156
Xenopus
each called a subgenome. According to whole genome
analysis of X. laevis (Session et al., 2016), in the subgenus
Xenopus lineage, bifurcating speciation of a diploid species
occurred 34 million years ago (Mya) to generate two species, the so-called “L” and “S” species (see subsequently for
the explanation of “L” and “S”). Then, allotetraploidization
occurred 17~18 Mya between the two species to generate
a new allotetraploid species. This is the common ancestor
of the subgenus Xenopus and had subgenomes “L” and “S.”
The “same” genes in the original L and S species are called
orthologs, but after allotetraploidization, those genes came
to reside in a single species, and now the former orthologs
became “homeologs” in subgenomes L and S. After allotetraploidization, repeated specifcation occurred to generate
several Xenopus species, as shown in Figure 10.1A. Some of
them underwent additional rounds of allopolyploidization to
generate octoploid and dodecaploid species.
WGD is one of the driving forces behind evolution. This
is postulated to have occurred twice in the common ancestor of vertebrates 500~600 Mya (Meyer and Van de Peer,
2005) and a third time in the ancestor species of bony f shes
about 306 Mya (Inoue et al., 2015) and is considered to have
contributed greatly to diversity. However, over the years
since WGD, chromosomes have reorganized and become
diploid, referred to as “diploidization,” and it is now impossible to distinguish whether the ancestral vertebrate species
was autopolyploid or allopolyploid. Compared to those species, allotetraploidization in the Xenopus lineage occurred
relatively very recently, and thereby subgenomes L and S
were clearly identifed in X. laevis by whole genome analysis. Thus, though allotetraploidy of X. laevis has long been
considered a disadvantage for genetics, the identif cation of
subgenomes made X. laevis a very useful model for studying
subgenome evolution after allopolyploidization.
10.1.1. HETEROSIS
Heterosis, or hybrid vigor, is observed in hybrids that have superior phenotypes such as more vigorous growth or higher progeny yields compared to their parents. This phenomenon is also
observed in allopolyploid plants, for example, wheat (Chen,
2013). In animals, heterosis is commonly known in mules
whose sire is a donkey and dam is a horse, and they are particularly useful as working animals for their physical strength.
However, interspecies hybrids are generally infertile, mostly
due to the impairment of meiosis. If WGD occurs in F1 hybrids
(i.e. allopolyploidization), they could be fertile and possibly
lead to the generation of a new species (Chen, 2013). Therefore,
allopolyploid animals are expected to exhibit heterosis, but
this has not been actually demonstrated. But still, this can be
inferred from two lines of circumstantial evidence for Xenopus.
First, the parent species L and S of the subgenus Xenopus are
extinct, possibly due to domination of their descendant allotetraploids over the parental species; second, the habitat of diploid
(not polyploid) X. tropicalis is limited to the region including
Nigeria and Ivory Coast, whereas species of subgenus Xenopus
inhabit a much wider region (Tinsley et al., 1996; Evans et al.,
2004). Therefore, it may have been an advantage for Xenopus
species to have become allotetraploid (or higher).
