| Nucleic Acids Research | Pages |
TAR-RNA binding by HIV-1 Tat protein is selectively inhibited by its l-enantiomer
Introduction
Materials And Methods
Oligoribonucleotides synthesis
UV experiments
CD experiments
Gel-shift experiments
Results
Chemical synthesis and characterization of TAR-oligoribonucleotides
Tat binds d- and l-TAR with the same apparent affinity and discriminates mutants
Discussion
Acknowledgement
References
TAR-RNA binding by HIV-1 Tat protein is selectively inhibited by its l-enantiomer
ABSTRACT
INTRODUCTION
The HIV-1 Tat protein, required for efficient virus production, stimulates the elongation efficiency of RNA polymerase II (1). Moreover, Tat can be released from acutely infected cells as a biologically active protein (2), which, as indicated by a substantial body of evidence, contributes to the pathogenesis of AIDS-associated diseases (3) like Kaposi's sarcoma, and is suggested to participate in the induction of apoptosis even in uninfected lymphocytes (4). Because of the pleiotropic role played by this viral protein, any Tat ligand endowed with good affinity and specificity could be valuable both as a research tool and as a potential lead compound for therapeutical application.
The primary mechanism of Tat activation of HIV-1 replication is mediated by its binding to a stem-loop region (TAR) present near the 5[prime] terminus of all retroviral mRNAs (5,6). Details of the stoichiometric binding of Tat to TAR have become available which define a trinucleotide bulge within TAR together with flanking base pairs to form a unique binding pocket (7-11). Given the very tight association of the Tat-TAR complex (Kd of 1-3 nM), a TAR decoy, i.e. a short synthetic RNA, which features the wild-type sequence of the Tat binding site, should have the potential to function as an efficient and specific ligand for the viral Tat protein. In fact, this strategy has been applied successfully to inhibit HIV-1 replication in cells transduced with retroviral vectors and overexpressing the TAR sequence (12,13). Since our initial aim was to develop a Tat ligand that could be delivered to cells in culture and bind both intracellular and secreted protein, a native TAR decoy would have been unsuitable, given the well known instability of the phosphodiester backbone in biological media (14). The so called `mirror-image' DNAs and RNAs are probably the most nuclease-resistant oligonucleotide analogs presently known (15). These compounds are also called enantio- or l-nucleic acids, since they contain, in place of the natural d-ribose, its enantiomer l-ribose. We (16) and others (17-22) have already evaluated their potential as antisense and antigene agents for the control of gene expression, but, with the exception of highly biased sequences (20-22), they do not anneal to natural nucleic acids (23). To our knowledge, a potential interaction of l-DNAs and -RNAs with nucleic acid binding proteins has never been reported, we tested the l-enantiomer of a shortened form of the native TAR sequence for its affinity to the viral protein. According to a generally accepted view (24), the structural complementarity between natural nucleic acids and proteins, i.e. d-nucleic acid forming complexes with l-proteins (as d-l), would imply a highly reduced affinity of the diastereoisomeric complexes l-l and d-d. Nevertheless, aside from this inference, we took into account the following considerations to support our working hypothesis: (i) the structural complementarity between d-nucleic acids and l-proteins, an extension of the `key and lock' analogy (25), coined by Emil Fisher in 1894 to illustrate enzyme specificity, is considered particularly compelling when the enzyme substrate is a chiral molecule. However, exceptions to the enantiospecific behavior of enzymes have been described for the phosphotransferases of carrot and human prostate acting on uridine (26), and for HSV-1 TK and human dCK which phosphorylate efficiently both enantiomers of thymidine and 2[prime]-deoxycytidine, respectively (27,28). Moreover l-form nucleosides have been described as RT inhibitors and anti-HIV drugs (29). These findings can be interpreted by Koshland's `induced fit' theory of enzyme activity, which, while maintaining the Fisher's concept of shape complementarity of enzyme and substrate, introduced the idea of enzyme flexibility, whereby the mutual fitting of the interacting partners is described by the `hand-in-glove' metaphor (30). (ii) The idea that natural nucleic acids have generally higher affinities for l- than for d-proteins had already been experimentally disproved in at least two cases, in which l- or d-polypeptides were shown to bind nucleic acids with equal affinities (31,32). Moreover, while the present study was progressing, it was reported that the l- and d-enantiomers of a 36 amino acid peptide, which contains the RNA-binding domain of Tat, bind TAR RNA with very close affinities (33). (iii) The interaction of the M13 ssDNA binding protein with d-(dA)6 and l-(dA)6 is merely selective for the chirality of the sugar-phosphate backbone (A.Garbesi, unpublished observations). (iv) It has been shown that the RNA binding domain of HIV Tat forms in solution an extended, unstructured loop (34). Upon contact with the flexible TAR-RNA (35), both interacting molecules assume specific structural rearrangements, which lead to the observed tight complex. Hence, the high stability of the Tat-TAR complex can be seen as the end point of a process of mutual fitting and refolding, rather than as the result of a built-in shape complementarity between the interacting partners. This view, reminiscent of the `induced-fit' model for enzyme-substrate interactions, led us to the idea that the very flexible arginine-rich Tat domain could recognize the `moldable' contact site of TAR, independently of its particular chiral sense.
