Racemic phosphorothioate as a tool for NMR investigations of protein-DNA complexes

Abstract

A major driving force for protein-nucleic acid association is electrostatic interactions via ion pairs of the positively charged basic side chains and negatively charged phosphates. For a better understanding of how proteins scan DNA and recognize particular signatures, it is important to gain atomic-level insight into the behavior of basic side chains at the protein-DNA interfaces. NMR spectroscopy is a powerful tool for investigating the structural, dynamic, and kinetic aspects of protein-DNA interactions. However, resonance assignment of basic side-chain cationic moieties at the molecular interfaces remains to be a major challenge. Here, we propose a fast, robust, and inexpensive approach that greatly facilitates resonance assignment of interfacial moieties and also allows for kinetic measurements of protein translocation between two DNA duplexes. This approach utilizes site-specific incorporation of racemic phosphorothioate at the position of a phosphate that interacts with a protein side chain. This modification retains the electric charge of phosphate and therefore is mild, but causes significant chemical shift perturbations for the proximal protein side chains, which facilitates resonance assignment. Due to the racemic nature of the modification, two different chemical shifts are observed for the species with different diastereomers R_P and S_P of the incorporated phosphorothioate group. Kinetic information on the exchange of the protein molecule between R_P and S_P DNA duplexes can be obtained by ¹⁵N_z exchange spectroscopy. We demonstrate the applications of this approach to the Antennapedia homeodomain–DNA complex and the CREB1 basic leucine-zipper (bZIP)–DNA complex.

Introduction

Protein-DNA interactions are essential for gene regulation, transcription, DNA repair, and replication. In these processes, DNA-binding proteins dynamically scan DNA, recognize particular signatures, and cooperate with other proteins to perform their functions.^1–4 For the molecular association, arginine (Arg) and lysine (Lys) side chains play an important role through electrostatic interactions with DNA phosphates, forming ion pairs.^5,6 Interestingly, Arg and Lys appear to have considerably different properties in DNA recognition by proteins. Compared to Lys side chains, Arg side chains exhibit a stronger propensity to form hydrogen bonds to DNA bases.^7,8 Cation–π interactions with DNA base are also more common for Arg side chains.^9,10 For a better understanding of the functions of DNA-binding proteins, it is important to gain insight at an atomic level into how the interfacial basic side chains behave while the proteins are bound to DNA.

NMR spectroscopy can provide atomic-level information about the structure, dynamics, and kinetics of protein-DNA interactions.^11,12 Some NMR studies showed that interfacial Arg and Lys side chains exhibit differences in mobility at protein-nucleic acid interfaces.^13–19 Interfacial Lys NH₃⁺ groups were found to generally be more mobile than interfacial Arg guanidinium groups. The mobility of Arg side chains appears to be more sensitive to interactions with nucleic acids and can become rigid more easily. Although the NMR studies have provided important insight into how individual side chains behave, one bottleneck in such investigations is the difficulty in resonance assignment of the Lys NH₃⁺ or Arg guanidinium groups. There are double or triple resonance methods that should in principle enable resonance assignment of the cationic moieties of Lys and Arg side chains.²⁰ In practice, however, these NMR methods may not suffice due to rapid hydrogen exchange, rapid transverse relaxation (e.g. due to conformational exchange or slow molecular tumbling), or poor dispersion of CH₂ resonances. Resonance assignment of the Lys NH₃⁺ and Arg guanidinium groups are often challenging and time-consuming, especially for protein-DNA complexes larger than 20 kDa, even when these cationic moieties of Lys and Arg side chains exhibit well-resolved ¹H-¹⁵N correlation spectra.

To overcome this difficulty, Anderson et al. proposed an alternative approach that utilizes site-specific oxygen-to-sulfur substitutions of the O_P1 and O_P2 atoms of DNA phosphate.²¹ The phosphorodithioate (-O-PS₂⁻-O-) modification retains the charge on the phosphate and does not diminish the protein-DNA interactions but does cause large perturbations on ¹H and ¹⁵N resonances for the side-chain groups that interact with the phosphate at the modification site. These perturbations facilitate resonance assignment for the interfacial side-chain groups in a close proximity to the DNA phosphate. Despite its potential efficacy, the approach utilizing phosphorodithioate has a severe practical disadvantage in that the incorporation of this modification into DNA requires special skills and, as far as we know, is currently unavailable in commercial custom DNA synthesis services.

