Biolayer Interferometry: Novel Analysis of Binding Events

Biolayer interferometry, commonly called BLI, measures molecular binding events in real time by detecting shifts in light interference patterns at the surface of a fiber-optic biosensor. The technique works without labels or fluorescent tags, giving researchers a direct readout of how fast molecules associate, how tightly they stick together, and how quickly they come apart. What makes it genuinely useful across a range of fields is its “dip-and-read” simplicity combined with the ability to process hundreds or thousands of samples in a day, qualities that have pushed BLI into antibody development, gene therapy manufacturing, and basic biochemistry research alike.

How Light at a Fiber Tip Reveals Molecular Interactions

At its core, BLI relies on a straightforward optical principle. A biosensor tip is coated with a thin layer of molecules, and white light is sent down the fiber. Some of that light reflects off the inner surface of the coating, and some reflects off the outer surface where binding happens. These two reflected beams interfere with each other, creating a characteristic pattern. When target molecules from the surrounding solution bind to the coated surface, the optical thickness of that layer changes, shifting the interference pattern by a measurable amount. That shift, reported in nanometers, becomes the real-time signal.

Because the measurement depends only on what is physically accumulating at the sensor tip, it does not require the target molecules to carry a fluorescent dye, a radioactive label, or any other tag. This is a practical advantage. Labels can alter how molecules fold or interact, which means a label-free approach tends to give a more faithful picture of the binding event as it would occur naturally. The sensor tip simply dips into a microplate well containing the sample, records the binding, then moves to a well containing buffer to watch the molecules dissociate. That dip-and-read workflow is one reason BLI has become so accessible: there is no microfluidic plumbing to maintain, no flow cells to unclog, and no complex sample injection system.

What a BLI Experiment Actually Tells You

A typical BLI experiment produces a sensorgram, a curve that traces the nanometer shift over time. The curve has a recognizable shape. During the association phase, the signal climbs as molecules from solution bind to the sensor. If enough time passes and the surface becomes saturated, the signal levels off at steady state. When the sensor is moved to a buffer-only well, the dissociation phase begins and the signal drops as bound molecules release. By running the experiment at several different concentrations of the analyte, researchers can extract an association rate, a dissociation rate, and from those, an equilibrium dissociation constant that describes overall binding strength. These measurements can be made in minutes using very small amounts of material.¹

In some experiments, the steady-state plateau itself is used directly to determine binding affinity, particularly when one binding partner is small or when kinetics are very fast. By plotting the steady-state response at each concentration, researchers generate a binding curve and fit it to extract the affinity constant.² The choice between kinetic fitting and steady-state analysis depends on the interaction being studied, and using the wrong approach for a given system can produce misleading numbers.

Screening Thousands of Antibodies in a Day

One of BLI’s most impactful roles is in antibody discovery and development. When a pharmaceutical team generates thousands of antibody-producing cell lines, they need a fast way to identify which clones produce antibodies that bind tightly and release slowly from their target. Slow dissociation generally correlates with therapeutic potency, so ranking clones by their off-rate is a practical shortcut. Using a 384-sensor platform, one group screened 2,000 hybridoma clones within 24 hours and selected high-affinity candidates for further development.³ That throughput would be difficult to achieve with techniques requiring individual flow-cell injections.

Beyond simple affinity ranking, BLI is also used for epitope binning, a process that sorts antibodies into groups based on which part of a target protein they recognize. By loading one antibody onto a sensor, letting it bind the antigen, and then testing whether a second antibody can also bind, researchers can map out which antibodies compete with each other. This reveals the landscape of available binding sites on a given target and helps design antibody panels or combination therapies. BLI-based epitope binning has been applied to complex antibody mixtures, including polyclonal serum antibodies and monoclonal antibodies raised against influenza hemagglutinin.⁴

