Differences in secondary structure between packaged and unpackaged single-stranded DNA of bacteriophage phi X174...

The single-stranded packaged genome (ssDNA) of bacteriophage phi X174 is shown by Raman spectroscopy to lack both the ordered phosphodiester backbone and base stacking, which are demonstrated for unpackaged, protein-free ssDNA. In solutions of moderate ionic strength, unpackaged ssDNA contains 36 +/- 7% of deoxyribosyl phosphate groups with conventional B-type backbone geometry [i.e., gauche- and trans orientations, respectively, for the 5'O-P (alpha) and 3'O-P (zeta) torsions], indicative of hairpin formation and intramolecular base pairing. Additionally, the bases of unpackaged ssDNA are extensively stacked. Estimates from Raman band hypochromic effects indicate that unpackaged ssDNA contains approximately 70% of the maximal base stacking exhibited in the linear, double-stranded, replicative form III of phi X174 DNA. Conversely, for the packaged phi X174 genome, ordered (B-type) phosphodiester groups are not present, and only 40% of the base stacking in RFIII DNA is observed. These results are interpreted as evidence that the substantial hairpin-forming potential of ssDNA is eliminated by specific and extensive ssDNA-protein interactions within the phi X174 virion. Comparison of the present results with studies of other packaged single-stranded nucleic acids suggests that proteins of the capsid shell (gpF + gpG + gpH) do not fully account for the conformational constraints imposed on ssDNA of phi X174. Accordingly, we propose a model for ssDNA packaging in which the small basic gpJ protein, which is packaged along with the genome, is involved stoichiometrically in binding to the ssDNA (approximately 90 nucleotides per subunit). The proposed gpJ-DNA interactions could prevent helical hairpin formation, restrict base stacking, and disfavor fortuitous base pairing within the capsid. The present analysis is based upon use of model nucleic acids of known conformation for calibration of the Raman intensity in the region 810-860 cm-1 in terms of specific secondary structures. The calibration curve allows quantitative determination of the percentage of ssDNA nucleotides for which the 5'O-P-O3' group is configured (g-,t) as in the B-form of DNA. The method proposed here is analogous to that employed by Thomas and Hartman (1973) for ssRNA and should be applicable to single-stranded DNA and to partially denatured forms of double- and multiple-stranded DNAs.