Standard Protocol for Mass Photometry Experiments

Abstract

Mass photometry (MP) is a relatively new experimental technique with a quickly expanding list of applications. Using optical detection, MP measures the mass of individual molecules to obtain molecular mass distributions of proteins and other biomolecules in solution. The combination of speed, sensitivity, and very low sample consumption with label- and immobilization-free detection sets MP apart from other analytical methods. An increasing number of laboratories incorporates mass photometry as a routine sample analysis technique. However, MP measurements can sometimes be challenging, especially for users without previous experience with single-molecule techniques. Here we present a protocol for the determination of protein molecular mass distributions by MP. It describes the sample and materials preparation as well as data collection and analysis. The advantages and limitations of this technique and the potential sources of artifacts are also given. This protocol can be used by new MP users and serve as a checklist for laboratories routinely performing MP experiments to guide consistent data collection and documentation.

Introduction

Mass photometry (MP) is a light scattering based technique that detects individual, unlabeled molecules in dilute solutions. MP can accurately measure molecular masses in the 40 kDa to 5 MDa range, and provides information on the relative abundance of species by molecular counting (Cole et al. 2017). The final result of the MP measurement is a molecular mass distribution reflecting the molecular composition of the sample. Since its introduction (Young et al. 2018), MP has been successfully used to study a broad range of biological systems, and new applications of this technique are still being developed. To date, MP has been used to measure molecular masses of proteins and nucleic acids (Li et al. 2020; Young et al. 2018), characterize membrane proteins and micellar aggregates (Lebedeva et al. 2020; Olerinyova et al. 2021), study the heterogeneity of large protein complexes and measure protein-protein binding affinities (Sonn-Segev et al. 2020; Wu and Piszczek 2020).

While the MP measurements are fast and relatively easy to execute (Wu and Piszczek 2021), achieving reliable experimental results requires consideration of several factors that can affect data quality. This includes avoiding possible instrumental errors, selecting appropriate data analysis parameters, and careful handling of sample solutions at nanomolar concentrations. The protocol described here outlines a standard procedure used to collect molecular mass distributions of protein samples. With minor modifications it can also be used to study other molecules, including nucleic acids. The first area the protocol focuses on is sample preparation. The label- and immobilization-free detection is one of the main advantages of the MP technique but provides no detection specificity. For this reason, preparation of samples free of light scattering impurities is critical. The protocol also outlines data analysis and data presentation guidelines. MP data presentation should allow readers to easily compare results obtained in different laboratories and assess the quality of the experimental results. Moreover, the experimental conditions and data analysis parameters necessary to fully understand the results and their potential limitations should also be included. Since MP is a relatively new technique, this protocol can provide a basis for the further development of standards for MP data analysis and MP data presentation.

MP Standard Operating Procedure

1 Prepare the instrument and materials

• 1.1 Turn on the instrument to allow for thermal equilibration of the optics. For precise molecular mass determination, a warmup time of at least one hour is recommended.

• 1.2 Prepare sample chambers

  Note: Both glass flow chambers and silicon gasket wells can be used to hold the MP sample. Gasket wells are easier to prepare, but samples are diluted with buffer during the well loading procedure. Use flow chambers to avoid sample dilution during loading.

  Note: Coverslips quality varies between manufacturers, and it is necessary to test the glass surface before use. In some cases, only one side of the 24 x 50 mm coverslips has the optical quality suitable for MP measurements. If a new box of 24 x 50 mm coverslips is used, identify the high-quality side (working side) of the coverslips as described in step 1.2.2 and mark the working side orientation on the box. All coverslips in the box should have the same orientation.

  • 1.2.1 Clean coverslips with water, ethanol, isopropanol and dry with a stream of clean nitrogen (Wu and Piszczek 2021). During the cleaning procedure, track the orientation of the coverslips. If the working sides of the coverslips in the box have been previously identified, proceed to step 1.2.3.

  • 1.2.2 Position the coverslip in the instrument as described in step 2.1 and place a drop of distilled water in the sample region. Perform the focusing procedure (step 2.2). Examine the MP image - the surface of the glass should be free of large imperfections and the signal value (RMS deviation of the MP image) should be less than or equal to 0.05% (Fig. 1A, 1B). If the signal value is larger than 0.05%, discard the used coverslip and repeat the test on the opposite side of a new coverslip. If neither side of the coverslips has a signal within acceptable limits, the cleaning procedure has failed. Restart the experiment at step 1.2.1 with a new coverslip and a more thorough cleaning.

  • 1.2.3 Assemble flow chambers from stacked coverslips and double-sided tape as demonstrated in the video tutorial (Wu and Piszczek 2021). If using gasket wells, cut out a segment of silicone gasket well sheet, rinse it with water, and place the gasket on top of the working side of the coverslip. Gently press the silicone onto glass using a clean pipette tip. Use square gasket segments of 4 or 9 wells.

