Studying how proteins bind to DNA is not a single technique but a sprawling toolkit, with each method answering a slightly different version of the question. Some tell you whether a protein touches a specific stretch of DNA. Others map every binding site across an entire genome. Still others measure how tightly the two partners grip each other, or reveal the three-dimensional shape of the resulting complex. Choosing the right approach depends on what you actually want to know, and researchers routinely combine several methods because no single one gives the full picture.
Gel Shift Assays and DNase Footprinting
The electrophoretic mobility shift assay, almost universally called EMSA or simply a “gel shift,” has been a workhorse since the 1980s. The idea is straightforward: you mix a protein with a short, labeled piece of DNA, then run the mixture through a gel. A protein-DNA complex is bulkier than naked DNA, so it migrates more slowly through the gel matrix. You end up with two bands on the resulting image, one for free DNA and one for the slower-moving complex, and the relative intensity of each band tells you how much binding occurred.
1PubMed Central. Electrophoretic mobility shift assay (EMSA) for detecting protein-nucleic acid interactionsEMSA’s appeal is its simplicity. You can run it with standard lab equipment, and it works for both DNA- and RNA-binding proteins. Competition experiments, where you add excess unlabeled DNA to see whether it steals the protein away, let you test how specific the interaction is. The assay does have quirks, though. Buffers commonly used in EMSA contain EDTA, a chelating agent that strips zinc ions from zinc-finger proteins and can cause them to unfold. Somehow the assay still detects zinc-finger binding under these conditions, likely because the complex forms before EDTA has time to disassemble it, but researchers studying zinc-finger proteins need to be aware of the chemistry at play.
2PubMed Central. Electrophoretic mobility shift assay of zinc finger proteins: competition for Zn(2+) bound to Sp1 in protocols including EDTAWhere EMSA tells you that a protein binds, DNase I footprinting tells you exactly where on the DNA it sits. The logic is elegant: if a protein is clamped onto a stretch of DNA, the enzyme DNase I cannot chew through that stretch. You treat the protein-DNA mixture with DNase I, then separate the resulting fragments on a gel. The region shielded by the protein shows up as a gap, or “footprint,” in the ladder of cut fragments.
3PubMed. DNase I footprint analysis of protein-DNA bindingBoth techniques are performed on purified components in a test tube, which means they can confirm a direct physical interaction but cannot tell you much about what happens inside a living cell. The protein you purify may behave differently once it is surrounded by thousands of competitors for the same DNA.
Mapping Binding Sites Across Whole Genomes With ChIP-seq
Chromatin immunoprecipitation followed by sequencing, known as ChIP-seq, transformed the field by taking the question from “does this protein bind this one piece of DNA?” to “where does this protein bind across the entire genome?” The method works inside cells. You first lock proteins in place on DNA using a chemical crosslinker such as formaldehyde. Then you shatter the DNA into short fragments, use an antibody to fish out the fragments attached to your protein of interest, and sequence whatever DNA comes along for the ride. Mapping those sequences back to the genome reveals the binding sites.
4Nature Protocols. Multiplexed chromatin immunoprecipitation sequencing for quantitative study of histone modifications and chromatin factorsChIP-seq has been used in organisms from bacteria to humans. In one study of the bacterium E. coli, researchers immunoprecipitated a tagged version of the sigma factor σS and identified 78 genomic regions significantly enriched in the pulled-down DNA, each a candidate binding site for that transcription factor.
5Scientific Reports. Characterization of the Escherichia coli σS core regulon by Chromatin Immunoprecipitation-sequencing (ChIP-seq) analysisThe technique does have a well-known blind spot when it comes to distinguishing direct from indirect binding. A protein that touches DNA through an intermediary partner can look identical to one that contacts DNA on its own. In-vitro methods like EMSA do not have this problem, but they also cannot capture the complex, crowded reality inside a nucleus. Researchers who want the best of both worlds typically run ChIP-seq first to find candidate sites, then verify them with in-vitro or biophysical assays.