MATERIALS AND METHODS
Oligoribonucleotides synthesis
All the oligonucleotides used in this work were synthesized by solid phase on a Pharmacia Gene Assembler II plus using standard phosphoramidite method at 0.2 and 1 µmol scales. The d-amidites were purchased from ABI, while l-amidites were prepared as follows. The l-nucleosides were obtained from l-arabinose by epimerization to l-ribose, followed by conversion in three steps to 1-O-acetyl-2,3,5-tri-O-benzoyl-[beta]-ribufuranoside (36), and reaction with in situ silylated N-protected bases (37). After complete deprotection, the nucleosides were protected at their exocyclic amino group with benzoyl for adenosine and cytidine, and with isobutyryl for guanosine (38). The base-protected nucleosides were finally converted in three steps to their 5[prime]-O-terbutyldimethylsylyl-3[prime]-O-(2-cyanoethyl-N,N[prime]-diisopropyl)-phosphoramidites following usual procedures for d-isomers (39,40). Removal of the protecting groups was performed as described by Sproat et al. (41). The crude oligonucleotides were purified by anion exchange HPLC on a Waters 600E HPLC with a DEAE 5PW column, at 30°C, flow rate 1 ml/min and gradient: 0.00 min, 100% buffer A; 40 min, 60% buffer A (buffer A: 20 mmol LiClO4, 20 mmol NH4OAc in H2O/CH3CN 90:10; buffer B: 600 mmol LiClO4, 20 mmol NH4OAc in H2O/CH3CN 90:10).
UV experiments
UV melting experiments were conducted on a Perkin-Elmer 554 spectrophotometer equipped with a MGW Lauda thermostat and a MGW Lauda R40/2 digital thermometer. An electronic device was used to generate a linear temperature gradient (0.6°C/min). The oligoribonucleotide was dissolved in a buffer containing 0.1 M Tris-HCl, 0.1 M NaCl at pH 7.0; the solution was heated at 90°C for ~30 min, allowed to cool slowly to room temperature, then kept at 4°C for a few hours, before starting the melting experiments. The half-melting temperature was determined from the first derivative of the denaturation curve.
CD experiments
Annealing of the oligoribonucleotides was performed as described for UV melting experiments. CD spectra of the annealed oligoribonucleotides were recorded with a Jasco J700A spectropolarimeter in 0.1 M NaCl and 0.1 M Tris-HCl buffer, pH 7.0, at room temperature.
Gel-shift experiments
Recombinant Tat protein was prepared and binding assays were carried out as previously described (9). 33mer d-TAR oligoribo-nucleotide was labeled using [[gamma]-32P] (3000 Ci/mmol) (Amersham, UK) and T4 Polynucleotide kinase (New England Biolabs, Beverly, MA). Competition binding assays (25 µl) included 0.1% Triton X-100, 50 mM Tris-HCl (pH 8.0), 20 mM KCl, 0.1 M DTT, 12.5 nM 32P-labelled 33mer d-TAR RNA (20 000 c.p.m.), 15-20 nM Tat protein and various concentrations of unlabelled competitor RNA. Binding reactions were incubated on ice for 15 min, and non-denaturing PAGE was carried out at 4°C (7.2% gel, 20 × 20 cm) in 0.5× TB (TB = 44.5 mM Tris base, 44.5 mM boric acid), 0.1% Triton X-100 for 1 h at 200 V. Gels were dried and complexes detected by autoradiography. Relative competition constants were calculated after densitometrically scanning the resulting autoradiographs using a Biorad GS-700 imaging densitometer.
