3-D models of the antigenomic ribozyme of the hepatitis delta agent with eight new contacts suggested by sequence analysis of 188 cDNA clones.

We mapped 359 mutations at 25 positions in synthetic variants of the antigenomic ribozyme of the hepatitis delta agent by analyzing the sequences of 188 cDNA clones. These data were used to identify three features of the ribozyme: highly conserved nucleotides, positions with restricted nucleotide substitutions and three-dimensional relationships between nucleotides. The distribution of mutations at the 25 positions was as follows: G-11 (the eleventh nucleotide from the cleavage site) was mutated in 56 clones; G-12 in 36; U-15 in 33; C-13 in 26; G-28 in 23; C-27 in 21; C-29 in 19; U-26 in 17; C-18 in 14; A-14 in 13; C-16 in 13; C-19 in 12; U-17 in 11; A-20 in 10; G-42 in 9; G-40 in 7; G-41 in 7; C-24 in 6; U-32 in 6; U-23 in 5; C-25 in 4; C-21 in 3; G-30 in 3; G-31 in 3; C-22 in 1. All clones containing a mutation at C-25 had an A at this position, suggesting that the extra cyclic amino group present in adenine and cytosine may function during the cleavage event. Mutations at certain positions were common in simple clones (containing only one or two mutations), while mutations at other positions were over-represented in more complex clones. Both compensatory base changes and co-mutational frequencies were used to identify eight pairs of nucleotides which may interact with each other: G-11 and C-18, G-12 and C-27, C-13 and G-28, C-21 and U-23/C-24, C-21 and G-30, U-23 and G-31/U-32, C24 and G-30, C-27 and G-42. These pairs, which involve some of the most conserved positions in the molecule, suggest interactions among nucleotides previously depicted in open-loop structures. The newly proposed points of contact between pairs of nucleotides are compatible with both the axehead and pseudoknot secondary structural models and were combined with previously proposed Watson-Crick base paired helices to produce two three dimensional models. In both of these, C-25 and C-76 are placed near the cleavage site.     

Identification of a novel HIV-1 TAR RNA bulge binding protein.

The Tat protein binds to TAR RNA to stimulate the expression of the human immunodeficiency virus type 1 (HIV-1) genome. Tat is an 86 amino acid protein that contains a short region of basic residues (aa49-aa57) that are required for RNA binding and TAR is a 59 nucleotide stem-loop with a tripyrimidine bulge in the upper stem. TAR is located at the 5' end of all viral RNAs. In vitro, Tat specifically interacts with TAR by recognising the sequence of the bulge and upper stem, with no requirement for the loop. However, in vivo the loop sequence is critical for activation, implying a requirement for accessory cellular TAR RNA binding factors. A number of TAR binding cellular factors have been identified in cell extracts and various models for the function of these factors have been suggested, including roles as coactivators and inhibitors. We have now identified a novel 38 kD cellular factor that has little general, single-stranded or double-stranded RNA binding activity, but that specifically recognises the bulge and upper stem region of TAR. The protein, referred to as BBP (bulge binding protein), is conserved in mammalian and amphibian cells and in Schizosaccharomyces pombe but is not found in Saccharomyces cerevisiae. BBP is an effective competitive inhibitor of Tat binding to TAR in vitro. Our data suggest that the bulge-stem recognition motif in TAR is used to mediate cellular factor/RNA interactions and indicates that Tat action might be inhibited by such competing reactions in vivo.     

Inhibition of human immunodeficiency virus type 1 Tat-dependent activation of translation in Xenopus oocytes by the benzodiazepine Ro24-7429 requires trans-activation response element loop sequences.

Two benzodiazepine compounds, [7-chloro-5-(2-pyrryl)-3H-1,4 benzodiazapin-2-(H)-one] (Ro5-3335) and [7-chloro-5-(1H-pyrrol-2-yl)-3H-benzo[e] [1,4] diazepin-2-yl]- methylamine (Ro24-7429), inhibit human immunodeficiency virus type 1 (HIV-1) replication via a specific effect on the function of the transactivator protein, Tat. To gain further insight into the mechanism of action of these compounds, we have tested their effects in an alternative assay for Tat activation in Xenopus oocytes. In this system, translation of trans-activation response element (TAR)-containing RNA is activated by Tat. Both compounds specifically blocked activation of translation in a dose-dependent fashion, with Ro24-7429 showing the greater potency. In the Xenopus oocyte system, as in mammalian cells, mutation of the TAR loop sequences abolishes Tat action. However, it is possible to obtain TAR-specific, Tat-dependent activation of a target RNA with a mutation in the loop provided that this target is in large excess. This result has been interpreted as indicating that a negative factor has been titrated (M. Braddock, R. Powell, A.D. Blanchard, A.J. Kingsman, and S.M. Kingsman, FASEB J. 7:214-222, 1993). Interestingly Ro24-7429 was unable to inhibit the TAR-specific but loop sequence-independent mode of translational activation. This finding suggests that a specific loop-binding cellular factor may mediate the effects of this inhibitor of Tat action. Consistent with this notion, we could not detect any effect of Ro24-7429 on the efficiency of specific Tat binding to TAR in vitro.     

