Download - Philsci

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Metagenomics wikipedia , lookup

Horizontal gene transfer wikipedia , lookup

Transcript
The Concept and Causes of Microbial Species
John S. Wilkins
Biohumanities Project, University of Queensland, Brisbane, Australia 4072. Email:
[email protected]
In … an asexual group, systematic classification would not be impossible,
for groups of related forms would exist which had arisen by divergence
from a common ancestor. Species, properly speaking, we could hardly
expect to find, for each individual genotype would have an equal right to
be regarded as specifically distinct, and no natural groups would exist
bound together like species by a constant interchange of their germ-plasm.
The groups most nearly corresponding to species would be those adapted
to fill so similar a place in nature that any one individual could replace
another. (Fisher 1958:135)
Abstract
Species concepts for bacteria and other microbes are contentious, because
they are often asexual. There is a Problem of Homogeneity: every
mutation in an asexual lineage forms a new strain, of which all
descendents are clones until a new mutation occurs. We should expect
that asexual organisms would form a smear or continuum. What causes
the internal homogeneity of asexual lineages, if they are in fact
homogeneous? Is there a natural “species concept” for “microbes”? Two
main concepts devised for metazoans and metaphytes have been applied
to bacteria. One is the Recombination Concept, a revised form of the
Biological Species Concept in which the homogenizing mechanism is the
sharing of genome fragments, somewhat akin to sexual recombination.
The other is the Ecological Species Concept, in which the ecological
niche is that which maintains lineages as cohesive. In this paper I will
discuss these two concepts, and offer an underlying model that conjoins
them, and consider the implications for species concepts in general. In
short, my argument is that asexual species are instances of the most
primitive and underived notion of species, which I will call
“quasispecies”, following Eigen, and that sexual species are merely one
derived kind of species. Moreover, I will argue that there is a continuum
of recombination from simple viral models in which each strain is a clone,
through to obligate recombination of 50% of the parents’ genome, and
that consequently there is no sharp division between “microbial” and
more familiar species.
Keywords: species concept, microbes, later genetic transfer, quasispecies
The Problem of Homogeneity
There is a longstanding debate over whether microbial species form species taxa. This
is predominantly due to operational difficulties of identification, but also because
microbial organisms usually lack sexual recombination that is considered by many to
be crucial for the evolution of reproductive isolation, which is according to the
dominant view of what species are, the defining character of being a species. What is
a microbial species, either a bacterial or other mostly asexual single celled species?
Let us begin by noting that this is a perpetual problem in microbiology. In 1962,
microbiologist G. D. Floodgate noted the following points about the state of thinking
about bacteria at the time. Arguably microbiologists are in much the same straits now
as then.
Of all the words in the taxonomist’s vocabulary, probably the most
difficult to define is “species.” The perplexities which bacteriologists
encounter in trying to give the word an exact meaning are well known … .
A species has been described as more real or as having “a greater degree
of objectivity” than any other taxon …, but it has also been called a manmade fiction … . Again, according to one author … a species is a
dynamic system, but according to another it is as outdated as phlogiston
… . Further, a meaning given to “species” in one biological discipline
may not be used at all in another. For example, the bacteriologist does not
usually refer to interbreeding or exchange of genetical material when
describing the meaning of species as other biologists sometimes do,
though recently an attempt has been made to introduce species in this
sense in bacteriology as well …. Yet again, some bacteriologists have
considered a species to be a discrete segment of a phyletic line evolving
independently of other segments … . (Floodgate 1962)
To address this issue, I want to pose a philosophical problem I shall call the Problem
of Homogeneity, or, Why Are There Microbial Species and Not Just A Mess of
Strains?
First, we should note that the literature on microbial species refers to a number of
groupings of organisms: prokaryotes, protists, monerans, bacteria, and so on. There
appears to be no single term to cover them all, so I will here adopt “microbe” as a
convenient label, following O’Malley and Dupré (2007), although it must not be
thought that this implies microbes are a natural group. Few of these groups appear to
be “natural” (in the cladistic sense of monophyletic). Almost all of them are defined
as some complement of Life after other groups, eukaryotes, animals or metazoans,
plants or metaphytes, and fungi, have been taken out. In other words, “microbe” is a
paraphyletic or polyphyletic grouping. Discussions of microbial species have often
focussed on the operational aspects of diagnosis. However, what we want is not an
operational convention to delimit microbial species, but a theoretical, or ontological,
account of what kinds of things microbial species could be. Several options have been
proposed:

Microbial species could be formed from an analogue of sexual recombination, via
lateral transfer between strains, or by other mechanisms of genetic exchange that
cause a “core genome” to be maintained which prevents some lateral transfer from
being successful, much in the same way that reproductive isolation acts to prevent
gene exchange.

Microbial species could be simple ecotypes, tracking the fitness peaks of hosts, in
the case of parasites, pathogens and single host exploiters.

