Characterization of cis-Regulatory Elements and Transcription Factor Binding

Summary

To understand how cardiac gene expression is regulated, the identification and characterization of cis-regulatory elements and their trans-acting factors by gel mobility shift assay (GMSA) or gel retardation assay are essential and common steps. In addition to providing a general protocol for GMSA, this chapter describes some applications of this assay to characterize cardiac-specific and ubiquitous trans-acting factors bound to regulatory elements [novel TCTG(G/C) direct repeat and A/T-rich region] of the rat cardiac troponin T promoter. In GMSA, the specificity of the binding of trans-acting factor to labeled DNA probe should be verified by the addition of unlabeled probe in the reaction mixture. The migratory property of DNA-protein complexes formed by protein extracts prepared from different tissues can be compared to determine the tissue specificity of trans-acting factors. GMSA, coupled with specific antibody to trans-acting factor (antibody supershift assay), is used to identify proteins present in the DNA-protein complex. The gel-shift competition assay with an unlabeled probe containing a slightly different sequence is a powerful technique used to assess the sequence specificity and relative binding affinity of a DNA-protein interaction. GMSA with SDS-PAGE fractionated proteins allows for the determination of the apparent molecular mass of bound trans-acting factor.

1. Introduction

The identification and characterization of cis-regulatory elements and transacting factors is essential for understanding the control of cardiac gene expression. Promoter studies on cardiac-restricted genes have suggested that the activation of these genes is likely controlled through interactions of a combination of numerous cardiac-restricted and ubiquitous transcriptional factors. Additionally, transcriptional repressors may also be present in noncardiac tissues and may play a role in reinforcing this cardiac specificity during differentiation. The gel mobility shift assay (GMSA) provides an approach to characterize the interactions between cis-regulatory elements and trans-acting factors that regulate cardiac gene expression. The fundamental principle behind this assay is that the binding of a protein to a DNA fragment, which has been radiolabeled for ease in analysis, will reduce the mobility of this DNA fragment when it is electrophoresed on a nondenaturing polyacrylamide gel. This differential mobility allows the complex to be distinguished from the unbound probe. This technique can be performed with whole cell extracts from cardiac tissue, as well as nuclear extracts from cardiac tissue, to investigate binding of trans-acting factors to the promoter regions of cardiac genes. This assay can further be applied to confirm interactions of suspected transcription factors with a regulatory region of the promoter.

This chapter discusses a specific application of the GMSA to characterize the binding of cardiac-specific and ubiquitous trans-acting factors to the regulatory region of rat cardiac troponin T (cTnT) promoter. Expression of the rat/mouse cTnT gene is clearly detectable in the lateral plate mesoderm of the heart forming fields at embryonic day 7.5 (E7.5), and in the linear heart tube at E8.0 (1,2). At E10.5, cTnT is strongly expressed in the heart and weakly expressed in a few of the most posterior somites. However, no skeletal muscle TnT gene expression is seen at this stage of development, suggesting that the expressed cTnT protein may be functional in these skeletal muscle precursors. At later fetal stages, transcription of the cTnT gene is specifically repressed in developing skeletal muscle. The cTnT proteins detected in the skeletal muscles decline progressively and disappear 2 wk after birth. In the adult, cTnT is expressed only in the cardiomyocytes. This expression pattern implies both temporal and spatial regulation of cTnT expression and suggests a tightly regulated expression profile. Using reporter assays, a cTnT proximal promoter region −497 bp from the transcriptional start site has been identified to drive cardiac-specific expression specifically in cultured cardiomyocytes and to recapitulate the endogenous cTnT expression pattern in transgenic mice (1–3). This indicates that the essential regulatory elements for driving cardiac specificity are contained within this −497 bp promoter (Fig. 1A).

Fig. 1.

Rat cardiac troponin T (cTnT) promoter. (A) Diagram of rat −497 bp cTnT promoter. Both D and F modules have a TCTG(G/C) direct repeat indicated by arrows and an A/T-rich site. In addition, the F module contains an A/T-rich core sequence within a consensus MEF2 site (MEF2-like motif). (B) Sequence comparison between D (D1 and D2, −335 bp to −289 bp) and F (F41, −249 bp to −209 bp) modules. The vertical lines represent identical sequence. Gaps are introduced to obtain maximal alignment. The TCTG(G/C) direct repeats in D1, D2, and F41 are indicated by arrows.

