In the four decades since Watson and Crick (1953) established the double helical nature of DNA, great advances have been made in the area of DNA chemistry, protein-DNA interactions, and gene expression. Only four years after Watson and Crick's announcement, G. Felsenfeld et al discovered that complementary homopolyribonucleotide duplexes can form a triple helical structure (1957). Yet, this field lay dormant for almost three decades, perhaps because of its "peculiar" nature (Letai et al 1988). However, the last several years have witnessed a surge in triple helix research, brought about by the recognition that a third-strand can selectively seek out a discrete sequence in the genome and deliver a covalently-bound passenger. This specificity is due to the "third-strand binding code" (Table 1) first described by the Fresco lab (Broitman et al 1986; Letai et al 1988). Moser and Dervan then showed that an EDTA/Fe2+ moiety covalently attached to a third-strand can specifically cleave DNA at the site of binding (1987). Since then, various molecular passengers, such as fluorescent markers, psoralen, and other intercalators have been attached to oligonucleotides designed for third-strand binding. One of the most important consequences of this piggy-back capacity is the eventual possibility of carrying out in vivo site-specific gene therapy.
Presently, the Fresco laboratory has developed a third-strand based methodology to target specific centromeric repetitive sequences containing homopurine runs. Eventually, these sequences will be exploited by M.D. Johnson and J.R. Fresco for isolating individual human metaphase chromosomes. Binding studies are being performed on plasmids (nicked and supercoiled), cosmids, and lambda phage DNA in order to evaluate the effects of the size of the DNA on the kinetics of binding to the short target. Preliminary results indicate that the size of the DNA surrounding the target does affect binding rates and equilibrium states when the ratios of third-strand to target is stoichiometric (MD Johnson, unpublished results).
As part of this project, an extensive database of over 300 separate sequence entries of human α-satellite DNA was created (Grasso 1994). Sixty-three percent (88,746 out of 140,564 bp) of the total α-satellite sequences evaluated were assignable to particular chromosomes, so that there are potential targets for third-strand binding within twenty-two of the twenty-four human chromosomes. α-satellite sequences are particularly good candidates for third-strand binding. They are one of the more heavily sequenced regions in the human genome and make up seven percent of the human genome (over several million base pairs). Furthermore, α-satellite sequences are chromosome specific down to nucleotide resolution and are highly repetitive (adapted from Grasso 1994).
Non-centromeric regions of the human genome have also been investigated for suitable third-strand binding sites. Such sequences are important for one of two reasons: