Papilio joanae mtDNA - a reply

[Felix Sperling]

Stan,
        Please see my responses interspersed below.
 
>Felix, I find this discussion of Papilio polyxenes and P. joanae very
>interesting.  I would like to give some of my views on the subject, and
>hopefully get some answers to my questions (as well as corrections to
>any misconceptions I may have).  I do not know anything about these
>butterflies except from what I have read in LEPS-L.
>
>First, I want to go over my understanding of mDNA in these kinds of
>studies, to make sure I am on the right base.  Please bare with me. To
>start, although mDNA  is part of the genome, it really operates at
>another level compared to nuclear DNA.  It does not undergo
>recombination, nor is a mitochondria formed by two gametes that were the
>product of recombination.  Consequently, its role in adaptive selection
>and speciation is highly limited.  Further, because the mitochondria's
>primary function is being the provider of cellular energy through
>respiration and oxidation, it has a limited range of response to
>selective pressures, i.e., the response would be to either increase or
>decrease providing the cell energy.  This type of response is pretty far
>removed from selection pressures related to most speciation events.
>Perhaps during past ice ages such selective pressures (to either
>increase or decrease the mitochondria's activity) would have been more
>important to the process of speciation.
 
Mitochondrial (mt) DNA is inherited in the cytoplasm of the egg, as
part of the mitochondria. Mitochondria are remnants of a 2+ billion
year old symbiosis between bacteria, with part of the original DNA of
the bacterium that became the mitochondria still being inherited
separately from nuclear DNA.
 
One interesting aspect of the maternal inheritance of mtDNA is that
it has a much smaller effective population size (Ne) than nuclear
genes. Only one version (haplotype) of mtDNA is ever inherited by the
progeny of a pairing, and that is the mtDNA of the mother (there are
exceptions, as in so much of nature, but none are known from
butterflies). On the other hand, there are 4 different versions of
most nuclear genes that could be passed to offspring, 2 from the
father and 2 from the mother. Thus mtDNA has only one fourth (25%) as
large an effective population size as most nuclear genes, and it is
more susceptible to genetic bottlenecks and other random sampling
effects. Thus mutations can more easily be fixed in a small
population (which many peripheral populations are) and mtDNA can be
expected to show differences between populations more easily than for
nuclear genes.
 
Mutations in mtDNA can be selectively neutral or not (and very weakly
selected mutations are not really distinguishable from neutral).
*Very* few observed changes in mtDNA are likely to have much
selective consequent; the vast majority is probably very neutral.
Presumably there are also plenty of mutations that cause a metabolic
change, but these are lethal and are immediately eliminated. The best
evidence for this is the fact that virtually all the differences
between closely related haplotypes are in the 3rd position of the
triplet codon for amino acids, where mutations are largely redundant
and end up coding for the same amino acid regardless of a change.
 
So if most sequence evolution is neutral, why does it occur in the
first place? Common explanations include the inexorable, incremental
changes due to sampling error (see above - expected to be faster in
mtDNA due to lower Ne). An analogous example is western surnames.
Other possibilities are genetic hitch-hiking, which means that
selection on one mutation will drag along many neutral mutations that
happen to be located near the selected mutation.
 
 
 
>As Patrick Foley pointed out, the so called 'evolution' of mDNA is
>primarily mutation, and I would guess that quite a large number of
>mutations are possible that do not affect the functioning of the
>mitochondria (neutral mutations), or mutations that affect the
>functioning very little.  mDNA genomic differences between
>non-interbreeding populations of the same species across large distances
>would be expected  to differ due to random mutational events in
>conjunction with genetic drift, and perhaps as the result of the founder
>effect, i.e., a few or one female starts a new population.  These are
>sampling effects.  When populations hybridize, perhaps determining the
>relative frequencies of the different mDNA genotypes may give some
>information as to the extent of  hybridization and introgression.
>Knowing the mutation rate can be used to calibrate the evolutionary time
>scale of differences between populations.  Other analyses could lead
>ultimately to the identification of the more 'primitive', or oldest,
>genotype.
 
Knowing the precise mutation rate is a very hard thing to do.
However, divergence rates in mtDNA are very roughly clock-like (at
least within many taxonomic groups), and so percent sequence
divergence between two lineages is useful for estimating how long ago
the lineages diverged. Interestingly enough, now that we
(specifically Andrew Mitchell, my former postdoc) have sequenced
several more copies of part of the mtDNA of joanae, brevicauda, and
some machaon, it is striking how very little divergence there is.
This suggests that these populations were very recently (less than a
few 10's of thousands of years) connected in some way.
 
