp rates (or for all pairs of Popmax and Fopmax parameters at different rates) defines the best initial parameter values. Using this method and fitting P_op(F) for TAR RNA opening distributions at all three rates yield as the most consistent pair of initial parameter values xop† = 10.0 ± 0.5 nm and ΔGop† = 26.5 ± 2.0 k_BT (for ν = 1/2) and xop† = 10.0 ± 0.5 nm and ΔGop† = 17.5 ± 2 k_BT (for ν = 2/3), in the absence of NC. Similar analysis in the presence of 50 nM NC yields the best-fit parameters xop† = 4.5 ± 0.5 nm and ΔGop† = 17 ± 2 k_BT (for ν = 1/2) and xop† = 4.5 ± 0.5 nm and ΔGop† = 11 ± 2 k_BT (for ν = 2/3). The results obtained using Eq. S8 are very close to their final fitted values based on global fitting for both P_op(F) and k_op(F) at all pulling rates, as summarized in Table S2 and in Fig. 5 of the main text (Fig. S3 shows the minimization landscapes). Note that the global fits to Eqs. S6 and S7 were obtained by choosing multiple starting parameters over a large range of values. The robustness of the fit to changes in each parameter is illustrated in Fig. S3, which shows the value of χ² as a function of each parameter while holding the other parameters at their fitted values. The final fitted parameter values presented in Table S2 were obtained by averaging between the values fitted for the two limiting analytical shapes of U_o(x) potentials (with ν = 1/2 and ν = 2/3) to Eqs. S6 and S7. As discussed above, changes in the fitted parameters due to TAR–NC binding are of principal interest for us in the present study.

Importantly, the transition state is not well defined at very low forces because the unfolded state is very high energy and the reaction is not driven under these conditions. The use of force allows us to characterize the unfolding alone and map the unfolding transition onto a single reaction coordinate, which in turn allows us to use mfold theory to determine NC binding and destabilization sites, which are independent of force. In the relevant HIV-1 reverse transcription minus strand transfer reaction the final state is stabilized by the formation of a complete duplex rather than by force. However, the transition pathway for minus strand transfer is likely different from that observed here.

4) Molecular Opening Rates, k_op(F), Are Determined from the Distribution of Opening Forces, P_op(F).

In addition to measurements of the total free energies of RNA opening, our FEC measurements performed at a number of constant force ramp rates can be analyzed to yield the probability distributions of the opening and closing forces, P_op(F) and P_cl(F). For a two-state system, these distributions can be used to calculate the universal pulling rate-independent average molecular opening and closing rates, k_op(F) and k_cl(F), as a function of applied force. Specifically, one can use the opening and closing force distributions to calculate S_op(F) and S_cl(F), the survival probabilities of the closed and opened states, respectively, which are (17, 18)Sop(F)=1−∫FminFPop(F′)dF′Scl(F)=1−∫FFmaxPcl(F′)dF′.[S9]We can further relate the derivative of the survival probabilities and the opening/closing force distributions through the force ramp rate, r = dF/dt, such that we can write explicit expressions for the opening and closing rates through our measured opening and closing force distributions and their corresponding survival probabilities as follows (17, 18):kop(F)=r⋅Pop(F)Sop(F)kcl(F)=r⋅Pcl(F)Scl(F).[S10]Presented in Fig. 3C of the main text are the opening rates k_op(F) calculated from our measured P_op(F), using both Eqs. S9 and S10 for all three force ramp rates, both in the absence and in the presence of NC. As expected, the calculated k_op(F) for all three experimental force ramp rates overlap with each other, resulting in a universal k_op(F) behavior for TAR RNA opening both in the presence and in the the absence of NC. However, the calculated k_cl(F) vary strongly with the pulling rate, due to the non-two-state nature of the TAR RNA closing transition, as discussed in the text and illustrated in Fig. S4. Future studies could elucidate the kinetics of refolding when these intermediates are present, as can now be done with unfolding intermediates (19).

5) The Effective TAR RNA Folded State Is the Frayed State.

It has been shown that the probability density of hairpin end-to-end length, p(x), is related to its opening landscape ΔG(x) at F_1/2 as follows: p(x) ∼ exp(−ΔG(x)/k_BT) (10). For the TAR hairpin, we can plot both the landscape and the probability density, after smoothing that reflects the influence of the DNA handles and normalization. This smoothing effectively blurs the sharper peaks of the probability density and is consistent with the length measurements of Fig. 1 F and G. These smoothed densities are shown in Fig. 3 of the main text for TAR and TAR with NC. In each case only two states can be discerned and these clearly correspond to the closed and open hairpin states, with equal probability associated with each state. Furthermore, the peak probability for the closed state is not centered around zero extension, but at about 5 nm, corresponding to the location of the lowest stem bubble. Physically, this is due to thermally driven fraying of the lowest part of the stem, further destabilizing this region from the features determined by mfold. Such fraying drives a similar shift in the probability density in the model with NC as well. These densities could be refined further, especially for the model with NC, by formally removing the frayed state from the landscape and recalculating F_1/2 = 7.7 ± 0.2 pN (there was no change observed for TAR alone). Although this refinement eliminates the need for probability normalization, it does not significantly improve the final result. Taking into account that the effective TAR folded state is in fact its frayed state with the lower 4-bp stem separated from the rest of the TAR by bulge, we arrive at the mfold-calculated values of ΔGop† and xop† that are entirely consistent with the values following from Dudko’s analysis of our optical tweezers (OT) TAR RNA unfolding data.

6) Solution Conditions of the Landscape Model Calculations Are Equivalent to Experimental Conditions.

The free-energy profile for unfolding n nucleotides of TAR RNA at zero force was calculated using the mfold program at standard conditions of 37 °C and 1.0 M NaCl (20). These conditions appear to be thermodynamically approximately equivalent to our experimental conditions of 23 °C and 0.1 M NaCl. According to recent work by Stephenson et al. (29), decreasing the temperature from 37 °C to 23 °C is expected to stabilize the CD4 RNA hairpin by ∼14 k_BT, whereas at the same time decreasing the solution ionic strength from 1.0 M to 0.1 M is expected to destabilize the same hairpin by ∼15 k_BT. Force spectroscopy studies have also shown that the variations in solution ionic strength and temperature do not affect the position of the opening transition state of the RNA (29) and DNA (21) hairpins, as they uniformly change the stability of all NA structural elements, at least in the first approximation. Other OT results from Vieregg et al. (22) measured ∼15 k_BT destabilization of a shortened TAR hairpin as ionic strength (Na⁺) decreased from 1.0 M to 0.1 M. Both results are in good agreement with known models for nucleic acid stabilities across varying solution conditions based on conventional thermal melting experiments (20). Thus, the effects of lower temperature and salt (23 °C and 0.1 M NaCl) in our experiments relative to the standard conditions of mfold (37 °C and 1 M NaCl) approximately cancel each other within the accuracy of the present calculations. The primary uncertainty in this calculation (which appears to be of minor significance, according to our results) is due to the unknown dependence of specific TAR secondary structure elements on temperature and salt, which may be different from the simplest dependence of base pair stability on these parameters (22, 28).

7) RNA/DNA Construct Preparation.

A double-stranded DNA template encoding the TAR RNA sequence in front of a T7 RNA polymerase promoter was generated by PCR, starting with a 103-nt synthetic DNA oligonucleotide (see Fig. S5 for a summary and placement of all of the oligos). The PCR product was used to prepare the 59-nt TAR RNA hairpin with additional flanking sequences by in vitro transcription using T7 RNA polymerase. The RNA was purified by denaturing 10% (wt/vol) polyacrylamide gel electrophoresis (PAGE) and the 3′ end of the RNA was ligated to a 30-nt DNA oligonucleotide, using T4 RNA ligase 1 (New England Biolabs). Ligation was facilitated by first annealing the 30-mer to a complementary 46-nt DNA that also contained additional 5′ nucleotides to allow ligation to the labeled DNA handles discussed below. Following purification by denaturing 10% PAGE, the 5′ end of the TAR RNA was ligated to a 17-mer DNA oligonucleotide, using T4 RNA ligase 2 (New England Biolabs). The 17-mer contained three ribonucleotides at the 3′ end and was annealed to a complementary 25-nt DNA oligonucleotide before the ligation reaction. Following purification by denaturing 10% PAGE, the resulting TAR RNA containing flanking DNA oligonucleotides (TAR + 30 + 17) was ligated to long DNA handles, which were labeled for bead attachment. The 3,400-bp biotin handle was prepared by PCR amplification of plasmid pBR322, using 5′-biotinylated primer 1 and primer 2 followed by EcoRI digestion and purification on a 0.8% agarose gel. The 3,100-bp digoxigenin handle was prepared by PCR amplification of plasmid pBR322, using primer 3 and 5′-DIG–labeled primer 4 followed by BspEI digestion and purification on a 0.8% agarose gel. Ligation of TAR RNA to the biotin handle was carried out by first annealing the TAR + 30 + 17 to the 46-mer DNA and then ligating to the biotin handle, using T4 DNA ligase. The reaction was extracted with phenol-chloroform and ethanol precipitated before ligation to the digoxigenin handle. For this reaction, the TAR RNA ligated to the biotin handle was annealed together with the digoxigenin handle and the 25-nt DNA oligonucleotide (Fig. S4 and the final product in the OT experiment in Fig. 1). Ligation was then carried out using T4 DNA ligase. The final product was purified on a 0.8% agarose gel. To serve as a control, a construct that ligated the two labeled DNA handles together omitting the RNA hairpin was also created.

8) Conversion Between Constant Pulling Rate and Constant Force Ramp Rates.

In these experiments, our OT instrument controls the end-to-end extension of the construct rather than the force. TAR RNA FECs are obtained by extending the construct at a constant rate dx/dt. However, our analysis of the rate-dependent probability distributions of the opening forces uses the force ramp rate, r = dF/dt = κ·dx/dt, rather than the extension rate. Here κ = dF/dL is the net construct rigidity, which is determined by the rigidity of three elements connected in series: the rigidity of the closed hairpin, κ_NA, the rigidity of the dsDNA linker, κ_dsDNA, and the rigidity of the optical trap, κ_OT:1κ=1κNA+1κOT+1κdsDNA.[S11]The hairpin rigidity is very high, so the first term in Eq. S11 can be always neglected. The third term, κ_dsDNA, depends on the length of the dsDNA linkers, described as a worm-like chain (WLC) rigidity, which is given by (18)1κdsDNA=LdFWLC/d(x/L)=2LβIp(1+FβIp)3+5FβIp+8(FβIp)5/2≈L4F(FβIp)1/2,[S12]where the last expression applies in the high-force regime (Fβl_p > 1 or F > 0.08 pN). At 15 pN, this term is 3.6 nm/pN. In contrast, the OT instrument used here has a stiffness of κ_OT = 0.08 pN/nm or 1/κ_OT = 12.5 nm/pN. Therefore, the force-independent stiffness of the trap dominates the pulling rate. However, there is a small force dependence due to the dsDNA term. All force ramp rates have therefore been calculated directly from the force-extension curves to obtain the actual rate of force change, r=F˙=κ(F)⋅dx/dt, where dx/dt is the pulling rate in nanometers per second and κ(F) is the weakly force-dependent overall stiffness of the construct including the OT.

9) Higher Concentrations of NC Lead to Only Minor Additional Effects on TAR RNA Opening.