Microbial species could be a disparate and ultimately illusory conventional
category with no underlying ontological category that applies to each and every
case. If this is true, then we should expect to find no principled cause for
microbial species, and each case would have a different account. I will seek to
show this is not the case.
We need to account for the cohesion and clustering of genotypes and phenotypes of
microbes. This gives us a “species phenomenon” (Brigandt 2003) to investigate, and
is the explanandum of microbial species. Operational criteria are useful only so far as
they are reliable markers of underlying biological realities. So we will ignore
operational issues here.
The problem facing microbiologists is that usually in “macrobial” biology a species is
defined as a group of related organisms that share genes (or is a gene pool, which
amounts to the same thing), the so-called Biological Species Concept. Sometimes
ecological considerations are also included (either in the form of shared natural
selection, or in terms of sharing an adaptive niche). But many bacterial and archaeal
species either do not exchange genes to reproduce, or they can but do not need to; that
is, they are either facultative gene exchangers or obligate non-exchangers. So the
question is sometimes raised whether these organisms form species at all, or if there is
some replacement taxon term or concept for groups of them. And while many recent
writers, such as Templeton (1998) or Coyne and Orr (2004) now, rather grudgingly,
accept the reality of asexual species, others such as Meier and Willmann (2000),
continue to deny species rank to asexual taxa, instead calling them “agamotaxa”.1
Even so, bacteriologists continued during the Synthesis and after to name and
describe species, even though they could not really make use of the Biological
Species Concept of Dobzhansky and Mayr. They relied, among other characters, on
their organism’s staining properties, the colony shape, the microscopic morphology of
the cells, and of course the ecological conditions under which they lived. What was
lacking, though, was a clear definition of what a species could be for these organisms
(Moreno 1997; Cohan 2001, 2002; Istock et al. 1996; Tibayrenc 2006).
Part of the difficulty is that if a species were obligately clonal, then each mutation
should make a new clonal lineage, a strain, and we might naively expect to find not
genetic clusters but a carpet of strains more or less evenly distributed (figure 1). How
can we account for this? I will call this the Problem of Homogeneity: why are asexual
lineages ever found as groups at all? Why are they homogeneous over time, and stable
enough to be called “species” (Hanage, Fraser, and Spratt 2006)?
There are two relevant conceptions of what species are for macrobial species, which
is to say, species of plants and animals.2 One is the suite of definitions that rely upon
reproductive isolation, such as the Biological Species Concept, the Recognition
Concept (Paterson 1985), the Genotypic Cluster Concept (Mallet 1995), and
Templeton’s Cohesion Concept (Templeton 1989). The last is interesting because it
combines both genetic (reproductive cohesion) and ecological similarity. These might
be grouped together as Reproductive Isolation Conceptions. The other set of
conceptions relies directly on ecological roles (Hutchinson 1968; Ghiselin 1974; Van
Valen 1976; Sterelny 1999). Although ecological species conceptions are rarely
proposed as a sufficient account of sexual species, they are invoked to explain why
1
Or binoms (Grant 1957), or agamospecies (Cain 1954).
2
That there are basically three conceptions of “species” is argued by Hull (1997).
sexual organisms that can freely interbreed between species remain constant as
species. Another set of conceptions, which rely on evolutionary or phylogenetic
differentiation, are not important in this context, as they merely restate what is our
problem here, to wit, that clusters of lineages form over evolutionary time. We require
an account of why they cluster before we can apply historical conceptions such as
phylospecies or evospecies.
Figure 1: The “carpet model” of asexual strains. This is a two-dimensional
representation of a high-dimensional genome space. See Gavrilets (2004) and Skipper
(2004) for a discussion of the concept and representation of a genome spaces and
fitness spaces. Although the distribution shown here is even, random strains are likely
to form a Poisson distribution from the initial genome coordinate.
Lateral and Horizontal Transfer
A further complication of the topic is the increasing recognition that microbial
organisms do share their genetic material often and in evolutionarily important ways.3
3
(Kunin et al. 2005; Beiko, Harlow, and Ragan 2005; Makarenkov and Legendre 2004; Raymond and
Blankenship 2003; Kurland, Canback, and Berg 2003; Wolf et al. 2002; Gogarten, Doolittle, and
Lawrence 2002; Ragan 2001; Koonin, Makarova, and Aravind 2001; Ochman, Lawrence, and
Groisman 2000; Woese 1998; Syvanen and Kado 1998)
“Lateral” transfer usually refers to the transfer of individual genetic components
across distant phylogenetic groups, while “horizontal” transfer often refers to the
interbreeding of genomes between species, but the basic problem is the same: how
can we define species (in microbes) if their genetic lineages do not generally match
their expected phylogeny? Is there even any sense in calling these phylogenetic
lineages?