In the current paper, we present a similar, but practically easier and more useful approach and apply it to NMR investigations of Arg side chains at protein-DNA interfaces. Our approach utilizes the incorporation of racemic phosphoromonothioate (-O-PSO⁻-O-) at a particular phosphate position of DNA (Figure 1). This modification is inexpensive and available in many commercial custom DNA synthesis services. Due to the racemic nature, phosphoromonothioate (hereafter, simply referred to as phosphorothioate) causes two signals from the species with two different diastereomers R_P and S_P (Figure 1). The simultaneous presence of these species in the same solution allows us to obtain kinetic information on the process of a protein’s translocation between two DNA duplexes.

Figure 1.

Racemic phosphorothioate in a DNA strand. This modification is incorporated at a specific phosphate site during chemical DNA synthesis.

Materials and Methods

Protein preparation

The synthetic DNA encoding the human CREB1 basic leucine-zipper (bZIP) domain (the residues P278-D341) bearing C300S, C310S, and C337S mutations and an HRV-3C protease recognition sequence at the N-terminus was inserted into the KpnI/EcoRI sites of the pET-32a vector. Escherichia coli strain BL21(DE3) cells were transformed with this plasmid and cultured at 37°C in a 4 L of M9 minimal medium (100 μg/ml ampicillin) supplemented with 1 g/L ¹⁵N ammonium chloride as the sole nitrogen source. The expression of the thioredoxin-bZIP fusion protein was induced at OD₆₀₀ ≈ 0.8 using 0.4 mM isopropyl β-D-thiogalactopyranoside. The culture was continued at 18 °C overnight (~16 hours). The harvested cells were suspended with a buffer of 50 mM Tris•HCl (pH 7.5), 1 M NaCl, 1% Triton X-100, and Roche complete EDTA-free protease inhibitors (3 tablets in 120 ml) and were disrupted by sonication. The supernatant of the lysate was loaded onto a HisPrep FF column (GE Healthcare) and eluted with a gradient of 5–400 mM imidazole in a buffer of 50 mM Tris•HCl (pH 7.5), 500 mM NaCl, 1 mM β-mercaptoethanol, and 10% glycerol. Fractions containing the fusion protein were pooled and concentrated to ~50 ml in a buffer of 20 mM Tris•HCl (pH 7.5), 500 mM NaCl, 3 mM β-mercaptoethanol, and 10% glycerol, and then cleaved using 100 units of the HRV-3C protease (Genway) at 4°C for 8 hours. The thioredoxin tag and the remaining thioredoxin-fusion protein were removed by passing the reaction mixture through a HisPrep FF column. The CREB1 bZIP domain was further purified through Resource-S cation-exchange column chromatography, in which the protein was eluted with a gradient of 0–1500 mM NaCl in a buffer of 20 mM Tris•HCl (pH 7.5), 1 mM EDTA, and 2% glycerol. The CREB1 bZIP concentration was measured using UV absorbance at 280 nm along with an extinction coefficient of 2980 M⁻¹cm⁻¹ calculated by the ExPASy protparam tool.²²

The ¹⁵N-labeled Antennapedia (Antp) homeodomain (60 residues) bearing the C39S mutation was prepared as previously described.²³

DNA preparation

The unmodified 15-bp DNA duplex used for the Antp homeodomain was prepared as previously described.²³ The unmodified palindromic 20-bp DNA duplex for the CREB1 bZIP was purchased as a HPLC-grade material from Integrated DNA Technologies (Coralville, Iowa). DNA strands containing a racemic phosphorothioate were purchased as desalt-grade materials from Sigma-Genosys (Woodlands, Texas). The sequences and the modification sites are given in figures shown below. The modified DNA strands were purified through an S-100 size-exclusion column (GE Healthcare; bed volume, 300 ml). To retain the intact molar ratio of R_P to S_P, size-exclusion chromatography was used for the purification instead of more commonly used anion-exchange chromatography because the latter may separate DNA strands with R_P from those with S_P species^24,25 The DNA duplexes were quantified using UV absorbance at 260 nm together with the extinction coefficients calculated from the sequences by the method of Tataurov et al.²⁶

