System and method for proteomic analysis of limited samples

Abstract

Top-down proteomics analysis of limited biological samples as small as a single cell is provided using ultralow flow liquid chromatography with novel porous layer open tubular (PLOT) columns combined with mass spectrometry. The PLOT columns are fabricated such that the porous polymer stationary phase has uniform thickness and improved durability, run- to-run and column-to-column reproducibility, and resists clogging. As a result, the new PLOT columns are especially well-suited for all modes of proteomics analysis.

Description

[0001] System and Method for Proteomic Analysis of Limited Samples

[0002] CROSS-REFERENCE TO RELATED APPLICATION

[0003] This application claims priority to U.S. Provisional Appl. No. 63 / 544,485, filed 17 October 2023, which is hereby incorporated by reference in its entirety.

[0004] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0005] This invention was made with government support under Grant Numbers 1 R35GM136421 and R41GM156145 awarded by the National Institutes of Health. The Government has certain rights in the invention.

[0006] BACKGROUND

[0007] Top-down proteomics (TDP) is a powerful approach for the analysis of purified proteins, protein complexes, and complex biological samples, which has been applied to answer diverse biological and clinical questions.1The term “proteoform”, introduced in 2013, describes all different molecular forms and the chemical diversity of proteins that originate from a single gene, taking into consideration changes caused by genetic variations, mutations, RNA splicing, proteolytic activity, and post-translational modifications (PTMs).2In contrast to the bottom-up proteomic approach, which relies on enzymatic digestion of proteins into peptides prior to performing mass spectrometry (MS), the top-down proteomics MS approach examines intact proteins at the MS1 level and utilizes various fragmentation strategies to gain sequence information for identification and structural characterization. Top-down analysis has demonstrated several advantages over bottom-up approaches in its potential to access minimally altered structures of biologically active proteins and non-covalent protein assemblies, and the ability to locate and characterize PTMs.3Also, sample preparation in top- down proteomic analysis typically requires less time and sample handling due to omitting timeconsuming protein digestion steps, which minimizes sample losses from additional handling. Despite recent advancements in top-down proteomic analysis of limited biological samples, many challenges remain, including the lack of sensitivity and the absence of universal and standardized sample preparation methods.4’5Additionally, the detection of proteins in their intact state by electrospray ionization (ESI) MS is challenging because of the splitting of the MS signal into multiple ion product channels as an outcome of several factors including the presence of multiple charge states of the same proteoform, adduct formation, PTMs and other modifications, and isotopic distribution. In addition, low protein solubility, poor protein separation, ionization suppression, slow data acquisition rate, and inefficiency of fragmentation for large protein species are further challenges preventing the broad adoption and success of top-down proteomics.6-9

[0008] Recently, several groups have demonstrated the benefits and feasibility of top-down analysis of limited samples (e.g., microneedle biopsies, populations of rare cells) and single cells in their proof-of-concept studies.4’5’8’10Zhou and coworkers applied their bottom-up microfluidic sample preparation system, termed nanoPOTS (nanodroplet processing in one pot for trace samples), to top-down proteomics, while being able to identify ~170 to ~620 proteoforms from ~70 to ~770 HeLa cells.8Another microfluidic sampling methodology that integrates single-cell extraction from MS-incompatible media with electrical cell lysis and nanoESI-MS was used in the analysis of single human erythrocytes.10Processing small numbers of intact cells (<10) and single cells directly in the separation capillary, online coupled with capillary electrophoresis-tandem mass spectrometry (CE-MS / MS) was shown to be an effective strategy to reduce sample handling and transfer steps, and a range of 17-40 proteins and 23-50 proteoforms was identified from single HeLa cells.4Melby and colleagues demonstrated the capabilities of one-pot sample preparation coupled with highly sensitive capillary LC-MS / MS in the analysis of single muscle cells while detecting large proteoforms (>200 kDa) and achieving single-cell resolution.5The above-mentioned studies highlight the potential and the importance of top-down proteomics analysis of limited samples and single cells and represent a significant step towards our better understanding of phenotypic heterogeneity and functional diversity of living systems.

[0009] The separation of proteins upstream of MS detection is critical to increasing the depth of proteomic profiling. In addition to pressure or electric field-driven separations, ion mobility separation is another way to boost the identification results. High-field asymmetric waveform ion mobility spectrometry (FAIMS) is an atmospheric pressure ion mobility technique that separates gas-phase ions by their characteristic charge and collisional cross-section differences in an asymmetric electric field.18Reportedly, FAIMS improves the dynamic range and detection limits in MS-based measurements by filtering out chemical noise.19 20While FAIMS has been shown to increase the signal-to-noise of analytes of interest through the application of a particular compensation voltage (CV) that preferentially transmits certain ions, there is typically a decrease in the ion signal of the analyte of interest, which may compromise the sensitivity in limited sample applications.20Nevertheless, it has been shown that FAIMS can be highly beneficial in bottom-up, middle-down, and top-down proteomics studies.19 21-24 The advantages of coupling FAIMS with chromatographic separations were demonstrated in the analysis of complex protein mixtures, showing up to 62% increase in protein identifications in top-down proteomics.21Another benefit of the FAIMS interface is its utilization as a gasphase prefractionation strategy upstream of MS with the incorporation of multiple CVs per analysis, which may help streamline and speed up proteomic protocols.21 24'26

[0010] SUMMARY

[0011] The present technology provides ultralow flow (ULF) separations of protein and peptide isoforms using porous layer open tubular liquid chromatography (PLOT LC) columns combined with mass spectrometry (LC / MS) in top-down proteomics analysis of a limited biological sample, such as a single cell or a small number of cells. The technology further provides use of a high-field asymmetric waveform ion mobility spectrometry (FAIMS) interface to remove contaminants and increase S / N ratios, as well as use of FAIMS as an online gasphase fractionation strategy to increase protein and proteoform identifications.

[0012] The technology provides novel, highly efficient polystyrene-divinylbenzene-based PLOT LC columns using temperature-induced polymerization that had an approximately 250 nm-thick layer of stationary phase attached to the capillary interior surface. The PLOT columns can optionally be alkyl-functionalized (e.g., containing C2 , to C30 groups, such as C18) with or without incorporation of additional chemical moieties, such as those commonly used in reverse-phase, ion exchange, or affinity LC The novel analytical PLOT columns as well as bead-packed trapping columns were used to build a robust and easy-to-use methodology for analysis of limited biological samples using only commercially available parts and instrumentation.

[0013] The present technology is the first in which ULF LC coupled with a FAIMS interface is applied to perform top-down proteomics analysis of limited samples of biological material or biomolecules. The FAIMS instrument interfaces nanoelectrospray (i.e., ESI conducted at nL / min flow rates) generated at the ESI emitter placed at the outlet of the PLOT LC column and the MS ion source. The optimized operating conditions of this interface enables online gas-phase proteoform fractionation to increase by about 2-fold the number of protein and proteoform species identifications compared to the same separation and analysis performed without FAIMS. The methodology shows an exceptional separation performance in top-down proteomic analysis of an aliquot of, for example, only -200 HeLa cells and resulted in a superior depth of proteomic profiling. The technology can be employed using limited biological samples such as microneedle biopsies, populations of rare cells, or single cells.

[0014] Compared to bottom-up proteomics (BUP), the field of top-down proteomics is still at an early stage and will benefit from the improvements and technical development in all areas of the proteomic workflow, including data acquisition and data analysis, provided by the present technology, including improvements in the liquid chromatography and its interfacing to MS. Further improvements include use of an Orbitrap Fusion Lumos for analysis of samples containing a low cell number, and the sample requiring long injection times and microscan averaging to collect sufficient ions for fragmentation and to increase S / N ratios.

[0015] The present technology can be further summarized in the following list of claimss.

[0016] 1 . A porous layer open tubular (PLOT) liquid chromatography column comprising a capillary tube, a porous polymer layer attached to an inner wall of the capillary tube, and an open lumen disposed along a central long axis of the tube, wherein the porous polymer layer has a thickness in an cross-sectional plane through the capillary tube, the cross-sectional plane being orthogonal to the central long axis, and wherein the porous polymer layer has an essentially uniform thickness which is constant within + / - about 200 nanometers or less, 150 nanometers or less, 100 nanomecters or less over a length of the column along the central long axis and lacks rugulose structures having a thickness greater than about 50%, greater than about 40%, greater than about 30%, greater than about 20%, or greater than about 10% of a mean thickness of the porous polymer layer.

[0017] 2. The PLOT column of feature 1 , wherein the column diameter is in the range from about 90 to about 360 micrometers and the porous polymer layer thickness is in the range from about 1 nanometer to about 500 nanometers.

[0018] 3. The PLOT column of feature 1 or 2, wherein any rugulose structures present in the porous polymer layer have a size less than about 5 times, or less than 3 times, or less than about 2 times, or less than about 1.5 times the mean layer thickness of the porous polymer layer.

[0019] 4. The PLOT column of any of the preceding features, wherein the column diameter is about 90 to about 360 micrometers, the lumen is about 1 to about 50 micrometers, and the average porous polymer layer thickness along the central long axis is about 100 to about 250 nanometers.

[0020] 5. The PLOT column of any of the preceding features, wherein the capillary tube comprises fused silica.

[0021] 6. The PLOT column of any of the preceding features, wherein the porous polymer layer comprises polystyrene-divinylbenzene polymer or a methacrylate polymer.

[0022] 7. The PLOT column of any of the preceding features, wherein the porous polymer layer comprises one or more hydrophobic moieties, such as C1 to C30 moieties, preferably C18 moieties.

[0023] 8. A method of fabricating a PLOT column, comprising

[0024] (a) providing a capillary tube and a liquid polymerization mixture;

[0025] (b) filling the capillary tube with the liquid; and (c) heating the filled capillary, whereby the one or more polymer precursors polymerize to form a porous polymer layer attached to an inner surface of the capillary; wherein the polymerization conditions in (c) are selected so that the formed porous polymer layer has an essentially uniform thickness which is constant within + / - about 200 nanometers or less, 150 nanometers or less, 100 nanometers or less over a length of the column along the central long axis and lacks rugulose structures having a thickness greater than about 50%, greater than about 40%, greater than about 30%, greater than about 20%, or greater than about 10% of a mean thickness of the porous polymer layer

[0026] 9. The method of feature 8, wherein the polymer precursors comprise 2,2'-azobis(2- methylpropionitrile) (Al BN), 1 -decanol, divinylbenzene (DVB), styrene, and 1 -octadecene and the porous polymer layer comprises polystyrene-divinylbenzene polymer, the polymer optionally further comprising C1 to C30 moieties, preferably C18 moieties.

[0027] 10. A PLOT column made by the method of any of features 8-9.

[0028] 11. A liquid chromatography-mass spectrometry (LC-MS) system, comprising

[0029] (i) the PLOT column of any of features 1-10;

[0030] (ii) a mass spectrometer; and

[0031] (iii) an electrospray ionization (ESI) interface coupled at its inlet to an output of the PLOT column and serving as ion source for the mass spectrometer.

[0032] 12. The LC-MS system of feature 11 , further comprising one or more of:

[0033] (iv) liquid chromatography equipment;

[0034] (v) autosampler equipment; and

[0035] (vi) one or more zero dead volume T-unions.

[0036] 13. The LC-MS system of feature 11 or 12, further comprising a high-field asymmetric waveform ion mobility spectrometry (FAIMS) interface.

[0037] 14. The LC-MS system of any of features 11 to 13, further comprising a microreactor trapping column as sample inlet prior to the PLOT column.

[0038] 15. A method of performing proteomics analysis, the method comprising

[0039] (a) providing the LC-MS system of any of features 11-14 and a protein- or peptide- containing sample for proteomics analysis;

[0040] (b) loading the sample onto the PLOT column of the LC-MS system;

[0041] (c) separating proteins and / or peptides using the PLOT column;

[0042] (d) ionizing the separated proteins and / or peptides and / or fragments thereof using the ESI interface and delivering the ionized proteins and / or peptides and / or fragments thereof to the mass spectrometer for separation by mass / charge ratio; and

[0043] (e) analyzing the separated ionized proteins and / or peptides and / or fragments thereof to identify individual proteins, peptides, and / or proteoforms. 16. The method of feature 15, whereby the sample comprises proteins and / or peptides derived from less than 1 cell, or from a single cell, or from up to 10 cells, or from up to 50 cells, or from up to 200 cells, or from up to about 500 cells.

[0044] 17. The method of feature 15, wherein the sample is from a biopsy, a bodily fluid, an extraca tissue homogenate, a cell homogenate, or a component removed from a cell, such as a cell nucleus, one or more mitochondria, or cytosol.

[0045] 18. The method of any of features 15 to 17, further comprising processing a biological sample to solubilize, release, or fragment proteins and / or peptides in a form suitable for LC- MS analysis.

[0046] 19. The method of any of features 15 to 18, wherein the LC-MS system comprises a FAIMS interface, and the FAIMS interface applies an electric field to separate ionized proteins, peptides, and / or fragments thereof based on ion mobility.

[0047] 20. The method of any of features 15-19, wherein proteoforms are detected comprising one or more of N-acetylation, N-acetylation with removal of N-terminal methionine, removal of N-terminal methionine, phosphorylation, N-glycosylation, O-glycosylation, or fragmentation.

[0048] 21. The method of any of features 15-20, wherein the type of proteomics analysis is selected from the group consisting of top-down proteomics, bottom-up proteomics, middle- up proteomics, middle-down proteomics, phosphoproteomics, and native form proteomics,

[0049] BRIEF DESCRIPTION OF DRAWINGS

[0050] Figs. 1A-1 B present a schematic overview of a ULF PLOT LC-MS experimental workflow. Fig. 1A shows schematics of plumbing and configuration of a ULF PLOT LC-MS system for sample loading and gradient delivery. Fig. 1B shows experimental workflows with and without the FAIMS Pro interface for top-down proteomic analysis of an aliquot of -200 injected cells from -2E4 processed HeLa cells.

[0051] Fig. 2 shows base peak chromatograms of six LC-MS replicates (no FAIMS). Six technical replicates of an aliquot of -200 injected cells from -2E4 processed HeLa cells show retention time and MS signal reproducibility. Maximum intensity levels are specified for individual analyses (on the right), while all base peak chromatograms are shown zoomed in to intensity levels of 2E6.

[0052] Fig. 3 shows base peak chromatograms acquired in the 27-35 min retention time window showing the background noise intensity levels in six no FAIMS LC-MS replicate analyses. Signal intensities were normalized to 4E5. The same analyses as in Fig. 2 are shown in the same order.

[0053] Fig. 4 shows extracted ion chromatograms (EICs) of ten representative histone proteins (where the corresponding ion species were extracted either all together or individually) in the analysis of an aliquot of -200 cells injected from 2E4 processed HeLa cells in no FAIMS experiments. The ten individual EICs were labeled according to the elution order of the detected histones, as shown in the top panel: 1 : m / z 603.7408, 2: m / z 730.9188, 3: m / z 716.0748, 4: m / z 612.7307, 5: m / z 625.6865, 6: m / z 621.3453, 7: m / z 643.3853, 8: m / z 782.7380, 9: m / z 699.5034, 10: m / z 629.0308. Qual browser automatic processing parameters: mass tolerance 10 ppm and mass precision set to four decimal points. *Peaks 4- 9 could not be assigned to a single protein accession number and single histone species based on data analysis by the TopPIC algorithm.

[0054] Figs. 5A and 5B show EIC of DDM ion species. EICs show the transmission of one singly charged DDM ion (m / z 1021.6161) through the FAIMS Pro interface at different CV values (ranging from -60 V to +30 V) and no FAIMS. The selected time range was from 80 to 110 minutes. In Fig. 5A signal intensities were normalized to 1.6E8. Fig. 5B shows individual intensity levels as specified.

[0055] Fig. 6 shows background noise intensity levels in the FAIMS and no FAIMS experiments. Each FAIMS experiment was done using one CV value (external CV stepping method) and is compared to the no FAIMS experiment. Each CV step has a different transmission efficiency of ions, resulting in different background signal intensity levels. Total ion current (TIC) traces of the background noise intensity levels are shown after the elution of sample peaks during the system equilibration.

[0056] Fig. 7 shows extracted ion chromatograms (EICs) and S / N ratios for representative four multiply charged protein ions in the analysis of an aliquot of -200 cells injected from -2E4 processed HeLa cells in the FAIMS (at different CV values from -60 V to +15 V) and no FAIMS experiments. Signal intensities were normalized to 6.69E5 (m / z 696.3791), 3.90E6 (m / z 779.6108), 1.31 E7 (m / z 723.7989), and 9.66E4 (m / z 654.8297). FAIMS experiments at CV +30 V are not shown for clarity reasons.

[0057] Figs. 8A-8F show top-down protein identification results for FAIMS and no FAIMS ULF PLOT LC-MS analyses of 200 cell-equivalent injections from 2E4 processed HeLa cells at 1% FDR. Figs. 8A-8B are bar charts showing average identification results for proteins and proteoforms, respectively. Error bars correspond to a standard deviation from 3 FAIMS (both, external and internal stepping methods) and 6 no FAIMS technical replicates. Figs. 8C-8D are Venn diagrams showing common and unique identifications for all nonredundant protein and proteoforms detected using different FAIMS CV methods (N = 3, the first technical replicate from each CV were used: 1x -45 V, 1x -30 V, and 1x -15 V) compared to no FAIMS (N = 3, the first three technical replicates were used). Fig. 8E is a Venn diagram illustrating redundancy and uniqueness in proteoform identifications between FAIMS experiments (N = 3 for each CV experiment: 3x -45 V, 3x -30 V, and 3x -15 V). Fig. 8F is a Venn diagram showing overlaps in proteoforms between three technical replicates (CV: -30 V). Figs. 9A-9D show top-down protein identification results for FAIMS and no FAIMS ULF PLOT LC-MS analyses of injections equivalent to 200 cells from 2E4 processed HeLa cells at 1 % FDR. Figs. 9A-9B are Venn diagrams showing common and unique identifications for nonredundant protein and proteoforms with FAIMS (N = 6, the first technical replicate from each CV were used: 1x -60 V, 1x -45 V, 1x -30 V, 1x -15 V, 1x 0 V, and 1x 15 V) compared to no FAIMS (N = 6) experiments. Figs. 9C-9D are Venn diagrams showing common and unique identifications for nonredundant protein and proteoforms with FAIMS (N = 6, the first two technical replicates from each CV were used: 2x -45 V, 2x -30 V, and 2x -15 V) compared to no FAIMS (N = 6) experiments.

[0058] Figs. 10A-10C show top-down proteoform identification results for FAIMS and no FAIMS ULF PLOT LC-MS analyses of aliquots of 200 injected cells from 2E4 processed HeLa cells at 1 % FDR. The Venn diagrams show overlaps in proteoforms between three technical replicates (Fig. 10A) CV: -15 V, (Fig. 10B) CV: -45 V, (Fig. 10C) no FAIMS (the first three technical replicates were used).

[0059] Figs. 11A-11 D explore the internal stepping methods for top-down proteomics. (Fig. 11 A) An example of the external and internal stepping method. (Fig. 11 B) BPCs at three FAIMS experiments (external stepping methods: -30 V, -15 V, and internal stepping method: -30 V & -15 V) and no FAIMS. Signal intensities were normalized to 1.4E7. (Fig. 11C) Venn diagram illustrating the overlap in proteoform identifications between FAIMS (external stepping methods: -30 V and -15 V; internal stepping method: -30 V & -15 V) and no FAIMS experiments. N = 3 for each type of experiment. (Fig. 11 D) Venn diagram shows the overlap in proteoforms in internal stepping method technical replicates (-30 V & -15 V, N = 3).

[0060] Fig. 12 shows base peak chromatograms of ULF PLOT LC-MS replicates (internal stepping method -30 V & -15 V). Three technical replicates of an aliquot of -200 cells injected from 2E4 processed HeLa cells show retention time and MS signal reproducibility. Signal intensities were normalized to 1.0E7.

[0061] Figs. 13-13E show the effect of the FAIMS gas flow settings at CV -30 V on protein identification results. (Fig. 13A) Schematic side view of the FAIMS interface and the nitrogen gas flow pathways through the FAIMS device at 4 L / min (modified) and 5 L / min (default). (Figs. 13B-13C) Bar charts show average identification results for proteins and proteoforms, respectively. Error bars correspond to a standard deviation from three technical replicates. (Figs. 13D-13E) Venn diagrams show common and unique identifications for nonredundant protein and proteoforms at 4 L / min compared to 5 L / min FAIMS gas flow settings.

[0062] Figs. 14A-14C show comparative assessment of the depth of proteomic profiling in FAIMS and no FAIMS experiments. (Fig. 14A) S-curve of protein copy numbers per cell from a previously reported deep proteomic study of the HeLa cell line from 2017 (ref), with proteins identified from raw files acquired in this study (872 proteins from 33 raw files) marked with a short vertical line. The data from the 2017 study were divided into 10 equally large subsections (deciles) and numbers describe the portion of proteins identified by the TDP ULF LC-MS platform mapped to each decile. (Fig. 14B) A boxplot shows the range of proteoform molecular mass distributions for FAIMS at different CVs and no FAIMS experiments (N = 3). The upper and lower hinges correspond to the 25th and 75th percentiles (the first and third quartiles). The upper whisker lengthens from the hinge to the largest value no further than the 1.5 x interquartile range (IQR, distance between the first and third quartiles). The lower whisker lengthens from the hinge to the smallest value at most 1 .5 x IQR of the hinge. The horizontal line is the median value. (Fig. 14C) A dotplot displays distributions of proteoform charge states identified in FAIMS and no FAIMS experiments. The ULF PLOT LC-MS system was operated in a positive mode, therefore only positive proteoform charges were detected. The dots show the mean of three replicates.

[0063] Fig. 15 shows molecular mass characteristics of identified proteoforms. A boxplot shows the proteoform molecular mass distribution using FAIMS at different CVs and no FAIMS. Three LC-MS technical replicates are shown for each type of experiment. The upper and lower hinges correspond to the 25th and 75th percentiles (the first and third quartiles). The upper whisker lengthens from the hinge to the largest value no further than the 1.5 x interquartile range (IQR, distance between the first and third quartiles). The lower whisker lengthens from the hinge to the smallest value at most, 1.5 x IQR of the hinge. The horizontal line is the median value.

[0064] Fig. 16 shows top-down protein identification results from analysis of aliquots of 200 injected cells from 2E4 processed HeLa cells at 1% FDR. The Venn diagram shows common and unique identifications for nonredundant proteoforms with FAIMS (CV +30 V, N = 3) compared to no FAIMS (N = 3) experiments.

[0065] Fig. 17 shows charge state characteristics of identified proteoforms. A dotplot displays distributions of proteoform charge states identified in FAIMS and no FAIMS experiments. The ULF PLOT LC-MS system was operated in a positive mode; therefore, only positive proteoform charges were detected. Three ULF PLOT LC-MS technical replicates are shown for each experiment type.

[0066] Fig. 18 shows retention time, molecular mass, and charge state characteristics of identified proteoforms. Proteoform mass and charge state distributions over retention time depending on the applied FAIMS CV value in comparison to no FAIMS experiment are shown. For clarity reasons, only the first technical replicate from each experiment was used to create the figure, although consistent results were observed across all replicates.

[0067] Fig. 19 shows retention time, hydrophobicity index, and charge state characteristics of identified proteoforms. GRAVY index and charge state of identified proteoforms over retention time depending on the used FAIMS CV value in comparison to no FAIMS experiment. For clarity reasons, only the first technical replicate from each experiment was used to create the figure, although similar results were observed across all replicates.

[0068] Figs. 20A-20D show PTM analysis of identified PrSMs from 33 raw files acquired in this study (blank runs excluded) in the ULF PLOT LC-MS analysis of an aliquot of -200 injected cells from 2E4 processed HeLa cells. (Fig. 20A) A bar chart shows assigned modifications in proteoform spectral matches. All raw files, excluding blank runs, were included in this summary. A custom list of expected molecular mass modifications was created and included in the TopPIC suite for the assignment of the mass shifts to selected modifications from the UniMod database. *Possible coordinated metal ions (Li+, Na+, Cu+, K+, Al3+, Fe3+, Ca2+, Ni2+, Ag+, Mg2+). (Fig. 20B) An exemplary MS / MS spectrum with mapped fragments, identification results, and fragmentation coverage indicating acetylated and phosphorylated parathymosin (P20962) identified in the no FAIMS experiment (Fig. 20C) A representative MS / MS spectrum of acetylated and dimethylated histone H4 (P62805) proteoform identified in the no FAIMS experiment, displaying the proteoform molecular mass, sequence coverage, and identified terminal b and y fragments. (Fig. 20D) A representative MS / MS spectrum of peptidyl-prolyl cis-trans isomerase A (P62937) identified in the no FAIMS experiment, displaying the glycoform molecular mass, sequence coverage, identified terminal b and y fragments, and oxonium ions. The asparagine residue highlighted in red localizes the N- glycosylation site on the protein backbone red.