The concentration of NC used in our TAR RNA unfolding experiments (50 nM) is below equilibrium dissociation constant values seen previously for NC binding to TAR-polyA (K_d ∼ 250 nM, once adjusted to match the solution conditions in these experiments) (34) and for NC binding to the tRNA^Lys,3 minihelix (K_d ∼ 200 nM, also adjusted for solution conditions) (35). Limited measurements that we have performed in 200 nM NC suggest rather similar effects of 50 nM and 200 nM NC on TAR RNA opening by force. This implies that nonspecific binding of NC molecules to TAR RNA has only a minor effect on TAR stability and opening kinetics, in comparison with the main effect produced by the few specifically bound NCs, which likely have a K_d lower than 50 nM. These specifically bound NC molecules act by destabilizing TAR at a few sites. At the same time, higher NC concentrations would be expected to lead to saturated NC–RNA binding with stoichiometry of roughly 1 NC molecule per 6 nt as well as nonspecific NA aggregation (27). Indeed, in our experiments with higher NC concentration we observed some RNA and DNA aggregation, as well as an increase in the flexibility of the dsDNA handles. To avoid these effects that are unrelated to the main duplex destabilizing function of NC, we used 50 nM NC for the TAR stretching studies presented here.

This combination of specific/nonspecific binding to multiple sites of the folded/unfolded TAR RNA means that although NC binding clearly destabilizes the TAR hairpin and unfolded ssRNA, it is difficult for these experiments to determine the free-energy change of NC binding to TAR RNA. A very rough calculation suggests that for TAR RNA if there are four binding sites, each destabilized by ∼4 k_BT, we may determine a ratio of KD(ssRNA)/KD(TAR)≈exp(4)≈50, so that binding to ssRNA is ∼50 times stronger than binding to the TAR hairpin. However, this ignores nonspecific binding and assumes all specific binding sites are equal. Thus, although the offset shown for the TAR RNA +NC landscape in Fig. 5 is qualitatively correct, we cannot determine the exact value of TAR destabilization upon binding of NC.

Discussion

TAR RNA Hairpin Stability Is Weakened by Multiple Duplex Defects.

Comparing the interrupted sequence of the TAR RNA hairpin stem to regular, “unpatterned” DNA hairpins of similar stem and loop length, which were studied extensively by force-unfolding with OT by Woodside et al. (10), we observe that the overall TAR RNA stability is lower, whereas the equilibrium extension change upon TAR unfolding is the same. Several defects along the stem are responsible for this overall decrease in the TAR stability. These defects also lead to a relatively low value of the measured and calculated transition force of F_1/2 = 10.1 pN, which is typical of much shorter hairpin stems only 10 bp in length. A TAR duplex with no defects (∼24 bp in the stem) would have an expected F_1/2 > 15 pN. A TAR RNA hairpin, modified by the replacement of the lowest 4 bp and the C bulge with a stabilizing G⋅C base pair and excluding the bases below the lowest bulge, was shown to be more stable than the wild-type hairpin studied here, with F_1/2 ∼12 pN or higher, depending on solution conditions (11, 22).

In the absence of protein, the TAR hairpin opening transition state is predicted by our landscape calculations to be located at the G²¹⋅C⁴¹ base pair, the second base pair below the UCU bulge. mfold calculations reproduce this TAR opening length to the transition state, which lies about half way to the unfolded state (Fig. 5). However, regular hairpins with unpatterned stems typically have their opening transition state close to the loop, such that for an ∼24-bp hairpin with a 6-base loop, the distance from the folded state to the transition state would include nearly the entire hairpin length less the length of the loop. For TAR, the low stability of the 4-bp helix separating the terminal loop from the nearby UCU bulge means that this entire region effectively behaves as a giant hairpin loop during TAR force-unfolding (10). The long effective size of this loop shifts the opening state to the midpoint between the closed and opened state of the hairpin. This defect must be also responsible for its faster zero-force opening rate of ∼10⁻⁸ s⁻¹, which is typical for 20-bp hairpins (10). In summary, these mfold-based landscape model calculations reproduce and explain all of the essential measured features of the full TAR RNA hairpin.

NC Targets Specific Locations on the TAR RNA Hairpin.

Comparing our results for HIV-1 TAR RNA unfolding by force in the absence and presence of NC, we conclude that only a few specifically bound NC molecules alter the TAR RNA opening pathway by destabilizing several sites on this interrupted hairpin. Our data allow us to estimate the destabilization at these sites to be 2.4 kcal/mol (∼4.0 k_BT), which implies complete melting of the G–C base pairs shown in Fig. 4, as well as some additional stabilization of the single-stranded state by NC. In general, weakly base-paired G bases were recently shown to be sites for preferential NC binding accompanied by duplex destabilization in 2-aminopurine fluorescence studies (6), in single-molecule FRET studies (23), in SHAPE footprinting studies of HIV-1 NC (8) and MLV NC (24), and in a recent NMR study of MLV NC (25). Our results support this observation while also identifying the specific subset of these sites on TAR that are responsible for altering its unfolding landscape. The observed specific sites of NC-induced duplex destabilization are not the most energetically favorable, which should instead correspond to G-rich single-stranded regions bordered by duplexes (8, 26, 27), such as those found in the major TAR loop. Although there is likely NC binding to the TAR loop, our results show that such binding has negligible effect on TAR stability. The differences in NC–RNA binding and structure destabilization reflect the two independent NC functions in viral RNA: selection and packaging vs. nucleic acid refolding during reverse transcription. The ability of NC to bind selectively to very specific sites during packaging and to also bind to multiple less specific sites to facilitate nucleic acid rearrangement constitutes independent activities of NC that contribute to its functioning at different stages of the virus life cycle.

Specific Binding Drives the Unusual Shift in the TAR Transition State Distance.