Figure 1. (a) Secondary structure of TAR33 RNA. The two terminal GC base pairs at the bottom of the stem are included for enhanced stability. A bold `T' at the 3[prime] terminus indicates that this residue for synthetic reasons is always in d-form. Underlined bases are involved in the formation of a Tat-binding pocket. An inactivating U->C mutation at position 23 is indicated. (b) Thermal denaturation profile of both enantiomeric forms of TAR33.
RESULTS
Chemical synthesis and characterization of TAR-oligoribonucleotides
All sequences were obtained by solid-phase synthesis using d- or l-phosphoramidites. Due to the use, in all syntheses, of a commercial support loaded with d-thymidine, two unpaired nucleotides are present at the 3[prime] terminus. The penultimate residue has the same chirality of the main sequence and was introduced to avoid skeletal deformation in the bottom part of the stem of the l-TAR RNAs. Sequence and characteristic features are depicted in Figure
Figure 2. CD (top) and absorption (bottom) spectra of equimolar solutions of l- and d-TAR33
Tat binds d- and l-TAR with the same apparent affinity and discriminates mutants
As a model of the natural TAR sequence (59 residues), we used, as described by others (6-10), a shorter synthetic oligoribonucleotide (Fig.
Figure This demonstrates clearly that both TAR RNAs from the l- and d-series provide a suitable binding site for Tat. In order to study the selective interaction between Tat and l-TAR RNA further, we introduced a more subtle change into synthetic l- and d-RNAs. Although we could demonstrate that Tat protein binds to synthetic l-TAR with wild-type affinity, it seemed crucial to confirm that the main contributor to this interaction was not just an unselective protein-bulged-RNA binding. Therefore, we performed competition binding experiments, this time comparing l- and d-TAR as mutants, in which the uridine residue at position 23 was replaced by a cytidine residue (Fig.
Figure 3. (a) Competition gel shift for the Tat-l-TAR complex. Control lanes, `-Tat' includes 15 nM of 32P-labeled d-TAR33 RNA; in the `0' controls and throughout the competition lanes, Tat is present at a constant concentration (~20 nM). In the adjacent lanes, cold competitor RNAs (as marked above the autoradiographies) were co-incubated at indicated concentrations. (b) Competition curves obtained by quantitative densitometric analysis of the gels shown in (a). Shift intensity is plotted as a function of concentration of RNA competitor as indicated.
DISCUSSION
Very recently, Nolte et al. reported an elegant application of the SELEX methodology, which led to the identification of l-oligonucleotides that interact with a nucleic acid binding peptide of HIV Tat (42). This and other published examples have proven that SELEX can be a powerful tool for the selection of nucleic acid sequences having very high affinity for a variety of biological molecules, including HIV-1 reverse transcriptase (43). Nolte's 38mer l-RNA, originally developed as a ligand for l-arginine, is sequence-wise unrelated to TAR and binds to a short peptide (12 residues) corresponding to the basic region of HIV-1 Tat protein with a Kd of 26 µM. In contrast, our competition binding experiments demonstrate that the affinity of the mirror-image (l-enantiomer) of wild-type TAR for the Tat protein is 1000-fold better, thereby matching the affinity of the d-sequence. Moreover, our data show that in both cases the stability of the complexes is significantly lowered upon changing the bulge residue U23 of TAR RNA to cytidine, i.e. Tat is able to discriminate a critical bulge mutation in both RNA enantiomers. Thus, a central-and somewhat surprising-finding of our work is that the selective binding of Tat to the TAR element is unaffected by the chirality of the receptor RNA sequence. In agreement with the principles of stereochemistry, which also apply to chiral biopolymers (44), our experimental results show that l- and d-TAR RNAs have the same secondary structure (same thermal stability) and the same spatial folding but in opposite direction (mirror-image CD spectra). This led to the following inference: a similar thermal stability (same Tm values) suggests an identical secondary structure (base pairing) of both enantiomers. The mirror image of the CD-spectra indicates that the spatial folding similarly is a mirror image. Thus, the logical consequence of same sequence (primary structure), same base pairing (secondary structure) and same, but mirror image, folding (tertiary structure) of l- and d-TAR is that they engage in a molecular interaction with the same intrinsic energy content.