Detailed evaluation of this −497 bp proximal promoter region uncovered two sequence homologous modules, module D and module F, which each contain TCTG(G/C) direct repeating units as well as an A/T-rich site (Fig. 1B). In addition, module F contains a myocyte-specific enhancer factor 2 (MEF2)-like motif, which is also an A/T-rich region. Analysis with deletion and religation mutant promoters suggests that module D serves as an enhancer to increase promoter activity but cannot totally substitute for the function of module F (4). GMSA and competition GMSA with probes (F41, D1, and D2) containing direct repeats and A/T-rich sites of modules D and F were used to demonstrate that the same protein factors bound to these cis-regulatory elements. Using application of the GMSA, the 42-kDa proteins were identified to bind the direct repeat regions of this promoter in a cardiac tissue-specific manner (Fig. 2). Furthermore, a 25-kDa ubiquitous protein was identified to bind the A/T-rich site and was later shown to be a high mobility group 2 (HMG2) family protein (4). Such application of the GMSA provides an effective tool for investigating the details of the complex regulatory process through which cTnT directs cardiac-specific expression.

Fig. 2.

Gel mobility shift assays (GMSAs) with the F41 probe. Protein extract (10 μg) prepared from heart or stomach was incubated with ³²P-labeled F41 probe in each reaction. For examination of the specificity, a 100-fold excess of the unlabeled F41 probe (cold probe) was added to the reaction mixture before incubation with the labeled probe. Complexes A and B formed by heart extracts appeared to have different mobility from complexes formed by stomach extracts (As and Bs). In contrast, the fast-migrating complexes (C and Cs) formed by heart and stomach extracts had the same mobility, suggesting the involvement of a ubiquitous factor in this complex.

In this chapter, we describe in detail the preparation of total protein extracts from cardiac tissue, the preparation of nondenaturing polyacrylamide gel and labeled probe, the DNA-protein binding reaction in the GMSA, and the specific applications of this technique.

2. Materials

2.1. Protein Preparation

1. 20 g of Each tissue (fresh or frozen): heart, liver, stomach, and skeletal muscle. Frozen rat tissues can be purchased from Pel-Freez Biologicals, Rogers, AR; www.pelfreez-bio.com.

2. Liquid nitrogen.

3. Mortar and pestle.

4. Lysis buffer: 15 mM HEPES, pH 7.6, 100 mM KCl, 5 mM MgCl₂, 1 mM dithio-threitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF), and 10% glycerol.

5. 400 mM NaCl extraction buffer: 15 mM HEPES, pH 7.6, 400 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 1 mM PMSF, and 10% glycerol.

6. Heparin-agarose column: 2.5 mL prepacked column (Sigma, St. Louis, MO, cat. no. HEP-1-5).

7. 1 M NaCl extraction buffer: 15 mM HEPES, pH 7.6, 1 M NaCl, 5 mM MgCl₂, 1 mM DTT, 1 mM PMSF, and 10% glycerol.

8. 1X GMSA binding buffer: 20 mM HEPES, pH 7.6, 100 mM KCl, 5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, and 10% glycerol.

9. Dialysis tubing: mol. wt. cutoff 12 to 14,000 or Spectra/por #1 mol. wt. cutoff 6 to 8,000 (Spectrum Medical Industries, Los Angeles, CA).

2.2. DNA Probe Preparation

The length of the DNA fragment can vary between 20 and 300 bp for production of a probe. The longer the fragment, the more likely it is that multiple binding sites are present, and nonspecific binding may occur.

1. Restriction enzyme-digested, or polymerase chain reaction (PCR)-amplified, end-labeled DNA fragments or annealed, end-labeled oligonucleotides (commercially synthesized) can be used as probes for GMSA.

2. [α-³²P]dNTP (>3000 Ci/mmol) for Klenow fill-in, end-labeling or [γ-³²P]ATP (>3000 Ci/mmol) for T4 polynucleotide kinase end-labeling.

3. Klenow fragment of E. coli DNA polymerase I, restriction enzymes, T4 polynucleotide kinase (Promega, Madison, WI or New England Biolabs, Beverly, MA).

4. 10X T4 polynucleotide kinase buffer: 700 mM Tris-HCl, pH 7.6, 100 mM MgCl₂, and 50 mM DTT.

