HENRY M. SOBELL

LENGTHENING AND UNWINDING OF DUPLEX DNA IN COMPLEXES WITH recA PROTEIN. A. Stasiak, E. DiCapua and T. Koller, Institute for Cell Biology, Swiss Federal Institute of Technology, Hoenggerberg, CH-8093 Zurich, Switzerland. Cold Spring Harbor Symposia on Quantitative Biology 47, 811-820 (1983).

This elegant electron microscopic study of the recA protein of Escherichia coli complexed to circular double-stranded nicked relaxed DNA and to negatively superhelical DNA molecules establishes the existence of a DNA form clearly different from A- or B-DNA.

RecA protein, in the presence of the nonhydrolyzable ATP analog, adenosine 5' gamma- thiotriphosphate, binds to nicked relaxed circular DNA with high cooperativity, forming an organized helical structure in which double-stranded DNA is stretched one and one-half times (see Figure 1a). The complex can be visualized with platinum-shadowed specimens (or with negative staining using phosphotungstic acid) and appears as a cross-striated (or as a zig-zag) protein-DNA complex.

There are 267 cross-striations on the plasmid, which contains 4,961 base pairs. This means that each striation (estimated to contain about 6.2 recA monomers) covers 18.6 base pairs, and that the helical structure (inferred from examination of the symmetrical diffraction patterns obtained from scanning transmission electron microscopic images of negatively stained segments of the complex) contains 18.6 base pairs per turn in approximately 99 Angstroms.

This helix repeat agrees very closely with the value of 95 Angstroms, obtained by dividing the average contour length of each circular complex (25,530 Angstroms) by the number of cross-striations (267).

The binding of recA to double-stranded DNA also unwinds superhelical DNA. This is shown in Figure 1d, which shows a negatively superhelical DNA molecule bound by the recA protein to form a gentle left-handed superhelical structure, until further binding is inhibited by the appearance (and the energetics) of a left-handed positively superhelical DNA tail.

One can quantitate this measurement by studying the binding of recA to negatively superhelical DNA molecules with known linking numbers. This is achieved by using the nicking-closing extract from rat liver on nicked relaxed circular DNA molecules, in the presence of increasing concentrations of ethidium bromide. After purification, it is possible to separate topoisomers having different linking numbers by gel electrophoresis, and to analyze their binding to recA (see Figures 4 and 5).

Topoisomers having smaller linking numbers can also be estimated by studying the percent covering of DNA molecules by recA as a function of ethidium bromide concentration in the nicking-closing reaction, in the presence of known amounts of circular DNA.



One can readily see that there is direct correlation between the percent covering of DNA by recA and the linking number. The smaller the linking number, the greater the percent covering of DNA by recA.


Notice that at 100% covering, the estimated linking number is 267. One can calculate the unwinding of DNA associated with the binding of recA to DNA in such a structure, assuming B-DNA in solution has 10.5 base pairs per turn (therefore, having a linking number of 472 for 4,961 base pairs). The unwinding predicted is [(472-267) x 360]/ 4,961 = 14.9 degrees per base pair, or about 30 degrees every other base pair.

Their model for the recA-DNA complex is shown below (Figure 8).