13. Hox Genes and Arthropod Diversity
201
arthropod biologists have not been able to agree about this for the last 100 years.
There are two primary competing hypotheses. One simply counts segments backwards from the eye segment, homologising chelicerae with the insect antenna (crustacean first antenna), and so on. The other assumes that an ancestral antennal
segment has been "deleted" in the chelicerate lineage, and equates the chelicerae
with the second antennal segment of Crustacea (insect intercalary segment), and
therefore the pedipalps with the mandible. The principle argument in favour of this
second model relates to the innervation of the segments, and the position of segmental ganglia in relation to the oesophagus (Weygoldt 1985). In Mandibulates the
mandibular ganglion is the first post-oral ganglion; in chelicerates it is the pedipalp ganglion.
There is a third possibility, not generally made explicit. This is that any attempt
to homologise segments is meaningless, because at the time when these two lineages diverged, the post-antennal segments were not individuated, or differentiated
from one another, and therefore deletion of a segment would have no meaning.
Two recent papers describing Hox gene expression patterns in chelicerates
(Damen et al. 1998; Telford and Thomas 1998) go a long way towards proving that
the question is meaningful, and that the first of the above hypotheses is correct.
We have seen that the individual anterior Hox genes themselves are older than
the protostome/deuterostome split, and therefore more ancient than any conceivable divergence of chelicerates and mandibulates. We also know that in insects
these genes serve as markers for specific head segments, and define at least some of
the differences between them. Characteristically, the insect antennal segment expresses no Hox gene, and may be regarded as a ground state for a trunk segment, at
least with respect to Box gene function. The intercalary segment expresses the most
anterior of the Hox genes, labial, and one segment behind this, the Deformed gene
is active in the mandible (Rogers and Kaufman 1997). An anterior limit to the
trunk is defined by a homeobox gene of a different class, Orthodenticle, which
defines fore-brain territory, and specifically the ocular segment.
The Hox expression data relate to two different chelicerate groups, an Oribatid
mite, Archegozites, and a spider, Cupiennius. As we would expect from earlier
work with insects, the Chelicerate Hox genes are expressed in specific sets of segments with defined anterior boundaries. Strikingly, insects .and chelicerates show a
corresponding set of Hox gene boundaries, if it is assumed that no segment is missing (Fig. 1). The anterior limit of Hox gene expression is the pedipalp segment in
chelicerates, suggesting that this is the homologue of the intercalary segment of
insects; The Deformed genes (class 4) are expressed from the first leg bearing segment, and genes of class 5 (Sex combs reduced) from the third leg segment. There
are differences between Drosophila and these chelicerates, particularly in the domains of the class 2 and 3 genes, but there are good reasons to believe that these
genes have changed within the insects, and that the patterns in the chelicerates are
ancestral. Other conserved details make it unlikely that these two corresponding
patterns have been acquired independently in the two groups. For example, genes
of class 1 and 4 appear to be expressed in register with the segment boundaries,
whereas genes of class 5 respect anterior limits that lie out of register with seg-
201
arthropod biologists have not been able to agree about this for the last 100 years.
There are two primary competing hypotheses. One simply counts segments backwards from the eye segment, homologising chelicerae with the insect antenna (crustacean first antenna), and so on. The other assumes that an ancestral antennal
segment has been "deleted" in the chelicerate lineage, and equates the chelicerae
with the second antennal segment of Crustacea (insect intercalary segment), and
therefore the pedipalps with the mandible. The principle argument in favour of this
second model relates to the innervation of the segments, and the position of segmental ganglia in relation to the oesophagus (Weygoldt 1985). In Mandibulates the
mandibular ganglion is the first post-oral ganglion; in chelicerates it is the pedipalp ganglion.
There is a third possibility, not generally made explicit. This is that any attempt
to homologise segments is meaningless, because at the time when these two lineages diverged, the post-antennal segments were not individuated, or differentiated
from one another, and therefore deletion of a segment would have no meaning.
Two recent papers describing Hox gene expression patterns in chelicerates
(Damen et al. 1998; Telford and Thomas 1998) go a long way towards proving that
the question is meaningful, and that the first of the above hypotheses is correct.
We have seen that the individual anterior Hox genes themselves are older than
the protostome/deuterostome split, and therefore more ancient than any conceivable divergence of chelicerates and mandibulates. We also know that in insects
these genes serve as markers for specific head segments, and define at least some of
the differences between them. Characteristically, the insect antennal segment expresses no Hox gene, and may be regarded as a ground state for a trunk segment, at
least with respect to Box gene function. The intercalary segment expresses the most
anterior of the Hox genes, labial, and one segment behind this, the Deformed gene
is active in the mandible (Rogers and Kaufman 1997). An anterior limit to the
trunk is defined by a homeobox gene of a different class, Orthodenticle, which
defines fore-brain territory, and specifically the ocular segment.
The Hox expression data relate to two different chelicerate groups, an Oribatid
mite, Archegozites, and a spider, Cupiennius. As we would expect from earlier
work with insects, the Chelicerate Hox genes are expressed in specific sets of segments with defined anterior boundaries. Strikingly, insects .and chelicerates show a
corresponding set of Hox gene boundaries, if it is assumed that no segment is missing (Fig. 1). The anterior limit of Hox gene expression is the pedipalp segment in
chelicerates, suggesting that this is the homologue of the intercalary segment of
insects; The Deformed genes (class 4) are expressed from the first leg bearing segment, and genes of class 5 (Sex combs reduced) from the third leg segment. There
are differences between Drosophila and these chelicerates, particularly in the domains of the class 2 and 3 genes, but there are good reasons to believe that these
genes have changed within the insects, and that the patterns in the chelicerates are
ancestral. Other conserved details make it unlikely that these two corresponding
patterns have been acquired independently in the two groups. For example, genes
of class 1 and 4 appear to be expressed in register with the segment boundaries,
whereas genes of class 5 respect anterior limits that lie out of register with seg-