[0069] Figs. 21A-21 B show a schematic representation of on-capillary TDP analysis of a few selected HeLa cells. (Fig. 21 A) Overview of the ULF PLOT LC-MS-based workflow. HeLa cells (6 cells) were manually injected into the microreactor with visualization and cell counting under the microscope. The microreactor was directly connected to a PLOT column-based platform (in the same way as the trap column, see Fig. 1A), and cells were lysed by a plug of lysis solution injected using an autosampler, eluted by a 30-min gradient, and introduced into the MS by nanoESI. (Fig. 21 B) S-curve of protein copy numbers per cell from a previously reported deep proteomic study of the HeLa cell line (ref Bekker-Jensen et al., 2017, Cell Systems 4, 587-599), with 44 proteins identified from raw files acquired in this study.

[0070] Fig. 22 shows cross-sectional images of six PLOT columns as described in Example 7. For each column, cross-sections were imaged at the indicated positions along the length of the column. The images were made using scanning electron microscopy (SEM).

[0071] Figs. 23A-23F show high magnification views of the PLOT columns shown in Fig. 22, with thickness of the porous polymer layer indicated at certain positions. The columns shown in Fig. 23A-23F correspond to Columns A-F of Fig. 22, respectively.

[0072] Fig. 24 shows the column to column and run to run consistency of separations performed using PLOT columns A1 , A2, and C as described in Fig. 22 and Example 7. Figs. 25A-25E show LC runs with PLOT columns according to the present technology performed in bottom-up (25A, 25B), fast bottom-up (25C), phosphoproteinomics (25D), and top-down (25E) modes.

[0073] DETAILED DESCRIPTION

[0074] High-efficiency porous layer open tubular (PLOT) nano-liquid chromatography (LC) columns, when operated at ultra-low flow (ULF) rates, significantly increase the sensitivity and depth of electrospray ionization-mass spectrometry (ESI-MS) based omics analysis of scarce biological and clinical samples. The present technology provides a novel methodology based on ULF PLOT columns and high-field asymmetric waveform ion mobility spectrometry (FAIMS) interfaced with an ultrasensitive mass spectrometry (MS)-based methodology for performing top-down proteomic analysis of a low number of mammalian cells.

[0075] In the examples described below, a model sample of -200 HeLa cells was analyzed. The developed robust and easy-to-operate methodology delivered remarkable reproducibility of retention times (RSD < 0.4%) and exceptional separation performance for intact proteins (-14 seconds peak full width at half maximum and peak capacity of over 100 for a 60-min active gradient). The compensation voltage applied on FAIMS electrodes ranged from -75 V to +30 V, and the protein identification results were compared to results obtained from control (no FAIMS) experiments. The power of the FAIMS interface in top-down proteomics analysis was demonstrated in that the signals of typical LC-MS contaminants (polysiloxanes and lysis buffer constituents) were mostly removed, while comparable protein signal intensities were detected compared to no FAIMS (control) experiments. Collectively, 454 proteins and 1 ,305 proteoforms were identified from the no FAIMS experiment compared to the FAIMS experiment that yielded 819 proteins and 2,645 proteoforms, representing an approximately 2-fold increase in identification results. In total, 872 proteins and 2,979 proteoforms were identified in all ULF LC-MS experiments (across both, FAIMS and no FAIMS experiments). Furthermore, altering the FAIMS nitrogen gas setting from 5 L / min to 4 L / min boosted the protein identification results by -14%. The subsequent proteoform analysis uncovered a high degree of post-translational modifications such as acetylation, methylation, phosphorylation, glycosylation, and other biologically relevant modifications. Notably, use of the present technology produced a superior depth of protein identification results for limited samples that is comparable to large-scale studies. The methodology described herein is the first example of coupling ULF LC with FAIMS in top-down proteomics analysis of limited samples, and makes possible single-cell top-down proteomics profiling.

[0076] Open-tubular (OT) and porous layer open-tubular (PLOT) nano-LC columns are specific types of nano-LC columns that are often overlooked due to the non-existence of commercially available, standardized, validated, and reproducible products, as well as previous reports mentioning sophisticated column preparation and operation and low column- to-column reproducibility.11-14Such columns are usually prepared in fused silica capillaries in an open-tubular format with a thin layer of stationary phase firmly attached (covalently or non- covalently) to the interior surface of the fused silica capillary. This column technology was originally developed and broadly implemented in the GC field. The thickness of the stationary phase can be anywhere from a monolayer of single molecule thickness (i.e., OT columns) to a porous layer having a thickness of a few pm (i.e., PLOT columns). The thickness of the stationary phase can affect the separation performance and column loading capacity. The OT configuration diminishes eddy diffusion and mass transfer in the stationary phase, resulting in excellent separation efficiency. High-efficiency ultra-narrow bore PLOT nano-LC columns hold promise for high-sensitivity and deep molecular profiling analysis of limited samples, including analysis of proteins in their intact state (i.e., top-down proteomics). PLOT / OT columns prepared in capillaries with an internal diameter <10 pm and operated at an ultra-low flow (ULF) rate (<20 nL / min) significantly increase the sensitivity of nanoESI-MS.13’15’16

[0077] ULF PLOT LC-FAIMS-MS System Attributes

[0078] The PLOT column-based setup developed by the inventors was coupled to an ultrasensitive MS via a FAIMS Pro interface, or directly to the MS without the FAIMS interface (“no FAIMS” experiment), using a nanoESI source to evaluate the high-sensitivity, robustness, and reproducibility of top-down proteomic profiling of a low number of mammalian cells (Fig 1A). The developed nanoLC methodology involved a short bead-packed trapping column employed for quick sample loading and relatively short 1-m-long PLOT column, in comparison to previously reported OT nLC columns, which allowed high separation efficiency of intact proteins using relatively short (<1 hr) gradients (Fig. 1A).

[0079] The trapping and separation columns, splitting capillaries, and other plumbing elements were connected via low dead-volume tee unions to minimize band broadening. Two splitting capillaries connected to two low dead-volume tee unions were used for a quick sample loading and delivery of an ultra-low LC flow while using only commercially available LC instrumentation. Trapping and analytical columns were prepared in-house. Any extra column effects significantly contribute to peak band broadening in nanoLC; therefore, it is important to minimize those effects in order to achieve a high separation efficiency. This is especially crucial in nanoLC operated at ULF (<50 nL / min), where any imperfections in plumbing and extra column effects can cause a significant decrease in separation performance. To minimize the extra column volume and to make the PLOT system more robust and reproducible, all capillary ends were polished, and only commercially available tee unions were used with the lowest possible through hole diameter and length as well as swept volume, in addition to a zero dead volume butt-to-butt connection between the outlet end of the PLOT column and the inlet of the nESI emitter.

[0080] Fig. 1 B describes the experimental workflow. Only ~2E4 cells were lysed directly in a glass insert of the autosampler vial in a miniaturized volume of 100 pL to minimize sample losses during sample processing and transfer steps; lower volumes also could be used. Then, 1 pL of the prepared sample was injected directly onto the trapping column, corresponding to approximately 200 HeLa cells. Any sample cleanup and fractionation steps were purposely omitted, in contrast with typical large-scale top-down proteomic experiments.24 32Streamlining the sample preparation in top-down proteomics analysis of limited samples and single cells can increase the recovery of proteins and minimize sample losses. The prepared sample was afterward analyzed using the developed top-down proteomics ULF PLOT LC-MS methodology described herein. Examples of base peak chromatograms corresponding to six consecutive LC-MS technical replicates are shown in Figs. 2 and 3. It is apparent that the top-down proteomics ULF PLOT LC-MS system provided high reproducibility of analysis, as the retention time and magnitude of MS signals were perfectly aligned for replicate injections. From these data, 10 peaks were extracted throughout the active portion of gradient elution that correspond to histone proteins and calculated the retention time, signal intensity, and S / N reproducibility values (Fig. 4, Table 1). The calculated RSDs show exceptionally high reproducibility of retention times (<0.401 %) and peak intensities (<0.003%) and acceptable reproducibility of S / N ratios.

[0081] Generally, the thickness of the stationary phase in OT columns influences the loading capacity (this applies for a majority of direct injections without the use of a precolumn), and any irregularities in the stationary phase’s layer contribute to the A-term of the van Deemter equation, thus decreasing the separation efficiency. The optimized protocol for PLOT column preparation described herein resulted in a -0.25 pm thickness of the stationary phase attached to the capillary wall. Fig. 23A shows the structure, homogeneity, thickness, and the quality of the gapless attachment of the stationary phase to the capillary wall. Despite the apparently mild unevenness, the SEM-imaged inner surface of the column is substantially smoother compared to other previous reports describing PS-DVB-based PLOT columns13’36’37while expectedly rougher and thicker compared to the reported OT columns without a porous layer38, where, in theory, only a non-porous single-molecule monolayer coating defines the stationary phase thickness.

[0082] Table 1. A custom list of modifications included in the TopPIC suite in the “advanced parameters” section and the “modification file for localization” column. A default MIScore threshold of 0.15 was selected.

[0083] Background Noise Reduction in Top-down Analysis of Limited Samples Using FAIMS Pro Interface.

[0084] It was shown previously that the FAIMS interface can be beneficial in filtering out “chemical noise” (based on selected CV values), thereby increasing S / N ratios without significantly decreasing the analyte intensity in conventional and ULF flow LC- and CE-MS applications.19’39The present inventors observed similar behavior in the top-down proteomics analysis of model mixtures and selected limited samples using ULF PLOT LC columns, as shown on the ion density maps in Fig. 1 D and Figs. 5-7. More specifically, several types of contaminant ions were observed that were significantly removed using the FAIMS interface: i) lysis buffer constituents that were not retained by the stationary phase and, therefore, did not interfere with the protein-derived LC-nESI MS signals, but can contaminate the MS ion optics and diminish the signal intensity over time; ii) polysiloxane ions (typical contaminants in MS) and other singly-charged ions, that interfere with the protein LC-nESI MS signal throughout the run; and iii) lysis buffer constituents (including DDM detergent ions), that are retained by the stationary phase and elute at a certain percentage of solvent B and therefore, interact, with the protein LC-nESI-MS signals (Fig. 1 D, Fig. S4). One of the main benefits of using FAIMS in the present technology can be illustrated in a “DDM detergent” example. FAIMS, to a large extent, removes the signal of DDM, the detergent that is considered MS- friendly in many BUP studies since it elutes at a higher concentration of organic solvent B (typically between 45%-80% ACN in 0.1% FA in BUP studies) and, therefore, does not significantly interfere with the peptide signals.35However, in top-down proteomics studies, where substantially higher organic content may be needed in the mobile phase to elute trapped proteins, it poses a challenge since it interferes with protein nESI MS detection. Fig. S4 shows how the DDM ions are removed at different CV values compared to the no FAIMS experiment. Consequently, the TIC background noise levels in FAIMS experiments were significantly decreased, as documented in Fig. S5 (from the average values of 2E6 for no FAIMS down to the range of 5E3-4E5 for FAIMS experiments) without significantly altering or decreasing the protein nESE MS1 signal (Fig. 7). From the subset of protein ions displayed in Fig. S6, it is apparent that the MS1 signal is not always increased with FAIMS. It is important to note that the ion transmission efficiency for a particular protein ion is highly dependent on the CV, and the optimal CV value for each ion may not have been tested if it is between two values selected for this work, or outside of the range (intervals of 15 V between -75 V and +30 V). In addition, Fig. 7 suggests that higher-charged ions are better transmitted at higher CV values. Comparison of Identification Results with FAIMS and no FAIMS Included in top-down proteomics ULF LC-MS Methodology.

[0085] In a typical top-down proteomics DDA analysis of complex biological samples, highly abundant proteins / proteoforms are expectedly more successfully detected in comparison to low abundance species. This is due to the stochastic nature of DDA, where precursor ions are selected in order from highest to lowest intensity levels. Given the capabilities of the FAIMS interface, it was expected that a different rate of transmission through the FAIMS interface could serve as an online gas-phase proteoform fractionation strategy without the need for time and labor extensive off-line fractionation in order to increase the depth of proteomic profiling.21’26Therefore, in initial experiments, FAIMS was first investigated with a single CV per LC-MS run (external CV stepping method). Based on previous publications in which CVs from -80 V to +30 V were applied for top-down analysis, CVs were examined from -75 V to +30 V in 15 V increments.21’25 26All FAIMS data were collected as triplicates, while the “no FAIMS” (control) experiment was done in six replicates. Figs. 2A-2B show the average numbers of identified proteins and proteoforms in a Gaussian-like distribution profile with the highest peak corresponding to CV of -30 V (for the FAIMS experiment). The experiment where CV of -75 V was used did not result in any proteins and proteoforms and, therefore, was excluded from further analysis. The analysis of the blank sample resulted in the detection of 16±2 proteins and 19±3 proteoforms, and 25±3 proteins and 34±2 proteoforms for FAIMS (CV -30 V) and no FAIMS experiment, respectively. The no FAIMS experiments resulted in an average of 319 proteins and 711 proteoforms, while the best CV (i.e., -30 V), yielded an average of 336 proteins and 737 proteoforms, representing a modest increase of 5% and 4% in identified proteins and proteoforms, respectively. It is expected that the modest increase in the FAIMS protein identification results is most likely caused by the limitations of the acquisition method (long or maxed-out injection times and microscan averaging). Long injection times are necessary using the current generation of the experimental MS equipment because of low sample input as only an aliquot of 200 cells was subjected to analysis in every run. However, similar modest increases in the FAIMS protein identification results vs. no FAIMS were reported for large-scale studies as well.21 24All of this, as a result, yields to an overall long total MS cycle time and missing MS2fragmentation of many precursor ions using the current instrumentation. The extracted average peak widths from six replicate no FAIMS runs at 10% of the peak height, and peak half height were 34 s and 15 s, respectively. The approximate MS cycle time in no FAIMS experiments was set to 6 s, resulting in a mismatch between elution time (peak width) and MS acquisition speed.

[0086] Due to the cycle time limitations of the MS data acquisition method (e.g., stochastic ion sampling) restricting the benefits that might be observed with FAIMS for a single analysis, the number of non-redundant identifications summed across multiple injections was investigated. In total, 454 proteins and 1 ,305 proteoforms were identified from the no FAIMS analyses (N = 6). The total identifications from the FAIMS analyses yielded 819 proteins and 2,645 proteoforms, representing a ~2-fold increase in the number of identifications compared to results with no FAIMS (27 raw files from external CV stepping, internal CV stepping, and FAIMS gas settings experiments, see below). From a total of 33 raw files (excluding blank sample runs, including all FAIMS and no FAIMS (control) experiments), 872 proteins and 2,979 proteoforms were collectively identified. Compared to other recently published top-down proteomics studies, where a significantly higher amount of starting material and larger protein amounts were subjected to analysis, a competitive or higher depth of proteomic profiling was obtained, considering the amount of utilized sample while processing only 2E4 HeLa cells and injected an aliquot equivalent to 200 cells, roughly corresponding to 100 ng of total protein analyzed in each LC-MS run.21 24'26

[0087] In the next steps of data analysis, the same number of ULF LC-MS injections was compared corresponding to FAIMS and no FAIMS experiments. The analysis of six replicate injections (3x FAIMS - the first technical replicate from each CV: -45 V, -30 V, -15 V, and the first three technical replicates from no FAIMS), resulted in a total of 628 proteins with 233 (37%) unique to FAIMS and only 59 (9%) unique to no FAIMS experiment (Fig. 8C). A similar trend was observed for detected proteoforms (Fig. 8D). In addition, 59% and 76% more unique proteins and proteoforms, respectively, were detected using FAIMS that were not detected in the no FAIMS experiments (Fig. 8C-8D). Fig. 9 shows two additional alternative combinations of Fig. 8C-8D, where in total 12 raw files (six with FAIMS and six without FAIMS) were compared. In every data analysis condition, the FAIMS experiment detected more proteins and proteoforms. These and previous results clearly describe the benefit of FAIMS and the power of PLOT columns in top-down proteomics ULF LC-MS analysis. To decrease the redundancy in proteoform identifications between CV values, these values should be spaced at least 10-15 V apart.19’26Fig. 2E illustrates the overlap between three subsequent CV values. The maximum overlap in ID results between any of the two CVs shown in Fig. 8E was ~7%, resulting in the majority of all proteoforms being unique to each CV. In addition, Fig. 8F and Fig. 10 describe the overlaps in proteoform identifications between three technical replicates. At least 38% of all proteoforms were detected between three replicates in all examined cases.

[0088] Investigating FAIMS with Internal CV Stepping Method.

[0089] As can be seen from Figs 28-8B, the majority of proteins and proteoforms were identified between -45 V and -15 V. One internal stepping method (-30 V and -15 V in one ULF LC-MS run) was compared it to two external stepping methods (-30 V in the first and -15 V in the second run) and a no FAIMS experiment. Fig. 11 describes the different data acquisition strategies for external and internal CV stepping methods. In the external stepping method, only one CV value is used throughout one LC-MS run, while in the internal stepping method, two or more CV values are used throughout one LC-MS run. Initially, base ion chromatograms (BICs) were examined using a normalized intensity scale, and in general, comparable peak intensities and high reproducibility of retention times were observed (Figs. 11 B and 12). Next, the numbers of proteins and proteoforms identified were examined. However, this experiment did not bring an additional increase in protein identification results, as comparable numbers of proteins and proteoforms were detected (Figs. 8A-8B and 11C), most likely caused by the limitations of the acquisition method utilizing relatively long ion accumulation times. Nevertheless, 8% to 17% of proteoforms were uniquely identified in each experiment (Fig. 11C). Finally, the overlaps in proteoforms were explored between three internal stepping method technical replicates. At least 43% of all proteoforms were detected in all technical replicates, demonstrating similar reproducibility of detected proteoforms to external stepping methods and no FAIMS experiments (Fig. 11 D).

[0090] Improving Identification Results in ULF LC-FAIMS MS by Optimizing FAIMS Gas Settings.

[0091] Typically, in an experiment employing the FAIMS Pro interface, the FAIMS “user” gas flow is set to 0 L / min while the FAIMS carrier gas flow is hard-coded to 5 L / min. These gas settings are most likely optimal for the majority of FAIMS Pro applications using relatively high flow rate in nanoflow and microflow chromatography applications (100 nL / min up to 25 pL / min). With the present PLOT column methodology, ULF chromatography was used (i.e., 25 nL / min), and it was expected that the FAIMS carrier gas flow of 5 L / min may be too high, which could cause instabilities of nanoESI and hampering the ion transmission efficiency into and through the FAIMS interface. To test this hypothesis, the FAIMS carrier gas flow settings was reset from the factory settings of 5 L / min to 4 L / min and compared to the control experiment (i.e., default FAIMS gas flow settings of 5 L / min), by adding back the 1 L / min gas flow using the FAIMS “user” gas flow setting (Fig. 13A). This small change in FAIMS gas flow settings resulted in a 14% boost in the average number of identified proteins and proteoforms (Figs. 13B-13C). The analysis of 6 ULF LC-MS replicates (3x FAIMS gas 4 L / min, 3x FAIMS gas 5 L / min, both CV -30 V), resulted in a total of 798 proteins with 109 (23%) unique proteins using the 4 L / min operation mode and only 49 (10%) identified proteins unique to the 5 L / min mode (Fig. 13D). A similar trend was observed for detected proteoforms, where in total, 1 ,238 proteoforms were detected across both experimental arms (Fig. 13E). In addition, 30% and 39% more unique proteins and proteoforms were detected using 4 L / min that were not detected in the 5 L / min experiments (Figs. 13D-13E).

[0092] In-depth Assessment of Proteomic Profiling across Different CVs in FAIMS and no FAIMS Experiments. Further detailed analysis of the acquired data revealed that the technique enabled the identification and characterization of proteins with the dynamic range over five orders of magnitude (2E2-6E7) in the sample of this scale, as estimated based on previously reported protein copy numbers per HeLa cell.44In Fig. 14A, protein identifications are mapped from top- down proteomics ULF LC-MS analysis to protein copy numbers per cell for -14,200 proteins identified and quantified in a study by Bekker-Jensen.44Next, the reference dataset was split into deciles and the percentage calculated of identified proteins mapped to the reference dataset. Most of the proteins (-78%) corresponded to the first and second decile in the reference dataset (corresponding to the top 20% most abundant proteins in HeLa cells). However, many of the proteins (-22%) that were detected were mapped to the remaining eight deciles of proteins with lower copy numbers per cell. These results suggest that ultrasensitive and trace analysis of intact proteins in complex samples is feasible using this methodology.

[0093] Using FAIMS, a correlation trend was found between more positive CVs and higher average molecular mass of identified proteoforms, ranging from -3 kDa at CV -60 V to -22 kDa at CV +30 V, which is in line with previous observations in other FAIMS top-down proteomics studies (Fig. 14B).21’24’25In addition, Fig. 15 shows the high reproducibility of detected proteoform molecular mass ranges in three technical replicates in each experimental condition. It is important to note that the way the sample is prepared and analyzed by LC-MS (e.g., sample fractionation included / excluded, type of stationary phase used, or RF lens settings applied) can also greatly influence the size of detected proteoforms. Fig. 14B and Fig. 15 also demonstrate the possibilities of the FAIMS Pro interface to better capture certain molecular masses based on the CV used (fractionation of proteoforms based on size), opening the door to more targeted analysis. Based on this observation, this methodology can be potentially used in various proteomic modes (e.g., bottom-up, peptidomic, middle-up, middledown, top-down, or native MS), greatly expanding its versatility. Besides, several larger proteins were uniquely identified in FAIMS using CV +30 V, that were not detected in the corresponding no FAIMS experiment (Fig. 14B, Fig. 15, Fig. 16). For example, interferon alpha-2 (P01563, 19.6 kDa) was detected in FAIMS CV +30 V experiment but was not detected in the no FAIMS experiment. Similarly, a larger proteoform (23.1 kDa) of Rho GDP- dissociation inhibitor 1 (P52565) was only detected in FAIMS CV +30 V experiment, while a smaller proteoform (6.4 kDa) was only detected in no FAIMS experiment. Fig. 16 shows that 1 ,033 proteoforms were detected in no FAIMS compared to only 28 proteoforms identified from FAIMS CV +30V. However, more than 71% (20 proteoforms) of those were unique to FAIMS CV +30 V that were not detected in the no FAIMS experiments.

[0094] The charge states were analyzed for identified proteoforms that were extracted from the TopPIC result file (Fig. 14C, Fig. 17). These charge state distributions include only precursors that met the DDA criteria and were selected for MS / MS scans; not all charge state distributions transferred through the FAIMS interface. The data show the following trend: lower charge states are more represented at more negative CV values, and higher charge states are more represented at more positive CV values. Interestingly, some of the tested experimental conditions, e.g., using CV -60 V, almost uniquely identified proteoforms with charge states below +10, while another condition, e.g., using CV -15 V detected almost a similar number of proteoforms in a broad range of charges (i.e., from <+10 to >+25). In addition, Fig. 17 confirms the reproducibility in identification results and the distribution of charge states between technical replicates for each experimental condition.

[0095] The following data analysis examines how the proteoform molecular mass, proteoform hydrophobicity (GRAVY index), and proteoform charge state change through gradient elution and at different CV values (Fig. 18, Fig. 19). In line with previous observations, the proteoform molecular masses and proteoform charge states were dependent on the CV used. Again, larger proteoforms and proteoforms with higher charge states were detected using more positive CV values and were also observed later during the gradient elution (Fig. 18). Similarly, the GRAVY index of each detected proteoform was calculated to demonstrate the effect of hydrophobicity on separation. As expected for RP-LC, the trend of greater retention for more hydrophobic proteoforms was observed, and FAIMS CV values did not appear to bias the results based on the proteoform hydrophobicity (Fig. 19).

[0096] PTM analysis of detected PrSMs.

