What Are Proximity Ligation Assays and How Do They Work?

A proximity ligation assay, or PLA, is a technique that detects proteins and their interactions by converting what happens at the molecular level into a DNA signal that can be copied millions of times over and then visualized or measured. The core idea is elegant: two antibody probes, each carrying a short strand of DNA, bind to targets sitting close together on or inside a cell. If and only if those probes land near enough to each other, their DNA tags get linked and amplified, producing a detectable signal. This makes PLA extraordinarily sensitive and specific, capable of spotting individual protein complexes in a single cell, and it has become a go-to method for studying everything from cancer signaling to how receptors on the cell surface pair up.

How the Assay Works, Step by Step

The basic mechanism involves a few key stages, each building on the last. First, researchers prepare what are called proximity probes. These are antibodies (or sometimes other binding molecules) that have been chemically attached to short pieces of single-stranded DNA. For a typical two-target experiment, one probe carries a DNA oligonucleotide called the “plus” strand and the other carries the “minus” strand. The probes are designed so that each one recognizes a different part of the target. If you want to detect a protein-protein interaction, one probe binds protein A and the other binds protein B.¹

Once both probes have bound their respective targets, the two DNA arms dangle close together in space. Connector oligonucleotides are then added to the sample. These short DNA strands are designed to hybridize (base-pair) with the ends of both probe arms at once, effectively bridging them. A DNA ligase enzyme seals the nicks, joining the connectors and probe-arm ends into a continuous circular DNA molecule. This ligation step is the critical gate: it only succeeds if the two probes are physically close enough for the connector oligos to reach both DNA arms simultaneously. If the target proteins are not interacting, or if only one probe binds, no circle forms and no signal is generated.

That distance threshold is tight. The probes typically need to be within about 40 nanometers of each other for ligation to occur.² For context, 40 nanometers is far smaller than the wavelength of visible light. This is what gives PLA its specificity: two proteins merely present in the same general area of the cell will not trigger a signal. They have to be interacting or sitting practically on top of each other.

Turning a Tiny Signal Into Something You Can See

Once the circular DNA template is formed, it gets amplified through a process called rolling circle amplification, or RCA. A DNA polymerase enzyme latches onto the circle and starts copying it, going round and round the template to produce a long, continuous single-stranded DNA molecule made up of hundreds of repeated copies of the circle’s sequence. This growing DNA strand stays tethered to the probe site, coiling up into a compact ball right where the original protein interaction took place.

To make this ball visible, fluorescently labeled detection oligonucleotides are added. These short probes hybridize to repeated sequences in the amplified DNA, stacking enough fluorescent molecules in one spot to create a bright dot that you can see under a fluorescence microscope. Each dot represents a single molecular event: one pair of proteins interacting at one specific location in the cell.³ Some researchers have also coupled PLA with other RCA-based strategies to image specific protein modifications on the cell surface, such as glycosylation of signaling receptors, by pairing aptamer-based probes with glycan-labeling chemistry.⁴

The result is a fluorescent image where each bright punctum corresponds to a single detected interaction. Researchers can then count those dots, map their locations within the nucleus or cytoplasm, and compare across experimental conditions. This single-molecule resolution is something most conventional protein-detection methods cannot match.

What PLA Can Actually Detect

The most common use of PLA is detecting protein-protein interactions directly inside fixed cells or tissue sections. If two proteins form a complex, proximity probes against each protein will bind close enough to generate a signal. This has been applied widely in cancer research, where signaling pathways depend on specific protein complexes forming at the right time and place. One group, for instance, used PLA to detect complexes involving the p53, p63, and p73 tumor suppressor proteins directly in tumor tissue, achieving what they described as single-molecule resolution for localizing and quantifying those complexes.⁵

PLA is not limited to interactions between two different proteins. It is also widely used to detect post-translational modifications, the chemical tags that cells add to proteins to change their behavior. Phosphorylation is the classic example. By using one probe that recognizes a protein and a second probe that recognizes a specific phosphorylated residue on that same protein, researchers can detect and localize activated forms of signaling receptors. This was demonstrated for the platelet-derived growth factor receptor, where PLA revealed individual phosphorylated receptors after cells were exposed to growth factor stimulation.⁶ A hybrid format called ELISA-PLA has extended this concept to measure phosphorylation dynamics quantitatively, tracking how proteins like ERK1/2 get phosphorylated over time.⁷

Beyond interactions and modifications, researchers have combined PLA with other molecular detection tools. One protocol pairs PLA with padlock probes, allowing simultaneous detection of individual mRNA molecules and protein complexes or modifications in the same cell. This combined readout gives information about gene expression and protein activity at the same time, within the same tissue section.⁸

In Situ Versus Solution-Phase Formats

PLA exists in two major formats that serve different purposes. In situ PLA, sometimes sold under the brand name Duolink, is the microscopy-based version described above. Cells or tissue sections are fixed on slides, probed, ligated, amplified, and imaged. The great advantage is spatial information: you see exactly where in the cell or tissue each interaction occurs. The trade-off is throughput. Processing slides, imaging fields of view, and counting dots is relatively slow.

