Ultrasensitive Proteome Profiling for 100 Living Cells by Direct Cell

Article pubs.acs.org/ac

Ultrasensitive Proteome Profiling for 100 Living Cells by Direct Cell Injection, Online Digestion and Nano-LC-MS/MS Analysis Qi Chen, Guoquan Yan, Mingxia Gao, and Xiangmin Zhang* Collaborative Innovation Center of Chemistry for Life Sciences, Department of Chemistry and Institutes of Biomedical Sciences, Fudan University, Shanghai 200433, China S Supporting Information *

ABSTRACT: Single-cell proteome analysis has always been an exciting goal because it provides crucial information about cellular heterogeneity and dynamic change. Here we presented an integrated proteome analysis device (iPAD) for 100 living cells (iPAD-100) that might be suitable for single-cell analysis. Once cells were cultured, the iPAD-100 could be applied to inject 100 living cells, to transform the living cells into peptides, and to produce protein identification results with total automation. Due to the major obstacle for detection limit of mass spectrometry, we applied the iPAD-100 to analyze the proteome of 100 cells. In total, 813 proteins were identified in a DLD-cell proteome by three duplicate runs. Gene Ontology analysis revealed that proteins from different cellular compartments were well-represented, including membrane proteins. The iPAD-100 greatly simplified the sampling process, reduced sample loss, and prevented contamination. As a result, proteins whose copy numbers were lower than 1000 were identified from 100-cell samples with the iPAD-100, showing that a detection limit of 200 zmol was achieved. With increased sensitivity of mass spectrometry, the iPAD-100 may be able to reveal bountiful proteome information from a single cell in the near future.

M

other hand, circulating tumor cells (CTCs) are responsible for the spread of cancer throughout the body,11 and identifying the protein components has great significance for diagnosis and monitoring of cancer. However, CTCs are extremely rare in cancer patients’ blood, sometimes as low as 5 CTCs are available per milliliter blood.12 Several groups have attempted to set up methods to analyze the proteome of hundreds of cell samples. The main obstacles in analyzing a low number of cells are considerable sample losses during preparation and incomplete protein extraction, especially water-insoluble membrane proteins. To transform whole cells into MS-identifiable peptides usually involves multiple steps, including protein extraction/solubilization, protein purification, reduction of disulfide bonds, alkylation of free cysteine residues, and overnight digestion with trypsin. Therefore, sampling is a key step. Wang et al.13 lysed 500 to 5000 cells in a microtube with a surfactant (NP-40) and then removed the MS-unfriendly surfactant by acetone precipitation of the proteins. Maurer et al.14 used sodium dodecyl sulfate (SDS) for cell lysis, followed by two different methods to remove SDS, namely, “pseudoshotgun” (PSG) and filter-aided sample preparation (FASP). Their methods proved feasible to analyze proteins from 250 cells per injection, which was

ass spectrometry-based proteomics has been a fastdeveloping field in the past decade,1,2 and proves to be the method of choice for obtaining bountiful information about the protein components in a biological system. We have made various efforts to develop novel methods to proteome sample handling, thus improving proteome analysis results and target protein assays.3−5 Single-cell analysis is of great importance, for it will provide crucial information about cellular heterogeneity and dynamic change.6Sweedler’s group made great contributions toward single-cell metabolome analysis on the basis of capillary electrophoresis (CE)-mass spectrometry (MS) technology.7−9However, proteome analysis is quite different from metabolome analysis because it involves a different sampling process and MS detection mode. Therefore, single-cell proteome analysis is an ambitious goal that is largely unfulfilled because of a lack of miniaturization, integration, and detection sensitivity. Even though single-cell proteome analysis is a longterm goal beyond immediate reach, analyzing the proteome of hundreds of cells would be a big step forward and has significance of its own. Normally, there is sufficient biological sample for proteome analysis. However, there are occasions when the cell number is limited, making low-number-cell proteomics very important. In the biological body, there are highly diversified types of cells making up so-called “miniorgan”. For example, Waanders et al.10 analyzed the proteome of a single kidney glomeruli and a single pancreatic islet containing 2000−4000 cells. On the © XXXX American Chemical Society