To address these questions, I will appeal to the notion of a genomic space, which is
effectively the adaptive landscape idea of Sewall Wright (see Pigliucci and Kaplan
2006 for a treatment of this; a more technical account is given in Gavrilets 2004). A
genomic space is an n-dimensional space of all possible genomes, and each
coordinate in that space represents one genome. Our explanandum is that instead of
finding genomes more or less evenly smearing over the possible genome space, we
find that microbial species, like viral species, cluster, which is to say, they form
homogeneous groups.
Is the Problem of Homogeneity an Illusion?
We must first set up expectations here: could clustering be the result of simple
stochasticity of mutation and extinction? Simulations give reasons for thinking that
random processes will not generate the sort of clustering actually observed (described,
e.g., in Konstantinidis, Ramette, and Tiedje 2006). Further, they indicate that logistic
diversification and phylogenetic entrenchment will give this sort of distribution.
It might be thought that clustering in genome space doesn’t need explanation.
Clustering of phylotypes might in the first case simply be due to stochastic processes
of elimination and divergence. To reject that assumption, I will appeal here to a recent
paper (Pie and Weitz 2005) on null models in morphospace. Given that the
morphological traits here are heritable, and that they are in effect asexually
transmitted (because the morphology here is typical of a species that diverges in a
branching random walk), it is applicable to our case. This is an argument by
simulation, and lacks as yet any empirical foundation, but as we have no clear prior
assumptions about what the default expectations for asexually reproducing lineages
may be, this may help to give us some. Further simulations might change these
expectations. The authors understood their simulation to apply in the case of asexual
lineages (Pie, pers. comm.).
Pie and Weitz simulated speciation in four cases. In the first case, they allowed
speciation without extinction, which generated the expected carpet model. In the
second case, they allowed stochastic extinction, and the surprising result here is that
the clustering, while tighter, still spreads out evenly. In the third case, however, they
allowed logistic diversification, equivalent to “adaptive niche packing” according to
carrying capacity, and now we see the scatter of clusters that we do find in genotypic
space in asexual lineages. A fourth case allowed developmental entrenchment, in
which the longer the trait (in our case, gene sequence) has been in the lineage, the less
likely it is to go extinct, and the clustering is even more diverse and tightly packed,
although mean separation decreased in this case after an initial increase. Since
speciation without extinction and with stochastic extinction are both stochastic
processes, while the other two cases answer to selection for niche occupancy and
internal cohesion respectively, it appears that simple stochasticity does not explain
clustering, and we must consider the other two cases separately. Other realistic factors
are not included here, such as differential extinction due to habitat fracturing, which
may still form clusters in a stochastic manner. Extinction can cause there to be
patches in genome space. Some varieties die out. While this can be because of
selection, often it will be due to plain old genetic drift, and contingency. Moreover, if
a genome evolves in a habitat that is sensitive to sudden changes in climate or even
geological change that degrades it, then that genome and its neighbors will become
extinct.
Even so, Pie and Weitz’s model gives us an initial expectation that when patchiness
occurs over large numbers of genome lineages, it will be due either to selection or to
the phylogenetic integration of certain genes in the developmental process (their
“developmental entrenchment”).
Sex versus Ecology as a Solution to Homogeneity
Sex
One solution to the Problem of Homogeneity for asexual organisms is what I will call
the Recombination Species Concept. Proposed by Dykhuizen and Green (1991), it
basically the Biological Species Concept for lineages that occasionally share genes or
gene fragments through lateral transfer:
“the phylogenies of different genes from individuals of the same species
should be significantly different, whereas the phylogeny of genes from
individuals of different species should not be significantly different. Thus
we have an operational criterion for the defining of bacterial species”
(Dykhuizen and Green 1991:7266)
“Horizontal gene transfer across species should be rare enough that it will
not be a problem.” (loc. cit p7267)
There are several mechanisms by which lateral transfer of DNA can occur among
“prokaryotes”. One is DNA fragment reuptake, in which DNA from a cell that has
lysed (its membrane or wall has disintegrated, releasing the cell contents into the
medium) is taken up by another cell, and rather than being digested it becomes active.
There are variations on this. Entire small chromosomal rings, called plasmids, can be
taken up this way. Or a cell can “bleb”, forming vesicles, or compartments, of lipids,
containing DNA (including plasmids), which then attach to the receiving cell, opening
up to the interior. A process that is directly analogous to sex in bacteria is
conjugation. This is a case where part of the genetic component, usually plasmids,
which are secondary small chromosomes, or the main nucleoid, can be inserted into
another cell via processes called pili, which are part of the Type IV secretory system
used for other purposes and which is homologous to flagella in Gram-negative
bacteria. The typical mode of conjugation is that one mating type (often called the
“male”) is activated by pheromones from another mating type (“female”) to attach the
pilus to the recipient, and insert the genetic material. This is “almost-sex”, because
there is no genetic reassortment, but other processes will tend to shuffle genes into the
nucleoid, as well as utilizing the taken-up plasmids. But again, while the mating types
seem to act as cohesive mechanisms, conjugation can be profligate across large, even
vast, phylogenetic distances (and hence genetic distances). It has been observed
between bacteria and yeast, bacteria and plants, and there has even been a case in
which it was observed between bacteria (E. coli) and mammalian (hamster) cells
(Waters 2001), although there is no evidence this caused any evolutionary outcome.