• 1.3 Prepare protein samples

  • 1.3.1 Filter all buffers with 0.22 μm syringe filters. Filter protein stocks or centrifuge for 10 min at the maximum speed of the tabletop centrifuge.

    Note: Most biological buffers are compatible with for the MP measurements, but buffers with salts concentration lower than 10 mM should be avoided. Glycerol concentrations higher than 5% v/v may affect MP focusing. Micelles formed in buffers containing detergents above their CMC concentration will usually produce unacceptable levels of the MP image background. Carrier proteins cannot be used for MP measurements since the typical carrier protein concentration is much higher than the acceptable MP protein concentration limit (high pM to mid nM). Measurements of DNA, nucleic acid binding proteins, and membrane proteins may require special buffer conditions or additional coverslip preparation. Buffers suitable for those measurements are described elsewhere (Li et al. 2020; Olerinyova et al. 2021).

  • 1.3.2 Determine molar concentrations of all protein stocks. The preferred method for the concentration determination is the 280 nm absorption measurement and calculation of the concentration based on the protein molar extinction coefficient value.

    Note: If the concentration of the sample cannot be determined, the required stock dilution ratio can be obtained experimentally by preparing a stock dilution series and identifying the solution with an optimal sample concentration by repeating MP measurements (step 2.4 note). This method, however, is more time consuming and requires a large amount of sample chambers.

  • 1.3.3 Prepare 50 μL protein solutions. For flow chambers, use 20 nM protein concentration. This will be the final measurement concentration when the flow chambers are used. When using sample wells, prepare 40 nM protein solutions. This solution will be further diluted by a factor of two during the well loading procedure.

    Note: The protein concentration range optimal for MP measurements is approximately 10 nM to 50 nM (Fig. 1C). Samples in this concentration range usually yield sufficiently high frequency of landing events without compromising their spatial separation. To achieve these conditions, the recommended 20 nM sample concentration can be adjusted if necessary—follow the note of step 2.4 of the SOP for the concentration adjustment procedure.

    Note: Protein solutions at nanomolar concentrations are prone to excessive material loss due to the surface adhesion. If maintaining a precise concentration value is important, or if an excessive loss of material is observed, an optimal vial material should be selected. Store the diluted protein solution in a vial for at least 60 minutes and compare the MP count rate with that obtained from a freshly prepared solution. Select the vial type with the smallest count loss. Vials can also be passivated with a Casein solutions (Soltermann et al. 2020). Repeated handling of dilute solutions with new pipette tips or vortexing can also contribute to the loss of material.

Fig. 1.

Representative screen shots from the MP camera. (A) Native (left) and ratiometric (right) buffer images on a clean coverslip. (B) Native (left) and ratiometric (right) images of an insufficiently cleaned coverslip surface. (C) Ratiometric view of the protein landing events at the optimal sample concentration.

2 Acquire MP data

• 2.1 Apply immersion oil on the microscope objective and position the coverslip on the microscope stage. Make sure that no air bubbles are trapped in the illuminated area of the immersion oil layer. Load 10 μL of clean buffer into to the sample compartment.

  Note: The back reflection from the surface of the liquid can form wave-like patterns in the image. To avoid this effect when using gasket chambers, place the gasket well slightly off-center over the microscope objective.

• 2.2 Focus the objective on the surface of the glass-buffer interface following the procure described in the data acquisition software manual. Details of the focusing procedure and the range of acceptable image quality parameter values may depend on the data acquisition software version. Engage the autofocus to maintain the optimal focus position.

• 2.3 Load the protein sample. When using flow chambers, load 20 μL of the 20 nM sample to fully replace the buffer in the flow chamber channel as demonstrated in the video tutorial (Wu and Piszczek 2021). When using gasket wells load 10 μL of the 40 nM solution into the buffer-containing well and use the pipette tip to mix the content of the well.

  Note: Larger dilution ratios for gasket wells are also possible but complicate the sample mixing procedure. For larger dilution ratios, increase the volume of the buffer loaded into the well in step 2.1 to load between 10 μL to 19 μL of buffer. During sample loading (step 2.3), add the protein stock volume required to achieve the final total sample volume of 20 μL.

  Note: Sample vials should be thermally equilibrated on the lab bench for 5 to 10 minutes before loading. Loading samples directly from vials kept on ice may affect the measured contrast values.

• 2.4 Select the desired acquisition time or the number of frames to be acquired and start the data acquisition. Typical data acquisition times are in the 60 s to 120 s range (approximately 60,000 to 120,000 video frames). See note for the adjustment procedure.

