The
daf-2 receptor tyrosine kinase,
age-1 PI 3-kinase, and
daf-16 forkhead genes form a nematode insulin-like signalling pathway(1-4), but potential insulin-like ligands that may regulate this pathway have not yet been elucidated. We performed a TBLASTN search of the C. elegans genomic and EST databases using the A chain of both human insulin and the silkworm insulin-like peptide bombyxin-D1 (BBX-D1), and identified 5 different nematode insulins. Further BLAST searches were performed with these worm insulins, leading to the discovery of a total of 15 insulin-like (ins) genes to date in C. elegans. One of the ins genes was already available as an EST, and we have shown using RT-PCR that an additional 13 of the predicted ins genes are expressed. The ins sequences were aligned by both the Clustal-W and the Divide-and-Conquer(5) multiple sequence alignment algorithms, with manual adjustment of the A and B chains. Based on these alignments, the ins genes can be divided into 6 groups. Group 1 (7 genes) has 8 cysteines, groups 2 (2 genes), 3 (2 genes) and 4 (2 genes) have 6 cysteines but group 3 has a longer A chain and group 4 has a longer B chain. Groups 1, 2, 3 and 4 all possess a short C-peptide and probably are not cleaved following pro-insulin folding. Group 5 (1 gene) possesses a novel repeat domain structure that suggests that it is highly derived. Group 6 (1 gene) appears to be a pseudogene. Three-dimensional modeling of the Group 1, 2, 3 and 4 ins genes was performed by threading onto the known structure of IGF-I (which has a small C-peptide) using the SwissPdbViewer. This analysis shows that these worm insulins possess a tertiary structure similar to mammalian insulin-like peptides and that their cysteines are properly aligned for the formation of disulfide bonds to maintain the structure. A neighbor-joining phylogenetic analysis of the insulin superfamily shows that the worm insulins form distinct phylogenetic clades, suggesting that the gene duplications leading to this family are a recent event. However, the cladistic relationship of the ins intron sequences is different from the relationships of their coding regions. We believe that this difference in the phylogenies can be explained by gene conversion, which could also explain the close phylogenetic relationship of the worm insulins. The majority of the worm insulins are found in clusters: two clusters of four and two ins genes in group 1 and a cluster of the two group 2 genes, while a group 3 member is within 700bp of the group 5 gene. Gene clustering is also characteristic of the multiple insulin-like genes found in silkworms(6) and freshwater snails(7). Why are there multiple copies of insulin-like ligands in these invertebrate species? In the silkworm, a large increase in the titer of BBX proteins after pupation lasts until eclosion, suggesting that large quantities of BBX are necessary for this metamorphosis, and it has been suggested that the increased copy number of bbx genes is necessary to produce enough BBX protein. It is possible that the multiple insulin genes in nematodes are serving a similar purpose and function in the transition between developmental stages. As a first test of this hypothesis, we are currently examining the in vivo tissue expression pattern of the C. elegans insulins. 1. Kimura et al, Science 1997;277:942. 2. Morris et al, Nature 1996;382:536. 3. Ogg et al, Nature 1997;389:994; 4. Lin et al, Science 1997;278:1319. 5. J. Stoye, PhD Dissertation, Universitaet Bielefeld, 1997. 6. Konda et al, J Mol Biol 1996;259:926. 7. Geraerts et al. Experientia 1992;48:464.