Enhancing Protein Binding Studies with a Light-Scattering Toolkit

September 1, 2015
Daniel Some

Daniel Some, PhD, is director of marketing and principal scientist, at Wyatt Technology Corporation, 6300 Hollister Ave, Santa Barbara, CA 93117-3253.

BioPharm International, BioPharm International-09-01-2015, Volume 28, Issue 9
Page Number: 40–46

imagewerksBiophysical binding studies utilizing surface plasmon resonance, biolayer interferometry, isothermal titration calorimetry, or related techniques are ce

Biophysical binding studies utilizing surface plasmon resonance, biolayer interferometry, isothermal titration calorimetry, or related techniques are central to the selection and optimization of biotherapeutic candidates based on proteins, immunoglobulin G (IgG), or advanced formats, including bispecifics and fusion proteins. Though these screening and characterization technologies are well established, in a variety of circumstances, binding measurements may be ambiguous or even fail to provide useful data. This article discusses a suite of complementary techniques, all based on light scattering, that are useful in troubleshooting many of the underlying characterization issues. These techniques can help investigators assess solution quality prior to running binding experiments, qualify aggregation behavior of analytes, and characterize complex interactions that may not be amenable to standard characterization methodology. Judicious use of a biophysical light-scattering toolkit is essential for robust and reliable interaction studies.

Common problems in interaction measurements
Experienced practitioners of biophysical techniques-designed to measure biomolecular interactions, such as surface plasmon resonance (SPR), bio-layer interferometry (BLI), isothermal titration calorimetry (ITC), and others-will recognize the following not-uncommon scenarios:

  • Poor fits of the signals to theoretical binding curves

  • Ambiguous determination of the appropriate association model

  • Erratic or irreproducible data

  • Results that vary with immobilization chemistry or chip coating

  • Results that depend on which binding partner is immobilized or titrated

  • Results that vary greatly from lot to lot of reagent

  • Fouling of microfluidic channels.

The causes of these difficulties are varied, so it is convenient to divide the most common into two categories: quality control and method limitations.

Quality control issues typically arise from suboptimal sample preparation, purification, or formulation. Manifestations may include aggregates, particulates, other impurities, concentration-dependent self-association, or undesirable analyte surface adhesion (e.g., stickiness). While sample quality control problems tend to be more prevalent during early phases of expression, purification, and process development of a biomolecule, established processes and commercially available reagents are by no means immune to such excursions from the normal standard of quality.

Method limitations include those imposed by surface-based methodologies and those imposed by indirect interaction reporter signals. The former are generally related to ligand immobilization, while the latter is common to many types of solution-based techniques. Every experimental technique is subject to specific pros and cons, and often, the application of several techniques is necessary to complete the picture of an interaction or to cross-validate results. This is particularly true when the complexes that form go beyond standard homo/heterodimer or trimer stoichiometries.

Light scattering: The solution for characterizing solutions
Many of these troublesome phenomena may be avoided or overcome by applying one or more characterization techniques based on light scattering. Described in the following examples are the light scattering tools most complementary to biophysical interaction analysis: size-exclusion chromatography–multi-angle light scattering (SEC–MALS), dynamic light scattering (DLS), composition-gradient–multi-angle light scattering (CG–MALS), and composition-gradient–dynamic light scattering (CG–DLS). Each of these tools consists of a sample preparation and delivery method combined with one of the two primary flavors of light scattering for solution characterization: multi-angle light scattering (MALS) and dynamic light scattering (DLS).

MALS is a first-principles technique for determining the molecular weight and size of macromolecules and nanoparticles in solution/suspension (1). A beam of light impinges on the solution. While most of the light traverses the solution volume unimpeded, a fraction of the photons meet solute molecules and scatter in all directions. The magnitude R of scattered intensity relative to the incident intensity (excess Rayleigh ratio) is directly related to the molecular weight of the solute M, the solute’s mass/volume concentration c, and the relative difference in refractive index of the solute and solvent dn/dc. There is also an angular dependence to the scattered intensity when the particles are larger than approximately 25 nm in diameter, but because most proteins and other biomolecules involved in interaction studies do not exceed this size limit, we will hereafter ignore the angular dependence.

The fundamental thermodynamic relation, which may be considered the “ideal gas law” for light scattering-representing the limit of a dilute system of point-like particles-is expressed as R=K·M·c·(dn⁄dc)where is a system constant. By measuring the scattered intensity and the concentration, MALS provides a first-principles determination of molar mass. When more than one species of solute is present, MALS measures the weight-average molar mass Mw=∑iMici/∑ici.

