21. Vectors for Using Nested Deletions to Sequence Either Strand of Cloned DNA 

John J. Dunn, Laura Praissman, Laura-Li Butler-Loffredo, John J. McNulty, and F. William Studier 
Biology Department, Brookhaven National Laboratory, Upton, NY 11973-5000 
jdunn@bnl.gov 

Regions of highly repeated DNA are encountered frequently in human DNA and are likely to be particularly troublesome near centromeres and telomeres. Highly repeated regions are difficult to assemble correctly by shotgun sequencing, but cloned fragments at least 10 kilobase pairs long can be sequenced and assembled easily by generating an ordered set of nested deletions whose ends are separated by less than the length of sequence read from a single priming site within the adjacent vector. Assembly of the overlapping sequences is guided by knowledge of the relative length of the portion of the fragment remaining in the clone, as determined by gel electrophoresis. 

We have made a series of plasmid vectors, the pZIP series, which allow rapid generation of an ordered set of nested deletions from either strand of a cloned DNA fragment. The vectors are based on the low-copy F replicon. The size of the vector DNA has been reduced to the 4.5-kbp range by removing the 2.5-kbp sop (stability of plasmid genes) region. The resulting plasmids have the low copy number typical of F plasmids and remain stable enough to be easily maintained by growth in the presence of kanamycin, the selective antibiotic. DNA in amounts convenient for sequencing is readily obtained by amplification from an IPTG-inducible P1 lytic replicon. 

Nested deletions are generated by cleavage near one end of the cloned fragment, using commercially available site-specific endonucleases (PI-PspI, I-CeuI or I-SceI) whose recognition sites span 18-30 bp and are therefore unlikely to occur in cloned DNA fragments. Cleavage by these nucleases generates four-base 3' overhangs that are resistant to digestion by E. coli exonuclease III. A second cleavage by one of several nucleases with 8-base recognition sites leaves the end adjacent to the cloned fragment susceptible to ExoIII digestion, permitting unidirectional 3' to 5' digestion across the cloned fragment. The resulting single-strand tails are digested with S1 nuclease, the ends are repaired and ligated with T4 DNA polymerase and ligase, and clones are obtained by electroporation. An range of digestion times with ExoIII can easily produce a distribution of deletion lengths extending across the entire cloned fragment. Cleavage sites for the site-specific endonucleases are positioned in the vectors so that nested deletions can be generated from either end of an individual cloned fragment. 

Conditions for routinely generating ordered sets of nested deletions and using them to sequence both strands of cloned fragments in the 5-kbp to 15-kbp size range are being developed by sequencing fragments of human DNA from BACs.