[0097] The PTMs of identified proteoforms was examined. In conducted database searches using the TopPIC software, one unexpected mass shift was allowed, from -160 Da to +500 Da, in addition to four N-terminal forms of proteins (N-terminal methionine excision (NME), N- terminal acetylation after the initiator methionine is removed (NME_ACETYLATION), N- terminal methionine acetylation (M_ACETYLATION), and no modification (NONE)). A custom list of modifications was created and included in the TopPIC suite in the “advanced parameters” section and the “modification file for localization” column before starting to process the data (Table 2). The TopPIC search engine first looks for a maximum number of mass shifts and maximum mass shift in Daltons, both defined by the user, followed by assigning these mass shifts to the provided list of modifications based on the software algorithm. As a result of these database search settings, the identified proteoform may have a) no modifications; b) N-terminal modification(s) only; c) one unexpected modification in a sequence only; d) N-terminal modification(s), and one unexpected modification anywhere in its sequence of amino acid residues. Fig. 20A shows the bar chart of assigned modifications in all proteoform spectral matches (PrSMs). From all -19,500 PrSMs, N-terminal modifications were the most common, with 4,394 NME_ACETYLATIONs, 3,423 NMEs, and 783 M_ACETYLATIONs identified, and 249 other acetylation not included in NME_ACETYLATIONs and M_ACETYLATIONs identified. Also identified were a high degree of methylation (mono-, di-, tri-), dehydration, and phosphorylation. Interestingly, based on the detection of characteristic oxonium ions by manual examination of RAW files (e.g., Hex, HexNAc - 2H2O, HexNAc - H2O, HexNAc, (Hex)2, HexHexNAc) glycan moieties could be identified in the presented dataset.

[0098] Table 2. Analytical parameters extracted from 10 selected histone peaks within the effective portion of the gradient in the developed TDP ULF PLOT LC-MS system. retention time experiment / m / z 603.7408 730.9188 716.0748 612.7307 625.6865 621.3453 643.3853 782.7380 699.5034 629.0309 no FAIMS #l 41.2 43.6 44.5 45.0 53.2 55.4 65.3 71.0 71.8 90.4 no FAIMS #2 41.2 43.4 44.5 44.9 52.6 55.2 65.2 70.8 71.7 90.1 no FAIMS #3 41.1 43.5 44.5 44.9 52.7 55.2 65.1 70.7 71.6 90.3 no FAIMS #4 41.1 43.5 44.5 45.0 52.8 55.2 65.3 70.8 71.7 90.2 no FAIMS #5 41.2 43.4 44.5 44.9 52.8 55.3 65.2 70.8 71.7 90.4 no FAIMS #6 41.1 43.5 44.5 44.8 53.0 55.3 65.3 70.8 71.8 90.6 average 41.1 43.5 44.5 44.9 52.8 55.3 65.2 70.8 71.7 90.3 stdevpa 0.0 0.1 0.0 0.0 0.2 0.1 0.1 0.1 0.1 0.2

[0099] RSD 0.1 0.2 0.0 0.1 0.4 0.1 0.1 0.1 0.1 0.2 intensity experiment / m / z 603.7408 730.9188 716.0748 612.7307 625.6865 621.3453 643.3853 782.7380 699.5034 629.0309 no FAIMS #l 4.7E+04 6.6E+04 6.3E+04 1.7E+05 9.4E+04 3.8E+05 8.4E+05 5.9E+05 5.6E+05 7.7E+05 no FAIMS #2 4.8E+04 5.7E+04 5.8E+04 1.8E+05 6.6E+04 3.1E+05 6.5E+05 4.4E+05 4.8E+05 4.3E+05 no FAIMS #3 5.2E+04 6.6E+04 7.7E+04 2.0E+05 8.4E+04 3.5E+05 7.4E+05 4.8E+05 6.1E+05 4.0E+05 no FAIMS #4 5.0E+04 7.8E+04 7.3E+04 2.0E+05 1.1E+05 4.6E+05 8.5E+05 5.9E+05 6.0E+05 3.6E+05 no FAIMS #5 5.1E+04 8.9E+04 7.7E+04 2.1E+05 8.7E+04 5.3E+05 8.3E+05 8.0E+05 6.8E+05 7.3E+05 no FAIMS #6 5.3E+04 7.7E+04 9.2E+04 1.7E+05 7.4E+04 5.0E+05 9.0E+05 6.4E+05 4.9E+05 7.8E+05 average 5.0E+04 7.2E+04 7.3E+04 1.9E+05 8.6E+04 4.2E+05 8.0E+05 5.9E+05 5.7E+05 5.8E+05 stdevpa 2.0E+03 1.0E+04 1.1E+04 1.7E+04 1.5E+04 7.9E+04 8.3E+04 1.2E+05 6.9E+04 1.8E+05

[0100] RSD 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

[0101] S / N experiment / m / z 603.7408 730.9188 716.0748 612.7307 625.6865 621.3453 643.3853 782.7380 699.5034 629.0309 no FAIMS #l 389 426 98 1651 712 991 2863 14910 2396 2923 no FAIMS #2 385 253 277 1121 330 1863 2260 INF 1388 2049 no FAIMS #3 451 252 194 1742 446 1052 2652 1308 32291 824 no FAIMS #4 435 851 335 1661 793 1726 5388 INF 5486 1026 no FAIMS #5 345 260 282 1743 635 1868 9033 INF 2925 3558 no FAIMS #6 413 166 180 1135 419 1277 6869 2054 1931 3604 average 403 368 228 1509 556 1463 4844 6091 7736 2331 stdevpa 35 229 79 272 168 370 2495 5363 11058 1120

[0102] RSD 9 62 35 18 30 25 52 88 143 48 The following are some representative examples of post-translationally modified proteoforms identified in the present disclosure.

[0103] Parathymosin (P20962, -11.5 kDa) is a nuclear protein that may mediate immune function by blocking the effect of prothymosin alpha, which confers resistance to certain opportunistic infections. Several PTMs in the parathymosine sequence were reported in the literature, mostly consisting of acetylations and phosphorylations. Here it was observed that there is an N-terminal acetylation after the initiator methionine was removed and an O- phosphorylation is present on threonine T52 that was previously not experimentally confirmed in the reported literature for human samples. In the human UniProt database, the modification site was manually inferred based on the sequence similarity between the rat and human versions of the protein (Fig. 6B).45

[0104] Histone proteins were largely represented in the acquired dataset. Histones are basic proteins with highly abundant lysine and arginine residues in their primary structure, and therefore multiple histone-derived peptides with or without PTMs are not easily detectable by conventional bottom-up proteomic approaches, highlighting an opportunity for top-down proteomics to increase understanding of histone protein biology.46All five major histone families, H1 / H5 (known as the linker histones), H2A, H2B, H3, and H4 (known as the core histones) were detected, and in total 245 histone proteoforms were identified in this study. One example is histone H4 (P62805) with N-terminal acetylation after the initiator methionine was removed and dimethylation on arginine (R20) (Figs. 20A-20D). Highly informative MS / MS spectra resulted in high sequence coverage and unequivocal identification and localization of PTMs in many detected histone proteoforms.

[0105] EXAMPLES

[0106] Example 1. Materials and Reagents.

[0107] Fused-silica capillary tubing (10 pm internal diameter (ID), 360 pm outer diameter (OD)) with polyimide outer coating was purchased from Molex (Polymicro Technologies, Phoenix, AZ). Optima LC-MS grade water, acetonitrile (ACN), and guanidine hydrochloride (GuHCI) were purchased from Fisher Scientific (Pittsburgh, PA). Sodium hydroxide (NaOH), methanol (MeOH), ethanol (EtOH), n-dodecyl-B-D-maltoside (DDM), HPLC-MS grade formic acid (FA), 2-diphenyl-1-picrylhydrazyl (DPPH), N,N-dimethylformamide (DMF), 3- (trimethoxysilyl)propyl methacrylate (TMSPMA), 2,2'-azobis(2-methylpropionitrile) (Al BN), 1- decanol, divinylbenzene (DVB), styrene, 1 -octadecene, and were obtained from Sigma- Aldrich (St. Louis, MO). Ammonium bicarbonate (ABC) was purchased from Honeywell Fluka (Charlotte, NC). Phosphate-buffered saline (PBS), tris(2- carboxyethyl)phosphine (TCEP), and Gibco penicillin-streptomycin (P / S) were purchased from Thermo Fisher Scientific (Waltham, MA). Premium fetal bovine serum (FBS) was obtained from R&D Systems (Minneapolis, MN). F-12K medium (Kaighn's Modification of Ham's F-12 Medium) was purchased from ATCC (Manassas, VA). Ultra-high purity nitrogen (grade 5.0) was purchased from Middlesex Gases & Technologies (Everett, MA).

[0108] Example 2. Cell Culture and Sample Preparation.

[0109] The HeLa-S3 cell line (ATCC, Manassas, VA) was initially cultured in F-12K medium supplemented with 10% FBS, 100 I.U. / mL penicillin, and 100 pg / mL streptomycin and converted to grow in suspension mode following a previously published protocol.27The cell suspension was collected, centrifuged at 300 x ref for 5 min, washed three times with ice-cold 1X PBS, and resuspended / diluted in ice-cold 1X PBS to 400 cells / pL. Cell count and viability were determined by a two-chip disposable hemocytometer (Bulldog Bio, Portsmouth, NH) and trypan blue staining, respectively. 50 pL of cell suspension was transferred into a glass insert and mixed with 50 pL of lysis buffer (400 mM GuHCI, 20 mM TCEP HCI, 0.10% DDM, 50 mM ABC, pH = -7.2). The mixture was sonicated for 10 min and incubated for 20 min at room temperature. The resulting sample was centrifuged at 10k ref for 10 min. 1 pL of the sample (an aliquot of -200 cells in 200 mM GuHCI, 10 mM TCEP HCI, 0.05% DDM, and 25 mM ABC) was injected in every LC-MS analysis. For a blank sample, 1X PBS was mixed with lysis buffer.

[0110] Example 3. Preparation of PLOT Column.

[0111] Polystyrene-divinylbenzene-based PLOT columns with incorporated C18 moieties (referred to as PS-C18-DVB-based PLOT columns) were made using temperature-induced polymerization. The PLOT column preparation procedure was adapted from a previously published protocol with modifications.16Briefly, a 10 m-long capillary (10 pm ID, 360 pm OD) was pretreated with 1 M NaOH (1 h) at room temperature, LC-MS-grade water (30 min), and MeOH (30 min). The capillary was then dried with nitrogen. In the next step, a solution of 30% (v / v) TMSPMA and 0.5% (w / v, g / mL) DPPH in DMF was prepared, and the capillary was filled with the mixture. Both ends of the capillary were enclosed with a septum and placed in an oven for 6 h at 110°C. Next, the silanized capillary was washed with MeOH and dried with nitrogen. In the following step, the capillary was filled with the polymerization mixture consisting of 5 mg AIBN, 425 pL of EtOH, 425 pL of 1-decanol, 19 pL of styrene, 56 pL of 1- octadecene, and 75 pL of DVB, and placed in an oven at 64°C for 16 h. The resulting PLOT column was cleaved with a capillary cutting stone into shorter lengths and washed with ACN. The column was stored dry if not in use. The total length of the PLOT column used in LC-MS analyses was 100 cm. To minimize the extra-column volume and to make the PLOT system more robust and reproducible, both PLOT column ends were polished with a capillary polishing station (CPS-2, ESI Source Solutions, Woburn, MA), and only commercially available tee unions (C360QTPK2, VICI Valeo Instruments, Houston, TX) with the lowest possible through hole diameter were used.

[0112] Example 4. Nano-Liquid Chromatography Combined with Mass Spectrometry.

[0113] An Ultimate 3000 nanoLC system (Thermo Fisher Scientific) was used in nanoLC-MS experiments with several adjustments to the system’s plumbing and connections (see Fig. 1A). The trapping column packed with ReproSil-Pur, C4, 5 pm, 300 A (Dr. Maisch, Ammerbuch, Germany) in a 50 pm ID, 360 pm OD, 7 cm long capillary with a porous Kasil frit31) at the distal (outlet) capillary end was connected to a 10-port switching valve via the 1stmicrotee union and a NanoViper LC Column to MS Tubing (20 pm ID, Thermo Fisher Scientific) (Fig. 1A). A splitting capillary (10 pm ID, 10 cm) was connected to the same (1st) microtee union to generate a ULF rate and another NanoViper LC Column to MS Tubing (20 pm ID). During the sample loading, this fluidic line was closed with a plug via a connection through the switching valve and left open during the rest of the chromatographic run to enable fast loading at 300 nL / min flow rate and ULF elution at ~25 nL / min flow rate (set pump flow: 750 nL / min, splitting ratio: 1 :30) (Fig. 1A). The outlet end of the trapping column was connected to another microtee union (the 2nd), which was coupled to a 1 m long PS-C18-DVB- based PLOT column (10 pm ID). Another NanoViper LC Column to MS Tubing (20 pm ID, Thermo Fisher Scientific) was coupled to the 2ndmicrotee union and connected to a waste line during the sample loading, or otherwise closed by a plug during the rest of the chromatographic run. The end of the PLOT column was connected butt-to-butt to a distal- coated ESI emitter (FS360-10-5-D, NewObjective, Woburn, MA) via a Teflon sleeve union to achieve a zero dead volume connection. The ESI voltage was applied to the distal-coated ESI emitter through a Nanospray Flex Ion Source (Thermo Fisher Scientific).

