Complete control

One of the things that we are learning in the doing of Science is that, fundamentally, all things are interconnected . Nowhere is this more apparent than in the familiar Central Dogma, the formation of protein from DNA through the intermediation of RNA.

For years the textbooks have viewed this process in three, discrete steps:

# Transcription (DNA -> pre-mRNA) ref¹
# Processing (pre-mRNA is capped, spliced, tailed and exported from the nucleus)
# Translation (RNA -> protein)

Although we have long believed that these events happen almost simultaneously to any given ‘message’ in bacteria, the confounding presence of the nucleus has led us to believe that the same steps in real cells are spatially and temporally distinct. It turns out that this view is untenable.

Not only do bacteria have a nucleoid, the structure of which can influence gene expression in much the same way as eukaryotic chromatin, but the process of transcription-processing-translation in eukaryotes is much more coordinated than was once thought. Capping and splicing seem to be simultaneous with transcription, and export from the nucleus is similarly coupled to transcription and splicing.

Naturally, you only want your ribosomes to see to capped, spliced and polyA-tailed mesenger RNA, which then must be prevented from returning to the nucleus (see Ratcheting mRNA out of the Nucleus by Murray Stewart, and references therein). But how to do this?

The first thing you can do is let everything out of the nucleus, look for stuff that hasn’t been correctly spliced &c. and destroy it. A more efficient method, and the one that the cell seems to favour, is to stop incorrectly (or incompletely) processed RNA from getting exported in the first place:

There is a nuclear pore-associated protein called Mlp1 that retains intron-containing RNA, i.e. unprocessed mRNA, at the nuclear pore. This binds to something called Nab2, that in turns binds RNA itself and the mRNA export factor Gfd1 .

Nab2 is potentially a marker for ‘mature’ (processed and export-ready) mRNA. It has a compact N-terminal domain (i.e., at the start of the protein sequence) that despite looking like a well-characterized RNA-binding domain actually is necessary and sufficient for binding to Mlp1.

By using my NMR structure (left) as a starting point for molecular replacement, Murray was able to solve the phase problem for the 1.8Å dataset obtained from crystals of it. Furthermore, a single mutation in the middle of the domain, that did not negate its binding to Gfd1, completely knackers binding to Mlp1 (much thanks to the yeast people in Atlanta).

So know we have another little piece of the RNA export puzzle. You can read all about it (and what the reviewer said ).

R GRANT, N MARSHALL, J YANG, M FASKEN, S KELLY, M HARREMAN, D NEUHAUS, A CORBETT, M STEWART (2008). Structure of the N-Terminal Mlp1-Binding Domain of the Saccharomyces cerevisiae mRNA-Binding Protein, Nab2 Journal of Molecular Biology, 376 (4), 1048-1059 DOI: 10.1016/j.jmb.2007.11.087

¹ -> = ‘goes to’. I am not dereferencing an array in Perl. Please don’t hurt me.