A Novel LC System Embeds Analytes in Pre-formed Gradients for Rapid, Ultra-robust Proteomics

Abstract

To further integrate mass spectrometry (MS)-based proteomics into biomedical research and especially into clinical settings, high throughput and robustness are essential requirements. They are largely met in high-flow rate chromatographic systems for small molecules but these are not sufficiently sensitive for proteomics applications. Here we describe a new concept that delivers on these requirements while maintaining the sensitivity of current nano-flow LC systems. Low-pressure pumps elute the sample from a disposable trap column, simultaneously forming a chromatographic gradient that is stored in a long storage loop. An auxiliary gradient creates an offset, ensuring the re-focusing of the peptides before the separation on the analytical column by a single high-pressure pump. This simplified design enables robust operation over thousands of sample injections. Furthermore, the steps between injections are performed in parallel, reducing overhead time to a few minutes and allowing analysis of more than 200 samples per day. From fractionated HeLa cell lysates, deep proteomes covering more than 130,000 sequence unique peptides and close to 10,000 proteins were rapidly acquired. Using this data as a library, we demonstrate quantitation of 5200 proteins in only 21 min. Thus, the new system - termed Evosep One - analyzes samples in an extremely robust and high throughput manner, without sacrificing in depth proteomics coverage.

Keywords: Automation; Clinical proteomics; HPLC; High Throughput Screening; Mass Spectrometry; Pre-formed gradient; Robustness; StageTip.

Graphical abstract

Fig. 1.

Evosep One flow diagram and time schedule. A, Almost all of the system runs at low pressure (10–20 bar), increasing the lifetime and robustness of the LC. Only a single pump and flow path operates at high pressure and this does not involve any solvent mixing. B, Stepwise timetable including all steps that the Evosep One is performing during a LC-MS run for the 60 samples/day method. The activities for the autosampler, the low-pressure pumps and the high-pressure pump are color-coded in green, yellow and blue, respectively. For a detailed flow diagram with highlighted flow path states see supplemental Fig. S2–S9.

Fig. 2.

UV set up to test gradient storage. A, Flow diagram for testing potential gradient mixing during storage in the capillary loop. B, Profiles of the acetonitrile and water plugs that were recorded by the UV detector for different storage times. Profiles were almost completely superimposed, consistent with minimal mixing of the two phases during storage.

Fig. 3.

Pre-formed gradient. A, Peptides are eluted from the C18 containing Evotip by pumps A and B. Low pressure pumps C and D form the final gradient, which is stored in the capillary loop together with the analytes. Subsequently, the valve switches and the high-pressure pump (H) simply pushes the gradient with its peptides over the analytical column. B, Composition of the gradient resulting from the confluence of the flows from pumps A, B and pumps C, D (x axis designates the volume entering the storage loop). The proportion of acetonitrile is indicated on the y axis. C, Analytes embedded in the storage loop are represented in red and as peak intensities. Because of the offset provided by pumps C and D, peptides are shortly retained at the head of the analytical column and elute with narrow peak widths. D, Comparison of three base peak chromatograms from a HeLa digest, demonstrating the reproducibility.

Fig. 4.

Robustness evaluation. A, Error frequency during the development phase of the system assessed by consecutive measurements of HeLa digests. B, The first and last base peak chromatogram of a HeLa digest in a series of 1500 measurements using a 22 min gradient. C, Pressure profiles over the gradient for the first and last three HeLa digests of the same experiment.

Fig. 5.

Clinical applicability to the plasma proteome. A, Retention time stability of selected peptides spanning a range of elution times over 96 plasma proteome runs. B, Pearson correlation matrix comparing all 96 plasma runs to each other. A single correlation graph with the median Pearson value is shown in the inset. C, Summed total peptide intensities in alternating plasma and blank runs.

Fig. 6.

Evosep One methods and chromatographic performance. A, Extracted peaks of synthetic peptides (colored) in a HeLa background (gray). The inset illustrates the extracted peak properties. B, For ease of use, five optimized methods have been pre-set to provide the best performance to time compromise. They are defined by the total number of samples that can be run per day rather than referring to the length of the gradient. The peak width and peak capacity values are averages on a HeLa digest with spiked in synthetic peptides (for details see supplemental Fig. S10–S14). C, Technical replicates of a digest of the UPS1 Proteomic Standard were injected 200 times with the 200 samples/day method. The number of identified proteins for each sample is shown as a bar graph in chronological order.

Fig. 7.

Rapid generation of mammalian cell line proteomes. A, Table for the comparison of the Evosep One with the Easy-nLC 1200, including total measurement time, gradient time and the numbers for identified proteins and peptides. B, Numbers of identified peptides per fraction over the 46 high pH reversed-phase fractions for both LC systems. C, Cumulative numbers of unique peptides across the fractions.

Fig. 8.

Rapid generation of mammalian cell line proteomes. A, Two scan modes for the acquisition of DIA data were devised and tested. B, Average number of precursors, identified peptides and protein groups for five HeLa measurements with 21 min gradients on the Evosep One. C, Number of proteins quantified with a coefficient of variation (CV) below 20 and 10%.