Using electrophoretic mobility shift assays to measure equilibrium dissociation constants: GAL4-p53 binding DNA as a model system

Authors


Abstract

An undergraduate biochemistry laboratory experiment is described that will teach students the practical and theoretical considerations for measuring the equilibrium dissociation constant (KD) for a protein/DNA interaction using electrophoretic mobility shift assays (EMSAs). An EMSA monitors the migration of DNA through a native gel; the DNA migrates more slowly when bound to a protein. To determine a KD the amount of unbound and protein-bound DNA in the gel is measured as the protein concentration increases. By performing this experiment, students will be introduced to making affinity measurements and gain experience in performing quantitative EMSAs. The experiment describes measuring the KD for the interaction between the chimeric protein GAL4-p53 and its DNA recognition site; however, the techniques are adaptable to other DNA binding proteins. In addition, the basic experiment described can be easily expanded to include additional inquiry-driven experimentation. © 2012 by The International Union of Biochemistry and Molecular Biology

We present an experiment, appropriate for an undergraduate biochemistry laboratory course, that will enable students to measure the equilibrium dissociation constant (KD) for a protein/DNA interaction using electrophoretic mobility shift assays (EMSAs), also known as gel shifts. This experiment introduces students to the theoretical and practical considerations behind measuring the binding affinity for a biological interaction. Equilibrium binding and dissociation constants are typically introduced in general chemistry; however, their application to biological interactions may not be part of biochemistry curricula. Indeed, many biochemistry students, even at the graduate level, do not firmly grasp these concepts. This experiment provides the opportunity to measure the binding affinity of a protein/DNA interaction.

The interaction between the protein GAL4-p53 and its DNA recognition sequence is used as a model system. Protein/nucleic acid interactions are fundamental to all of biology, therefore understanding how to quantitatively measure the affinity of such an interaction, and what that affinity means for biological regulation is important. In addition, this experiment will teach students how to perform EMSAs. This is a basic technique that is widely used in biochemical research laboratories, hence is a valuable technique to learn. Lastly, this experiment provides a framework that can be readily expanded, for example, to teach the relationship between KD and rate constants, or to allow for inquiry-driven experimentation.

BACKGROUND AND THEORY

Transcription Factors as a Model System for Protein/DNA Interactions

Proper transcription of protein-encoding genes (the synthesis of mRNA using DNA as a template) is essential for cellular viability. In eukaryotes, mRNA transcription is catalyzed by the enzyme RNA polymerase II, which requires numerous accessory factors to synthesize mRNA in a gene-specific and regulated fashion [1]. One such category of regulatory factors is transcriptional activator proteins. A prototypical transcriptional activator protein consists of a DNA binding domain (DBD) that binds to specific DNA sequences with high affinity, and at least one activation domain (AD) that is required to mediate transcriptional activation [2, 3]. The DBD alone will bind DNA, however cannot activate transcription, while the AD alone neither binds DNA nor activates transcription. A hallmark of these domains is their modularity. The DBD from one activator can be fused to the AD from a different activator, and the resulting chimeric protein will have the DNA binding properties of one protein, and the transcriptional activation properties of the other [2, 3]. The experiment described here uses the chimeric transcriptional activator protein GAL4-p53. It contains the DBD from the yeast transcriptional activator GAL4 fused to the AD from human p53. Because this experiment measures the affinity of DNA binding, the GAL4 DBD is the functional region of interest. Several excellent reviews and research articles focusing on the GAL4 DBD and how it has been used as a tool to study transcription have been published [4–7].

