«Joshua Frederick Coulcher UCL Submitted for the Degree of Doctor of Philosophy September 2011 Declaration I, Joshua Frederick Coulcher, confirm that ...»
Expression of cnc and Dfd are conserved in mandibulate arthropods. These observations indicate that the mandible patterning mechanism described for Tribolium is very likely to be conserved and probably originated once in the ancestor to all mandibulate arthropods. As the mandible has probably evolved from a maxilla-like precursor, the acquisition of a maxilla-to-mandible differentiating role for cnc resembles the likely evolutionary history of the mandibular appendage.
I have provided evidence based on expression of the Notch signalling pathway, which marks the formation of arthropod appendage segments, that the Tribolium mandible consists of two segments, a subcoxa and a coxa. The incisor and molar process are derived from the outer and inner lobes respectively, which develop from a solitary endite located on the mandibular coxa. There are significant similarities of the subcoxa and the coxa of the mandible to the subcoxa and coxa of other appendages based upon the timing and location of PD domain genes expression relative to the expression of a member of the Notch signalling pathway. These similarities are evidence of serial homology between the subcoxa of the mandible to the subcoxa of the maxilla, labial and leg appendages.
After discussing the results of the previous chapters in more detail, I will discuss the implications of the serial homology of the mandible subcoxa and coxa to other appendage segments. I will then present a summary of the mandibular patterning mechanism of Tribolium, followed by an evaluation of the likely ancestral expression of the Hox genes in the ancestor to the mandibulate arthropods and a model of mandible evolution via the acquisition of a mandible patterning function of cnc.
Mandible patterning genes in Tribolium
It was shown in chapter four that Tc cnc differentiates the mandible from a maxilla probably in part by repressing the maxilla patterning Hox genes. The Hox genes that are required to pattern the maxilla, Tc Dfd and mxp, are repressed by Tc cnc in regions of the mandibular segment. This is similar to what is observed in Drosophila. As the mandible has probably evolved by modification of a maxilla-like appendage, genetic patterning of the mandible by Tc cnc recapitulates evolution of the mandible from a maxilla. The expression of cnc is conserved in the mandibular segment of diverse mandibulate arthropods, which suggests that cnc shares a similar function and that the mandible patterning function of cnc may have evolved once in the stem lineage of Mandibulata.
cnc is a promising candidate gene to have evolved mandible patterning function in the ancestor to all mandibulates. cnc is the only known gene that differentiates the mandible segment from other segments in Tribolium and Drosophila, although I have shown in Tribolium that it is dependent on Dfd to achieve this function.
The function of cnc has only been studied in two arthropods, Tribolium and Drosophila, and only one of them a typical mandibulate, Tribolium. Therefore, to show that cnc was responsible for differentiation of the mandibular segment from the maxillary segment in ancestral mandibulates, it has to be shown that this role of cnc is ancestral to Mandibulata and that it does not have any such role in the sistergroup of Mandibulata. A likely corollary of this is that the function of cnc will be conserved in diverse mandibulates. In species in which cnc is shown not to pattern the mandible, it would have to be demonstrated that this loss of function is a derived condition.
cnc represses Dfd in the both Tribolium and Drosophila. In chapter five I showed that in Tribolium, Tc Dfd is required to pattern the mandible (a protopodite) and the protopodite of the maxilla and that it activates protopodite specific gene expression, such as Tc prd and the proximal domains of Tc dac and Tc Dll. The mandible gnathal edge evolved from the endite on the protopodite of a limb and therefore the regulation of Dfd by cnc is particularly relevant regarding the hypothesized ancestral patterning function of cnc to differentiate the mandible gnathal edge from the maxillary endites.
In chapter five I showed that Tc Dfd patterns the endite and protopodite of the maxilla but does not control limb segmentation. Tc Dfd activates Tc prd and the proximal domain of Tc dac and Tc Dll but does not activate Tc ser expression. It was not possible to show if this is the case in the mandible, as the mandible is transformed to antennal identity and therefore shows antennal specific expression domains of these genes. However, it is reasonable to hypothesize that Tc Dfd performs a similar function in the mandibular segment and that Tc cnc modifies expression of these genes to mandible segment specific domains of expression. Alternatively, Tc cnc may activate these genes directly. More detailed genetic experiments are required to unpick these specific genetic interactions. If these specific genetic interactions were demonstrated in other mandibulates it would provide more proof that the mandible patterning mechanism is conserved. Below I discuss the mechanism of mandibular segment patterning and the likelihood of its conservation in other mandibulates by examining expression of the Hox genes that are required to patterned gnathal appendages in other arthropods.
The role of cnc homologues in non-mandibulates
In order to prove that cnc acquired a new role to pattern the mandible in the ancestor to all mandibulates, it is necessary to determine that this role is shared by other mandibulates. To show that the mandibular segment patterning role of cnc (homologous to the first leg segment of chelicerates) was acquired once in the lineage leading to the mandibulate arthropods, it is important to study the expression of cnc in non-mandibulates. As there is a reasonable chance that the function of cnc will not be the same in these outgroups (as they do not have a mandible), it is important to study as many outgroups as possible to have a better idea of the ancestral developmental role, if any, of cnc.
