HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 45, 43233-43242, November 8, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/45/43233    most recent M205295200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Goldschmidt, V. Articles by Marquet, R. Search for Related Content PubMed PubMed Citation Articles by Goldschmidt, V. Articles by Marquet, R. Social Bookmarking                      What's this?

Direct and Indirect Contributions of RNA Secondary Structure Elements to the Initiation of HIV-1 Reverse Transcription^*

Valérie Goldschmidt, Mickaël Rigourd, Chantal Ehresmann, Stuart F. J. Le Grice^§, Bernard Ehresmann, and Roland Marquet^¶

From the  UPR 9002 du CNRS affiliée à l'Université Louis Pasteur, Institut de Biologie Moléculaire et Cellulaire, 15 rue René Descartes, 67084 Strasbourg cedex, France and the ^§ Resistance Mechanisms Laboratory, NCI-Frederick, National Institutes of Health, Frederick, Maryland 21702

Received for publication, May 29, 2002, and in revised form, August 21, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Initiation of human immunodeficiency virus type 1 (HIV-1) reverse transcription requires specific recognition between the viral RNA (vRNA), tRNA, which acts as primer, and reverse transcriptase (RT). The specificity of this ternary complex is mediated by intricate interactions between the HIV-1 RNA and tRNA. Here, we compared the relative importance of the secondary structure elements of this complex in the initiation process. To this aim, we used the previously published three-dimensional model of the initiation complex to rationally introduce a series of deletions and substitutions in the vRNA. When necessary, we used chemical probing to check the structure of the tRNA-mutant vRNA complexes. For each of them, we measured the binding affinity of RT and the kinetics of initial extension of tRNA and of synthesis of the () strand strong stop DNA. Our results were overall in keeping with the three-dimensional model of the initiation complex. Surprisingly, we found that disruption of the intermolecular template-primer interactions, which are not directly recognized by RT, more severely affected reverse transcription than deletions or disruption of one of the intramolecular helices to which RT directly binds. Perturbations of the highly constrained junction between the intermolecular helix formed by the primer binding site and the 3' end of tRNA and the helix immediately upstream also had dramatic effects on the initiation of reverse transcription. Taken together, our results demonstrate the overwhelming importance of the overall three-dimensional structure of the initiation complex and identify structural elements that constitute promising targets for anti-initiation-specific drugs.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reverse transcription is a key event in the retroviral replication cycle (1, 2). During this process, the single-stranded genomic RNA is converted into double-stranded DNA by reverse transcriptase (RT),¹ a multifunctional enzyme that possesses RNA- and DNA-dependent DNA polymerase and RNase H activities (3). Initiation of reverse transcription is primed by a cellular tRNA that is selectively encapsidated into the viral particles (reviewed in Refs. 4-6). Different classes of retroviruses use different primer tRNAs: tRNA is the natural primer of most immunodeficiency viruses, including the type 1 human immunodeficiency virus (HIV-1), whereas tRNA^Trp and tRNA^Pro are used by most avian and murine retroviruses, respectively (4-6). In all cases, the 18 nucleotides at the 3' end of the primer are complementary to the primer binding site (PBS) located in the 5' region of the genomic RNA.

In addition to this "general" tRNA-PBS interaction, "virus-specific" interactions between the primer tRNA and the genomic RNA have been demonstrated in avian retroviruses (7-10), HIV types 1 (11-18) and 2 (19, 20), feline immunodeficiency virus (21), and the yeast retrotransposons Ty1 (22-24) and Ty3 (25). These additional interactions are required for efficient replication (9, 24, 26-28) and in vivo initiation of reverse transcription (29-31). They also account for the specific binding of HIV-1 RT to the initiation complex (32-34) and the efficient extension of tRNA, as compared with other RNA primers, observed in vitro (32, 33, 35-37).

In HIV-1, the existence of intricate interactions between tRNA and the genomic RNA was initially demonstrated in vitro, using chemical and enzymatic probing (13, 15) and site-directed mutagenesis (14, 32). These data allowed us to propose a secondary structure model of the initiation complex for HIV-1 Mal (13) (see Fig. 1). In this model, in addition to the interaction between the 3' part of tRNA with the PBS (helix 7F in Fig. 1), parts of the anticodon stem-loop and of the variable loop of tRNA interact with viral sequences upstream of the PBS (forming helices 6C, 5D, and 3E in Fig. 1). The same interactions were observed when the primer-template complex was formed by heat annealing or with the nucleocapsid protein (38). Recently, an alternative interaction was proposed for HIV-1 HXB2 between the sequence involved in the 5' strand of helix 1 in our model and 5' part of the TC stem-loop of tRNA (39).