5. 10X Klenow buffer: 100 mM Tris-HCl, pH 7.5, 50 mM MgCl₂, and 75 mM DTT.

6. 0.5 mM each of dNTPs.

7. Razor blade.

8. Qiaex II gel extraction kit (Qiagen, Valencia, CA).

9. Micro Bio-spin 6 column (Bio-Rad, UK).

2.3. Polyacrylamide Gel Preparation

1. 0.1% Sodium dodecyl sulfate (SDS) solution for washing glass plates.

2. 95% Ethanol for cleaning plates.

3. 1% Agarose for sealing plates when pouring the acrylamide gel.

4. 5X Tris-glycine electrophoresis buffer stock: 60.56 g (0.25 M) Tris base, 285.4 g (1.9 M) glycine, 7.44 g (10 mM) EDTA disodium salt, dihydrate, and deionized, distilled H₂O to 2 L. Check that the pH is 8.5. Store for several months at room temperature.

5. 30% Acrylamide.

6. 1% Bis-acrylamide.

7. 50% Glycerol.

8. TEMED (N,N,N′,N′-tetramethylenediamine).

9. 10% Ammonium persulfate (freshly prepared).

10. Whatman 3MM filter paper. 11. Plastic wrap.

11. Gel dryer (Bio-Rad, UK).

12. AR X-ray film, intensifying screen and cassette or PhosphorImager (Molecular Dynamics).

2.4. DNA-Protein Binding Reaction

1. 2X GMSA binding buffer: 40 mM HEPES, pH 7.6, 200 mM KCl, 10 mM MgCl₂, 0.4 mM EDTA, 1 mM DTT, 20% glycerol.

2. 1 μg/μL poly(dI-dC).poly(dI-dC).

3. 10X Gel loading dye: 0.25% bromophenol blue in 1X GMSA binding buffer.

3. Methods

3.1. Preparation of Protein Extracts

The preparation of the protein extracts from frozen adult rat tissues such as heart, liver, stomach, and skeletal tissues for use in GMSA is described in Subheadings 3.1.1. and 3.1.2. This includes (1) a detailed description for preparing total protein extracts and (2) a brief description of the preparation of nuclear extracts.

3.1.1. Preparation of Total Protein Extracts

1. Grind 20 g of frozen adult rat heart, skeletal muscle, liver, or stomach (see Note 1) in liquid nitrogen using a mortar and pestle.

2. Incubate the pulverized tissue in 200 mL of lysis buffer at 4°C for 30 min while gently stirring.

3. Centrifuge lysate at 8000g at 4°C for 20 min.

4. Save a small amount of the supernatant for later footprinting or GMSA analysis to make sure that the supernatant does not contain any binding activity. Resuspend the pellet with 400 mM NaCl extraction buffer. Incubate at 4°C for 30 min.

5. Centrifuge the mixture at 8000g at 4°C for 30 min and collect the supernatant. Repeat centrifugation, if needed. Save a small amount of the supernatant for later footprinting or GMSA analysis.

6. Apply the supernatant to a 2.5-mL heparin-agarose column preequilibriated with 20 mL of 400 mM NaCl extraction buffer (see Note 2).

7. Wash the column with 10 mL of 400 mM NaCl extraction buffer.

8. Elute the bound proteins with 12 mL of 1 M NaCl extraction buffer.

9. Dialyze the eluent with two changes (2 L each) of 1X GMSA binding buffer for at least 4 h.

10. Check the DNase I footprinting or GMSA activity of the extracts before and after heparin-agarose chromatography to compare the specific activity of trans-acting proteins in each fraction. Based on our studies, the resulting protein extracts after heparin-agarose column purification contain binding activity for many known transcription factor-binding motifs, such as the AP2 site, TATA site, CArG box, M-CAT site, MEF2-like motif, and others, as well as unknown factor-binding sites (3,4).

11. Determine the protein concentration by the Bradford method (5) or the Bio-Rad Protein Assay (Bio-Rad, UK).

12. Aliquot and store the protein extracts at −70°C.

3.1.2. Preparation of Nuclear Extracts

The preparation of nuclear extract requires isolation of nuclei from tissues or cultured cells. To reduce the contribution of nuclear extract by cell types other than muscle cells, primary cardiomyocytes can first be isolated, cultured, and subsequently used for the isolation of nuclei (6–8). In striated muscle tissues, large quantities of myofibrils interfere with the isolation of nuclei. Therefore, a special step to relax the muscle tissue is included before homogenization of tissues (9).