[0114] The samples were loaded at 300 nL / min for 20 min using 1 % of solvent B (0.1 % FA in ACN) and desalted for an additional 10 min. Trapped proteins were eluted at 25 nL / min from the trapping and analytical column with a 60-min-long multistep gradient from 1% B to 50% B (1-10% B in 5 min, 10-33% B in 50 min, and 33-50% B in 5 min). The solvent composition was switched from 50% B to 95% B over 2 min and held constant at 95% B for 3 min. Finally, the mobile phase composition was changed from 95% B to 1 % B over 0.1 min and held constant at 1% B for 35 min to equilibrate the separation system. Solvent A consisted of 0.1% FA in water.

[0115] For in-capillary analysis of small populations of HeLa cells (i.e., six cells), trapping columns were used as microreactors. The only difference between the above-described and microreactor trap columns was in the bed length of the stationary phase. Microreactors were packed with the stationary phase only for ~1 cm of the total length of the 7 cm-long capillary, creating ample space at the inlet side of the microreactors, where cells or other biological material could be injected. The gradient elution was shortened to 30 min. Other conditions were kept the same as described above with LC-MS analyses using trapping columns.

[0116] NanoLC-MS / MS top-down experiments were performed using an Ultimate 3000 nanoLC system coupled via a Nanospray Flex Ion Source to an Orbitrap Fusion Lumos Tribrid MS operated in intact protein and standard pressure mode (all Thermo Fisher Scientific). LC injection and gradient programs and MS data acquisition parameters were controlled by Xcalibur software (v. 4.1.50, Thermo Fisher Scientific). LC-MS experiments were conducted with or without a FAIMS Pro interface (Thermo Fisher Scientific) placed between the ESI emitter and MS entrance to evaluate the changes in the proteomic profiling results. Corresponding experiments were marked in the figures and text as “FAIMS” and “no FAIMS” (i.e., control), respectively. The ion transfer tube (ITT) temperature was kept at 220°C, and the ESI voltage was set to 1 .5 kV.

[0117] The MS was set to acquire MS1spectra at 120,000 resolution (at m / z 200) with an automatic gain control (AGC) target value set at 1.0E6, a maximum injection time of 100 ms, a 600-2,000 m / z scan range, an RF Lens at 50%, a source fragmentation of 15 V and 4 microscans. Data were stored in profile data type. For MS / MS, the data-dependent acquisition (DDA) mode with a maximum of two dependent scans was used. Protein ions were isolated by a quadrupole mass filter using an isolation window of 3 m / z units, and fragmented using a fixed higher-energy collision dissociation (HCD) energy of 25%. MS / MS spectra were acquired in the Orbitrap mass analyzer at 60,000 resolution (at m / z 200) with an AGC target set at 1.0E6, a maximum injection time of 500 ms, a first mass at 150 m / z, and 4 microscans. Data were stored in centroid data type. Only ions with a charge state >5 were selected for MS / MS analysis. The intensity threshold of 2.0E4 was used. Dynamic exclusion was set to 90 sec after one MS / MS event. Three technical replicates were performed for each FAIMS experiment (PBS blank and cell sample analysis). For no FAIMS experiments, three and six technical replicates were performed for the analysis of PBS blank and cell samples, respectively.

[0118] Example 5. High-Field Asymmetric Waveform Ion Mobility Spectrometry (FAIMS).

[0119] High-field asymmetric-waveform ion mobility spectrometry (FAIMS) is used to separate ions by applying a high-voltage asymmetric waveform at radio frequency combined with a DC waveform applied between two electrodes. Only ions with specific mobility characteristics will pass through the FAIMS device. As a result, it can be used to filter out interfering or abundant ions.

[0120] The temperatures of the FAIMS inner electrode, outer electrode 1 , and outer electrode 2 were set to 100°C, and a Standard Resolution Mode was used. The minimum FAIMS carrier gas flow settings were “hard-code” modified in the software by Thermo Fisher Scientific representatives from the factory settings of 5 L / min (default) to 4 L / min (modified). This is a permanent change in software to FAIMS carrier gas flow settings in Xcalibur software v. 4.1.50, which became a tunable option in a later software version. Therefore, to perform the experiment with default FAIMS gas flow settings (i.e. , 5 L / min), an extra 1 L / min was added by changing the FAIMS “user” gas flow to 1 L / min. All experiments were done using a FAIMS gas flow setting of 5 L / min (i.e., 4 L / min + 1 L / min) unless otherwise specified. Both single CV per LC / MS run (i.e., external CV stepping method), and multiple CVs per LC / MS run (i.e., internal CV stepping method) were evaluated.

[0121] Example 6. Data Analysis.

[0122] The MS RAW data were converted to mzML format by MSConvert28and afterward deconvoluted using TopFD (v.1.5.2).29For deconvolution by TopFD, the following parameters were used. The maximum charge state was set to 30, and the maximum mass to 70 kDa. For MS1 scans and MS2 scans, the S / N ratio was set to 3 and 1 , respectively. The precursor window size of 3.0 m / z and m / z error of 0.02 were used. TopPIC (v.1.5.2), an open-source methodology, was used for protein identifications. UniProtKB / Swiss-Prot human database (Release 2021_04, containing 20,375 reviewed sequences) was used. Spectrum and proteoform level identifications were filtered using 1 % FDR cutoff settings. Mass error tolerance was set to 15 ppm, and proteoform spectrum match (PrSM) cluster error tolerance to 1.2 Da. One unexpected mass shift from -160 Da to +500 Da was allowed in the TopPIC methodology. No modification (NONE), N-terminal methionine excision (NME), N-terminal acetylation after the initiator methionine is removed (NME_ACETYLATION), and N-terminal methionine acetylation (M_ACETYLATION) were selected as N-terminal forms of proteins. Peak width values were extracted from the MS RAW files using RawTools (v.1.3.3), an open- source software.30Example 7. Comparison of PLOT Columns Fabricated by Different Methods.

[0123] PLOT columns made according to the present technology were compared in their structural and functional characteristics with PLOT columns made by earlier methods. Columns A and B were prepared using the protocol described in Example 3 on the same day by cutting one column into two pieces to test consistency along the length of a single fabricated column. Columns C and D) were prepared using the same protocol but on a different day than Columns A and B, to test batch to batch consistency. Column E was prepared using a previously published protocol (PS-DVB-C18, described in Li, S., Karger, B.L., Ivanov, A.R., et al., MCP 2015, 14, 1672-1683, PMID 25755294). Column F was prepared using a previously published protocol (PS-DVB, described in Yue, G., Karger, B.L., et al., Anal. Chem. 2007, 79, 938-946, PMID: 17263319, also described in US 20140033804A1 , which is hereby incorporated by reference in its entirety). The columns were sectioned at selected positions along their length, and images obtained using SEM (see Fig. 22). PLOT columns A-D prepared according to the new method showed porous polymeric layers of consistent, smooth appearance along their length and were consistent across their length and from batch to batch. However, Columns E and F showed a much thicker porous polymer layer with variable rugulose structures in the column lumen, consisting of a variety of different structures, often spheroidal or vesicle-like in shape and sometimes aggregated into various groupings.

[0124] Figs. 23A - 23F show close-up views of the same cross-sections as shown in Fig. 22, with measurements of porous polymer layer thickness as indicated. For Fig. 23A, Column a, the thickness of the smooth porous layer of the stationary phase was 149 nm (mean, N = 8). The iindicated thicker structures present in the porous layer of the stationary phase ranged from 196-237 nm (min to max thickness, N = 7). The ratio of minimum to mean thickness was 196nm / 149nm = 1.32, while the ratio of maximum to mean thickness was 237nm / 149nm = 1.59.

[0125] For Column B the mean thickness of the smooth porous layer of the stationary phase was 145 nm (mean, N = 9). Thicker areas ranged 222-253 nm. The ratio of minimum to mean thickness was 222nm / 145nm = 1.53, while the ratio of maximum to mean thickness was 253nm / 145nm = 1.74.

[0126] For Column C the mean thickness of the smooth porous layer of the stationary phase was 143 nm (mean, N = 8). Thicker areas ranged 176-323 nm. The ratio of minimum to mean thickness was 176nm / 143nm = 1.23, while the ratio of maximum to mean thickness was 323nm / 143nm = 2.26.

[0127] For Column D the mean thickness of the smooth porous layer of the stationary phase was 143 nm (mean, N = 9). Thicker areas ranged 176-225 nm. The ratio of minimum to mean thickness was 176nm / 143nm = 1.23, while the ratio of maximum to mean thickness was 225nm / 143nm = 1.57.

[0128] For Column E the mean thickness of the smooth porous layer of the stationary phase was 597 nm (mean, N = 7). Column E contained regulose structures in the lumen. These ranged in thickness from 702-6359 nm (N = 8). The ratio of minimum to mean thickness was 702nm / 597nm = 1.18, while the ratio of maximum to mean thickness was 6359nm / 597nm = 10.65.

[0129] For Column E the mean thickness of the smooth porous layer of the stationary phase was 597 nm (mean, N = 7). Column E contained regulose structures in the lumen. These ranged in thickness from 702-6359 nm (N = 8). The ratio of minimum to mean thickness was 702nm / 597nm = 1.18, while the ratio of maximum to mean thickness was 6359nm / 597nm = 10.65.

[0130] For Column F the mean thickness of the smooth porous layer of the stationary phase was 669 nm (mean, N = 7). Column F contained regulose structures in the lumen. These

[0131] T1 ranged in thickness from 716-4642 nm (N = 8). The ratio of minimum to mean thickness was 716nm / 669nm = 1.07, while the ratio of maximum to mean thickness was 4642nm / 669nm = 6.94.

[0132] The reproducibility of peptide separations by LC-MS was investigated for Columns A- F. LC-MS base peak chromatograms (BPCs) were run in bottom-up proteomics mode, using as sample 1 ng of HeLa cell protein digest. Column length was 1 m. Run conditions were the same as described in Li, S., Karger, B.L., Ivanov, A.R., et al., MCP 2015, 14, 1672-1683. The results are shown in Fig. 24 for Columns A1 (A), A2 (B), and C. The results showed high run- to-run and column-to-column reproducibility, with similar and stable column backpressure (backpressure was measured at 20 nL / min flow rate using 1% acetonitrile in 0.1 % formic acid.

[0133] The plots show the performance stability and reproducibility in numbers of peptide and protein identifications per LC-MS run using identical sample amounts and elution conditions for Columns A-C. In contrast, Columns E and F showed loss of resolution and increased backpressure after several runs (not shown),

[0134] Example 8. Use of PLOT Columns for Proteomics Analysis.

[0135] Figs. 25A-25E show LC runs with PLOT columns according to the present technology performed in bottom-up (25A, 25B), fast bottom-up (25C), phosphoproteinomics (25D), and top-down (25E) modes.

[0136] Fig. 25A shows a base peak chromatogram (BPC) from ULF PLOT LC-MS analysis of 0.2 ng Jurkat cell protein digest. Separation column was 2 m x 10 pm i.d. PS-C18-DVB PLOT column according to the present technology. The system was operated at 20 nL / min, using a 30 min gradient 1-20%B, split mode, DIA LC-MS acquisition mode.

[0137] Fig. 25B shows a BPC from ULF PLOT LC-MS analysis of 0.2 ng HeLa cell protein digest. The separation column was a 2 m x 10 pm i.d. PS-C18-DVB PLOT column according to the present technology. The system was operated at 20 nL / min, using a 30 min gradient 1- 20%B, split mode, DIA LC-MS acquisition mode.

[0138] Fig. 25C shows a fast bottom-up proteomic (BUP) analysis using a short gradient and a column prepared using the present technology. The BPC from ULF PLOT LC-MS analysis was performed using 0.2 ng HeLa cell protein digest. Separation column was 2 m x 10 pm i.d. PS-C18-DVB PLOT column according to the present technology. The system was operated at 20 nL / min, 10 min gradient 1-20%B, split mode. DIA LC-MS acquisition mode.

