cok at mole.bio.cam.ac.uk wrote:
> A phenomemon that I can't seem to rationalise is the (relative)
> confinement of mobile self-splicing introns to organelles. I appreciate
> that:
By "relative" confinement perhaps you intend to imply that these introns
are not absolutely restricted to organelles. Group I introns are
sometimes present in nuclear ribosomal RNA genes in protists, most
famously in Tetrahymena (the Tom Cech intron), but they have never
been seen in nuclear protein-coding genes. They are also found
in eubacterial tRNAs and in some bacteriophage genes. Group II introns
are found very rarely in eubacterial genomes and have never been
seen in nuclear genomes (but fragments of them have been detected).
This doesn't make the question any less interesting. In fact it
makes it more interesting because it shows that the dearth of
self-splicing introns in non-organellar genomes is not due to
lack of exposure to these introns.
> - they may have come into eukaryotes with organelles; BUT this hasn't
> stopped loads of other organelle genes moving to the nucleus.
And they may have brought their group II introns with them. Because
the splice donor and acceptor sites are very similar, it has been
suspected that these introns would rapidly degrade _in situ_ into
spliceosomal introns by loss of self-splicing ability. I
believe it was John Rogers who suggested this in 1989. So we wouldn't
necessarily expect group II introns in nuclear genes to be maintained
as such for long periods of time.
> - they may have evolved into spliceosomal introns in the nucleus, as their
> host organism would develop an interest in making damned sure that it
> retained the machinery to splice them effectively out of coding regions.
> BUT I don't see any reason why mobile self-splicing introns couldn't be
> maintained in the longer term in nuclear genomes just as well as they have
> been in many mitochondrial genomes.
>> - long-term maintenance of mobile elements in a genome usually requires
> sex, so that elements can spread to unpopulated genomes. BUT this should
> be even less of a barrier for nuclear mobile introns than for organelle
> ones.
I agree completely. The rarity of group I introns in non-organellar
genomes is not due to lack of exposure, and for most nuclear genomes,
it is not due to the lack of opportunities for selfish spread, which
would be provided by sexual reproduction.
So, we can conclude that the relative barrier is either they cannot
get established when introduced, or once established, they face some
sort of negative selection that is not present in organellar genomes.
My pet idea is that the distribution of self-splicing introns
is largely due to a barrier to establishment due to polymerase
sensitivity to highly structured transcripts, which cause polymerase
pausing, which is often the first step in termination. Eukaryotes
use RNA Pol II for mRNA genes, but RNA Pol I & III for structural
RNA genes. Eubacteria (E. coli, at least) transcribe rRNA genes
from a special promoter that signals termination-resistant
transcription.
Bacteriophage polymerases are an entirely different class of
polymerases.
Organellar RNA polymerases, curiously, are related to phage polymerases.
Perhaps structured introns tend to be limited to genes that are
transcribed by structure-resistant polymerases. Just a thought (but
at least its an experimentally tractable one).
Arlin
--
Arlin Stoltzfus, Ph.D. (arlin at is.dal.ca)
Department of Biochemistry, Dalhousie University
Halifax, Nova Scotia B3H 4H7 CANADA
phone: 902-494-2968 fax: 902-494-1355