DLS is a first-principles technique for measuring the diffusion coefficient of macromolecules and nanoparticles in solution/suspension and for estimating particle sizes and size distributions (2). As in a MALS measurement, in DLS, a laser beam impinges on the solution and scatters from particles. Unlike in MALS, a DLS measurement does not consist of simply measuring the average scattered intensity, but in determining its rate of fluctuation. These fluctuations arise from the scatterers’ Brownian motion, the rapid jittering of particles buffeted by solvent molecules. Because the rate of fluctuation is correlated to the particle’s diffusion rate, diffusion coefficient(s) may be determined by analyzing the DLS fluctuations. A measure of particle size known as the hydrodynamic radius rh is calculated from the Stokes-Einstein equation rh=(kBT)⁄(6πηDt).



When more than one population of particle size is present in the solution, DLS analysis may reflect their distribution. Depending on the range of sizes, the distribution is typically described in one of two ways:

  • Polydispersity: If the size range is limited around a central value, DLS reports the average size and the polydispersity index, which represents the width of the distribution.

  • Multi-modal distributions: If size populations are separated by a factor of 3–5x, a distribution exhibiting distinct peaks around each size range may be calculated. The polydispersity index of each peak may also be assessed.

SEC–MALS uses SEC coupled with MALS to determine an accurate distribution of solution molecular weights of a sample from first principles. SEC separates molecules according to hydrodynamic size, and in standard SEC analysis, the retention time is related to molecular weight via a series of calibration standards. Hydrodynamic size may or may not be related directly to molecular weight; the specific size/weight relationship depends on a molecule’s conformation. Additionally, molecules may interact non-ideally with column-packing materials, or columns may age, leading to skewed retention times. Hence standard SEC calibration using globular, hydrophilic proteins often does not reflect the true molecular weight of the eluting sample, as shown in Figure 1. SEC–MALS overcomes these limitations by determining independently the molar mass of each elution volume, regardless of elution time (3). While there are a variety of important uses of SEC–MALS for protein characterization, including analysis of the solution molecular weights and conjugation state of glycoproteins and membrane proteins, the applications most useful in the context of biomolecular interactions are:

  • Determination of a protein’s oligomeric state in solution

  • Assessment of aggregates

  • Testing for self-association

  • Preliminary analysis of heterocomplex stoichiometry, especially when binding affinity is high enough to survive dilution in the SEC column.

CG–MALS couples a MALS detector with a composition gradient (CG) system, which automatically prepares a series of concentrations or compositions and injects them sequentially to the MALS and concentration detectors, with no accompanying separation step (4, 5). Each composition is analyzed by MALS to determine molecular weight (Mw), which increases with the formation of complexes in a direct and intuitive fashion with respect to the nature of the complexes formed. Hence, dimer formation results in complexes that scatter twice as much as the two individual monomers combined; trimer formation results in complexes that scatter three times as much as the three individual monomers. 

dynamic light scattering (HT–DLS) analysis of proteins samples in situ in a 96-well plate (DynaPro Plate Reader, Wyatt Technology). The analysis is color-coded to highlight different levels of purity, indicating which solutions will provide high-confidence results and which are not suitable for further use.A complete CG–MALS analysis determines equilibrium dissociation constants Kd as well as absolute molecular stoichiometries-the number of molecules of each type in the complex, rather than the ratio of moles of each type that react to form the complex. Hence, CG–MALS excels at analyzing self-association to determine the oligomer(s) formed, as well as complicated interactions involving multiple complexes, simultaneous self-and hetero-association, and cooperative interactions-all in solution without labeling or immobilization. As a general rule of thumb, CG–MALS is appropriate when the molar mass of the complex is at least 10% greater than the largest constituent monomer’s molar mass.

Batch DLS simply refers to DLS measurements in a cuvette or other vessel with no separation. High-throughput–dynamic light scattering (HT–DLS) performs standard DLS analysis in high throughput format, in situ, in standard microwell plates, making possible analyses that otherwise would be too onerous to carry out on a regular basis. Figure 2 illustrates a common solution quality visualization scheme.