NMR samples of protein-DNA complexes

The NMR samples for 0.3 mM Antp homeodomain-DNA complexes in a buffer of 20 mM succinate-d₄•KOH (pH 5.8), 100 mM KCl, and 0.4 mM NaF was prepared as previously described.¹⁷ In preparation of the NMR samples for the CREB1 bZIP-DNA complexes, both the purified protein and DNA samples were buffer-exchanged separately using an Amicon Ultra filter with a 3 kDa cutoff (Millipore) and a buffer containing 20 mM cacodylate•NaOH (pH 5.8), 15 mM MgCl₂, and 300 mM NaCl. The protein and DNA (~20 μM each) in this buffer were mixed at a 1:1.5 molar ratio of the CREB1 bZIP dimer to DNA. The sample was buffer-exchanged again using an Amicon Ultra filter with a 10 kDa cutoff and a buffer of 20 mM cacodylate•NaOH (pH 5.8), 15 mM MgCl₂, and 100 mM NaCl. The solution of the complex was then concentrated to obtain a 450 μl solution of 0.15 mM CREB1 bZIP-DNA complex. 5 % D₂O was added for the NMR lock signal. Each sample was transferred into a 5-mm NMR tube (Norell).

NMR experiments

All NMR experiments were performed using a Bruker Avance III spectrometer equipped with a TCI cryogenic probe operated at the ¹H frequency of 800 MHz. The sample temperature was set to 25˚C for the NMR experiments on the Antp homeodomain–DNA complex and to 35˚C for the experiments on the CREB1 bZIP–DNA complexes. The two-dimensional (2D) ¹H-¹⁵N heteronuclear in-phase single quantum coherence (HISQC) experiments for Arg side-chain NεH groups of the protein-DNA complexes were performed as previously described.²⁰ For moieties that undergo rapid hydrogen exchange (e.g. Lys NH₃⁺ and Arg guanidinium groups), HISQC has significant advantages over HSQC, as pointed out previously.^20,27,28 The ¹⁵N_z exchange experiments,²⁹ which were modified to implement the HISQC principle to maintain the in-phase ¹⁵N transverse magnetizations, were conducted using mixing times between 0.05 and 1.8 s. The kinetic rate constants for protein translocation between the two DNA duplexes were determined using MATLAB software, as previously described.³⁰ The spectra were processed with the NMRPipe software³¹ and analyzed using the NMRView software³².

Results

We utilized racemic phosphorothioate as a tool for NMR investigations and applied it to the 17 kDa complex of the Antp homeodomain with a 15-bp DNA duplex and the 29 kDa complex of the CREB1 bZIP homodimer with a 20-bp DNA duplex. Since the Antp homeodomain–DNA complex had been characterized extensively in previous NMR studies,^{15,17,23,25,30,33–37} this complex served as an ideal model case for us to examine the feasibility of the approach. Based on our proof of concept using the model system, we applied the same approach to the CREB1 bZIP–DNA complex and investigated the Arg side chains at the protein-DNA interfaces.

Phosphorothioate at the R43 interaction site in the Antp homeodomain–DNA complex

We incorporated a racemic phosphorothioate into a 15-bp DNA duplex at the position indicated in Figure 2A. In the crystal structures of Antp homeodomain–DNA complexes (PDB 4XIC, 4XID, 5JLX, 5JLW, and 9ANT),^23,25,38 the R43 side chain forms an intermolecular ion pair with the phosphate group at this position. The R3 side chain is also located near the same phosphate in two crystal structures (4XIC and 4XID), though this side chain is disordered and unresolved in the other three crystal structures.

Figure 2.

Phosphorothioate at the R43 interaction site of the Antp homeodomain–DNA complex. (A) The 15-bp DNA duplex with phosphorothioate. The modified DNA involves a racemic mixture of the R_P and S_P diastereomers at the phosphorothioate (see Figure 1). One of the crystal structures of the Antp HD- DNA complex (PDB 4XIC) is shown to indicate the Arg side chains and the modification site. (B) Overlaid ¹H-¹⁵N HISQC spectra recorded for the Arg side-chain N_εH groups in the original (red) and modified (blue) DNA complexes of the ¹⁵N-labeled Antp HD. (C) A ¹H-¹⁵N correlation spectrum recorded through the ¹⁵N_z exchange experiment using a mixing time of 0.9 s. Note that the R43 N_εH group exhibited exchange cross peaks, which arise from translocation of the Antp homeodomain between the R_P-and S_P-phosphorothioate DNA duplexes.