A reduction of the hairpin unfolding free-energy barrier may be caused by increasing concentrations of NC or by decreasing the concentration of salt in solution. Previous investigations have examined effects on hairpin force-unfolding due to changing solution ionic strength (22, 28), temperature (29), and stabilizing ligand (30). Importantly, in all of these previous studies the location of the opening transition state was not affected by the variation of solution conditions or the addition of ligand. However, the overall unfolding free energy, the transition state free energy, and the transition kinetics were all affected in these experiments. Solution factors that uniformly affect the stability of all nucleic acid structural elements generally leave the position of the transition state unchanged, although position shifts due to multistate unfolding have been observed in multistem hairpins (31). In contrast, NC binding to TAR alters the two-state transition pathway, specifically shortening the transition state distance, and this indicates that NC’s effect on the stability of structural elements of TAR must be selective. NC binds to guanines adjacent to unstable stem regions, such as mismatches, loops, and bulges. Binding further destabilizes these weak base pairs, leading to a longer destabilized region running from the apical loop down to the wobble base pair beyond the new transition site (G¹¹–C⁵⁰). By itself this destabilization is insufficient to induce full duplex unfolding, but can lead to opening of the base pairs destabilized by nearby duplex imperfections. Therefore, in the presence of NC unfolding requires only the destabilization of the lowest part of the stem structure up to the location of the transition state, whereas the net elongation during the transition is still the full TAR RNA length. This model quantitatively elucidates the biophysical mechanism responsible for the experimental observation of a shortened transition state but an unchanged overall elongation upon unfolding.

Conclusions

These experiments form an important novel case study of a protein that locally destabilizes nucleic acid secondary structures at specific locations. HIV-1 NC preferentially destabilizes specific G-containing base pairs adjacent to defects in the secondary structure, thereby leading to further RNA structure fragmentation. We have shown that the effects are particularly significant for TAR as the hairpin already contains several destabilized regions, which are enhanced by the binding of NC. Thus, NC facilitates TAR RNA annealing to its cDNA hairpin by further fragmentation of the interrupted TAR hairpin stem, dramatically shifting the transition state and leading to a shorter critical unfolding helix, while having little effect on the stability of the final annealed long nucleic acid duplex. Our measured change in the unfolding free-energy barrier height as well as shift in position of the transition state drives a 10⁴ increase in the zero-force rate of unfolding, which we determine here. Unlike ligands or solution conditions that uniformly alter duplex stability, NC’s targeted, structure- and sequence-specific activity appears optimal for the purpose of facilitating nucleic acid rearrangements into the lowest free-energy conformation. This activity is known to be critical for reverse transcription in retroviruses. By identifying the locations and magnitude of TAR RNA destabilization by NC, as well as quantifying the resulting unfolding landscape, the present study adds important insight into the conceptual picture of the molecular mechanism of NC’s nucleic acid chaperone activity.

Materials and Methods

Both control and TAR hairpin constructs were tethered between labeled beads as shown in Fig. 1C. See SI Materials and Methods, section 7 and Fig. S5 for construct preparation. A 2.1-μm diameter anti–digoxigenin-coated bead (Spherotech) was fixed onto a micropipette tip (WPI), while a 5.4-μm diameter streptavidin-coated bead (Bangs Labs) was held in a dual laser (Lumics) optical trap, described previously (32). Constructs were extended using a subnanometer resolution piezoelectric transducer (nPoint), and forces were recorded on lateral effect detectors (SpotOn; National Instruments). Experiments were performed at constant pulling rates, which are equivalent to force-ramp rate experiments, as discussed in SI Materials and Methods, section 8. Chosen pulling rates (0.8 pN/s, 10 pN/s, and 30 pN/s) span the tradeoff between detector noise, piezoelectric stage feedback, and instrument baseline stability. Experimental buffer was 10 mM Hepes, pH 7.5, and 100 mM Na⁺, and experimental temperature was 23 °C. Both the elasticity of the DNA handles and unfolded RNA segments are characterized by models of polymer elasticity (SI Materials and Methods, section 1). Hairpins were also extended after exposure to a solution of NC prepared as described previously (33). The specific binding sites on TAR RNA appear to be saturated at 50 nM NC (34, 35), as discussed further in SI Materials and Methods, section 9.

Fig. S5.

RNA/DNA construct preparation. (A) RNA/DNA construct (not to scale), where the TAR RNA (blue) includes 12-nt RNA flanking sequences (highlighted in red and yellow). These sequences are annealed/ligated to short DNA oligos (green), and then to long, labeled DNA handles, as described in SI Materials and Methods. (B) Sequences of all oligonucleotides used, with bases shaded to match A. The T7 promoter for RNA synthesis is also shown (purple).

Acknowledgments

This work was supported by National Institutes of Health (NIH) Grant GM072462 and National Science Foundation Grant MCB-1243883 (to M.C.W.) and NIH Grant GM065056 (to K.M.-F.). Additional funds were provided by the National Cancer Institute, NIH, under Contract HHSN261200800001E with Leidos Biomedical Research, Inc. (to R.J.G.).

Footnotes

↵¹To whom correspondence should be addressed. Email: mark{at}neu.edu.

Author contributions: K.M.-F. and M.C.W. formulated the experimental concept; M.J.M. built the apparatus, performed experiments, and analyzed data; K.A.M. synthesized TAR hairpins; R.J.G. provided NC protein; I.R. interpreted results and provided feedback; and M.J.M., I.R., R.J.G., K.M.-F., and M.C.W. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1510100112/-/DCSupplemental.

