to differentiate between classical HGT and EGT. Consequently, numbers can be overestimated. Nevertheless,
available data suggest that they can represent a very significant proportion of the eukaryotic genomes (Doolittle et al.
2003). For example, the analysis of the genome of the yeast
Saccharomyces cerevisiae has shown that ~75 % of genes of
bacterial origin are likely of mitochondrial origin (Esser et al.
2004). In the case of the chloroplast endosymbiosis, a large
quantity of EGTs has been detected in the host genomes. For
example, more than 15 % of nuclear genes in Arabidopsis
thaliana are close to cyanobacterial homologues and could
thus have been acquired from the plastid endosymbiont (Martin et al. 2002).
Even if all known mitochondrial and plastidial genomes
only contain a small number of genes, their reduction by
EGT appears to be an ongoing process. The identification of
multiple recent and independent EGTs of the plastidial gene
infA towards the nuclear genome in different lineages of
angiosperms offers a good example (Millen et al. 2001).
Mitochondrial genomes also illustrate this phenomenon, as
shown by the study of 40 mitochondrial genes in 280 angiosperm species, which reveals 16 cases of parallel EGT during the evolution of these organisms (Adams et al. 2002).
This contrasts with the extreme homogeneity of the animal
mitochondrial genomes, which suggests that very few EGTs
have occurred in this lineage since their first diversification.
Thus, EGTs occur at different rates in different eukaryotic
phyla. Similarly, certain mitochondrial or plastidial genes
can be transferred in certain lineages but not in others.
Mitochondria and plastids are not the only actors in EGT
events since this phenomenon is general in all
endosymbioses. In fact, the study of the nuclear genomes
of diverse eukaryotes has revealed numerous gene
acquisitions from other types of endosymbionts. For example, fragments of the genome of the alphaproteobacterium
Wolbachia, an endosymbiont of nematodes and arthropods,
have been detected in the nuclear genomes of some of its
hosts (Hotopp et al. 2007). An even more surprising example
comes from the discovery of an almost complete copy of the
Wolbachia chromosome inserted in the genome of a beetle
(Callosobruchus chinensis) hosting this bacterium. This
insertion is probably very recent since it is not found in
another beetle species closely related to C. chinensis. The
possibility that this is an artifact or a nonfunctional sequence
can be discarded since it has been shown that at least 2 % of
the Wolbachia genes inserted in the beetle genome are
transcribed (Hotopp et al. 2007; Nikoh et al. 2008). These
results show not only that, in contrast with what is commonly assumed, there is no strong barrier that impedes the
incorporation of exogenous genes into the eukaryotic
genomes but also that this is possible even in animals,
where, as stated above, only transfers occurring in the
germ line can be fixed.
These examples illustrate that during the evolution of the
endosymbiotic relationships, transfers of genes from the
endosymbiont towards the host nuclear genome occur. Nevertheless, if in the case of mitochondria and plastids the
selective advantage of such transfers is clear, for many
other genes acquired from other endosymbionts a possible
selective advantage remains to be identified. Even the
hypothesis that they are kept in the nuclear genome just in
a transient way cannot be discarded.
EGT is a major source of gene acquisition by eukaryotes
and plays an important role in the evolution of these
organisms. However, eukaryotes can also acquire exogenous
DNA by other mechanisms. Many eukaryotic phyla have
phagotrophic species, namely, they feed on other cells (prokaryotic and/or eukaryotic). This mode of nutrition may
offer to these species the opportunity to acquire DNA from
their preys. This idea is summarized in an axiom, derived
from the German proverb “Man ist was man isst,”
reformulated by Doolittle in 1998 as “You are what you
eat” (Doolittle 1998). The implicit hypothesis is that a larger
number of exogenous genes has to be found in phagotrophic
eukaryotes than in non-phagotrophic ones. Moreover, it is
expected that these genes should have very diverse origins in
the nonspecialized phagotrophs, namely, those with a large
spectrum of preys. Even if few studies have analyzed this
hypothesis, some elements appear to support it. It has been
shown that typical phagotrophic eukaryotes, such as the
amoebas, seem to have acquired a variety of genes from
bacterial preys. This is the case for Dictyostelium
discoideum, which possesses at least 18 bacterial genes in
its genome. Some of them assure essential functions, as the
one coding for a glycoprotein involved in cell aggregation
(Eichinger et al. 2005). The number of genes of bacterial
origin detected in this amoeba is certainly small, but some
have also been detected in other amoebas very distantly
related to D. discoideum, such as species of the genus
Naegleria. This suggests independent HGT events and
points to an important role of those proteins in these
amoebas. Higher HGT rates have been found in another
group of phagotrophic eukaryotes, the ciliates, especially
in anaerobic species that thrive in the rumen of ruminants.
These ciliates share a relatively large number of genes (148)
with prokaryotes. Those genes are absent in other sequenced
eukaryotic genomes, suggesting that they are specific
acquisitions (Ricard et al. 2006). Some of these genes have
allowed these ciliates carrying out essential function for their
life in the rumen, for example, the capacity to degrade
complex carbohydrates, which are very abundant in the
rumen due to the herbivorous nutrition of ruminants.
12 Horizontal Gene Transfer in Microbial Ecosystems
477
Précédent

- 485/933

Suivant