Mutants in the C. elegans gene
unc-89 have disorganized myofibrils in which thick filaments are not organized into A-bands, and for most alleles, there are no M-lines.
unc-89 encodes a giant 732,000 Da polypeptide that is a member of the intracellular branch of the Ig superfamily (Benian et al., 1996). UNC-89 is composed of 53 Ig domains and several signal transduction domains, including a dbl-homology (DH) domain near its N-terminus. Recently, a possible human analog of UNC-89 called "obscurin" has been identified (Young et al., 2001). Teichmann and Chothia (2000) and WormBase predicted that gene C24G7.5 encodes an Ig superfamily member that is expressed in muscle. The predicted initiator methionine codon for C24G7.5 is only 2.7 kb downstream of the 3' end of
unc-89. With a combination of GeneMark.hmm exon predictions and RT-PCR, we have been able to demonstrate that C24G7.5 is not a separate gene, but rather a set of alternatively spliced exons for
unc-89. Splicing treats the 680 bp 3'UTR and stop codon of the previous
unc-89 as an intron, and continues into the C24G7.5 exons. Whereas the previously characterized UNC-89 (UNC-89-S) is 6,632 residues, the new isoform (UNC-89-L) is potentially 8,082 residues. The C-terminal extension of UNC-89-L contains a protein kinase, 640 residues of low sequence complexity, an Ig, a Fn3, and a second protein kinase domain. The two kinase domains of UNC-89-L have highest homology to twitchin, MLCKs, and titin, especially in the large subunit. However, they show significant dissimilarity in subdomains I and II that suggest that both kinase domains might lack phosphotransferase activity. Using RT-PCR we can show that both kinase domains are included in an UNC-89-L polypeptide.