Content uploaded by David S. Thaler
Author content
All content in this area was uploaded by David S. Thaler on May 21, 2018
Content may be subject to copyright.
Stoeckle M.Y.
Program for the Human Environment
The Rockefeller University
1230 York AVE
New York, NY 10065
USA
Email: mark.stoeckle@rockefeller.edu
Thaler D.S.
Biozentrum, University of Basel
Klingelbergstrasse 50/70
CH - 4056 Basel
Switzerland
Email: david.thaler@unibas.ch
davidsthaler@gmail.com
DOI: 10.14673/HE2018121037
Why should mitochondria dene species?
Key words: Species evolution,
mitocondrial evolution, speciation,
human evolution.
Vol. 33 - n. 1-2 (1-30) - 2018HUMAN EVOLUTION
More than a decade of DNA barcoding encompassing
about ve million specimens covering 100,000 animal
species supports the generalization that mitochondrial
DNA clusters largely overlap with species as dened by
domain experts. Most barcode clustering reects synony-
mous substitutions. What evolutionary mechanisms ac-
count for synonymous clusters being largely coincident
with species? The answer depends on whether variants
are phenotypically neutral. To the degree that variants are
selectable, purifying selection limits variation within spe-
cies and neighboring species may have distinct adaptive
peaks. Phenotypically neutral variants are only subject
to demographic processes—drift, lineage sorting, genetic
hitchhiking, and bottlenecks. The evolution of modern
humans has been studied from several disciplines with
detail unique among animal species. Mitochondrial bar-
codes provide a commensurable way to compare modern
humans to other animal species. Barcode variation in the
modern human population is quantitatively similar to that
within other animal species. Several convergent lines of
evidence show that mitochondrial diversity in modern
humans follows from sequence uniformity followed by
the accumulation of largely neutral diversity during a
population expansion that began approximately 100,000
years ago. A straightforward hypothesis is that the extant
populations of almost all animal species have arrived at
a similar result consequent to a similar process of expan-
sion from mitochondrial uniformity within the last one to
several hundred thousand years.
Precis
1. Mitochondrial Cytochrome Oxidase Subunit I DNA barcodes (COI barcodes,
often shortened to “DNA barcodes” or “barcodes” in this article) began as an
aid to animal species identication and made no claims of contributing to evo-
lutionary theory. Five million DNA barcodes later the consistent and commen-
surable pattern they present throughout the animal kingdom is one of the most
general in biology. In well-studied groups the majority of DNA barcode clusters
agree with domain experts’ judgment of distinct species.
2STOECKLE, THALER2
2. The tight clustering of barcodes within species and unlled sequence space
among them are key facts of animal life that evolutionary theory must explain.
Many aspects of speciation are complex. Barcodes are unique in being quanti-
ably commensurable across all animal species and almost always yielding the
same single simple answer [1].
3. Either of two evolutionary mechanisms might account for the facts: a) species-
specic selection, or b) demographic processes acting independently of pheno-
type.
4. Most barcode variation consists of synonymous codon changes. Since the as-
sumption of neutrality of mitochondrial synonymous codons was asserted,
many exceptions in nuclear genes and prokaryotic systems have been found.
5. New arguments are presented that synonymous codon changes in mitochon-
drial genes are neutral to a greater extent than nuclear genes.
6. Extensive data on modern humans make our species a valuable model system
for animal evolution as a whole. The mitochondrial variation within the modern
human population is about average when compared to the extant populations
of most animal species.
7. Similar neutral variation of humans and other animals implies that the extant
populations of most animal species have, like modern humans, recently passed
through mitochondrial uniformity.
History of COI barcoding
DNA barcoding was rst proposed as a tool for practical taxonomy and to democ-
ratize actionable biological knowledge [2, 3]. At its origin DNA barcoding made no
claim of contributing to evolutionary theory. Previous work bode well for mitochondrial
genomes being reliably similar within animal species yet in many cases distinct among
neighbor species [4, 5]. The particular mitochondrial sequence that has become the most
widely used, the 648 base pair (bp) segment of the gene encoding mitochondrial cyto-
chrome c oxidase subunit I (COI), reached a tipping point because widely applicable
reliable primers and methods useful for both vertebrates and invertebrates were adopted
by a critical mass of the community [6, 7].
Skeptics of COI barcoding [8] raised a number of objections about its power and/or
generality as a single simple metric applicable to the entire animal kingdom, including:
1) the small fraction of the genome (about 5% of the mitochondrial genome and less than
one millionth of the total organism’s genome) might not be sensitive or representative
[9, 10]; 2) since animal mitochondria are inherited maternally the apparent pattern of
speciation from mitochondria is vulnerable to distortion when females and males roam
differently [11]; 3) the mitochondrial chromosome is subject to types of selection not
3WHY SHOULD MITOCHONDRIA DEFINE SPECIES? 3
experienced by the nuclear genome [12]: replicon competition within each organelle
[13], among organelles inside each cell [14-16], including differential segregation of or-
ganelles at cell division [17]; and 4) mitochondria in some groups are sensitive to agents
such as Wolbachia that are not known to affect nuclear genes [18]. Mitochondrial pseu-
dogenes in the nucleus sometimes confused analysis [19]. Anecdotally, some domain
experts felt that only specialists can reliably recognize species in each group and that
“DNA taxonomy” was felt as necessarily inferior or a threat.
The current eld of COI barcodes is no longer fragile but neither is it complete. As
of late 2016 there were close to ve million COI barcodes between the GenBank and
BOLD databases. Objections can now be seen in the cumulative light of these data and
more than a decade’s experience. There is no longer any doubt that DNA barcodes are
useful and practical (Figs. 1,2). The agreement with specialists encompasses most cases
in several important animal domains. Many cases where DNA barcodes and domain
specialists do not agree reect geographic splits within species or hybridization between
species. Others upon further investigation been attributed to mislabeling or sequence
error [20]. Some may represent bona de exceptions to the rule that mitochondrial se-
quence clusters coincide with species dened by other means. In the great majority of
cases COI barcodes yield a close approximation of what specialists come up with after a
lot of study. Birds are one of the best characterized of all animal groups and COI barcode
clusters have been tabulated as agreeing with expert taxonomy for 94% of species [21].
Exceptions to the rule that each species is a single unique cluster
Most exceptions to the generality that COI clusters represent species are also excep-
tions to the general rule that species are single interbreeding populations. These include
cases with phylogeographic divisions within species and those with shared or overlap-
ping barcode clusters (Figs. 2,3).
In most well-studied cases of shared or overlapping barcodes, nuclear genome anal-
ysis demonstrates these anomalies are due to hybridization resulting in mitochondrial
introgression from one species into the other. If recent, and complete across the whole
population, introgression erases mitochondrial differences between species. Introgres-
sion events in the more distant past and those involving only part of a species produce
more complex patterns, as illustrated by Ursus bears (Fig. 3). Based on nuclear and
mitochondrial genome analysis, polar bears (U. maritimus) hybridized with “ABC is-
land” brown bears (U. arctos) about 50,000 years ago, with introgressive replacement
of ABC arctos mitogenomes by maritimus mitogenomes. The mitochondrial lineages
subsequently diverged, but ABC island brown bear mtDNA remains closer to polar bear
than to mainland brown bears. Nuclear genomic analysis supports taxonomic classica-
4STOECKLE, THALER
Fig. 1. Low intraspecic COI barcode variation is the norm in animals, not an artifact of
handpicking examples or small sample size. Variation is expressed as average pairwise dif-
ference (APD) between individuals.
5WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
Fig. 2. Relatively large interspecic differences, as compared to uniformly small intra-
specic differences, are the norm in animals. Together these yield the familiar clustering
pattern that enables DNA barcode species identication. Shown are neighbor-joining (NJ)
trees (with scale bars for number of individuals and percent K2P distance) and average pair-
wise distance (APD) within and between sets of closely-related congeneric species. At top,
NJ trees with bars marking species clusters. Exceptions to the one species/one cluster rule
include cases with multiple clusters within species, corresponding to geographically isolated
populations [marked as (W)estern and (E)astern], and cases with clusters shared between
species, marked by double vertical lines. At bottom, APDs for the same congeneric sets,
with average (horizontal bar) and range (vertical bar) of intraspecic and interspecic APDs
shown.
6STOECKLE, THALER
Fig. 3. Clustering of 0.6 kbp COI barcode segments accurately represents the complete
12 kbp coding mitogenome. At top, COI and mt genome NJ trees exhibit similar clustering
patterns. At bottom, average pairwise differences within and between species in each set are
about the same whether calculated from COI barcodes or coding mitogenomes. As in Fig. 2
legend, apparent exceptions with phylogeographic divisions (locusts) or shared or overlap-
ping clusters (bears, fruit ies) are noted. NJ tree scale bars for number of individuals and
percent K2P distance are shown.
7WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
tion of ABC island and mainland populations as subspecies of brown bear.
Incomplete lineage sorting with retention of ancestral polymorphisms is a plausible
mechanism for shared or overlapping mitogenomes, also called paraphyly. However, in
all cases we know of, when analyzed for nuclear and mitochondrial differences, ongoing
or historical hybridization is the likely cause (see Fig. 1 and Table S3 in reference [20]).
General across the animal kingdom
DNA barcodes based on mitochondrial sequences might have failed to be sensitive,
general, practical, or to agree with the judgment of experts in each domain. Five million
DNA barcodes later some exceptions have been found, however, the power, generality,
and validity of the COI barcode approach for identifying animal species is no longer in
question, at minimum, for several major groups (Figs. 1-3). A general observation is that
barcode clusters correspond best to species in well-studied animal groups, where taxono-
mists have mostly decided and agreed upon what species are. Thus there is good support
in several major phyla, including Chordata, Arthropoda, Mollusca, Echinodermata. We
note that these phyla are estimated to contain about ¾ of named animal species.
Incompletely studied groups
In the remaining 23 animal phyla, there are examples where clusters match spe-
cies, but the overall picture is muddier. Many are small animals, difcult to distinguish
morphologically, and have attracted relatively little taxonomic or DNA barcode study.
Major incompletely studied groups include Annelida, Nematoda, Platyhelminthes, Porif-
era, and Rotifera. We expect that with further study these phyla will t a pattern similar
to that in more established groups. However, at this stage it takes cherry-picking to nd
examples that match the better-studied phyla and one cannot make a data-based case for
the general validity of DNA barcoding in these phyla.
Beyond using the DNA barcode as an aid to taxonomy, the enormity of data now
available make it appropriate to extend the applications of the “broad but not deep”
vista that COI barcodes uniquely provide [20, 22, 23]. In a founding document of phy-
logeography, Avise and colleagues noted the long-standing divide in biology between
the intellectual lineages of Linnaeus for whom species are discrete entities and those of
Darwin who emphasize incremental change within species leading to new species [4].
They presciently proposed that mitochondrial analysis would provide a way to bridge the
intellectual gap. DNA barcoding now provides the most comprehensive database allow-
ing a kingdom-wide and quantitative realization of that vision.
8STOECKLE, THALER
Differing denitions of species
There are approximately 30 different denitions of species in the biological, philo-
sophical, and taxonomical literatures [24]. Almost all of them share the idea that spe-
cies are distinct entities in biology and the corollary idea that there are discontinuities
among species [25]. In their clarifying and valuable analyses Mayr [25] and de Queiroz
[11] point out that all denitions of species involve separate monophyletic evolution-
ary lineages (with important exceptions where symbiosis or horizontal gene transfer are
key [26]). Different distinguishing factors such as mating incompatibility, ecological
specialization, and morphological distinctiveness evolve, in various cases, in a different
temporal sequence. During the process, as species diverge and emerge some of these
characteristics will be fullled while others are not. Disagreement is inevitable when
different properties are considered necessary and sufcient to t one or another deni-
tion of “species”.
There are two important observations regarding how COI barcodes t into the dif-
fering denitions of species. First, the cluster structure of the animal world found in
COI barcode analysis is independent of any denition(s) of species. Second, domain
experts’ judgments of species tend to agree with barcode clusters and many apparent
deviations turn out to be “exceptions that prove the rule”. Controversy around the edges,
e.g. disagreements about whether or not borderline cases constitute species or subspe-
cies [27, 28] should not obscure visualizing the overall structure of animal biodiversity.
It is unavoidable that some cases will be considered as species by one denition and not
another. Controversial cases can illuminate in the context of William Bateson’s adage to
“treasure your exceptions” [29] but they should not obscure the agreement for most cases
and an appreciation of the overall structure within the animal kingdom. This pattern of
life, close clustering within individual species with spaces around clusters, can be visu-
alized and demonstrated in different ways and with different statistics (e.g., Figs. 1-4).
It qualies as an empirically-determined evolutionary law [30]. Barcode distribution is
arrived at independently but consistent with a view of biology as composed of discrete
entities that on different levels include organisms [31] and species [32].
The pattern of DNA barcode variance is the central fact of animal life that
needs to be explained by evolutionary theory.
In ‘The Structure of Scientic Revolutions’ Thomas Kuhn makes the point that
every scientic model takes certain facts of nature or experimental results as the key
ones it has to explain [33]. We take the clustering structure of COI barcodes—small
variance within species and often but not always sequence gaps among nearest neighbor
9WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
Fig. 4. Species are islands in sequence space. COI barcode NJ tree and Klee diagram of
American Robin (Turdus migratorius) and closely related Turdus species. To generate data-
set, a single American robin COI barcode was used to search GenBank using BLAST, and
the top 100 matches were downloaded. In Klee diagram, numbers indicate species, asterisk
marks T. migratorius sequences, and indicator vector correlation scale is at right, with 1 rep-
resenting 100% sequence identity.
10 STOECKLE, THALER
species—as the primary fact that a model of evolution and speciation must explain. The
pattern of life seen in barcodes is a commensurable whole made from thousands of in-
dividual studies that together yield a generalization. The clustering of barcodes has two
equally important features: 1) the variance within clusters is low, and 2) the sequence
gap among clusters is empty, i.e., intermediates are not found. Beyond the qualitative
descriptor “low” for the variance within species there is a quantitative statement. The
average pairwise difference among individuals (APD; equivalent to population genetics
parameter π) within animal species is between 0.0% and 0.5%. The most data are avail-
able for modern humans, who have an APD of 0.1% calculated in the same way as for
other animals (See Fig. 2 in [34] and Fig. 7 in this paper).
The agreement of barcodes and domain experts implies that explaining the origin of
the pattern of DNA barcodes would be in large part explaining the origin of species. Un-
derstanding the mechanism by which the near-universal pattern of DNA barcodes comes
about would be tantamount to understanding the mechanism of speciation.
The clustering pattern of life was elegantly articulated by Dobzhansky in his 1937
book Genetics and the Origin of Species [35] from which an extensive quote is merited.
Only through DNA barcodes can the same metric be used so that the “feeling that it
must be right” can now be given a single quantitative meaning across the entire animal
kingdom:
If we assemble as many individuals living at a given time as we can, we notice that
the observed variation does not form a single probability distribution or any other kind
of continuous distribution. Instead, a multitude of separate, discrete, distributions are
found. In other words, the living world is not a single array of individuals in which any
two variants are connected by unbroken series of intergrades, but an array of more or
less distinctly separate arrays, intermediates between which are absent or at least rare.
Each array is a cluster of individuals, usually possessing some common characteristics
and gravitating to a denite modal point in their variation.… Therefore the biological
classication is simultaneously a man-made system of pigeonholes devised for the prag-
matic purpose of recording observations in a convenient manner and an acknowledge-
ment of the fact of organic discontinuity.
Two models have the potential to explain the structure of COI barcodes in the
extant animal kingdom
Either 1) COI barcode clusters represent species-specic adaptations, OR 2) ex-
tant populations have recently passed through diversity-reducing regimes whose conse-
quences for sequence diversity are indistinguishable from clonal bottlenecks. This way
11WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
Fig. 5. mtDNA clusters reect synonymous substitutions. Charts depict nucleotide and
amino acid differences from the mode for congeneric COI barcode sets in Fig. 3. Nucleotide
differences are colorized (A=green; C=blue; G=black; T=red). To minimize contribution of
sequence errors and missing data, the 648 bp barcode region is trimmed by 10% at either end,
leaving 519 nt/173 amino acids. At right, synonymous (S) and non-synonymous (N) average
pairwise distances within (W) and between (B) species. Horizontal bar indicates mean and
vertical line indicates maximum and minimum.
12 STOECKLE, THALER
of framing the problem is similar to that raised by the analysis of isozymes by electro-
phoresis more than half a century previous [36]. The key difference being that the COI
barcode data are vast and commensurable across the animal kingdom. “Commensura-
ble” means using the same measurement and being directly comparable. “The “awesome
power of” [37] mitochondrial COI barcodes allows the same metric to be used animal
kingdom-wide. Without DNA barcodes, generalizations to all animals have to be based
on putting together data sets gathered and analyzed by different methods.
Most barcode variation among neighboring species—and also within species—con-
sists of synonymous codon changes (Fig. 5). The question that determines which of the
two mechanisms is most plausible is whether or not synonymous codons in the mito-
chondrial genome are selectively neutral. If purifying selection does not act on synony-
mous codons in the mitochondrial genome then demographic processes must be acting
to suppress neutral variance.
