Characterization of a herpes simplex virus sequence which binds a cellular protein as either a single-stranded or double-stranded DNA or RNA.

Earlier we reported that herpes simplex virus 1 DNA contains a sequence which binds a host protein in a sequence-specific manner as either a single-stranded or a double-stranded DNA or RNA and that this sequence is located in a transcriptional unit whose RNA traverses the origin of viral DNA replication (OriSRNA) (R.J. Roller, L. McCormick, and B. Roizman, Proc. Natl. Acad. Sci. USA 86:6518-6522, 1989). The protein reacts with both DNA and RNA in band-shift assays and protects the single-stranded RNA sequence from digestion by RNase. We report that the minimal cognate sequence required for these interactions consisted of [N(GTGGGTGGG)2(N less than or equal to 10)]. The ninemer repeat sequence was located at nucleotides -1 to -18 relative to the transcription initiation of the major species of OriSRNA. The activity of the cognate sequence required at least three guanines between thymines and tolerates the insertion of additional thymines, but it was inactivated by the insertion of adenines or by the substitution of some of the guanines with cytosines in one repeat. Replacement of the 10 3' nucleotides has no effect on binding activity, whereas deletion of these sequences abolished it. Among the related sequences with no demonstrable binding activity were some telomeric sequences which interact with known cognate proteins. The electrophoretic mobility of the herpes simplex virus cognate sequences in nondenaturing gels suggests that they may be able to form higher-order structures, but the conditions under which they were formed were different from the optimal conditions for binding the protein. UV light cross-linking studies of labeled RNA-protein complexes following digestion with RNases indicated that the electrophoretic mobility of the protective activity corresponded to that of a protein with an apparent molecular weight of 100,000.