tRNA Binding to Antitumor Drug Doxorubicin and Its Analogue

The binding sites of antitumor drug doxorubicin (DOX) and its analogue N-(trifluoroacetyl) doxorubicin (FDOX) with tRNA were located, using FTIR, CD, fluorescence spectroscopic methods and molecular modeling. Different binding sites are involved in drug-tRNA adducts with DOX located in the vicinity of A-29, A-31, A-38, C-25, C-27, C-28, G-30 and U-41, while FDOX bindings involved A-23, A-44, C-25, C-27, G-24, G-42, G-53, G-45 and U-41 with similar free binding energy (-4.44 for DOX and -4.41 kcal/mol for FDOX adducts). Spectroscopic results showed that both hydrophilic and hydrophobic contacts are involved in drug-tRNA complexation and FDOX forms more stable complexes than DOX with K _DOX-tRNA = 4.7 (±0.5)×10⁴ M⁻¹ and K _FDOX-tRNA = 6.3 (±0.7)×10⁴ M⁻¹. The number of drug molecules bound per tRNA (n) was 0.6 for DOX and 0.4 for FDOX. No major alterations of tRNA structure were observed and tRNA remained in A-family conformation, while biopolymer aggregation and particle formation occurred at high drug concentrations.     

Characterization of the solution conformations of unbound and Tat peptide-bound forms of HIV-1 TAR RNA.

Basic peptides from the carboxy terminus of the HIV-1 Tat protein bind to the apical stem-loop region of TAR RNA with high affinity and moderate specificity. The conformations of the unbound and 24 residue Tat peptide (Tfr24)-bound forms of TAR RNA have been characterized by NMR spectroscopy. The unbound form of TAR exists in major and minor forms having different trinucleotide bulge conformations. A specific TAR RNA conformational change is observed upon complex formation with Tfr24, consisting of coaxial stacking of helical stems and base triple formation. A U23-A27-U38 base triple is proposed based on exchangeable proton NMR data, where U23 forms a base pair with A27 in the major groove. No evidence for base triple formation was found for Tat peptides in which lysine residues are extensively substituted for arginine.     

Evidence for a base triple in the free HIV-1 TAR RNA

We propose the existence of a novel base triple in the HIV-1 TAR hairpin. This triple is supported by covariation of loop residue 31 with residue 22, which is part of an unusual base pair with U40 below the 3-nucleotide bulge. A set of mutants was constructed to test the involvement of bases A22, U31, and U40 in a triple interaction. RNA structure probing, trans-activation assays, and structure modeling are consistent with the existence of this base triple in a bent conformation of the free TAR element. However, disruption of the base triple does not affect binding of a Tat-derived peptide. We therefore compared the structure of free and Tat-bound TAR RNA by footprinting and site-specific cross-linking analyses. These studies indicate that the Tat arginine-rich motif, in addition to its known binding site at the bulge, is in close contact with U31 in the TAR loop. Because binding of Tat to TAR is known to coincide with the formation of a base triple with residues U23, A27, and U38, we hypothesize that Tat binding and the associated straightening of TAR triggers the disruption of the (A22-U40)U31 triple.     

Selective modification of cytidine and uridine residues in Escherichia coli formylmethionine transfer ribonucleic acid

The specificity of methoxyamine for the cytidine residues in Escherichia coli formylmethionine tRNA is described in detail. Of the nine cytidine residues not involved in hydrogen-bonding in the clover leaf model of the tRNA, three are very reactive (C-1, 75 and 76), three less so (C-16, 17 and 35) and three unreactive (C-33, 49 and 57). Surprisingly, residue C-35 at the 3′ end of the anticodon triplet is not completely modified by methoxyamine.The specificity of 1-cyclohexyl 3-[2-morpholino (4)-ethyl] carbodiimide methotosylate for the uridine and guanosine residues of this tRNA is also described in detail. Of the twelve uridine and guanosine residues not involved in hydrogen-bonding in the secondary structure of the molecule, two are reactive (U-37 and48), one less so (U-18), one partially (U-34), and eightare unreactive (U-8 and 61; G-9, 15, 19, 20, 27 and 46). No guanosine residues in the tRNA are modified by the carbodiimide. The ribosylthymine and pseudouridine residues in loop IV are also unreactive. The extent and position of the carbodiimide modification as a function of time is also described.The importance of particular residues being modified or not under the reaction conditions used is discussed in terms of transfer RNA conformation. A reduction from 10 to 4 mm-magnesium ions in the modification experiments has no apparent effect on the extent and position of the carbodiimide or methoxyamine reactions.     

TAR-RNA binding by HIV-1 Tat protein is selectively inhibited by its L-enantiomer.

