In continuing our attempts to clone the elusive nob genes
nob-1 1 and
unc-62 2 (pka
nob-5), we undertook representational difference analysis (RDA) 3. These two genes have proven resistant to standard cloning techniques, and
nob-1 maps to one of the remaining gaps in the physical map on LGIIIR. RDA is a PCR-enhanced subtractive hybridization method in which denatured fragments of "tester" DNA to which linkers have been added are allowed to reanneal in the presence of a 20- to 400,000-fold excess of denatured "driver" DNA fragments without these linkers. Reannealed double-stranded tester DNA, which should be enriched for sequences that are altered or missing in the driver DNA population, can be specifically amplified by PCR. RDA has several advantages as a cloning technique: 1) it requires only 100 ng of starting driver DNA; 2) roughly 50% of the genome can be analyzed in each round; and 3) location of the mutation is unimportant, i.e. lesions in introns, splice sites, UTR's or nearby regulatory regions are as valuable for drivers as coding region mutations. We used N2 or
dpy-11 for tester DNAs and annealed them with driver DNA from viable
nob-1(
ct230),
unc62(ct344), or
unc-22(
ct174) homozygotes. All three driver alleles are psoralen-induced mutations; therefore, the first two may be small deletions just as the third, included as a control, is known to
be4. We cut the tester and driver starting genomic DNA's with Sau3A1 and performed three rounds of selection and amplification exactly as described by Lisitsyn and Wigler 3. Each difference product resulted in at least one unique band, from which we isolated and sequenced subclones. The 350bp
unc-22 difference product corresponded to a sequence from the
unc-22 gene, showing that the method works. Analysis of the
unc-62(
ct344) difference products is underway. From the 600bp
nob-1(
ct230) difference product we sequenced twelve subclones; of these, four appeared to be the same sequence, which had no exact match in the DNA sequence database but generated hits in the Kohara EST database. The others represented a variety of sequences, only one of which appears to represent a known gene, the 18S rRNA sequence on LGX. Given that ribosomal genes tend to be clustered in multiple copies, it is conceivable that such a region might be prone to rearrangements that could lead to polymorphisms. In PCR experiments using primers specific to the multiply represented sequence, we found that a fragment of the expected size could be amplified reproducibly from single N2 L1's and the nonviable L1's resulting from the two lethal alleles
nob-1(
ct223) and
nob-1(
ct351), but not from viable
nob-1(
ct230) L1's. We showed that the RDA product represented a sequence in the
nob-1 region using PCR experiments with single nonviable embryos homozygous for one of three previously mapped deletions. Using the RDA fragment as a probe for Southern blots, we observed a clear polymorphism consistent with an approximately 4.5kB deletion in DNA from animals heterozygous for the lethal allele
ct223, a different allele than that used to obtain the RDA probe. Based on this evidence, the probe is very likely to be
nob-1-specific. It gave no signal when used to probe the standard YAC grid, consistent with location in a gap, and we have sent it to Alan Coulson for further physical mapping. Cloning cDNAs recognized by the probe in an early embryonic cDNA library and identification of the predicted gene product are in progress. 1. Van Auken, K. M., L. Edgar, and B. Wood (1997) International C. elegans Meeting Abstracts, p. 198. 2. Weaver, D.C., L. Edgar, and B. Wood (1997) International C. elegans Meeting Abstracts, p. 177. 3. Lisitsyn, N., N. Lisitsyn, and M. Wigler (1993) Science 259: 946-951. 4. Yandell, M. D., L. G. Edgar, and W. B. Wood (1992) Proc. Nat. Acad. Sci. 91: 1381-1385.