  Note: Molecules landing on the surface of the coverslip will appear as dark spots in the MP image (Fig. 1C). Monitor the density of the landing events after sample loading. The optimal sample concentration should allow the camera to register approximately 500 to 5,000 landing events in a 100 s data acquisition. Acquisition time can be extended to collect more landing events, but the density of the events decreases with time and long acquisition times are not practical. If the landing event density is too low, increase the sample concentration and repeat the measurement. Data from several acquisitions can also be combined to obtain the desired total number of landing events from low concentration samples. If the landing event density is too high, the molecules will spatially overlap resulting in a poor data quality. Decrease sample concentration until well spatially-separated landing events are observed. When the landing density of a sample is too high, but the landing density rate decreases over time, the start of the acquisition can be delayed until landing density decreases to a desired level.

  Note: If the acquisition camera region of interest (ROI) is software-selectable, the ROI can be adjusted for measurements of large molecular mass complexes. Expanding the ROI decreases instrument mass sensitivity but allows for monitoring of a larger image area. This helps to increase the number of detected landing events for larger, and therefore more slowly diffusing molecules, that exhibit lower landing event frequencies.

• 2.5 After data collection, the coverslip can be moved to position the next channel or sample well over the objective for a consecutive measurement. After the final data acquisition, remove the coverslip and immediately clean the objective with isopropyl alcohol using the cotton tipped applicator.

3 Analyze the MP data

• 3.1 Process the video file using the data analysis software to identify and quantify the landing events.

  Note: The data analysis software usually has adjustable filter parameters that can be modified to control the algorithms for the landing events identification and noise rejection. Identification of proteins in the 40 kDa to 50 kDa molecular mass range may require increasing the sensitivity of the analysis. When studying large molecular mass complexes sensitivity can be decreased to reject the background generated by smaller molecules. Reanalyze data modifying the filter parameters to optimize their values.

• 3.2 Examine changes of the signal level and the frame image appearance with time (frame index) to identify data collection artifacts. If necessary, spatial and temporal masks can be applied to eliminate the high noise areas.

• 3.3 Convert the contrast values to molecular masses using the instrument calibration function. The accuracy of the calibration function should be regularly validated by measuring a protein standard mixture (unstained protein ladder for electrophoresis, Fig. 2) and verifying the molecular masses of mixture components. As good laboratory practice, it is recommended to include the standard sample measurement with every series of experiments.

• 3.4 Plot the MP data as histograms or kernel density estimate (KDE) distributions. Adjust the histogram bin width or the KDE bandwidth value to account for all distribution components without amplifying the noise of the measurement. A good starting value for the bandwidth and the bin width parameters is 1/100 of the main region of interest (desired range) of the mass distribution.

  Note: KDE distributions are usually preferable when presenting several distributions on the same plot. Since KDE is a smoothing algorithm, the KDE bandwidth value has to be selected carefully to avoid oversmoothing.

• 3.5 Fit the distribution peaks with Gaussian functions to obtain the average molecular mass of each distribution component. The KDE peak finder can also be used, if available, to obtain the values of the components’ molecular masses.

• 3.6 Save the analysis results and export the distribution plot to document the results. The result files contain settings used in the data processing, but they should also be noted for the data report.

• 3.7 When documenting the experiment, note the sample and the buffer identity, the stock dilution rate, the final sample concentration, and the incubation time after the final dilution, if any. For the data acquisition, note the type of chambers used (and the dilution rate for the gasket wells), the approximate delay time between the sample loading and the start of the acquisition, and the length of the data acquisition. Indicate the ROI used if applicable. For the data analysis step, note what parameters were modified from their default values and the bin width or KDE bandwidth settings used. Indicate if the distribution plot shows the complete range of events or if the abscissa mass range was truncated. Note the total number of events in the acquisition and in the exported distribution.

Fig. 2.

Molecular mass distribution histogram of a standard protein ladder sample. The solid line represents major species fit with four Gaussian functions. Inset: linear fit of the contrast-to-mass calibration.

4 Calculate the relative species populations in the distribution (optional step)

Note: Relative species populations are directly calculated from the areas of the Gaussian peaks (step 3.5). This is usually sufficient, but when the accuracy of that calculation has to be verified, follow the procedure outlined in this step. This is particularly important when the relative species populations are used to calculate the binding affinities.

Note: Procedure described in this step requires a data file with a sufficiently large number of counts in each peak of interest. The precise value of the time interval between the sample loading and the start of the acquisition is also required.

• 4.1 Extract the list of landing events with their mass and time information from the result files saved in step 3.6.

• 4.2 Identify the peak limits for each species of interest from the mass distribution plot (step 3.5). For partially overlapping peaks, use the value of the distribution plot minimum between the two peaks or the crossing point of the peak’s Gaussian fits.

• 4.3 In the list created in step 4.1, select only the events within a mass range identified in step 4.2 for the first species of interest. Plot the histograms of the counts versus time for that species. The events time stamps have to be corrected to account for the acquisition start delay: t_corrected = t_recorded + τ_delay, where τ_delay is the time interval between the sample loading and the start of the acquisition.

• 4.4 Fit the histogram with the exponential decay function, f = a · e^−kt.

• 4.5 Integrate the best fit exponential function in the [0,∞] range. The integral represents the corrected total counts value for that species.

• 4.6 Repeat steps 4.3 to 4.5 for all species of interest.

Discussion

This protocol can be used to detect protein molecules in the 40 kDa to 5 MDa molecular mass range when using the MP instrument equipped with a green illumination laser (520 nm). Since the MP measurements are fast and require a very small amount of material, this technique has a broad analytical utility. It can be used, among other applications, to quickly determine the composition of protein solutions, verify sample purity, test the heterogeneity of protein preparations, study the oligomeric state of proteins, and detect the presence of protein-protein interaction complexes and determine their stoichiometry. MP is invaluable for the rapid evaluation of samples prepared for a more sample- and time-consuming analysis like analytical ultracentrifugation or cryo-electron microscopy. Besides the sample quality control applications, MP is an innovative research technique in its own right and is able to deliver data difficult to obtain with other analytical methods. Some of the applications particularly suited for MP are the analysis of multi-component mixtures and detection of low abundance molecular species.

To assist readers just starting to use the MP technique, the most common instrumental and sample preparation errors have been listed in the troubleshooting guide (Tab. 1). Table of materials (Tab. 2) has also been provided, but it should serve only as a guide. Other, equivalent supplies can also be used for MP experiments without negatively affecting experimental results.

Table 1.

MP troubleshooting guide.

Problem	Cause	Solution
Inaccurate molecular mass results	• Inaccurate focus	• Repeat the focusing procedure, making sure that the global maximum of the “sharpness” parameter has been found
	• Temperature shifts	• Equilibrate the sample in room temperature before loading
		• Increase the warm-up time of the instrument and check the stability of the lab temperature
	• Calibration and sample buffer mismatch	• Measure protein standards for mass-to-contrast calibration in the sample buffer
Noisy contrast distribution	• Number of landing events too small	• Increase the sample loading concentration
	• Histogram bin size too small	• Adjust the histogram bin size
	• Impurities in the sample or buffer	• Filter buffer and/or improve sample purity
Poor peak resolution	• Large number of overlapping landing events	• Decrease the sample concentration
	• Image noise generated by buffer components	• Reduce concentration of buffer additives and co-solutes
	• Impurities in the sample or buffer	• Filter buffer and/or improve sample purity
Too many detaching events/white spots in MP images	• Landing events frequency too high	• Decrease the loading concentration
	• Charge and other buffer effects reducing non-specific binding to the coverslip	• Modify buffer composition, increase ionic strength of the buffer if salt concentration is low

Table 2.

Materials

Name	Company	Catalog number
Cotton tipped applicators	Thorlabs	CTA10
Coverslips 24 x 24 mm	Globe Scientific	1405–10
Coverslips 24 x 50 mm	Fisher Scientific	12–544-EP
Immersion oil	Thorlabs	MOIL-30
Syringe filter 0.22 μm	Millipore	Millex-GS MCE SLGSV255F
Reusable gasket wells	Grace Bio-labs	CultureWell™ CW-50R-1.0
Mass photometer	Refeyn	OneMP

The procedure outlined in steps 1–3 of this protocol is usually sufficient for most applications listed above. Special care has to be taken when mass photometry data are used to calculate protein-protein binding affinity values or protein oligomerization constants. In that case, it is advisable to follow the correction procedure described in step 4 of the protocol to account for differences in the landing rate decays that may exist between different molecular species in the sample. Another factor that has to be considered when using MP to determine binding affinity is the time required to reach chemical equilibrium after the protein stock dilution steps. Since MP is used to measure relatively strong binding affinities (Sonn-Segev et al. 2020; Wu and Piszczek 2020), longer sample incubation times may be required. And finally, when performing affinity measurements, the protein concentrations have to be selected to obtain solutions with sufficient populations of both the free and complex species. This is often difficult due to the relatively narrow optimal MP concentration range. For very strong interactions, sample concentrations in a sub-nanomolar range can be measured by MP using constantly perfused cells (Soltermann et al. 2020). Measurements at a total protein concentration above the typical 100 nM MP concentration limit can sometimes be performed when one of the proteins has a molecular mass below the MP mass detection limit of 40 kDa (Sonn-Segev et al. 2020; Wu and Piszczek 2020).