Case 10
We tend to imagine the genome as a relatively discrete entity -- large,
complicated, variously mutable and recombined through sex, yes, but at
least something for which we can draw boundaries to say what is part of
it and what isn't. A bacterial infection, for example, is clearly
not part of it, just a parasite coming along for the ride. Except
that, in various interesting cases, things are not that clear cut at
all.
Inherited bacteria are found quite widely in invertebrates, and their
behaviour is interesting both in its own right and for the light it
casts on the now widely-accepted theory that eukaryotic cellular
organelles, like mitochondria and chloroplasts, were once separate
organisms that were adopted as symbionts and subsequently evolved to
become inextricable parts of their hosts.
The bacteria under consideration here are still identifiably separate
entities, which live inside the cells of their insect companions and
seldom travel beyond. Horizontal transmission -- that is, between mature
organisms -- is rare. Rather, they are passed from a mother to her
offspring in the gametes: her eggs are infected before they are even
fertilized. And not by accident.
Two classes of heritable bacteria are distinguished: primary
symbionts are obligate, which is to say the host cannot survive
without them, while secondary symbionts are conditionally
beneficial -- there are hosts without them, but infection confers
some potential advantage. In evolutionary terms, primary symbionts must
once have been secondary, until the host discovered it could dispense
with its own facilities for doing those things that its guests did
better, and live more efficiently by living in harmony.
The bacteria retain their own lifecycle and can be killed off by
antibiotics. In the case of primary symbionts, such treatment leads to
death or, at best, sterility in the host. Often the symbionts' role is
anabolic, manufacturing chemicals such as amino acids that the
host needs but cannot create on its own. This is particularly common in
insects whose diets are very restricted -- those that feed on wood, for
example, which is markedly lacking in interesting nutritional content.
Some dependent hosts actually have special organs in which the bacteria
are stored and maintained and their products used.
Since the offspring of these creatures cannot survive uninfected, the
mothers must guarantee transmission. To this end, bacteria are
transported from the bacteriome organs into the ovum. Such intricate
relationships are evolutionarily long-standing, the hosts and their
necessary diseases bound together for hundreds of millions of years.
Even though their genes remain in separate units, the bacteria must be
considered quasi-genomic now.
Secondary symbiosis is less cut and dried, and uninfected hosts may
prosper. The advantages of infection will often be environmentally
determined, so in some contexts it is better to be infected whereas in
others it is not. Typical benefits include resistance to fungal
infections or parasitoid wasp larvæ, or the ability to make use of
different diet plants. Infected and uninfected hosts may then not be
competing for exactly the same resources and the species as a whole --
if we can still regard it as just one -- can occupy a larger niche.
For the symbiont, its environmental niche is the host. Since
transmission typically occurs only maternally, male hosts are a
reproductive dead end -- and many inherited bacteria have evolved to
cause sex-ratio distortion in their hosts, leading to a
disproportionate number of female children.
One common mechanism for this is male killing: the mother still
lays approximately equal numbers of male and female eggs, but the males
do not hatch. With offspring battling for the same resources, removing
the sons improves the prospects for the daughters, and hence also for
their bacterial passengers. The process may provide a more immediate and
concrete gain: ladybirds, for example, exhibit dead larval cannibalism:
eggs that do not hatch are eaten by the larvæ, so the nutritional
resources that went into the boys are transferred directly to the girls.
Another way to shift the balance is feminization: the bacteria
cause genetically male offspring to develop as females. In woodlice, for
example, which are female heterogametic -- meaning that females
have two different sex chromosomes (ZW) while males have a duplication
(ZZ), in contrast to humans where it is the males that are heterogametic
(XY) -- the phenotypic difference between the sexes is caused by an
androgenic gland in the males; the inherited bacteria inhibit the
function of this gland, changing the males into phenotypic females.
Yet a third approach is induced parthenogenesis, whereby the
females wind up producing daughters asexually. Parthenogenesis is common
in social insects such as ants and bees, which are haplodiploids:
males have half as many chromosomes as females and are developed from
unfertilized eggs. Fertilization provides the second set of chromosomes
and leads to females. The inherited bacterial infection disrupts this
process, causing duplication of the maternal chromosomes in the
offspring so that the would-be haploid males instead become homozygous
diploid females. This can lead to the loss of sex altogether: the entire
population becomes female (and treatment with antibiotics leads to an
entirely male population that cannot reproduce at all).
There are also strategies that act through the males as well as the
females to increase the chances of infected offspring. In this case,
both sexes are produced, but infected males are modified so that their
sperm is incompatible with the eggs of uninfected females. Infected
females can thus be fertilized by all males, while uninfected females
are limited to uninfected males. As the prevalence of infection
increases, the uninfected females find it more difficult to locate a
compatible mate.
This strategy is interesting because it can lead to essentially disjoint
populations, infected by two different kinds of bacteria each of which
induces incompatibility to mates hosting the other. Even though the
populations are still notionally the same species, they can no longer
exchange genetic material and may subsequently diverge. In this case,
since it is important for individuals to be able to recognise compatible
mates, one obvious way they might diverge is in appearance, which is
often quite variable between isolated populations of the same species
even when they remain genetically compatible.
Male killing and other sex ratio distorting behaviour is not, in
general, particularly advantageous to the host, since male offspring are
good at spreading the host genes. Consequently, there is an ongoing
evolutionary tension between behaviour modification by the bacteria and
host adaptation, which can lead to rapid changes in particular
populations. Observations of Pacific island butterflies over the last
century or so have found drastic swings one way or the other over
remarkably short time periods. In Samoa, for example, male killing was
observed with very high prevalence for the whole of the 20th century,
but since the end of 2001 sex ratios have returned to 1:1. Males and
females are now all infected, but the MK behaviour is suppressed.