Equilibrium Dissociation Constants

The quantitative measurement used to assess the affinity of a biological interaction is the equilibrium dissociation constant (KD). As it relates to the experiment described here, the KD is the concentration of GAL4-p53 at which 50% of the DNA is in a complex with the protein, hence, it is expressed in units of molar (moles/liter). The relationship between KD and affinity is reciprocal; the lower the KD, the higher the affinity. To measure the KD for a protein/DNA interaction, a series of binding reactions are performed in which the concentration of the DNA is below the KD for the interaction, and the protein is titrated from below to above the KD. This requires an initial estimate of the KD, which can be determined from a preliminary experiment in which the DNA concentration is set low and the protein is titrated over a broad range. For high affinity complexes (e.g. low nM), it may be problematic to set the concentration of DNA well below the KD because the DNA might be difficult to detect under these conditions. In this case, the lowest reliably detectable DNA concentration can be used, and the measured KD is an upper limit of the affinity of the interaction (e.g. KD < 1 nM). After performing binding reactions in which the protein is titrated, the fraction of DNA bound at each concentration of protein is calculated and the data are fit with a binding equation using nonlinear regression (described in detail in the Experimental Procedures section). A more thorough discussion of the theory, experimental considerations, and equations behind measuring a KD can be found elsewhere [8]. Computer simulations that run in Excel and allow the user to understand the relationship between affinity and the concentrations of reaction components are downloadable from the referenced website [8]. The Supporting Information contains pre-lab questions for students that focus on equilibrium binding reactions.

Electrophoretic Mobility Shift Assays to Monitor Protein/DNA Interactions

To monitor the interaction between GAL4-p53 and its DNA recognition sequence, the experiment described here uses EMSAs. The EMSA, or gel shift, monitors the migration of the DNA through a native gel under a current in both the absence and presence of GAL4-p53. The rate at which the DNA migrates through the gel is related to its size and overall charge, as well as shape. The size of the DNA/GAL4-p53 complex is larger than the DNA alone, therefore, the protein-bound DNA migrates more slowly through a gel causing the position of the DNA to shift. To visualize the DNA in the EMSA described here, the DNA is labeled with a fluorophore, making the assay simple to perform in a teaching lab that does not allow the use of radioactivity. The 21 basepair double-stranded DNA is modified on the 5′ end of one strand with a Cy5 fluorophore that can be detected by scanning the native gel with a Typhoon scanner (or other imaging system that can detect a fluorescent signal in a gel).

During an EMSA, the gel matrix stabilizes protein/DNA complexes so that if the complex dissociates while the gel is running, the protein and DNA are unable to diffuse away from one another before rebinding occurs; this is known as the “caging” effect. Therefore, the ratio of bound DNA to unbound DNA after the EMSA should reflect the ratio in the binding reaction at the time the gel was loaded. Additional discussion of the theory and practical applications of EMSAs can be found elsewhere [9].

EXPERIMENTAL PROCEDURES

Time and Teams

This experiment is appropriate for students in upper-level undergraduate biochemistry laboratory courses. These students are typically familiar with the general principles encompassed by the experiment such as DNA and protein structure; however, it is not necessary that students have been introduced to EMSAs or transcriptional activator proteins. This experiment can be taught as a stand-alone experiment that can be completed in a single 4–5 hour laboratory session. Alternatively, it can be expanded to encompass several laboratory sessions, using the suggestions described in the Concluding Remarks section. The Supporting Information contains set-up instructions with lists of reagent needs, equipment needs, and buffer/gel recipes.

Protein

The experiment used recombinant GAL4-p53, which was expressed in Escherichia coli and purified as described previously [10]. Purified protein was quantitated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE); several dilutions were run on a 10% gel alongside a standard curve of bovine serum albumin (BSA). The gel was stained with Coomassie (0.2% Coomassie brilliant blue R-250, 50% methanol, 10% acetic acid; destain with 25% methanol, 10% acetic acid). The amount (ng) of GAL4-p53 in each lane was determined by comparing the band intensities to the standard curve. Protein was stored in small aliquots at −80 °C.

The experiment described here can also be performed with the GAL4 DBD alone [7] or other GAL4 DBD fusion proteins. The GAL4 DBD is commercially available (Santa Cruz Biotechnology, Santa Cruz, CA), as are various GAL4 DBD-containing chimeric proteins (for example, GAL4-VP16 from ProteinOne, Rockville, MD which contains the GAL4 DBD fused to the AD of VP16). When using purchased protein of unknown purity it is important to quantitate the amount of GAL4-containing protein using SDS-PAGE, as described above, in case there are other proteins present that would contribute to the absorbance at 280 nm.

DNA

To generate the 21 basepair double-stranded DNA necessary for the experiment, the following two single-stranded oligonucleotides were ordered from Integrated DNA Technologies, Coralville, IA: 5′-Cy5-TCCGGAGGACTGTCCTCCGGC-3′ and 5′-GCCGGAGGACAGTCCTCCGGA-3′. Prior to the experiment the instructor prepared the double-stranded DNA as follows. The GAL4 binding sequence in the oligo is self-complementary so intramolecular annealing can interfere with the efficiency of the annealing reaction. To mitigate this problem, a large excess of unlabeled oligonucleotide is used in the annealing reaction to drive the labeled oligonucleotide to anneal. A 50 μL annealing reaction was assembled containing 100 pmol of the fluorescently labeled DNA strand, 500 pmol of the unlabeled DNA strand, 20 mM Tris (pH 7.9), 2 mM MgCl2, and 50 mM KCl. The annealing reaction was heated to 95 °C for 3 min, and then cooled to 25 °C at a rate of 0.1 °C per second using a thermocycler. (If a thermocycler is not available, the reactions can be heated to 95 °C then slowly cooled to room temperature by turning off the heating device.). Supporting Information Fig. 1 shows a native gel resolving annealing reactions that contained different ratios of unlabeled:labeled oligonucleotide. Even with a large excess of the unlabeled oligonucleotide, single stranded DNA remains; therefore, the double stranded DNA was gel purified.

Figure 1.

GAL4-p53 decreases the migration of fluorescently labeled DNA in a native gel. Representative data obtained from a student.

The double-stranded annealed DNA was purified away from remaining single-stranded oligonucleotide as follows. The annealing reaction was run on a native gel (described below). The DNA was visualized by UV shadowing. The slowest migrating band corresponds to the double-stranded DNA, and was cut out of the wet gel. The gel slice was crushed and soaked overnight in 300 μL of 10 mM Tris (pH 7.9), 0.1 mM EDTA, protecting it from light. Solid gel matrix was removed by centrifugation through a filter, and the DNA was ethanol precipitated and resuspended in 200 μL. The concentration was determined by running a titration of the purified dsDNA (2–20 μL) on a native gel next to known amounts (1–100 ng) of a similar sized DNA standard. The gel was stained with SYBR Gold, and imaged on a Typhoon scanner. The band intensities were quantitated and the GAL4 DNA was compared to the standard curve. The purified double-stranded DNA was diluted to 10 nM in distilled water and stored in small aliquots at 20 °C, protected from light. Even after gel purification, it is not uncommon to observe some ssDNA in the samples, which runs below (faster) the double stranded DNA and does not shift or change upon addition of GAL4-p53.

Native Gel

The native gel was prepared before assembling the binding reactions. Low fluorescence glass gel plates (20 cm × 22 cm, Owl, Asheville, NC) with 1.5 mm thick spacers (Owl) were taped together. A mixture containing 5% acrylamide (37.5:1 acrylamide:bis ratio), 1× Tris-Glycine buffer, and 5% glycerol was prepared. To initiate polymerization, 450 μL of 10% ammonium persulfate and 100 μL of N′,N′,N′,N′-tetramethylethylenediamine (TEMED) were added, then immediately the gel plates were filled with the acrylamide mix and the comb was inserted. After the gel was polymerized, it was placed in a gel running apparatus and gel running buffer (1× Tris-Glycine buffer, 5% glycerol) was added to the upper and lower reservoirs. The gel was pre-run a minimum of 30 min at 150 V. Pre-cast native gels are commercially available (BioRad, Invitrogen, CBS Scientific), however, would have to be tested for use in the assays described here.

Binding Reactions

20 μL binding reactions were assembled on ice according to Table I, adding reaction components in the order listed from top to bottom. Notice that the first reaction does not contain protein, and the GAL4-p53 is titrated up in subsequent reactions. Buffer A + DNA contains 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.9), 8 mM MgCl2, 1 mM dithiothreitol (DTT), 50 μM ZnCl2, and 0.2 nM DNA. Buffer B contains 20% glycerol, 20 mM Tris-HCl (pH 7.9), 100 mM KCl, 1 mM DTT, and 50 μg/mL BSA. The 400, 40, and 4 nM GAL4-p53 solutions were made using Buffer B. (Note that these are the concentrations of GAL4-p53 dimers, not monomers, since the protein binds DNA as a dimer). The use of two buffers allows the volume of protein added to each reaction to be easily manipulated; GAL4-p53 is stored and diluted in Buffer B, and GAL4-p53 plus Buffer B should always equal 10 μL.

Table I. Experimental set up for 11 binding reactions
 Reaction number
Reaction component (μL)1234567891011
Buffer A + DNA1010101010101010101010
Buffer B109.5989.5985985
400 nM GAL4-p53125
40 nM GAL4-p530.5125
4 nM GAL4-p530.512

After addition of GAL4-p53, the reactions were incubated 15 min at room temperature. 18 μL of each reaction was loaded on the polymerized 5% native gel, while the gel was running at 150 V. Students should exercise caution not to touch the running buffer when loading a running gel. The gel was run for 45–75 min.

Fluorescence Measurement

The gel was scanned for fluorescence while still in the plates using a Typhoon Imager 9400 (GE biosciences) to detect the Cy5 fluorophore with the following settings: 633 nm excitation, 670 nm emission (with a 30 nm bandpass), and the focal plane at +3 mm. Any imaging system that will detect Cy5 fluorescence can be used, or the dye can be changed to accommodate other imaging systems.

If a Typhoon or other fluorescent imager is not available, the alternative approach of performing the experiment with unlabeled DNA oligonucleotides and staining the gel with a nucleic acid binding dye can be considered. Conceptually, this approach should work; however, we have not tested it with GAL4-p53 binding to DNA. Given the low amount of DNA used in the experiment, ethidium bromide will not be sensitive enough to visualize the DNA. A number of more sensitive dyes are commercially available. These include, for example, GelRed and GelGreen (Phenix Research, Candler, NC), GelStar (Cambrex, Charles City, IA), SYBR Gold and SYBR Green (Invitrogen, Grand Island, NY). Several of these can be used with a standard UV trans-illuminator. In addition, a mutant DNA sequence that decreases the affinity of the interaction can be used [7], which would allow for a higher concentration of DNA in the assay while still keeping it below the KD of the interaction.

Data Analysis

The value of the KD can be approximated from looking at the gel and determining the concentration of GAL4-p53 at which half of the DNA looks to be unbound and half looks to be bound. A precise value for the KD was determined by first using ImageJ software (freely downloadable from NIH) to quantitate the fluorescent signal in each DNA band. Background signals from blank regions of the gel were subtracted from the signal intensities obtained from bands. The fraction of DNA bound was determined from the background-subtracted signal intensities using the expression: bound/(bound + unbound). The fraction of DNA bound in each reaction was plotted versus the concentration of GAL4-p53. The data were fit with the following binding equation using Prism software to perform non-linear regression and obtain a value for KD and Bmax (the fraction bound at which the data plateaus):

equation image

If a software program that performs nonlinear regression is not available, Excel can be set-up to perform this type of analysis using the Solver add-in [11, 12].

RESULTS AND DISCUSSION

The results of an EMSA to measure the affinity of the interaction between GAL4-p53 and DNA are shown in Fig. 1. Lane 1 of the gel shows the migration of the unbound DNA and Lanes 2–11 show the shift in migration that occurred upon titrating GAL4-p53 into the binding reactions. Approximately, 50% of the DNA was in a complex with GAL4-p53 in reactions containing 2 nM protein, which provided an estimate of the KD for the interaction. Shown in Figs. 2a and 2b are plots of the data from Fig. 1. The fluorescent signal in each bound and unbound band was quantitated, and the fraction of DNA bound in each reaction was plotted versus GAL4-p53 concentration (in nM). The data were fit with a binding equation to obtain values for the KD and maximum fraction bound (Bmax) of 2.0 ± 0.8 nM and 1.05 ± 0.05, respectively. The errors are the 95% confidence interval obtained from the curve fit; the R2 is 0.978. In Fig. 2a the data are plotted with a linear X-axis, and in Fig. 2b with a logarithmic the X-axis. In the latter case the binding curve becomes sigmoidal, which allows the concentration of GAL4-p53 at which 50% of the DNA is bound (e.g. KD) to be more easily visualized.

