CD Spectroscopy to Study DNA-Protein Interactions

Summary

The interaction of an ATP-dependent chromatin remodeler with a DNA ligand is described using CD spectroscopy. The induced conformational changes on a gene promoter analyzed by the peaks generated can be used to understand the mechanism of transcriptional regulation.

Abstract

Circular dichroism (CD) spectroscopy is a simple and convenient method to investigate the secondary structure and interactions of biomolecules. Recent advancements in CD spectroscopy have enabled the study of DNA-protein interactions and conformational dynamics of DNA in different microenvironments in detail for a better understanding of transcriptional regulation in vivo. The area around a potential transcription zone needs to be unwound for transcription to occur. This is a complex process requiring the coordination of histone modifications, binding of the transcription factor to DNA, and other chromatin remodeling activities. Using CD spectroscopy, it is possible to study conformational changes in the promoter region caused by regulatory proteins, such as ATP-dependent chromatin remodelers, to promote transcription. The conformational changes occurring in the protein can also be monitored. In addition, queries regarding the affinity of the protein towards its target DNA and sequence specificity can be addressed by incorporating mutations in the target DNA. In short, the unique understanding of this sensitive and inexpensive method can predict changes in chromatin dynamics, thereby improving the understanding of transcriptional regulation.

Introduction

Circular dichroism (CD) is a spectroscopic technique that relies on the inherent chirality of biological macromolecules that leads to differential absorption of right-handed and left-handed circularly polarized light. This differential absorption is known as circular dichroism. The technique, therefore, can be used to delineate the conformation of biological macromolecules, such as proteins and DNA, both of which contain chiral centers^1,2.

Electromagnetic waves contain both electric and magnetic components. Both the electrical and the magnetic fields oscillate perpendicular to the direction of wave propagation. In the case of unpolarized light, these fields oscillate in many directions. When the light is circularly polarized, two electromagnetic fields are obtained at 90° phase difference to each other. Chiral molecules show circular optical rotation (birefringence) such that they will absorb the right-handed circularly polarized light and the left-handed circularly polarized light to different extents³. The resulting electrical field will be traced as an ellipse, a function of the wavelength. The CD spectrum is, thus, recorded as ellipticity (q), and the data are presented as Mean Residue Ellipticity as a function of wavelength.

In the case of proteins, the Cα of amino acids (except glycine) is chiral, and this is exploited by CD spectroscopy to determine the secondary structure of this macromolecule⁴. The CD spectra of protein molecules are typically recorded in the Far UV range. α-helical proteins have two negative bands at 222 nm and 208 nm and one positive peak at 193 nm⁴. Proteins with anti-parallel β-sheet secondary structure show a negative peak at 218 nm and a positive peak at 195 nm⁴. Proteins with disordered structures show low ellipticity near 210 nm and a negative peak at 195 nm⁴. Thus, the well-defined peak/bands for different secondary structures make CD a convenient tool to elucidate the conformational changes occurring in the secondary structure of the proteins during denaturation as well as ligand binding.

Nucleic acids have three sources of chirality: the sugar molecule, the helicity of the secondary structure, and the long-range tertiary ordering of DNA in the environment^5,6. The CD spectra of nucleic acids are typically recorded in the 190 to 300 nm range^5,6. Each conformation of DNA, just like proteins, gives a characteristic spectrum, although the peaks/bands can vary by some degrees due to solvent conditions and differences in DNA sequences⁷. B-DNA, the most common form, is characterized by a positive peak around 260-280 nm and a negative peak around 245 nm⁶. The peaks/bands of B-form DNA are generally small because the base pairs are perpendicular to the double helix, conferring weak chirality to the molecule. A-DNA gives a dominant positive peak at 260 nm and a negative peak around 210 nm⁶. Z-DNA, the left-handed helix, gives a negative band at 290 nm and a positive peak around 260 nm⁶. This DNA also gives an extremely negative peak at 205 nm⁶.

In addition to these conformations, DNA can also form triplexes, quadruplexes, and hairpins, all of which can be distinguished by CD spectroscopy. The parallel G-quadruplex give a dominant positive band at 260 nm, while the anti-parallel G-quadruplex gives a negative band at 260 nm and a positive peak at 290 nm, making it easy to distinguish between the two forms of quadruplex structures⁶. Triplexes do not give a characteristic spectrum⁸. For example, the spectra of a 36 nucleotide-long DNA with the potential to form an intramolecular triple helix containing G.G.C and T.A.T base pairs in the presence of Na⁺ show a strong negative band at 240 nm and a broad positive peak. The broad positive peak shows contributions at 266, 273, and 286 nm. The same oligonucleotide in the presence of Na⁺ and Zn⁺ shows four negative bands (213, 238, 266, and 282 nm) and a positive peak at 258 nm. Thus, the spectra of triplex DNA can vary depending upon salt conditions⁸.