In addition, at least the initial event in Tat-TAR recognition, i.e. the fit of the Arg side-chain into the unique binding pocket (formed by G26:C39, U23, and A22:U40) (35,45) and the subsequent conformational RNA change (46), are not expected to differ significantly from natural d-TAR to its mirror image l-TAR. In our study, beyond this entropy-conserving step of cation-[pi]* van der Waals interaction, the peculiar observation that the apparent affinity of the natural complex is retained was less expected. Obviously, our results do not allow us to conclude that every single electrostatic interaction between protein and RNA is kept as such; nevertheless, the sum of binding energies is the same for both enantiomers. This would imply a great `cost-neutral' flexibility of the structure of Tat, since it can accommodate two enantiomeric RNAs and still depend, for both, on the presence of a unique binding pocket (discrimination of U23 mutants).
The findings reported here, combined with those recently published by Huq et al. (33) on the high-affinity interaction of a d-Tat-peptide with d-TAR-RNA, strongly support the view that both partners in the Tat-TAR interaction are moldable. Thus, rather than being modelled as a rigid key-lock, or even as a hand-in-glove fitting, we suggest the Tat-TAR interaction to be best represented as a `hand-in-mitten', in which the thumb would be the intrusion of an Arg side-chain in the binding-pocket of TAR, with the rest of the hand (Tat) fitting, not in the exact same way, but as well as, both right- or left-hand mitten (l- or d-TAR RNA).
Whatever the detailed structure of the Tat-l-TAR complex, which could be further elucidated e.g. by NMR studies,the present finding is, to the best of our knowledge, the first example of a nucleic acid-protein interaction that depends strongly upon a specific tertiary structure of the binding pocket, but not upon its particular chiral sense.
The apparent flexibility of the Tat structure invites the speculation that it could be utilized by the viral protein to alternatively act on other nucleic acid targets like cis-regulatory DNA elements in the viral genome and/or of the genome of the host. If such interactions were to be identified, they could shed light on yet unexplained effects of Tat on the infected cell. Then, beyond preventing the Tat-TAR trans-activation, the l-TAR could be considered as a `multi-task' decoy to divert the protein from other potential function(s) and associated pathologies.
ACKNOWLEDGEMENT
A.G. dedicates this work to the memory of Prof. Antonino Fava.
REFERENCES
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 4 Jun 1998
Copyright©Oxford University Press, 1998.
![]()
CiteULike
Connotea
Del.icio.us What's this?
![]()
![]()

![]()
![]()
![]()
A. LITOVCHICK and R. R. RANDO
Stereospecificity of short Rev-derived peptide interactions with RRE IIB RNA
RNA,
August 1, 2003;
9(8):
937 - 948.
[Abstract]
[Full Text]
[PDF]
![]()
![]()
![]()

![]()
![]()
![]()
F. Hamy, E. Felder, K. Lipson, and T. Klimkait
Merged Screening for Human Immunodeficiency Virus Tat and Rev Inhibitors
J Biomol Screen,
June 1, 2001;
6(3):
179 - 187.
[Abstract]
[PDF]
![]()
![]()
![]()

![]()
![]()
![]()
O. Mauffret, C. El Amri, F. Santamaria, G. Tevanian, B. Rayner, and S. Fermandjian
A two B-Z junction containing DNA resolves into an all right-handed double-helix
Nucleic Acids Res.,
November 15, 2000;
28(22):
4403 - 4409.
[Abstract]
[Full Text]
[PDF]
![]()
This Article ![]()
![]()
Abstract
![]()
Print PDF (140K)
![]()
Alert me when this article is cited
![]()
Alert me if a correction is posted
![]()
Services ![]()
![]()
Email this article to a friend
![]()
Similar articles in this journal
![]()
Similar articles in ISI Web of Science
![]()
Similar articles in PubMed
![]()
Alert me to new issues of the journal
![]()
Add to My Personal Archive
![]()
Download to citation manager
![]()
Search for citing articles in:
ISI Web of Science (24)
![]()
Request Permissions ![]()
Commercial Re-use Guidelines
for Open Access NAR Content
![]()
Google Scholar ![]()
![]()
Articles by Garbesi, A.
![]()
Articles by Klimkait, T.
![]()
Search for Related Content
![]()
PubMed ![]()
![]()
PubMed Citation
![]()
Articles by Garbesi, A.
![]()
Articles by Klimkait, T.
![]()
Social Bookmarking ![]()
![]()
What's this?