Received: March 1, 2015 Accepted: June 10, 2015

A

DOI: 10.1021/acs.analchem.5b00808 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Scheme 1. Illustration of the iPAD-100 for Cell Digestiona

a

Position A (1−10 connected) was for cell injection and LC-MS analysis. Position B (1−2 connected) was for cell digestion and peptide trapping.

equivalent to 60 ng of protein. Liu et al.15 developed an integrated method for cell lysis, protein extraction, and digestion for phosphoproteome analysis of 100 000 cells. They combined phosphopeptide enrichment and MS-incompatible surfactant removal into one step. Further steps would be needed to remove the surfactant, if the complete proteome instead of phosphoproteome were to be analyzed. Ethier et al.16 developed a microfluidic proteomic reactor for a low number of cells on the basis of a small strong cation exchange column. Rapid extraction and enrichment of proteins directly from cells as well as chemical and enzymatic treatments of proteins were performed in 50 nL effective volume. Tian et al.17 also used it for protein relative quantification. In this paper, an integrated proteome analysis device (iPAD) for 100 living cells (iPAD-100) was elaborately established for 100 cell samples. Living cells were directly injected into the system, where cells’ proteins were digested, enriched, and analyzed by nano LC-MS. Once the cells were introduced into the system, all procedures were performed automatically. Therefore, iPAD-100 was able to avoid complicated sample treatment, lower sample loss, and prevent sample contamination, leading to a high sensitivity.

Longerpump, Baoding, China) that could accurately control the volume that the syringe drew. Between Port 3 and 10 was a 40 cm long, 100 μm id capillary loop, which was ca. 3.2 μL in volume. A 1 mm-long frit was fabricated using the sol−gel technology on the loop end at Port 3 as a filter to keep away cell debris. The protocol for making the frit was previously developed in our lab, using 5 μm silica gel as packing material.18 Between Port 4 and 7 was a monolithic C8 trap column made according to the protocol also developed in our lab with modifications.19,20 The trap capillary (75 μm id) was 10 cm long and had a 3 mm long C8 column bed on the end of Port 4. The C8 particle size was 5 μm (300 Å bore size). Port 5 was connected to 75 μm id capillary C18 column (15 cm, 5 μm, 100 Å) and mass spectrometry detector. The column was homemade according to the protocol described in the previous article.21 Port 6 was connected to the binary pump (nanoACQUITY Ultra Performance LC system, Waters, Milford, MA). Port 8 was an outlet to waste. Port 9 was connected to a loading pump from the Waters pump system. Cell Culture and Preparation. DLD-1 (Duke’s type C colorectal adenocarcinoma, Chinese Academy of Sciences, Shanghai, China) cells were cultured at 37 °C in DMEM medium supplemented with 10% fetal calf serum. Cells were harvested following treatment with trypsin and washed three times with PBS. Preparation of the Digestion Solution. For the 100-cell sample, cells were originally dispersed in PBS solution in a microtube at a concentration of around 100cells/μL. NH4HCO3, guanidine hydrochloride and EDTA were added to it to reach a final concentration of 50, 200, and 10 mM, respectively. Trypsin was added to reach a concentration of 5 ng/μL. The microtube was placed on an ice-bath. Cell counting was performed immediately by Accuri C6 flow cytometer (BD, Oxford, UK) to get the exact cellular concentration, which should be around 50 cells/μL. Integrated Sample Treatment with the iPAD-100. The whole sample treatment process consisted of four steps. At Step 1, the valve was at Position A (1−10 connected), and sample solution containing 100 cells was drawn into the loop. The detailed procedures were as follows. The syringe pump drew 0.25 μL of air into the needle. Next, the cell solution was vortexed for 10 s, and the needle was immediately placed in the solution. Then, the syringe pump drew a proportion of sample

