Detecting and Characterizing Protein Self-Assembly In Vivo by Flow Cytometry

Understanding the physical basis of protein self-assembly in cells is limited by lack of appropriate tools. Current methods suffer from low sensitivity, indirect readouts, limited throughput, and/or population level resolution. In contrast, our assay detects and quantifies the propensity of a protein for self-assembly in vivo with a sensitive, single cell, high throughput readout irrespective of protein localization or solubility.

When first trying the assay, getting the photo conversion right may require some empirical testing to obtain optimal conditions for achieving a good readout across the entire plate. The sensitivity of the assay could also be compromised if the cytometer does not have separate detection filters for the FRET and acceptor signals. Demonstrating the procedure with me will be Andrew Box, the lab manager of the cytometry core.

To prepare a yeast culture for DAmFRET, for every query protein, inoculate the transformed yeast colonies in triplicate in 200 microliters of an appropriate non-inducing growth medium in individual wells of a 96 well round bottom culture plate. Incubate the yeast cells with shaking with a 1.5 millimeter orbit at 1, 200 rotations per minute at 30 degrees Celsius for 16 hours. At the end of the incubation, sediment the yeast cells by centrifugation and remove the supernatants by forceful inversion.

Add 200 microliters of an appropriate induction medium and return the cells to the shaking incubator for another 12 hours. At the end of the incubation, sediment the yeast cells by centrifugation and forceful inversion, then add 200 microliters of fresh induction medium. Then incubate the cells with shaking for an additional four hours to reduce auto-fluorescence.

To photo convert the samples, place the plate without the cover under an ultraviolet lamp fitted with a 320 to 500 nanometer filter and a beam collimator positioned 45 centimeters above the plate. Turn on the lamp to photo convert the cells for 25 minutes with shaking. For DAmFRET data collection, load the photo converted cells onto an imaging flow cytometer with non-colinear 488 and 561 nanometer lasers and gate four single unbudded cells by high circularity and a small cell area.

Then collect data for two to five times 10 to the four single cells per sample within a fluorescence positive gate. At the end of the analysis, use a non-photo converted mEos3.1 sample for the pure donor signal and monomeric DsRed two for the pure acceptor signal to set the compensation. For more sensitivity, ensure that the FRET detector channel is also considered as the spillover target in the analysis program.

Calculate the AmFRET parameter as the ratio of the total FRET signal divided by the total FRET acceptor signal. The AmFRET profiles of yeast cells expressing fluorescent protein alone exhibit negligible AmFRET while yeast cells expressing as stable homooligomer fused to the fluorescent protein demonstrate uniform positive AmFRET values. Images of the cells acquired by imaging flow cytometry reveal a diffused fluorescence in all of the channels.

Confocal microscopy of Z stacks for multiple fields of cells shows a uniform distribution of the fluorescence throughout the cytosol of each cell with no detectable puncta. Yeast cells expressing the fluorophore alone or the fluorophore plus the mutant form of a human inflammasome protein exhibit negligible AmFRET over the entire concentration range indicating an inability to self-interact. In contrast, the wild type inflammasome protein demonstrates a DAmFRET profile with one negligible AmFRET population and one high AmFRET population.

Note that the wild type inflammasome protein resists self-assembly even at high concentrations with the protein switching to the self-assembled form in only a fraction of the cells. This is a clear indication of the existence of a nucleation barrier to self-assembly. As confirmed by fluorescence imaging, these populations represent cells that contain either soluble protein only or mostly self-assembled protein respectively.

Indeed, DAmFRET corroborates prior structural data that the point mutant in the inflammasome disrupts nucleation across the concentrations achievable by this expression system. Note that protein expression profiles in mammalian cells that lack Caspase-1 expression resemble those observed in yeast cells. The most important step to remember is the photo conversion.

Other crucial steps include collecting a sufficient number of events that are not saturated for fluorescence signal and getting the compensation correct. Following up the assay with appropriate biochemistry, rational mutagenesis and structural biology approaches can help to fully elucidate the mechanism of the self-assembly.