Box 6.1 (continued)
species. A value of 70 % RBR roughly corresponds to
nucleotide mismatch rates on the order of 5 % in
housekeeping genes. This is exactly what the ΔTm
measures, with a drop of 1
C per % genomics mismatch.
However, it is the significance of phylogenetic
groupings measured, for instance, by the iterative
“bootstrap” resampling method – and not an absolute
threshold value – that must be used to delimit species
in MLSA (Gevers et al. 2005). Other criteria, such as
the average nucleotide identity or ANI (Goris et al.
2007) or other methods also based on the alignments
of sequenced genomes, reveal not only discontinuities
between species but also between genera. Because of
easy access to complete genome sequencing, it can be
expected that in the future, depositing the genomic
reference sequence will be a prerequisite to the definition of new bacterial species.
As shown in Box Fig. 6.1, all these methods show
that there is a discontinuity in the distribution of divergence measures (AFLP, MLSA, ΔTm) – or symmetrically of the similarity (RBR, ANI) – between genomes.
This determination is empirical. What may be arbitrary
is the choice of the degree of genomic difference to
define entities called “species” about which one may
wonder how they compare or not to eukaryotic species.
Speciation of Genomic Species Is “Fixed” in the Past
It must be understood that this definition implies that
genomic species are bacterial lineages that have
differentiated and have been isolated long enough that
their genomic differences have led to the thresholds
described above. An immediate consequence of this is
that the events that led to this divergence have occurred
long ago: speciation of genomic species was fixed in the
past. In addition, since the infraspecific divergence is of
the same order of magnitude for all species, speciation
would have occurred roughly at the same time in various
lineages! However, the phenomenon of speciation
occurred and still occurs now, as is the case for many
bacterial pathogens such as Yersinia pestis, Mycobacterium tuberculosis, or Bacillus anthracis that belong to
wider “genomic” species. Since the genomic bacterial
species definition does not integrate contemporary
speciation, it is therefore irreconcilable with the BSC of
eukaryotes.
(continued)
Box 6.1 (continued)
Homologous Recombination and Sexual Isolation
of Genomic Species
In eukaryotes, BSC is based on the sexual isolation of
species. In prokaryotes, “sexuality” is both more promiscuous because it can involve very distantly related
taxa (e.g., Firmicutes vs. Proteobacteria) and, very partial, because it concerns only a fraction of the genome.
However, it has been suggested that the genetic divergence between species is such that it could result in a
significant decrease in the frequency of homologous
recombination (i.e., the mechanism by which foreign
DNA is integrated into the bacterial genome) and thus
lead to relative sexual isolation of species. Attempts
have been made to explore this idea because it was
attractive to be able to reconcile the two concepts.
Briefly, it appears that in a model organism (the
Agrobacterium tumefaciens species complex) and for
a given marker gene, the decrease in the rate of homologous recombination was from 8 times, between very
distant strains belonging to the same species, to 9 times,
between strains belonging to different but closely
related species (Costechareyre et al. 2009). This difference does not appear very significant, and homologous
recombination is probably not sufficient to explain the
genetic isolation of bacterial species. A sexual isolation,
however, can have other causes. Studies of population
genetics, for example, the work of Bailly et al. (2006),
showed that sympatric species of Sinorhizobium
are sufficiently isolated that gene exchanges occur
significantly more between members of the same
species than between species. The nature of the barrier,
physiological, geographical, or otherwise, which leads
to the sexual isolation of these species, is not known.
The Cause of Discontinuities Between Genomic Species
The existence of genomic groups is the result of forces
that have swept or “purged” the diversity among these
groups. Two types of models are available for these
forces. One is based on selection. The other involves
genetic drift with a predominant role of the founder
effect.