■

EXPERIMENTAL SECTION Materials and Reagents. Ammonium bicarbonate (NH4HCO3), phosphate buffered saline (PBS), guanidine hydrochloride, and ethylenediaminetetraacetic acid disodium salt (EDTA) were purchased from Sigma (St. Louis, MO). Sequence-grade trypsin was purchased from Promega (Madison, WI). Acetonitrile (ACN, chromatographic grade) was purchased from Merck (Darmstadt, Germany). DMEM (Dulbecco’s Modified Eagle’s Medium) and fetal bovine serum were from Thermo Scientific (San Jose, CA). All capillaries used were from Idex (Oak Harbor, WA). Water used in the experiments was ultrapure water from Milli-Q50SP system (Millipore, Milford, MA). Construction of the iPAD-100. The iPAD-100 was built upon a two-position 10-port valve (Scheme 1) purchased from Vici-Valco (part no. C82H-1670EUDC, Houston, TX). Port 1 was connected to an 8 cm long, 250 μm id capillary, which served as a cell-solution inlet needle. The needle volume was ca. 4 μL. Port 2 was connected to a 10 μL syringe (Gaoge, Shanghai, China) with a syringe pump (part no. TS-2A, B


Article

Analytical Chemistry

the autosampler. Between Port 3 and 10 was a 40 cm long, 100 μm id capillary loop. The loop end at Port 3 had a 1 mm-long frit to filter out cell debris. The inner diameter of the loop was wide enough to prevent cell congestion. When the cell sample was completely transferred into the loop, it should contain around 2 μL sample, with 0.25 μL of air segments on both ends. Therefore, the loop volume of 3.2 μL was adequate. Between Port 4 and Port 7 was a C8 trap column for peptide concentration and desalting. The trap column was built upon a 10 cm, 75 μm id capillary, which was around 500 nL in volume. The length of the column bed was 3 mm, and this trap column could concentrate peptide mass of around 1ug. Because one cell had a protein mass of around 0.1 ng, the trap column was quite sufficient even for the 500-cell sample. During trapping, the flow direction was from Port 4 to Port 7, although during LCMS analysis, the flow direction would be reversed. Therefore, the 500 nL volume in the capillary would not be dead volume but could cause gradient delay. The dead volume at connections on the valve was acceptable, considering that the flow rate of nano LC was 300 nL/min. Analyzing Cell Samples with the iPAD-100. In a previous study,22 we found that when cells were immersed in trypsin solution, the cell membrane would break apart and cellular proteins would be digested into peptides in a single step. On the basis of this finding, we designed the iPAD-100 to digest the cells into peptides in the system. At the same time, precautions were taken to make sure that cells stay intact before being introduced into the system, because once the cells were broken, some proteins would stick to the inner wall of the vessel and the needle. In our digestion formula, there were two ingredients that might contribute to the breaking of cells, namely, trypsin and guanidine hydrochloride. We put our cellular solution on an ice-bath to lower the activity of trypsin. Guanidine hydrochloride was added to the solution to increase solubilities of hydrophobic proteins. Even though trypsin could stand up to 1 M guanidine hydrochloride,23 we reduced the concentration to 200 mM to slow down the process of cell breaking. We did the following experiment to show the effects of trypsin and guanidine on cell breaking. Before trypsin and guanidine hydrochloride were added to the cell solution, we counted the cell number per microliter with flow cytometry. Once trypsin and guanidine hydrochloride were added to the cell solution, we counted the cell number again. The cells were counted for the third time after 10 min. Flow cytometry results showed that the cell numbers did not change. Therefore, once trypsin and guanidine hydrochloride were added to cell solution, the cells would stay intact for at least 10 min. That would give us enough time to perform cell counting and cell solution injection. (Refer to Figure S-1 in Supporting Information for more experimental evidence.) Before injection, the cell solution was vortexed for 10 s to help cells disperse evenly in the solution. The cells in the solution could be completely transferred into the loop due to the following reasons. First, the cell solution contained 10 mM EDTA, which would be chelated to metal ions, and 5 ng/μL trypsin, which would truncate proteins inclined to stick to the wall. Therefore, the interaction between cells and the inner wall of needle or valve was greatly reduced. Second, the inner diameters of needle and loop were 250 and 100 μm, respectively, and the bore size of the valve was medium (250 μm). This pathway was wide enough to prevent cell congestion. In order to determine the optimal temperature and duration of the digestion process, we performed a series of parallel