In addition to these conformations, CD spectra have enabled the identification of another form of DNA called X-DNA. X-DNA is formed when the DNA sequence contains alternate adenine and thymine residues. The CD spectra of X-DNA contain two negative peaks at 250 and 280 nm. Very little information is available about X-DNA, although it has been speculated to function as a sink for positive supercoiling^6,9. Changes in CD spectra can also reveal details about ligand-protein interactions and, therefore, have been added to the arsenal of molecular methods for detecting drug-protein interactions^{10,11,12,13,14}. CD spectra have also been used to monitor the changes in the secondary structure of proteins during the folding process¹⁵. Similarly, CD spectra can also be used for probing ligand-DNA interactions^16,17.

CD spectroscopy, thus, is an easy, inexpensive method to distinguish between the different forms of DNA conformation, provided there is access to not-so-inexpensive equipment and software. The method is exceedingly sensitive and quick. It only requires a small amount of DNA, giving it an edge over the alternate technique of nuclear magnetic resonance (NMR) spectroscopy. Titrations with ligands and substrates are also easy to perform. The major constraint is that the DNA should be highly pure. It is advisable to use polyacrylamide gel electrophoresis (PAGE)-purified DNA.

The information obtained by CD spectra has been mainly used to deduce protein structural features and to identify distinct DNA conformers. In this study, CD spectra have been used to integrate the results obtained from an in vivo Chromatin Immunoprecipitation (ChIP) experiment to delineate whether the protein of interest/predicted transcription factor can bring about a conformational change in the promoter region of its effector genes. This collaboration aids in the progress of traditional CD spectroscopic techniques by predicting the mechanism of transcription regulation by the predicted transcription factor on and around the transcription start site (TSS) of a promoter.

Chromatin remodeling is a well-defined mechanism known to regulate DNA metabolic processes by making the tightly packed chromatin accessible to various regulatory factors such as transcription factors, components of DNA replication, or damage repair proteins. The ATP-dependent chromatin remodelers, also known as the SWI/SNF family of proteins, are key remodeler proteins present in eukaryotic cells^18,19. Phylogenetic clustering has categorized the SWI/SNF family of proteins into 6 sub-groups²⁰: Snf2-like, Swr1-like, SSO1653-like, Rad54-like, Rad5/16-like, and distant. SMARCAL1, the protein of interest in this study, belongs to the distant sub-group²⁰. This protein has been used to investigate its mode of transcriptional regulation using CD spectroscopy.

Most of the members of the ATP-dependent chromatin remodeling proteins have been shown to either reposition or evict nucleosomes or mediate histone variant exchange in an ATP-dependent manner^21,22. However, some members of this family have not been shown to remodel nucleosomes, e.g., SMARCAL1. Even though studies have shown that SMARCAL1 associates with polytene chromosomes, experimental evidence regarding its ability to remodel nucleosomes is lacking²³. Therefore, it was postulated that SMARCAL1 may regulate transcription by altering the conformation of DNA²⁴. CD spectroscopy provided an easy and accessible method to validate this hypothesis.

SMARCAL1 is an ATP-dependent chromatin remodeling protein that primarily functions as an annealing helicase^25,26,27. It has been postulated to modulate transcription by remodeling the DNA conformation²⁴. To test this hypothesis, the role of SMARCAL1 in regulating gene transcription during doxorubicin-induced DNA damage was studied. In these studies, SMARCAL1 was used for in vivo analysis and ADAAD for in vitro assays^28,29. Previous studies have shown that ADAAD can recognize DNA in a structure-dependent but sequence-independent manner^30,31. The protein binds optimally to DNA molecules possessing double-strand to single-strand transition regions, similar to stem-loop DNA, and hydrolyzes ATP ^30,31.

In vivo experiments showed that SMARCAL1 regulates the expression of MYC, DROSHA, DGCR8, and DICER by binding to the promoter regions^28,29. The region of interaction was identified by ChIP experiments^28,29. The ChIP technique is used to analyze the interaction of a protein with its cognate DNA within the cell. Its goal is to determine whether specific proteins, such as transcription factors on promoters or other DNA binding sites, are bound to specific genomic areas. The protein bound to DNA is first cross-linked using formaldehyde. This is followed by isolation of the chromatin. The isolated chromatin is sheared to 500 bp fragments either by sonication or nuclease digestion, and the protein bound to DNA is immunoprecipitated using antibodies specific to the protein. The cross-linking is reversed, and the DNA is analyzed using either polymerase chain reaction (PCR) or quantitative real-time PCR.