It has to be shown how cnc functioned in outgroups to the mandibulates. For example, it has to be shown that cnc does not in some way differentiate or pattern the segment homologous to the mandibular segment in non-mandibulates.
Unfortunately, I was unable to determine a specific expression pattern in the spider Achaearanea in order to show whether or not cnc has a first leg segment domain of expression in the spider. It is possible that cnc does not have a developmental role in the spider, either because embryonic roles of cnc function were never acquired in the chelicerate lineage or because developmental functions have been lost. To prove that a gene, like cnc, does not have a developmental role (as opposed to ubiquitous expression) by in situ hybridisation is difficult without additional evidence as there will be no specific expression pattern that would confirm that the in situ hybridisation experiment was successful.
One means to do this experimentally is to shown that levels of cnc transcription are maintained at constant levels throughout embryogenesis by quantitative PCR. If it is shown that cnc is not upregulated at any stage of embryogenesis, this would provide some support for the notion that cnc has no embryonic developmental role.
Transcripts of At cnc were detected at all stages of embryogenesis by RT-PCR (with appropriate controls proving that there was no genomic DNA contamination), however, this is to be expected as cnc has a conserved role across metazoans in oxidative and xenobiotic stress responses. It is necessary to use a quantitative method of detecting transcript abundance to show that a gene does not show stage specific upregulation, and therefore a likely developmental role.
I was unsuccessful in my attempts to clone a homologue of cnc in an onychophoran, which is a more distantly related outgroup to the mandibulates.
The question is still open therefore as to whether cnc homologues in nonmandibulate arthropods have a developmental function, and it is not yet known whether cnc is expressed in the homologous segment to the mandibular segment, the first leg segment of chelicerates and onychophorans. Skn-1, the homology of cnc in Caenorhabditis elegans, has been shown to have embryonic developmental role in patterning mesodermal derived structures like the pharynx and body-wall muscle and intestines derived from endoderm (Bowerman et al., 1992; Walker et al., 2000)
Characterization of the mandibular endite.
In order to study the developing mandible in Tribolium embryos I studied the expression of genes i) that are expressed in the endites, ii) a gene that marks the developing segments of appendages and iii) the PD domain genes to identify appendage segments.
By studying genes that are expressed in the endites, Tc prd and Tc dac, I demonstrated in chapter two that the inner and outer lobes of the mandible correspond to the developing molar and incisor process respectively. Comparison of the expression of these genes in the mandible to the maxillary endites showed that these two lobes are very likely to be derived from one endite as the expression pattern in the mandible resembles the expression pattern seen in each endite of the maxilla.
Understanding that the gnathal edge is derived from one endite enables comparison of the gnathal edge, as a single endite, to other endites on other appendages. Machida hypothesized that the incisor and molar processes were homologous to the galea and lacinea respectively, which implies that the incisor and molar processes are derived from two separate endites. As the incisor and molar processes are derived from one endite as shown in chapter 2, Machida’s hypothesis of mandibular endite homology is incorrect.
Locating the endite to a particular segment, creates the possibility of homologizing it to endites present on segments of other appendage types.
7.2 The mandibular subcoxa I was interested in investigating any evidence of segmentation of the mandible by studying the Notch signalling pathway, in particular by studying the expression of Tc ser which is expressed in the distal part of each appendage segment. These results were shown in chapter three. Machida studied SEMs of the developing mandible of a bristletail and found evidence of a vestigial subcoxal/coxal segment boundary and suggested the subcoxa was homologous to the cardo of the maxilla (Machida, 2000).
Study of the expression of Tc ser in Tribolium showed that the mandible is indeed divided into a subcoxa and coxa segment. These segments do not form a joint in the larval or adult mandible. The presence of a ring of Tc ser expression in the mandible provides molecular evidence of a subcoxal/coxal division in the embryonic mandible.
This result supports Machida’s hypothesis that the mandible is subdivided into two segments. The mandible gnathal edge derived from the endite – both incisor and molar processes - is therefore present on the more distal coxal segment of the mandible.
Expression of Tc ser suggests that there is a subcoxal segment of the Tribolium mandible. This subcoxa has fused with the coxa to form the mandible present in Tribolium. Other apparently unsegmented mandibles may, like Tribolium, also have hidden subcoxal segments in their embryonic stages. Only some diplopods (millipedes) possess mandibles with an obvious subcoxal segment in postembryonic stages. Apart from the diplopods, there is no visible subcoxa present on the mandibles of any other mandibulate arthropod. It remains to be tested whether there is evidence of a hidden subcoxal segment present in other mandibles (shown in fig. 7.1G,D and fig.7.2C) by studying the Notch signalling pathway. Interestingly, if the presence of a mandibular subcoxa is primitive the segmented diplopod mandible could represent the ancestral state although this would require multiple independent losses of the subcoxa in other mandibulates.
Homology of the subcoxa and coxa of the mandible and maxilla
The presence of a mandibular subcoxa means that it is possible to homologize it to the cardo of the maxilla. This is assuming that both subcoxal segments are primitive characters that were present in the ancestral gnathal appendage (see fig.