CONFORMATIONAL CHARACTERISTICS OF HIGH-AFFINITY SP1 BINDING ENHANCER ELEMENTS OF HIV-LTR BY HIGH-RESOLUTION 2D-NMR

The understanding of early events in the expression of genes has vastl y improved in recent years with the identification of a variety of gen e- and sequence-specific DNA binding transcription factors. One such p rotein, Sp1, has been implicated in activating transcription of variou s cellular and viral genes including those of HIV, and SIV types of re troviruses. The basic recognition site for Sp1 has been identified as variants of a 10 base-pairs long GC-rich DNA, often containing a hexan ucleotide segment 5'-GGGCGG (termed CIC-box). However, variations in b oth the relative protein-DNA binding affinity and the nature of bindin g sequences have been noted. Two-dimensional IH-NMR experiments (500 M Hz) were employed for conformational studies of two decadeoxyribonucle otide duplexes, d(GAGGCGTGGC)-d(GCCACGCCTC), termed Sp1-III, and d(GGG AGTGGCG)-d(CGCCACTCCC), termed Sp1-I. These are two of the highest aff inity Sp1 binding sites and consist of diverse positioning: of the tri - and tetranucleotide segments GAG, GTG, GCG, GGCG, GTGG and GGAG, tha t occur frequently in other Sp1 binding sites as well, and may form sp ecific contacts with the protein. Phase-sensitive nuclear Overhauser e nhancement (2D-NOESY and MINSY) and correlation (COSY) spectra were ob tained for the assignment of the exchangeable and nonexchangeable prot ons in a sequence-specific fashion. As a prelude to determination of t he detailed solution structures of the selected sequences, numerous st ructural constraints were obtained from angle-dependent coupling const ants and relative intensities of distance-dependent intra- and internu cleotide NOEs. Overall, each duplex adopts a structure similar to B-DN A with predominantly C-2'-endo/S-type sugar conformation and anti-glyc osidic torsion angles. A selective disruption of sequential NOE connec tivities at the GAG-CAC and GTG-CAC steps, irrespective of the flankin g sequence, suggests that conformational changes at these sites may ac t as unique determinants of sequence specific recognition/binding of S p1. Implications for a specific inhibition of Sp1-mediated transcripti on by minor groove binding class of drugs, designed to recognize GC-ri ch sequences, are also briefly discussed.