solution containing 100 cells into the needle at a speed of 20 μL/min. After that, the syringe pump further drew air to transfer the sample into the loop, with 0.25 μL of air segments on both ends of sample solution. At Step 2, the valve was switched to Position B (1−2 connected). The sample loop was heated in a water bath at 50 °C for 1 h. At Step 3, keeping the valve at Position B, the loading pump at Port 9 delivered solvent (2% ACN, 0.1% methanoic acid) at 5 μL/min for 2 min to preconcentrate and desalt cellular peptides on the trap column. At Step 4, the valve was switched back to Position A, and the gradient was started to elute peptides gradually for LC-MS analysis. LC-MS Analysis. Nano-HPLC-MS experiments were performed using the Waters nanoflow system connected to a Q-Exactive mass spectrometer (Thermo Scientific, San Jose, CA) equipped with a nanoelectrospray ion source. Peptides were separated with a 60 min gradient from 5% acetonitrile to 40% acetonitrile at a flow rate of 300 nL/min. Full MS scans were acquired in the Orbitrap mass analyzer over the range m/z 350−1200 with a mass resolution of 70 000 (at m/z 200). The 15 most intense ions were fragmented and tandem mass spectra were acquired in the Orbitrap mass analyzer with a mass resolution of 17 500 at m/z 200. The dynamic exclusion time was set to 30 s. The ion selection threshold was 1.3 × 104, and the maximum allowed ion accumulation times were 10 ms for MS scans and 150 ms for MS/MS. Data Analysis. The raw data acquired were processed with MaxQuant software (version 1.3.0.5) according to the standard workflow. The data were searched against the Human UniProtKB/Swiss-Prot database (Release 2012_12_07, with 20 233 sequences). No fixed modification was applied, and oxidation of methionine was chosen as a variable modification. Initial precursor mass tolerance was 6 ppm and fragment mass deviation was 20 ppm. Up to two missed cleavages were allowed for trypsin digestion. Identifications were filtered by the 1% false discovery rate both on the peptide and protein level using a reverse database. Only peptides with a minimum of seven amino acids in length were considered for identification.

■

RESULTS AND DISCUSSION Fabrication of the iPAD-100. The protein content of 100 cells was a relatively low (10 ng), which was less than 1% of the usual starting material for proteome analysis. Therefore, we could not follow the routine procedures for microgram-grade protein samples, which usually involved multiple steps such as buffer exchange, centrifugation, or precipitation. Here we presented an integrated system iPAD-100, which showed the advantages of automation and high efficiency. The iPAD-100 could prevent sample contamination and reduce sample loss to a minimum level, therefore achieving extremely high sensitivity. The iPAD-100 was centered around a two-position 10-port valve. The valve could stand up to the high pressure of 600 bar. With respect to the cells that had a particle size around 10 μm, the bore of the valve had a medium size of 250 μm, which would allow cells to travel smoothly while not adding much dead volume to the system. Port 1 was connected to a cellsolution inlet needle. The needle was a wide pathway through which the cells could travel. Port 2 was connected to a 10 μL syringe together with a syringe pump, which could both draw and deliver solvent. If a commercial LC autosampler was available, the syringe and syringe pump could be replaced by C


Article


Figure 1. 100-cell samples were placed in sample loops and digested at different temperatures (37 or 50 °C) and for different durations (0.5, 1, 1.5, or 2 h), followed by LC-MS analysis on a LTQ-Orbitap XL mass spectrometer. (A) Cells were digested for 1 h at 37 and 50 °C. Protein number, peptide number, and total peptide intensity were normalized to that of 50 °C, showing a faster digesting speed at higher temperature. Then protein number (B), peptide number (C), and total peptide intensity (D) from 50 °C was recorded, reaching a plateau at approximately 1 h.

