Distinguishing specific from nonspecific complexes on southwestern blots by a rapid DMS protection assay.

Southwestern blotting has been extensively used in studies aiming to the identification of sequence-specific DNA-binding activities in cellular extracts (1). While the sensitivity of this methodology aids to the detection of factors with low, specific DNA-binding affinities, the odd equilibrium and kinetics of the interaction of a DNA segment with a matrix of immobilized protein molecules, allow the simultaneous detection of additional proteins with high, nonspecific DNA-binding affinities, that are present in excess to the probing DNA sequence. Binding to these latter proteins is sometimes abrogated by competition with unlabelled specific DNA (and/or is not affected by unspecific DNA) and manipulating parameters such as ionic strength and pH may not enable the differentiation between specific and nonspecific interactions. In this report we utilized dimethyl sulfate (DMS) in a rapid in situ protection assay (DNA elution/ cleavage/purification performed in a single step) to distinguish specific from superfluous DNA—protein complexes on blotting membranes. Apart from providing complete information about a particular DNA-protein interaction (molecular mass of the protein component and protein-contact positions on the DNA), the small size and hydrophobic nature of this chemical reagent render it perfectly suitable for an assay of this kind because of the distinct microenvironments existing in the binding interface of specific and nonspecific complexes. Nonspecific DNA binding to immobilized targets is driven electrostatically (multiple, strong salt bridges with the polyanionic backbone of phosphate and sugar groups, lacking directionality) and by the release of counterions and water bound loosely to it. This transfer of ordered water and hydrated cations to bulk solvent strengthens the ionic DNA-protein interactions by making a favorable contribution to the reaction entropy, and creates an extended, virtually hydrophobic interface readily penetrable by DMS, which will methylate all accessible reaction sites. On the other hand, specific DNA binding involves primarily gradual formation of oriented hydrogen bonds and van der Waals contacts; as water is recruited from bulk solvent to correct 'imperfections' in complementarity, it builds a three-dimensional network linking DNA and protein which increases the number of direct and water-mediated hydrogen bonds, hence reinforcing binding and further driving the association. A 'compact' hydrophilic interface is thus generated which will exclude DMS by severely impairing its diffusional freedom, limiting its activity to only distorted and exposed base nitrogens. Conceivably, the methylation/cleavage pattern of the interface in DNA derived from nonspecific complexes will be markedly enhanced compared to its unbound counterpart, whereas that from specific complexes will exhibit protections and enhancements due to the specific contacts and specific binding-triggered perturbations, respectively. Although DNase I can in principle substitute for DMS as a probe in this assay (2), its bulky dimensions, mode of target searching and binding to cleave (2), and the requirement for Mg (which often stabilizes both specific and nonspecific complexes) highly reduce its potential to 'sense' the aforementioned differences. Total or partially purified nuclear proteins (100—200 fig) are resolved in a preparative SDS polyacrylamide gel and electrotransferred on to a supported hydrophilic membrane [i.e., Hybond-N (Amersham); see below]. The filter-immobilized proteins are renatured under established conditions (3, 4, 5) and probed with 10—20/tg (~2—4X10 cpm) of an asymmetrically P-labelled restriction fragment or oligonucleotide bearing the putative binding site. The membrane sheet is exposed (wet) to X-ray film to locate regions of radioactive signal which are then cut with a razor blade and immediately immersed in 500 /J of 20 mM Tris-HCl, pH 8.0. After a 10' equilibration at 25°C, the radioactivity retained on the strips is measured for Cherenkov counts and a comparable amount of cpm of the same probe in solution is added to 250 /tl of 20 mM Tris-HCl, pH 8.0 in a separate tube and put aside awaiting treatment. The membrane strips are transferred with forceps into 500 pi of freshly made 0.18% (v/v) DMS (reagent grade) solution in 20 mM Tris-HCl, pH 8.0 and incubated for 25 -40" at 25°C (to avoid variations in the exposure of samples to DMS, during the DMS addition step no more than one sample should be handled at a time). Although the longer the DNA probe the shorter the length of incubation to achieve the desirable single modification per molecule, we have found (employing probes of various lengths) that the above combination of reaction time and DMS concentration ensures that single-hit kinetics is followed (the percentage of uncut DNA in the final analysis is greater than 70%, Figure 1) and sufficient cleavage for a good signal-to-noise ratio is generated. Due to the increased kinetic stability of a filterimmobilized DNA -protein complex (reversible binding to even low-affinity proteins is enhanced because the excess unbound DNA fraction has been washed out and hence is not exchangable), these reaction parameters are not determined by the dissociation rate of the complex, which interferes with DMS protection assays performed in solution. Methylations are quenched by quickly removing the strip (with forceps) and immersing it in a tube