Some microbes can have multiple mating types at different parts of the developmental
cycle. Plasmodium falciparum, the pathogen that causes malaria in humans, for
instance, has seven, and many of the Plasmodia have more.
The Recombination model of microbial species is based on the claim that the greater
the genetic distance between strains, the less likely it is that the transferred genes will
be functional and useful in the receiving strain, and this is what serves to maintain the
homogeneity of bacterial and other microbial species (the phylogenetic entrenchment
of Pie and Weitz). One of the claims of Dykhuizen and Green is in fact that the
differences in genetic structure, and in some cases differences in restriction enzymes,
that break DNA sequences, will make the DNA fragment non-coding, or insert it in a
nonfunctional location in the target genome. This is sometimes referred to as the Core
Genome Hypothesis or Genomic Island Hypothesis (Wertz et al. 2003; Coleman et al.
2006) – what makes laterally transferred genes functional are the core “housekeeping
genes” that do not get transferred because they are critical to the functioning and
viability of the organism, and are closely entrenched in that organism’s genome. This
means that they will likely not be useful in distantly related organisms with their own
entrenchments.
These compatibility issues act in the same way that sex does in sexual species to
maintain the overall “location” in genome space of the population (figure 2). Those
strains that deviate too far from the mode will be unable to take up the functionally
useful lateral genes, and so will be more susceptible to extinction through genetic
load. Hence, on the Recombination Model, compatibility of gene exchange will
generate patchiness in gene space.
Figure 2: In the Recombination Model, functional lateral or horizontal transfer (thick
horizontal lines) maintains the cohesion of the cluster of genomes.
So the Recombination Model is a mix of Maynard Smith’s theory of sex, and Mayr’s
notion of biological species. The problem is that it isn’t consistently true.
Recombination via lateral transfer is rather more profligate than it first seemed
(Beiko, Harlow, and Ragan 2005), while there are some “species” of bacteria, such as
the Lyme Disease-causing spirochete Borrelia burgdorferi, that do not share
autapomorphic genes much, if at all (as Dykhuizen himself observes: Dykhuizen and
Baranton 2001). So if in those latter cases clustering occurs, it is not due to lateral
transfer, but some other processes. While it may be that recombination of lineages
through partial lateral genetic transfer operates as a reason for some phylotypes
occurring, it does not account for all, and LGT is therefore not a sine qua non of
specieshood, or homogeneity, among all microbes.
Ecology
An alternative explanation for the coherence of the genome cluster of asexuals is that
they are, in Templeton’s words (1989), “demographically replaceable”. That is, they
tend to occupy the same fundamental ecological niche. Cohan, for instance (Cohan
2002) has proposed a “stable ecotype” conception of bacterial species. This would be
a sufficient explanation, were it not for the fact that “species within species” in
bacteria and other organisms, that is, clusters that fall within other clusters of
genomes, can occupy quite divergent niches or have different phenotypes; e.g.,
Burkholderia mallei, which causes glanders, and B. pseudomallei, which causes
melioidosis (Hanage et al. 2006), though the former is a genomic strain of the latter.
Ecological adaptation alone fails to account for the clustering, although geographical
isolation and local adaptation could. For instance, a pathogen that switches host
species or cell types may fall within the overall cluster so long as the ecologically
relevant genes do not have to vary too far from the region which the ancestral species
occupies in genome space. Hence, there is no unique genome cluster for some
differing ecotypes.
In the case of a hypothetical purely asexual species, which lacks any lateral or
horizontal transfer, ecological adaptation may suffice to explain the coherence of the
cluster. And if the species are sympatric (that is, they share the same host or
ecological environment), divergent selection could be expected to shift the location of
the strain’s genome, and thus break the homogeneity of the cluster (Cohan 2001). In
the case of an organism that is competent to occupy many such niches, however,
coherence cannot be explained simply in terms of ecological adaptation.
So, there are two main mechanisms that might be applied to microbial species, neither
of which quite fits the all the facts. The problem lies, I believe, in the attempt to make
only one of them a universal conception of species. To address this, we must
reconsider what it is we are trying to explain; what the phenomena are.
