From Downscaling to Single-Cell Proteomic: Understanding and minimizing the downscaling effect

INTRODUCTION: Cellular systems consist of a variety of cells with distinct molecular and functional properties based on their location and timing. The characterization of proteome heterogeneity in these systems is a key to enhancing medical research and precision medicine, which requires quantification of proteins at the single-cell level. Liquid chromatography mass spectrometry (LC-MS) is a well-suited technique for proteomics analysis. However, the drop in performance is an inherent effect of decreasing the starting sample material amount. This downscaling effect is strongly related to the sensibility of MS instrument and the peptide loss during sample preparation. A comprehensive study of all contributions to the sample downscaling effect is essential to properly optimize single-cell proteomic methods and thus for minimizing the performance drop.

METHODS: Here we present results from downscaled proteomic analysis together with a software-assisted strategy to evaluate and furthering the understanding of the sample downscaling effect on the performance drop for bottom-up proteomic analysis. In this approach, a sample condition (e.g., sample preparation protocol, LC-MS method, MS instruments) is evaluated by monitoring performance when the injected quantity of peptides is reduced. Each peptide signal intensity is monitored in function of the injected quantity in LC-MS. The signal drop for each peptides can then be study independently in function of their intrinsic properties (e.g., mass, charge, hydrophobicity factor, acidic/basis residue ratio) to highlight the causes of peptide loss.

PRELIMINARY DATA: This approach was first used to evaluate and compare the performances of LC-MS methods (e.g. sensitivity, feature detection) in a high throughput context on QExactive (Thermo) and timsTOF (Bruker) instruments. This study led to a comprehensive comparison of the performances of these two MS for proteomics of low amount to single-cell level samples. The same workflow was then applied to evaluate specific peptide interactions (binding) with the vial surface. Peptides from a HeLa tryptic digest standard were chosen as peptide mix model for this study to avoid the contribution of sample preparation on the analysis performance. This study was first conducted with Total Recovery glass vials from Waters. The loss of peptide signal was assessed by a downscaling experiment on a set of vials containing peptides at different concentration levels (from 180ng/µL to 10ng/µL). As expected, the total peptide signal decreases as the total peptide concentration is reduced. However, the drop in peptide signal was not homogenous in regards to peptide hydrophobicity factors. This observation has been related to preferential and significant peptide binding on the vial surface. These interactions becoming non-negligible when peptide concentration is downscaled. Based on these results, vials molded in different polymeric materials (e.g., glass-filled polymers, Cyclic olefin polymers, polypropylene, poly(methyl methacrylate, polyether ether ketone) were tested with our downscaling approach. The results helped determine the best candidate polymeric material for vials or other laboratory consumables, minimizing peptide-surface interactions, for single cell proteomic analysis or low starting material experiments LC-MS analyses. As preliminary results, poly(methyl methacrylate), PMMA, vials showed promising behaviors for downscaled proteomics increasing the number of hydrophobic peptides detected by LC-MS compared to glass vial. This improvement leads to a 15% increase in identified proteins.

NOVEL ASPECT: Influence of microtube material for single-cell sample preparation and related bioinformatic tools for performance check and protocol optimization.