This issue contains three varying perspectives on a specific hypothesis for the origin of the genetic code. RNA is pleased to provide a forum for constructive “debates” such as this one. The editors welcome suggestions for additional topics that could benefit from an open discussion of divergent views.

In plant mitochondria, editing of messenger RNA by C-to-U conversions is essential for correct gene expression as it usually improves the protein-sequence conservation between different species or sometimes affects the reading frames (for a review, see Maier et al., 1996). Editing sites have been identified in mitochondrial (mt) RNA of all major groups of land plants, including Bryophytes, Pteridophytes, Prespermaphytes, and Spermaphytes (Hiesel et al., 1994a,b; Malek et al., 1996). Editing mainly affects messenger RNA, but editing sites have also been identified in three transfer RNAs. In dicot mitochondria a C-to-U editing event corrects a C:A mismatch into a U:A base pair in the acceptor stem of tRNAPhe(GAA) (Maréchal-Drouard et al., 1993; Binder et al., 1994). In the gymnosperm Larix leptoeuropaea, three C-to-U conversions restore a U:A base pair in the acceptor stem, D stem, and anticodon stem of tRNAHis(GUG), respectively (Maréchal-Drouard et al., 1996b). The third example described is the Oenothera berteriana mt tRNACys(GCA), where a C28:U42 mismatch is converted into a U28:U42 noncanonical base pair (Binder et al., 1994). In the case of both tRNAPhe and tRNAHis, editing of precursors is a prerequisite for 5′ and 3′ processing to generate a mature tRNA (Marchfelder et al., 1996; Maréchal-Drouard et al., 1996a, 1996b; Kunzmann et al., 1998). The role of editing in the case of tRNACys has not been studied so far, although it has been shown that it occurs at the precursor level (Binder et al., 1994). In this letter, we report an evolutionary and functional study of mt tRNACys(GCA) editing in plant mitochondria.

In the genetic code, triplet codons and amino acids can be shown to be related by chemical principles. Such chemical regularities could be created either during the code's origin or during later evolution. One such chemical principle can now be shown experimentally. Natural or particularly selected RNA binding sites for at least three disparate amino acids (arginine, isoleucine, and tyrosine) are enriched in codons for the cognate amino acid. Currently, in 517 total nucleotides, binding sites contain 2.4-fold more codon sequences than surrounding nucleotides. The aggregate probability of this enrichment is 10−7 to 10−8, had codons and binding site sequences been independent. Thus, at least some primordial coding assignments appear to have exploited triplets from amino acid binding sites as codons.

It has been suggested that in vitro selection experiments can provide information not only on what might have occurred during the evolution of the RNA world, but can in fact yield insights into particular features of the RNA world. In particular, it has been suggested that the sequences of anti-amino acid aptamers can provide clues to the origin of the genetic code, and that there is a statistically significant association between motifs found in aptamers and codons. We argue that the suggested connections between modern motifs and ancient sequences are logically tenuous, and show that there is no statistically meaningful association between motifs found in aptamers and codons.

If the genetic code arose in an RNA world, present codon assignments may reflect primordial RNA–amino acid affinities. Whether aptamers selected from random pools to bind free amino acids do so using the cognate codons at their binding sites has been controversial. Here we defend and extend our previous analysis of arginine binding sites, and propose a model for the maintenance of codon–amino acid interactions through the evolution of amino acids from ribozyme cofactors into the building blocks of proteins.

The ribonuclease P ribozyme (RNase P RNA), like other large ribozymes, requires magnesium ions for folding and catalytic function; however, specific sites of metal ion coordination in RNase P RNA are not well defined. To identify and characterize individual nucleotide functional groups in the RNase P ribozyme that participate in catalytic function, we employed self-cleaving ribozyme–substrate conjugates that facilitate measurement of the effects of individual functional group modifications. The self-cleavage rates and pH dependence of two different ribozyme–substrate conjugates were determined and found to be similar to the single turnover kinetics of the native ribozyme. Using site-specific phosphorothioate substitutions, we provide evidence for metal ion coordination at the pro-RP phosphate oxygen of A67, in the highly conserved helix P4, that was previously suggested by modification-interference experiments. In addition, we detect a new metal ion coordination site at the pro-SP phosphate oxygen of A67. These findings, in combination with the proximity of A67 to the pre-tRNA cleavage site, support the conclusion that an important role of helix P4 in the RNase P ribozyme is to position divalent metal ions that are required for catalysis.