Are synonymous codons in mitochondria neutral?
Comparative rates along phylogenies have been used to argue that amino acid
changes in the mitochondrial genome are subject to purifying selection but synonymous
substitutions are not [23, 38, 39]. Across the animal kingdom the preponderance of SNP
variation in mitochondrial sequences consists of synonymous codon changes. Are these
synonymous codons targets for purifying and/or adaptive selection strong enough to
be responsible for the low variance within species and/or the different consensus se-
quence among neighboring species? Codon bias in the mitochondrial genome has been
shown at the phylogenetic level of order but there is no evidence for different codon bias
among neighboring species [40, 41]. Furthermore, the number of synonymous codons
relevant to the discussion of DNA barcodes (0.0%-0.5% within species, 0.0%-5.0% for
neighboring species) is not enough to alter codon bias. Nearby species do not differ in
overall codon bias or GC ratios [40, 41], an observation in contrast to a prediction of the
hypothesis that GC bias is an important factor in speciation [42]. If synonymous codons
are differentially selected in DNA barcodes, this selection must be acting at the level of
the placement of individual codons rather than their cumulative average effects on base
composition.
13WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
An assumption of neutrality for synonymous codons is no longer a “slam dunk”
[43], i.e., is not a certain conclusion.
Avise et al in 1987 made an absolutist statement on the irrelevance of synonymous
codons to selective tness:
First, in a mechanistic sense, we already “know” that most of the particular mtDNA
genotypic variants segregating in populations probably have, by themselves, absolutely
no differential effect on organismal tness. These include, for example, base substitu-
tions in silent positions of protein-coding genes, and some substitutions and small addi-
tion/ deletions in the nontranscribed D-Ioop region. These changes are disproportion-
ately common in mtDNA [25] and are ones for which only the most ardent selectionist
would argue a direct link to organismal tness.
This statement merits critical evaluation in light of three intervening decades of
molecular genetics. (Spoiler alert: in this case re-examination strengthens the original
assertion for reasons that the authors themselves could not have anticipated.)
Since 1987 numerous examples have emerged where even very few synonymous
codon changes make important and selectable differences in organismal, cellular, or viral
physiology [44-50].
Synonymous codons may also modulate protein folding or membrane insertion
concomitant with translation [51], as suggested for synonymous codons that modulate
phenotypes of the cystic brosis transmembrane conductance regulator [52] or the p-
glycoprotein (multidrug resistance protein) [53]. The cited cases involve cytoplasmic
translation of bacteria, bacteriophage, and drosophila nuclear genes. However, they re-
quire one to critically re-examine the assertion of absolute and universal neutrality with
regard to mitochondria.
New justication for an old assertion.
The proven cases where synonymous codons have phenotypes are all consequent to
their differential rate of translation. Codon-specic translation rates are in turn attributed
to different concentrations of codon-specic tRNAs, known as “isoacceptors.” This has
been proven directly in an experimental system where overexpression of the cognate
tRNA changes a low-expression codon to a high one [54, 55]. Isoacceptor concentra-
tions differ between species and even within individuals in a tissue-specic manner [56,
57]. The frequencies of synonymous codons and the concentration of their isoaccep-
tor tRNAs with complementary anticodons coevolve [58]. Codon-specic isoacceptor
tRNAs are tightly and dynamically regulated; they play important roles in the differential
14 STOECKLE, THALER
regulation of gene expression [59]. The human nuclear genome encodes tRNAs with
51 distinct anticodons for the 20 amino acids [60]. In addition to isoacceptors there are
dozens to hundreds of “isodecoder” genes in the nuclear genome. Isodecoders are tRNAs
that share an anticodon sequence but differ elsewhere. The different sequences of iso-
decoders are often also associated with different post-transcriptional modications [60].
It is likely that isodecoders add a further important layer to differential gene expression
depending on a codon’s sequence and tissue context but this remains to be proven.
In striking contrast to the multitude of different nuclear tRNA genes and cytoplas-
mic tRNAs animal mitochondria have only 22 different tRNA types to translate the 20
amino acids [61, 62]. With two exceptions, isoacceptor tRNAs are not available inside
animal mitochondria. Leucine and serine each have two mitochondrial tRNAs with dif-
ferent anticodons; the remaining 18 amino acids are each translated by a single tRNA
that covers all cognate codons. The best documented mechanism for altering the ef-
ciency of translation is when the changed codon(s) are near to the rst-translated end
of the gene [54]. The approximately even distribution of synonymous variation among
mitochondrial genes in modern humans [34] is most compatible with neutrality [63].
Speculatively, synonymous codon changes could affect gene expression by mecha-
nisms independent of tRNAs. These include changes in mRNA secondary structure and
stability and the binding of specic factors, protein or miRNA. Modication of splic-
ing is a candidate and possibly important mechanism by which synonymous codons
alter protein structure and function. However, in contrast to most nuclear genes in most
animals, animal mitochondrial protein-encoding genes do not have introns. In animal
mitochondria there are no alternative spliced forms whose ratios could be modulated by
synonymous codons near splice sites.
Kimura’s insight that a preponderance of synonymous substitutions is evidence for
neutral evolution [64] now appears to be more universally valid for the mitochondrial
than for the nuclear genome. For nuclear genomes one nds a growing number of cases
and mechanisms where synonymous codons have phenotypes and are subject to selec-
tion. In contrast, for the mitochondrial genomes of animals there is not a single example
of any synonymous codon having a phenotype. Furthermore, the known mechanisms
that allow synonymous codons to alter the phenotypes of nuclear genes are impossible
in mitochondria. Mitochondrial sequences yield straightforward and uncomplicated phy-
logenetic analysis and species-level identication for reasons beyond those known by
those who originally proposed them. Sometimes you get lucky.
15WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
The case of the missing G’s.
Codons that end in G are underrepresented by a factor of about ten in animal mi-
tochondria. Previously we interpreted the lack of third position G’s in mitochondrial
coding sequences is evidence of a role for extreme purifying selection in determining the
COI DNA barcoding gap [20] but we now nd this argument awed. On the one hand
there appears to be purifying selection against codons that end in G but this apparent
selection is similar in neighboring species. With selection against G-ending codons in all
animal species it could not be a source of species-specic adaptive peaks. Further insight
into the lack of G in the third position follows from a focused review of the wobble hy-
pothesis in the context of mitochondria.
Francis Crick set forth a set of stereochemical models in which a single tRNA an-
ticodon pairs with multiple codons for the same amino acid. Crick called his idea the
“Wobble hypothesis” because it postulated exible pairing between the 3’ base of the co-
don with the 5’ base of the anticodon [65]. The Wobble hypothesis is brilliantly insight-
ful, however, details of the pairing scheme have changed with knowledge of extensive
post-transcriptional chemical modications of tRNAs. More than 150 different chemical
modications of RNAs are now characterized [66]; the greatest concentration of RNA
modication is found on anticodons. Only certain modications at the U at the 5’ posi-
tion in the anticodon allow efcient Wobble-pairing with G [67-70]. Wobble G is rare
in animal mitochondrial codons[20] consistent with the fact that Wobble G recognition-
specic modications have not been found in animal mitochondrial tRNAs [71, 72].
Several human pathologies are correlated by Genome Wide Association Studies
(GWAS) to SNPs that change a synonymous codon and it is expected that more will be
found [73]. Two cautions apply when considering Genome Wide Association Studies
linking synonymous codons with human mitopathologies: 1) So far as we are aware
there are no inferences of pathologies based on synonymous substitutions in the mi-
togenome [74], 2) GWAS are subject to artifacts of inference that encourage erroneous
condence [75, 76]. GWAS are hypothesis generators, not proof. Anecdotally, workers
in mitochondrial pathologies are well aware of synonymous codons, and so the absence
of evidence for any human pathology due to a synonymous codon change in the mi-
tochondrial genome is not due to a lack of looking. In contrast, missense mutations in
mitochondrial tRNA genes are “hotspots” for human pathologies [77]; they lead to large
scale insertion of inappropriate amino acids.
16 STOECKLE, THALER
Selectionist arguments.
Thomas Kuhn points out that science stalls when different camps that study the
same aspects of nature use different vocabularies, cite only within their own discipline
and ignore or disparage each other. There appears to be an unfortunate isolation in the
literature between camps that advocate mitochondrial selection and those that rely on
demographic reasoning. Here we argue that selectionist perspectives are valid in some
mitochondrial systems and plausible in others. However, exceptions abound and as a
whole we nd that a selectionist perspective is not robust enough to account for the
animal-kingdom wide facts of the barcode gap.