10.1.2. RECOGNITION OF ALLOTETRAPLOIDY
What led to the current understanding that X. tropicalis is
diploid and X. laevis is allotetraploid? Analyses of various
species from the Xenopus genus and other genera in Pipidae
showed that their number of chromosomes and the DNA
content per cell were in proportions of approximately 1:2:4:6
and that the basal level is those of X. tropicalis, which has
20 chromosomes (n = 10) and 3.55 pg DNA/cell, indicating
the existence of polyploid species (Kobel, 1981). The f nding
that the numbers of X. laevis was about twice of X. tropicalis suggested tetraploidy. Since around 1990, X. laevis has
become widely recognized as being allotetraploid, after gene
cloning became popular in the late 1980s and the presence of
two different sequences for many genes were revealed. For
example, for the hoxb7 gene (formerly known as XlHbox2),
the clone called p52 identifed by Wright et al. (1987) was a
different version of the gene identifed by Müller et al. (1984),
called MM3. These two sequences showed high identity to
each other, raising the possibility that this ref ected polymorphism, but analyses in X. laevis/X. borealis interspecif c
hybrids showed that these two genes do not segregate from
each other, which led to the conclusion that these are not
alleles of the same gene (Fritz et al., 1989). At that time, the
f rst identifed gene was given a postfx “a,” and the second
gene was given “b”; thus, they were usually called the a- and
b-genes. However, analysis of one of the two genes was (or
appeared to be) practical for developmental investigations;
because of the high nucleotide identities of coding sequences
as well as translated amino acid identities between the two
versions, cross-hybridization was inevitable for standard
Northern blots and whole-mount in situ hybridization, thus
making distinguishing the two diffcult. The difference in the
functions of the two versions was probably not always studied, but in the case of GATA-1, precise analyses showed that
GATA-1a and GATA-1b share the same function in stimulating erythropoiesis, but only the latter inhibits neurogenesis
when overexpressed (Xu et al., 1997). As another example,
two gbx2 genes, gbx2.2.L (Xgbx2a) and gbx2.1.S (Xgbx2b),
when overexpressed, caused malformation of the head and
notochord but showed differences in their temporal and spatial expression patterns (Tour et al., 2001).
Now, after the completion of the whole genome sequencing of X. laevis, the two versions a and b are acknowledged
as homeologs that arise from the different subgenomes
and are called the L and S genes. Back to the two versions
of XlHbox2/hoxb7, p52 and MM3 are now identif ed as
hoxb7.L and hoxb7.S, respectively (see Kondo et al., 2017).
GATA-1a and GATA-1b are gata1.L and gata1.S , respectively
(Watanabe et al., 2017). Note that a- and b-genes do not necessarily correspond to L and S genes, respectively, because
of the difference between their defnitions: the chronological
order of identifcation for a and b versus the position on the L
or S chromosomes.
Xenopus
each called a subgenome. According to whole genome
analysis of X. laevis (Session et al., 2016), in the subgenus
Xenopus lineage, bifurcating speciation of a diploid species
occurred 34 million years ago (Mya) to generate two species, the so-called “L” and “S” species (see subsequently for
the explanation of “L” and “S”). Then, allotetraploidization
occurred 17~18 Mya between the two species to generate
a new allotetraploid species. This is the common ancestor
of the subgenus Xenopus and had subgenomes “L” and “S.”
The “same” genes in the original L and S species are called
orthologs, but after allotetraploidization, those genes came
to reside in a single species, and now the former orthologs
became “homeologs” in subgenomes L and S. After allotetraploidization, repeated specifcation occurred to generate
several Xenopus species, as shown in Figure 10.1A. Some of
them underwent additional rounds of allopolyploidization to
generate octoploid and dodecaploid species.
WGD is one of the driving forces behind evolution. This
is postulated to have occurred twice in the common ancestor of vertebrates 500~600 Mya (Meyer and Van de Peer,
2005) and a third time in the ancestor species of bony f shes
about 306 Mya (Inoue et al., 2015) and is considered to have
contributed greatly to diversity. However, over the years
since WGD, chromosomes have reorganized and become
diploid, referred to as “diploidization,” and it is now impossible to distinguish whether the ancestral vertebrate species
was autopolyploid or allopolyploid. Compared to those species, allotetraploidization in the Xenopus lineage occurred
relatively very recently, and thereby subgenomes L and S
were clearly identifed in X. laevis by whole genome analysis. Thus, though allotetraploidy of X. laevis has long been
considered a disadvantage for genetics, the identif cation of
subgenomes made X. laevis a very useful model for studying
subgenome evolution after allopolyploidization.
