How DSP Crosslinker Isolates Weak Protein Complexes

DSP, or dithiobis(succinimidyl propionate), is a chemical crosslinker widely used in molecular biology to lock proteins and other biomolecules into place so researchers can study their interactions and structures. Sometimes called Lomant’s Reagent, DSP is popular because it does something most fixatives cannot: it forms strong covalent bonds between nearby molecules and then lets you break those bonds later with a simple reducing agent. That reversibility makes DSP one of the most versatile tools in the crosslinker toolkit, turning up in applications from immunoprecipitation and mass spectrometry to single-cell RNA sequencing and chromatin analysis.

What DSP Actually Does to Molecules

DSP belongs to a class of reagents called homobifunctional crosslinkers, meaning both ends of the molecule carry the same reactive group. In DSP’s case, those groups are NHS esters (N-hydroxysuccinimide esters), which react with primary amines, the kind found on lysine side chains and at the free N-terminus of proteins. When a DSP molecule encounters two proteins sitting close together, one NHS ester can grab a lysine on one protein while the other grabs a lysine on the neighboring protein, forming stable amide bonds and effectively stapling the two proteins together.¹ The bridge between those two attachment points, the spacer arm, spans about 12 ångströms, which is roughly the width of a single alpha-helical turn in a protein.² That short reach means DSP only links molecules that are genuinely close neighbors, a feature that helps researchers distinguish real binding partners from bystanders floating nearby in the cell.

Why Reversibility Sets DSP Apart

Buried in the middle of DSP’s spacer arm is a disulfide bond, a sulfur-sulfur linkage that snaps apart when exposed to a reducing agent such as DTT (dithiothreitol) or beta-mercaptoethanol. This is the feature that gives DSP most of its practical value. You can crosslink proteins together in a living cell, lyse the cell, pull down your target protein with antibodies, and then cleave the crosslinks before running the sample on a gel or a mass spectrometer. The result is clean, unmodified protein that behaves normally in downstream analysis, rather than a tangled web of covalently fused complexes.³

Formaldehyde, the most commonly used fixative in biology, also crosslinks proteins through amine groups, but its crosslinks are much harder to reverse cleanly. Heating and detergent treatment can partially undo formaldehyde fixation, but the process is harsh and incomplete. DSP, by contrast, is cleaved under mild reducing conditions, which preserves both protein integrity and, in many cases, RNA quality. One direct comparison found that cleaving DSP crosslinks improved protein recovery from fixed tissue roughly 18-fold and increased the number of identifiable proteins by about 20 percent compared with standard formaldehyde-fixed, paraffin-embedded samples under the same mild extraction conditions.⁴

Stabilizing Weak and Transient Protein Complexes

Some protein-protein interactions are robust enough to survive cell lysis and immunoprecipitation buffers on their own. Others are not. Signaling complexes, for example, often assemble briefly and fall apart the moment cell conditions change. If you lyse a cell without crosslinking first, those fleeting interactions dissociate before you can capture them. DSP solves this by freezing the interaction in place while the proteins are still in their native environment inside the cell. After lysis, the crosslinked complex survives the purification steps, giving researchers a snapshot of what was bound to what at the moment of fixation.⁵

This approach has been particularly useful for studying receptor-protein interactions at the cell surface. Membrane receptors often associate with scaffolding proteins and signaling partners in ways that depend on whether the receptor is active or inactive at that moment. Adding a crosslinker before cell lysis captures those state-dependent complexes, letting researchers compare the interactome of a receptor in its resting state versus after stimulation by a drug or hormone.⁶ Without crosslinking, many of those partners would be lost during the wash steps of a standard affinity purification.

Preparing and Using DSP in the Lab

DSP is not water-soluble on its own. The standard preparation is to dissolve it in DMSO at a concentrated stock, typically 20 mM, and then dilute that stock into an aqueous buffer just before use.⁷ Because the NHS ester groups begin to hydrolyze (lose their reactivity) the moment they encounter water, fresh preparation matters. A stock solution made an hour ago may have noticeably less crosslinking activity than one made minutes before the experiment.

A typical crosslinking protocol involves washing cells with a simple phosphate-buffered saline to remove serum proteins (which would soak up the crosslinker nonproductively), incubating with DSP for about 30 minutes at room temperature, and then quenching the reaction. Quenching uses a primary amine like Tris buffer or L-cysteine, which reacts with any remaining NHS esters and stops further crosslinking.⁸ Skipping the quench step is a common beginner mistake that leads to continued crosslinking during lysis, increasing background noise.

Getting the Concentration Right

Choosing the right DSP concentration is one of the trickiest parts of a crosslinking experiment, and there is no single answer that works for every cell type and application. Too little crosslinker, and the interactions you care about are not stabilized. Too much, and you create massive crosslinked aggregates that are impossible to immunoprecipitate or resolve on a gel.

The ReCLIP method, which was developed specifically for capturing labile protein complexes, tested a range of DSP concentrations and settled on 0.5 mM as a good general-purpose working concentration. At that level, crosslinking was efficient enough to stabilize known interactions without excessive nonspecific capture. A higher concentration of 1.0 mM was somewhat more effective at stabilizing target complexes, but the researchers chose the lower dose to limit false positives.⁹ Most published protocols fall in the 0.25 to 2.0 mM range, with the specific optimum depending on cell density, the abundance of the target protein, and how aggressively you plan to wash during purification. If you are setting up DSP crosslinking for the first time, titrating across three or four concentrations in a pilot experiment is the standard advice.

Reading Crosslinked Samples on Gels

One straightforward way to confirm that DSP crosslinking worked is to run samples on an SDS-PAGE gel under non-reducing conditions, meaning without adding DTT or other reducing agents to the sample buffer. Under those conditions, crosslinked complexes hold together and migrate at a higher apparent molecular weight than the individual monomers. A study of the MPYS protein (also known as STING, a protein involved in innate immune signaling) demonstrated this clearly: DSP-treated MPYS appeared as a single band at roughly 80 kDa on a non-reducing gel, consistent with the expected homodimer, while untreated or DTT-treated samples ran at the monomer size of about 40 kDa.¹⁰

Running a parallel lane under reducing conditions then collapses the crosslinks, and the proteins should return to their expected monomeric sizes. Comparing the two lanes gives you immediate visual confirmation that your crosslinking is working and that the higher-molecular-weight band is indeed a crosslinked complex, not an artifact of incomplete denaturation. This reducing-versus-non-reducing gel comparison is a standard quality-control step in any DSP-based protocol.

Single-Cell RNA Sequencing and Tissue Preservation

One of DSP’s more recent and perhaps surprising applications is in single-cell transcriptomics. The challenge in single-cell RNA sequencing is that dissociating tissue into individual cells and processing them is time-consuming, and cells can change their gene expression profile during that process. Mechanical and enzymatic dissociation activates stress-response genes, meaning the transcriptome you sequence may not reflect the cell’s true in-tissue state. Fixing cells with DSP before dissociation locks RNA in place and prevents these stress artifacts from distorting the data.

Researchers adapted DSP, or Lomant’s Reagent, as a fixative for single-cell applications and found that it preserves RNA integrity and yield in a manner compatible with downstream sequencing. DSP-fixed single cells did show some characteristic differences from fresh cells, including slightly reduced cDNA yield and a detectable bias toward the 3′ end of transcripts, but the overall RNA complexity at the gene level was not substantially reduced.¹¹ A more recent methodology called FixNCut built on this principle, using reversible fixation followed by tissue dissociation and demonstrating that RNA integrity, library complexity, and cellular composition were all preserved across human and mouse tissues, while stress-related artifacts were diminished. The method proved compatible with multiple single-cell and spatial transcriptomic platforms.¹²

This fix-then-dissociate approach also brings practical flexibility to study design. Tissues can be fixed immediately upon collection, stored, and processed later without the pressure of getting fresh samples onto a sequencing platform within hours. For clinical samples especially, where collection timing is unpredictable, that flexibility can be the difference between usable data and a failed experiment.

Probing Three-Dimensional Protein Structure With Mass Spectrometry

Beyond capturing interactions between different proteins, DSP can provide information about the shape of a single protein or complex. The logic is simple: if two lysine residues on the same protein can be bridged by DSP’s 12-ångström spacer arm, those residues must be within roughly that distance of each other in the folded structure. By identifying which residue pairs become crosslinked and feeding those distance constraints into structural modeling software, researchers can generate or refine three-dimensional protein models.

A technique combining DSP crosslinking with electrochemistry and mass spectrometry takes this further. After crosslinking, the disulfide bond in DSP’s spacer arm can be cleaved electrochemically rather than chemically, allowing the process to happen online during mass spectrometry analysis. This approach enables researchers to identify crosslinked peptide pairs and map distance constraints in a single streamlined workflow.¹³ A related method uses isotope-labeled versions of DSP (light DSP-d0 mixed with heavy DSP-d8 in equal amounts) to make crosslinked peptide pairs easier to identify in mass spectra, since they appear as distinctive doublet peaks separated by a known mass difference.¹⁴

These structural mass spectrometry approaches are particularly useful for proteins that resist crystallization or are too large for NMR spectroscopy, the two traditional methods for determining protein structure. While crosslinking mass spectrometry does not produce atomic-resolution structures on its own, the distance restraints it generates can substantially improve the accuracy of computational models, especially for protein complexes where the structures of individual subunits are known but the arrangement of those subunits relative to each other is not.

Chromatin Immunoprecipitation and Epigenetics

Chromatin immunoprecipitation (ChIP) is a workhorse technique for studying which proteins bind to which stretches of DNA inside cells. The standard version uses formaldehyde to crosslink DNA-binding proteins to DNA, but formaldehyde does not work equally well for all targets. Some transcription factors and chromatin-associated proteins resist immunoprecipitation after formaldehyde fixation, possibly because the fixation alters the epitope that the antibody recognizes or because the protein’s association with DNA is too indirect for formaldehyde’s very short crosslinking reach.

DSP has shown potential as a supplementary or alternative fixative for ChIP experiments targeting these difficult proteins. In one study, researchers used DSP to stabilize DNA-associated protein complexes and then performed immunoprecipitation followed by next-generation sequencing. The DSP-stabilized complexes yielded recognizable transcription factor binding motifs, including known RUNX2-binding motifs detected in a CBFB ChIP-seq experiment. While a direct head-to-head comparison with conventional formaldehyde-based ChIP-seq was noted as still needed, the results suggested that DSP-based fixation could supplement formaldehyde for targets that are otherwise difficult to capture.¹⁵

Common Pitfalls and Artifacts to Watch For

DSP is powerful, but it is not foolproof. Several artifacts are well-recognized in the literature, and understanding them helps researchers interpret their results honestly.

Over-crosslinking: Using too much DSP or incubating too long creates large aggregates that do not enter a gel properly, resist immunoprecipitation, and artificially inflate the apparent size of protein complexes. The fact that higher concentrations sometimes capture more target does not mean higher is always better, because nonspecific background rises in parallel.
Hydrolysis of NHS esters: DSP’s reactive groups begin degrading as soon as they contact water. Old stock solutions or slow experimental workflows lead to inconsistent crosslinking between replicates. Always prepare fresh stock immediately before use.
Incomplete quenching: Failing to quench residual DSP with a primary amine like Tris means crosslinking continues during cell lysis, potentially linking proteins that were never in the same complex inside the cell.
DMSO effects: Because DSP must be dissolved in DMSO before dilution into buffer, the final working solution contains a small percentage of DMSO. At high DSP concentrations, the DMSO itself can affect cell viability and membrane permeability, complicating experiments that aim to capture complexes in a near-native cellular state.
3′ bias in RNA applications: As noted in single-cell sequencing work, DSP fixation can introduce a detectable shift toward capturing the 3′ ends of transcripts more efficiently than the 5′ ends. For gene-level quantification this is usually tolerable, but for analyses that depend on full-length transcript coverage, it may matter.

One practical advantage of DSP over formaldehyde, even when things go slightly wrong, is that DSP introduces chemically defined modifications on proteins. When you know exactly what the crosslinker adds to each lysine, you can account for that mass shift during database searches in proteomics experiments. Formaldehyde, by contrast, creates a more heterogeneous set of modifications that are harder to model computationally.¹⁶

DSP Versus Other Crosslinkers

DSP is far from the only crosslinker available, and choosing the right one depends on what you are trying to capture. A few common alternatives help illustrate where DSP fits in the landscape.

DTME (dithio-bismaleimidoethane) is another thiol-cleavable crosslinker, but instead of targeting primary amines like DSP does, it reacts with sulfhydryl groups on cysteine residues. The ReCLIP method deliberately uses DSP and DTME together, reasoning that combining an amine-reactive and a sulfhydryl-reactive crosslinker captures a broader range of interaction surfaces than either alone.¹⁷ DSS (disuccinimidyl suberate) has the same NHS ester chemistry as DSP and a similar spacer arm length, but lacks the central disulfide bond, making it irreversible. DSS is preferred when you want permanent crosslinks, for instance in structural studies where the crosslinked peptides themselves are the analyte. BS3 (bis(sulfosuccinimidyl) suberate) is essentially a water-soluble version of DSS, useful when you need to crosslink proteins on the cell surface without the reagent entering the cell.

The spacer arm length matters for structural studies because it defines the maximum distance between two residues that can be bridged. DSP’s 12-ångström arm sets a fairly tight distance constraint. Longer crosslinkers like EGS (ethylene glycol bis(succinimidyl succinate)), with a spacer arm of about 16 ångströms, reach further but provide less precise distance information. The optimal spacer length for a given structural modeling application depends on the size of the target protein and the specific residues being connected.

For researchers who want DSP’s amine reactivity but need to work in aqueous solution without DMSO, a sulfonated version called DTSSP (3,3′-dithiobis(sulfosuccinimidyl propionate)) is available. DTSSP is water-soluble and membrane-impermeable, making it useful for crosslinking proteins on the extracellular face of the plasma membrane without affecting intracellular complexes. DSP, being membrane-permeable, crosses into the cell and crosslinks proteins throughout the cytoplasm and nucleus, a distinction that matters when you need spatial selectivity.

When Formaldehyde Might Still Be the Better Choice

Despite DSP’s advantages in reversibility and protein recovery, formaldehyde is not going anywhere. Formaldehyde is cheap, water-soluble, and works with essentially every tissue and cell type with decades of optimized protocols. Its crosslinks are extremely short (about 2 ångströms), which means it captures only the most intimate molecular contacts, including protein-DNA interactions that longer crosslinkers might bridge artificially. For conventional histology and standard ChIP-seq of well-characterized transcription factors, formaldehyde remains the default because the established workflows are reliable and the bioinformatics pipelines are designed around its properties.

DSP’s niche is situations where formaldehyde falls short: targets that resist immunoprecipitation after formaldehyde fixation, experiments that require clean protein recovery for mass spectrometry, workflows that benefit from fixing tissue and processing it later, and applications where the ability to remove the crosslinker after purification adds analytical power. The two reagents are increasingly used together in dual-crosslinking protocols, where formaldehyde captures tight DNA-protein contacts and DSP stabilizes the broader protein-protein interactions within the same complex. Thinking of DSP and formaldehyde as competitors misses the point. They are complementary tools that happen to share the word “crosslinker” in their description.