Even within a single species, different external environments may select for par-
ticular alleles of mitochondrial-encoded enzymes [78-81]. In modern humans there are
two different amino acid alleles of the mitochondrial-encoded ATPase. This subunit par-
titions the proton gradient of mitochondria in two ways: it can use the gradient to form
the covalent to join inorganic phosphate to ADP in order to make ATP. Alternatively,
if the proton gradient runs down without storing energy in the synthesis of ATP heat
is immediately released. The allele predisposed to ATP synthesis is more frequent in
human populations who inhabit tropical regions. Conversely, the allele biased toward
instantaneous heat generation is more frequent in colder regions [82]. The argument for
environmentally-driven selection for this allele is logical and inspires interesting experi-
ments [83]. The plausible but unproven possibility of selection for a single allelic case
of amino acid substitution is a small pebble in the scale when compared to the evidence
for the apparent neutrality of most mitochondrial variation. Most human mitochondrial
variation, similar to that of other animal species, consists of synonymous codon changes.
However, in principle, the linkage of a single selected amino acid could drive a species-
wide sweep of the entire linked mitochondrial genome.
The model of an optimum sequence for each species has two subcategories: a) op-
timal for the external environment, and b) optimal integration with other genes of the
organism [84]. These two mechanisms can work together and various permutations have
been suggested with more or less emphasis on selection for external conditions such
as environmental temperature or internal compatibility with nuclear genes [80, 85-87].
Compatibility among the thousand or so nuclear genes whose products enter the mito-
chondria and the 13 gene products coded for by the mitochondrial genome can lead to
reproductive isolation and incompatibility [21, 86, 88]. Incompatibility of mitochondrial
and nuclear genes can cause reproductive isolation either immediately or via decreased
tness of progeny [87, 89-92]. Mitochondrial introgression in some cases has been pro-
posed to favor the co-introgression of compatible nuclear alleles that form subunits of
mitochondrial complexes [93, 94]. This is a potentially important perspective but its
generality is unclear.
17WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
We see arguments against stepping from specic examples of nuclear-mitochondri-
al incompatibilities to a general theory of this effect as the major driver of animal specia-
tion: 1) There are many examples of fertile and t interspecies hybridizations including
cases with mitochondrial sequences that diverge by 4% or more [95-102]. 2) Some well-
studied cases of mating incompatibility have nothing to do with nuclear-mitochondrial
incompatibility. Examples include chromosomal inversion [103, 104] and the activation
of endogenous retroviruses-transposons [105]. 3) Nuclear-mitochondrial incompatibility
is a subset of physiological incompatibility. Other denitions of species, e.g. behav-
ioral, geographical, ecological, do not require physiological mating incompatibility in
any form [11]. COI barcode clustering is more widely a fact than can be accounted for
by mating incompatibility in general and nuclear-mitochondrial incompatibility in par-
ticular. 5) Finally, there is no example in which mating incompatibility or weakness of
inter-species hybrids is attributable to the synonymous codons that constitute the major
fraction of barcode gaps.
The average pairwise difference of the COI barcode in modern humans is 0.1%, i.e.,
about average for the animal kingdom. However, the most extreme differences between
individual humans approach 1%. This difference is as great as many distinctions among
neighboring species. Modern humans are a single population. Darwin made this point
with respect to visible phenotypes and it applies even more strongly when neutral vari-
ants are considered:
Hereafter, we shall be compelled to acknowledge that the only distinction between
species and well-marked varieties is, that the latter are known, or believed, to be con-
nected at the present day by intermediate gradations, whereas species were formerly
thus connected [106].
The possibility of preferred combinations of nuclear and mitochondrial alleles with-
in a species is intriguing and there is one example of experimental support. An inbred
strain of mouse was shown to have non-optimal physiology when the mitochondrial
genome from a different inbred line was crossed in (10 backcrosses to the nuclear line
all using the female descendent from the rst mitochondrial donor) [107]. This nding
has been extrapolated as justication to urge studies of nuclear-mitochondrial compat-
ibility in human three-parent IVF (in vitro fertilization) [108]. On the other hand, hu-
man mitochondrial transfer experiments have found no analogous effect [109], arguably,
owing to the different genetic structure of our species when compared to inbred mouse
strains [110]. The differences in the two mouse mitochondrial genomes at issue include
missense in the coding region, tRNA alterations and ori-region changes as well as syn-
onymous codon changes. There are no data to pinpoint which sequences make a differ-
ence, in particular no evidence for a phenotype of synonymous codon changes, which the
authors mark as “silent” (Extended Data Table 1 in [107]).
18 STOECKLE, THALER
Fig. 6. Fertile hybrids. Mytilus mussels exhibit complex patterns of mitochondrial and
nuclear introgression, reecting multiple historical and recent hybridization events,
some following introduction of non-native species for aquaculture. F1 hybrids are fertile
even though parental species differ by 10-20% in COI nucleotide sequence. This sup-
ports view that mtDNA clustering is not due to species-specic adaptations.
19WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
This line of work and controversy adds evidence that in some cases mitochondrial-
nuclear incompatibility may interfere with mating or health of offspring. However, the
work does not show any effect of synonymous codon changes. No matter which mecha-
nism for speciation is responsible in any specic case, the 0.0%-0.5% accumulation
of synonymous variance independent of population size or apparent species age is a
biological fact. The variable distance between the most closely related living species
presumably reects differing numbers of extinct intermediate sequences.
Conditions that favor clonal uniformity are frequent in biology
Bottlenecks, founder effects, lineage sorting, and gene sweeps decrease genetic di-
versity in a population [111]. The question is how widespread these effects are in the
context of dening animal species and if it is possible distinguish them in other than a
rhetorical manner. Here we emphasize the overlap—in fact the near congruence—in the
conditions that favor each of these mechanisms.
Based on contemporary mitochondrial sequence data alone it is impossible to distin-
guish an organismal bottleneck from mitochondrial and Y chromosome specic lineage
sorting since both mechanisms make the same prediction of a uniform mitochondrial
sequence in the past [112].
A positively selected allele has the potential to sweep through a population and
by hitchhiking [113, 114] or genetic draft [115] carry the entirety of the linked genome
along thereby resetting mitochondrial variation to zero. This scenario requires a single
maternal lineage replace all others [113]. It is reasonable to hypothesize that somewhere
on the mitochondrial genome there arises a positively selected amino acid substitution
leading to the replacement of the entire linked genome in the entire population. One
should not mince words about what a mitochondrial genome sweep requires: the entire
population’s mitochondrial genome must re-originate from a single mother.
These three pathways toward sequence uniformity should not be thought of as en-
emies because they converge in both cause and effect. Lineage sorting is most efcient
when the population is small, when the number of different mitochondrial genotypes
is small, and when the population is either stable or shrinking [116]. An extreme di-
minishment of population size followed by population expansion is the denition of a
bottleneck. Lineage sorting is diminished during periods of population growth and does
not occur at all during exponential growth when all neutral lineages leave progeny [117,
118]. The same conditions that favor lineage sorting also favor gene sweeps, which in
the context of a totally-linked genome means one mitochondrial genome. The concept
of “gene sweeps” emphasizes positive selection whereas “lineage sorting” emphasizes
neutral events. Bottlenecks are extreme forms of the same conditions.
20 STOECKLE, THALER
Bottlenecks followed by expansion are the dominant mechanism for evolution in
the microbial majority of life and it might seem odd to think animals should be excep-
tional [119]. Ever since Koch, microbiologists have streaked out their bacteria to begin
experiments with pure, i.e., clonal cultures [120]. The rst experiments showing evolu-
tion of new mutants from clonal starting populations were the classical cases of proof
that bacterial genes follow the patterns expected from random mutation that grow indis-
tinguishably from sibs when unselected ([117, 121-123]). Clonal outgrowth and replace-
ment of the inoculating population was inferred from the earliest chemostat [124] as well
as later serial transfer experiments [125]. Epidemiology shows that repeated bottlenecks
play dominant roles in the natural evolution of microbial pathogens including protists,
bacteria and viruses [126-132]. A visually impressive demonstration of successive clonal
selection and population outgrowth is seen in time lapse studies of bacteria serially mu-
tating to new heights of antibiotic resistance [133]. On the host side of the equation, the
Fig. 7. Kimura’s equilibrium model alone is insufcient to account for usual levels of
intraspecic variation in animal species. APD and census population size for 112 bird spe-
cies without phylogeographic clusters are shown. Dashed line is expected APD limit due to
(AVP = 2 N μ, where N = population size and μ = mutation rate, using 10-8 substitutions/site/
generation, or 1% per My, assuming generation time is 1 y). Average effective population
size in the birds shown is 70 thousand (range 0-300 thousand); average census population
size is 30 million (range 5 thousand to 500 million). Human mitochondrial variation (popula-
tion 7.5 billion, APD 0.1%) is typical of that in other animal species.
21WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
clonal selection theory of immune system development was controversial when rst
proposed but its logic and experimental support proved compelling [134, 135].
Mayr made a specic proposal for the role of extreme bottlenecks in speciation that
followed from a founder effect (originally 1942, here quoted from a reprise based on
interviews in 2004 [136]):
The reduced variability of small populations is not always due to accidental gene
loss, but sometimes to the fact that the entire population was started by a single pair or
by a single fertilized female. These “founders” of the population carried with them only
a very small proportion of the variability of the parent population. This “founder” prin-
ciple sometimes explains even the uniformity of rather large populations, particularly if
they are well isolated and near the borders of the range of the species.
Eldredge and Gould used this idea of allopatric speciation in small isolated popula-
tions that then rapidly expanded to account for the abrupt transitions seen in the broad
range of the fossil record [137].
Models of allopatric or peripatetic speciation invoke a bottleneck with an additional
feature: What emerges from the bottleneck looks or acts differently, i.e., it is a bona de
new species. It may be more frequent that what emerges from a bottleneck looks and acts
like a middling representative of what went in.
If mitochondria are considered “honorary prokaryotes” then the dominant mode in
prokaryotes of frequent processes that lead to clonal outgrowth either by selection or ran-
dom processes [138] are not counterintuitive. A number of different processes could lead
to the mitochondrial sequence becoming clonal. Candidate processes include bottlenecks
and lineage sorting on three different levels: Within organelles, among organelles in the
same cells, among cells in an organism (particularly in the germ line) and among organ-
isms. Not certain is whether different processes have led to a similar result throughout
the animal kingdom or if a single process operates throughout. Occam’s razor, the prin-
ciple of parsimony, suggests that a single explanation should be considered.
Purifying selection in linked genomes slows but does not stop the accumulation of
neutral variation [139]. Drift and lineage sorting during population stasis or shrinkage
decrease variation. The efciency of decrease depends on the number of haplotypes
in the population, as well as the numbers and distributions of female offspring among
parents with different haplotypes [10]. A key prediction of naïve neutral theory that does
not hold up against extensive barcode data from across the animal kingdom is that larger
populations or older species should harbor more neutral variation [20, 140, 141]. The
key incompatibility of naïve neutral theory with biological fact is that the theory consid-
ers populations at equilibrium in the sense that the population be at stable numbers for
22 STOECKLE, THALER
approximately as many generations as the mutation rate per generation. The evolution
of modern humans offers a specic solution to the animal-kingdom-wide dilemma of
missing neutral mutations.
Modern humans
More approaches have been brought to bear on the emergence and outgrowth of
Homo sapiens sapiens (i.e., modern humans) than any other species including full ge-
nome sequence analysis of thousands of individuals and tens of thousands of mitochon-
dria, paleontology, anthropology, history and linguistics [61, 142-144]. The congruence
of these elds supports the view that modern human mitochondria and Y chromosome
originated from conditions that imposed a single sequence on these genetic elements
between 100,000 and 200,000 years ago [145-147]. Contemporary sequence data cannot
tell whether mitochondrial and Y chromosomes clonality occurred at the same time, i.e.,
consistent with the extreme bottleneck of a founding pair, or via sorting within a found-
ing population of thousands that was stable for tens of thousands of years [116]. As Kuhn
points out unresolvable arguments tend toward rhetoric.
Summary and conclusion
Science greedily seizes simplicity among complexities. Speciation occurs via alter-
native pathways distinct in terms of the number of genes involved and the abruptness
of transitions [148]. Nuclear variance in modern humans varies by loci in part due to
unequal selection [149] and the linkage of neutral sites to those that undergo differential
selection. Complexity is the norm when dealing with variance of the nuclear ensemble
[150-154]. It is remarkable that despite the diversity of speciation mechanisms and path-
ways the mitochondrial sequence variance in almost all extant animal species should be
constrained within narrow parameters.
Mostly synonymous and apparently neutral variation in mitochondria within spe-
cies shows a similar quantitative pattern across the entire animal kingdom. The pattern
is that that most—over 90% in the best characterized groups—of the approximately ve
million barcode sequences cluster into groups with between 0.0% and 0.5% variance as
measured by APD, with an average APD of 0.2%.
Modern humans are a low-average animal species in terms of the APD. The molecu-
lar clock as a heuristic marks 1% sequence divergence per million years which is consis-
tent with evidence for a clonal stage of human mitochondria between 100,000- 200,000
years ago and the 0.1% APD found in the modern human population [34, 155, 156]. A
conjunction of factors could bring about the same result. However, one should not as a
23WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
rst impulse seek a complex and multifaceted explanation for one of the clearest, most
data rich and general facts in all of evolution. The simple hypothesis is that the same
explanation offered for the sequence variation found among modern humans applies
equally to the modern populations of essentially all other animal species. Namely that
the extant population, no matter what its current size or similarity to fossils of any age,
has expanded from mitochondrial uniformity within the past 200,000 years.
Nonhuman animals, as well as bacteria and yeast, are often considered “model sys-
tems” whose results can be extrapolated to humans. The direction of inference is re-
versible. Fossil evidence for mammalian evolution in Africa implies that most species
started with small founding populations and later expanded [157] and sequence analysis
has been interpreted to suggest that the last ice age created widespread conditions for a
subsequent expansion [158]. The characteristics of contemporary mitochondrial vari-
ance may represent a rare snapshot of animal life evolving during a special period. Al-
ternatively, the similarity in variance within species could be a sign or a consequence of
coevolution [159].
Mitochondria drive many important processes of life [160-162]. There is irony but
also grandeur in this view that, precisely because they have no phenotype, synonymous
codon variations in mitochondria reveal the structure of species and the mechanism of
speciation. This vista of evolution is best seen from the passenger seat.
Acknowledgements. Bruce Levin suggested the phrase “honorary prokaryote” in reference to mito-
chondria. Others have used the phrase in reference yeast or to phage introns. Thanks to Glen Bjork,
Manny Goldman, Ken Zahn for discussions on tRNAs, Jesse H. Ausubel, Frank Stahl for encourage-
ment and comments and the Alfred P. Sloan Foundation and Monmouth University/Rockefeller Univer-
sity Marine Science and Policy Initiative for support.
24 STOECKLE, THALER
References
1. Avise, J.C. and K. Wollenberg, Phylogenetics and the origin of species. Proc Natl Acad Sci U S A,
1997. 94(15): p. 7748-55.
2. Hebert, P.D.N, et al., Biological identications through DNA barcodes. Proc Biol Sci, 2003.
270(1512): p. 313-21.
3. Stoeckle, M., Taxonomy, DNA, and the bar code of life. BioScience, 2003. 53: p. 2-3.
4. Avise, J.C., et al., Intraspecic phylogeography: the mitochondrial bridge between population ge-
netics and systematics. Ann Rev Ecol Syst, 1987. 18: p. 489–522.
5. Moore, W.S., Inferring phylogenies from mtDNA variation: Mitochondrial-gene trees versus nucle-
ar-gene trees. Evolution, 1995. 49: p. 718–726.
6. Gladwell, M., The Tipping Point: How Little Things Can Make a Big Difference. 2000: Little
Brown. p. 279.
7. Hebert, P.D.N., et al., Identication of birds through DNA barcodes. PLOS Biol, 2004. 2(10): p.
e312.
8. Galtier, N., et al., Mitochondrial DNA as a marker of molecular diversity: a reappraisal. Molec
Ecol, 2009. 18: p. 4541–4550.
9. Hickerson, M.J., C.P. Meyer, and C. Moritz, DNA barcoding will often fail to discover new animal
species over broad parameter space. Syst Biol, 2006. 55(5): p. 729-39.
10. Avise, J.C., B.W. Bowen, and F.J. Ayala, In the light of evolution X: Comparative phylogeography.
Proc Natl Acad Sci U S A, 2016. 113(29): p. 7957-61.
11. de Queiroz, K., Species concepts and species delimitation. Syst Biol, 2007. 56: p. 879–886.
12. Albert, B., et al., Dynamics of plant mitochondrial genome: model of a three-level selection process.
Genetics, 1996. 144(1): p. 369-82.
13. Taylor, D.R., C. Zeyl, and E. Cooke, Conicting levels of selection in the accumulation of mito-
chondrial defects in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A, 2002. 99(6): p. 3690-4.
