The molecular basis of Mullerian mimicry

Mimicry among butterfly species is a paradigm example of evolution and adaptation, but its molecular genetic basis has never been studied. Here we will use modern genomic techniques to identify genes controlling mimicry patterns in the genus Heliconius, a well-known evolutionary model system. What is particularly striking is our recent discovery that three Heliconius species use the same genetic region to control variation in colour patterns. In two of the species, H. melpomene and H. erato, these are convergent patterns that evolved through mutual mimicry. In the third species, H. numata, the patterns are very divergent and mimic species from a different sub-family, the Ithomiinae. Furthermore in H. numata the locus is a 'supergene' i.e. a single locus that controls all aspects of wing pattern and is polymorphic within populations (many different-looking forms are found flying together in the same population). Thus, the same region of the genome has somehow evolved to control both convergent and divergent patterns. Here we will use markers already developed in H. melpomene to identify the region of interest in the other two species and obtain DNA sequence for a sufficiently large region that we are sure contains the gene of interest in all three species. This will allow a comparison of genome organisation in this region between the three species and a test for reorganisation of the genome that may underlie the evolution of wing pattern. In particular there is a long-standing hypothesis that a 'supergene' such as that seen in H. numata might arise by gradually moving several genes together (the evolutionary advantage of this is that unfit intermediate patterns are avoided in wild populations). This study will allow the first explicit test of this hypothesis at a molecular level. We will then determine which locus in this region is controlling the yellow band pattern in H. melpomene. We will first identify all the possible genes in this region as candidates for being the patterning locus, and determine which ones are expressed in developing wings and whether they show any patterns that correlate with the wing patterns. We therefore aim to identify the Yb gene that controls a band on the hindwing of H. melpomene, and compare the organisation of the genome in the region of this locus between the three species. This work offers an unusual opportunity to link a trait whose importance in natural populations is well studied with a change at the level of DNA sequences. It will represent the first time that a gene controlling natural variation in such a well studied trait has been characterised at a molecular level, and will stimulate future developmental and population genetic studies of mimicry in Heliconius. In particular, the results will facilitate future studies of how strong selection on a gene affects variation in the surrounding genome.

Mimicry among butterfly species is a commonly cited example of evolution and adaptation and Heliconius are perhaps the best studied example, but the molecular genetic basis of mimicry has never been studied. Here we will first clone a region containing several linked elements that control yellow patterns in H. melpomene. BAC clones spanning the region of interest will be identified and then fingerprinted to assemble contigs. Linkage mapping will be used to confirm the position of patterning genes relative to these physical markers. A tile path across the contig will then be chosen, sequenced and assembled. We will annotate these sequences to identify all the possible genes in this region that are candidates for being the patterning locus, and determine which ones are expressed in developing wings and whether they show any patterns that correlate with the wing patterns using transcriptomics and RNAi gene knock-down experiments. We expect to demonstrate spatial patterns of expression of the candidate locus that correlate with wing patterns, or alternatively temporal patterns of expression of the candidate locus during wing development that can be disrupted using RNAi experiments. Either way, we will identify a specific candidate for the Yb gene that controls a band on the hindwing of H. melpomene. We have recently shown that a homologous region controls pattern segregation in two further species, H. erato and H. numata. In H. numata the locus is a 'supergene' i.e. a single locus that controls all aspects of wing pattern and is polymorphic within populations We will then conduct a comparative analysis of this region in the three species, by repeating the chromosome walk in H. erato and H. numata and sequencing a similarly sized tile path in these two species. This will allow us to test whether genomic rearrangements have played a role in pattern evolution, as predicted by the theory of 'supergene' evolution.