References

↵
1. Levin JG,
2. Guo J,
3. Rouzina I,
4. Musier-Forsyth K
(2005) Nucleic acid chaperone activity of HIV-1 nucleocapsid protein: Critical role in reverse transcription and molecular mechanism. Prog Nucleic Acid Res Mol Biol 80:217–286
.
OpenUrl CrossRef PubMed
↵
1. Berkhout B,
2. Jeang KT
(1992) Functional roles for the TATA promoter and enhancers in basal and Tat-induced expression of the human immunodeficiency virus type 1 long terminal repeat. J Virol 66(1):139–149
.
OpenUrl Abstract/FREE Full Text
↵
1. Vo MN,
2. Barany G,
3. Rouzina I,
4. Musier-Forsyth K
(2006) Mechanistic studies of mini-TAR RNA/DNA annealing in the absence and presence of HIV-1 nucleocapsid protein. J Mol Biol 363(1):244–261
.
OpenUrl CrossRef PubMed
↵
1. Cruceanu M,
2. Gorelick RJ,
3. Musier-Forsyth K,
4. Rouzina I,
5. Williams MC
(2006) Rapid kinetics of protein-nucleic acid interaction is a major component of HIV-1 nucleocapsid protein’s nucleic acid chaperone function. J Mol Biol 363(5):867–877
.
OpenUrl CrossRef PubMed
↵
1. Wu H, et al.
(2013) Aromatic residue mutations reveal direct correlation between HIV-1 nucleocapsid protein’s nucleic acid chaperone activity and retroviral replication. Virus Res 171(2):263–277
.
OpenUrl CrossRef PubMed
↵
1. Godet J, et al.
(2013) Site-selective probing of cTAR destabilization highlights the necessary plasticity of the HIV-1 nucleocapsid protein to chaperone the first strand transfer. Nucleic Acids Res 41(9):5036–5048
.
OpenUrl Abstract/FREE Full Text
↵
1. Grohman JK, et al.
(2013) A guanosine-centric mechanism for RNA chaperone function. Science 340(6129):190–195
.
OpenUrl Abstract/FREE Full Text
↵
1. Wilkinson KA, et al.
(2008) High-throughput SHAPE analysis reveals structures in HIV-1 genomic RNA strongly conserved across distinct biological states. PLoS Biol 6(4):e96
.
OpenUrl CrossRef PubMed
↵
1. Hyeon C,
2. Thirumalai D
(2007) Mechanical unfolding of RNA: From hairpins to structures with internal multiloops. Biophys J 92(3):731–743
.
OpenUrl CrossRef PubMed
↵
1. Woodside MT, et al.
(2006) Nanomechanical measurements of the sequence-dependent folding landscapes of single nucleic acid hairpins. Proc Natl Acad Sci USA 103(16):6190–6195
.
OpenUrl Abstract/FREE Full Text
↵
1. Li PT,
2. Collin D,
3. Smith SB,
4. Bustamante C,
5. Tinoco I Jr
(2006) Probing the mechanical folding kinetics of TAR RNA by hopping, force-jump, and force-ramp methods. Biophys J 90(1):250–260
.
OpenUrl CrossRef PubMed
↵
1. Greenleaf WJ,
2. Frieda KL,
3. Foster DAN,
4. Woodside MT,
5. Block SM
(2008) Direct observation of hierarchical folding in single riboswitch aptamers. Science 319(5863):630–633
.
OpenUrl Abstract/FREE Full Text
↵
1. Collin D, et al.
(2005) Verification of the Crooks fluctuation theorem and recovery of RNA folding free energies. Nature 437(7056):231–234
.
OpenUrl CrossRef PubMed
↵
1. Crooks GE
(1999) Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences. Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics 60(3):2721–2726
.
OpenUrl PubMed
↵
1. Jarzynski C
(1997) Nonequilibrium equality for free energy differences. Phys Rev Lett 78:2690–2693
.
OpenUrl CrossRef
↵
1. Bennett CH
(1976) Efficient estimation of free energy differences from Monte Carlo data. J Comput Phys 22:245–268
.
OpenUrl CrossRef
↵
1. Dudko OK,
2. Hummer G,
3. Szabo A
(2006) Intrinsic rates and activation free energies from single-molecule pulling experiments. Phys Rev Lett 96(10):108101
.
OpenUrl CrossRef PubMed
↵
1. Dudko OK,
2. Hummer G,
3. Szabo A
(2008) Theory, analysis, and interpretation of single-molecule force spectroscopy experiments. Proc Natl Acad Sci USA 105(41):15755–15760
.
OpenUrl Abstract/FREE Full Text
↵
1. Zhang Y,
2. Dudko OK
(2013) A transformation for the mechanical fingerprints of complex biomolecular interactions. Proc Natl Acad Sci USA 110(41):16432–16437
.
OpenUrl Abstract/FREE Full Text
↵
1. Zuker M
(2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31(13):3406–3415
.
OpenUrl Abstract/FREE Full Text
↵
1. Woodside MT, et al.
(2006) Direct measurement of the full, sequence-dependent folding landscape of a nucleic acid. Science 314(5801):1001–1004
.
OpenUrl Abstract/FREE Full Text
↵
1. Vieregg J,
2. Cheng W,
3. Bustamante C,
4. Tinoco I Jr
(2007) Measurement of the effect of monovalent cations on RNA hairpin stability. J Am Chem Soc 129(48):14966–14973
.
OpenUrl CrossRef PubMed
↵
1. Cosa G, et al.
(2004) Secondary structure and secondary structure dynamics of DNA hairpins complexed with HIV-1 NC protein. Biophys J 87(4):2759–2767
.
OpenUrl CrossRef PubMed
↵
1. Gherghe C, et al.
(2010) Definition of a high-affinity Gag recognition structure mediating packaging of a retroviral RNA genome. Proc Natl Acad Sci USA 107(45):19248–19253
.
OpenUrl Abstract/FREE Full Text
↵
1. Miller SB,
2. Yildiz FZ,
3. Lo JA,
4. Wang B,
5. D’Souza VM
(2014) A structure-based mechanism for tRNA and retroviral RNA remodelling during primer annealing. Nature 515(7528):591–595
.
OpenUrl CrossRef PubMed
↵
1. Bazzi A, et al.
(2012) Intrinsic nucleic acid dynamics modulates HIV-1 nucleocapsid protein binding to its targets. PLoS One 7(6):e38905
.
OpenUrl CrossRef PubMed
↵
1. Vuilleumier C, et al.
(1999) Nucleic acid sequence discrimination by the HIV-1 nucleocapsid protein NCp7: A fluorescence study. Biochemistry 38(51):16816–16825
.
OpenUrl CrossRef PubMed
↵
1. Bizarro CV,
2. Alemany A,
3. Ritort F
(2012) Non-specific binding of Na+ and Mg2+ to RNA determined by force spectroscopy methods. Nucleic Acids Res 40(14):6922–6935
.
OpenUrl Abstract/FREE Full Text
↵
1. Stephenson W, et al.
(2014) Combining temperature and force to study folding of an RNA hairpin. Phys Chem Chem Phys 16(3):906–917
.
OpenUrl CrossRef PubMed
↵
1. Anthony PC,
2. Perez CF,
3. García-García C,
4. Block SM
(2012) Folding energy landscape of the thiamine pyrophosphate riboswitch aptamer. Proc Natl Acad Sci USA 109(5):1485–1489
.
OpenUrl Abstract/FREE Full Text
↵
1. Liphardt J,
2. Onoa B,
3. Smith SB,
4. Tinoco I Jr,
5. Bustamante C
(2001) Reversible unfolding of single RNA molecules by mechanical force. Science 292(5517):733–737
.
OpenUrl Abstract/FREE Full Text
↵
1. Chaurasiya KR,
2. Paramanathan T,
3. McCauley MJ,
4. Williams MC
(2010) Biophysical characterization of DNA binding from single molecule force measurements. Phys Life Rev 7(3):299–341
.
OpenUrl CrossRef PubMed
↵
1. Wu W, et al.
(1996) Human immunodeficiency virus type 1 nucleocapsid protein reduces reverse transcriptase pausing at a secondary structure near the murine leukemia virus polypurine tract. J Virol 70(10):7132–7142
.
OpenUrl Abstract/FREE Full Text
↵
1. Webb JA,
2. Jones CP,
3. Parent LJ,
4. Rouzina I,
5. Musier-Forsyth K
(2013) Distinct binding interactions of HIV-1 Gag to Psi and non-Psi RNAs: Implications for viral genomic RNA packaging. RNA 19(8):1078–1088
.
OpenUrl Abstract/FREE Full Text
↵
1. Jones CP,
2. Datta SA,
3. Rein A,
4. Rouzina I,
5. Musier-Forsyth K
(2011) Matrix domain modulates HIV-1 Gag’s nucleic acid chaperone activity via inositol phosphate binding. J Virol 85(4):1594–1603
.
OpenUrl Abstract/FREE Full Text