14. Birky, C.W., Jr., On the origin of mitochondrial mutants: evidence for intracellular selection of
mitochondria in the origin of antibiotic-resistant cells in yeast. Genetics, 1973. 74(3): p. 421-32.
15. Otten, A.B., et al., Differences in strength and timing of the mtDNA bottleneck between zebrash
germline and non-germline cells. Cell Rep, 2016. 16(3): p. 622-30.
16. Rebolledo-Jaramillo, B., et al., Maternal age effect and severe germ-line bottleneck in the inheri-
tance of human mitochondrial DNA. Proc Natl Acad Sci U S A, 2014. 111(43): p. 15474-9.
17. Vevea, J.D., et al., Inheritance of the ttest mitochondria in yeast. Trends Cell Biol, 2014. 24(1):
p. 53-60.
18. Ballard, J.W., B. Chernoff, and A.C. James, Divergence of mitochondrial DNA is not corroborated
by nuclear DNA, morphology, or behavior in Drosophila simulans. Evolution, 2002. 56(3): p. 527-
45.
19. Song, H., et al., Many species in one: DNA barcoding overestimates the number of species when
nuclear mitochondrial pseudogenes are coamplied. Proc Natl Acad Sci U S A, 2008. 105(36): p.
13486-91.
20. Stoeckle, M.Y. and D.S. Thaler, DNA barcoding works in practice but not in (neutral) theory. PLOS
ONE, 2014. 9(7): p. e100755.
21. Hill, G.E., Mitonuclear coevolution as the genesis of speciation and the mitochondrial DNA bar-
code gap. Ecol Evol, 2016. 6(16): p. 5831-42.
22. Hebert, P.D., P.M. Hollingsworth, and M. Hajibabaei, From writing to reading the encyclopedia of
life. Philos Trans R Soc Lond B Biol Sci, 2016. 371(1702).
25WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
23. Pentinsaari, M., et al., Molecular evolution of a widely-adopted taxonomic marker (COI) across the
animal tree of life. Sci Rep, 2016. 6: p. 35275.
24. Zachos, F.E., Species Concepts in Biology: Historical Development, Theoretical Foundations and
Practical Relevance. 2016, Switzerland: Springer International. p. 220.
25. Mayr, E. The species problem: A symposium presented at the Atlanta meeting of the American Asso-
ciation for the Advancement of Science, December 28–29, 1955. 1957. Washington, DC: American
Association for the Advancement of Science.
26. Margulis, L., Symbiosis in Cell Evolution. 1981: WH Freeman. p. 419.
27. Groves, C., Species concept in primates. Am J Primatol, 2012. 74(8): p. 687-91.
28. Zachos, F.E., Taxonomy: Species splitting puts conservation at risk. Nature, 2013. 494(7435): p. 35.
29. Bateson, W., The Methods and Scope of Genetics. 1908: Cambridge University Press.
30. Van Valen, L., A new evolutionary law. Evolutionary Theory, 1973. 1: p. 1-30.
31. Buss, L.W., The Evolution of Individuality. 1987: Princeton University Press. p. 201.
32. Gould, S., The Structure of Evolutionary Theory. 2002: Belknap Press. p. 1464.
33. Kuhn, T., The Structure of Scientic Revolutions 1962: University of Chicago Press. p. 226.
34. Thaler, D.S. and M.Y. Stoeckle, Bridging two scholarly islands enriches both: COI DNA barcodes
for species identication versus human mitochondrial variation for the study of migrations and
pathologies. Ecol Evol, 2016. 6: p. 6824–6835.
35. Dobzhansky, T., Genetics and the Origin of Species 3rd ed. 1937, New York: Columbia University
Press.
36. Lewontin, R.C., Twenty-ve years ago in Genetics: electrophoresis in the development of evolution-
ary genetics: milestone or millstone? Genetics, 1991. 128(4): p. 657-62.
37. Stahl, F. and C. Gross, Ira Herskowitz (1946-2003). Cell, 2003. 114: p. 9-10.
38. Castellana, S., S. Vicario, and C. Saccone, Evolutionary patterns of the mitochondrial genome in
Metazoa: exploring the role of mutation and selection in mitochondrial protein coding genes. Ge-
nome Biol Evol, 2011. 3: 1067-79.
39. Kerr, K.C., Searching for evidence of selection in avian DNA barcodes. Mol Ecol Resour, 2011.
11(6): p. 1045-55.
40. Selva Kumar, C., et al., Inuence of certain forces on evolution of synonymous codon usage bias
in certain species of three basal orders of aquatic insects. Mitochondrial DNA, 2012. 23(6): p.
447-60.
41. Wei, L., et al., Analysis of codon usage bias of mitochondrial genome in Bombyx mori and its rela-
tion to evolution. BMC Evol Biol, 2014. 14: p. 262.
42. Forsdyke, D., Base composition, speciation, and why the mitochondrial barcode precisely classi-
es. Biological Theory, 2017. 12: p. 157-168.
43. Scientic language is becoming more informal. Nature, 2016. 539(7628): p. 140.
44. Saikia, M., et al., Codon optimality controls differential mRNA translation during amino acid star-
vation. RNA, 2016. 22(11): p. 1719-1727.
45. Komar, A.A., The yin and yang of codon usage. Hum Mol Genet, 2016. 25(R2): p. R77-R85.
46. Supek, F., The code of silence: widespread associations between synonymous codon biases and
gene function. J Mol Evol, 2016. 82(1): p. 65-73.
47. Bull, J.J., Evolutionary reversion of live viral vaccines: Can genetic engineering subdue it? Virus
Evol, 2015. 1(1): p. vev005.
48. Li, M., et al., Codon-usage-based inhibition of HIV protein synthesis by human schlafen 11. Nature,
2012. 491(7422): p. 125-8.
49. Napolitano, M.G., et al., Emergent rules for codon choice elucidated by editing rare arginine co-
dons in Escherichia coli. Proc Natl Acad Sci U S A, 2016. 113(38): p. E5588-97.
26 STOECKLE, THALER
50. Lajoie, M.J., D. Soll, and G.M. Church, Overcoming challenges in engineering the genetic code. J
Mol Biol, 2016. 428(5 Pt B): p. 1004-21.
51. Quax, T.E., et al., Codon bias as a means to ne-tune gene expression. Mol Cell, 2015. 59(2): p.
149-61.
52. Bali, V., et al., A synonymous codon change alters the drug sensitivity of DeltaF508 cystic brosis
transmembrane conductance regulator. FASEB J, 2016. 30(1): p. 201-13.
53. Bartoszewski, R., et al., Codon bias and the folding dynamics of the cystic brosis transmembrane
conductance regulator. Cell Mol Biol Lett, 2016. 21: p. 23.
54. Zahn, K. and A. Landy, Modulation of lambda integrase synthesis by rare arginine tRNA. Mol
Microbiol, 1996. 21(1): p. 69-76.
55. Powell, J.R. and K. Dion, Effects of codon usage on gene expression: empirical studies on Dro-
sophila. J Mol Evol, 2015. 80(3-4): p. 219-26.
56. Bulmer, M., The selection-mutation-drift theory of synonymous codon usage. Genetics, 1991.
129(3): p. 897-907.
57. Bulmer, M., Coevolution of codon usage and transfer RNA abundance. Nature, 1987. 325(6106):
p. 728-30.
58. Krakauer, D.C. and V.A. Jansen, Red queen dynamics of protein translation. J Theor Biol, 2002.
218(1): p. 97-109.
59. Wilusz, J.E., Controlling translation via modulation of tRNA levels. Wiley Interdiscip Rev RNA,
2015. 6(4): p. 453-70.
60. Parisien, M., X. Wang, and T. Pan, Diversity of human tRNA genes from the 1000-genomes project.
RNA Biol, 2013. 10(12): p. 1853-67.
61. Lott, M.T., et al., mtDNA Variation and Analysis Using Mitomap and Mitomaster. Curr Protoc Bio-
informatics, 2013. 44: p. 1.23.1-26.
62. Salinas-Giege, T., R. Giege, and P. Giege, tRNA biology in mitochondria. Int J Mol Sci, 2015. 16(3):
p. 4518-59.
63. Kimura, M., The neutral theory of molecular evolution and the world view of the neutralists. Ge-
nome, 1989. 31(1): p. 24-31.
64. Kimura, M., Preponderance of synonymous changes as evidence for the neutral theory of molecular
evolution. Nature, 1977. 267(5608): p. 275-6.
65. Crick, F.H., Codon--anticodon pairing: the wobble hypothesis. J Mol Biol, 1966. 19(2): p. 548-55.
66. Machnicka, M.A., et al., MODOMICS: a database of RNA modication pathways--2013 update.
Nucleic Acids Res, 2013. 41(Database issue): p. D262-7.
