204
M. Akam
pression domains, and more importantly, show that even between quite closely
related arthropods, the boundaries of Hox gene expression vary. I am more inclined
to interpret the Chelicerate data as showing some degree of convergent similarity.
It is important to remember that the anterior segments of the opisthosoma are highly
differentiated in chelicerates, containing the male and female genitalia (Brusca and
Brusca 1990). It is therefore hardly surprising that they are distinguished by distinct and precisely defined domains of Hox gene expression. Distinguishing between these interpretations will again require sampling a much wider range of taxa
(and, I fear, resolving the phylogeny of the major arthropod groups).
The only relevant data so far available have been obtained by using a crossreacting antibody (Kelsh et al. 1994) that detects one subclass of Hox proteinsthe Ubx and abd-A gene products, in all tested arthropods, and in some closely
related phyla. The availability of this antibody makes it possible to survey a range
of species without having to clone genes from each. Using this antibody, Grenier
and colleagues (1997) have shown that the Ubx/abd-A class Hox proteins are expressed throughout the trunk of centipedes, except in the first specialised poison
claw segment. Our own work with a different species of centipede (Lithobius
forficatus) suggests that these genes are initially expressed in all trunk segments,
including the poison claw segment, and only later in embryogenesis become restricted to the leg bearing segments (Smith 1998). At first sight this pattern is
directly comparable with that of the same genes in Artemia, defining a domain of
the body lying immediately behind the head. However, the chelicerate data described above, and a very different pattern observed for these same genes in
Onychophora (Grenier et al. 1997), suggests that it would be premature to posit this
as all ancestral state for all arthropods.
Onychophora are probably the closest living relatives of the arthropods proper,
and are generally assumed to share a common segmented ancestor. Behind the
head, they have a uniform trunk with many similar lobo pod bearing segments,
which might therefore be expected to show a pattern of Hox gene expression ancestral to an unspecialised trunk segment. However, the cross-reacting Ubx/abd-A
antibody stains only the last appendage-bearing segment and the terminal region of
the Onychophoran trunk - a domain that does not obviously reflect any known
segment specialisation. (In this case the relevant genes have been cloned, confirming that the epitope recognised by the antibody is conserved in the expected Hox
gene products.) Unfortunately, we do not yet know which Hox genes are expressed
in the more anterior segments.
On the basis of these observations, Grenier and colleagues suggest that "expression domains of Ubx/abd-A do not demarcate homologous body regions in these
taxa [onychophorans, chelicerates, crustaceans], but instead, evolutionary changes
in Ubx/abd-A deployment underlie the diversification of arthropod body plans that
are evident in the Cambrian fossil record". This is perhaps correct, but too negative
as an overall assessment of the significance of Hox expression domains. Demarcation of body regions within the arthropod trunk probably occurred independently
within the most basal lineages, and for these lineages, we do not expect to be able to
M. Akam
pression domains, and more importantly, show that even between quite closely
related arthropods, the boundaries of Hox gene expression vary. I am more inclined
to interpret the Chelicerate data as showing some degree of convergent similarity.
It is important to remember that the anterior segments of the opisthosoma are highly
differentiated in chelicerates, containing the male and female genitalia (Brusca and
Brusca 1990). It is therefore hardly surprising that they are distinguished by distinct and precisely defined domains of Hox gene expression. Distinguishing between these interpretations will again require sampling a much wider range of taxa
(and, I fear, resolving the phylogeny of the major arthropod groups).
The only relevant data so far available have been obtained by using a crossreacting antibody (Kelsh et al. 1994) that detects one subclass of Hox proteinsthe Ubx and abd-A gene products, in all tested arthropods, and in some closely
related phyla. The availability of this antibody makes it possible to survey a range
of species without having to clone genes from each. Using this antibody, Grenier
and colleagues (1997) have shown that the Ubx/abd-A class Hox proteins are expressed throughout the trunk of centipedes, except in the first specialised poison
claw segment. Our own work with a different species of centipede (Lithobius
forficatus) suggests that these genes are initially expressed in all trunk segments,
including the poison claw segment, and only later in embryogenesis become restricted to the leg bearing segments (Smith 1998). At first sight this pattern is
directly comparable with that of the same genes in Artemia, defining a domain of
the body lying immediately behind the head. However, the chelicerate data described above, and a very different pattern observed for these same genes in
Onychophora (Grenier et al. 1997), suggests that it would be premature to posit this
as all ancestral state for all arthropods.
Onychophora are probably the closest living relatives of the arthropods proper,
and are generally assumed to share a common segmented ancestor. Behind the
head, they have a uniform trunk with many similar lobo pod bearing segments,
which might therefore be expected to show a pattern of Hox gene expression ancestral to an unspecialised trunk segment. However, the cross-reacting Ubx/abd-A
antibody stains only the last appendage-bearing segment and the terminal region of
the Onychophoran trunk - a domain that does not obviously reflect any known
segment specialisation. (In this case the relevant genes have been cloned, confirming that the epitope recognised by the antibody is conserved in the expected Hox
gene products.) Unfortunately, we do not yet know which Hox genes are expressed
in the more anterior segments.
On the basis of these observations, Grenier and colleagues suggest that "expression domains of Ubx/abd-A do not demarcate homologous body regions in these
taxa [onychophorans, chelicerates, crustaceans], but instead, evolutionary changes
in Ubx/abd-A deployment underlie the diversification of arthropod body plans that
are evident in the Cambrian fossil record". This is perhaps correct, but too negative
as an overall assessment of the significance of Hox expression domains. Demarcation of body regions within the arthropod trunk probably occurred independently
within the most basal lineages, and for these lineages, we do not expect to be able to