6Briefings in Functional Genomics. Experimental strategies for studying transcription factor–DNA binding specificitiesThe Artifact Problem in ChIP-seq
For all its power, ChIP-seq carries systematic biases that can generate convincing but wrong results. Two independent studies in yeast found that unrelated proteins, and even proteins that should have no reason to bind DNA at all, appeared to be enriched at highly transcribed genes. In one striking control experiment, a nuclear-localized green fluorescent protein, which is not a DNA-binding protein, showed enrichment at actively transcribed loci. When the researchers induced expression of the GAL genes, the GFP signal at those loci increased accordingly, with no biological explanation.
7PubMed Central. Highly expressed loci are vulnerable to misleading ChIP localization of multiple unrelated proteinsThe underlying cause appears to be that highly transcribed regions of the genome are more accessible and produce more DNA fragments during the ChIP procedure, inflating their representation in the sequenced library. Library-construction steps can make the bias worse. The practical consequence is that a known transcriptional repressor can erroneously look like an activator if the analysis does not account for these artifacts. Proper input controls and careful statistical treatment are essential, but even then, researchers should be cautious about interpreting ChIP-seq peaks at very highly expressed genes.
8PubMed Central. Widespread misinterpretable ChIP-seq bias in yeastCUT&RUN and CUT&Tag as Lower-Background Alternatives
Some of ChIP-seq’s weaknesses stem from its brute-force approach: crosslink everything, shatter the entire genome, then try to pull out the signal from a sea of noise. CUT&RUN takes a fundamentally different strategy. Instead of solubilizing all the chromatin and then immunoprecipitating, it works on intact nuclei. An antibody finds the target protein in place, and a tethered enzyme (micrococcal nuclease) is guided to cut the DNA only around the bound protein. The released fragments diffuse into the surrounding liquid and are collected for sequencing. Because only the target-adjacent DNA is released, background is dramatically lower.
9PubMed Central. An efficient targeted nuclease strategy for high-resolution mapping of DNA binding sitesCUT&Tag builds on this in-situ principle but swaps the nuclease for a transposase (Tn5) that simultaneously cuts and tags DNA fragments with sequencing adapters in one step. This makes library preparation faster and has enabled single-cell versions of the assay. One protocol using combinatorial barcoding, called sciCUT&Tag, can profile chromatin marks across thousands of individual cells in a single run.
10Nature Protocols. Scalable single-cell profiling of chromatin modifications with sciCUT&TagThe platform continues to evolve. A variant called Nano-CUT&Tag uses nanobody-transposase fusions to simultaneously profile multiple epigenetic marks from the same single cell.
11PubMed. Nano-CUT&Tag for multimodal chromatin profiling at single-cell resolution Another recent iteration, Af-CUT&Tag, eliminates the need for traditional antibodies entirely by using genetically encoded peptide tags recognized by engineered binders fused to the transposase, improving the signal-to-noise ratio over conventional antibody-based approaches.
12PubMed Central. Af-CUT&Tag: a sensitive and antibody-free chromatin profiling method using genetically encoded tags and high-affinity binders fused to Tn5Measuring How Tightly Proteins Grip DNA
Knowing where a protein binds is only part of the story. Researchers also need to know how strong the interaction is, how fast the protein latches on, and how quickly it falls off. Biophysical techniques fill this gap.
Surface plasmon resonance (SPR) measures binding in real time without needing to label either partner. One molecule, usually the DNA, is attached to a sensor chip. When the protein in solution binds, it changes the way light reflects off the chip surface, and the instrument tracks that change continuously. This gives you association and dissociation rates as the interaction forms and breaks apart.
13PubMed Central. Real-Time Analysis of Specific Protein-DNA Interactions with Surface Plasmon Resonance SPR has been used to characterize everything from replication protein A binding single-stranded DNA to zinc-finger proteins engaging their target sequences.
14Journal of Immunological Methods. Surface plasmon resonance based kinetic studies of zinc finger-DNA interactionsIsothermal titration calorimetry, or ITC, takes a different approach entirely: it measures the heat released or absorbed when a protein binds DNA. You slowly inject one binding partner into a solution containing the other and record the tiny temperature changes. From these data you can extract the binding strength, the number of binding sites, and the thermodynamic forces driving the interaction, all in a single experiment without any labels or surface attachment. Researchers consider ITC particularly valuable because its thermodynamic readout can reveal whether an interaction is driven mainly by electrostatic attraction, by the shape of the DNA groove the protein recognizes, or by the protein folding around the DNA upon binding.