[0139] Fig. 25D shows LC-MS analysis of stable isotope-labeled (SIL) O-phosphopeptides using a ULF PLOT LC column prepared according to the present technology. The BPC shown was obtained from ULF PLOT LC-MS analysis of 250 SIL O-phosphopeptides. A total of 100 amol was injected on column (2 m x 10 pm i.d. PS-C18-DVB PLOT column). The system was operated at 20 nL / min, using a 30 min gradient 1-20%B, split mode. Fig. 25E shows an example of top-down proteomic analysis using a ULF PLOT LC column prepared according to the present technology. The BPC shown was obtained using LC-MS analysis of a 10 ng protein sample of yeast lysate. The separation column was 4 m x 10 pm i.d. PS-C18-DVB PLOT column, and the system was operated at 50 nL / min, using a 30 min gradient 1-40%B, direct injection mode using Vanquish Neo and DDA LC-MS acquisition mode.

[0140] As used herein, "consisting essentially of" allows the inclusion of materials or steps that do not materially affect the basic and novel characteristics of the claim. Any recitation herein of the term "comprising", particularly in a description of components of a composition or in a description of elements of a device, can be exchanged with "consisting essentially of' or "consisting of".

[0141] While the present invention has been described in conjunction with certain preferred embodiments, one of ordinary skill, after reading the foregoing specification, will be able to effect various changes, substitutions of equivalents, and other alterations to the compositions and methods set forth herein.

[0142] REFERENCES

[0143] (1) Nickerson, J. L.; Baghalabadi, V.; Rajendran, S.; Jakubec, P. J.; Said, H.; McMillen, T. S.; Dang, Z.; Doucette, A. A. Recent advances in top-down proteome sample processing ahead of MS analysis. Mass Spectrom Rev 2021. DOI: 10.1002 / mas.21706 From NLM Publisher.

[0144] (2) Smith, L. M.; Kelleher, N. L.; Consortium for Top Down, P. Proteoform: a single term describing protein complexity. Nat Methods 2013, 70 (3), 186-187. DOI: 10.1038 / nmeth.2369 From NLM Medline.

[0145] (3) Smith, L. M.; Kelleher, N. L. Proteoforms as the next proteomics currency. Science 2018, 359 (6380), 1106-1107. DOI: 10.1126 / science.aat1884 From NLM Medline.

[0146] (4) Johnson, K. R.; Gao, Y.; Gregus, M.; Ivanov, A. R. On-capillary Cell Lysis Enables Top- down Proteomic Analysis of Single Mammalian Cells by CE-MS / MS. Anal Chem 2022, 94 (41), 14358-14367. DOI: 10.1021 / acs.analchem.2c03045 From NLM Medline.

[0147] (5) Melby, J. A.; Brown, K. A.; Gregorich, Z. R.; Roberts, D. S.; Chapman, E. A.; Ehlers, L. E.; Gao, Z.; Larson, E. J.; Jin, Y.; Lopez, J. R.; et al. High sensitivity top-down proteomics captures single muscle cell heterogeneity in large proteoforms. Proc Natl Acad Sci U S A 2023, 720 (19), e2222081120. DOI: 10.1073 / pnas.2222081120 From NLM Medline.

[0148] (6) Donnelly, D. P.; Rawlins, C. M.; DeHart, C. J.; Fornelli, L.; Schachner, L. F.; Lin, Z.; Lippens, J. L.; Aluri, K. C.; Sarin, R.; Chen, B.; et al. Best practices and benchmarks for intact protein analysis for top-down mass spectrometry. Nat Methods 2019, 16 (7), 587-594. DOI: 10.1038 / S41592-019-0457-0 From NLM Medline.

[0149] (7) Yang, Z.; Shen, X.; Chen, D.; Sun, L. Toward a Universal Sample Preparation Method for Denaturing Top-Down Proteomics of Complex Proteomes. J Proteome Res 2020, 79 (8), 3315-3325. DOI: 10.1021 / acs.jproteome.0c00226 From NLM Medline.

[0150] (8) Zhou, M.; Uwugiaren, N.; Williams, S. M.; Moore, R. J.; Zhao, R.; Goodlett, D.; Dapic, I.; Pasa-Tolic, L.; Zhu, Y. Sensitive Top-Down Proteomics Analysis of a Low Number of Mammalian Cells Using a Nanodroplet Sample Processing Methodology. Anal Chem 2020, 92 (10), 7087-7095. DOI: 10.1021 / acs.analchem.0c00467 From NLM Medline.

[0151] (9) Melby, J. A.; Roberts, D. S.; Larson, E. J.; Brown, K. A.; Bayne, E. F.; Jin, S.; Ge, Y. Novel Strategies to Address the Challenges in Top-Down Proteomics. J Am Soc Mass Spectrom 2021, 32 (6), 1278-1294. DOI: 10.1021 / jasms.1c00099 From NLM Medline.

[0152] (10) Vaclavek, T.; Foret, F. Microfluidic device integrating single-cell extraction and electrical lysis for mass spectrometry detection of intracellular compounds. Electrophoresis 2023, 44 (1-2), 313-322. DOI: 10.1002 / elps.202100379 From NLM Medline.

[0153] (11) Lam, S. C.; Sanz Rodriguez, E.; Haddad, P. R.; Pauli, B. Recent advances in open tubular capillary liquid chromatography. Analyst 2019, 144 (11), 3464-3482. DOI: 10.1039 / c9an00329k.

[0154] (12) Wilson, S. R.; Olsen, C.; Lundanes, E. Nano liquid chromatography columns. Analyst 2019, 744 (24), 7090-7104. DOI: 10.1039 / c9an01473j From NLM PubMed-not-MEDLINE.

[0155] (13) Li, S.; Plouffe, B. D.; Belov, A. M.; Ray, S.; Wang, X.; Murthy, S. K.; Karger, B. L.; Ivanov, A. R. An Integrated Methodology for Isolation, Processing, and Mass Spectrometrybased Proteomic Profiling of Rare Cells in Whole Blood. Mol Cell Proteomics 2015, 14 (6), 1672-1683. DOI: 10.1074 / mcp.M114.045724.

[0156] (14) Burato, J. S. D.; Borsatto, J. V. B.; Lancas, F. M. Practical aspects of preparation, optimization, and evaluation of poly (styrene-divinylbenzene)porous-layer open tubular (PLOT) columns for capillary liquid chromatography. Taianta 2023, 253. DOI: ARTN 124106

[0157] 10.1016 / j. taianta.2022.124106.

[0158] (15) Vehus, T.; Roberg-Larsen, H.; Waaler, J.; Aslaksen, S.; Krauss, S.; Wilson, S. R.; Lundanes, E. Versatile, sensitive liquid chromatography mass spectrometry - Implementation of 10 mum OT columns suitable for small molecules, peptides and proteins. Sci Rep 2016, 6, 37507. DOI: 10.1038 / srep37507.

[0159] (16) Xiang, P.; Zhu, Y.; Yang, Y.; Zhao, Z.; Williams, S. M.; Moore, R. J.; Kelly, R. T.; Smith, R. D.; Liu, S. Picoflow Liquid Chromatography-Mass Spectrometry for Ultrasensitive Bottom- Up Proteomics Using 2-mum-i.d. Open Tubular Columns. Anal Chem 2020, 92 (7), 4711- 4715. DOI: 10.1021 / acs.analchem.9b05639.

[0160] (17) Xiang, P. L.; Yang, Y.; Chen, H.; Chen, A. P.; Liu, S. R. Liquid chromatography using <= 5 mu m open tubular columns. Trac-Trend Anal Chem 2021 , 142. DOI: 10.1016 / j.trac.2021.116321.

[0161] (18) Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Mass spectrometric characterization of a high-field asymmetric waveform ion mobility spectrometer. Rev. Sci. Instrum. 1998, 69 (12), 4094-4105, Article. DOI: 10.1063 / 1.1149255.

[0162] (19) Gregus, M.; Kostas, J. C.; Ray, S.; Abbatiello, S. E.; Ivanov, A. R. Improved Sensitivity of Ultralow Flow LC-MS-Based Proteomic Profiling of Limited Samples Using Monolithic Capillary Columns and FAIMS Technology. Anal Chem 2020, 92 (21), 14702-14712. DOI:

[0163] 10.1021 / acs.analchem.0c03262.

[0164] (20) Hebert, A. S.; Prasad, S.; Belford, M. W.; Bailey, D. J.; McAlister, G. C.; Abbatiello, S.

[0165] E.; Huguet, R.; Wouters, E. R.; Dunyach, J. J.; Brademan, D. R.; et al. Comprehensive Single-Shot Proteomics with FAIMS on a Hybrid Orbitrap Mass Spectrometer. Anal Chem 2018, 90 (15), 9529-9537. DOI: 10.1021 / acs.analchem.8b02233.

[0166] (21) Gerbasi, V. R.; Melani, R. D.; Abbatiello, S. E.; Belford, M. W.; Huguet, R.; McGee, J. P.; Dayhoff, D.; Thomas, P. M.; Kelleher, N. L. Deeper Protein Identification Using Field Asymmetric Ion Mobility Spectrometry in Top-Down Proteomics. Anal Chem 2021, 93 (16), 6323-6328. DOI: 10.1021 / acs.analchem.1c00402 From NLM Medline.

[0167] (22) Johnson, K. R.; Gregus, M.; Ivanov, A. R. Coupling High-Field Asymmetric Ion Mobility Spectrometry with Capillary Electrophoresis-Electrospray Ionization-Tandem Mass Spectrometry Improves Protein Identifications in Bottom-Up Proteomic Analysis of Low Nanogram Samples. J Proteome Res 2022, 21 (10), 2453-2461. DOI: 10.1021 / acs.jproteome.2c00337 From NLM Medline.

[0168] (23) Shliaha, P. V.; Gorshkov, V.; Kovalchuk, S. I.; Schwammle, V.; Baird, M. A.;

[0169] Shvartsburg, A. A.; Jensen, O. N. Middle-Down Proteomic Analyses with Ion Mobility Separations of Endogenous Isomeric Proteoforms. Anal Chem 2020, 92 (3), 2364-2368. DOI : 10.1021 / acs.analchem.9b05011.

[0170] (24) Xu, T.; Wang, Q.; Wang, Q.; Sun, L. Coupling High-Field Asymmetric Waveform Ion Mobility Spectrometry with Capillary Zone Electrophoresis-Tandem Mass Spectrometry for Top-Down Proteomics. Anal Chem 2023, 95 (25), 9497-9504. DOI: 10.1021 / acs.analchem.3c00551 From NLM Medline.

[0171] (25) Fulcher, J. M.; Makaju, A.; Moore, R. J.; Zhou, M.; Bennett, D. A.; De Jager, P. L.; Qian, W. J.; Pasa-Tolic, L.; Petyuk, V. A. Enhancing Top-Down Proteomics of Brain Tissue with FAI MS. J Proteome Res 2021, 20 (5), 2780-2795. DOI: 10.1021 / acs.jproteome.1c00049 From NLM Medline.

[0172] (26) Kaulich, P. T.; Cassidy, L.; Winkels, K.; Tholey, A. Improved Identification of Proteoforms in Top-Down Proteomics Using FAIMS with Internal CV Stepping. Anal Chem 2022, 94 (8), 3600-3607. DOI: 10.1021 / acs.analchem.1c05123 From NLM Medline.

[0173] (27) Abdeen, S. H.; Abdeen, A. M.; El-Enshasy, H. A.; El Shereef, A. A. HeLa-S3 Cell Growth Conditions in Serum-Free Medium and Adaptability for Proliferation in Suspension Culture. Journal of Biological Sciences 2011, 11 (2), 124-134. DOI: 10.3923 / jbs.2011.124.134.

[0174] (28) Chambers, M. C.; Maclean, B.; Burke, R.; Amodei, D.; Ruderman, D. L.; Neumann, S.; Gatto, L.; Fischer, B.; Pratt, B.; Egertson, J.; et al. A cross-methodology toolkit for mass spectrometry and proteomics. Nat Biotechnol 2012, 30 (10), 918-920. DOI: 10.1038 / nbt.2377 From NLM Medline.

[0175] (29) Kou, Q.; Xun, L.; Liu, X. TopPIC: a software tool for top-down mass spectrometry-based proteoform identification and characterization. Bioinformatics 2016, 32 (22), 3495-3497. DOI: 10.1093 / bioinformatics / btw398 From NLM Medline.