Probing of the ribose-phosphate backbone of the tRNA-viral RNA (vRNA) complex and enzymatic footprinting of RT on this complex allowed us to build a three-dimensional model of the tRNA-vRNA:RT complex (16). This model suggested that HIV-1 RT directly interacts with helices 7F, 1, and 8 and with the three-nucleotide junction between helices 7F and 2. Modeling also indicated that helix 2 is in close proximity to the finger and thumb subdomains of RT and could contribute to binding of the polymerase. Importantly, we observed no interaction between RT and the intermolecular helices 6C, 5D, and 3E, whose importance for viral replication has been demonstrated (26-28, 31, 40). These helices likely play an indirect role by preventing steric clashes between the primer-template complex and RT and/or by imposing a correct orientation of the structural elements directly interacting with the polymerase (16).

In the present study, we used the three-dimensional tRNA-vRNA:RT model to design a series of deletion and substitution mutants of the vRNA, to compare the importance of the RNA structural elements on the early steps of reverse transcription. When required, we used chemical probing to check the structure of the tRNA-mutant vRNA complexes, and for each of them, we measured the binding affinity of RT and the kinetics of initial extension of tRNA and of synthesis of the () strand strong stop DNA (() ssDNA). Unexpectedly, our results indicated that perturbing the global folding of the initiation complex was more detrimental than deleting individual structural elements involved in RT binding. They also identified structural elements that constitute promising targets for anti-initiation-specific drugs.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Primer, Templates, and RT-- Natural tRNA was purified from beef liver as described (41). It was internally labeled at its ultimate 3' phosphate or at its 5' end according to published procedures (16).

Wild type and S162-167 RNAs were obtained by in vitro transcription of plasmids pJCB (42) and pICA1 (18), respectively, linearized with RsaI. RNA 123-217 was obtained by in vitro transcription of the PCR product obtained by amplification of nucleotides 123-217 of pJCB using primers 5'-GGAATTCTAATACGACTCACTATAGGGCTCTGGTAACTAGAGATCCC-3' (sense) and 5'-GGGCCCTGTTACTTTCACTTTAAAGTCCC-3' (antisense). To obtain the other RNA templates, plasmid pJCB was mutated with the QuikChange^TM site-directed mutagenesis kit (Stratagene) using the protocols provided by the supplier, and the mutant vRNAs were obtained as described above. All of the RNAs were purified as described previously (43). Wild type HIV-1 RT was purified essentially as described (44)

Chemical Probing of vRNA-- Viral RNA (4 pmol) and purified natural tRNA were incubated in water for 2 min at 90 °C, cooled on ice, and incubated at 70 °C for 20 min in sodium cacodylate (pH 7.5) 50 mM, 300 mM KCl. After hybridization, the samples were incubated at 20 °C for 15 min in the same buffer supplemented with 5 mM MgCl₂. After addition of 1 µg of yeast total tRNA, 1 µl of 10-fold diluted DMS was added and allowed to react for 5 or 10 min. RNA modification was stopped with 200 µl of ethanol and 50 µl of sodium acetate 0.3 M (pH 5.3) containing 1 µg of glycogen. A DMS reaction was conducted in parallel on the free vRNA. Modified bases were detected by primer extension with reverse transcriptase as described previously (45).

() Strand Strong Stop DNA Synthesis-- In a standard experiment, vRNA (final concentration, 30 nM) was annealed with ³²P-labeled tRNA (final concentration, 10 nM) as described above and preincubated for 4 min with 25 nM RT in 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 6 mM MgCl₂, 1 mM dithioerythritol. Reverse transcription was initiated by adding a mixture of the four deoxynucleotide triphosphates (50 µM each) in the same buffer. Formamide containing 50 mM EDTA was added to aliquots of the reaction mixture at times ranging from 15 s to 30 min, and the reaction products were analyzed on 8% denaturing polyacrylamide gels and quantified with a BioImager BAS 2000 (Fuji) using the MacBas software.