15PubMed. Analysis of Protein-DNA Interactions Using Isothermal Titration Calorimetry: Successes and FailuresA newer option, microscale thermophoresis (MST), works on much smaller sample volumes than ITC and tracks how molecules move through a temperature gradient, a property that changes when a binding partner is present. MST and ITC are often treated as complementary methods: ITC delivers a richer thermodynamic profile, while MST is more forgiving when material is scarce.
16PubMed. Measurements of Protein-DNA Complexes Interactions by Isothermal Titration Calorimetry (ITC) and Microscale Thermophoresis (MST)Seeing the Complex in Three Dimensions
For the most detailed view of how a protein contacts DNA, structural biology methods provide atomic-resolution snapshots. X-ray crystallography has been the traditional route. A curated database of protein-DNA complex structures solved by crystallography contained 922 entries at resolutions of 2.5 Å or better, each documenting a distinct protein-DNA interface.
17BMC Bioinformatics. The Protein-DNA Interface databaseCryo-electron microscopy has increasingly taken over for large, dynamic complexes that resist crystallization. Researchers have used cryo-EM to capture transcription initiation complexes in the act of opening DNA, resolving structures at resolutions down to roughly 3 Å, fine enough to trace individual amino acid side chains contacting individual DNA bases.
18PubMed. Transcription initiation complex structures elucidate DNA opening In one case, cryo-EM captured a yeast mitochondrial RNA polymerase at two different stages of transcription initiation, including a previously unknown partially melted intermediate.
19Molecular Cell. Structural Basis of Transcription Initiation and DNA Scrunching in Yeast Mitochondrial RNA PolymeraseFlipping the Question With Locus-Specific Proteomics
ChIP-seq and its descendants start with a known protein and ask “where does it bind?” But sometimes you want to start with a known piece of DNA and ask “which proteins are sitting on it?” This reverse question requires a different family of methods.
Targeted chromatin purification (TChP) uses a genetically encoded “hook” to isolate a specific genomic locus from living cells, then identifies the associated proteins by mass spectrometry. The technique was used to identify transcription factors bound to regulatory regions of the silenced γ-globin gene in mammalian cells.
20PubMed. Locus-specific proteomics by TChP: targeted chromatin purificationHyCCAPP takes a different approach, using synthetic single-stranded DNA probes to fish out specific genomic sequences from crosslinked chromatin. Mass spectrometry of the captured material revealed hundreds of proteins at each targeted locus, with statistical filtering narrowing the list to dozens of genuinely enriched candidates per gene.
21PubMed Central. Multiplexed Sequence-Specific Capture of Chromatin and Mass Spectrometric Discovery of Associated ProteinsA related strategy called ChroP combines chromatin immunoprecipitation with mass spectrometry to catalog not just the proteins at a locus but also the histone modifications and histone variants present, giving a more complete portrait of the local chromatin environment.
22PubMed Central. The ChroP approach combines ChIP and mass spectrometry to dissect locus-specific proteomic landscapes of chromatinGenetic Screening With Yeast One-Hybrid Assays
If your goal is to cast a wide net and find any transcription factor that can bind a regulatory DNA sequence, the yeast one-hybrid (Y1H) system is a classic genetic screen. You place your DNA sequence of interest upstream of a reporter gene in yeast, then introduce a library of candidate transcription factors. Any factor that binds the DNA sequence activates the reporter, and you identify it by sequencing.
Y1H screens have been used productively in plant biology to identify upstream regulators of pigment genes in sweet potato and lily, among many other systems.
23PubMed Central. Yeast One-Hybrid Screening to Identify Transcription Factors for IbMYB1-4 in the Purple-Fleshed Sweet Potato (Ipomoea batatas [L.] Lam.) A more recent extension, the paired yeast one-hybrid (pY1H) assay, can detect not just individual binding but also cooperativity and antagonism between pairs of transcription factors at the same DNA region, something the standard version misses.