Article Alerts

Email Article

Citation Tools

Request Permissions

Article Classifications

Biological Sciences
Biophysics and Computational Biology

Proceedings of the National Academy of Sciences: 112 (44)

Table of Contents

Submit

[1] ↵
Levin JG,
Guo J,
Rouzina I,
Musier-Forsyth K
(2005) Nucleic acid chaperone activity of HIV-1 nucleocapsid protein: Critical role in reverse transcription and molecular mechanism. Prog Nucleic Acid Res Mol Biol 80:217–286
.
OpenUrl CrossRef PubMed

[2] Levin JG,

[3] Guo J,

[4] Rouzina I,

[5] Musier-Forsyth K

[6] ↵
Berkhout B,
Jeang KT
(1992) Functional roles for the TATA promoter and enhancers in basal and Tat-induced expression of the human immunodeficiency virus type 1 long terminal repeat. J Virol 66(1):139–149
.
OpenUrl Abstract/FREE Full Text

[7] Berkhout B,

[8] Jeang KT

[9] ↵
Vo MN,
Barany G,
Rouzina I,
Musier-Forsyth K
(2006) Mechanistic studies of mini-TAR RNA/DNA annealing in the absence and presence of HIV-1 nucleocapsid protein. J Mol Biol 363(1):244–261
.
OpenUrl CrossRef PubMed

[10] Vo MN,

[11] Barany G,

[12] Rouzina I,

[13] Musier-Forsyth K

[14] ↵
Cruceanu M,
Gorelick RJ,
Musier-Forsyth K,
Rouzina I,
Williams MC
(2006) Rapid kinetics of protein-nucleic acid interaction is a major component of HIV-1 nucleocapsid protein’s nucleic acid chaperone function. J Mol Biol 363(5):867–877
.
OpenUrl CrossRef PubMed

[15] Cruceanu M,

[16] Gorelick RJ,

[17] Musier-Forsyth K,

[18] Rouzina I,

[19] Williams MC

[20] ↵
Wu H, et al.
(2013) Aromatic residue mutations reveal direct correlation between HIV-1 nucleocapsid protein’s nucleic acid chaperone activity and retroviral replication. Virus Res 171(2):263–277
.
OpenUrl CrossRef PubMed

[21] Wu H, et al.

[22] ↵
Godet J, et al.
(2013) Site-selective probing of cTAR destabilization highlights the necessary plasticity of the HIV-1 nucleocapsid protein to chaperone the first strand transfer. Nucleic Acids Res 41(9):5036–5048
.
OpenUrl Abstract/FREE Full Text

[23] Godet J, et al.

[24] ↵
Grohman JK, et al.
(2013) A guanosine-centric mechanism for RNA chaperone function. Science 340(6129):190–195
.
OpenUrl Abstract/FREE Full Text

[25] Grohman JK, et al.