The ChIP results led to the hypothesis that SMARCAL1 possibly mediates transcriptional regulation by inducing a conformational change in the promoter regions of these genes. QGRS mapper and Mfold software were used to identify the potential of these promoter regions to form secondary structures^28,29. QGRS mapper is used for predicting G-quadruplexes³², while Mfold³³ analyzes the ability of a sequence to form secondary structures such as stem-loops.

After secondary structure analysis, further in vitro experiments were performed with recombinant 6X His-tagged Active DNA-dependent ATPase A Domain (ADAAD), the bovine homolog of SMARCAL1, purified from Escherichia coli³⁴. ATPase assays were performed using ADAAD to establish that the identified DNA sequences could act as effectors^28,29. Finally, CD spectroscopy was performed to monitor the conformational changes induced in the DNA molecule by ADAAD^28,29.

To prove that the ATPase activity of the protein was essential for inducing a conformational change in the DNA molecule, either ethylenediamine tetraacetic acid (EDTA) was added to chelate Mg⁺² or Active DNA-dependent ATPase A Domain Inhibitor Neomycin (ADAADiN), a specific inhibitor of the SWI/SNF protein, was added^35,36. This CD spectroscopic technique can be utilized with any purified protein that has been demonstrated by ChIP or any other relevant assay to bind to a predicted genomic region of a promoter.

Protocol

1. Working concentration of the reaction components

1. Prepare the working concentrations of buffers for CD and other reaction components freshly (see Table 1) and keep them at 4 °C before setting up the reactions.
   NOTE: For the CD reactions described in this paper, the working concentrations of components are as follows: Sodium phosphate buffer (pH 7.0) 1 mM, ATP 2 mM, DNA 500 nM, Protein 1 µM, MgCl₂ 10 mM, EDTA 50 mM, ADAADiN 5 µM.

2. ATPase activity

1. Before CD spectroscopy, establish the ATPase activity of the protein in the presence of the DNA molecules to ensure that the protein used in the CD spectroscopy is active and to identify the DNA molecules that are optimally effective in eliciting ATP hydrolysis.
2. Measure the ATPase activity of the protein in the presence of different DNA molecules by an NADH-coupled oxidation assay consisting of the following two reactions.
   1. Mix 0.1 µM ADAAD, 2 mM ATP, 10 nM DNA, and 1x REG buffer in a 96-well plate to a final volume of 250 µL.
      NOTE: The pyruvate kinase enzyme uses the ADP and Pi to convert phosphoenolpyruvate to pyruvate, thus regenerating ATP. This ensures that ATP is always in a saturating concentration in the reaction. In the second reaction, the pyruvate formed by the action of pyruvate kinase is converted by lactate dehydrogenase to lactate. In this reaction, one NADH molecule is oxidized to NAD⁺. The consumption of NADH is measured by measuring the absorbance of the molecule at 340 nm.
   2. Incubate for 30 min at 37 °C in an incubator.
   3. Measure the amount of NAD⁺ at 340 nm using a microplate reader.
   4. To measure the amount of NAD⁺, use the software provided along with the microplate reader.
      1. Click on the NADH assay to measure the absorbance at 340 nm.
      2. Place the 96-well plate on the plate holder in the instrument. Click on the Read Plate button to record the absorbance.
         NOTE: The concentration of NAD⁺ is calculated using the molar extinction coefficient of NADH as 6.3 mM⁻¹ by using eq (1).
         A = εcl (1)
         
         Here, A = Absorbance
         ε = Molar extinction coefficient
         c = Molar concentration
         l = Optical path length in cm

3. Choosing and preparation of CD cuvettes

1. Collect CD spectra in high-transparency quartz cuvettes. Use rectangular or cylindrical cuvettes.
   NOTE: A CD quartz cuvette (nominal volume of 0.4 mL, path-length of 1 mm) was used for all the reactions described in this paper.
2. Use a cuvette cleaning solution to clean the cuvette. Add 1% cuvette cleaning solution in water to make 400 µL of the solution, pour it in the cuvette, and incubate it at 37 °C for 1 h.
3. Wash the cuvette with water several times to clean the cuvette. Take a scan of the water or buffer in the cuvette to check whether it is clean.
   NOTE: The water or buffer must give a reading in the 0 to 1 mdeg range.

