living in an rna world

The central dogma of molecular biology — DNA ⇒ RNA ⇒ protein — implies a passive role for RNA. In truth, RNA may be the most dynamic step of the process. Biologists have spent decades elucidating the roles of RNA and the processes by which genes are copied into mRNA and then translated into protein. Now, reams of data are becoming available from high-throughput sequencing projects.

Our research uses bioinformatics and molecular biology to understand how evolution has shaped post-transcriptional gene regulation.

translation

A new method, ribosome profiling, has made it possible to measure translation genome-wide by capturing information about the position of ribosomes in vivo. Fragments of mRNA protected by ribosomes are sequenced and mapped back to the genome to show the positions (or footprints) of millions of individual ribosomes in the original population of cells. This can show how much translation is occurring and on what genes.

As a postdoc with Pat Brown at Stanford, I made an unexpected discovery: ribosome footprints reveal two distinct ribosome states. Along with the known 28-nt footprints, we noticed many small (21 nt) footprints clearly spaced codon by codon and representing translating ribosomes. The ribosome undergoes major structural rearrangements during elongation, rotating open after peptide bond formation then rotating closed to translocate along the mRNA. After finding that we could stabilize one or the other conformation with different translation inhibitors, we concluded that the two footprints, large and small, represent non-rotated and rotated conformations of pre- and post-peptide bond ribosomes (Lareau et al, eLife, 2014).

The counts of small and large footprints should reflect how long the ribosome spends in each conformation at each codon (with additional variation from other causes). Thus, ribosome profiling gives us a simple, fast readout of features that affect translation.

unproductive splicing

The prevalence of alternative splicing was a surprising outcome of the human genome sequencing project — almost all human genes can be processed into distinct mRNA isoforms. What is all of this alternative splicing doing?

In graduate school, my colleagues in the Brenner group at Berkeley made an interesting discovery: a third of the alternative splicing of human genes leads to mRNAs that are probably degraded by a cellular surveillance pathway called nonsense-mediated mRNA decay (NMD). Why is so much RNA produced only to be destroyed? In some cases, the unproductive splicing may be important for gene expression. Unproductive splicing (that is, coupled alternative splicing and NMD) could allow the cell to 'turn down' expression of a gene after the mRNA has already been transcribed from the genome.

SR proteins and poison exons

We showed that a particularly interesting family of splicing factors, the SR proteins, may all be regulated by unproductive splicing (Lareau et al, Nature, 2007). We found that all of the human genes in this family have alternative splice forms that are degraded by NMD. Most of them include an alternative exon that marks the mRNA for decay — we call these 'poison exons' because their only role seems to be to 'kill' the mRNA.

Ultraconserved poison

This study opened up fascinating questions about evolutionary genomics. The alternative regions of the SR genes are extremely conserved between human and mouse, more conserved than the protein-coding regions of the same genes. Such high conservation is usually a sign of functional constraints on the sequence, and we're working to identify the pressures that have kept these regions remarkably static.