Determination of the Equilibrium Dissociation Constant of the Template-Primer:RT Complexes-- To determine the equilibrium dissociation constant of the wild type and mutant vRNA-tRNA:RT complexes, ~5 nM ³²P-labeled tRNA-vRNA complex were preincubated for 4 min with increasing concentrations of RT (0 to 400 nM) in 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 6 mM MgCl₂, 1 mM dithioerythritol. The addition of 500 µM dCTP, together with poly(rA)·(dT)₁₅ at a final concentration of 1 µM (dT)₁₅ allowed the extension of the primer by one nucleotide in the preformed ternary complexes, while preventing recycling of RT. Prior to addition to the reaction mixture, poly(rA) and (dT)₁₅ at a 10:1 (w/w) ratio were annealed for 20 min at 70 °C. The extension reaction was stopped after 15 s, and the reaction products were separated and quantified as described above. The efficiency of poly(rA)·(dT)₁₅ in trapping free RT was controlled by checking that primer extension did not significantly increase when the reaction time was increased from 15 s to 15 min.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Design of the vRNA Mutants-- The wild type RNA template used in this study encompassed nucleotides 1-311 of the HIV-1 Mal isolate. It served as a reference, because the structural studies conducted on the tRNA-vRNA and tRNA-vRNA:RT complexes were performed using this isolate (13-16,32), and structural probing data for the HXB2 strain (39, 46, 47) concern only the free RNA. During formation of the tRNA-vRNA complex, most of the structural rearrangements were observed between nucleotides 123 and 217 of the vRNA, and previous structural studies showed that a template encompassing these nucleotides formed the same primer-template complex as RNA 1-311 (14, 16) (Fig. 1). Therefore, we tested whether RNA 123-217 contains all of the elements required for efficient initiation of reverse transcription.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Secondary structure model of the wild type and mutant vRNA-tRNA complexes used in this study. The portion of the vRNA encompassing nucleotides 123-217 of HIV-1 Mal is shown in black, and tRNA is drawn in red. The helices are numbered according to Ref. 13. The mutations introduced in the vRNA are shown in blue. The schematic drawings of the mutant complexes are not intended to represent the actual structure of these complexes but only to indicate the mutated regions. Nucleotides whose increased (*) or decreased (o) reactivity are the hallmark of the primer-template complex formation are indicated on the wild type complex.

To test the importance of helices 1 and 8, we constructed a template corresponding to nucleotides 131-196 that lacks these two helices and mutants corresponding to deletion of nucleotides 123-130 and 208-217 (Hx 1) and 200-208 (Hx8), in the context of RNA 123-217. All other mutations were introduced in the context of RNA 1-311.

To test the function of helix 2, we first substituted nucleotides 132-139 by AUCUCUAG, a copy of the 3' strand of this helix (Fig. 1, mutant S132-139). We then restored this helix by further substituting nucleotides 168-175 by the original 5' strand of helix 2, generating the compensatory mutant CompCG (Fig. 1). To evaluate the influence of the sequence and stability of helix 2 on reverse transcription, we substituted the third base pair of the CompCG helix2 (a C-G base pair) by a U-A base pair (CompUA).

To study the importance of the intermolecular tRNA-vRNA interactions (helices 6C, 5D, and 3E in Fig. 1), we used mutant S162-167, in which the wild type sequence GUAAAA is replaced by CUAUG. A previous structural study of this mutant showed that not only helix 6C but also helices 5D and 3E were disrupted in the corresponding binary complex (14). Because the intermolecular interactions require the post-transcriptional modifications of natural tRNA to be stable (13, 15), it was not possible to use compensatory mutations in the primer to restore helix 6C. To test whether nucleotides 140-167 could hinder reverse transcription by producing steric clashes with RT when not engaged in intermolecular helices, we compared mutants S162-167 and 140-167, in which these nucleotides were replaced by a UUCG tetraloop. A mutant further deleted from helix 2 was also constructed (mutant 132-175).

The last series of mutants was designed to test the importance of the junction between helices 2 and 7F, which imposes strong structural constraints to the tRNA-vRNA complex (16). This junction, which is three nucleotides in length in the wild type complex, was entirely deleted (mutant J0), shortened to one nucleotide (mutant J1), or extended to five nucleotides (mutant J5). Nondenaturing polyacrylamide gel electrophoresis was used to ensure that the heat-annealing protocol we used allowed quantitative (> 95%) hybridization of tRNA to all RNA templates (data not shown).

Effects of the Mutations on the Structure of the Primer-vRNA Complexes-- Conducting a complete structural study on all of the mutant complexes was beyond the scope of this work. Nevertheless, we wanted to check whether the mutations induced unexpected rearrangements of the tRNA-vRNA complexes. In particular, because the intermolecular interactions were previously shown to affect reverse transcription (32), we tested their existence in the mutants not designed to disrupt them. Probing of the template RNA with DMS, which modifies single-stranded As and Cs at one of their Watson-Crick positions (48), was particularly useful in this respect.

We tested the conformation of mutants S132-139, CompCG, and CompUA, either free or engaged in the primer-template complex (Fig. 2). The main signature of the intermolecular interactions was the partial protection of A164-167 upon formation of the wild type binary complex, with A165 and A166 being more efficiently protected than A164 and A167, whereas A168, which was engaged in helix 2, was not modified (Fig. 2). In addition, formation of the wild type complex was accompanied by an increased reactivity of C176, A157, and A150 that reflected rearrangement of the vRNA (Figs. 1 and 2).

View larger version (71K): [in this window] [in a new window]   Fig. 2.   Probing of the wild type and some selected mutant vRNA-tRNA complexes with dimethyl sulfate. All templates were probed free (tRNA) and engaged in the binary complex (+tRNA). In each case, a control lane without DMS (lanes 1, 4, 7, 10, 13, 16, 19, and 22) and two lanes corresponding to RNA modification with DMS for 5 and 10 min are shown. Nucleotides whose increased (*) or decreased (o) reactivity are the hallmark of the primer-template complex formation are indicated.