4. Preparation of proteins and DNA oligonucleotide

1. Keep the volume of the protein below 50 µL in the reaction to minimize the amounts of the buffer components that sometimes cause the formation of ambiguous peaks. Keep the protein on the ice throughout the experiment to avoid any degradation.
2. Use PAGE-purified DNA oligonucleotides in the reactions.
   NOTE: In the reactions described here, DNA was used both in native as well as heat-cooled forms (fast-cooled (FC) and slow-cooled (SC)). Fast cooling promotes intramolecular bonding in the DNA, yielding more secondary structures. In contrast, slow cooling promotes intermolecular bonding in the DNA, resulting in fewer secondary structures.
3. For fast-cooling, heat DNA at 94 °C for 3 min on the heating block and immediately cool it on ice. For slow-cooling, heat DNA at 94 °C for 3 min and allow it to cool to room temperature at a rate of 1 °C per minute.

5. Setting up control experiments to record the baseline spectra

1. Keep the reaction volume at 300 µL in all the reactions. Set up a total of 5 baseline reactions in 1.5 mL centrifuge tubes, one by one, as follows: i) Buffer + Water; ii) Buffer + MgCl₂ + ATP + Water; iii) Buffer + MgCl₂ + ATP + Protein + Water; iv) iii + EDTA or ADAADiN; v) Buffer + Protein +Water.

6. Setting up the experiments to record CD spectra

1. Set up a total of 5 reactions, one by one, in 1.5 mL centrifuge tubes as follows: i) Buffer + DNA + Water; ii) Buffer + DNA + MgCl₂ + ATP + Water; iii) Buffer + DNA + MgCl₂ + ATP + Protein + Water; iv) iii + EDTA or ADAADiN; v) Buffer + DNA + Protein +Water.

7. Recording scan

1. Turn on the gas and switch on the CD spectrometer.
2. Switch on the lamp after 10-15 min. Switch on the water bath and set the holder temperature at 37 °C.
3. Open the CD spectrum software.
   1. Set the temperature to 37 °C.
   2. Set the wavelength range at 180 - 300 nm.
   3. Set the time per point to 0.5 s.
   4. Set the scan number to 5.
   5. Click on Pro-Data Viewer, make a new file, and rename it with details about the experiment and date.
4. Keep all the reaction components on ice to avoid any degradation. Make the baselines and reactions, one by one, in centrifuge tubes and mix them by pipetting. Transfer the reaction mix to the cuvette carefully, ensuring that there are no air bubbles.
5. If performing a time-course experiment, incubate the reactions at 37 °C for the required time and take the scan. Add EDTA to the buffer containing the DNA, ATP, Mg⁺², and protein to stop ATP hydrolysis.
6. Increase the concentration of EDTA and its incubation time to inhibit ATPase activity completely.
7. Subtract the baselines from the corresponding reactions in the software (e.g., subtract reaction 1 from baseline 1). Smoothen the data either in the CD spectrum software or in the data plotting software. Plot the data in the data plotting software.
   NOTE: Subtracting the baselines from the corresponding reactions will give the net CD spectra of only DNA.

8. Data analysis and interpretation

1. Use the formula given by eq (2) to convert the values obtained in millidegrees to mean residue ellipticity.
     (2)
   Here, S is the CD signal in millidegrees, c is the DNA concentration in mg/mL, mRw is the mean residue mass, and l is the path length in cm.
2. Plot a graph against wavelength and mean residue ellipticity using the data plotting software and analyze the peaks.
3. To plot the graph, select the mean residue ellipticity on the Y-axis and wavelength on the X-axis and plot a straight line graph.
   NOTE: This graph will provide the characteristics peaks of different forms of DNA. The forms of DNA corresponding to the peaks can be identified using existing literature⁶.

Representative Results

ADAAD stabilizes a stem-loop like structure on the MYC promoter
Previous experimental evidence showed that SMARCAL1 is a negative regulator of MYC²⁹. Analysis of the 159 bp long promoter region of the MYC gene by QGRS mapper showed that the forward strand had the potential to form a G-quadruplex (Table 2). Mfold showed that both strands of the MYC DNA could form a stem-loop-like structure (Table 2). A 34 bp long DNA sequence containing the G-quadruplex (G_EC_E) was synthesized. The Mfold structures of the forward and the reverse sequence of the G_EC_E oligonucleotide are shown in Figure 1A,B.