Genomic Species Versus Ecological Species
For Cohan (2001), in the world of prokaryotes, each
genomic group could correspond to an “ecotype”
defined as a population of cells occupying the same
(continued)
148
P. Caumette et al.
species. A value of 70 % RBR roughly corresponds to
nucleotide mismatch rates on the order of 5 % in
housekeeping genes. This is exactly what the ΔTm
measures, with a drop of 1
C per % genomics mismatch.
However, it is the significance of phylogenetic
groupings measured, for instance, by the iterative
“bootstrap” resampling method – and not an absolute
threshold value – that must be used to delimit species
in MLSA (Gevers et al. 2005). Other criteria, such as
the average nucleotide identity or ANI (Goris et al.
2007) or other methods also based on the alignments
of sequenced genomes, reveal not only discontinuities
between species but also between genera. Because of
easy access to complete genome sequencing, it can be
expected that in the future, depositing the genomic
reference sequence will be a prerequisite to the definition of new bacterial species.
As shown in Box Fig. 6.1, all these methods show
that there is a discontinuity in the distribution of divergence measures (AFLP, MLSA, ΔTm) – or symmetrically of the similarity (RBR, ANI) – between genomes.
This determination is empirical. What may be arbitrary
is the choice of the degree of genomic difference to
define entities called “species” about which one may
wonder how they compare or not to eukaryotic species.
Speciation of Genomic Species Is “Fixed” in the Past
It must be understood that this definition implies that
genomic species are bacterial lineages that have
differentiated and have been isolated long enough that
their genomic differences have led to the thresholds
described above. An immediate consequence of this is
that the events that led to this divergence have occurred
long ago: speciation of genomic species was fixed in the
past. In addition, since the infraspecific divergence is of
the same order of magnitude for all species, speciation
would have occurred roughly at the same time in various
lineages! However, the phenomenon of speciation
occurred and still occurs now, as is the case for many
bacterial pathogens such as Yersinia pestis, Mycobacterium tuberculosis, or Bacillus anthracis that belong to
wider “genomic” species. Since the genomic bacterial
species definition does not integrate contemporary
speciation, it is therefore irreconcilable with the BSC of
eukaryotes.
(continued)
Box 6.1 (continued)
Homologous Recombination and Sexual Isolation
of Genomic Species
In eukaryotes, BSC is based on the sexual isolation of
species. In prokaryotes, “sexuality” is both more promiscuous because it can involve very distantly related
taxa (e.g., Firmicutes vs. Proteobacteria) and, very partial, because it concerns only a fraction of the genome.
However, it has been suggested that the genetic divergence between species is such that it could result in a
significant decrease in the frequency of homologous
recombination (i.e., the mechanism by which foreign
DNA is integrated into the bacterial genome) and thus
lead to relative sexual isolation of species. Attempts
have been made to explore this idea because it was
attractive to be able to reconcile the two concepts.
Briefly, it appears that in a model organism (the
Agrobacterium tumefaciens species complex) and for
a given marker gene, the decrease in the rate of homologous recombination was from 8 times, between very
distant strains belonging to the same species, to 9 times,
between strains belonging to different but closely
related species (Costechareyre et al. 2009). This difference does not appear very significant, and homologous
recombination is probably not sufficient to explain the
genetic isolation of bacterial species. A sexual isolation,
however, can have other causes. Studies of population
genetics, for example, the work of Bailly et al. (2006),
showed that sympatric species of Sinorhizobium
are sufficiently isolated that gene exchanges occur
significantly more between members of the same
species than between species. The nature of the barrier,
physiological, geographical, or otherwise, which leads
to the sexual isolation of these species, is not known.
The Cause of Discontinuities Between Genomic Species
The existence of genomic groups is the result of forces
that have swept or “purged” the diversity among these
groups. Two types of models are available for these
forces. One is based on selection. The other involves
genetic drift with a predominant role of the founder
effect.
Genomic Species Versus Ecological Species
For Cohan (2001), in the world of prokaryotes, each
genomic group could correspond to an “ecotype”
defined as a population of cells occupying the same
(continued)
148
P. Caumette et al.