Probing experiments clearly indicated that, in addition to disrupting helix 2, the substitution introduced in mutant S132-139 also prevented formation of the intermolecular interactions 3E, 5D, and 6C. Modification of A132, C134, C136, and A138 indicated that the substituted sequence was single-stranded. In addition, A164-167 remained equally modified by DMS upon formation of the primer-template complex, and reactivity of C176 and A157 did not increase (Fig. 2). However, a structural rearrangement involving nucleotides 145-150 took place, reflecting formation of stem-loop 4. Analysis of mutant CompCG was consistent with the restoration of helix 2, as indicated by the absence of modification of A132, C134, C136, and A138 (Fig. 2). It also indicated that the intermolecular interactions were restored, as judged by the reactivity of C176, A164-167, A157, and A150 (Fig. 2). Expectedly, mutant CompUA yielded the same DMS modification pattern (Fig. 2), indicating that both compensatory mutants adopted the same structure.

Structural probing of mutants J0, J1, and J5 showed that the complexes formed by these mutant templates all adopt the wild type structure (Fig. 2). The decreased reactivity of A164-167 and the increased reactivity of A157 and A150 supported the existence of the intermolecular interactions and helix 4. Furthermore, the absence of reactivity of nucleotides 132-139 was in agreement with formation of helix 2. The only significant difference we observed among those mutants concerned A164, which was as reactive as A167 in mutant J0 but less reactive in J1 and J5. This difference might reflect increased structural stress when deleting the entire junction.

Binding of RT to the Mutant vRNA-tRNA Complexes-- Determination of the equilibrium dissociation constant of RT to wild type and mutant template-primer complexes required distinguishing between bound and free enzyme. This distinction was possible by using a trap that bound to the free enzyme while allowing primer extension by RT bound to the tRNA-vRNA complexes (33, 49-52). Such methodology requires extension of tRNA to be faster than dissociation of the bound enzyme or at least of the same order of magnitude. Thus, to maximize the sensitivity of this assay, we used a high dCTP concentration (500 µM). Indeed, we found the K_d of dCTP for the natural tRNA-wild type vRNA to be ~250 µM,² i.e. more than 1 order of magnitude lower than when using a synthetic RNA primer (36). This method for K_d determination took only productively bound RT into account. This was especially important because our large template-primer complexes might contain unspecific RT-binding sites.

Examples of binding curves are shown in Fig. 3, and the data obtained with all of the mutant templates are summarized in Table I. For most templates, the maximal fraction of extended primer was about 0.7 (Fig. 3 and data not shown). Noticeable exceptions were RNAs 131-196, J0 (Fig. 3), J5, and to a lesser extent Hx1 (not shown). In these experiments, an increased dissociation rate and/or a decreased polymerization rate result in a decreased primer extension. The extension of the primer can be further reduced if a significant amount of the primer-template complex adopts a conformation to which RT cannot bind.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Determination of the equilibrium binding constant (Kd) of RT to the wild type and mutant vRNA-tRNA complexes. The fraction of extended primer was plotted as a function of the RT concentration and used to fit the following theoretical function. (Eq. 1) where the fitted parameters were Kd and f, the maximal fraction of extended primer, and E and S represent the RT and primer-template concentrations, respectively. The Kd values obtained from the fits were 21 nM for RNA 1-311 (closed circles), 12 nM for RNA CompUA (open squares), 14 nM for RNA J0 (open circles), and 94 nM for RNA 131-196 (closed squares).

View this table: [in this window] [in a new window]   Table I Summary of the binding and kinetic parameters of the WT and mutant tRNA/vRNA complexes

All of the K_d measurements were performed in parallel with a single batch of RT and freshly purified RNA templates. Under these conditions, the reproducibility of the experiments was very good, and the errors on the K_d values were low, thus allowing a relative comparison between the wild type and mutant RNAs. The K_d of the RNA 1-311-tRNA:RT complex was 21 ± 4 nM. Overall, the K_d values for the mutant complexes were all within 1 order of magnitude. Because when the natural RNA is replaced by an 18-mer RNA complementary to the PBS, the K_d of RT for the 18-mer RNA-vRNA is much greater than 200 nM (36), unspecific interactions alone cannot confer strong RT binding to the RNA-RNA complexes. Thus, our results suggest that RT binding to the initiation complex involves a number of RNA structural elements that independently contribute to binding.

A rather unexpected finding was the strong contribution of the template regions outside nucleotides 123-217 to RT binding. Indeed, the K_d of RT for the RNA 123-217-tRNA complex was 3-fold higher than for the RNA 1-311-tRNA complex (Table I), even though these two complexes adopt the same conformation, and the latter one contains the most regions of the template protected from RNase cleavage upon RT binding (16). To evaluate the contribution of the domains outside the 123-217 region in RT binding, we measured the dissociation constant of the RNA 123-311-tRNA:RT complex (data not shown). This template lacks the trans-activating region and polyadenylation stem-loops (53), the former having been implicated in reverse transcription (54, 55). The intermediate K_d of this complex, 39 nM, suggested that sequences both upstream and downstream of the 123-217 region stabilized RT binding to the template-primer complex.