[26] ↵
Wilkinson KA, et al.
(2008) High-throughput SHAPE analysis reveals structures in HIV-1 genomic RNA strongly conserved across distinct biological states. PLoS Biol 6(4):e96
.
OpenUrl CrossRef PubMed

[27] Wilkinson KA, et al.

[28] ↵
Hyeon C,
Thirumalai D
(2007) Mechanical unfolding of RNA: From hairpins to structures with internal multiloops. Biophys J 92(3):731–743
.
OpenUrl CrossRef PubMed

[29] Hyeon C,

[30] Thirumalai D

[31] ↵
Woodside MT, et al.
(2006) Nanomechanical measurements of the sequence-dependent folding landscapes of single nucleic acid hairpins. Proc Natl Acad Sci USA 103(16):6190–6195
.
OpenUrl Abstract/FREE Full Text

[32] Woodside MT, et al.

[33] ↵
Li PT,
Collin D,
Smith SB,
Bustamante C,
Tinoco I Jr
(2006) Probing the mechanical folding kinetics of TAR RNA by hopping, force-jump, and force-ramp methods. Biophys J 90(1):250–260
.
OpenUrl CrossRef PubMed

[34] Li PT,

[35] Collin D,

[36] Smith SB,

[37] Bustamante C,

[38] Tinoco I Jr

[39] ↵
Greenleaf WJ,
Frieda KL,
Foster DAN,
Woodside MT,
Block SM
(2008) Direct observation of hierarchical folding in single riboswitch aptamers. Science 319(5863):630–633
.
OpenUrl Abstract/FREE Full Text

[40] Greenleaf WJ,

[41] Frieda KL,

[42] Foster DAN,

[43] Woodside MT,

[44] Block SM

[45] ↵
Collin D, et al.
(2005) Verification of the Crooks fluctuation theorem and recovery of RNA folding free energies. Nature 437(7056):231–234
.
OpenUrl CrossRef PubMed

[46] Collin D, et al.

[47] ↵
Crooks GE
(1999) Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences. Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics 60(3):2721–2726
.
OpenUrl PubMed

[48] Crooks GE

[49] ↵
Jarzynski C
(1997) Nonequilibrium equality for free energy differences. Phys Rev Lett 78:2690–2693
.
OpenUrl CrossRef

[50] Jarzynski C

[51] ↵
Bennett CH
(1976) Efficient estimation of free energy differences from Monte Carlo data. J Comput Phys 22:245–268
.
OpenUrl CrossRef

[52] Bennett CH

[53] ↵
Dudko OK,
Hummer G,
Szabo A
(2006) Intrinsic rates and activation free energies from single-molecule pulling experiments. Phys Rev Lett 96(10):108101
.
OpenUrl CrossRef PubMed

[54] Dudko OK,

[55] Hummer G,

[56] Szabo A

[57] ↵
Dudko OK,
Hummer G,
Szabo A
(2008) Theory, analysis, and interpretation of single-molecule force spectroscopy experiments. Proc Natl Acad Sci USA 105(41):15755–15760
.
OpenUrl Abstract/FREE Full Text

[58] Dudko OK,

[59] Hummer G,

[60] Szabo A

[61] ↵
Zhang Y,
Dudko OK
(2013) A transformation for the mechanical fingerprints of complex biomolecular interactions. Proc Natl Acad Sci USA 110(41):16432–16437
.
OpenUrl Abstract/FREE Full Text

[62] Zhang Y,

[63] Dudko OK

[64] ↵
Zuker M
(2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31(13):3406–3415
.
OpenUrl Abstract/FREE Full Text

[65] Zuker M

[66] ↵
Woodside MT, et al.
(2006) Direct measurement of the full, sequence-dependent folding landscape of a nucleic acid. Science 314(5801):1001–1004
.
OpenUrl Abstract/FREE Full Text

[67] Woodside MT, et al.

[68] ↵
Vieregg J,
Cheng W,
Bustamante C,
Tinoco I Jr
(2007) Measurement of the effect of monovalent cations on RNA hairpin stability. J Am Chem Soc 129(48):14966–14973
.
OpenUrl CrossRef PubMed

[69] Vieregg J,

[70] Cheng W,

[71] Bustamante C,

[72] Tinoco I Jr

[73] ↵
Cosa G, et al.
(2004) Secondary structure and secondary structure dynamics of DNA hairpins complexed with HIV-1 NC protein. Biophys J 87(4):2759–2767
.
OpenUrl CrossRef PubMed

[74] Cosa G, et al.

[75] ↵
Gherghe C, et al.
(2010) Definition of a high-affinity Gag recognition structure mediating packaging of a retroviral RNA genome. Proc Natl Acad Sci USA 107(45):19248–19253
.
OpenUrl Abstract/FREE Full Text

[76] Gherghe C, et al.

[77] ↵
Miller SB,
Yildiz FZ,
Lo JA,
Wang B,
D’Souza VM
(2014) A structure-based mechanism for tRNA and retroviral RNA remodelling during primer annealing. Nature 515(7528):591–595
.
OpenUrl CrossRef PubMed

[78] Miller SB,

[79] Yildiz FZ,

[80] Lo JA,

[81] Wang B,

[82] D’Souza VM

[83] ↵
Bazzi A, et al.
(2012) Intrinsic nucleic acid dynamics modulates HIV-1 nucleocapsid protein binding to its targets. PLoS One 7(6):e38905
.
OpenUrl CrossRef PubMed

[84] Bazzi A, et al.

[85] ↵
Vuilleumier C, et al.
(1999) Nucleic acid sequence discrimination by the HIV-1 nucleocapsid protein NCp7: A fluorescence study. Biochemistry 38(51):16816–16825
.
OpenUrl CrossRef PubMed

[86] Vuilleumier C, et al.