10.1.1. HETEROSIS
Heterosis, or hybrid vigor, is observed in hybrids that have superior phenotypes such as more vigorous growth or higher progeny yields compared to their parents. This phenomenon is also
observed in allopolyploid plants, for example, wheat (Chen,
2013). In animals, heterosis is commonly known in mules
whose sire is a donkey and dam is a horse, and they are particularly useful as working animals for their physical strength.
However, interspecies hybrids are generally infertile, mostly
due to the impairment of meiosis. If WGD occurs in F1 hybrids
(i.e. allopolyploidization), they could be fertile and possibly
lead to the generation of a new species (Chen, 2013). Therefore,
allopolyploid animals are expected to exhibit heterosis, but
this has not been actually demonstrated. But still, this can be
inferred from two lines of circumstantial evidence for Xenopus.
First, the parent species L and S of the subgenus Xenopus are
extinct, possibly due to domination of their descendant allotetraploids over the parental species; second, the habitat of diploid
(not polyploid) X. tropicalis is limited to the region including
Nigeria and Ivory Coast, whereas species of subgenus Xenopus
inhabit a much wider region (Tinsley et al., 1996; Evans et al.,
2004). Therefore, it may have been an advantage for Xenopus
species to have become allotetraploid (or higher).
10.1.2. RECOGNITION OF ALLOTETRAPLOIDY
What led to the current understanding that X. tropicalis is
diploid and X. laevis is allotetraploid? Analyses of various
species from the Xenopus genus and other genera in Pipidae
showed that their number of chromosomes and the DNA
content per cell were in proportions of approximately 1:2:4:6
and that the basal level is those of X. tropicalis, which has
20 chromosomes (n = 10) and 3.55 pg DNA/cell, indicating
the existence of polyploid species (Kobel, 1981). The f nding
that the numbers of X. laevis was about twice of X. tropicalis suggested tetraploidy. Since around 1990, X. laevis has
become widely recognized as being allotetraploid, after gene
cloning became popular in the late 1980s and the presence of
two different sequences for many genes were revealed. For
example, for the hoxb7 gene (formerly known as XlHbox2),
the clone called p52 identifed by Wright et al. (1987) was a
different version of the gene identifed by Müller et al. (1984),
called MM3. These two sequences showed high identity to
each other, raising the possibility that this ref ected polymorphism, but analyses in X. laevis/X. borealis interspecif c
hybrids showed that these two genes do not segregate from
each other, which led to the conclusion that these are not
alleles of the same gene (Fritz et al., 1989). At that time, the
f rst identifed gene was given a postfx “a,” and the second
gene was given “b”; thus, they were usually called the a- and
b-genes. However, analysis of one of the two genes was (or
appeared to be) practical for developmental investigations;
because of the high nucleotide identities of coding sequences
as well as translated amino acid identities between the two
versions, cross-hybridization was inevitable for standard
Northern blots and whole-mount in situ hybridization, thus
making distinguishing the two diffcult. The difference in the
functions of the two versions was probably not always studied, but in the case of GATA-1, precise analyses showed that
GATA-1a and GATA-1b share the same function in stimulating erythropoiesis, but only the latter inhibits neurogenesis
when overexpressed (Xu et al., 1997). As another example,
two gbx2 genes, gbx2.2.L (Xgbx2a) and gbx2.1.S (Xgbx2b),
when overexpressed, caused malformation of the head and
notochord but showed differences in their temporal and spatial expression patterns (Tour et al., 2001).
Now, after the completion of the whole genome sequencing of X. laevis, the two versions a and b are acknowledged
as homeologs that arise from the different subgenomes
and are called the L and S genes. Back to the two versions
of XlHbox2/hoxb7, p52 and MM3 are now identif ed as
hoxb7.L and hoxb7.S, respectively (see Kondo et al., 2017).
GATA-1a and GATA-1b are gata1.L and gata1.S , respectively
(Watanabe et al., 2017). Note that a- and b-genes do not necessarily correspond to L and S genes, respectively, because
of the difference between their defnitions: the chronological
order of identifcation for a and b versus the position on the L
or S chromosomes.