The K_d of the RNA 131-196-tRNA:RT complex further increased to 94 nM (Table I), indicating a role for helices 1 and/or 8 in RT binding, in agreement with our footprinting data (16). Selective deletion of either helix 1 or 8 indicated that helix 1 played a major role in the RT binding affinity, whereas the K_d of RNA Hx8-tRNA:RT complex was lower than that of RNA 123-217-tRNA:RT complex (Table I). The low affinity of RT for the RNA Hx1-tRNA and RNA 131-196-tRNA complexes might explain the low extension of the primer observed at the plateau (Fig. 3 and data not shown).

Substitution of nucleotides 132-139 decreased the RT binding affinity by 2.7-fold, indicating the importance of helix 2 for RT binding. The wild type structure but not the sequence of helix 2 appeared important for binding; indeed, RT bound with a 2-fold increased affinity to the RNA CompCG-tRNA and RNA CompUA-tRNA complexes (Table I).

Mutant templates S162-167, 140-167, and 132-175 showed that the intramolecular helices 6C, 5D, and 3E significantly contributed to RT binding to primer-template complex (Table I). Furthermore, the binding affinity of RT gradually decreased as the deletion of the viral RNA increased. This result indicates that in mutant S162-167, nucleotides 132-175 did not produce a steric clash preventing binding of RT when not engaged into intermolecular interactions, because their deletion did not favor binding. Furthermore, the weaker RT binding to the RNA 132-175-tRNA complex, as compared with the RNA 140-167-tRNA complex, is consistent with the importance of helix 2 for this process.

Finally, analysis of mutants J0 and J1 revealed that the junction between helices 2 and 7F can be shortened or even totally deleted without reducing the affinity of RT for the resulting template-primer complexes (Table I). On the contrary, decreasing the length of this junction increased the binding affinity by 1.5-2.3-fold. On the other hand, increasing the length from three to five nucleotides, as in mutant J5, strongly decreased RT binding. For mutant J0, the high affinity of RT for the primer-template complex, combined with the decreased amplitude at the plateau of the binding curve (Fig. 3) suggested either a decreased polymerization rate or/and a significant portion of the complex adopting an inactive conformation (see below).

Rate of the Initial Extension of the Wild Type and Mutant Template-Primer Complexes-- We next compared the synthesis of the () ssDNA using either the wild type or mutant template-primer complexes (Fig. 4). We were particularly interested in the rates of the initial extension of the primer and of the () ssDNA synthesis and in the pausing pattern observed during addition of the first nucleotides to tRNA.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Analysis of the kinetics of () ssDNA synthesis on representative mutant templates by gel electrophoresis. Radiolabeled tRNA (10 nM) was hybridized to the wild type or mutant template RNA indicated above the gels, and DNA synthesis was performed in the presence of 25 nM RT and 50 µM of each dNTP. The aliquots were taken at increasing time points and analyzed by denaturing polyacrylamide gel electrophoresis. Lanes 1-12 correspond to reverse transcription for 0, 15, 30, 45, 60, 150, 300, 600, 900, 1200, 1500, and 1800 s, respectively.

The semi-logarithmic plots of the fraction of unextended primer as a function of time revealed a bi-exponential process (Fig. 5). Previous presteady state and steady state kinetics of the wild type HIV-1 RNA/tRNA indicated that the fast reaction most likely corresponds to extension of tRNA by the preformed primer-template:RT complex, whereas the slow step requires RT recycling and a conformational change of the primer-template complex (33, 35). Accordingly, the rate constant of fast process (k_fast) we measured here during () ssDNA synthesis (Table I) is close to the polymerization rate of the first nucleotide (k_pol) that we previously determined using similar nucleotide concentrations (33, 35).

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Kinetics of the initial extension of tRNA. Semi-logarithmic plots of the fraction of unextended tRNA as a function of time. A, the primer was annealed to RNA 1-311 (closed circles), RNA CompUA (open circles), RNA 132-175 (closed squares), or RNA S162-167 (open squares). B, the primer was annealed to RNA 1-311 (closed circles), RNA J0 (open circles), RNA J1 (closed squares), and RNA J5 (open squares). The curves correspond to the best fit to the following bi-exponential equation: (Eq. 2) where kfast and kslow are the rate constants of the fast and slow processes, respectively, and A and B are the amplitudes of these reactions. The parameters derived from these fits are listed in Table I.

