The zinc finger protein TFHIA, a positive transcription factor of the SS RNA gene, binds to an internal control region of 50 nucleotides. Two modes of binding have been considered for the TFIIIA-DNA complex, one of which has been proposed on the basis of nuclease and chemical protection experiments and the other on model building. Since then, evidence has accumulated on the structures of individual components of the complex-for example, zinc finger polypeptides studied by NMR and a segment of the binding site analyzed by x-ray crystallography, but no high-resolution structural data on the TFIIIA-DNA complex itself are available. Probes used previously to study the TFIIIA-DNA complex do not react with every nucleotide ofDNA, unlike hydroxyl radical, which cleaves DNA at every backbone position. We describe here the quantitative analysis of high-resolution hydroxyl radical footprints and suggest how the array of zinc fingers might interact with the double helix.Transcription factor IIIA (TFIIIA) is the archetype of a class of proteins that use the zinc finger motif for DNA binding (1)(2)(3). No complexes of zinc finger proteins with DNA have so far yielded to structural analysis, so we must rely on mutagenesis (4-6), chemical and nuclease probing in solution (7-9), and model building (10, 11) to suggest how binding might occur. The nine zinc fingers ofTFIIIA bind to -"50 base pairs (bp) of the internal control region of the 5S ribosomal RNA gene of . DNase I (6,15) and hydroxyl radical (15) footprints of TFIIIA deletion mutants on the 5S RNA gene show that TFIIIA protects the internal control region in a collinear fashion. Two models for the TFIIIA-DNA interaction have been considered (9): (i) the "wrapping around" model, where successive zinc fingers make structurally equivalent contacts in the major groove without crossing over the minor groove (10), and (ii) the "alternating" model, where alternate fingers bind on one face of the DNA in an equivalent manner to the major groove, so that successive minor grooves must be crossed (9). Protection experiments that use DNase I and DNase II (8), micrococcal nuclease, and dimethyl sulfate (9) suggest that TFIIIA binds in the major groove, mainly to one face ofthe DNA helix and, therefore, support model ii. However, a model of the zinc finger polypeptide proposed by Berg (10), which agrees in its essentials with recent two-dimensional NMR structure determination (16, 17), has been considered to favor the "wrapping around" mode of binding to DNA. To help settle this question we have analyzed quantitatively hydroxyl radical footprints of the intact TFIIIA-DNA complex. We discuss the feasibility of the two models and propose a mode of binding consistent with the results from all of the studies on the TFIIIA-DNA complex.Quantitative Hydroxyl Radical Footprinting. Footprinting, i.e., protection against chemical or enzymatic attack, has been used to determine the DNA-binding sites of many drugs and proteins (18)(19)(20). The use of hydroxyl radical as the cleavage age...