We have identified two collagen genes linked to
myo-3: col-? Iies approximately 80 kb to the left of
myo-3, while col-l lies approximately 120 kb to its right. We have determined the DNA sequence of 1771 nucleotides which contains the coding region of the col- ? gene. A putative initiating met codon occurs at position 701 and is immediately preceded by a potential splice acceptor site (TTTTAGG) with no corresponding upstream donor site, suggesting that the co 1- ? message may have a trans-spliced leader. The coding region continues uninterrupted by introns to a stop codon at position 1646, predicting a 315 aa protein with an isoelectric point of 3.59 and a typical cuticle collagen structure. Comparisons of the three conserved cysteine-rich domains in col- ? with these domains in other C. elegans collagen genes reveal that this new collagen belongs in the
col-7, col- 8, col-l 9 group. However, the COO- domain of col- ? contains 41 aa, while other members of this group have only 14 or 15 aa. A search for conserved DNA sequences in the S' flanking region of the gene reveal homologies with two repeats found upstream of most collagen genes and with one repeat found upstream of only dauer-specific collagen genes. Linked to
myo-3 are four cuticle-defective loci (
sqt-3,
rol4, ali-l,
lon-3). To determine which of these genetically defined loci are encoded by col-? and col-l, we have generated 500-600 bp PCR fragments of the COO- end of each gene from eight mutants. We have cloned these fragments and are currently sequencing them. A missense mutation (Gly to Glu) in the triple helical domain of col-l has been detected in DNA from
sqt-3(
sc63), although we need to verify that this G to A transition was not generated during the amplification or cloning.