Table 1. Unique Proteins Numbers Identified from Different Numbers of Cells and the distributions according to cell compartment GO terms percentage of proteins according to different GO cellular compartment categories cell number

unique protein groups

average number of unique proteins

number of proteins common to all three replicate experiments

number of proteins identified at least once

cytoplasm

nucleus

plasma membrane

100

614 639 651 786 741 750 943 1042 1195

635

476

813

54%

57%

17%

759

552

985

54%

57%

17%

1060

758

1385

45%

54%

17%

250

500

the temperature of 50 °C and digestion duration of 1 h was chosen for all the subsequent experiments. For a typical offline method, desalted peptide solutions were transferred into a sample vial before injected into LC-MS. Suppose the peptide solution volume were 20 μL. Because 2 μL of residual was required to be left in the sample vial, each injection loaded 90% of the sample. In our system, 100% of the sample was to be utilized. All the peptides were trapped and ready for LC-MS analysis. Under extreme conditions when the sample was very limited, the number of identified peptides using direct sample injection would be several times higher than using conventional sample injection system with an autosampler.25 With our iPAD-100, direct sample injection was fulfilled. LC gradient needs to be optimized considering the nature of the sample. Briefly, if the gradient is too fast, many peptide ions

experiments using 100-cell samples. These samples were transferred to sample loops and digested at different temperatures (37 or 50 °C) and for different durations (0.5, 1, 1.5, or 2 h) and then analyzed on a LTQ-Orbitap XL mass spectrometer (Thermo Scientific, San Jose, CA). We used protein number, peptide number, and total peptide intensity as three indicators for digestion efficiency, as shown in Figure 1. Total peptide intensity was obtained by adding all peptide intensities using MaxQuant software.24 The digestion speed was much faster at 50 °C than at 37 °C. This might be due to faster breaking of cells and higher vitality of trypsin at a higher temperature. At 50 °C, protein number, peptide number, and total peptide intensity all reached a plateau at 1 h, showing that the cell digestion process was completed within 1 h. Therefore, D


Article


Figure 2. (A) Venn diagram of number of identified proteins from three replicate experiments. Each sample contained 100 cells. The identified proteins from three 100-cell samples were combined and categorized according to molecular weight (B), theoretical pI (C), and grand average of hydropathy (GRAVY) values (D).

There have been several reports on cellular protein copy number estimation using label-free approaches,26−29 and protein expression patterns were quite similar among cell lines from the same tissue origin.26 Now that we analyzed a colon-cancer cell line, the estimated protein copy number could be retrieved from a database of another colon-cancer cell line, due to the proteome similarities between cell lines from the same tissue. The number of identified proteins from 100-cell samples in our experiment was compared with a previous report by Wisniewski et al.29 as a function of the protein copy number per cell (pc/c), as shown in Figure 3. Wisniewski et al. quantified 7630 proteins from 100 000 cells. From our 100-cell samples, 814 proteins were identified, among which 801 proteins were included in the 7630-protein list. The figure showed that our method with the iPAD-100 could identify proteins over the complete range of less than 1000 to more than 10 000 000 pc/c, achieving a dynamic range of 6 orders. However, due to the extremely low number of cells, which was merely 100, our results were biased toward high-abundance proteins. Around 90% of our identified proteins had copy numbers higher than 100 000 pc/c. On the other hand, proteins less than 1000 pc/c were identified from 100-cell sample. The number of each protein was less than 100 000, showing that the detection limit of our system was less than 200 zmol. Cellular Component Distribution Analysis. In order to tell whether identified proteins from 100-cell samples were biased toward any category other than protein abundance, we also analyzed 250- and 500-cell samples. The protein identification results from the 100, 250, and 500 cells were summarized in Table 1. Gene Ontology analysis was performed using STRAP software30 to determine the cellular component and molecular

would coelute, leaving inadequate time for the acquirement of MS/MS spectra. If the gradient is too slow, due to the very limited sample, many ions may get diluted and not reach a minimum intensity to acquire a database-searchable MS/MS spectrum. Considering these potential disadvantages, we applied a 60 min gradient to samples containing 100 cells. At the same time, we set the ion selection threshold for MS/MS analysis to 1.3 × 104, 10 times lower than the usual value. Identification Results from 100-Cell Samples. The protein identification results are listed in Table 1. (Refer to Supporting Information for the detailed protein list of each cell sample.) These data were acquired from three experimental replicates, showing good reproducibility (Figure 2A). This was largely because once cells were introduced into the system, no extra human labor was needed, and all following procedures were performed by instruments with precision. Though we started with a very low number of cells, the numbers of identified proteins were quite high, showing a high sensitivity of the iPAD-100. The identified proteins of three duplicate experiments from 100-cell samples were combined and analyzed according to their molecular weight, theoretical pI, and grand average of hydropathy (GRAVY) values (Figure 2). Most proteins had a molecular weight in the small to middle range; however, proteins of large molecular weight (>250 kDa) could also be identified (Figure 2B). Because whole cells were digested and no proteins were lost in the sampling process, both very acidic (pI < 5.0) and very basic (pI > 11.0) proteins could be confidently identified (Figure 2C). Most identified proteins were hydrophilic, but very hydrophobic proteins were also identified (GRAVY > 0.4), partially due to the help of guanidine hydrochloride (Figure 2D). E


