Def. of evol: non-allelic changes are necessary

arlin at arlin at
Wed Oct 30 14:59:52 EST 1991

In a previous posting in response to the comments of Stewart Schultz
regarding the definition of evolution, I argued that isolation is a
necessary component of speciation that is distinct from allele
replacement (see "Re: Def of evol: isolation is necessary").  The
arguments below are an attempt to show that any definition of
evolution that strives to be complete must also (in addition to
including mechanisms of isolation) refer to non-allelic changes in
hereditary factors.  I can talk more specifically about the
non-allelic nature of molecular drive at a later date, should this be

In my original posting I proposed a sampling of evolutionary patterns
for consideration, briefly as follows:

Based on comparisons of archaebacteria, eubacteria, and eukaryotes
(especially archezoan eukaryotes), the most recent common ancestor of
all extant cellular life appears to have been a prokaryotic organism
with a single circular DNA genome, probably few introns, little
repetitive DNA, and on the order of a thousand genes (the exact
details don't matter as long as everyone agrees that this scenario is
in the realm of the possible).   Now we have a vast array of 100's of
millions of species, some of which still have small circular genomes
with on the order of a thousand genes, others of which have dozens of
huge linear chromosomes, many introns, and many thousands of genes.
In the descent of these organisms from their common ancestor, we can
distinguish splitting of one population into many populations, changes
in the sequences of genes, increases in the size and information
content of genomes, acquistion of endosymbionts, propagation of
transposable elements and other repetitive sequences, changes in
chromosome number, changes in the form (linear or circular) of
chromosomes, and so on, all of which can rightly be called
evolutionary changes.

Are all of these changes "allele replacements" in some fundamental
sense?   Although this contention may be correct, *it must be tested
by a logical inquiry*: we are not obliged by some _a priori_ necessity
to define all evolutionary change as "allele replacements."   "Allele
replacement" was already defined before molecular biologists and
introductory textbook authors started defining evolution as a
succession of allele replacements.  Allele replacement refers to a
mechanistic model of genetic change in a population, in which one
nucleic acid sequence shared by the members of a population is
replaced by a different nucleic acid sequence at the same chromosomal
location; once the new allele is created by mutation, changes in its
frequency in the population take place primarily as a result of the
cumulative effects of differences in the survival and reproduction of
the organisms that carry it; these differences in survival and
reproduction are either deterministic (for allele replacements
motivated by selection) or non-determinisic (for allele replacements
motivated by random drift).   The very precision of the allele
replacement paradigm is what has made it so successful as the basis
for explanatory and predictive models.  When we set out to evaluate
the completeness the allele-replacements definition of evolution, let
us not throw out the precision and turn this elegant paradigm into a
fuzzy concept, so that it becomes a moving target.

I hope my arguments in previous postings have made clear that the
isolation required to split a population is a necessary part of
evolution but cannot be equated with or subsumed by allele
replacement, regardless of how "allele replacement" is defined (see my
posting "Re: def of evol: isolation is necessary").   This is not a
novel finding  in evolutionary genetics.

Evolution is yet more than allele replacements and isolation.  In the
course of eukaryotic evolution, endosymbionts have been gained.
Consider the membrane of the ancestral mitochondrion: this membrane is
not an allele, since alleles are nucleic acids.   However, the
membrane of the mitochondrion is a hereditary factor for an essential
trait.  No organism can build a mitochondrion without first having the
membrane: the membrane itself provides a template for the growth of
more membrane by the addition of lipids, and it contains signals
(either in lipid composition or integral proteins) necessary for
transport of building materials to the mitochondrion.  The acquisition
of this membrane and the cell it bounded was an important step in
eukaryotic evolution, yet it was not an allele replacement, because a
membrane is not an allele and the cytoplasm where it exists is not a
chromosomal locus.

Similarly, many eukaryotic organisms have centrioles.  Centrioles
cannot be built from centriolar proteins and other raw materials alone
because centrioles must be templated on other centrioles  (just as
genes cannot be made solely with enzymes and raw materials: they must
be templated on other genes).  Centrioles, membranes, the cortex of
ciliates:  these are non-genetic hereditary factors that have played a
role in evolution.

In the preceding paragraphs it was argued that non-genetic evolutionary
changes (which are by definition not allelic) have occurred.  Of
course, we could expand our definition of alleles and include all
hereditary determinants, not just nucleic acids.  Personally, I like
this idea, but I think it would cause confusion and make the
"geneticists" (especially the "molecular geneticists") upset, so let's
stick with "hereditary determinants" or "hereditary factors" as the
general term, and use "genes" or "genetic material" to refer
specifically to hereditary factors that are nucleic acids.

Even if the geneticists would let us get away with a broader
definition of "gene," we would not be out of the woods: some
evolutionary changes in hereditary factors are not allelic changes.
For example, not all organisms have the same set of genetic loci: new
loci have been created in the course of evolution.  If we were forced
to take the position that all genetic changes are allele replacements,
we would have to describe the creation of a new locus as the
replacement of one allele with another.  Thus, we would first have to
define each point on the chromosome that could at some unspecified
time in the future become the home of a gene as a
locus-waiting-to-happen.  These loci-waiting-to-happen could become
loci that themselves contain additional loci-waiting-to-happen.  The
cytoplasm that could eventually harbor an endosymbiont and its genes
would have to be a chromosome-waiting-to-happen.  By treating non-loci
as loci-waiting-to-happen, we would be able to say that each newly
formed locus was not newly formed, but that all organisms have always
had the same loci, with each individual organism having either null
information-less alleles or functional /informative alleles at a
"locus-waiting-to-happen" or "locus".

Similarly, we could try to treat the splitting of one chromosome ABCDE
into two chromosomes, AB and CDE, as an allele replacement,  the two
variant alleles being
presense-of-the-phosphodiester-bond-connecting-B-and-C and
absence-of-the-bond-connecting-B-and-C.   The same treatment would
work equally well for turning a circular chromosome  (with A, B, C, D,
and E in that order) into the linear chromosome CDEAB: the "alleles"
are B-C-bond and no-B-C-bond.   What the "locus" would be in these
cases is entirely unclear.

The above examples suffice to show that the conceptual framework of
allele replacements is inadequate to the task of accounting for all
evolutionary change.  Is this conclusion revolutionary?   Is it
something to feel insecure about?  Does it lay waste to decades of
population genetic theory?  No, I don't think so.  It is much better,
in my opinion, to let go of the stilted modern nucleicacidocentric
notion that everything must be an allele replacement; it is better not
to sacrifice the precision of the allele-replacements paradigm, and
instead to freely admit (into our formal repertoire of mechanisms for
evolutionary change) separate categories for isolation, for changes in
non-genetic hereditary material, and for non-allelic changes in
hereditary materials such as chromosome-splitting and the creation of
new hereditary loci.

Arlin Stoltzfus
Department of Biochemistry
Dalhousie University

Arlin at
usual disclaimer

More information about the Mol-evol mailing list