ATPase assays using 6X His-ADAAD showed that fast-cooled G_EC_E was a better effector than the native and the slow-cooled forms. Therefore, fast-cooled G_EC_E was used to record the CD spectra in the absence and presence of ATP and ADAAD. The CD spectra showed that ADAAD induces two positive peaks-one at 258 nm with a shoulder at 269 nm and a larger peak at 210 nm in the DNA (Figure 1C). A dip towards the negative around 240 nm was also observed. This spectrum was similar to the one obtained when a synthetic stem-loop DNA, the optimal effector of ADAAD, was incubated with the protein and ATP (Figure 2A,B). Triplex DNA can give a similar spectrum³⁷, leading to the hypothesis that the protein could be inducing such a structure in this case. ATP forms a coordination complex with Mg⁺², and this cation is essential for ATP hydrolysis. The addition of EDTA chelates Mg⁺², leading to the inhibition of ATP hydrolysis³⁸. Therefore, EDTA was added to the reaction mix to understand whether ATP hydrolysis by ADAAD was important for conformational change. The addition of EDTA to the reaction abrogates this conformation. The CD spectra now have a negative 210 nm peak and a broad positive band with peaks at 230 and 250 nm (Figure 1C).

The importance of ATPase activity was also confirmed using an ATPase-dead mutant of ADAAD. The K241A mutation occurs in the conserved GKT box of motif I, and this mutant has been shown previously to lack the ability to hydrolyze ATP in the presence of DNA³¹. The mutant protein was expressed with a GST tag and purified using glutathione affinity chromatography. The conformational change induced in MYC DNA by this mutant was different from that induced by the wild-type ADAAD. The CD spectrum of the MYC DNA in the presence of the mutant protein possessed a positive 210 nm peak and a negative 260 nm peak (Figure 1D).

ADAAD induces A-form of conformation in DROSHA promoter
The promoter regions of DROSHA, DGCR8, and DICER too were analyzed by QGRS mapper and Mfold software. Both QGRS and MFold showed that the promoter regions possess the potential to form G-quadruplex and stem-like structures (Table 2). The Mfold structures of the forward and reverse oligonucleotides are shown in Figure 3A,B. The ATPase activity showed that the native and heat-cooled DNA behaved similarly. Therefore, the slow-cooled form of these DNA sequences was used for CD studies. The CD spectra showed that ADAAD induces a negative peak at 210 nm and a positive peak at 260 nm in the DROSHA promoter (Figure 3C). This spectrum is a characteristic of A-DNA⁶.

ADAAD induces B-X transition and G-quadruplex formation in the DGCR8 promoter
The Mfold structures of the forward and the reverse strands of the oligonucleotides used are shown in Figure 4A,B. A positive peak at 210 nm and a broad negative peak at 260 nm were observed for DGCR8 pair 1 (Figure 4C). This spectrum is characteristic of B-X transition⁶. The Mfold structures of the forward and the reverse strands of the oligonucleotides used are shown in Figure 4D,E. The CD spectra of DGCR8 pair 7 showed a strong positive peak at 210 and 270 nm and a negative peak at 250 nm (Figure 4F). This spectrum is characteristic of parallel G-quadruplex DNA structures⁶.

ADAAD induces A-X transition in DICER promoter
The Mfold structures of the forward and reverse oligonucleotides are shown in Figure 5A,B. A positive peak at 210 nm and two negative peaks-one at 230 nm and the other at 260 nm peak-were observed for the DICER pair 1 (Figure 5C). These peaks are characteristic of A-X DNA transition^6,9. All the CD spectra peaks and the forms of DNA having specific roles in the transcription process have been summarized in Table 3.

Figure 1: ADAAD alters the conformation of G_EC_E DNA. Mfold structures were predicted for the (A) forward strand and (B) reverse strand. (C) CD spectra of G_EC_E alone (black), G_EC_E incubated with ATP and ADAAD before (red) and after adding EDTA (blue). (D) CD spectra of G_EC_E incubated with ATP and GST-tagged ADAAD before (black) and after adding EDTA (red) as well as CD spectra of G_EC_E incubated with ATP and GST-tagged K241A mutant (blue). This figure has been modified from ²⁹. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; CD = circular dichroism. Please click here to view a larger version of this figure.

Figure 2: ADAAD alters the conformation of slDNA. (A) Mfold structure predicted for the stem-loop DNA. (B) CD spectra of slDNA alone (black), slDNA incubated with ATP and ADAAD (red). This figure has been modified from²⁹. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; slDNA = stem-loop DNA; CD = circular dichroism. Please click here to view a larger version of this figure.