We report the evolution of an RNA aptamer to change its binding specificity. RNA aptamers that bind the free amino acid tyrosine were in vitro selected from a degenerate pool derived from a previously selected dopamine aptamer. Three independent sequences bind tyrosine in solution, the winner of the selection binding with a dissociation constant of 35 μM. Competitive affinity chromatography with tyrosine-related ligands indicated that the selected aptamers are highly L-stereo selective and also recognize L-tryptophan and L-dopa with similar affinity. The binding site was localized by sequence comparison, analysis of minimal boundaries, and structural probing upon ligand binding. Tyrosine-binding sites are characterized by the presence of both tyrosine (UAU and UAC) and termination (UAG and UAA) triplets.

An iterative in vitro splicing strategy was employed to select for optimal 3′ splicing signals from a pool of pre-mRNAs containing randomized regions. Selection of functional branchpoint sequences in HeLa cell nuclear extract yielded a sequence motif that evolved from UAA after one round of splicing toward a UACUAAC consensus after seven rounds. A significant part of the selected sequences contained a conserved AAUAAAG motif that proved to be functional both as a polyadenylation signal and a branch site in a competitive manner. Characterization of the branchpoint in these clones to either the upstream or downstream adenosines of the AAUAAAG sequence revealed that the branching process proceeded efficiently but quite promiscuously. Surprisingly, the conserved guanosine, adjacent to the common AAUAAA polyadenylation motif, was found to be required only for polyadenylation. In an independent experiment, sequences surrounding an optimal branchpoint sequence were selected from two randomized 20-nt regions. The clones selected after six rounds of splicing revealed an extended polypyrimidine tract with a high frequency of UCCU motifs and a highly conserved YAG sequence in the extreme 3′ end of the randomized insert. Mutating the 3′ terminal guanosine of the intron strongly affects complex A formation, implying that the invariant AG is recognized early in spliceosome assembly.

The transfer RNA 5′ maturation enzyme RNase P has been characterized in Bacteria, Archaea, and Eukarya. The purified enzyme from all three kingdoms is a ribonucleoprotein containing an essential RNA subunit; indeed, the RNA subunit of bacterial RNase P RNA is the sole catalytic component. In contrast, the RNase P activity isolated from spinach chloroplasts lacks an RNA component and appears to function as a catalytic protein. Nonetheless, the chloroplast enzyme recognizes a pre-tRNA substrate for E. coli RNase P and cleaves it as efficiently and precisely as does the bacterial enzyme. To ascertain whether there are differences in catalytic mechanism between an all-RNA and an all-protein RNase P, we took advantage of the fact that phosphodiester bond selection and hydrolysis by the E. coli RNase P ribozyme is directed by a Mg2+ ion coordinated to the nonbridging pro-RP oxygen of the scissile bond, and is blocked by sulfur replacement of this oxygen. We therefore tested the ability of the chloroplast enzyme to process a precursor tRNA containing this sulfur substitution. Partially purified RNase P from spinach chloroplasts can accurately and efficiently process phosphorothioate-substituted pre-tRNAs; cleavage occurs exclusively at the thio-containing scissile bond. The enzymatic throughput is fivefold slower, consistent with a general chemical effect of the phosphorothioate substitution rather than with a metal coordination deficiency. The chloroplast RNase P reaction mechanism therefore does not involve a catalytic Mg2+ bonded to the pro-RP phosphate oxygen, and hence is distinct from the mechanism of the bacterial ribozyme RNase P.