[87] ↵
Bizarro CV,
Alemany A,
Ritort F
(2012) Non-specific binding of Na+ and Mg2+ to RNA determined by force spectroscopy methods. Nucleic Acids Res 40(14):6922–6935
.
OpenUrl Abstract/FREE Full Text

[88] Bizarro CV,

[89] Alemany A,

[90] Ritort F

[91] ↵
Stephenson W, et al.
(2014) Combining temperature and force to study folding of an RNA hairpin. Phys Chem Chem Phys 16(3):906–917
.
OpenUrl CrossRef PubMed

[92] Stephenson W, et al.

[93] ↵
Anthony PC,
Perez CF,
García-García C,
Block SM
(2012) Folding energy landscape of the thiamine pyrophosphate riboswitch aptamer. Proc Natl Acad Sci USA 109(5):1485–1489
.
OpenUrl Abstract/FREE Full Text

[94] Anthony PC,

[95] Perez CF,

[96] García-García C,

[97] Block SM

[98] ↵
Liphardt J,
Onoa B,
Smith SB,
Tinoco I Jr,
Bustamante C
(2001) Reversible unfolding of single RNA molecules by mechanical force. Science 292(5517):733–737
.
OpenUrl Abstract/FREE Full Text

[99] Liphardt J,

[100] Onoa B,

[101] Smith SB,

[102] Tinoco I Jr,

[103] Bustamante C

[104] ↵
Chaurasiya KR,
Paramanathan T,
McCauley MJ,
Williams MC
(2010) Biophysical characterization of DNA binding from single molecule force measurements. Phys Life Rev 7(3):299–341
.
OpenUrl CrossRef PubMed

[105] Chaurasiya KR,

[106] Paramanathan T,

[107] McCauley MJ,

[108] Williams MC

[109] ↵
Wu W, et al.
(1996) Human immunodeficiency virus type 1 nucleocapsid protein reduces reverse transcriptase pausing at a secondary structure near the murine leukemia virus polypurine tract. J Virol 70(10):7132–7142
.
OpenUrl Abstract/FREE Full Text

[110] Wu W, et al.

[111] ↵
Webb JA,
Jones CP,
Parent LJ,
Rouzina I,
Musier-Forsyth K
(2013) Distinct binding interactions of HIV-1 Gag to Psi and non-Psi RNAs: Implications for viral genomic RNA packaging. RNA 19(8):1078–1088
.
OpenUrl Abstract/FREE Full Text

[112] Webb JA,

[113] Jones CP,

[114] Parent LJ,

[115] Rouzina I,

[116] Musier-Forsyth K

[117] ↵
Jones CP,
Datta SA,
Rein A,
Rouzina I,
Musier-Forsyth K
(2011) Matrix domain modulates HIV-1 Gag’s nucleic acid chaperone activity via inositol phosphate binding. J Virol 85(4):1594–1603
.
OpenUrl Abstract/FREE Full Text

[118] Jones CP,

[119] Datta SA,

[120] Rein A,

[121] Rouzina I,

[122] Musier-Forsyth K

Main menu

User menu

Search

Targeted binding of nucleocapsid protein transforms the folding landscape of HIV-1 TAR RNA

Significance

Abstract

Results

TAR Hairpin Unfolding Is a Two-State Process.

NC Reduces the Equilibrium Free Energy of TAR RNA Hairpin Unfolding.

NC Shifts the Transition State of Hairpin Unfolding.

mfold Quantifies Destabilization of TAR Due to Specific NC Binding.

SI Materials and Methods

1) Models of Polymer Elasticity.

2) Comparisons of Unfolding Free-Energy Measurements.

3) Opening Force Distribution Analysis to Obtain Transition State Parameters.

4) Molecular Opening Rates, k_op(F), Are Determined from the Distribution of Opening Forces, P_op(F).

5) The Effective TAR RNA Folded State Is the Frayed State.

6) Solution Conditions of the Landscape Model Calculations Are Equivalent to Experimental Conditions.

7) RNA/DNA Construct Preparation.

8) Conversion Between Constant Pulling Rate and Constant Force Ramp Rates.

9) Higher Concentrations of NC Lead to Only Minor Additional Effects on TAR RNA Opening.

Discussion

TAR RNA Hairpin Stability Is Weakened by Multiple Duplex Defects.

NC Targets Specific Locations on the TAR RNA Hairpin.

Specific Binding Drives the Unusual Shift in the TAR Transition State Distance.

Conclusions

Materials and Methods

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information

Main menu

User menu

Search

Significance

Abstract

Results

TAR Hairpin Unfolding Is a Two-State Process.

NC Reduces the Equilibrium Free Energy of TAR RNA Hairpin Unfolding.

NC Shifts the Transition State of Hairpin Unfolding.

mfold Quantifies Destabilization of TAR Due to Specific NC Binding.

SI Materials and Methods

1) Models of Polymer Elasticity.

2) Comparisons of Unfolding Free-Energy Measurements.

3) Opening Force Distribution Analysis to Obtain Transition State Parameters.

4) Molecular Opening Rates, kop(F), Are Determined from the Distribution of Opening Forces, Pop(F).

5) The Effective TAR RNA Folded State Is the Frayed State.

6) Solution Conditions of the Landscape Model Calculations Are Equivalent to Experimental Conditions.

7) RNA/DNA Construct Preparation.

8) Conversion Between Constant Pulling Rate and Constant Force Ramp Rates.

9) Higher Concentrations of NC Lead to Only Minor Additional Effects on TAR RNA Opening.

Discussion

TAR RNA Hairpin Stability Is Weakened by Multiple Duplex Defects.

NC Targets Specific Locations on the TAR RNA Hairpin.

Specific Binding Drives the Unusual Shift in the TAR Transition State Distance.

Conclusions

Materials and Methods

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

Sign up for Article Alerts

Jump to section

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information

4) Molecular Opening Rates, k_op(F), Are Determined from the Distribution of Opening Forces, P_op(F).