We prepared the Antp homeodomain-DNA complexes with and without the phosphorothioate modification and compared the NMR spectra. Figure 2B shows the overlaid ¹H-¹⁵N HISQC spectra²⁰ recorded for Arg side-chain N_εH groups for both the unmodified and modified complexes at 25˚C. The majority of side-chain N_εH groups exhibited virtually no chemical shift perturbation by the modification. However, in the spectrum for the modified complex, two additional signals were observed and the signal from the R3 N_εH group was significantly perturbed. The presence of additional signals is reasonable because the modification involves two diastereomers R_P and S_P of phosphorothioate. The interfacial Arg NεH groups near the modification site should in principle exhibit two signals due to these two species.

It should be noted that the DNA-bound proteins undergo exchange between different DNA duplexes in solution. In fact, by ¹⁵N_z exchange spectroscopy, we were able to observe exchange cross peaks indicating the exchange of the Antp homeodomain between the R_P-phosphorothioate DNA duplex and the S_P-phosphorothioate DNA duplex (Figure 2C). The ¹⁵N_z exchange data correlate signals from the R43N_εH group. The ¹⁵N chemical shift difference between the two auto-cross peaks are relatively large (0.6 ppm), which is qualitatively consistent with the location of the phosphorothioate modification. Unfortunately, because one of the exchange cross peaks overlaps with the signal from R29, it is difficult to determine the kinetic rate constants for the exchange process through a simultaneous fitting to two auto and two exchange cross peaks. As shown in Figure 1C, the ¹⁵N_z exchange spectrum recorded at the mixing time of 0.9 s showed comparable signal intensities for the auto and exchange cross peaks. This suggests that the timescale of the exchange between R_P and S_P complexes are qualitatively consistent with the previously determined residence time (0.2 s) of the Antp homeodomain on the unmodified DNA duplex at 100 mM KCl.³⁰ These results represent a proof of concept for our approach to utilize racemic phosphonothioate as a tool for NMR investigations.

Application to interfacial Arg side chains of the CREB1 bZIP-DNA complex

We applied the same approach to the interfacial Arg side chains of the CREB1 bZIP-DNA complex. The CREB1 bZIP domain forms a homodimer and bind to palindromic DNA. Figure 3A shows the crystal structure of the CREB1 bZIP-DNA complex (PDB 1DH3), the sequence of the 20-bp palindromic DNA, and a 2D ¹H-¹⁵N HISQC spectrum recorded for Arg side-chain N_εH groups of the complex. Initially, we tried to assign ¹H and ¹⁵N resonances of Arg N_εH groups using conventional methods;²⁰ however, this attempt was difficult due to rapid NMR relaxation that diminished the quality of HCCH-TOCSY³⁹ and other side-chain ¹³C correlation spectra. We were able to resolve this issue by using the phosphorothioate-based approach. Four modified palindromic DNA duplexes containing phosphorothioates at different and sequential positions (Sites 1–4) were prepared, as indicated in Figure 3A. These modification sites were chosen because the phosphate groups at those positions are close to 5 interfacial Arg side chains R286, R289, R294, R298, and R301. Site 1 is close to R301 and R298; Site 2 is close to R298 and R294; Site 3 is close to R284; and Site 4 is close to R286 and R289.

Figure 3.

Application of racemic phosphorothioate for the resonance assignments of the interfacial Arg N_εH groups of the CREB1 bZIP-DNA complex. (A) The 20-bp DNA duplex used for the CREB1 bZIP-DNA complex. Modification sites 1–4 are indicated by arrows. The Arg ¹H-¹⁵N HISQC spectra and the crystal structure (PDB 1DH3)⁴⁰ of the unmodified DNA complex of the CREB1 bZIP dimer are also shown. (B, C, D, E) Arg HISQC and ¹⁵N_z exchange spectra recorded for the complexes with modified DNA duplexes. Each spectrum is overlaid with the spectrum recorded with the unmodified complex as a reference. Black annotations shown in the HISQC spectra indicate the signals that were significantly perturbed by the modification. Red annotations in the ¹⁵N_z exchange spectra indicate the sets of auto and exchange cross peaks arising from protein translocation between R_P- and S_P-phosphorothioate DNA sites. The ¹⁵N_z exchange spectra were recorded with a mixing time of 0.7 s. The structure around each modification site is also shown.