1. Nuclei from the homogenated tissues/cells are pelleted through a sucrose cushion by centrifugation to separate the cytoplasmic contents from the nuclei.

2. The protein extraction from the collected nuclei is generally carried out by the careful dropwise addition of a high-salt buffer into the resuspended nuclei, as described in detail elsewhere (10).

3. The high-salt extracts are subsequently clarified by centrifugation, and the supernatant is saved.

4. This supernatant now contains nuclear extracts and is dialyzed into GMSA binding buffer.

5. The nuclear extracts can then be aliquoted and stored in liquid nitrogen or at −70°C.

3.2. Preparation of Nondenaturing Polyacrylamide Gel

1. Wash glass plates thoroughly with 0.1% SDS solution, rinse well with deionized water, and dry (see Note 3).

2. Clean the plate with 95% ethanol.

3. Assemble two plates with 1.5-mm spacers on the sides and bottom and clip plates together securely to prepare for casting the gel.

4. Seal the sides and bottom of the plates with warm 1% agarose, and allow the agarose to solidify.

5. Prepare a low-ionic-strength gel solution (see Note 4). For a 6% native polyacrylamide gel, combine: 4.6 mL 5X Tris-glycine electrophoresis buffer stock, 4.6 mL 30% acrylamide, 3.0 mL 1% bis-acrylamide, 1.15 mL 50% glycerol, 22 μL TEMED, 9.43 mL H₂O, and 220 μL 10% ammonium persulfate, for a total of 23 mL. This amount can be scaled appropriately depending on the size of gel needed.

6. Pour the gel solution into the space between the plates, insert the comb into the gel, and allow gel to polymerize completely for approx 30 min.

7. Remove bottom spacer and attach plates to electrophoresis tank. Fill the lower and upper reservoirs of the tank with 1X Tris-glycine electrophoresis buffer (see Note 5). Remove the comb, and use a bent-needle syringe to remove air bubbles trapped between the plates below the gel, and flush out and straighten the wells.

8. Prerun the gel for 30 to 60 min at 100 V.

3.3. Preparation of Labeled DNA Probe

3.3.1. Preparation of Restriction Fragment Probes or PCR-Amplified DNA Probes

Prior to constructing the radiolabeled probe, the DNA fragment of interest must be selected (see Note 6).

1. Isolate a DNA fragment containing the binding sites of interest using a standard restriction enzyme digestion (see Note 7). Restriction endonucleases that leave 5′ overhangs are preferable for use in the subsequent end-labeling step.

2. End-label the restriction fragment by a Klenow fill-in reaction (see Note 8): 25 μL of 1X Klenow buffer containing 1 to 50 pmol of DNA fragment, 150 μCi of a suitable [α-³²P]dNTP (>3000 Ci/mmol), 40 μM each of the other three unlabeled dNTPs, and 10 U of the Klenow fragment of E. coli DNA polymerase I. Incubate for 15 min at room temperature. Then chase with the unlabeled form of the fourth dNTP and incubate for an additional 5 min.

3. Separate the desired labeled restriction fragment from the plasmid using gel electrophoresis. Identify the labeled DNA fragment by briefly exposing this electrophoresed gel to X-ray film. Excise the identified radiolabeled band from the gel with a razor blade and elute it by the Qiaex II gel extraction kit (Qiagen).

4. If PCR is used to amplify the region of interest, separate the PCR product by gel electrophoresis. Excise the DNA band from the gel and elute it by the Qiaex II gel extraction kit (Qiagen). The isolated DNA fragment is subsequently end-labeled using T4 polynucleotide kinase as described in Subheading 3.3.2.

3.3.2. Preparation of Oligonucleotide Probes

1. Synthesize commercially complementary oligonucleotides (see Note 9).

2. Anneal them to generate the double-stranded oligonucleotide probe containing the specific binding site of interest.

3. End-label the probe using T4 polynucleotide kinase in the following reaction conditions: 30 μL of 1X T4 polynucleotide kinase buffer containing dephosphorylated DNA fragments (1–50 pmol of PCR-amplified or double-stranded oligonucleotide probe), 150 μCi of [γ-³²P]ATP (6000 Ci/mmol), and 20 U of T4 polynucleotide kinase, 37°C for 1 h.

