Genetic Code Evolution Paper

I came across this paper which is quite similar to the postings I
made on this subject a few months ago.
http://www.pnas.org/cgi/content/abstract/102/12/4442

The big difference is that they postulate a set of two-nucleotide
'adapter' molecules (much like Crick's hypothesis) that functioned
like tRNAs before ribosomal translation got started.

They even suggest, as I did, that alpha-keto acids were the molecules
charged onto the adapters and that they were reductively aminated there.

One of the authors - Morowitz - has written much that I admire
regarding the priority of membranes in the OOL and the plausibility
of an autotrophic origin. This is the first time, AFAIK, that he
has written anything specifically about the code.

One of the other authors is from the Santa Fe Institute, so I will
blame him for the mistakes and faulty logic that appear in the paper.
;-)

One example of this may be found in the first paragraph, where they
write:
...codons for Glu and Pro both begin with C, and those for Cys and
Leu begin with U. Codons beginning with G encode amino acids that
can be formed by direct reductive amination of a simple alpha-keto acid.
These include glycine, alanine, aspartate, and glutamate, which can
be formed by reductive amination of glyoxalate, pyruvate, OAA, and
alpha-KG, respectively.

Well, I'm pretty sure that codons for Glu don't begin with C. Perhaps
they meant Gln. And ALL amino acids except proline can be (and are)
formed by reductive amination of simple alpha-keto acid. The ones
they list are merely the best known alpha-keto acids.

While their theory is quite similar to mine regarding code evolution,
it is dramatically different regarding the origin of ribosomal
translation. They have mRNAs before ribosomes. I don't. Their
theory means that some coded proteins would have been available
when ribosomes and tRNAs were inventing modern translation. But
it makes it incomprehensible why modern translation would be so
mechanically different from primitive translation. And why it makes
so little use of coded proteins.

However, the thing that I like least about their theory is that
their adapters are small soluble molecules. I don't believe that
such molecules played any part in metabolism until the inversion of
'obcells' as described by Cavalier-Smith. And that could not have
taken place until at least some coded protein synthesis was taking
place. I think that their small adapter molecules would simply
diffuse away from the organism. Or else, they could never be formed
inside the organism.

Also, they suggest that ornithine might once have been a coded
amino acid. Unlikely. It would detach itself from the adapter
by forming a lactam. (I learned this from Stanley Miller, of all
people. Cavalier-Smith makes the same point.).

Overall, though, they get things mostly right, though they for some
reason rotate the code matrix from its most familiar orientation.
They miss the connection in the UNN row whereby both phospho-serine
and PEP are interconvertible, and hence that Ser, Cys, Seleno-Cys,
and Trp are biosynthetic cousins of Phe and Tyr. But that may be
forgiven.

It is nice to have something close to my ideas finally in print.
Now, all that remains is to prove my theory. The ideal proof would
be to find an organism that synthesizes His from an activated Glu
already attached to a CAR-accepting tRNA. It would most likely be
something that uses CO or HCN as a carbon source. But even if such
organisms exist, I am not very hopeful that they will be cultured and
studied anytime soon.



.