67. Yokoyama, S., et al., Molecular mechanism of codon recognition by tRNA species with modied
uridine in the rst position of the anticodon. Proc Natl Acad Sci U S A, 1985. 82(15): p. 4905-9.
68. Osawa, S., et al., Recent evidence for evolution of the genetic code. Microbiol Rev, 1992. 56(1): p.
229-64.
69. Nasvall, S.J., P. Chen, and G.R. Bjork, The wobble hypothesis revisited: uridine-5-oxyacetic acid is
critical for reading of G-ending codons. RNA, 2007. 13(12): p. 2151-64.
70. Rozov, A., et al., Novel base-pairing interactions at the tRNA wobble position crucial for accurate
reading of the genetic code. Nat Commun, 2016. 7: p. 10457.
71. Watanabe, K., Unique features of animal mitochondrial translation systems. The non-universal
genetic code, unusual features of the translational apparatus and their relevance to human mito-
chondrial diseases. Proc Jpn Acad Ser B Phys Biol Sci, 2010. 86(1): p. 11-39.
72. Suzuki, T. and T. Suzuki, A complete landscape of post-transcriptional modications in mammalian
mitochondrial tRNAs. Nucleic Acids Res, 2014. 42(11): p. 7346-57.
27WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
73. Bali, V. and Z. Bebok, Decoding mechanisms by which silent codon changes inuence protein bio-
genesis and function. Int J Biochem Cell Biol, 2015. 64: p. 58-74.
74. Alston, C.L., et al., The genetics and pathology of mitochondrial disease. J Pathol, 2016.
75. Wood, A.R., et al., Another explanation for apparent epistasis. Nature, 2014. 514(7520): p. E3-5.
76. Fish, A.E., J.A. Capra, and W.S. Bush, Are interactions between cis-regulatory variants evidence
for biological epistasis or statistical artifacts? Am J Hum Genet, 2016. 99(4): p. 817-830.
77. Abbott, J.A., C.S. Francklyn, and S.M. Robey-Bond, Transfer RNA and human disease. Front Gen-
et, 2014. 5: p. 158.
78. Ruiz-Pesini, E., et al., Effects of purifying and adaptive selection on regional variation in human
mtDNA. Science, 2004. 303(5655): p. 223-6.
79. Wallace, D.C., The mitochondrial genome in human adaptive radiation and disease: on the road to
therapeutics and performance enhancement. Gene, 2005. 354: p. 169-80.
80. Wallace, D.C., A mitochondrial paradigm of metabolic and degenerative diseases, aging, and can-
cer: a dawn for evolutionary medicine. Annu Rev Genet, 2005. 39: p. 359-407.
81. Dingley, S.D., et al., Mitochondrial DNA variant in COX1 subunit signicantly alters energy me-
tabolism of geographically divergent wild isolates in Caenorhabditis elegans. J Mol Biol, 2014.
426(11): p. 2199-216.
82. Wallace, D.C., Mitochondria as chi. Genetics, 2008. 179(2): p. 727-35.
83. Kenney, M.C., et al., Mitochondrial DNA variants mediate energy production and expression levels
for CFH, C3 and EFEMP1 genes: implications for age-related macular degeneration. PLOS ONE,
2013. 8(1): p. e54339.
84. Orr, H.A., Dobzhansky, Bateson, and the genetics of speciation. Genetics, 1996. 144(4): p. 1331-5.
85. Lane, N., Biodiversity: On the origin of bar codes. Nature, 2009. 462(7271): p. 272-4.
86. Bar-Yaacov, D., et al., Mitochondrial involvement in vertebrate speciation? The case of mito-nucle-
ar genetic divergence in chameleons. Genome Biol Evol, 2015. 7(12): p. 3322-36.
87. Gibson, J.D., et al., Genetic and developmental basis of F2 hybrid breakdown in Nasonia parasitoid
wasps. Evolution, 2013. 67(7): p. 2124-32.
88. Gershoni, M., A.R. Templeton, and D. Mishmar, Mitochondrial bioenergetics as a major motive
force of speciation. Bioessays, 2009. 31(6): p. 642-50.
89. Burton, R.S. and F.S. Barreto, A disproportionate role for mtDNA in Dobzhansky-Muller incompat-
ibilities? Mol Ecol, 2012. 21(20): p. 4942-57.
90. Willett, C.S. and R.S. Burton, Evolution of interacting proteins in the mitochondrial electron trans-
port system in a marine copepod. Mol Biol Evol, 2004. 21(3): p. 443-53.
91. Gebiola, M., et al., “Darwin’s corollary” and cytoplasmic incompatibility induced by Cardinium
may contribute to speciation in Encarsia wasps (Hymenoptera: Aphelinidae). Evolution, 2016.
70(11): p. 2447-58.
92. Chou, J.Y. and J.Y. Leu, Speciation through cytonuclear incompatibility: insights from yeast and
implications for higher eukaryotes. Bioessays, 2010. 32(5): p. 401-11.
93. Beck, E.A., et al., Gene ow between Drosophila yakuba and Drosophila santomea in subunit V of
cytochrome c oxidase: A potential case of cytonuclear cointrogression. Evolution, 2015. 69(8): p.
1973-86.
94. Cong, Q., et al., When COI barcodes deceive: complete genomes reveal introgression in hairstreaks.
Proc Biol Sci, 2017. 284(1848). p. 20161735.
95. Weckstein, J., et al., Anomalous variation in mitochondrial genomes of white-crowned (Zonotrichia
leucophrys) and golden-crowned (Z. atricapilla) sparrows: Pseudogenes, hybridization, or incom-
plete lineage sorting? The Auk 2001. 118: p. 231–36.
96. Toews, D.P., et al., Plumage genes and little else distinguish the genomes of hybridizing warblers.
Curr Biol, 2016. 26(17): p. 2313-8.
28 STOECKLE, THALER
97. Dabrowski, A., et al., Geographic variability in mitochondrial introgression among hybridizing
populations of golden-winged (Vermivora chrysoptera) and blue-winged (V. pinus) warblers. Con-
servation Genetics, 2005. 6: p. 843–853.
98. Brelsford, A., B. Milá, and D. Irwin, Hybrid origin of Audubon’s warbler. Mol Ecol, 2011. 20: p.
2380–89.
99. McGuire, J.A., et al., Mitochondrial introgression and incomplete lineage sorting through space
and time: phylogenetics of crotaphytid lizards. Evolution, 2007. 61(12): p. 2879-97.
100. Llopart, A., et al., Sequential adaptive introgression of the mitochondrial genome in Drosophila
yakuba and Drosophila santomea. Mol Ecol, 2014. 23(5): p. 1124-36.
101. von Beeren, C., et al., Interbreeding among deeply divergent mitochondrial lineages in the Ameri-
can cockroach (Periplaneta americana). Sci Rep, 2015. 5: p. 8297.
102. Hogner, S., et al., Deep sympatric mitochondrial divergence without reproductive isolation in the
common redstart Phoenicurus phoenicurus. Ecol Evol, 2012. 2(12): p. 2974-88.
103. Noor, M.A., et al., Chromosomal inversions and the reproductive isolation of species. Proc Natl
Acad Sci U S A, 2001. 98(21): p. 12084-8.
104. Navarro, A. and R. Faria, Pool and conquer: new tricks for (c)old problems. Mol Ecol, 2014.
23(7): p. 1653-5.
105. Hill, T., C. Schlotterer, and A.J. Betancourt, Hybrid dysgenesis in Drosophila simulans associated
with a rapid invasion of the P-element. PLOS Genet, 2016. 12(3): p. e1005920.
106. Darwin, C.A., The Origin of Species. 2nd ed. 1860 (reprinted 1958), New York: New American
Library. p. 448.
107. Latorre-Pellicer, A., et al., Mitochondrial and nuclear DNA matching shapes metabolism and
healthy ageing. Nature, 2016. 535(7613): p. 561-5.
108. Wallace, D.C., Genetics: Mitochondrial DNA in evolution and disease. Nature, 2016. 535(7613):
p. 498-500.
109. Kang, E., et al., Mitochondrial replacement in human oocytes carrying pathogenic mitochondrial
DNA mutations. Nature, 2016. 540: p. 270-7.
110. Shoubridge, E., Biomedicine: Replacing the cell’s power plants. Nature, 2016. 540: p. 210-1.
111. Templeton, A.R., The reality and importance of founder speciation in evolution. Bioessays, 2008.
30(5): p. 470-9.