Figure 3: ADAAD alters the conformation of DROSHA pair 5 DNA. Mfold structure predicted for the (A) forward strand and (B) reverse strand. (C) CD spectra of DROSHA pair 5 DNA alone (black), DROSHA pair 5 DNA incubated with ATP and ADAAD (red). This figure has been modified from²⁸. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; CD = circular dichroism. Please click here to view a larger version of this figure.

Figure 4: ADAAD alters the conformation of DGCR8 pair 1 and 7 DNA. Mfold structures predicted for the (A) forward strand and (B) reverse strand of DGCR8 pair 1 oligonucleotide. (C) CD spectra of DGCR8 pair 1 alone (black), DGCR8 pair 1 incubated with ATP and ADAAD (red). Mfold structures predicted for the (D) forward strand and (E) reverse strand of DGCR8 pair 7 oligonucleotide. (F) CD spectra of DGCR8 pair 7 alone (black), DGCR8 pair 7 incubated with ATP and ADAAD (red). This figure has been modified from²⁸. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; CD = circular dichroism. Please click here to view a larger version of this figure.

Figure 5: ADAAD alters the conformation of DICER pair 1 DNA. Mfold structure predicted for the (A) forward strand and (B) reverse strand of DICER pair 1 oligonucleotide. (C) CD spectra of DICER pair 1 alone (black), DICER pair 1 incubated with ATP and ADAAD (red). This figure has been modified from²⁸. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; CD = circular dichroism. Please click here to view a larger version of this figure.

5x REG buffer
Component	Working concentration
Lactate dehydrogenase (LDH)	50 units/mL
Magnesium acetate (Mg(OAc)2)	30 mM
Phosphoenolpyruvate (PEP)	6.8 mg/mL
Pottasium acetate (KOAc)	300 mM
Pyruvate kinase (PK)	50 units/mL
Tris acetate (Tris-OAc)	125 mM
β-mercaptoethanol (β-ME)	25 mM

Table 1: Buffer components.

Oligonucleotides		Forward sequence		ΔG (m-Fold) Kcal/mol	G-score (QGRS mapper)	Reverse sequence		ΔG kcal/mol (Mfold prediction)	G-score (QGRS mapper)
slDNA		GCGCAATTGCGCTCGA CGATTTTTTAGCGCAA TTGCGC		-16.36	-	-		-	-
MYC GECE		CGCGCGTGGCGTGG CGGTGGGCGCGCA GTGCGTT		-8.64	19	AACGCACTGCGCGCC CACCGCCACGCCA CGCGCG		-5.74	-
DROSHA Pair 5		GGCCAGGCACGGTG GCTTATGCCTGTAAT CCCAGCCCTTTGGGA GGCTGAGGCAGG		-10.18	19, 19	CCTGCCTCAGCCTCC CAAAGGGCTGGGATT ACAGGCATAAGCCAC CGTGCCTGGCC		-10.95	-
DGCR8 Pair 1		GCCACCGTGCCCGG CCGAACCACCGTGC CCGGCCGAACCC		-7.9	-	GGGTTCGGCCGGG CACGGTGGTTCGG CCGGGCACGGTGGC		-9.4	21
DGCR8 Pair 7		TCATACTGCCGCTG GGTTGGGCGGGAG GCTGCAGCGGGAG		-7.99	33	TCCCGCTGCAGCCT CCCGCCCAACCCAG CGGCAGTATGAC		-5.43	-
DICER Pair 1		AAGTGCTCAGGAG CCATGTGGGGCTG GTGGCCCCAAACTG		-8.82	19	CAGTTTGGGGCCAC CAGCCCCACATGGC TCCTGAGCACTT		-9.16	-

Table 2: Oligonucleotide sequences. All the sequences are in the 5'-3' direction. Abbreviations: slDNA = stem-loop DNA.

Oligonucleotides	CD Spectra Peaks in nm (After incubating with ATP and ADAAD)	Form of DNA	Role in Transcription
slDNA	+215, +250, +272	Triplex	Repression
MYC GECE	+210, +258, +269	Triplex	Repression
DROSHA Pair 5	-210, +260	A-DNA	Initiation/Activation
DGCR8 Pair 1	+210,-260	B-X Transition	Positive supercoiling/Activation
DGCR8 Pair 7	+210, -250, +270	G-quadruplex	Activation
DICER Pair 1	+210, -230, -257	A-X Transition	Positive supercoiling/Activation

Table 3: CD spectra peak corresponding to different forms of DNA with their role in transcription. Abbreviations: ADAAD = Active DNA-dependent ATPase A Domain; CD = circular dichroism.