Getting Capture Density Right

A recurring challenge in BLI experiments is controlling how much of the capture molecule is loaded onto the sensor surface. If the surface is too densely packed, incoming analyte molecules can rebind immediately after dissociating rather than diffusing away. This creates an artifact called mass transport limitation that makes the dissociation rate appear slower than it really is, leading to an artificially tight affinity measurement. Careful work on antibody-antigen systems has shown that keeping the capture density low, below about 0.6 nanometers of signal on the sensor, significantly reduces this problem and enables accurate measurement even for very tight binders in the picomolar range.¹ Surfaces loaded to just 0.1 nanometers of density gave the cleanest kinetic data.⁵

The flip side is that very low loading levels produce small signals, which makes noise a bigger problem. Researchers constantly balance these trade-offs, loading enough molecule to see a clean signal but not so much that the kinetics become distorted. Orbital shaking speed, buffer composition, and sensor chemistry all factor in, and experienced users develop an intuitive feel for how these variables interact. It is one of those areas where the technology is conceptually simple but practically demanding.

Taming Nonspecific Binding

When the interaction being studied is weak or involves low concentrations of analyte, nonspecific binding (molecules sticking to the sensor surface for the wrong reasons) becomes a serious problem. It adds a false signal on top of the real binding event, and for weak interactions, that noise can swamp the measurement entirely. A systematic study of blocking agents found that saccharides, particularly sucrose, are effective at suppressing this background. Combining sucrose at 0.6 molar concentration with bovine serum albumin and 20 millimolar imidazole created a cocktail that outperformed any two-component mixture.⁶ This kind of practical optimization may sound unglamorous, but it is often the difference between publishable data and unusable noise.

Reference subtraction also matters. Running a parallel sensor through the same steps without the capture molecule, then subtracting that trace from the experimental one, removes bulk refractive index changes and systematic drift. Most BLI software supports this automatically, but the quality of the reference sensor preparation directly affects how clean the corrected data look.

Interactions Beyond Protein-to-Protein

BLI was originally developed with protein-protein interactions in mind, but the technology has proven remarkably flexible. Researchers have used it to study interactions between proteins and nucleic acids, including DNA-protein binding and the behavior of functional aptamers, short nucleic acid sequences designed to bind specific targets.⁷ In one well-characterized example, an anti-thrombin RNA aptamer was loaded onto a biosensor and dipped into thrombin-containing wells, generating clean kinetic data using the same straightforward dip-and-read approach.

Lipid interactions have also been brought into the BLI fold. By reconstituting membrane proteins into lipid nanodiscs, tiny disc-shaped membrane patches stabilized by scaffold proteins, researchers have measured the binding kinetics of a membrane-embedded enzyme with its inhibitors. This approach was used to study a yeast vacuolar enzyme, immobilizing the nanodisc-reconstituted complex on streptavidin-coated sensors and measuring both inhibitor binding and nucleotide-dependent enzyme disassembly.⁸ Membrane proteins have historically been among the hardest targets for binding assays because they tend to misfold or aggregate outside their native lipid environment. Nanodiscs largely solve that problem, and BLI’s tolerance for complex sample matrices makes it a natural partner for this strategy.

Small molecule fragment screening is another area where BLI has gained ground. Fragment-based drug discovery starts by testing large libraries of tiny chemical fragments for weak binding to a drug target, then builds those fragments into larger, tighter-binding compounds. BLI has been validated for this purpose, producing affinity constants and binding profiles that are generally comparable to those obtained by surface plasmon resonance, the traditional workhorse of fragment screening.⁹

Watching Enzymes Work, Not Just Bind

One of the more creative applications of BLI goes beyond measuring binding and instead uses the sensor to track enzymatic activity in real time. Ubiquitination, the process by which cells tag proteins for degradation or signaling, traditionally requires western blotting, a labor-intensive technique that gives a snapshot rather than a movie. A BLI-based alternative immobilized a ubiquitin ligase on the sensor and then dipped it into a solution containing the full set of ubiquitination machinery. As ubiquitin molecules were added to the enzyme on the sensor, the accumulating mass produced a rising BLI signal. The assay achieved a Z-factor of 0.642, indicating it was robust enough for screening applications, and control reactions lacking key components showed minimal background signal.¹⁰