We previously showed that the extension rate constant of an 18-mer RNA complementary to the PBS is similar to that of natural tRNA, indicating that the complex structure of the natural primer-template complex only modestly affects k_pol. However, the amplitude of the reaction was dramatically reduced as a result of inefficient formation of the ternary complex (33, 35). Thus, we expected only minor effects on the k_pol of the mutations we introduced in the viral RNA. Indeed, most mutant template-primer complexes displayed only moderate reductions (up to 4-fold) in k_pol (Table I). However, templates S162-167 and J0 were two noticeable exceptions for which k_pol was reduced by 20-30-fold (Table I and Fig. 5).

The slow extension rate observed with RNA S162-167 suggested that nucleotides 140-167 prevented correct positioning of RT when they were not involved in intermolecular interactions (even though they did not prevent RT binding per se, see above). Accordingly, the k_pol values of tRNA hybridized to RNAs 140-167 and 132-175 were within the same range as the k_pol of the wild type complex (Table I and Fig. 5A). Similarly, our data could suggest that deleting the complete junction between helices 2 and 7F prevented correct positioning of RT, even though K_d did not increase. Alternatively, the slow polymerization rate observed with J0 might be linked to the fact that the first template nucleotide to which the incoming nucleotide must base pair is involved in helix 2. According to previous studies, the second explanation is most likely to be the correct one (35, 56). Indeed, insertion of a single nucleotide between helices 2 and 7F was sufficient to increase k_pol by 1 order of magnitude (Table I and Fig. 5B).

The k_pol value was not the only factor that determined the overall primer extension rate. Indeed, the amplitude associated with the fast extension reaction (A_fast) varied by a 3.5-fold among the templates we tested (Table I). The compensatory mutants CompCG and CompUA had A_fast values similar to that of RNA 1-311. On the other hand, all of the mutants that were unable to form the extended intermolecular interactions had A_fast values that were significantly decreased compared with the 1-311 template. In addition, all of the mutant templates with increased K_d displayed decreased A_fast values, even though the binding affinity was not the sole factor determining the amplitude of the fast process (Table I). The very low A_fast value observed with RNA J5, combined with the intermediate K_d value, suggests that a significant fraction of this primer-template complex adopted a conformation that could not productively bind RT, in keeping with the low plateau observed in the binding curve of this mutant (see above).

Pausing Pattern during Initiation of Reverse Transcription-- Next, we examined the pausing pattern arising during the initiation of reverse transcription and the efficiency of () ssDNA synthesis. Indeed, strong pausing is a hallmark of the initiation phase of reverse transcription (32). With the wild type primer-template complex strong pausing was observed after addition of the third nucleotide (position +3) and to a lesser extent at position +5 (32) (Fig. 4). In addition, pausing was observed at +14 to +16, in the A-rich sequence forming helix 6C in the secondary structure model, the main pausing site being at +16 (32) (Fig. 4).

Strong pausing during initiation was also observed with the mutant templates, even though the position and relative intensity of the pauses varied between templates (Fig. 4). When using mutants J0, J1, and J5, the position of these pauses was shifted: with J5, strong pausing was observed at +5, and weak pauses were observed at +7 and +16 to +18 (Fig. 4). With J1, the strongest pauses were observed at +1, +3, and +14, and weaker pauses were observed at +2, +4, +12, and +13. Finally, with J0, a strong pause was observed at +2, and a weak one was observed at +13. In the latter case, the slow extension of the primer (Figs. 4 and 5B) may also be regarded as a pause at position 0.

The question arises of whether these pauses are sequence-dependent, structure-dependent, or a combination of both. In the wild type, J5, J1, and J0, strong pausing was observed just before entering helix 2 (at +3, +5, +1, and 0, respectively) and after copying two bases of it (at +5, +7, +3, and +2, respectively), despite variations in the junction sequence. On template S132-139, in which helix 2 was disrupted while keeping the wild type sequence around the pausing sites (Fig. 1), strong pausing was observed at +3 but not at +5 (Fig. 4). These results indicated that helix 2 was not required to observe pausing at +3 and suggested that pausing at +3, but not at +5, was mainly directed by the template sequence.

Efficiency of () ssDNA Synthesis-- Strong pausing during the initiation of reverse transcription was observed with all RNA templates (Fig. 4). Thus, differences in pausing could not be the main factor explaining the differences in the efficiency of () ssDNA synthesis we observed among the templates we tested (Fig. 4 and Table I). Indeed, the results listed in Table I suggested that the efficiency of () ssDNA synthesis was a complex function of K_d, k_fast, and A_fast. To evaluate the relative importance of each of these factors, we plotted the efficiency of () ssDNA synthesis as a function of 1/K_d, k_fast, and A_fast (Fig. 6). The best correlation was observed when the relative amount of () ssDNA was plotted versus A_fast (r = 0.73), and the worst correlation was observed when it was plotted versus 1/K_d (r = 0.42), with k_fast yielding an intermediate correlation (r = 0.60).