CG–DLS is similar in concept to CCG–MALS, but the average size (rather than the average molar mass, as in CG–MALS) is measured as a function of concentration and/or concentration to assess interactions (6). While the range of Kd that may be quantified via CG–DLS is several orders of magnitude less that CG–MALS, and analysis of sizes is not as rigorous as analysis of masses to calculate interaction parameters, CG–DLS does offer a few advantages, including much lower sample consumption, simple operation in a microwell-plate format, and the ability to readily carry out temperature ramps to determine entropy and enthalpy of the interaction via van’t Hoff analysis. Perhaps one of the most useful aspects of CG–DLS is its ability to quickly screen hundreds of binding partners, in solution, with a simple determination or validation of complex stoichiometry.

Quality control with light scattering
Clean protein solutions with minimal aggregates, particulates, or other impurities are crucial to obtaining accurate and repeatable binding data. A publication (7) outlined how poor solution quality may impact SPR or BLI, with phenomena that include noisy signals, unusual/poorly fitting binding curves, low active concentrations, and even fouling of microfluidic channels. The presence of two analytes (e.g., monomer and dimer) that bind to the immobilized ligand, for example, will lead to a binding curve that includes two “on” rates and therefore will not be well represented by a single exponential fit. In another example, an aggregate containing multiple binding sites will often exhibit avidity, or anomalously enhanced binding and slow “off” rates not governed by the usual fit that assumes a single binding site.

Other interaction analysis techniques are no less susceptible to skewed and erroneous results caused by poor solution quality. All solutions (including pure buffer) should be prescreened to test for aggregation, particulates, etc. The quickest and easiest means for assessing solution quality is batch DLS, which can be performed with as little as 1 µL of sample (4–20 µL is more typical) and just a few seconds of measurement time (10–30 seconds is typical). Batch DLS provides a low-resolution size distribution covering 0.2–2500 nm in rh, highlighting the presence of large particles or aggregates as well as of low oligomers. 

In a quality-driven workflow, all solutions would be prescreened by batch DLS. Solutions with large particulates would be filtered or centrifuged to remove the particles, and the solution rechecked. Buffers should be cleaned up to see no appreciable particles above ˜0.3 nm. For protein solutions, if no large particles (i.e., more than 3–5 times the size of the monomer) remain, then the polydispersity of the protein peak should be checked. Because it is possible to filter protein solutions to 0.02 µm using syringe-tip filters (Whatman), most larger oligomers may be removed as well. 

When large numbers of solutions are to be prescreened, HT–DLS offers a ready solution. Because it uses standard microwell plates and measures in situ in the plate with no liquid handling, HT–DLS is ideal for integrating with interaction analysis technologies that sample from these plates (SPR, ITC) or make the measurements directly in the plates (BLI).

The final quality check of a protein solution consists of a SEC–MALS measurement to verify the oligomeric state of the protein and determine size and quantities of low oligomers. More than one interaction measurement has been led astray by assuming that the protein was monomeric in solution because a single peak appeared in the SEC run, or because a denaturing gel assay indicated the monomeric molecular weight. SEC–MALS provides confidence in the reagents to be used in the binding measurement, and the workflow should only proceed if an investigator is satisfied with the SEC–MALS results.



Overcoming method limitations with light scattering
Surface-based techniques such as SPR and BLI necessarily rely on assumptions regarding the exposure of relevant epitopes to the solution, which may not hold in certain circumstances. More common, though, are confusing phenomena such as avidity (the anomalously strong binding to immobilized ligands of analytes presenting two binding sites), surface-chemistry-dependent results (often related to the charge of the immobilization layer such as dextran), and complex multi-valent or cooperative interactions that cannot occur when one binding partner is immobilized.
While the limitations of surface interaction techniques are generally overcome by solution-based techniques such as ITC, fluorescence anisotropy, or microscale thermophoresis (MST), most of these solution-based assays are subject to their own limitations. A common drawback is the need for fluorescent labeling, potentially modifying the interaction. In addition, most solution-based measurements involve an indirect reporter signal that is assumed to represent the effect of a binding interaction, but cannot be unequivocally assigned to the formation of a specific biomolecular complex. For example, ITC measures the release or uptake of heat; while there is a good probability that this thermal signal is the result of association or dissociation, ITC does not offer direct proof of complex formation or dissociation and cannot indicate unambiguously which complex(es) form, especially when characterizing self-association. In instances of purely entropic binding, no thermal signal is available to report the interaction.

Analytical ultracentrifugation–sedimentation equilibrium is quite useful for analyzing a variety of interactions. Its primary limitation is the long time required to equilibrate, during which it is possible that sensitive proteins degrade.

Use of SEC–MALS, CG–MALS, and CG–DLS helps overcome most of aforementioned limitations and are excellent complements to other interaction analysis technologies. In some cases, CG–MALS may be the only technique capable of fully teasing out a complex interaction with relative ease. 

For self-association, it is often simplest to begin with SEC–MALS. Plots of molecular weight that vary along the peak in a concentration-dependent manner and that vary similarly with the concentration of the injected aliquot are good indicators of a self-associating protein, as shown in Figure 3. Under certain assumptions, it is even possible to determine Kd of simple oligomerization such as dimer or trimer formation (8). Alternatively, other techniques may suggest that self-association occurs. In a robust approach, the workflow would continue from an initial self-association indication to a CG–MALS analysis, which would accurately determine the oligomers that form-even if multiple oligomeric states are involved-and the binding affinity for each oligomer.

For hetero-association, too, SEC–MALS may be an excellent initial indicator of complex stoichiometry. A series of pre-incubated aliquots prepared at different stoichiometric ratios may be tested, with analysis consisting of determining the molar mass of the eluting complexes and unbound monomers (2). For example, if the complex is 1:1, and the aliquot is prepared at a 1:1 ratio, then little to no unbound monomer should remain, but if the complex is 2:1, then there will be an excess of B and a significant amount of unbound B should elute in a separate peak from the complex. While surface-based techniques are quite limited in this respect, other solution-based techniques also give good initial indications of hetero-association stoichiometry. 

A robust workflow designed for fully characterizing a hetero-association should proceed to a CG–MALS analysis such as that of Figure 4. CG–MALS will validate the true molecular stoichiometry of the complex and determine Kd by fitting the data of Mw vs. composition under the correct association model. 

A variety of advanced interactions are not addressed well by most techniques. These include simultaneous self- and hetero-association, interactions between two multi-valent molecules, higher-order self-assembly, and cooperative effects that lead to formation of higher-order complexes. Because CG–MALS is a first-principles, solution-based technique that provides a direct reporter signal (molecular weight), it is well suited for tackling such challenging interactions and providing a comprehensive analysis. In parallel, SEC–MALS may provide an initial indication of what forms are present in solution, guiding the CG–MALS method design.

CG–DLS may be used in a similar manner as CG–MALS, with a lower range of measurement and less robust analysis albeit much lower sample consumption. CG–DLS becomes limited in terms of analysis when the order of the interaction becomes high, since the relationship between stoichiometry and rh may become ambiguous.

Interactions between molecules are complicated. Careful characterization of reagents, as well as a critical examination of binding studies, is necessary to guarantee accurate and repeatable results, regardless of the techniques applied. Many of the challenges related to both reagent characterization and validation of interaction measurements may be addressed successfully through a suite of analytical tools involving light scattering. Light-scattering analyses are generally straightforward, intuitive, and informative. They solve analytical questions with minimal time and effort, enhance productivity, and minimize ambiguities in interaction studies.

1. P.J. Wyatt, Analytica Chimica Acta 272, pp. 1–40 (1993).
2. I. Teraoka, Polymer Solutions: An Introduction to Physical Properties, (John Wiley & Sons, Inc. ISBN 0-471-38929-3, New York, NY, 2002).
3. J. Wen, T. Arakawa, and J.S. Philo, Analytical Biochemistry 240, pp. 155–166 (1996).
4. D. Some and S. Kenrick, “Characterization of Protein-Protein Interactions via Static and Dynamic Light Scattering” in Protein Interactions, J. Cai and R. Wang Eds. (InTech, 2012), pp. 401–426.
5. D. Some, Biophysical Reviews 5 (2), pp. 147–158 (2013).
6. A. D. Hanlon, M. I. Larkin, and R. M. Reddick, Biophysical Journal 98 (2), pp. 297–304 (2010).
7. D. Some, “Protein Quality Control in SPR and BLI High-Throughput Screening Studies,” accessed Aug. 14, 2015. 
8. S. Das et al., J. Bacteriol. 190, pp. 7302–7307 (2008).

About the Author
Daniel Some, PhD, is principal scientist and director of marketing at Wyatt Technology Corp., dsome@wyatt.com.

Article DetailsBioPharm International
Vol. 28, No. 9
Pages: 40–46
Citation: When referring to this article, please cite it as D. Some, "Enhancing Protein Binding Studies with a Light-Scattering Toolkit," BioPharm International, 28 (9) 2015.