While in general we have few quibbles over the substance of R. Lewin's clear description of the genetic system of molecular
drive (Science 1982 218 552-3), we feel that several comments in the report merit further discussion. Lewin's report basically
asks two questions: how real and how important; and some commentators have offered answers.
In our two papers (1, 2) in which we have detailed the factual basis and theoretical implications of molecular drive, we have
defined it as a process of fixing a mutation within multigene and non-genic families in a population, as a consequence of
DNA turnover. Considerations of rates of turnover indicate that individuals of a sexual population would change in unison
with respect to the changing composition of a family. At the heart of molecular drive is the widespread phenomenon of concerted
evolution. Although the reality of this phenomenon is incontestable, we cannot accept the definitive statement of Lewin's,
drawing in particular on remarks made by A. Jeffreys on the human globin cluster and Alu family, "that it is not a universal
phenomenon". Concerted evolution is occurring in the globin cluster; indeed this phenomenon was first defined as such
in this cluster due to the homogenization of pairs of [alpha] and [gamma] genes, and their flanking sequences, by unequal
exchange or gene conversion. In reviewing such events in the globin cluster Jeffreys has written, "clearly, concerted
evolution is not a rare phenomenon, and seems to occur between even distantly related genes and between active genes and pseudogenes"
(3). In the case of very large families, such as Alu, detailed consideration needs to be given to the rates of homogenization
relative to the mutation rate. A 10% level of sequence variation between 10 cloned Alu repeats from the human genome (4)
reflects the constraints on homogenization imposed by the presence of 500,000 copies finely dispersed over 46 chromosomes.
Despite these constraints the very low levels of homology revealed by hybridization between human and mouse Alu families reflects
a much greater between-species than within-species divergence. Furthermore the human Alu family has been homogenized throughout
by an imperfect dimer whilst the mouse Alu family consists only of monomers (4). Turnover is occurring in the Alu family,
albeit slowly. We are not aware of families, whether tandem or interspersed, genic or non-genic that are immune from such
processes. The evolutionary progress of each family under molecular drive and the subsequent interaction with natural selection
is expected to be very different (1,2).
The importance of molecular drive as a genetic system can only be assessed by consideration of the way in which the genetic
and phenotypic cohesion of a population is maintained. An instructive example is provided by the phenomenon of hybrid dysgenesis
in Drosophila. In this example the molecular process is one of transposition; one of the three mechanisms underlying molecular
drive. A slow rate of transposition of P elements would lead to a genetic situation in which there would be little variation
in the number of P's in each individual at any one time during the initial accumulation of the element. The small variance
in P number would not lead to dysgenesis within the population, as is observed. A large difference, however, in P number between
a P population and a non-P population does lead to dysgenesis. Precisely the same low variance pattern of fixation would result
from the slow rates of unequal exchange or gene conversion involving the homogenization of existing families for one variant
Given this cohesive system of genetics, which contrasts remarkably with the classical population genetics of single-copy genes,
we allowed ourselves some freedom in speculating on its involvement in the origin of the ontogenetic and reproductive differences
between species. So far as we are aware, there are few experimental tests of the genetic mechanisms which are thought to underlie
species differences. We do not disagree with the conventional viewpoint that such differences might be consequential upon
natural selection and genetic drift working within mendelian populations. Nevertheless such external processes of fixation
are inadequate in explaining species differences in multiple-copy families, i.e. the phenomenon of concerted evolution. The
evolution of such families and their manifold phenotypic effects can be partly explained by the genetics of molecular drive,
which is precisely based on internal molecular mechanisms of turnover. Consequently, we are perplexed that Drs. Doolittle
and Selander consider our speculation on the evolutionary biology of molecular drive to be unhelpful. We consider that all
evolutionary biology may be, in essence, a manifestation of molecular events, and the artificial separation of molecular
and evolutionary biology is itself unhelpful.
Part of the problem seems to stem from a mistaken supposition that turnover is only observed in non-genic families whose biological
effects have yet to be ascertained. It would be a pity if this misunderstanding was widespread. Concerted evolution is an
extensively documented observation in many multigene families. The biological effects and evolutionary significance of changes
in these families cannot be seriously challenged. It could well be that even the species differences in behavior emphasized
by John Maynard-Smith, are under multigene control. A population could undergo a long-term collective transformation in behavior
under the aegis of the genetic system of molecular drive.
We do not consider molecular drive to be a catch-all for all genomic rearrangements and exchanges. If some rearrangements,
for example inversions, deletions, or duplications turn out to be one-off events, then they are analogous to most point mutations
which rely for their evolutionary progress on selection and drift. They do not contribute to the process of molecular drive.
From what we now understand of the activities of unequal exchange, gene
conversion and transposition in so many different families, the evolutionary differences between species must be considered
a complex outcome of three processes of fixation - adaptive, accidental and cohesively driven. Despite the seeming pitfalls
in trying to promote a new perspective, we see no reason to be unenthusiastic about the implications of molecular drive.
G. A. Dover, T. Strachan, E. S. Coen, [address as above] S. D. M. Brown* *Dept. Biochemistry, St. Mary's Hospital, London.
1. Dover, G. A., Brown, S. D. M., Coen, E. S., Dallas, J., Strachan, T. and Trick, M. (1982) in 'Genome Evolution'
(eds. G. A. Dover and R. B. Flavell) Acad. Press. p. 343.
2. Dover, G. A. (1982). Nature 299 111.
3. Jeffreys, A. (1982) in "Genome Evolution" (eds. G. A. Dover and R. B. Flavell) Acad. Press. p. 157.
4. Jelinek, W. R. & Schmid, C. W. (1982). Ann. Rev. Biochem. 51 813.