We have isolated the
sqt-1 gene by a combination of Tc1 tagging and chromosome walking/deficiency mapping. In the last WBG we reported on five strains derived from TR679 that carry new recessive left roller alleles of
sqt-1. Four of the strains have been carefully analyzed for useful new Tc1's, Tc2's, and Tc3's. Only one strain, BE143
sqt-1(
sc143), had a new element that could not be shown to be irrelevant ( Tc1-BE143). This Tc1 element was shown to be the result of a recent transposition, and was not separated from
sqt-1 after recombination with the flanking markers
unc-4 and
lin-29. However, it seemed unlikely that it was actually inserted into
sqt-1 because: it has the same phenotype as the other four TR679 derived strains, recessive left- roller, which is not that expected for a
sqt-1 null; none of the other three characterized strains have any detectable genomic alterations in the region where Tc1-BE143 is inserted; no transcripts are detectable using DNA flanking Tc1-BE143 as a probe. The
sqt-1(
sc143) left roller allele was put back into a mutator background (Mike Finney's HH*6 strain) to look for non-roller revertants.
mut-2;
sc143 animals were placed on race plates - bacterial lawn on one side, worms on the other. Non-roller animals generally cross the plate faster than rollers, so screening the lawn enriches for non-rollers. Four independent non-roller animals were recovered, backcrossed to remove
mut-2, and genetically tested. Two of the resultant strains still carry
sc143, but also carry an unlinked recessive suppressor. One strain carries an unlinked, dominant left roller allele (?
sqt-3 ?), that supresses
sc143. Finally, one strain ( CH1) acts like a
sqt-1 null strain, it fails to complement
sqt-1 recessive alleles, and gives a wild-type phenotype when heterozygous with other putative
sqt-1 null alleles. So far, only CH1 has been further characterized (see below ). A cosmid contig was established from Tc1-BE143 (Thanks, Alan and John) and whole cosmid DNA's have been used as probes to genomic DNA's from several of the mnDf strains in the region. Judging from relative hybridization intensities, Tc1-BE143 lies under Df's 77 and 87, but is outside of Df's 75, 76 and 86. This indicates that it lies near, but not in the
sqt-1 gene. (Note: this approach has worked well with strains carrying small to moderate size deficiencies, but has been much harder to interpret when strains with large deficiencies (e.g., mnDf's 83 and 89) are used. This problem is probably due to growth disadvantage of the heterozygotes carrying large deficiencies.) Both ends of Df86, the left ends of Df's 75, 76, and the right end of Df77 have been identified in the contig as restriction fragments of altered size (see map).
sqt-1 must lie within the overlap region of these four deficiencies, since it is not complemented by any of them. mnDf86 is a deletion of approximately 85 kb, and the overlap region of the four deficiencies is approximately 40 kb. Cosmid DNA's were probed under low stringency conditions with
col-2 and a single collagen gene was found in the contig, this collagen is contained on a 3.6kb EcoRI fragment, in cosmid C25B12, and is situated within the region of deficiency overlap. The CH1 strain has a Tc1 element inserted into this same 3.6kb EcoRI fragment (Tc1-CH1). We are currently determining whether the Tc1 insertion is in the collagen gene, but are willing to predict that the collagen gene is
sqt-1. None of the other TR679 or HH*6 derived
sqt-1 alleles have any detectable alterations in this region of the genome. So, only one out of six
sqt-1 alleles generated in mutator strains is due to the presence of a transposable element. Since
sqt-1 alleles are generated in these strains at high frequencies (10+E-4), we are interested in knowing what the mutator generated alleles are the result of. We have performed preliminary Northern blot analysis using the 3.6kb EcoRI fragment as a probe under very high stringency wash conditions ( 0.03 M Na+, 20% formamide, 65 C). We detect a transcript of approx. 1.2kb in dauer, L4, and adult molt RNA's, no transcript Is detected in late embryo RNA. These results are consistent with the temporal expression of the
sqt-1 phenotype, but we cannot rule out possible cross-hybridization. We are in the process of sequencing the wild- type and mutant alleles of
sqt-1 and will be able to generate gene- specific probes from the sequence. Also, we are making RNA's from CH1 to determine if the 1.2kb transcript is present in this strain. The evidence so far indicates that
sqt-1 is simply another member of the collagen gene family, though obviously a rather crucial one, at least when it's there. The map below indicates the position of
vim-1, which is the putative vitelline membrane protein gene isolated by Paolo Bazzicalupo. It seems possible that
vim-1 could actually be
ooc-3, since
ooc-3 lies quite close to the left of
sqt-1 and has a phenotype that could result from an abnormal vitelline membrane. It appears, however, that
vim-1 lies under mnDf77, which is supposedly not the case for
ooc-3. This question merits further investigation. [See Figure 1]