Quasispecies and species conceptions
The “species phenomenon” (Brigandt 2003) here is the clustering of genomes in a
state space akin to the adaptive landscape model of genetics. A genomic cluster of this
kind in purely asexual cases is called a “quasispecies”(devised to deal with virus
species, Eigen 1993), which is a definite cloud of genomes that cluster about a “wild
type”. A phenotypic version would be the phenetic “operational taxonomic unit”, but
we will discuss only genotypic quasispecies. According to the concept as developed
by Manfred Eigen4 for viral species, a quasispecies (“as-if-species”) is a cluster of
4
Eigen and those who have followed him have presented a hypercycle model of self-organisation for
why there are quasispecies. However, while this is consonant with the view I present here, what is
genomes in a genome space of the dimensionality of the number of loci. He defined
them as a
… region in sequence space [which] can be visualized as a cloud with a
center of gravity at the sequence from which all mutations arose. It is a
self-sustaining population of sequences that reproduce themselves
imperfectly but well enough to retain a collective identity over time.
(Eigen 1993:35)
A quasipecies is in effect a cluster of genomes (figure 3), with a “wild-type”
coordinate (that is, genome) that may or may not actually have an extant or extinct
instance. The cluster is measured by the modulus of the Cartesian, or Hamming,
distance from the wild-type.
Figure 3: A quasispecies is a cluster of genomes in a space determined by the number
of loci in the genomes, and the alleles at those loci, about a “wild-type” mode.
A genetic cluster of the quasispecies kind is therefore not unlike James Mallet’s
Genotypic Cluster Concept (Mallet 1995), although his is also primarily an
operational definition than a substantive underlying account of species, relying on
diagnosis. However, we still need to account for microbial quasispecies existing in the
first place. We have considered one possible mechanism – gene sharing by lateral
important at this point is the phenomenal nature of quasispecies’ genomes – that they are clusters of
genotypes in genome-space – not that they are caused in any manner offered by quasispecies
researchers.
transfer – and found it to be insufficient. Is there something else we might make use
of? There is one proposal by Alan Templeton (Templeton 1989), devised for sexual
organisms, which defines a species as a genetic cluster, as Mallet’s does, but accounts
for it both by genetic exchange, as in the recombination model, and ecological
(“demographic”) interchangeability.
For any cluster of genomes, each coordinate in genome space, that is to say each
genome, has a fitness value associated with it that is imposed by ecological factors. If
the adaptive landscape is relatively smooth, which means that adjacent coordinates
are correlated in their fitness values, we should expect in the absence of all other
causes of clustering that the cloud of genomes will tend to centre upon the most
adaptive genome (figure 4). Of course, this is an abstraction and a gross simplification
– genomes are not independent of each other, or from fitness values. Organisms create
their ecological conditions to a degree, and how fit a genome is depends also upon
what other genomes exist in a population (at least, in sexual organisms), but we can
leave these complications to one side for the moment.
Figure 4: Clustering could be explained in terms of tracking fitness peaks –
deviations from the local adaptive peak would be eliminated by selection.
So one potential reason why quasispecies exist, why genomes cluster, is that they
track local fitness peaks. Take a pathogen that is clonal. It needs to employ certain
features of the host species in order to infect and exploit that host. Assuming these
features remain constant – say they are recognition molecules on a cell surface – the
quasispecies will cluster about those points in the genome space that are more
effective than others with respect to the capacity to infect and exploit.
Now we have the two cohesion mechanisms proposed by Templeton – cohesion due
to shared genes, and cohesion due to the need to exploit the environment better than
competitors. Anything that can do well at the latter will tend to be better represented
in the average population. Hence quasispecies. But fitness peaks typically do not
remain constant or decoupled from the populational structure. Does this mean that
quasispecies are not real species because they are ephemeral? Of course not, as all
species are, over a suitably extended timescale, ephemeral. That is in the nature of
evolution. What the fitness peak conception means is that a quasispecies will remain
homogeneous so long as there is a more or less unitary fitness peak. If the peaks shift,
there will be “speciation” or, stopping the pretense that quasispecies aren’t real
species, there will be real speciation.
Prototype Species Concepts
We have problems when we attempt to apply to organisms that are not obligately
sexual (that is, which don’t have to have sex to reproduce) concepts that were
specified to use with those that are obligately sexual. The Recombination Model is
one such attempt. It is true that some microbial species exchange genes. Others do it
more frequently and more completely. However, there appears to be a continuum of
gene exchange all the way from almost never to almost every generation. So why
should we expect that gene transfer will provide us with the sort of homogeneity of
lineages and quasispecies that it does in obligate sexual species?
In part this is because we always start from what we know. The existence and
ubiquity of asexual organisms has been resisted by macrobial biologists for a long
time, and treated as exceptional rather than the rule, because biology began with large
scale plants and animals where the paradigm cases were encountered for the
biospecies concept. As exceptional cases were encountered, these concepts were
stretched, and modified, to serve the increasingly “deviant” cases, until now we
realize that deviance is relative to the exemplary conceptions rather than a fact about
the organisms. Instead, we should take seriously the idea that genetic exchange forms
a continuum, as the basis from which the metazoan and metaphyte conceptions are
drawn out. It is worth considering how the two factors of the Templeton conception –
exchange and ecological niche – play differing roles according to the degree of sex a
lineage has.
Figure 5: The contribution of the homogeneity of a microbial species of reproductive
cohesion increases with the percentage of gene exchange between lineages, while the
proportion of the contribution of niche tracking will decline.
So, in respect of the notion that sex is not an all-or-nothing affair, if the continuum
ranges from 0% gene exchange (total asexuality) to 50% gene exchange (total
sexuality), then it follows that at the asexual end, exchange can have only limited to
no effect on maintaining homogeneity, while at the sexual end, it can have a very
great role in maintaining homogeneity (figure 5). Likewise, at the asexual end,
homogeneity not due to stochastic effects or geographical fracturing will be due
largely to ecological selection (fitness peak tracking), while for sexuals this will not
be so important a cause. For sexual organisms at the upper end, reproductive cohesion
permits the taxon to track many ecological fitness peaks when the reproductive
cohesion is sufficient to overcome the divergent selection of multiple fitness peaks
(figure 6). However, in cases where so-called sympatric speciation is possible, set out
by Gavrilets, this will not prevent speciation.
Figure 6: Pure asexual species are maintained by niche selection, while pure sexuals
are maintained also by reproductive cohesion, allowing multiple fitness peaks to be
tracked to a degree.
This plurality of cluster causation is why gene exchange doesn’t have a simple
relationship to genomic homogeneity. It will depend on the rate and amount of genes
shared across lineages case by case, as well as the degree to which the quasispecies is
maintained by ecological interchangeability. Following on from a discussion about
speciation (Wilkins 2007), and an earlier paper on species concepts (Wilkins 2003),
where I employed the phylogeny of sex to argue that being a biospecies is an evolved
trait, not a natural kind, we ought to expect that each species and group of species has
its own unique evolutionary history and therefore properties, just as limbs and lungs
and livers do. In effect, the modality of “being a species” is an evolved property.
There will be a more general theoretical context of adaptive landscapes, gene
dynamics, and so on, but given that each evolutionary group has encountered different
conditions, this means that each modality will be shared in a fairly limited way.
Templeton’s “Genetic Cohesion” concept for sexual species involves both kinds of
mechanisms – cohesion mechanisms of genetic and reproductive compatibility and
isolation from other groups, and ecological or “demographic” exchangeability, in
which organisms play the same general role ecologically. The view of Templeton is
that species are conjunctively defined by being genetic clusters, any member of which
is “demographically replaceable” and “genetically exchangeable”; i.e., they can
exchange genes, and do the same thing ecologically. But this conjunction cannot be
true for facultatively sexual and obligately asexual species, which leads us to a
disjunct for those properties.
Microbial species therefore are maintained either by ecological selection or by
successful genetic exchange, or both. This leads to a view of the role of ecological
selection in the absence of genetic exchange as the sole cause of microbial species,
but that in the case of increasing genetic exchange, the importance of local adaptation
declines as a factor in the quasispecies identity.
To return to the operational issue once more, a reviewer asks whether this will assist
us in identifying the ecological niches for microbial (and, mutatis mutandis,
macrobial) species. In medical cases like the Burkholderia bacteria mentioned above,
the ecological niche is the host cells, infection of which cause the disease. Hence the
disease is a marker for a distinct ecological niche. We are not always so “fortunate” as
to have these markers, or for them to be unambiguous. In mostly vertical genetic
transfer quasispecies there will on this account be an ecological niche, and so we
might expect that the trophic role they play when discovered is indeed that niche
(although the entire lifecycle will need to be followed through). However, in mixed
modalities, this is going to be increasingly harder the more genetic material is, on
average, exchanged per generation, as it will mean that more adaptive peaks can be
tracked. All I can say here is that if they are almost entirely asexual, they cluster
because they have a single ecotype.
The reviewer also asks if species, being the phenomenological outputs of these two
processes of selection and exchange, are also inputs into biological processes.
Certainly, those who allow for species selection believe species are in fact causal
actors. I do not. In each case, unless the species is coterminous with a single deme,
the causal activity is at the population level at most. I do not think, although I cannot
argue for it here, that species are either theoretical objects of biology in the sense of
elements of a theory or required by an ensemble of theories, or that they are the
explanans of any other explanandum in biology apart from the classificatory activities
of taxonomists and philosophers of biology. Both microbial and macrobial species are
phenomena of genomic clusters and ecological inhabitants.
Conclusion
The basic notion of a phenomenal species, in microbes or in microbes, is a genetic
cluster in genome space, called a “quasispecies”. This is the explanandum for which
ecological or reproductive explanations are given. The role of reproductive
compatibility, leading to isolation from other groups, applies increasingly as the
amount of genetic material exchanged increases, and its frequency. To account for
clusters in cases of low genetic exchange, whether by lateral transfer or sexual
recombination, ecological selection has to be the other fork of the disjunct.
“Microbial” species form a continuum of gene exchange from 0% to a possible 50%
in the case of obligately sexual microbes, and as that percentage drops, the
explanation for the cohesion of homogeneous quasispecies clusters increasingly
requires ecological niche adaptation, in the absence of stochastic explanations. As
gene exchange approaches 50% exchange, the explanation must also include some
role for the coadaptation of mating types and the lack of fitness of interspecific
hybrids. All species are primarily adaptive quasispecies, and sexual species
secondarily, as it were.
References
Beiko, Robert G., Timothy J. Harlow, and Mark A. Ragan. 2005. Highways of gene
sharing in prokaryotes. Proceedings of the National Academy of Sciences,
USA 102 (40):14332-14337.
Brigandt, I. 2003. Species pluralism does not imply species eliminativism. Philosophy
of Science 70 (5):1305-1316.
Cain, Arthur J. 1954. Animal species and their evolution. London: Hutchinson
University Library.
Cohan, Frederick M. 2001. Bacterial species and speciation. Systematic Biology 50
(4):513-524.
———. 2002. What are bacterial species? Annu Rev Microbiol 56:457-487.
Coleman, Maureen L., Matthew B. Sullivan, Adam C. Martiny, Claudia Steglich,
Kerrie Barry, Edward F. DeLong, and Sallie W. Chisholm. 2006. Genomic
Islands and the Ecology and Evolution of Prochlorococcus. Science 311
(5768):1768 - 1770.
Coyne, Jerry A., and H. Allen Orr. 2004. Speciation. Sunderland, Mass.: Sinauer
Associates.
Dykhuizen, D. E., and G. Baranton. 2001. The implications of a low rate of horizontal
transfer in Borrelia. Trends in Microbiology 9 (7):344-350.
Dykhuizen, D. E., and L. Green. 1991. Recombination in Escherichia coli and the
definition of biological species. Journal of Bacteriology 173 (22):7257-7268.
Eigen, Manfred. 1993. Viral quasispecies. Scientific American July 1993 (32-39).
Fisher, Ronald Aylmer. 1958. The genetical theory of natural selection. 2nd rev. ed.
New York: Dover.
Floodgate, G. D. 1962. Some remarks on the theoretical aspects of bacterial
taxonomy. Bacteriological Review 26:277-291.
Gavrilets, Sergey. 2004. Fitness landscapes and the origin of species, Monographs in
population biology; v. 41. Princeton, N.J.; Oxford, England: Princeton
University Press.
Ghiselin, Michael T. 1974. A radical solution to the species problem. Systematic
Zoology 23:536-544.
Gogarten, J. Peter, W. Ford Doolittle, and Jeffrey G. Lawrence. 2002. Prokaryotic
Evolution in Light of Gene Transfer. Mol. Biol. Evol. 19 (12):2226-2238.
Grant, Verne. 1957. The Plant Species in Theory and Practice. In The Species
Problem, edited by E. Mayr. Washington DC: American Association for the
Advancement of Science.
Hanage, W., C. Fraser, and B. Spratt. 2006. Sequences, sequence clusters and
bacterial species. Philosophical Transactions of the Royal Society B:
Biological Sciences 361 (1475):1917-1927.
Hanage, W., B. Spratt, K. Turner, and C. Fraser. 2006. Modelling bacterial speciation.
Philosophical Transactions of the Royal Society B: Biological Sciences 361
(1475):2039-2044.
Hull, David L. 1997. The ideal species concept – and why we can’t get it. In Species:
The units of diversity, edited by M. F. Claridge, H. A. Dawah and M. R.
Wilson. London: Chapman and Hall.
Hutchinson, G. E. 1968. When are species necessary? In Population biology and
evolution, edited by R. C. Lewontin. Syracuse NY: Syracuse University Press.
Istock, C. A., J. A. Bell, N. Ferguson, and N. L. Istock. 1996. Bacterial species and
evolution: Theoretical and practical perspectives. Journal of Industrial
Microbiology and Biotechnology 17 (3 - 4):137-150.
Konstantinidis, K., A. Ramette, and J. Tiedje. 2006. The bacterial species definition in
the genomic era. Philosophical Transactions of the Royal Society B:
Biological Sciences 361 (1475):1929-1940.
Koonin, E. V., K. S. Makarova, and L. Aravind. 2001. Horizontal gene transfer in
prokaryotes: quantification and classification. Annu Rev Microbiol 55:709742.
Kunin, V., L. Goldovsky, N. Darzentas, and C. A. Ouzounis. 2005. The net of life:
reconstructing the microbial phylogenetic network. Genome Res 15 (7):954959. Epub 2005 Jun 2017.