Article


even though we did not analyze the proteome of less than 100 cells, this system in principle could be applied to single-cell analysis. All that we need is further miniaturization of the system and a more sensitive MS detector.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional Information as noted in text. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b00808.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone/Fax: +86-21-6564 1740. Notes

The authors declare no competing financial interest.

■

Figure 3. Identified proteins as a function of protein copy number per cell demonstrated by Wisniewski et al.29 The reference (blue) exhibits data from a human cell line presented by Wisniewski et al.; the red bars are data from triplicate analysis of our 100-cell samples using the iPAD-100.

ACKNOWLEDGMENTS This work was supported by the National High-Tech R&D Program of China (Project: 2012AA020202), the National Basic Research Program of China (Project: 2012CB910604, 2013CB911201) and the National Natural Science Foundation of China (Project: 21175026).

function distributions of the proteins identified from different cell number samples. Proteins identified once or twice by three experimental replicates were also included. (Refer to Supporting Information for the detailed GO terms of each protein.) The percentage distributions of proteins for each investigated cellular component are listed in Table 1. It is worth noting that a substantial percentage of proteins belonged to plasma membrane (17%). No surfactant was used in our method to dissolve the membrane part of cells, but the percentage of plasma membrane proteins was similar to the result of another group’s work, using SDS.14 Our sampling method digested whole cells, and proteins in different cellular compartments were well-represented, including cytoplasm, endosome, chromosome, ribosome, ER, mitochondria, nucleus, peroxisome, cytoskeleton, plasma membrane, cell surface, extracellular, other intracellular organelles, macromolecular complex, and so forth. With respect to the molecular function, the identified proteins of different cell number samples predominantly belonged to binding, catalytic activity, and structural molecule activity. These categories normally do not offer much clinical information. Therefore, we still need to strive for deeper proteome coverage for a low number of cells.