However, calling a particular haplotype "ancestral" or the "oldest"
is not really valid. It may have some features that are primitive,
but the real ancestor is long dead, and it's descendants have kept
evolving on different branches. Nothing stops dead in it's tracks;
some morphological features may not have changed from the ancestor,
but others have. Modern ginkos and coelocanths may look like the
fossils from hundreds of millions of years ago, but I guarantee that
their DNA has kept on evolving to some extent.
 
 
>
>While this process is going on in the background, recombination and
>segregation of nuclear genes are being subjected to adaptive selection
>pressures, and other pressures that can result in speciation events.
>Because the so called 'evolution' of mDNA is pretty much independent of
>these processes, an analysis of mDNA can help establish if two species
>are related, and to help determine when speciation events could have
>occurred, either on an evolutionary time scale, or relative to other
>speciations.
>
>Now to P. joanae.  Based on mDNA lineages, you said "This says that one
>part of the total genome of P. joanae is a lot more like that of P.
>machaon than P. polyxenes."  Based on my very limited understanding of
>mDNA analysis, this argues against  P. joanae  being "Ö the most
>primitive ancestor from which polyexenes, machaon, brevicauda, bairdii
>all arose."
 
I did say the first part about the total genome of P. joanae, but
definitely did not say the second part. You have mixed up quotes
somehow with someone else. There is no way that I would say that
joanae is the most primitive ancestor from which the others arose.
See my comments above about evolution not standing still.
 
>
>However, you also say "Ö another part of the genome of  P. joanae, which
>is the genes for the black color pattern, is a lot more like P.
>polyxenes than most populations of P. machaon."  If these genes came
>from P. polyxenes due to introgression, than a difficulty would be
>explaining away the differences in their mDNA.  Much has been said here
>about Haldane's rule on this topic, but Haldane's rule, from my
>understanding, is not a certainty in all hybridizations.  This raises a
>question - are all hybrid females sterile when P. joanae females are
>crossed with P. polyxenes males, as well as the reverse mating, i.e.,
>males with females?  If they are then it would be hard to explain how
>polyxenes black color pattern genes introgressed into P. joanae.  The
>only possible way it could happen is if F1 hybrid males mated with wild
>type P. joanae females which seems a little more improbable, but maybe
>not.
 
Yes, Haldane's rule is not absolute by any means. But it is a useful
generalization. I would love to see Heitzman's data on his hybrid
crosses with joanae published. As for introgression, that can be
asymmetric due to a variety of factors. One would be larger
population sizes in polyxenes, which could help drive polyxenes genes
into joanae. Another may be that the black wing pattern is
selectively advantageous in some way (mating advantage or perhaps
better ability to warm up in marginal sunshine).
 
>
>If Haldane's rule does not apply 100% to either sex combination mating,
>then introgression of black color polyxenes genes could have
>introgressed into P. joanae populations from a P. joanae female x  P.
>polyxenes  male mating.  I feel the mDNA lineages so clearly indicate
>that either species did not arise from the other (based on my,
>admittedly, very limited knowledge of the subject), that this gives rise
>to the next question, which is:  If Haldane's rule does not  apply 100%
>in hybridizations between these two species, then would not one expect
>to also find introgressions of certain P. joanae genes in P. polyxenes
>populations through P. joanae male x P. polyxenes female mating?  Are
>there any recognizable joanae genes, or gene complexes, in P. polyxenes
>populations that could have been introgressed through hybridization?
 
 
Except for one nuance (mtDNA does *not* indicate a clear difference
between joanae and machaon), the scenerios you allude to sound
plausible. Situations analogous to this may be happening between P.
machaon and P. hospiton in the Mediterranean (see work by Aubert et
al. 1997 in Biol. J. Linn. Soc)
 
>
>
>Finally, since butterflies have wing patterns determined by essentially
>the same genetic mechanisms, and especially since P. joanae and P.
>polyxenes are so closely related, could not the similarity between the
>black color genes of P. joanae to P. polyxenes be the result of the same
>genetic solution to similar selective pressures (keep in mind that I
>know nothing about the genetics of these butterflies and know only what
>I have read in LEPS-L)?
 
Convergence in the wing/body pattern genes? Possibly. But I doubt it.
There are just so many similarities in detail that I think it is less
likely than introgression. There is some reason to believe that the
black morph is a linked set of genes. For example, this linked set
appears to be coming apart in some P. polyxenes coloro that have
intermediate wing and body character combinations. The more complex
an overall character is, the less likely it is to arise
independently, all else being equal.
 
Felix Sperling
 
>
>Stan
>
 
 
 ------------------------------------------------------------
 
   For subscription and related information about LEPS-L visit:
 
   http://www.peabody.yale.edu/other/lepsl