Ribonuclease P is the enzyme responsible for removing the 5′-leader segment of precursor transfer RNAs in all organisms. All eukaryotic nuclear RNase Ps are ribonucleoproteins in which multiple protein components and a single RNA species are required for activity in vitro as well as in vivo. It is not known, however, which subunits participate directly in phosphodiester-bond hydrolysis. The RNA subunit of nuclear RNase P is evolutionarily related to its catalytically active bacterial counterpart, prompting speculation that in eukaryotes the RNA may be the catalytic component. In the bacterial RNase P reaction, Mg(II) is required to coordinate the nonbridging phosphodiester oxygen(s) of the scissile bond. As a consequence, bacterial RNase P cannot cleave pre-tRNA in which the pro-RP nonbridging oxygen of the scissile bond is replaced by sulfur. In contrast, the RNase P reaction in plant chloroplasts is catalyzed by a protein enzyme whose mechanism does not involve Mg(II) coordinated by the pro-RP oxygen. To determine whether the mechanism of nuclear RNase P resembles more closely an RNA- or a protein-catalyzed reaction, we analyzed the ability of Saccharomyces cerevisiae nuclear RNase P to cleave pre-tRNA containing a sulfur substitution of the pro-RP oxygen at the cleavage site. Sulfur substitution at this position prohibits correct cleavage of pre-tRNA. Cleavage by eukaryotic RNase P thus depends on the presence of a thio-sensitive ligand to the pro-RP oxygen of the scissile bond, and is consistent with a common, RNA-based mechanism for the bacterial and eukaryal enzymes.

The monoclonal anti-dsRNA antibody J2 binds double-stranded RNAs (dsRNA) in an apparently sequence-nonspecific way. The mAb only recognizes antigens with double-stranded regions of at least 40 bp and its affinity to poly(A) poly(U) and to dsRNAs with mixed base pair composition is about tenfold higher than to poly(I) poly(C). Because no specific binding site could be determined, the number, the exact dimensions, and other distinct features of the binding sites on a given antigen are difficult to evaluate by biochemical methods. We therefore employed scanning force microscopy (SFM) as a method to analyze antibody–dsRNA interaction and protein–RNA binding in general. Several in vitro-synthesized dsRNA substrates, generated from the Dictyostelium PSV-A gene, were used. In addition to the expected sequence-nonspecific binding, imaging of the complexes indicated preferential binding of antibodies to the ends of dsRNA molecules as well as to certain internal sites. Analysis of 2,000 bound antibodies suggested that the consensus sequence of a preferential internal binding site is A2N9A3N9A2, thus presenting A residues on one face of the helix. The site was verified by site-directed mutagenesis, which abolished preferential binding to this region. The data demonstrate that SFM can be efficiently used to identify and characterize binding sites for proteins with no or incomplete sequence specificity. This is especially the case for many proteins involved in RNA metabolism.

Ribonuclease-resistant RNA molecules that bind to infectious human cytomegalovirus (HCMV) were isolated in vitro from a pool of randomized sequences after 16 cycles of selection and amplification. The two ligands (L13 and L19) characterized exhibited high HCMV-binding affinity in vitro and effectively inhibited viral infection in tissue culture. Their antiviral activity was also specific as they only reacted with two different strains of HCMV but not with the related herpes simplex virus 1 and human cells. These two ligands appeared to function as antivirals by blocking viral entry. Ultraviolet (UV) crosslinking studies suggested that L13 and L19 bind to HCMV essential glycoproteins B and H, respectively. Thus, RNA ligands that bind to different surface antigens of HCMV can be simultaneously isolated by the selection procedure. Our study demonstrates the feasibility of using these RNA ligands as a research tool to identify viral proteins required for infectivity and as an antiviral agent to block viral infection.