4. If the double-stranded oligonucleotides are designed to have a 5′ overhang, they can be labeled using Klenow fragment as previously described.

5. Remove unincorporated [γ-³²P]ATP by passing the reaction mixture through a micro Bio-spin 6 column (Bio-Rad, UK).

6. Collect the flow through fraction and determine the radioactivity.

3.4. Gel Mobility Shift Assay

Numerous conditions can affect DNA-protein binding (11), and thus many parameters of the binding reaction can be altered to obtain optimal conditions for the DNA-protein interaction of interest. A range of conditions varying in amounts of DNA probe and binding proteins, ionic strength, pH, divalent ion, temperature, glycerol content, and nonionic detergent content, as well as including carrier proteins and/or carrier DNA, may have to be tested to identify a potential interaction. The following is a GMSA protocol that has been used effectively in identifying the cis-regulatory elements and trans-acting factor bindings on the cTnT proximal promoter (2–4) (see Note 10).

3.4.1. Reaction Setup

To show the specificity of the binding reaction, adequate positive and negative controls must be established. It is also important to show specificity using a competition assay by including specific unlabeled probe as well as nonspecific unlabeled probe.

1. Negative controls include using the labeled DNA probe of interest with no protein added or with a protein that should not bind the probe.

2. A positive control would be a known DNA-protein interaction, which may not always be available.

3. If competition assays are to be performed, conditions should be optimized so that 20 to 30% of the probe is bound. Specific competitors (specific unlabeled probes or self-probes) are used at increasing molar excess of the labeled probe to show the specificity of the binding reaction (see Note 11).

4. Nonspecific competitors include an unlabeled DNA fragment that is different from the sequence of interest in the binding reaction. Nonspecific competitors are usually used at increasing molar excess of the labeled probe. These reactions show the specificity of the desired DNA fragment with the protein. Nonspecific competitors can also be used to identify the exact nucleotide sequence needed for binding. A probe with a sequence similar to the fragment of interest, but with an altered nucleotide(s) incorporated, can be used to show exactly which nucleotides are necessary for binding.

3.4.2. Binding Reaction

1. While the gel is prerunning, add the following reagents to a 1.5-mL micro-centrifuge tube in the order listed: 5000 cpm labeled probe (0.1–0.5 ng), 1 μL of 1 μg/μL carrier DNA poly(dI-dC).poly(dI-dC), 10 μL of 2X GMSA binding buffer, protein/nuclear extracts (10–20 μg protein extract or 5–25 ng purified protein). Adjust total volume to 20 μL with deionized distilled water.

2. Mix by tapping, and incubate at room temperature for 30 min.

3.4.3. Running the Gel

1. Add 2 μL of 10X gel loading buffer to each reaction.

2. Immediately load each reaction on a 6% nondenaturing polyacrylamide gel, which has been prerun in 1X Tris-glycine buffer for 30 min to 1 h.

3. Run the gel at 100 V for 1 to 1.5 h until the tracking dye is near the bottom of the gel. The tracking dye (bromophenol blue) migrates roughly at the same position as a 70-bp DNA probe. For probes <70 bp, the dye should not be run to the bottom of the gel (see Note 12).

3.4.4. Drying the Gel

1. Remove the gel from the gel tank and carefully remove side spacers.

2. Slowly pry the glass plates apart, allowing the gel to remain on one plate.

3. Lay the glass plate with the gel attached on the bench (gel side up) and cover the gel with a sheet of Whatman 3MM filter paper.

4. Flip the gel sandwich so that the filter paper is on bottom and glass plate is on top.

5. Remove the glass plate.

6. Cover the gel with plastic wrap and dry under a vacuum at 80°C for 90 min.

7. Autoradiograph the dried gel on X-ray film with an intensifying screen at −70°C or by PhosphorImager and ImageQuant software (Molecular Dynamics).

Figure 2 illustrates a typical example of GMSA results.

3.5. Applications

The GMSA has been adapted to be used for many additional applications; such as the antibody supershift assay, gel-shift competition assay, and GMSA with SDS-PAGE gel fractionated proteins. These applications are discussed in Subheadings 3.5.1. to 3.5.3.