View larger version (10K): [in this window] [in a new window]   Fig. 6.   Factors effecting the efficiency of () ssDNA synthesis. The relative () ssDNA yield obtained with the wild type and mutant RNA template was plotted as the function of 1/Kd (A), kfast (B), and Afast (C). A line corresponding to the best linear fit is drawn in each panel. The value of the regression coefficient (r) was 0.42 (A), 0.60 (B), and 0.73 (C), respectively. In C, the open symbols correspond to templates 123-217, 131-196, 1, and 8.

In the plot of () ssDNA versus A_fast, three RNA templates displayed similar low () ssDNA synthesis (8-10% of wt), despite having A_fast values ranging from 26 to 62% (Fig. 6C). The two templates that generated low levels of () ssDNA despite having high A_fast values were those with the lowest k_fast values (templates S162-167 and J0) (Table I and Fig. 6C). In the plot presented in Fig. 6C, templates 123-217, 131-196, Hx1, and Hx8 (open symbols) were all situated to the left of the straight line; they generated more () ssDNA than expected from their A_fast values. These were the shortest templates. However, the polymerization length only moderately affected the yield of () ssDNA.

The fact that the efficiency of the () ssDNA synthesis was not strongly correlated with the RT affinity for the template-primer complex might appear surprising. This was probably due to the relative RT and template-primer complex concentrations used in our assay (25 and 10 nM, respectively). Indeed, from the plateau in the binding curves (Fig. 3), one can deduce that at least 70% of RT was active. Thus, active RT was in excess, as compared with the template-primer complex. On the other hand, the predominant importance of the amplitude of the initial primer extension (that took place within seconds) on the efficiency of () ssDNA synthesis that was measured after a 25-min reaction was unpredicted. It suggests that after extension of tRNA and dissociation of RT at the strong pausing sites (33), RT rebound preferentially to the elongated complexes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Initiation of reverse transcription is crucial for HIV-1 replication. We previously proposed a secondary structure model of the HIV-1 Mal template-primer complex (13) (Fig. 1) that gained further support from cell culture experiments (26-28, 31, 40, 57-61). Later, we proposed a three-dimensional model of the HIV-1 initiation complex accounting for additional probing and footprinting data (16). Modeling and footprinting data suggested that HIV-1 RT directly interacts with helices 7F, 1, and 8, and with the three nucleotide junction between helices 7F and 2 and that helix 2 could also contribute to RT binding. In addition, we observed no interaction between RT and the intermolecular helices 6C, 5D, and 3E.

Several groups, including ours, introduced mutations in the vRNA in the vicinity of the PBS and studied their effects on the initiation of reverse transcription. However, these studies were conducted using disparate systems and experimental conditions, thus preventing quantitative comparison. In most studies the structural effects of the mutations was not examined, further complicating the interpretation of the results. Therefore, we estimated the relative contribution of the secondary structure elements of the initiation complex, using the two- and three-dimensional models to design the mutations that we introduced in the vRNA.

Deletion of helix 8 (Fig. 1) affected neither the affinity of RT for the template-primer complex nor the rate of initial tRNA extension but significantly reduced the efficiency of () ssDNA. The limited contribution of helix 8 to RT binding is consistent with our modeling study. Even though helix 8 was close enough to the RT to be partially protected from cleavage by bulky nucleases, no strong contact could be established between the protein and this helix (16). This helix is rather unstable (13), and its sequence is poorly conserved among HIV-1 isolates (hiv-web.lanl.gov/seq-db.html). This helix is external to the central core of the template-primer complex, and no rearrangement of this core was expected upon deletion of helix 8, hence explaining its moderate effects.

By comparison, deletion of helix 1 had more severe adverse effects; it strongly reduced the affinity of the polymerase for the template-primer complex and decreased synthesis of () ssDNA by more than 2-fold. These results are in keeping with our three-dimensional model of the initiation complex that displayed helix 1 simultaneously interacting with the two RT subunits, even though the interaction surface was rather limited (16). In addition, helix 1 is formed by long distance base pairing, and it locks the secondary structure template-primer model (Fig. 1). Thus, its deletion might also affect the relative orientation of the other structural elements interacting with RT.

Alternatively, our results could be accounted for by the intermolecular interaction proposed by Beerens et al. (39). These authors proposed that initiation of reverse transcription is enhanced by an interaction between nucleotides ¹²¹GACUCUGG¹²⁸ of the template, which are involved in helix 1 in our secondary structure model (numbering is according to the Mal isolate), and nucleotides ^48 m5C^m5CAGGGT_m⁵⁵ in tRNA. However, this interaction is not in agreement with our probing experiments, because tRNA A50 was strongly reactive toward chemical probes, in agreement with our model (13). These results might be due to the use by Beerens et al. (39) of the Lai isolate, whereas probing of the binary complex was performed on the Mal isolate (13). Alternatively, the interaction proposed by Beerens et al. (39) might only exist transiently and hence could not be detected by chemical probing. We therefore analyzed the sequence of HXB2 HIV-1 mutants adapted to replicate using either tRNA^His (26) or tRNA^Met (27) after long term culture. We found no evidence of evolution of the mutant viruses to adapt their sequence to the T-arm of their primer. Thus, the intermolecular interaction proposed by Beerens et al. (39) remains to be proven.