Similarly, BLI can track conformational changes indirectly. In studies of a bacterial enzyme, the binding and release of a regulatory subunit was monitored on the sensor, and the effects of catalytic-site ligands on that subunit’s conformational state were measured in real time. Because BLI is not overly sensitive to what is dissolved in solution, researchers could add substrates, nucleotides, and various ligands without worrying about buffer-mismatch artifacts that plague some other optical methods.¹¹ That tolerance for varied solution conditions is an underappreciated advantage of the platform.

Counting Viral Vectors for Gene Therapy

Gene therapy has created an entirely new demand for BLI. Adeno-associated virus (AAV) vectors are among the most widely used vehicles for delivering therapeutic genes, and manufacturing them requires accurate measurement of how many viral capsids are in a given sample. The traditional method, ELISA, works but involves multiple incubation and wash steps that consume hours. BLI-based AAV titer assays use sensors functionalized with a capsid-binding ligand and can measure concentrations across a wide range, from about 10 billion to 1 trillion capsids per milliliter, for multiple AAV serotypes.¹²

Head-to-head comparisons have found that BLI and ELISA produce comparable capsid titers, but BLI offers a roughly 2.8-fold wider linear measurement range and significantly less hands-on work.¹³ For a manufacturing floor where process decisions need to be made quickly, shaving hours off a titer measurement is meaningful. BLI has been applied to both upstream monitoring (tracking virus production during cell culture) and downstream processing (confirming titers after purification steps), making it a useful tool throughout the production pipeline.

How BLI Stacks Up Against Other Methods

The most frequent comparison is between BLI and surface plasmon resonance, or SPR. Both are label-free, real-time methods that produce kinetic data. SPR uses a gold-coated chip and measures changes in the angle of reflected light as molecules bind near the surface, while BLI uses a fiber-optic sensor and measures interference shifts. Both are described as common approaches for quantifying binding events and providing kinetic parameters.¹⁴

In practice, the differences are mostly about workflow and throughput. SPR instruments typically use a continuous-flow microfluidic system, which delivers precise analyte concentrations to the chip surface but limits how many samples can be run in parallel. BLI’s dip-and-read format allows 8, 16, or even 96 sensors to run simultaneously, making it faster for screening large panels. SPR, on the other hand, tends to have higher sensitivity for very small molecules and extremely weak interactions, because the gold surface produces a cleaner signal with less drift. For most protein-protein and antibody-antigen work, the two technologies give similar results. For fragment screening, BLI has been validated but SPR remains the more established choice.⁹

Isothermal titration calorimetry, or ITC, is a completely different animal. It measures the heat released or absorbed when molecules interact in solution, with neither partner attached to a surface. This makes ITC the gold standard for measuring binding in its most native-like state. But ITC and BLI can give quite different affinity values for the same interaction. In one comparison, the affinity constant for a protein pair was measured at 0.12 micromolar by BLI but 5.0 micromolar by ITC, a roughly 40-fold difference. The discrepancy likely reflects differences in how the proteins are tagged, the buffer conditions, and whether conformational changes release heat that confounds the ITC measurement. In some cases, ITC failed to produce a usable measurement entirely, while BLI still returned clean data for the same interaction.¹⁵ The takeaway is not that one method is right and the other wrong. Different methods measure binding under different conditions, and the “true” affinity depends on which conditions best reflect biology.

