CORRESPONDENCE

Origin of the Genetic Code

Sir,—I should like to make two points about the account
of my talk to the British Biophysical Society which you
published recently?. I do not in fact believe that the idea
of a stereochemical relationship between all amino-acids
and their anticodons, as suggested by Dunnill*, is likely.
There is, in my opinion, suggestive evidence against it, but
at the moment it is not enough to disprove the idea.

My second point concerns the main-substance of my
talk, which was an attempt to show that a plausible
theory could be constructed without necessarily assuming
any stereochemical interaction of amino-acids with either
codons or anti-codons. I imagined the code to go through
three phases:

1. The Primitive Code, in which a small number of
amino-acids were coded by a small number of triplets.

2. The Intermediate Code, in which these primitive
amino-acids took over most of the triplets of the code in
order to reduce nonsense triplets to a minimum?. The
codons produced by this process for any one amino-acid
were likely to have been related.

Woese‘ has pointed out that this state of affairs could
also have been produced by reading only a single base ofa
triplet, or by considerable inaccuracy in the reading of a
few triplets.

3. The Final Code, as we have it today.

The crucial idea, already mentioned by Jukes', concerns
the transition from 2 to 3, which I certainly do not think
was “unlikely to have taken place’’?. Evolutionary theory
suggests that a new amino-acid was incorporated into the
developing code only if its introduction at that time gavea
selective advantage to the primitive organism. This
implies that its introduction did not disturb too much the
proteins then being produced, and in addition made a
significant improvement to at least one of them. This
would have happened most easily if

(a) the new amino-acid was “related” to the one pre-
viously coded by the triplet(s) in question*;

(6) the organism coded rather few proteins;

(c) these proteins were rather primitive in their con-
struction.

Eventually as the number of proteins coded became
larger, and their design more sophisticated, no possible
new amino-acid could, on balance, be an advantage and
the code would be frozen.

Such a theory could thus explain in a general way the
non-random nature of the present code, since “related”
amino-acids might well have acquired related codons. It is
quite distinct from theories*? which postulate that the
code evolved as it did in order to minimize the damaging
effects of present-day mutations on individual proteins.

F. H. C. Crick

Medical Research Council,
Laboratory of Molecular Biology,
Hills Road, Cambridge.

1 Nature, 212, 1897 (1966).

*Dunnill, P., Natura, 210, 1267 (1966),

* Sonneborn, T. M.,in £rolving Genes and Proteins (edit. by Bryson, Vernon,
and Vogel, Henry J.) (Academic Preas, New York and London, 1965).

‘ Woese, C. R., Proc. U.S. Nat. dead. Set., 54, 1546 (1965).

* Jukes, T. H., Molecules and Evolution, 70 (Colambia University Press,
1966).

* Epstein, C. J., Nasure,210, 25 (2066),
* Goldberg, A. L., and Wittes, B. E., Science, 158, 420 (1966).