Kurland, C. G., B. Canback, and Otto G. Berg. 2003. Horizontal gene transfer: A
critical view. PNAS 100 (17):9658-9662.
Makarenkov, V., and P. Legendre. 2004. From a phylogenetic tree to a reticulated
network. J Comput Biol 11 (1):195-212.
Mallet, J. 1995. The species definition for the modern synthesis. Trends in Ecology
and Evolution 10 (7):294-299.
Meier, Rudolf, and Rainer Willmann. 2000. The Hennigian species concept. In
Species concepts and phylogenetic theory: A debate, edited by Q. Wheeler and
R. Meier. New York: Columbia University Press.
Moreno, E. 1997. In search of a bacterial species definition. Revista de Biologia
Tropical 45 (2):753-771.
O’Malley, Maureen, and John Dupré. 2007. Size doesn’t matter: towards a more
inclusive philosophy of biology. Biology and Philosophy 22 (2):155-191.
Ochman, Howard, Jeffrey G. Lawrence, and Eduardo A. Groisman. 2000. Lateral
gene transfer and the nature of bacterial innovation. Nature 405:299-304.
Paterson, Hugh E. H. 1985. The recognition concept of species. In Species and
speciation, edited by E. Vrba. Pretoria: Transvaal Museum.
Pie, Marcio R., and Joshua S. Weitz. 2005. A Null Model of Morphospace
Occupation. Am Nat 166 (1):E1-E13.
Pigliucci, Massimo, and Jonathon Kaplan. 2006. Making sense of evolution: the
conceptual foundations of evolutionary biology. Chicago and London:
University of Chicago Press.
Ragan, Mark A. 2001. Detection of lateral gene transfer among microbial genomes.
Curr Opin Genet Dev. 11 (6):620-626.
Raymond, Jason, and Robert E. Blankenship. 2003. Horizontal gene transfer in
eukaryotic algal evolution. PNAS 100 (13):7419-7420.
Skipper, Robert A. 2004. The Heuristic Role of Sewall Wright's 1932 Adaptive
Landscape Diagram. Philosophy of Science 71 (5):1176-1188.
Sterelny, Kim. 1999. Species as evolutionary mosaics. In Species, New
interdisciplinary essays, edited by R. A. Wilson. Cambridge, MA:
Bradford/MIT Press.
Syvanen, Michael, and Clarence I. Kado, eds. 1998. Horizontal gene transfer. Boston:
Kluwer Academic Publishers.
Templeton, Alan R. 1998. Species and speciation: geography, population structure,
ecology, and gene trees. In Endless forms: species and speciation, edited by
D. J. Howard and S. H. Berlocher. New York: Oxford University Press.
Templeton, Alan R. 1989. The meaning of species and speciation: A genetic
perspective. In Speciation and its consequences, edited by D. Otte and J.
Endler. Sunderland, MA: Sinauer.
Tibayrenc, M. 2006. The species concept in parasites and other pathogens: a
pragmatic approach? Trends in Parasitology 22 (2):66-70.
Van Valen, L. 1976. Ecological species, multispecies, and oaks. Taxon 25:233-239.
Waters, V. L. 2001. Conjugation between bacterial and mammalian cells. Nature
Genetics 29 (4):375-376.
Wertz, J. E., C. Goldstone, D. M. Gordon, and M. A. Riley. 2003. A molecular
phylogeny of enteric bacteria and implications for a bacterial species concept.
J Evol Biol 16 (6):1236-1248.
Wilkins, John. 2007. The dimensions, modes and definitions of species and
speciation. Biology and Philosophy 22 (2):247-266.
Wilkins, John S. 2003. How to be a chaste species pluralist-realist: The origins of
species modes and the Synapomorphic Species Concept. Biology and
Philosophy 18:621-638.
Woese, C. 1998. The universal ancestor. Proceedings of the National Academy of
Sciences of the United States of America 95 (12):6854-6859.
Wolf, Y. I., I. B. Rogozin, N. V. Grishin, and E. V. Koonin. 2002. Genome trees and
the tree of life. Trends Genet 18 (9):472-479.
Acknowledgements
I am enormously indebted to Professor Paul Griffiths, ARC Federation Fellowship
No. FF0457917, for financial support and to him and Dr Stefan Linquist for much
discussion and advice. Professor Margaret A. (Peg) Riley provided figures and
advice, and Dr Stephen Bent some references. Dr Maureen O’Malley and Professor
John Dupré generously gave an invitation and funding to the Philosophy of
Microbiology workshop in July 2006, and I thank its attendees and my blog readers
for comment and discussion. John Dupré, Staffan Müller-Wille and Jan Sapp also
offered extensive and useful comments. Obviously, none of them can be blamed for
the paper’s remaining failings.