Figure 3B–E shows the ¹H-¹⁵N HISQC and ¹⁵N_z exchange spectra recorded for Arg side-chain N_εH groups of the CREB1 bZIP-DNA complexes with phosphorothioates at Sites 1–4. The ¹H-¹⁵N HISQC spectra of the modified complexes are overlaid with the corresponding spectrum recorded for the unmodified complex. Compared to the unmodified complex, the modified complexes exhibited a larger number of HISQC signals due to the presence of the R_P and S_P diastereomers that causes the splitting of signals from the Arg side chains near the modification site. Some signals observed for the unmodified complex were clearly perturbed. As seen for the Antp homeodomain–DNA complex, the ¹⁵N_z exchange spectra of the CREB1 bZIP-DNA complexes containing the modification clearly showed exchange cross peaks arising from the exchange of the protein between R_P-phosphorothioate and S_P-phosphorothioate DNA sites.

In conjunction with the proximity information from the crystal structure, these NMR spectra for the modified and unmodified complexes allowed us to unambiguously assign the ¹H_ε and ¹⁵N_ε resonances of R294, R298, and R301 side chains at the protein-DNA interface. Although the other two interfacial arginine side chains, R286 and R289, exhibited exchange cross peaks (Figure 3E), it was difficult to unambiguously distinguish two possibilities regarding the resonances of R286 and R289 because these two side chains exhibited similar perturbations upon the modification of Site 4. Additional information from 3D ¹⁵N-edited NOESY and 3D ¹H_ε/¹⁵N_ε/¹³C_δ/¹³C_γ correlation spectra⁴¹ recorded for the unmodified complex allowed us to distinguish the two possibilities. The phosphorothioate-based approach greatly facilitated the assignment of the ¹H_ε and ¹⁵N_ε resonances of all the 5 interfacial Arg side chains.

Kinetics of protein translocation between DNA sites

Although the simultaneous presence of the R_P and S_P diastereomers of phosphorothioate causes an increase in complexity of NMR spectra, these species allows us to conduct quantitative kinetic investigations of protein translocation between two DNA duplexes. By analyzing the ¹⁵N_z exchange spectra recorded using various mixing times, we can determine the kinetic rate constants for the protein translocation processes.⁴² Figure 4A shows a series of the ¹⁵N_z exchange spectra and intensities of the auto and exchange cross peaks at 8 different mixing times for the R298 and R301 N_εH groups of the CREB1 bZIP–DNA complex containing phosphorothioate at Site 1. The data show that the exchange cross peaks initially intensified and subsequently weakened, while the auto cross peaks continuously weakened. This is a classical pattern for exchange spectroscopy. Through simultaneous fitting of the auto and exchange cross-peak intensities, we determined the kinetic rate constant k for protein translocation. The fitting was obtained using an in-house MATLAB script that solves the McConnell equations for a two-site exchanging system, assuming the same rate constant for R_P → S_P and S_P → R_P processes. The kinetic rate constant k, the intensity of the auto peaks at time 0, and the intrinsic longitudinal relaxation rate were optimized as fitting parameters. The best-fit curves are shown as solid lines in the time course data (Figure 4B). The rate constant k was determined to be 0.53 ± 0.02 s⁻¹ and 0.46 ± 0.02 s⁻¹ from the datasets for R298 and R301, respectively. The slight difference between these rate constants are probably due to the fact that the R298 exchange cross-peak intensities were impacted by the auto peaks because the ¹H chemical shift difference between the two species are relatively small for this residue (see Figure 4A).

Figure 4.

Kinetic analysis of protein translocation between R_P-phosphorothioate and S_P-phosphorothioate DNA sites using the ¹⁵N_z exchange spectroscopy data. Shown are data for the R298 and R301 side-chain N_εH groups in the CREB1 bZIP-DNA complex containing a racemic phosphorothioate at Site 1 (see Figure 3). (A) Change of ¹H-¹⁵N correlation spectra recorded in the ¹⁵N_z exchange experiment using some different mixing times T_mix. Auto and exchange cross peaks are annotated. The stronger auto signals are annotated in black and the weaker auto signals are annotated in red. It is not possible to tell which auto cross peaks correspond to R_P (or S_P) species. (B) Signal intensities of the auto- and exchange cross peaks. The solid lines represent the best-fit curves obtained using the McConnell equations for a two-site exchange model. The rate constant for the protein’s translocation between R_P and S_P phosphorothioate DNA sites is indicated for each data set.