[0176] (30) Kovalchik, K. A.; Colborne, S.; Spencer, S. E.; Sorensen, P. H.; Chen, D. D. Y.; Morin, G. B.; Hughes, C. S. RawTools: Rapid and Dynamic Interrogation of Orbitrap Data Files for Mass Spectrometer System Management. J Proteome Res 2019, 18 (2), 700-708. DOI: 10.1021 / acs.jproteome.8b00721.

[0177] (31) Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D. J.; Inuganti, A.; Griss, J.; Mayer, G.; Eisenacher, M.; et al. The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res 2019, 47 (D1), D442-D450. DOI: 10.1093 / nar / gky1106.

[0178] (32) Drown, B. S.; Jooss, K.; Melani, R. D.; Lloyd-Jones, C.; Camarillo, J. M.; Kelleher, N. L. Mapping the Proteoform Landscape of Five Human Tissues. J Proteome Res 2022, 21 (5), 1299-1310. DOI: 10.1021 / acs.jproteome.2c00034 From NLM Medline.

[0179] (33) Petelski, A. A.; Emmott, E.; Leduc, A.; Huffman, R. G.; Specht, H.; Perlman, D. H.; Slavov, N. Multiplexed single-cell proteomics using SCoPE2. Nat Protoc 2021, 16 (12), 5398-5425. DOI: 10.1038 / s41596-021 -00616-z From NLM Medline.

[0180] (34) Kostas, J. C.; Gregus, M.; Schejbal, J.; Ray, S.; Ivanov, A. R. Simple and Efficient Microsolid-Phase Extraction Tip-Based Sample Preparation Workflow to Enable Sensitive Proteomic Profiling of Limited Samples (200 to 10,000 Cells). J Proteome Res 2021, 20 (3), 1676-1688. DOI: 10.1021 / acs.jproteome.0c00890.

[0181] (35) Liang, Y.; Acor, H.; McCown, M. A.; Nwosu, A. J.; Boekweg, H.; Axtell, N. B.; Truong, T.; Cong, Y.; Payne, S. H.; Kelly, R. T. Fully Automated Sample Processing and Analysis Workflow for Low-Input Proteome Profiling. Anal Chem 2021, 93 (3), 1658-1666. DOI: 10.1021 / acs.analchem.0c04240 From NLM Medline.

[0182] (36) Rogeberg, M.; Vehus, T.; Grutle, L.; Greibrokk, T.; Wilson, S. R.; Lundanes, E.

[0183] Separation optimization of long porous-layer open-tubular columns for nano-LC-MS of limited proteomic samples. J Sep Sci 2013, 36 (17), 2838-2847. DOI: 10.1002 / jssc.201300499 From NLM Medline.

[0184] (37) Nikolaeva, O. A.; Patrushev, Y. V.; Sidelnikov, V. N. Chromatographic properties PLOT multicapillary columns. J Chromatogr A 2017, 1488, 126-133. DOI:

[0185] 10.1016 / j. chroma.2017.01.087 From NLM Medline. (38) Yang, Y.; Liu, S. R. Non-porous thin dense layer coating: Key to achieving ultrahigh peak capacities using narrow open tubular columns. Taianta Open 2020, 1. DOI: ARTN 100003

[0186] 10.1016 / j.talo.2020.100003.

[0187] (39) Cong, Y.; Motamedchaboki, K.; Misal, S. A.; Liang, Y.; Guise, A. J.; Truong, T.; Huguet, R.; Plowey, E. D.; Zhu, Y.; Lopez-Ferrer, D.; et al. Ultrasensitive single-cell proteomics workflow identifies >1000 protein groups per mammalian cell. Chem Sci 2020, 12 (3), 1001- 1006. DOI: 10.1039 / d0sc03636f From NLM PubMed-not-MEDLINE.

[0188] (40) Johnson, K. R.; Gregus, M.; Kostas, J. C.; Ivanov, A. R. Capillary Electrophoresis Coupled to Electrospray Ionization Tandem Mass Spectrometry for Ultra-Sensitive Proteomic Analysis of Limited Samples. Anal Chem 2022, 94 (2), 704-713. DOI:

[0189] 10.1021 / acs.analchem.1 C02929.

[0190] (41) Specht, H.; Emmott, E.; Petelski, A. A.; Huffman, R. G.; Perlman, D. H.; Serra, M.; Kharchenko, P.; Koller, A.; Slavov, N. Single-cell proteomic and transcriptomic analysis of macrophage heterogeneity using SCoPE2. Genome Biol 2021 , 22 (1), 50. DOI:

[0191] 10.1186 / s 13059-021 -02267-5.

[0192] (42) McCool, E. N.; Xu, T.; Chen, W.; Beller, N. C.; Nolan, S. M.; Hummon, A. B.; Liu, X.;

[0193] Sun, L. Deep top-down proteomics revealed significant proteoform-level differences between metastatic and nonmetastatic colorectal cancer cells. Sci Adv 2022, 8 (51), eabq6348. DOI: 10.1126 / sciadv.abq6348 From NLM Medline.

[0194] (43) Melani, R. D.; Gerbasi, V. R.; Anderson, L. C.; Sikora, J. W.; Toby, T. K.; Hutton, J. E.; Butcher, D. S.; Negrao, F.; Seckier, H. S.; Srzentic, K.; et al. The Blood Proteoform Atlas: A reference map of proteoforms in human hematopoietic cells. Science 2022, 375 (6579), 411- 418. DOI: 10.1126 / science.aaz5284 From NLM Medline.

[0195] (44) Bekker-Jensen, D. B.; Kelstrup, C. D.; Batth, T. S.; Larsen, S. C.; Haldrup, C.; Bramsen, J. B.; Sorensen, K. D.; Hoyer, S.; Orntoft, T. F.; Andersen, C. L.; et al. An Optimized Shotgun Strategy for the Rapid Generation of Comprehensive Human Proteomes. Cell Syst 2017, 4 (6), 587-599 e584. DOI: 10.1016 / j.cels.2017.05.009 From NLM Medline.

[0196] (45) Lundby, A.; Secher, A.; Lage, K.; Nordsborg, N. B.; Dmytriyev, A.; Lundby, C.; Olsen, J.

[0197] V. Quantitative maps of protein phosphorylation sites across 14 different rat organs and tissues. Nat Commun 2012, 3, 876. DOI: 10.1038 / ncomms1871 From NLM Medline.

[0198] (46) Daled, S.; Willems, S.; Van Puyvelde, B.; Corveleyn, L.; Verheist, S.; De Clerck, L.;

[0199] Deforce, D.; Dhaenens, M. Histone Sample Preparation for Bottom-Up Mass Spectrometry:

[0200] A Roadmap to Informed Decisions. Proteomes 2021, 9 (2). DOI: 10.3390 / proteomes9020017 From NLM PubMed-not-MEDLINE.

[0201] (47) Jeong, K.; Babovic, M.; Gorshkov, V.; Kim, J.; Jensen, O. N.; Kohlbacher, O. FLASHIda enables intelligent data acquisition for top-down proteomics to boost proteoform identification counts. Nat Commun 2022, 73 (1), 4407. DOI: 10.1038 / s41467-022-31922-z From NLM Medline.

Claims

CLAIMSWhat is claimed is1. A porous layer open tubular (PLOT) liquid chromatography column comprising a capillary tube, a porous polymer layer attached to an inner wall of the capillary tube, and an open lumen disposed along a central long axis of the tube, wherein the porous polymer layer has a thickness in an cross-sectional plane through the capillary tube, the cross-sectional plane being orthogonal to the central long axis, and wherein the porous polymer layer has an essentially uniform thickness which is constant within + / - about 200 nanometers or less, 150 nanometers or less, 100 nanometers or less over a length of the column along the central long axis and lacks rugulose structures having a thickness greater than about 50%, greater than about 40%, greater than about 30%, greater than about 20%, or greater than about 10% of a mean thickness of the porous polymer layer.

2. The PLOT column of claim 1 , wherein the column diameter is in the range from about 90 to about 360 micrometers and the porous polymer layer thickness is in the range from about 1 nanometer to about 500 nanometers.

3. The PLOT column of claim 1, wherein any rugulose structures present in the porous polymer layer have a size less than about 5 times, or less than 3 times, or less than about 2 times, or less than about 1.5 times the mean layer thickness of the porous polymer layer.

4. The PLOT column of claim 1, wherein the column diameter is about 90 to about 360 micrometers, the lumen is about 1 to about 50 micrometers, and the average porous polymer layer thickness along the central long axis is about 100 to about 250 nanometers.

5. The PLOT column of claim 1 , wherein the capillary tube comprises fused silica.

6. The PLOT column of claim 1, wherein the porous polymer layer comprises polystyrene-divinylbenzene polymer or a methacrylate polymer.

7. The PLOT column of claim 1, wherein the porous polymer layer comprises one or more hydrophobic moieties, such as C1 to C30 moieties, preferably C18 moieties.

8. A method of fabricating a PLOT column, comprising(a) providing a capillary tube and a liquid polymerization mixture;(b) filling the capillary tube with the liquid; and(c) heating the filled capillary, whereby the one or more polymer precursors polymerize to form a porous polymer layer attached to an inner surface of the capillary; wherein the polymerization conditions in (c) are selected so that the formed porous polymer layer has an essentially uniform thickness which is constant within + / - about 200 nanometers or less, 150 nanometers or less, 100 nanometers or less over a length of the column along the central long axis and lacks rugulose structures having a thickness greater than about 50%, greater than about 40%, greater than about 30%, greater than about 20%, or greater than about 10% of a mean thickness of the porous polymer layer9. The method of claim 8, wherein the polymer precursors comprise 2,2'-azobis(2- methylpropionitrile) (Al BN), 1 -decanol, divinylbenzene (DVB), styrene, and 1 -octadecene and the porous polymer layer comprises polystyrene-divinylbenzene polymer, the polymer optionally further comprising C1 to C30 moieties, preferably C18 moieties.

10. A PLOT column made by the method of claim 8.

11. A liquid chromatography-mass spectrometry (LC-MS) system, comprising(i) the PLOT column of claim 1 ;(ii) a mass spectrometer; and(iii) an electrospray ionization (ESI) interface coupled at its inlet to an output of the PLOT column and serving as ion source for the mass spectrometer.

12. The LC-MS system of claim 11, further comprising one or more of:(iv) liquid chromatography equipment;(v) autosampler equipment; and(vi) one or more zero dead volume T-unions.

13. The LC-MS system of claims 11 , further comprising a high-field asymmetric waveform ion mobility spectrometry (FAIMS) interface.

14. The LC-MS system of claim 11, further comprising a microreactor trapping column as sample inlet prior to the PLOT column.

15. A method of performing proteomics analysis, the method comprising(a) providing the LC-MS system of claim 11 and a protein- or peptide-containing sample for proteomics analysis;(b) loading the sample onto the PLOT column of the LC-MS system;(c) separating proteins and / or peptides using the PLOT column;(d) ionizing the separated proteins and / or peptides and / or fragments thereof using the ESI interface and delivering the ionized proteins and / or peptides and / or fragments thereof to the mass spectrometer for separation by mass / charge ratio; and(e) analyzing the separated ionized proteins and / or peptides and / or fragments thereof to identify individual proteins, peptides, and / or proteoforms.

16. The method of claim 15, whereby the sample comprises proteins and / or peptides derived from less than 1 cell, or from a single cell, or from up to 10 cells, or from up to 50 cells, or from up to 200 cells, or from up to about 500 cells.

17. The method of claim 15, wherein the sample is from a biopsy, a bodily fluid, an extraca tissue homogenate, a cell homogenate, or a component removed from a cell, such as a cell nucleus, one or more mitochondria, or cytosol.

18. The method of claim 15, further comprising processing a biological sample to solubilize, release, or fragment proteins and / or peptides in a form suitable for LC-MS analysis.

19. The method of claim 15, wherein the LC-MS system comprises a FAIMS interface, and the FAIMS interface applies an electric field to separate ionized proteins, peptides, and / or fragments thereof based on ion mobility.

20. The method of claim 15, wherein proteoforms are detected comprising one or more of N-acetylation, N-acetylation with removal of N-terminal methionine, removal of N-terminal methionine, phosphorylation, N-glycosylation, O-glycosylation, or fragmentation.

21. The method of claim 15, wherein the type of proteomics analysis is selected from the group consisting of top-down proteomics, bottom-up proteomics, middle-up proteomics, middle-down proteomics, phosphoproteomics, and native form proteomics,

Images