Solution-phase PLA works differently. Proximity probes are added to a liquid sample such as blood serum or cell lysate. If the probes bind their targets in close proximity, ligation happens in solution, and the resulting DNA product is detected and quantified by real-time PCR rather than microscopy. This format lends itself to high-throughput biomarker screening. A multiplexed version allows simultaneous detection of several proteins in one sample, each identified by a unique DNA sequence and quantified by PCR.⁹ One adaptation eliminated all washing steps entirely, creating what researchers called a homogeneous assay suitable for screening large numbers of serum samples for protein biomarkers.¹⁰

Which format you choose depends on your question. If you need to know where inside a tumor cell two signaling proteins are meeting, in situ PLA is the way to go. If you need to screen thousands of patient serum samples for a panel of disease-related proteins, solution-phase PLA is far more practical.

Working with Clinical Tissue Samples

One of the practical strengths of PLA is that it works on formalin-fixed, paraffin-embedded (FFPE) tissue, the standard format in which hospital pathology labs store biopsies. These samples can be years old and the proteins in them are cross-linked by fixation, which makes many standard assays difficult. PLA handles this surprisingly well, allowing researchers to go back to archived tumor collections and study protein interactions that were never measured when the tissue was first collected.¹¹

This compatibility opens the door to cohort studies, where researchers examine FFPE tissue from large groups of patients to determine whether a particular protein interaction correlates with disease outcomes. The technique also works on cultured cells and frozen tissue sections, giving researchers flexibility depending on what material is available.¹² A brightfield version of PLA, which uses enzyme-based chromogenic detection instead of fluorescence, has also been developed for tissue sections. Brightfield PLA produces a stain visible under an ordinary light microscope, making it more accessible to pathology labs that may not have fluorescence imaging equipment. This approach was used to detect signaling complexes of the TGF-β and BMP pathways directly in mouse tissue sections.¹³

Combining PLA with Flow Cytometry

A microscope image is great for spatial detail but less ideal when you need statistics on thousands of individual cells. Researchers solved this by combining in situ PLA with flow cytometry, which passes cells one by one through a laser beam and records their fluorescence. The PLA reaction happens inside the cells first, amplifying each interaction into a fluorescent dot, and then the cells are run through the flow cytometer. This setup lets you measure protein interactions and modifications on a single-cell basis with the kind of statistical power that comes from analyzing large cell populations.

The approach has been validated for receptor dimerization studies. One group used it to detect interleukin-7 receptor heterodimers on the cell surface, demonstrating that PLA combined with flow cytometry was sensitive enough to pick up cytokine receptor pairing events that would be invisible to conventional methods.¹⁴ Another study applied the same combination to the epidermal growth factor receptor (EGFR) family, quantifying both the dimerization and the activation state of EGFR and HER2 in individual cells.¹⁵ This matters for cancer biology, where EGFR and HER2 are drug targets and their activation status can vary dramatically from one cell to the next within the same tumor.

Counting the Dots and Analyzing Images

A PLA experiment on a microscope slide produces fluorescent images full of bright dots. Turning those images into reliable numbers requires image analysis software, and this step can be a bottleneck. Counting dots by eye is tedious and subjective, so automated tools have been developed specifically for PLA data.

BlobFinder was one of the early free tools designed for this purpose. It performs batch processing of fluorescence microscopy images, automatically identifying and counting point-like signals from PLA and similar assays while also localizing cells in the image.¹⁶ More recently, a set of open-access image analysis pipelines built for the FIJI platform (a widely used distribution of ImageJ) included a dedicated PLA algorithm. This pipeline walks the user through smoothing, thresholding, segmenting, and counting PLA foci in both the nucleus and cytoplasm, aiming for reproducible high-throughput quantification even for researchers with no prior image analysis experience.¹⁷

Getting the analysis right matters because the raw number of PLA dots per cell is the primary quantitative output of in situ experiments. Inconsistent thresholding or failure to exclude background noise can inflate or deflate the count, making rigorous software and standardized settings essential for any study that compares conditions or patient groups.

Limitations and Common Pitfalls

PLA is powerful, but it has real weaknesses that researchers need to plan around. The most frequently discussed pitfall is antibody quality. Because the assay depends on two probes binding their targets with high specificity, any cross-reactivity can generate false positive signals. If one antibody binds nonspecifically to a protein it was not designed for, the probe may end up close to the second probe by accident, producing a spurious PLA dot that looks exactly like a real interaction. The standard advice is to use highly specific, validated antibodies and to include rigorous controls: single-probe-only conditions, knockout cell lines where one target protein is absent, and blocking experiments.¹⁸

Another limitation is that PLA typically works on fixed (dead) cells or tissue. It captures a snapshot of protein interactions at the moment of fixation but cannot follow dynamic changes in living cells in real time. Researchers who want to track how an interaction changes over minutes or hours need to fix separate samples at each time point, which introduces variability.

The 40-nanometer proximity requirement, while excellent for specificity, also means PLA cannot distinguish a true direct binding interaction from a situation where two proteins are merely part of the same large complex but do not touch each other directly. If protein A binds scaffold C and protein B also binds scaffold C, PLA may report an A-B “interaction” that is actually an indirect co-localization. This is worth keeping in mind when interpreting results, and orthogonal methods like co-immunoprecipitation or structural studies are often used alongside PLA to confirm direct binding.

Multiplexing in situ PLA is also harder than multiplexing solution-phase PLA. Running several probe pairs simultaneously on a single tissue section increases the risk of cross-ligation between unrelated probe arms, and the number of spectrally distinct fluorophores available for imaging limits how many interactions you can visualize at once.

Proximity Extension Assay and Other Variations

PLA inspired a related technique called the proximity extension assay, or PEA, which replaces the DNA ligase step with a DNA polymerase extension step. In PEA, when two probes bind targets in close proximity, the DNA strands on the probes hybridize and a polymerase extends one strand using the other as a template, creating a new double-stranded DNA product. This matters for practical reasons: the DNA ligase enzyme used in classical PLA can be inhibited by components of blood serum and plasma, which sometimes reduces signal in clinical samples. DNA polymerases are more robust to these inhibitors, making PEA better suited for running directly on blood-derived samples. PEA has been multiplexed to measure up to 384 proteins simultaneously.¹⁹

Other creative variations have pushed the proximity concept further. One group developed an aptamer-based version that bypasses antibodies entirely. In this approach, a split DNA aptamer, two DNA fragments that only assemble in the presence of a specific small molecule, was used to detect cocaine. One aptamer fragment was immobilized on a plate, the sample was added with the second fragment, and if the target molecule was present, the aptamer halves came together and were joined by a chemical ligation reaction. This “aptamer proximity ligation” approach demonstrates that the core principle of proximity-dependent signal generation extends well beyond antibodies and proteins, opening the door to detecting small molecules, drugs, and metabolites.²⁰

Why PLA Has Become So Widely Adopted

Conventional methods for studying protein interactions each have trade-offs that PLA sidesteps in useful ways. Co-immunoprecipitation, the traditional workhorse, requires cells to be lysed, destroying all spatial information and averaging across millions of cells. Yeast two-hybrid assays take proteins out of their normal cellular context entirely. Standard immunofluorescence can show where two proteins are located but cannot confirm they are physically interacting, only that they are in the same general area of the cell. PLA occupies a sweet spot: it works in intact fixed cells and tissues, preserves spatial context, and adds a molecular-proximity gate that distinguishes genuine interactions from mere co-localization.

The amplification step is the other major advantage. Because each interaction event generates hundreds of copies of DNA decorated with fluorescent probes, PLA can detect proteins at their natural endogenous levels without the need to overexpress tagged versions.² Overexpression artifacts are a persistent headache in cell biology: forcing cells to make abnormally high levels of a protein can create interactions that would never happen under normal conditions. PLA avoids this entirely by detecting whatever is already there at physiological concentrations.

The technique has also proven adaptable to tissue types that many other assays struggle with. Its compatibility with archival paraffin-embedded clinical specimens means that large retrospective studies become feasible without collecting new tissue, which is especially valuable in rare diseases where fresh patient samples are hard to come by.¹² For researchers working at the intersection of basic biology and clinical translation, this combination of sensitivity, spatial resolution, and practical flexibility explains why PLA has moved from a specialist tool to a standard item in the molecular biology toolkit.