Discussion

General applicability

Our new strategy for resonance assignment of interfacial side chains at the protein-DNA interfaces can be summarized as follow: 1) the incorporation of phosphorothioate at the positions of the phosphates that interact with the side chain of interest; 2) comparison of the NMR spectra recorded for the modified and unmodified complexes; 3) use of exchange spectroscopy to identify signals from the side chains that are influenced by the exchange between R_P- and S_P-phosphorothioate sites, which facilitates spectral analysis; and 4) iterations of steps 1–3 with consecutive modifications along the DNA phosphate backbone, which help resolve ambiguity among some different possibilities. Although this approach was used exclusively for Arg side-chain N_εH groups in our current study, the same approach should be applicable for Lys NH₃⁺ groups and other types of side chains that interact with DNA backbone phosphates. This approach can readily be applied to protein-RNA complexes as well, since phosphorothioate is also commercially available in custom RNA syntheses.

Cost & time effectiveness

The new strategy proposed here presents several advantages over the conventional methods for assignment. Typically, arginine side-chain resonances are assigned using various ¹H-detected^41,43,44 or ¹³C-detected^45–47 double/triple resonance spectra, which require expensive ¹³C,¹⁵N -labeled proteins. Our current approach requires only ¹⁵N-labeled proteins and is cost effective. Incorporation of racemic phosphorothioate is inexpensive (~$10 per phosphate in 1-μmol scale DNA/RNA synthesis) and commercially available at many custom DNA/RNA synthesis companies. The current approach is also time effective as only two 2D experiments (i.e., ¹H-¹⁵N HISQC and ¹⁵N_z exchange experiments in the current case) are acquired for each modification. Lastly, the assignment process is straightforward and does not require a lengthy analysis.

Exchange timescale

Another remarkable feature of the current approach with racemic phosphorothioate is that it allows kinetic analyses of protein translocation between DNA duplexes. In the current case, ¹⁵N_z exchange spectroscopy was effective for the determination of the kinetic rate constant for protein translocation between DNA duplexes. Such a kinetic analysis was possible because the exchange process was in the slow exchange regime. For weaker protein-DNA or protein-RNA association, the exchange between the R_P and S_P complexes may occur in the fast exchange regime. In such a case, the spectra of the modified complexes become simpler due to the absence of the peak splitting. In such a case of fast exchange, other methods such as ¹⁵N CPMG relaxation dispersion experiments may be applicable to determine the kinetic rate constants for the protein translocation between the R_P and S_P species. When the translocation process occurs in the intermediate exchange regime, the signals could be broadened beyond detection. However, because the kinetics of protein-DNA interactions depend strongly on ionic strength,^48–50 it may be possible to shift the timescale of the process to the fast or slow exchange regime by varying the salt concentration.

Limitations

We should also point out some caveats of the current approach. First, this approach allows for the resonance assignment only when structural information is available for the protein-nucleic acid complex of interest. However, the kinetic analysis of the exchange between the R_P and S_P complexes would be possible without such structural information. Secondly, only interfacial side chains near DNA/RNA phosphates can be assigned using the current approach. Although interfacial side chains are often the most important in NMR investigations of macromolecular complexes, this limitation could pose a problem when NMR data on the interfacial side chains should be compared to those on non-interfacial side chains. Finally, while it is easy to individuate signals from complexes with different diastereomers of phosphorothioate, it is virtually impossible to obtain stereospecific assignment regarding the R_P and S_P configurations for the individual NMR signals.

Concluding Remarks

In the current work, we have presented a fast, robust and inexpensive approach for NMR investigations of interfacial side chains of protein-DNA complexes. The site-specific incorporation of racemic phosphorothioate in nucleic acids provides straightforward means for the resonance assignment and kinetic measurements on protein translocation from one DNA molecule to another. This approach can facilitate NMR studies of protein-nucleic acid complexes, especially when the conventional NMR methods are unable to provide sufficient data on the interfacial side chains or when a quick characterization of only the interface residues is needed. Due to its simplicity and practical ease, we foresee this approach to have extensive applications for various protein-DNA and protein-RNA complexes.