Scanning force microscopy of Escherichia coli RNA polymerase.sigma54 holoenzyme complexes with DNA in buffer and in air.

Scanning force microscopy (SFM) was used to visualize complexes of Escherichia coli RNA polymerase.sigma54 (RNAP.sigma54) and a 1036 base-pair linear DNA fragment containing the glnA promoter. In order to preserve the native hydration state of the protein-DNA complexes, the samples were injected directly into the SFM fluid cell and imaged in buffer. With this protocol, an apparent bending angle of 26(+/-34) degrees was determined for the specific complexes at the promoter. The bending angle of the unspecifically bound RNAP.sigma54 showed a somewhat broader distribution of 49(+/-48) degrees, indicating the existence of conformational differences as compared to the closed complex. In about two-thirds of the closed complexes, the RNA polymerase holoenzyme was located in a lateral position with respect to the DNA and the bend of the DNA was pointing away from the protein. This conformation was consistent with the finding that for the complexes at the promoter, the apparent contour length was reduced by only about 6 nm in buffer as compared to the free DNA. From these results we conclude that in the closed complex of RNAP. sigma54, the DNA was not wrapped around the polymerase, and we present a model for the trajectory of the DNA with respect to the RNA polymerase. The images acquired in buffer were compared to samples that were washed with water and then dried before imaging. Two artefacts of the washing and drying process were detected. First, extensive washing of the sample reduced the number of the specific complexes bound at the promoter (closed complex of RNAP.sigma54) from about 70% to 30%. This is likely to be a result of sliding of the RNAP.sigma54 holoenzyme along the DNA induced by the washing process. Second, the apparent DNA shortening of the contour length of RNAP.sigma54-DNA complexes at the promoter as compared to the contour length of the free DNA was 22 nm for the dried samples as opposed to only 6 nm for the undried samples imaged in buffer. This suggests an artefact of the drying process.