An oligoribonucleotide, corresponding to the Tat-interactive top half of the HIV-1 TAR RNA stem-loop, was synthesized in both the natural D- and the enantiomeric L-configurations. The affinity of Tat for the two RNAs, assessed by competition binding experiments, was found to be identical and is reduced 10-fold for both, upon replacement of the critical bulge residue U23 with cytidine. It is suggested that this interaction of the flexible Tat protein depends strongly upon the tertiary structure of a binding pocket within TAR, but not upon its handedness, and may be described by a 'hand-in-mitten' model.     

Enzymatic 2''-O-methylation of the wobble nucleoside of eukaryotic tRNAPhe: specificity depends on structural elements outside the anticodon loop.

We have investigated the specificity of the enzyme tRNA (wobble guanosine 2'-O-)methyltransferase which catalyses the maturation of guanosine-34 of eukaryotic tRNAPhe to the 2'-O-methyl derivative Gm-34. This study was done by micro-injection into Xenopus laevis oocytes of restructured yeast tRNAPhe in which the anticodon GmAA and the 3' adjacent nucleotide 'Y' were substituted by various tetranucleotides. The results indicate that the enzyme is cytoplasmic; the chemical nature of the bases of the anticodon and its 3' adjacent nucleotide is not critical for the methylation of G-34; the size of the anticodon loop is however important; structural features beyond the anticodon loop are involved in the specific recognition of the tRNA by the enzyme since Escherichia coli tRNAPhe and four chimeric yeast tRNAs carrying the GAA anticodon are not substrates; unexpectedly, the 2'-O-methylation is not restricted to G-34 since C-34, U-34 and A-34 in restructured yeast tRNAPhe also became methylated. It seems probable that the tRNA (wobble guanosine 2'-O-)methyltransferase is not specific for the type of nucleotide-34 in eukaryotic tRNAPhe; however the existence in the oocyte of several methylation enzymes specific for each nucleotide-34 has not yet been ruled out.     

Conserved nucleotides in the TAR RNA stem of human immunodeficiency virus type 1 are critical for Tat binding and trans activation: model for TAR RNA tertiary structure.

Interaction between the human immunodeficiency virus type 1 (HIV-1) trans-activator Tat and its cis-acting responsive RNA element TAR is necessary for activation of HIV-1 gene expression. We investigated the hypothesis that the essential uridine residue at position 23 in the bulge of TAR RNA is involved in intramolecular hydrogen bonding to stabilize an unique RNA structure required for recognition by Tat. Nucleotide substitutions in the two base pairs of the TAR stem directly above the essential trinucleotide bulge that maintain base pairing but change sequence prevent complex formation with Tat in vitro. Corresponding mutations tested in a trans-activation assay strongly affect the biological activity of TAR in vivo, suggesting an important role for these nucleotides in the Tat-TAR interaction. On the basis of these data, a model is proposed which implicates uridine 23 in a stable tertiary interaction with the GC pair directly above the bulge. This interaction would cause widening of the major groove of the RNA, thereby exposing its hydrogen-bonding surfaces for possible interaction with Tat. The model also predicts a gap between uridine 23 and the first base pair in the stem above, which would require one or more unpaired nucleotides to close, but does not predict any other role for such nucleotides. In accordance with this prediction, synthetic propyl phosphate linkers of equivalent length to 1 or 2 nucleotides, were found to be fully acceptable substitutes in the bulge above uridine 23, demonstrating that neither the bases nor the ribose moieties at these positions are implicated in the recognition of TAR RNA by Tat.     

RNA recognition by Tat-derived peptides: interaction in the major groove?

Replication of human immunodeficiency virus requires binding of the viral Tat protein to its RNA target sequence TAR; peptides derived from Tat bind to a TAR "contact site" spanning 5 bp and a trinucleotide pyrimidine bulge. We find that high affinity binding requires a U residue in the bulge loop and 2 specific adjacent base pairs. Other bulged RNAs bind in a lower affinity nonspecific manner; sequence-specific binding requires a bulge loop of more than 1 nucleotide. Reaction with diethyl pyrocarbonate indicates that one effect of the bulge is to make the otherwise deep and narrow RNA major groove accessible. A model consistent with these data involves local distortion of A-form geometry at the bulge, which bends the helix and permits protein binding and interactive access in the RNA major groove.     

Recruitment of a protein complex containing Tat and cyclin T1 to TAR governs the species specificity of HIV-1 Tat.

Human cyclin T1 (hCycT1), a major subunit of the essential elongation factor P-TEFb, has been proposed to act as a cofactor for human immunodeficiency virus type 1 (HIV-1) Tat. Here, we show that murine cyclin T1 (mCycT1) binds the activation domain of HIV-1 Tat but, unlike hCycT1, cannot mediate Tat function because it cannot be recruited efficiently to TAR. In fact, overexpression of mCycT1, but not hCycT1, specifically inhibits Tat-TAR function in human cells. This discordant phenotype results from a single amino acid difference between hCycT1 and mCycT1, a tyrosine in place of a cysteine at residue 261. These data indicate that the ability of Tat to recruit CycT1/P-TEFb to TAR determines the species restriction of HIV-1 Tat function in murine cells and therefore demonstrate that this recruitment is a critical function of the Tat protein.