3.5.1. Antibody Supershift Assay

The antibody supershift assay is an adapted variant of the GMSA in which antibodies are used to identify proteins present in the DNA-protein complex. This variant is limited because the researcher must have prior knowledge of what the binding protein may be, in order to test with accurate antibodies. Owing to this limitation, this technique would not be suitable for identifying novel proteins. If the antibody is included in the GMSA, three outcomes are possible in regard to the mobility of the DNA-protein complex: (1) no effect on the mobility if the protein recognized by the antibody is not involved in complex formation, (2) no complex formation and the probe runs like free probe if the protein is necessary for complex formation and the antibody interferes with the DNA-interacting sites of the protein, or (3) reduced mobility (supershift) if the protein that forms the complex is recognized by the antibody and the antibody does not perturb DNA-protein complex formation (see Note 13).

When one is investigating the regulatory mechanisms of the rat cTnT (2) and the rat cardiac ventricular myosin light-chain 2 (MLC-2v) (12,13), the antibody supershift assay has been successfully used to demonstrate the absence or presence of specific transcription factors within DNA-protein complexes. Previously, MEF2A and HF1b transcription factors had both been shown to bind to the MEF2 consensus sequence and regulate transcription of cardiac-restricted genes (12,14). When the antibody supershift assay was used to investigate whether MEF2A or HF1b was present at the MEF2-like site within the rat cTnT promoter, neither supershift nor effect on the DNA-protein complexes was observed. This result strongly suggests that neither MEF2A nor HF1b is present in the complex formed by cardiac proteins and DNA probe containing the MEF2-like site (2). Previously, a 250-bp MLC-2v promoter fragment containing three conserved regulatory elements (HF1a, HF1b/MEF2, and HF3) has been shown to confer ventricular-specific expression of a lacZ reporter gene (7,13,15). However, evidence also suggests that cardiac-specific expression of MLC-2v requires combinatorial interactions between various elements located within the 250-bp fragment, which led to the discovery of a novel upstream stimulating factor (USF)-binding site named MLE1 (13). When using the antibody supershift assay, USF antibodies supershifted both endogenous MLE1 and the HF1a radiolabeled probes, suggesting that USF can bind specifically to the MLE1 site and the HF-1a site within the MLC-2v promoter (13).

3.5.2. Gel-Shift Competition Assays

Gel-shift competition assay is a powerful technique used to asses the sequence specificity of a DNA-protein interaction, especially since most protein extracts contain both specific and nonspecific DNA binding proteins. When one is assessing a specific DNA-protein interaction, an unlabeled competitor DNA probe with the same sequence as the protein binding site can be used to sequester/compete the protein away from labeled probe. As a control for binding specificity, a nonspecific competitor probe must be used to show that the unlabeled nonspecific probe cannot sequester/compete the protein away from the labeled probe. Generally, a nonspecific competitor can be any fragment of unrelated sequence with the same size and configuration at the ends as the labeled probe, but ideally a nonspecific competitor should be identical to the labeled probe except for a key mutation in the binding site known to disrupt protein binding.

When we investigated the proteins that interact with the D and F modules of the rat cTnT promoter, competition GMSAs were used to demonstrate that the same proteins could bind to different cTnT promoter modules with differential affinities (4). Using F41, D1, or D2 as labeled probes, GMSAs were carried out with the absence (probe alone) and presence of an increased amount of non-self-competitors as well as self-competitors. When assays were performed with the F41 probe (Fig. 3A), both D1 and D2 competitors were able to compete off complexes A and B, but only D2 could compete off complex C because D1 contained no A/T-rich site (see Fig. 1B for sequence difference). When assays were performed with the D1 probe (Fig. 3B), only D2 but not F41 competitor could effectively compete off complexes A and B. When assays were performed with the D2 probe (Fig. 3C), neither D1 nor F41 competitors could compete off the complexes formed by D2. These results together suggest that the TCTG(G/C) direct repeats within these different probes have different binding affinities for the same protein factors and that the relative affinity is D2>D1>F41.

Fig. 3.

