Positive strand RNA virus replication & RNA binding by eEF1A

This laboratory studies the molecular biology of plant RNA viruses, focussing on two systems: turnip yellow mosaic virus (TYMV) and Qbeta bacteriophage. Our broad aim is to understand the specificity mechanisms governing the selective amplification of viral RNAs. A long-standing interest is understanding the role of the transfer RNA mimicry of the 3'-end of the TYMV genome, an interest that has led to separate studies on the RNA binding properties of the translation elongation factor eEF1A (formerly EF-1a).
 

Studies on TYMV replication and the role of tRNA mimicry

TYMV has a single mRNA (6.3 kb) as its encapsidated genetic material. The 5'-end is capped, while the 3'-end comprises an 83-nt element that folds like a transfer RNA and confers on the viral RNA the biochemical properties typical of tRNAs: i.e., specific aminoacylation (with valine) and tight binding of the valylated RNA to eEF1A·GTP. We have shown that these properties are siminlarly efficient for TYMV RNA and tRNA^Val. Among the proteins encoded by the viral RNA is an RNA-dependent RNA polymerase (RdRp), whose activity is represented in extracts solubilized from membranes derived from TYMV-infected Chinese cabbage plants. This partially purified TYMV RdRp activity can be used to study how this viral enzyme selects appropriate viral templates from within the many RNAs in a cell, and how the initiation site on a template is chosen to accomplish appropriate end-to-end copying that leads to replication. 

The TYMV RdRp extract is able to transcribe the 3'-tRNA-like structures (TLS) as well as the full-length genome RNA to produce double-stranded products. A search within the TLS for features necessary for this transcription failed to identify elements that could be considered as conventional promoter features. Providing the CCA initiation box at the  3'-end is not base-paired, upstream sequences appear to be of only minor importance in supporting transcription. Given such unconventional control over transcriptional specificity (specific upstream recognition elements have been described in other systems), we are very interested in the mechanisms TYMV RdRp uses to ensure specific replication of the viral RNA. We have presented evidence, that initiation boxes related in sequence to CCA function efficiently only when they remain non-base-paired and accessible within the secondary and tertiary folding of the RNA. We are attempting to understand the definition of "accessibility" in this model.

• Our studies on the role of tRNA mimicry have established the following:

• mutation of the valine anticodon simultaneously results in loss of aminoacylation and replication/infectivity;
• replication/infectivity is regained if the anticodon is changed to permit aminoacylation of the viral RNA with methionine;
• these results indicate a requirement for aminoacylation but no specific requirement for valine, consistent with a required interaction in vivo with eEF1A. Our working model states that this interaction with eEF1A acts as a negative repressor of minus-strand synthesis, regulating the progression of the infection cycle and ratio of minus:plus sense genomic RNAs (see Dreher, Annu. Rev. Plant Pathol., 1999, 37:151-174). We have observed that eEF1A binding inhibits transcription by TYMV RdRp in vitro;

it has been possible to construct three chimeric TYMV genomes with heterologous 3' ends replacing the normal TLS but with no other changes. These genomes are interesting because they support only a background level of aminoacylation, yet are quite infectious. In two of these cases, the valine anticodon remains important for infectivity, although it does not control significant valylation. Our hypothesis is that for such non-aminoacylatable genomes valyl-tRNA synthetase or some other proteins substitute for eEF1A as the negative regulator.

Studies on transcription by Qbeta replicase

Qbeta replicase is unique among the replication enzymes of the positive strand RNA viruses in being capable of performing the entire replication cycle in vitro, resulting in the amplification of large amounts of RNA. The replicase is produced in E. coli cells infected with Qbeta bacteriophage. We have observed that, like TYMV RdRp, Qbeta replicase is able to respond to CCA initiation boxes with transcription of RNAs lacking other known specificity elements (Yoshinari et al., RNA, 2000, 6:698-707). Indeed, such essential and specific promoter elements have been hard to define in Qbeta RNA.
We find that Qbeta replicase has a strong tendency to begin transcription from a CCCA initiation box at the 3'-end of an RNA. This is a strong cis-specificity factor ensuring productive end-to-end copying. We are analyzing the properties of short templates that optimize end-to-end copying without the synthesis of unproductive truncated products arising from initiation of a small number of nucleotides from the 3'-end.

Further collaboration with: InVitro Diagnostics, Inc., NY.

Studies on RNA binding by eEF1A (EF-1a)

Although bacterial EF-Tu (renamed EF1A) has been extensively studied, the biochemistry of its eukaryotic homologue now being called eEF1A has been rather neglected. This is unfortunate, since this protein is one of the most abundant in eukaryotic cells, and it has been the subject of some intriguing but unexplained observations: overexpression leads to cellular transformation, and eEF1A interacts with the cytoskeleton. We have shown that the RNA binding activities of eEF1A are very reminiscent of EF-Tu. Aminoacylated tRNAs are bound with low nanomolar Kd's, while uncharged tRNA is bound with Kd = 15,000 nM. tRNA half molecules comprising only the acceptor/T arm are bound by eEF1A tightly, indicating that contact with the protein occurs with this part of the tRNA, as is the case for EF-Tu. eEF1A discriminates against plant and yeast initiator tRNA^Metby recognizing a phosphoribosyl modification of nucleotide 64 in the T-stem as an antideterminant. We are identifying the anti-determinants present in mammalian initiator tRNA and selenocystein-specific tRNA that prevent tight interaction with eEF1A.

Surprisingly, valylated TYMV RNA is bound by eEF1A with almost the same affinity as valylated tRNA, despite the presence of a pseudoknot in the acceptor stem. We are attempting to crystallize the co-complex with Dr. Jens Nyborg, Arhus, Denmark.