Gene regulation occurs at all stages of the information transfer process: during replication, transcription and translation, as well as the steps that occur between these fundamental processes. Our laboratory is interested in gene regulation occurring post-transcriptionally. Much of our recent work has concerned a process called programmed translational frameshifting, in which the process of protein synthesis is modified to allow expression of multiple alternative protein products from a single gene. More recently, we have become interested in how the protein synthetic machinery distinguishes correct from incorrect start sites for protein synthesis.

Errors in reading frame occur extremely rarely during translation, yet some genes have evolved sequences that efficiently induce frameshifting. These sequences, termed programmed frameshift sites, manipulate the translational apparatus to promote non-canonical decoding, and therefore provide tools to probe the mechanism by which the translational apparatus maintains frame during elongation. We study the mechanism of frameshifting in a lower eukaryote, the yeast Saccharomyces cerevisiae. Frameshifting occurs when the ribosome, the RNA•protein machine responsible for translating nucleic acid sequences into protein, changes the frame it uses in reading the 3 nucleotide mRNA sequences called codons that specify which amino acid should be inserted. Such a shift causes the ribosome to read the same RNA sequence but to produce a totally different protein product.

Viruses, transposable genetic elements and a few known cellular genes use this mechanism to encode alternative forms of proteins. We have studied a family of ttransposable elements in the yeast Saccharomyces cerevisiae, called Ty elements, that use programmed frameshifting. We find that Ty frameshifting occurs as part of a dual-error mechanism in which the ribosome first recruits the wrong tRNA to read an in-frame codon, and this errant tRNA then induces the ribosome to recognize the next codon in the shifted reading frame. We are actively engaged in determining the mechanism of this dual-error, and in finding the molecular factors which are responsible for its efficiency.

Because of our work on these programmed errors, we have become interested in the question of how errors are regulated during protein synthesis in general. This interest has led to experiments in two areas. First, we are studying the role of the ribosomal proteins rpS4, rpS5 and rpS12 in regulating error. Work by Gorini and colleagues in the 1960s and 1970s identified these three proteins as playing a significant role in error control. Mutations affecting rpS12 cause resistance to the error-inducing antibiotic streptomycin by making the ribosome hyperaccurate. Mutations of rpS4 and rpS5, identifed as suppressors of the rpS12 mutations, have the opposite effect of causing the ribosome to become error-prone. We are testing models to explain how these proteins modulate error rates in vivo.

The second area involves determine the rates of codon-specific errors. We have developed a system to measure the rate of all possible errors by a particular tRNA in vivo. We have published a paper describing the results of that work with firefly luciferase expressed in the bacterium Escherichia coli. Our data show that codon-specific error frequency varies drastically with the major determinant being the availability of the competing correct (cognate) tRNA. We are extending this work by measuring errors of tRNAs in other contexts using both luciferase and the E. coli lacZ gene.