Our three-dimensional model of the initiation complex suggested that helix 2 and the junction between helices 2 and 7F might directly interact with RT (16). The mutations we introduced in helix 2 showed that the structure but not the sequence of this helix is important for RT binding. This conclusion is in agreement with a recent publication by Rong et al. (62). These authors showed that disrupting helix 2 dramatically decreased synthesis of long DNA products. We observed only a limited effect on () ssDNA synthesis, but this apparent contradiction is explained by the fact that these authors used much more stringent conditions, including a very low nucleotide concentration, than we did (62). DNA synthesis was restored when compensatory mutations restoring base pairing were introduced (62).

Our results also highlight the pivotal role played by the three-nucleotide junction between helices 2 and 7F. This junction, together with the intermolecular interactions, dictates the relative orientation of helices 1, 2, and 7F (the PBS helix) (16). It can be reduced to one nucleotide without noticeable effects, except for the () ssDNA synthesis, which was significantly diminished. However, complete deletion of the junction or its lengthening by only two nucleotides dramatically compromised DNA synthesis. The adverse effects on K_d and A_fast that we observed when the junction length was increased by two nucleotides suggest that the relative orientation of helices 2 and 7F is crucial for productive positioning of RT on the initiation complex. Similar effects were observed by Rong et al. (62) when they increased the junction length by three or six nucleotides.

The last mutants we analyzed were designed to test the importance of the intermolecular helices 6C, 5D, and 3E (Fig. 1). All three mutants (S162-167, 140-167, and 132-175) displayed strongly reduced RT binding and () ssDNA synthesis. In addition, mutant S162-167 had a 25-fold reduced k_fast. The three-dimensional model of the initiation complex suggested two possible nonexclusive roles for the intermolecular template-primer interactions (16). Together with the junction between helices 2 and 7F, they might fix the position of the structural elements that directly interact with RT in a correct orientation. In addition, by structuring template nucleotides 140-167 and primer nucleotides 33-46, they might also prevent steric clashes between RT and these sequences. Our results suggested that both hypotheses are indeed correct. All three mutant-destroying helices 6C, 5D, and 3E dramatically reduced RT affinity and () ssDNA synthesis, even though RT does not interact directly with these helices. On the other hand, mutant S162-167, but not 140-167 and 132-175, had a strongly reduced k_fast (Table I), suggesting that template nucleotides 140-167 prevented correct positioning of RT when the intermolecular interactions were not formed. The pronounced defects observed with this mutant are in line with the numerous in vivo studies pointing toward the importance of the complementarity between the so-called A-rich loop and the anticodon loop of the primer tRNA (26-28, 31, 40, 57-61).

Taken together, our results are in keeping with our previously derived three-dimensional model (16). In addition, they point toward an unexpected and important conclusion; they demonstrate the overwhelming importance of the overall three-dimensional structure of the initiation complex. Indeed, perturbations of the intermolecular vRNA-tRNA interactions or of the junction between helices 2 and 7F, which lock the tertiary structure of the complex, more severely affected reverse transcription than mutations of one of the helices to which RT directly binds. Thus, this junction and the intermolecular interactions constitute promising targets for anti-initiation specific drugs. Unmodified and modified antisense oligonucleotides have demonstrated the validity of this approach (63-65). Undoubtedly, increasing knowledge on the initiation complex and diversification of the high throughput screening techniques should allow identification of other classes of inhibitors selectively targeting the initiation of reverse transcription.

	ACKNOWLEDGEMENTS

We thank Catherine Isel for helpful discussions and critical reading of the manuscript and Guillaume Bec and Gérard Keith for purification of tRNA.

	FOOTNOTES

^* This work was supported by a grant from the Agence Nationale de la Recherche contre le Sida and a Jeunes Equipes grant from the CNRS.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom correspondence should be addressed. Tel.: 33-3-88-41-70-91; Fax: 33-3-88-60-22-18; E-mail: r.marquet@ibmc.u-strasbg.fr.

Published, JBC Papers in Press, August 22, 2002, DOI 10.1074/jbc.M205295200

² M. Rigourd, unpublished results.

	ABBREVIATIONS

The abbreviations used are: RT, reverse transcriptase; HIV-1, human immunodeficiency virus, type 1; vRNA, viral RNA; () ssDNA, () strand strong stop DNA; PBS, primer binding site; DMS, dimethyl sulfate.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES