13. Hox Genes and Arthropod Diversity
203
4.2 The Arthropod Trunk
The situation in the arthropod trunk must be different. Segment number is not
conserved between all arthropods, and even within closely related groups, segment
numbers vary within well defined homologous regions of the body [e.g., in branchiopod crustaceans, and in geophilomorph centipedes (Schram 1986; Minelli and
Bortoletto 1988)]. This alone suggests that there must be some flexibility in the
relationship between the mechanisms of segment formation and those that regulate
Hox genes, leading to segment identity. Here it makes no sense simply to count
segments back from the head to establish "homologies". There need be no such
thing as a one-to-one homology between segments in two different species. A more
relevant comparison may be to ask how domains of Hox gene expression relate, not
to individual segments, but to the "regiments" of similar segments long recognised
as defining the pattern of tagmosis in the arthropod body. Our expectation is that
boundaries of Hox gene expression will define positions along the body axis where
segment morphologies change.
A clear correlation between tagmosis and Hox gene expression was revealed by
our studies of Hox gene expression in the Branchiopod crustaceanArtemia (Averof
and Akam 1995a). Artemia is a crustacean with a relatively simple body plan, with
clearly defined trunk tagmosis, and little segment variation within tagma. All the
11 segments of the thorax are of similar form, differing only in size. These uniform
segments all express the same set of Hox genes, albeit at different levels. The distinct genital segments express a Hox gene of the more posterior Abd-B class, and
the so-called abdomen of Artemia, lying posterior to the genitalia, expresses none
of the Hox genes homologous to those in insects.
If we interpret these gene expression patterns as marking homology between
body regions, we would conclude that the branchiopod thorax is homologous to the
whole of the insect trunk anterior to the genital segments, while the branchiopod
abdomen is either unrepresented in insects, or is represented only by the post-genital tenth abdominal segment (a'segment which never bears appendages in insects).
This was our interpretation when we published these data, and it remains my favoured
hypothesis today. However, we will only know whether or not this interpretation
makes sense when we have data on more genes, and have sampled more species.
Damen et al. (1998) have described Hox gene expression in the more posterior
parts of the chelicerate body. In their spider, the transition from prosoma to
opisthosoma ("abdomen") is marked by activation of the middle group Hox genes
(Antennapedia, two distinct Ultrabithorax-like genes, and abdominal-A). In this
regard Hox gene expression reflects tagmosis. Within the opisthosoma, each of
these Hox genes is activated at a different segmental boundary. These authors suggest that this series of boundaries can be homologised with the boundaries observed
in insects, as suggested above for the head segments, and so used to establish a oneto-one correspondence between insect and chelicerate segments throughout much
of the trunk.
I find it hard to reconcile this interpretation with observations in crustacea and
other arthropods (see below), which show no such precise correspondence in ex-
203
4.2 The Arthropod Trunk
The situation in the arthropod trunk must be different. Segment number is not
conserved between all arthropods, and even within closely related groups, segment
numbers vary within well defined homologous regions of the body [e.g., in branchiopod crustaceans, and in geophilomorph centipedes (Schram 1986; Minelli and
Bortoletto 1988)]. This alone suggests that there must be some flexibility in the
relationship between the mechanisms of segment formation and those that regulate
Hox genes, leading to segment identity. Here it makes no sense simply to count
segments back from the head to establish "homologies". There need be no such
thing as a one-to-one homology between segments in two different species. A more
relevant comparison may be to ask how domains of Hox gene expression relate, not
to individual segments, but to the "regiments" of similar segments long recognised
as defining the pattern of tagmosis in the arthropod body. Our expectation is that
boundaries of Hox gene expression will define positions along the body axis where
segment morphologies change.
A clear correlation between tagmosis and Hox gene expression was revealed by
our studies of Hox gene expression in the Branchiopod crustaceanArtemia (Averof
and Akam 1995a). Artemia is a crustacean with a relatively simple body plan, with
clearly defined trunk tagmosis, and little segment variation within tagma. All the
11 segments of the thorax are of similar form, differing only in size. These uniform
segments all express the same set of Hox genes, albeit at different levels. The distinct genital segments express a Hox gene of the more posterior Abd-B class, and
the so-called abdomen of Artemia, lying posterior to the genitalia, expresses none
of the Hox genes homologous to those in insects.
If we interpret these gene expression patterns as marking homology between
body regions, we would conclude that the branchiopod thorax is homologous to the
whole of the insect trunk anterior to the genital segments, while the branchiopod
abdomen is either unrepresented in insects, or is represented only by the post-genital tenth abdominal segment (a'segment which never bears appendages in insects).
This was our interpretation when we published these data, and it remains my favoured
hypothesis today. However, we will only know whether or not this interpretation
makes sense when we have data on more genes, and have sampled more species.
Damen et al. (1998) have described Hox gene expression in the more posterior
parts of the chelicerate body. In their spider, the transition from prosoma to
opisthosoma ("abdomen") is marked by activation of the middle group Hox genes
(Antennapedia, two distinct Ultrabithorax-like genes, and abdominal-A). In this
regard Hox gene expression reflects tagmosis. Within the opisthosoma, each of
these Hox genes is activated at a different segmental boundary. These authors suggest that this series of boundaries can be homologised with the boundaries observed
in insects, as suggested above for the head segments, and so used to establish a oneto-one correspondence between insect and chelicerate segments throughout much
of the trunk.
I find it hard to reconcile this interpretation with observations in crustacea and
other arthropods (see below), which show no such precise correspondence in ex-