In eubacteria, base pairing between the 3′ end of 16S rRNA and the ribosome-binding site of mRNA is required for efficient initiation of translation. An interaction between the 18S rRNA and the mRNA was also proposed for translation initiation in eukaryotes. Here, we used an antisense RNA approach in vivo to identify the regions of 18S rRNA that might interact with the mRNA 5′ untranslated region (5′ UTR). Various fragments covering the entire mouse 18S rRNA gene were cloned 5′ of a cat reporter gene in a eukaryotic vector, and translation products were analyzed after transient expression in human cells. For the largest part of 18S rRNA, we show that the insertion of complementary fragments in the mRNA 5′ UTR do not impair translation of the downstream open reading frame (ORF). When translation inhibition is observed, reduction of the size of the complementary sequence to less than 200 nt alleviates the inhibitory effect. A single fragment complementary to the 18S rRNA 3′ domain retains its inhibitory potential when reduced to 100 nt. Deletion analyses show that two distinct sequences of approximately 25 nt separated by a spacer sequence of 50 nt are required for the inhibitory effect. Sucrose gradient fractionation of polysomes reveals that mRNAs containing the inhibitory sequences accumulate in the fractions with 40S ribosomal subunits, suggesting that translation is blocked due to stalling of initiation complexes. Our results support an mRNA–rRNA base pairing to explain the translation inhibition observed and suggest that this region of 18S rRNA is properly located for interacting with mRNA.

Three classes of RNA, represented by atpB and petD mRNAs, Arg and Glu tRNAs, and 5S rRNA, were found to exist in polyadenylated form in Chlamydomonas reinhardtii chloroplasts. Sequence analysis of cDNA clones derived from reverse transcriptase-polymerase chain reaction protocols used to select polyadenylated RNAs revealed that, at least for the mRNAs and tRNAs, there are three apparent types of polyadenylation. In the first case, the poly(A) tail is added at or near the mature 3′ end, even when this follows a strong secondary structure. In the second case, the tail is added to pre-mRNA or pre-tRNA, suggesting a possible competition between polyadenylation and RNA-processing pathways. Finally, in all cases, the poly(A) tail can be added internally, possibly as a part of an RNA-decay pathway. The tails found in Chlamydomonas chloroplasts differ from those of spinach chloroplasts in adenine content, being nearly homopolymeric (>98% adenine) versus 70% in spinach, and are similar in length to those of Escherichia coli, being mostly between 20 and 50 nt. In vitro assays using a Chlamydomonas chloroplast protein extract showed that a 3′ end A25 tail was sufficient to stimulate rapid degradation of atpB RNA in vitro, with a lesser effect for petD, and only minor effects on trnE. We therefore propose that polyadenylation contributes to mRNA degradation in Chlamydomonas chloroplasts, but that its effect may vary.

To investigate the accuracy of a model [Giese et al., 1998, Biochemistry 37:1094–1100 and Mathews et al., 1999, J Mol Biol 288:911–940] that predicts the stability of RNA hairpin loops, optical melting studies were conducted on sets of hairpins previously determined to have unusually stable thermodynamic parameters. Included were the tetraloops GNRA and UNCG (where N is any nucleotide and R is a purine), hexaloops with UU first mismatches, and the hairpin loop of the iron responsive element, CAGUGC. The experimental values for the GNRA loops are in excellent agreement (ΔG°37 within 0.2 kcal/mol and melting temperature (TM) within 4 °C) with the values predicted by the model. When the UNCG hairpin loops are treated as tetraloops, and a bonus of 0.8 kcal/mol included in the prediction to account for the extra stable first mismatch (UG), the measured and predicted values are also in good agreement (ΔG°37 within 0.7 kcal/mol and TM within 3 °C). Six hairpins with unusually stable UU first mismatches also gave good agreement with the predictions (ΔG°37 within 0.5 kcal/mol and TM within 8 °C), except for hairpins closed by wobble base pairs. For these hairpins, exclusion of the additional stabilization term for UU first mismatches improved the prediction (ΔG°37 within 0.1 kcal/mol and TM within 3 °C). Hairpins with the iron-responsive element loop were not predicted well by the model, as measured ΔG°37 values were at least 1 kcal/mol greater than predicted.