Discussion

The purpose of this article is to introduce the CD spectroscopy technique as an approach to study the conformational changes occurring in the DNA in the presence of ATP-dependent chromatin remodeling proteins and to link these conformational changes to gene expression. CD spectroscopy provides a fast and easily accessible method to study the conformational changes in DNA.

A crucial point to be considered for this technique is the purity of the DNA and protein. It is advisable to ensure that both DNA and protein are >95% pure. PAGE-purified oligonucleotides must be used in the assay, and the protein should be preferably affinity-purified to >95% purity. The other critical parameter is that the cuvette should be clean such that the baseline reading does not exceed 1 mdeg. The buffers should be made using autoclaved water, and the baseline reading of the buffer should not exceed 1 mdeg. To study the conformation of the promoter, it is essential to identify the regions where the protein binds. Therefore, it is advisable to perform ChIP experiments using the protein of interest as this process helps to identify DNA sequences present in the promoter region of the effector gene bound by the protein. Once the region is identified, the ability of the sequence to adopt specific structures can be analyzed using available bioinformatics tools. This is important as the ChIP primers are usually 200 bp long and may have multiple conformations. Therefore, using bioinformatics tools to identify the structures would help shorten the length of the oligonucleotide to one structure.

Finally, if the protein of interest is an ATP-dependent chromatin remodeling protein, the ability of the oligonucleotides to act as an effector must be checked using ATPase assays. In both CD spectroscopy and ATPase assays, care should be taken to ensure that saturating concentrations of ligands are used in the reaction. If possible, the dissociation constant (K_d) for the protein-ligand interaction should be calculated before proceeding with CD spectroscopy. Numerous methods are available for calculating the binding parameter. Using ATPase assays, the Michaelis-Menten constant (K_M) can be calculated by titrating increasing concentrations of DNA. The K_M, in many cases, can be approximated to the binding constant. If the protein is fluorescent, binding constants can be calculated using fluorescence spectroscopy. If neither of these techniques is feasible, the electrophoresis mobility shift assay (EMSA) can be used.

The main problem with CD spectroscopy arises when the cuvettes are not cleaned or when the reagents are impure. If the baseline is too high, it is advisable to clean the cuvettes. Numerous cuvette cleaning solutions are available. Placing the cuvettes in a dilute acid solution for 16-24 h also helps clean the cuvette. It is advisable to purchase reagents that are >95% pure and to use double-distilled and autoclaved water. Baseline drift is another potential problem. If doing a long-term experiment, it is advisable to periodically check the baseline. The peaks of DNA when studying protein-DNA interactions may not exactly correspond to the peaks/bands obtained with DNA alone. Protein peaks are usually observed in the Far UV range between 190 and 230 nm. Therefore, peaks below 250 nm might have interference from the protein peak and might not provide reliable information. DNA can adopt a variety of non-B conformations depending upon the sequence. The longer the DNA sequence, the higher the chance of multiple conformations co-existing within the DNA oligonucleotide. This can make analysis difficult. Hence, it is advisable to use shorter oligonucleotides corresponding to the potential structures predicted by the bioinformatic tools.

The other major drawback of CD spectroscopy is that it does not allow for atomic-level structure analysis, and the obtained spectrum is insufficient to identify the only viable structure. For example, both X-ray crystallography and protein NMR spectroscopy provide atomic resolution data, whereas CD spectroscopy provides less detailed structural information. However, CD spectroscopy is a rapid approach that does not necessitate vast quantities of proteins or considerable data processing. As a result, CD may be used to investigate a wide range of solvent variables such as temperature, pH, salinity, and the presence of various cofactors. It can also be used to monitor structural changes (due to complex formation, folding/unfolding, denaturation due to temperature, denaturants, and changes in amino acid sequence/mutation) in dynamic systems. By connecting it to the stop-flow apparatus, it can also be used to study the kinetics of protein/DNA-ligand interactions.

Well-characterized DNA conformers include A/B/Z DNA, triplex, hairpin, and G-quadruplexes. All these forms of DNA are associated with an open DNA conformation, i.e., unwound DNA that serves as the sink for negative supercoiling. Transcription is associated with negative supercoiling as the formation of an open complex is a prerequisite for the movement of RNA polymerase. Therefore, transcription of most genes involves increased negative supercoiling in the promoter region. Studies have shown that nucleosome unfolding leads to A-DNA conformation, which, however, is unstable³⁹. One possibility is that the protein of interest, e.g., SMARCAL1, binds to such structures and stabilizes them, thus facilitating transcription, as seen in the case of DROSHA promoters. The guanine quadruplex is based on guanine tetrads bound by Hoogsteen hydrogen bonds. In silico analysis has confirmed that G4-forming sequences are notably enriched proximal to gene promoters and at transcription start sites. These G4 sequences can both activate and repress transcription.