Competition GMSAs using the F41, D1, and D2 probes. (A) GMSA with labeled F41 probe (lane 1). Competition assays were carried out with 100-, 200-, and 400-fold molar excess of D1 (lanes 2, 3, and 4, respectively), D2 (lanes 5, 6, and 7, respectively) and 100- and 200-fold molar excess of self-competitor (lanes 8 and 9, respectively). Both D1 and D2 can compete off complexes A and B, whereas only D2 can compete off complex C. (B) GMSA with labeled D1 probe (lane 1). A 100- and 200-fold molar excess of F41 (lanes 2 and 3, respectively), D2 (lanes 4 and 5, respectively), and D1 itself (lanes 6 and 7, respectively) served as competitors. D2 but not F41 effectively competes off complexes A and B formed by the D1 probe. (C) GMSA with labeled D2 probe (lane 1). Unlabeled D1 (lanes 2 and 3), F41 (lanes 4 and 5), and D2 itself (lanes 6 and 7) at 100- and 200-fold molar excess were used as competitors. Neither D1 nor F41 can compete off complexes A and B formed by D2.

3.5.3. GMSA with SDS-PAGE Gel Fractionated Proteins

When we investigated the proteins that interact with the D and F modules of the rat cTnT promoter, GMSAs were used to identify the proteins involved in binding to the TCTG(G/C) direct repeat and A/T-rich site (4). First, F41 probed GMSAs using protein extracts from heart and stomach formed an identical complex C, suggesting the involvement of a ubiquitous factor. In contrast, tissue-specific factors are responsible for the formation of complexes A and B (minus lanes in Fig. 2). Competitive GMSAs using 100-fold molar excess of unlabeled F41 demonstrated that the DNA-protein interactions were specific for all complexes that were formed from both heart and stomach protein extracts (plus lanes in Fig. 2). To characterize the proteins involved in complex formation further, protein extracts from adult rat hearts were fractionated by SDS-PAGE into many fractions, eluted, renatured (16), and then incubated with either F41 probe (Fig. 4A) or D1 probe (Fig. 4B) for the subsequent GMSAs (see Note 14).

Fig. 4.

GMSAs of renatured proteins after separation by SDS-PAGE. Protein extracts from adult rat hearts were fractionated by SDS-PAGE into 23 fractions. Proteins in each fraction were eluted, renatured, and incubated with labeled F41 (A) or D1 (B). Lane C in each panel represents GMSA with protein extracts before gel fractionation. Proteins with 42 kDa found in fraction 10 are capable of forming complexes A and B with F41 or D1 probe. On the other hand, protein with 25 kDa in fraction 15 forms complex C only with F41 but not with D1.

1. Briefly, 50 μg of cardiac protein extracts were heated at 37°C for 5 min in SDS-PAGE loading buffer and subsequently separated on a 12.5% SDS-PAGE gel.

2. The gel was then sliced into 23 pieces, each of which was then mashed and incubated in 240 μL of elution/renaturation buffer (1% Triton X-100, 20 mM HEPES, pH 7.6, 1 mM EDTA, 100 mM NaCl, 5 mg/mL bovine serum albumin, 2 mM DTT, and 1 mM PMSF) at 37°C for 3 h and then 4°C overnight.

3. The eluent was separated from gel residues by centrifugation at 12,000g for 10 min at 4°C and stored at −70°C before use.

4. Twelve microliters of eluent from each fraction was then used for GMSAs.

Compared with protein extracts before gel fractionation as seen in lane C of Fig. 4, the 42-kDa proteins in fraction 10 were shown to form complexes A and B with either the F41 or D1 probe. However, the 25-kDa protein found in fraction 15 could form complex C with the F41 probe. Furthermore, when the D2 probe was used in GMSAs with these gel-fractionated proteins, the same results were obtained as when F41 was used as a probe (data not shown). Since both F41 and D2, but not D1, contain the A/T-rich site in addition to the TCTG(G/C) direct repeat, these results further confirmed that the 25-kDa protein bound to an A/T-rich site, whereas the 42-kDa proteins recognized the TCTG(G/C) direct repeat. The appreciation of this GMSA technique allowed for the identification of a 25-kDa ubiquitous protein bound to the A/T-rich site. This protein was subsequently shown to be an HMG2 protein (4).

4. Notes

1. When one is characterizing the cis-regulating elements and the trans-activating factors of the rat cTnT gene, protein extracts are taken from multiple tissues for the following reasons. First, multiple tissue types may use different transacting factors to either enhance or suppress the cTnT expression. Both cardiac and skeletal muscles express cTnT during embryonic and fetal stages (1,17–22), but cTnT expression is restricted and upregulated in cardiac muscle during later fetal stages and adult (23). Furthermore, rat cTnT expression has also been identified in regenerating adult skeletal muscle after cold injury and in mature skeletal muscle fibers after denervation (19). To understand the mechanism regulating cTnT expression and whether there is a mechanism specific to cardiac muscle, extracts are made from cardiac and skeletal muscle for further examination. Second, since cardiac tissue primarily contains cardiac muscle cells and fibroblasts, liver tissue is used as a control for the nonmuscle cell contamination in the cardiac tissue extracts. Since cardiac and skeletal muscles are both striated muscle types, stomach tissue is used as a control to establish that cTnT is not expressed in smooth muscle.