112. Mende, M.B. and A.K. Hundsdoerfer, Mitochondrial lineage sorting in action--historical bioge-
ography of the Hyles euphorbiae complex (Sphingidae, Lepidoptera) in Italy. BMC Evol Biol,
2013. 13: p. 83.
113. Smith, J.M. and J. Haigh, The hitch-hiking effect of a favourable gene. Genet Res, 1974. 23(1):
p. 23-35.
114. Strobeck, C., J.M. Smith, and B. Charlesworth, The effects of hitchhiking on a gene for recombina-
tion. Genetics, 1976. 82(3): p. 547-58.
115. Oliveira, D.C., et al., Rapidly evolving mitochondrial genome and directional selection in mito-
chondrial genes in the parasitic wasp Nasonia (Hymenoptera: Pteromalidae). Mol Biol Evol,
2008. 25(10): p. 2167-80.
116. Avise, J.C., J.E. Neigel, and J. Arnold, Demographic inuences on mitochondrial DNA lineage
survivorship in animal populations. J Mol Evol, 1984. 20(2): p. 99-105.
117. Luria, S.E. and M. Delbruck, Mutations of bacteria from virus sensitivity to virus resistance. Ge-
netics, 1943. 28: p. 490-510.
118. Stewart, F., D.M. Gordon, and B.R. Levin, Fluctuation analysis: the probability distribution of the
number of mutants under different conditions. Genetics, 1990. 124: p. 175-85.
119. Tibayrenc, M., J.C. Avise, and F.J. Ayala, In the light of evolution IX: Clonal reproduction: Alter-
natives to sex. Proc Natl Acad Sci U S A, 2015. 112(29): p. 8824-6.
29WHY SHOULD MITOCHONDRIA DEFINE SPECIES?
120. Koch, R., Classics in infectious diseases. The etiology of tuberculosis: Robert Koch. Berlin, Ger-
many 1882. Rev Infect Dis, 1982. 4(6): p. 1270-4.
121. Stewart, F.M., Fluctuation analysis: the effect of plating efciency. Genetica, 1991. 84(1): p. 51-5.
122. Newcombe, H.B., Origin of bacterial variants. Nature, 1949. 164(4160): p. 150.
123. Lederberg, J. and E.M. Lederberg, Replica plating and indirect selection of bacterial mutants. J.
Bact., 1952. 63: p. 399-406.
124. Novick, A. and L. Szilard, Experiments with the chemostat on spontaneous mutations of bacteria.
Proc Natl Acad Sci U S A, 1950. 36(12): p. 708-19.
125. Maddamsetti, R., R.E. Lenski, and J.E. Barrick, Adaptation, clonal Interference, and frequen-
cy-dependent Interactions in a long-term evolution experiment with Escherichia coli. Genetics,
2015. 200(2): p. 619-31.
126. Grenfell, B.T., et al., Unifying the epidemiological and evolutionary dynamics of pathogens. Sci-
ence, 2004. 303(5656): p. 327-32.
127. Weir, W., et al., Population genomics reveals the origin and asexual evolution of human infective
trypanosomes. Elife, 2016. 5: p. e11473.
128. Tibayrenc, M. and F.J. Ayala, Reproductive clonality in protozoan pathogens--truth or artifact? A
comment on Ramirez and Llewellyn. Mol Ecol, 2015. 24(23): p. 5778-81.
129. Tomasini, N., et al., How often do they have sex? A comparative analysis of the population struc-
ture of seven eukaryotic microbial pathogens. PLOS ONE, 2014. 9(7): p. e103131.
130. Tibayrenc, M. and F.J. Ayala, How clonal are Neisseria species? The epidemic clonality model
revisited. Proc Natl Acad Sci U S A, 2015. 112(29): p. 8909-13.
131. Diaz Caballero, J., et al., Selective sweeps and parallel pathoadaptation drive Pseudomonas aeru-
ginosa evolution in the cystic brosis lung. MBio, 2015. 6(5): p. e00981-15.
132. Strelkowa, N. and M. Lassig, Clonal interference in the evolution of inuenza. Genetics, 2012.
192(2): p. 671-82.
133. Baym, M., et al., Spatiotemporal microbial evolution on antibiotic landscapes. Science, 2016.
353(6304): p. 1147-51.
134. Burnet, F.M., Immunological recognition of self. Science, 1961. 133(3449): p. 307-11.
135. Lederberg, J., Genes and antibodies. Science, 1959. 129(3364): p. 1649-53.
136. Provine, W.B., Ernst Mayr: Genetics and speciation. Genetics, 2004. 167(3): p. 1041-6.
137. Eldredge, N. and S.J. Gould, Punctuated equilibria: an alternative to phyletic gradualism, in
Models in Paleobiology, T.J.M. Schopf, Editor. 1972, Freeman, Cooper and Co.: San Francisco.
p. 82-115.
138. Straub, T.J. and O. Zhaxybayeva, A null model for microbial diversication. Proc Natl Acad Sci U
S A, 2017. 114(27): p. E5414-23.
139. Birky, C.W., Jr. and J.B. Walsh, Effects of linkage on rates of molecular evolution. Proc Natl Acad
Sci U S A, 1988. 85(17): p. 6414-8.
140. Bazin, E., S. Glemin, and N. Galtier, Population size does not inuence mitochondrial genetic
diversity in animals. Science, 2006. 312(5773): p. 570-2.
141. Goodall-Copestake, W., et al., Swarms of diversity at the gene cox1 in Antarctic krill. Heredity
(Edinb), 2010. 104: p. 513-8.
142. 1000 Genomes Project Consortium, A global reference for human genetic variation. Nature, 2015.
526(7571): p. 68-74.
143. Qiu, J., How China is rewriting the book on human origins. Nature, 2016. 535: p. 22-5.
144. Smith, F. and J. Ahern, eds. The Origins of Modern Humans: Biology Reconsidered. 2013, Wiley:
New Jersey. p. 454.
145. Cann, R.L., M. Stoneking, and A.C. Wilson, Mitochondrial DNA and human evolution. Nature,
1987. 325(6099): p. 31-6.
30 STOECKLE, THALER
146. Cann, R., All about mitochondrial Eve: an interview with Rebecca Cann. Interview by Jane
Gitschier. PLOS Genet, 2010. 6(5): p. e1000959.
147. Gibbons, A., Y chromosome shows that Adam was African. Science, 1997. 278(5339): p. 804-5.
148. Nosil, P., et al., Tipping points in the dynamics of speciation. Nature Ecol Evol, 2017. 1: p. 1-8.
149. Sudmant, P.H., et al., An integrated map of structural variation in 2,504 human genomes. Nature,
2015. 526(7571): p. 75-81.
150. Templeton, A.R., The theory of speciation via the founder principle. Genetics, 1980. 94(4): p.
1011-38.
151. Nei, M., Bottlenecks, genetic polymorphism and speciation. Genetics, 2005. 170(1): p. 1-4.
152. Johnson, T., Benecial mutations, hitchhiking and the evolution of mutation rates in sexual popu-
lations. Genetics, 1999. 151(4): p. 1621-31.
153. Hernandez, R.D., et al., Classic selective sweeps were rare in recent human evolution. Science,
2011. 331(6019): p. 920-4.
154. Campos, J.L., et al., The relation between recombination rate and patterns of molecular evolution
and variation in Drosophila melanogaster. Mol Biol Evol, 2014. 31(4): p. 1010-28.
155. Zuckerkandl, E. and L. Pauling, Molecules as documents of evolutionary history. J Theor Biol,
1965. 8(2): p. 357-66.
156. Lee, M.S. and S.Y. Ho, Molecular clocks. Curr Biol, 2016. 26(10): p. R399-402.
157. Vrba, E.S. and D. DeGusta, Do species populations really start small? New perspectives from
the Late Neogene fossil record of African mammals. Philos Trans R Soc Lond B Biol Sci, 2004.
359(1442): p. 285-92; discussion 292-3.
158. Grant, W.S., Problems and cautions with sequence mismatch analysis and Bayesian skyline plots
to infer historical demography. J Hered, 2015. 106(4): p. 333-46.
159. Thompson, J., Relentless Evolution 2013: University of Chicago Press.
160. Horvitz, H.R., Nobel lecture. Worms, life and death. Biosci Rep, 2003. 23(5-6): p. 239-303.
161. Mitchell, P., Chemiosmotic coupling in oxidative and photosynthetic phosphorylation. 1966. Bio-
chim Biophys Acta, 2011. 1807(12): p. 1507-38.
162. Pagliarini, D.J. and J. Rutter, Hallmarks of a new era in mitochondrial biochemistry. Genes Dev,
2013. 27(24): p. 2615-27.