NaSSU1 is a complex nuclear group I intron found in several species of Naegleria, consisting of a large self-splicing group I ribozyme (NaGIR2), which itself is interrupted by a small, group I-like ribozyme (NaGIR1) and an open reading frame (ORF) coding for a homing endonuclease. The GIR1 ribozyme cleaves in vitro transcripts of NaSSU1 at two internal processing sites about 400 nt downstream of the 5′ end of the intron, proximal to the endonuclease ORF. Here we demonstrate that self-cleavage of the excised intron also occurs in vivo in Naegleria gruberi, generating an ORF-containing RNA that possesses a short leader with a sequence element likely to be involved in gene expression. To assess the functional significance of self-cleavage, we constructed a genetic system in Saccharomyces cerevisiae. First, a mutant yeast strain was selected with a mutation in all the rRNA genes, rendering the rDNA resistant to cleavage by the Naegleria endonuclease. Active endonuclease, which is otherwise lethal, could be expressed readily in these cells. Endonuclease activity also could be detected in extracts of yeast harboring plasmids in which the endonuclease ORF was embedded in its native context in the intron. Analysis of the RNA from these yeast cells showed that the excised intron RNA was processed as in N. gruberi. A mutant intron constructed to prevent self-cleavage of the RNA failed to express endonuclease activity. These results support the hypothesis that the NaGIR1-catalyzed self-cleavage of the intron RNA is a key event in expression of the endonuclease.

Small catalytic RNAs like the hairpin ribozyme are proving to be useful intracellular tools; however, most attempts to demonstrate trans-cleavage of RNA by ribozymes in cells have been frustrated by rapid cellular degradation of the cleavage products. Here, we describe a fluorescence resonance energy transfer (FRET) assay that directly monitors cleavage of target RNA in tissue-culture cells. An oligoribonucleotide substrate was modified to inhibit cellular ribonuclease degradation without interfering with ribozyme cleavage, and donor (fluorescein) and acceptor (tetramethylrhodamine) fluorophores were introduced at positions flanking the cleavage site. In simple buffers, the intact substrate produces a strong FRET signal that is lost upon cleavage, resulting in a red-to-green shift in dominant fluorescence emission. Hairpin ribozyme and fluorescent substrate were microinjected into murine fibroblasts under conditions in which substrate cleavage can occur only inside the cell. A strong FRET signal was observed by fluorescence microscopy when substrate was injected, but rapid decay of the FRET signal occurred when an active, cognate ribozyme was introduced with the substrate. No acceleration in cleavage rates was observed in control experiments utilizing a noncleavable substrate, inactive ribozyme, or an active ribozyme with altered substrate specificity. Subsequently, the fluorescent substrates were injected into clonal cell lines that expressed cognate or noncognate ribozymes. A decrease in FRET signal was observed only when substrate was microinjected into cells expressing its cognate ribozyme. These results demonstrate trans-cleavage of RNA within mammalian cells, and provide an experimental basis for quantitative analysis of ribozyme activity and specificity within the cell.

Vertebrate TAP and its yeast ortholog Mex67p are involved in the export of messenger RNAs from the nucleus. TAP has also been implicated in the export of simian type D viral RNAs bearing the constitutive transport element (CTE). Although TAP directly interacts with CTE-bearing RNAs, the mode of interaction of TAP/Mex67p with cellular mRNAs is different from that with the CTE RNA and is likely to be mediated by protein–protein interactions. Here we show that Mex67p directly interacts with Yra1p, an essential yeast hnRNP-like protein. This interaction is evolutionarily conserved as Yra1p also interacts with TAP. Conditional expression in yeast cells implicates Yra1p in the export of cellular mRNAs. Database searches revealed that Yra1p belongs to an evolutionarily conserved family of hnRNP-like proteins having more than one member in Mus musculus, Xenopus laevis, Caenorhabditis elegans, and Schizosaccharomyces pombe and at least one member in several species including plants. The murine members of the family directly interact with TAP. Because members of this protein family are characterized by the presence of one RNP-motif RNA-binding domain and exhibit RNA-binding activity, we called these proteins REF-bps for RNA and export factor binding proteins. Thus, Yra1p and members of the REF family of hnRNP-like proteins may facilitate the interaction of TAP/Mex67p with cellular mRNAs.