2. If the lysate is too thick to go through the heparin-agarose column by gravity, the lysate loaded column can be centrifuged using a low-speed setting (2–3000g) at 4°C.

3. The plates must be thoroughly rinsed to remove all detergent, as detergent can interfere with the binding reaction.

4. Typically gels in the range of 4 to 6% acrylamide are best for this application. A 6% polyacrylamide gel worked well under these outlined conditions but will be variable depending on the sizes of the DNA-protein complexes.

5. The stability of the DNA-protein complex is also affected by the choice of buffer used for running the polyacrylamide gel, and this parameter must be optimized for individual binding reactions. Generally, Tris-glycine (as recommended by this protocol) works well, but alternatively Tris-acetate (TAE), and Tris-borate (TBE) electrophoresis buffers can be used. The buffer contained in the gel and the electrophoresis buffer should be the same.

6. Although the length of the DNA fragment can vary, our lab has had success with probes ranging from 16 to 92 bp (3,4). The longer the fragment, the more likely it is that multiple binding sites are present, which could result in nonspecific binding (11).

   Consensus sequence databases provide a useful tool for identifying transcription factor binding sites, which can then be used for determination of a DNA fragment to use in probe construction. Examples of such sites are available to the public (www.gene-regulation.com or www.cbrc.jp/research/db/TFSEARCH.html). Additionally, DNase I footprinting can also be employed to identify a binding region of interest more specifically for use in probe generation.

7. Digestion of 1 μg of plasmid DNA will yield sufficient restriction fragment for use in 200 binding reactions.

8. It is most convenient to choose restriction digest sites that leave 5′ overhangs, as these can be filled in using the Klenow fragment and an appropriate [α-³²P]dNTP and other nonlabeled dNTPs and thereby radiolabeled.

9. Oligonucleotide probes are generally designed to be conveniently labeled and are typically 20 to 40 nucleotides in length (11).

10. Various components can be added to the binding reaction to optimize or achieve the desired DNA-protein interaction. The inclusion of a carrier protein, such as BSA, in the binding reaction can increase the stability of some complexes during electrophoresis. When one is using protein extracts, increasing the salt concentration to 50 to 100 mM can reduce the number of nonspecific DNA-protein interactions (24). In contrast, when one is using purified proteins, lower salt concentrations can be used. A final component to consider is the amount of carrier DNA used. At a given protein concentration, too little carrier DNA will cause the entire probe to be bound and not enter the gel. If too much carrier DNA is added, none of the probe will be bound (11). The sequence of the carrier DNA should not resemble the DNA probe used.

    Various parameters of the binding reaction can also be adjusted to optimize the DNA-protein interaction. An important parameter is the temperature at which the reaction takes place. Most binding reactions are performed at 30°C, yet optimal temperature ranges can be from room temperature to 37°C (24). A second variable is the incubation time of the binding reaction. Most reactions will reach equilibrium within 30 min; therefore incubations are typically done from 10 to 30 min (24). Another important parameter to consider is pH. Various DNA-protein interactions are dependent on pH. Most reactions will be stable at pH 7.9, yet interactions need to be empirically determined at ranges of 6.5 to 8.5 (24).

11. Typical amounts of competitor in these assays are 50X, 100X, and 200X molar excess relative to the labeled probe (11).

12. While you are running the gel, it is important that the plates do not get too warm, as denaturation can occur, which interferes with protein binding. The voltage must be decreased if the plates become too warm, or, alternatively, the gel can also be run in a cold room if the gel is run at higher voltages.

13. The amount of antibody used in the supershift assays should be the minimal amount needed to produce a detectable effect. Possible complications in antibody supershifts occur if the antibody recognition site overlaps with the DNA binding domain.

14. Although it is typical to heat-denature protein extracts at 100°C prior to running them on a SDS-PAGE protein gel, it is important to note that this treatment may change the protein structure and alter the DNA-protein complex migration in the gel.