In the case of the MYC promoter⁴⁰, the formation of G-quadruplex acts as a repressor, while in the case of human vascular endothelial growth factor (VEGF), the G-quadruplex structure functions as a docking site for transcription factors⁴¹, thus activating the expression of this gene. In the case of SMARCAL1, the G-quadruplex structure was observed when ADAAD interacted with DGCR8 pair 7 promoter sequences. As the occupancy of SMARCAL1 and RNAPII increased on this primer pair, it is hypothesized that the formation of G-quadruplex, in this case, correlates with transcription activation of this gene. DNA can also be positively supercoiled, and the progress of RNA polymerase is known to generate positive supercoiling in front of it. This transcription-generated (+) supercoiling can disrupt or eliminate road-block proteins, destabilizing nucleosome structures to make the DNA more accessible to RNA polymerase. The X-DNA is a conformation of DNA that acts as sinks for positive supercoiling. The striking feature of an X-DNA is it can form in a sequence-specific manner on the promoter of a gene. In the case of SMARCAL1, ADAAD induced A-X and B-X transitions in an ATP-dependent manner in the DICER and DGCR8 promoters, respectively. Combined with in vivo data where increased SMARCAL1 and RNAPII occupancy was found on these promoters in the presence of doxorubicin-induced DNA damage, it can be hypothesized that X-DNA formation facilitates transcription by removing barriers/blocks. Triple DNA helices do not have a characteristic spectrum. The CD spectra of the DNA sequences present in the MYC promoter and the synthetic stem-loop DNA showed two positive peaks-one at 258 nm with a shoulder at 269 nm and a larger peak at 210 nm in the DNA. This type of spectrum can be obtained in the case of triplexes³⁷. Triplexes are difficult to unwind and, therefore, are known to block transcription⁴². Hence, it is hypothesized that the formation of this structure in the c-MYC promoter by SMARCAL1 leads to repression of transcription.

It should be noted that ATP also binds to ATP-dependent chromatin remodeling proteins. The conserved arginine present in the motif VI of the helicase domain of these proteins interacts via electrostatic interactions with the γ-phosphate of the protein⁴³. In the case of ADAAD, the K_d of protein-ATP interaction is (1.5 ± 0.1) × 10^-6 M³⁸. The binding of ATP induces a conformational change in the protein such that the affinity of the DNA increases. The binding of DNA also induces a conformation change in the protein leading to an increased 10-fold affinity for ATP³¹. For example, in the case of ADAAD, bands/peaks are observed at -212 nm and -222 nm. ATP also gives bands at 197 nm, +210 nm, -222 nm, -247 nm, and -270 nm. These must be subtracted from the spectra of DNA + ADAAD + ATP to obtain the "net" conformation of the DNA in the presence of the ligands.

Thus, this paper shows the convenience of CD spectroscopy for the study of the conformational changes occurring in the DNA in the presence of ATP-dependent chromatin remodeling proteins. Correlating the changes in the DNA conformation with ChIP data can provide the investigators with information regarding how the DNA conformers activate/repress transcription.

Materials

Name	Company	Catalog Number	Comments
2-Mercaptoethanol	Fisher scientific	O3446I-100
Adenosine 5′-triphosphate disodium salt hydrate	Sigmaaldrich	A2383
CD Quartz Cuvette	STARNA	21-Q-1
Chirascan V100 CD spectrometer	Applied Photophysics	Not available
EDTA Disodium Salt Dihydrate	SRL	43272
Glutathione Sepharose 4B	GE Healthcare	17-0756-01	Glutathione affinity chromatography
Hellmanex III cleaning solution	Hellma	9-307-011-4-507
L-Lactic Dehydrogenase	Sigmaaldrich	L2625
Magnesium Acetate Tetrahydrate	Fisher scientific	BP215-500
Magnesium Chloride Hexahydrate	Fisher scientific	M33-500
NADH disodium salt	Sigmaaldrich	10107735001
Phosphoenolpyruvate Monocyclohexylammonium Salt	SRL	40083
Potassium Acetate	Fisher scientific	P178-3
Pyruvate Kinase	Sigmaaldrich	P1506
Sodium Phosphate Dibasic Anhydrous	Fisher scientific	S374-500
Sodium Phosphate Monobasic Monohydrate	Fisher scientific	S369-500
Synergy HT microplate reader	BioTek	Not available
Tris Base	Fisher scientific	BP152-500