Background

Nuclease/chemical footprinting is a classic method to probe protein-DNA interactions (Galas and Schmitz, 1978; Sasse-Dwight and Gralla, 1989; Hampshire et al., 2007). In this technique, the binding of a protein to a particular region of DNA inhibits (protection) or increases (enhancement) the ability of the nuclease or chemical to cleave the DNA. Consequently, if using 5’-end labeled DNA, a specific pattern of protection/enhancement will be revealed by incubation of the protein-DNA complex with the cleaving reagent and the subsequent electrophoresis of the DNA on a sequencing gel. However, in order to obtain a clear, interpretable pattern, the protein must bind the DNA relatively tightly during the cleavage process. If the protein(s) of interest binds to the DNA weakly and/or if the protein has a fast off-rate, traditional footprinting methods are likely to fail. Furthermore, sometimes multiple complexes can be formed with differing protein-DNA contacts. In this case, the footprint pattern will only reveal the full ensemble of interactions rather than those associated with a discrete complex.

Here, we have developed a straightforward method to treat protein-radiolabeled DNA complexes with DNase I or KMnO₄ followed by isolation of the bound complex using an Electrophoretic Mobility Shift Assay (EMSA). In this assay, the DNA is incubated with the protein(s) of interest, and free DNA is separated from protein-bound DNA by electrophoresis on a native gel. This separation occurs because the protein-DNA complex migrates more slowly than the free DNA on the gel. The protein-DNA complex of interest is then extracted from the excised polyacrylamide gel slice, and the DNA is isolated and electrophoresed on a sequencing gel to determine the footprint. In many instances, this protocol allows users to obtain clear footprint patterns even when investigating less stable and/or weakly bound complexes. For DNase I reactions, the protection pattern with or without protein(s) indicates protein binding site(s) while the hypersensitivity site(s)/enhancement pattern represents DNA bending, kinking, and/or looping. For KMnO₄ reactions, enhancement bands represent regions containing single-stranded thymine residues, such as within an open transcription bubble. However, the procedure could be adapted to other types of enzymatic or chemical reagents, which would yield different types of patterns.

In the lab, we have successfully used this protocol to obtain clear footprints of protein-DNA complexes for transcriptional activators in two different bacterial systems: Bordetella pertussis BvgA (Boulanger et al., 2015) and Vibrio cholerae VpsR (Hsieh et al., 2018 and 2020) in the presence and absence of RNA polymerase (RNAP). In the case of VpsR, we separated the protein-DNA complexes formed in the presence or absence of the small molecule cyclic-di-GMP (c-di-GMP) from the free DNA (Figure 1A, 1B) in order to obtain direct patterns of the complexes themselves (Figure 1C). In the case of phosphorylated response regulator BvgA (BvgA~P), while only one complex, the open complex (RPo), was formed by RNAP in the absence of ribonucleoside triphosphates (rNTPs) (Figure 2A, lanes 1 vs. 2), two complexes, both RPo and the initiating complex (RPi), were formed by RNAP in the presence of rNTPs (Figure 2A, lane 3). Consequently, we used this technique to separate RPi from the free DNA and from RPo, allowing us to visualize the DNase I footprint of each complex separately (Figure 2B, lanes 2 and 3). We also employed this method together with KMnO₄ footprinting to determine the position of the transcription bubble within RPo formed by RNAP with non-phosphorylated BvgA (Figure 2A, lane 4) or within RPo formed by RNAP and BvgA~P (Figure 2A, lane 2). This yielded the KMnO₄ footprints seen in Figure 2C, lanes 1 and 2, respectively. We were then able to compare these results to those obtained without the complex isolation step (Figure 2D, lanes 1 and 2), allowing us to determine how the stable complexes differed from the mixture of complexes that were formed in solution.


Figure 1. Example of EMSA/DNase I footprinting. Figure has been adapted and reprinted with permission from Hsieh et al., 2018. (A) and (B) EMSA gels showing complexes formed by the V. cholerae regulator VpsR bound to the promoter DNA for the V. cholerae gene vpsL (P_vpsL) in the presence and absence of the small molecule c-di-GMP and in the absence (A) and presence (B) of RNA polymerase (RNAP), as indicated. The position of the free DNA is indicated by grey arrows while shifted complexes are represented by black arrows. Binding reactions in (A) contained 5 nM ³²P-labeled nontemplate P_vpsL harboring -97 to +103 relative to the transcription start site, 1 μg poly (dI-dC) as competitor, and increasing amounts of VpsR from 0 to 2 μM in the absence (lanes 1-5) and presence (lanes 6-10) of 50 μM c-di-GMP. Binding reactions in (B) contained 0.04 μM P_vpsL, 0.14 μM reconstituted RNAP (σ:core ratio of 2.5:1), 1.4 μM VpsR, and 50 μM c-di-GMP, as indicated, and the reactions were competed with 500 ng heparin for 15 s. (C) Sequencing gels showing the DNase I products from isolated complexes. In these cases, reactions were incubated with 0.3 U of DNase I at 37 °C for 30 s before loading on the EMSA gel, the free DNA and complexes were excised and isolated, and after extraction, the DNA was electrophoresed on the gel. GA indicates G+A ladder. A schematic of the -10 element, -35 element, and the +1 position of P_vpsL is shown to the right of the image. Protection pattern and hypersensitivity sites are indicated by black rectangular boxes and thin black arrows, respectively. Dashed red box represents the VpsR binding site.


Figure 2. Example of EMSA/DNase I footprinting and of KMnO₄ footprinting with and without isolation of protein-DNA complexes from EMSA gel. Figure has been adapted and reprinted with permission from Boulanger et al., 2015. (A) EMSA gel showing complexes formed by the B. pertussis response regulator BvgA either phosphorylated (BvgA~P) or non-phosphorylated (BvgA), the promoter for the B. pertussis gene fim3 (P_fim3), and RNAP. Binding reactions contained 0.05 pmol DNA, 5 pmol BvgA or BvgA~P, and 0.75 pmol reconstituted RNAP (σ:core ratio of 2.5:1). All samples were competed with 200 μg/ml heparin and as indicated, rNTPs (GTP, ATP, and CTP) were added. Free DNA is represented by a grey arrow while the open complex (RP_O) is labeled with a black arrow and the initiating complex (RP_I) is represented by a dashed black arrow. (B) Sequencing gels showing DNase I footprints at P_fim3 obtained from binding reactions with 0.36 U of DNase I and the subsequent EMSA isolation step. A schematic of the binding region as well as -35 , -10, and +1 is to the right of the image. Hypersensitive sites at -15 and -80, observed with RPi, but not with RPo, are indicated with the ‘*’. (C) and (D) Sequencing gels showing KMnO₄ footprints in the region of position +1 of P_fim3 obtained by treating the binding reactions with 2.5 mM KMnO₄ for 2.5 min at 37 °C and then either obtaining complexes from an EMSA gel and isolating the DNA before running the sequencing gel (C) or using the binding reaction products directly without the EMSA isolation step (D). Binding reactions contained 0.5 pmol DNA, 12 pmol BvgA or BvgA~P and 1.125 pmol reconstituted RNAP (σ:core ratio of 2.5:1). GA represents G+A ladder.