24Nature Communications. Paired yeast one-hybrid assays to detect DNA-binding cooperativity and antagonism across transcription factorsHigh-Throughput Motif Discovery
Sometimes you want to characterize which short DNA sequence a protein prefers without committing to a specific genomic context. Two in-vitro methods have been widely adopted for this. Protein-binding microarrays (PBMs) expose a transcription factor to a chip containing every possible short DNA sequence and measure binding to all of them simultaneously. HT-SELEX (high-throughput systematic evolution of ligands by exponential enrichment) uses rounds of selection to enrich the best-binding sequences from a random pool, then sequences the winners.
A comparative analysis found that models built from PBM and HT-SELEX data agree for most transcription factors. Where they differ, HT-SELEX tends to capture longer binding motifs, sometimes identifying extended versions of the shorter motifs seen on PBMs. PBM-derived rankings of individual short sequences were more accurate in pure ranking tests, but HT-SELEX-derived models were better at predicting actual binding in living cells.
25PubMed Central. A comparative analysis of transcription factor binding models learned from PBM, HT-SELEX and ChIP dataWatching Proteins Move on DNA in Real Time
All of the methods described so far provide either a snapshot or an average. Single-molecule techniques let you watch individual protein molecules interacting with individual DNA molecules as it happens. Fluorescence microscopy, optical tweezers, and magnetic tweezers have all been applied to track how proteins slide, hop, or stall on DNA strands.
26PubMed Central. Single-molecule views of protein movement on single-stranded DNAMoving from the test tube into living cells, reflected light sheet microscopy has achieved single-molecule sensitivity inside mammalian nuclei. Researchers used this approach to measure how long individual glucocorticoid receptor and estrogen receptor molecules stayed bound to DNA, resolving different binding modes and oligomeric states that would be invisible in any bulk measurement.
27Nature Methods. Single-molecule imaging of transcription factor binding to DNA in live mammalian cellsCrosslinking Strategies That Distinguish Direct From Indirect Binding
A persistent challenge across many of these methods is distinguishing a protein that directly contacts DNA from one that is merely nearby, hitching a ride on another protein. Crosslinking chemistry is the usual way to freeze interactions in place, but different crosslinkers capture different things.
Formaldehyde, the standard crosslinker for ChIP, is promiscuous: it links proteins to DNA but also proteins to proteins, which means indirect interactions get captured alongside direct ones. UV light crosslinking is more selective, creating covalent bonds only between amino acids in direct contact with nucleic acid bases. A recent study found that chemical crosslinkers preferentially link amino acids to purine bases in DNA, while UV crosslinks mainly form at pyrimidine bases, meaning the two approaches give complementary information about the same interface.
28Nucleic Acids Research. Chemical crosslinking extends and complements UV crosslinking in analysis of RNA/DNA nucleic acid–protein interaction sites by mass spectrometryUV laser crosslinking pushes this selectivity further. By delivering a UV pulse of defined duration, researchers can control the crosslinking window and compare the results with formaldehyde crosslinking of the same sample. A combination of the two has been shown to distinguish direct from indirect DNA-protein interactions in small numbers of living cells, which is something formaldehyde alone cannot do reliably.
29PubMed Central. Time-resolved analysis of DNA-protein interactions in living cells by UV laser pulsesComputational Prediction of Binding Affinity
With experimental data accumulating rapidly, machine-learning models are increasingly being trained to predict protein-DNA interactions from sequence alone. One such tool, DeePNAP, was built on a dataset of over 14,000 entries covering both wild-type and mutant protein-nucleic acid complexes from a wide range of organisms. The model predicts binding affinity and the effects of mutations using only the protein and DNA sequences as input, without requiring a solved three-dimensional structure.
30PubMed. DeePNAP: A Deep Learning Method to Predict Protein-Nucleic Acid Binding Affinity from Their SequencesThese models are not replacing experiments, but they are becoming useful for prioritizing which interactions to test. If you have thousands of candidate mutations in a transcription factor and want to know which ones are likely to disrupt binding, running them through a sequence-based predictor first can save significant time and reagent costs. The field is still young enough that predictions should be confirmed experimentally for any high-stakes application, but the trajectory is clear: computational tools are becoming a standard part of the workflow rather than a novelty.