Figure 2.

GAL4-p53 binds to DNA with high affinity. (a) The plot shows the fraction of DNA bound as GAL4-p53 was titrated. The values for the KD and Bmax obtained from the curve-fit are 2.0 ± 0.8 nM and 1.05 ± 0.05, respectively. (b) The data in panel (a) were plotted with a logarithmic X-axis.

This experiment is designed such that the DNA should be fully bound at the highest concentrations of GAL4-p53 in Table I. For various experimental reasons, however, complete binding may not occur. For example, if a given protein preparation is only partially active then it may take significantly higher concentrations of the protein preparation to saturate the DNA. Such a result can stimulate useful discussions about quantitating the fractional activity of a protein preparation or predicting other experimental reasons why binding may not have been complete. Importantly, even if the DNA is not fully saturated, students can usually still detect changes in the fraction bound as the experimental conditions are varied (i.e. using mutant templates or changing salt concentrations), therefore, still making the experiment useful as an assay to monitor a protein/DNA interaction.

Lastly, a student's success in performing this experiment is typically related to his/her ability to accurately assemble and manipulate small reaction volumes. Similarly, the quality of the curve-fit (a good R2 regardless of the value of the KD) largely depends on a student's pipetting skills, which can vary quite a bit between students. If necessary, simple tests of pipetting can help students practice prior to the experiment. For example, pipetting water onto a microbalance, or even adding small volumes of dyes to water, are helpful exercises. In addition, tips to help avoid common experimental pitfalls are listed in Table II.

Table II. Tips to avoid common experimental pitfalls
Mix well after each addition by finger tapping or a light, short vortex; Buffer B can be difficult to mix in due to the glycerol.
Thaw the GAL4-p53 immediately prior to adding it to reactions.
Ensure the reactions are in the bottoms of the tubes, using a short spin in a microfuge if necessary.
Clean the gel plates prior to use with either ethanol or water.
When quantitating the fluorescence in gels make sure the regions used for background subtraction do not contain stray spots of fluorescence.
If necessary practice pipetting small reaction volumes
Dyes can be loaded in the native gel next to the sample lanes in order to visually monitor the progress of the electrophoresis.
Keep the lids to all tubes closed unless a reagent is being added or removed.
If DNA that has been frozen for a long period of time is not shifting well, a new annealing and purification is recommended

CONCLUDING REMARKS

The experiment described here will teach students how to determine the KD for a protein/DNA interaction using EMSAs. It provides students a hands-on opportunity to learn a technique widely used in research settings, and principles that underpin biological interactions. Obtaining an understanding of both affinity measurements and the technique of EMSA is valuable to emerging scientists.

The experimental framework described is adaptable to designing further inquiry-driven experiments. For example, specific mutations in the DNA could be tested to ask how the sequence at select positions influences binding affinity. Published studies have investigated this relationship and could be used either as a guide for designing mutations or as a cross-check for results obtained by students [7]. Binding affinity is often influenced by salt concentration. Each group of students in a laboratory could measure the KD at a different salt concentration and compile their data to understand how salt dictates the interaction between the GAL4 DBD and DNA. Alternatively, the off-rate of the protein/DNA complex could be measured by monitoring the decrease in the fraction of DNA bound over time after adding excess non-fluorescent DNA as a competitor [8]. This kinetic experiment could also be performed under varying experimental conditions. By coupling it to the experiment described here, the relationship between binding affinity and rate constants can be investigated.

Lastly, students can take advantage of 3D images of the GAL4 DBD bound to DNA that were generated from crystal structures [6, 13]. The referenced webpages contain links to viewing platforms that allow students to visualize the protein/DNA structure and highlight specific features. Manipulating these structures will allow students to understand the specific amino acid–DNA contacts that mediate the interaction, and can also help generate hypotheses for how the protein/DNA complex might change as experimental conditions are varied or certain amino acids are mutated.

Acknowledgements

Authors thank Ben Gilman for preparation of the GAL4-p53.

Ancillary