Pushing Sensitivity with Signal Amplification

Standard BLI measures mass accumulation, which means very small or very low-abundance targets can fall below the detection threshold. This is a genuine limitation for applications like cytokine quantitation, where the molecules of interest may be present at picomolar or sub-picomolar concentrations. To address this, multi-step signal amplification strategies have been developed. In one approach, cytokines are sandwiched between a biotinylated capture antibody and an enzyme-conjugated detection antibody. The sensor is then dipped into a substrate reagent that produces an insoluble precipitate, amplifying the mass change far beyond what the cytokine molecule alone would produce.⁵

Other creative assay architectures have emerged for bispecific antibodies, engineered molecules designed to bind two different targets simultaneously. Bridging assays and dual-binding formats evaluated by BLI can determine whether both arms of a bispecific are functional and estimate the fraction of correctly assembled molecules in a manufacturing sample. These are questions that matter for drug quality control but would be difficult to answer with any technique that does not provide real-time, multi-step binding data.

Antibody-Free Detection and Molecularly Imprinted Polymers

Most BLI assays rely on biological capture agents: antibodies, streptavidin, or protein ligands coated onto the sensor. But biological capture agents can be expensive, fragile, and batch-variable. An emerging alternative uses molecularly imprinted polymers, synthetic materials formed around a template molecule that leave behind cavities shaped to rebind that same molecule. One group developed an antibody-free BLI biosensor for detecting immunoglobulin G1 in human serum by coating sensors with a molecularly imprinted polynorepinephrine layer. The platform was tested with different chemical linkers and evaluated for selectivity against other immunoglobulin classes and across the four IgG isotypes.¹⁶ If this approach matures, it could reduce the cost and variability of BLI sensors for clinical or industrial applications where antibody-based reagents are a bottleneck.

Whole Virus Characterization

Studying how intact viruses interact with their host cell receptors has traditionally been complicated by the fact that viruses present multiple copies of their binding proteins on a single particle. A single viral subunit tested in isolation may behave differently from the same protein displayed on the surface of a whole virion, where avidity effects (the combined strength of many simultaneous binding events) come into play. Both BLI and SPR have seen growing use for characterizing whole virus-ligand interactions, as these platforms can accommodate the larger particle sizes involved and provide a more accurate picture than subunit-only experiments.¹⁴ For emerging pathogens, where speed matters and viral material may be limited, BLI’s low sample consumption and rapid turnaround make it a practical first-pass tool for mapping receptor usage.

Understanding whole-virus binding is not just an academic exercise. Vaccine developers use this information to evaluate whether engineered antigens faithfully mimic the receptor-binding behavior of native virus, and public health researchers use it to track how mutations in circulating strains alter tropism. BLI’s ability to run these experiments in a high-throughput format, dipping rows of sensors into samples from different viral isolates in parallel, makes it well suited for surveillance-scale work.

Where BLI Reaches Its Limits

No technique does everything well. BLI’s reliance on mass accumulation at the sensor means it struggles with analytes that are too small to produce a detectable shift. Interactions involving molecules under a few hundred daltons often fall below the noise floor unless amplification strategies are used. Very fast kinetics can also be hard to resolve, since the dipping and mixing steps introduce a brief dead time before measurement begins. For the fastest association events, SPR’s continuous-flow format has an edge.

The surface-based nature of BLI means that one binding partner is always immobilized, which can alter its behavior. A protein tethered to a sensor by a tag at one end may not adopt the same conformation it would in free solution. The discrepancies between BLI and ITC affinity values noted earlier partly reflect this reality. Immobilization artifacts are manageable with careful experimental design, such as testing multiple tag placements or using oriented capture strategies, but they cannot be completely eliminated.

Sensor cost is another practical consideration. Each biosensor tip is typically used once, and in a high-throughput experiment running hundreds of sensors, consumable costs add up. Some users regenerate sensors by stripping and reloading the capture surface, but not every surface chemistry tolerates this well, and regeneration introduces its own variability. For labs running occasional experiments, the per-sensor cost may be negligible. For manufacturing groups running thousands of measurements per week, it is a line item worth negotiating.