■

REFERENCES

(1) Aebersold, R.; Mann, M. Nature 2003, 422 (6928), 198−207. (2) Nilsson, T.; Mann, M.; Aebersold, R.; Yates, J. R. Nat. Methods 2010, 7 (9), 681−685. (3) Li, Y.; Zhang, X. M.; Deng, C. H. Chem. Soc. Rev. 2013, 42 (21), 8517−8539. (4) Yao, G. P.; Deng, C. H.; Zhang, X. M.; Yang, P. Y. Angew. Chem. Int. Ed. 2010, 49 (44), 8185−8189. (5) Zhang, X. Y.; Zhu, S. C.; Xiong, Y.; Deng, C. H.; Zhang, X. M. Angew. Chem. Int. Ed. 2013, 52 (23), 6055−6058. (6) Wang, D. J.; Bodovitz, S. Trends Biotechnol. 2010, 28 (6), 281− 290. (7) Aerts, J. T.; Louis, K. R.; Crandall, S. R.; Govindaiah, G.; Cox, C. L.; Sweedler, J. V. Anal. Chem. 2014, 86 (6), 3203−3208. (8) Nemes, P.; Knolhoff, A. M.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2011, 83 (17), 6810−6817. (9) Nemes, P.; Rubakhin, S. S.; Aerts, J. T.; Sweedler, J. V. Nat. Protoc 2013, 8 (4), 783−799. (10) Waanders, L. F.; Chwalek, K.; Monetti, M.; Kumar, C.; Lammert, E.; Mann, M. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (45), 18902−18907. (11) Yoon, H. J.; Kim, T. H.; Zhang, Z.; Azizi, E.; Pham, T. M.; Paoletti, C.; Lin, J.; Ramnath, N.; Wicha, M. S.; Hayes, D. F.; Simeone, D. M.; Nagrath, S. Nat. Nanotechnol. 2013, 8 (10), 735−741. (12) Nagrath, S.; Sequist, L. V.; Maheswaran, S.; Bell, D. W.; Irimia, D.; Ulkus, L.; Smith, M. R.; Kwak, E. L.; Digumarthy, S.; Muzikansky, A.; Ryan, P.; Balis, U. J.; Tompkins, R. G.; Haber, D. A.; Toner, M. Nature 2007, 450 (7173), 1235−1239. (13) Wang, N.; Xu, M. G.; Wang, P.; Li, L. Anal. Chem. 2010, 82 (6), 2262−2271. (14) Maurer, M.; Muller, A. C.; Wagner, C.; Huber, M. L.; Rudashevskaya, E. L.; Wagner, S. N.; Bennett, K. L. J. Proteome Res. 2013, 12 (2), 1040−1048. (15) Liu, F. J.; Ye, M. L.; Pan, Y. B.; Zhang, Y.; Bian, Y. Y.; Sun, Z.; Zhu, J.; Cheng, K.; Zou, H. F. Anal. Chem. 2014, 86 (14), 6786−6791. (16) Ethier, M.; Hou, W. M.; Duewel, H. S.; Figeys, D. J. Proteome Res. 2006, 5 (10), 2754−2759. (17) Tian, R. J.; Wang, S. A.; Elisma, F.; Li, L.; Zhou, H.; Wang, L. S.; Figeys, D. Mol. Cell. Proteomics 2011, 10 (2), DOI: 10.1074/ mcp.M110.000679 −1. (18) Zhang, X. M.; Huang, S. J. Chromatogr. A 2001, 910 (1), 13−18.

■

CONCLUSION In this work, we established an integrated proteome analysis device for 100 living cells (iPAD-100) for digesting whole cells into MS-identifiable peptides in a single step with automation. We used our system to digest 100 cells. With 100 cells as starting material, an average of 635 proteins was identified. Good reproducibility was achieved, showing the stability and reliability of the whole system. GO analysis revealed that a substantial portion of those proteins were membrane proteins, even though no surfactant is used. Our system has great potential for application. One outstanding advantage is time-saving. Once the cells were acquired, proteome information would be revealed within 2 h (1 h for cell digestion and 1 h for LC-MS analysis). If deeper proteome coverage is desired, we can apply two-dimensional liquid chromatography for peptide separation and prolong the single RPLC gradient to 4 h or even longer. Last but not least, F


Article

Analytical Chemistry (19) Gu, X.; Wang, Y.; Zhang, X. M. J. Chromatogr. A 2005, 1072 (2), 223−232. (20) Guan, X.; Yan, G. Q.; Gao, M. X.; Hong, G. F.; Deng, C. H.; Zhang, X. M. J. Chromatogr. A 2010, 1217 (44), 6875−6881. (21) Yang, X. H.; Zhang, X. M.; Li, A. Z.; Zhu, S. Y.; Huang, Y. P. Electrophoresis 2003, 24 (9), 1451−1457. (22) Chen, Q.; Yan, G. Q.; Gao, M. X.; Zhang, X. M. Anal Bioanal. Chem. 2015, 407 (3), 1027−1032. (23) Jiang, X. G.; Jiang, X. N.; Feng, S.; Tian, R. J.; Ye, M. L.; Zou, H. F. J. Proteome Res. 2007, 6 (3), 1038−1047. (24) Cox, J.; Mann, M. Nat. Biotechnol. 2008, 26 (12), 1367−1372. (25) Masuda, T.; Sugiyama, N.; Tomita, M.; Ishihama, Y. Anal. Chem. 2011, 83 (20), 7698−7703. (26) Gholami, A. M.; Hahne, H.; Wu, Z. X.; Auer, F. J.; Meng, C.; Wilhelm, M.; Kuster, B. Cell Rep. 2013, 4 (3), 609−620. (27) Beck, M.; Schmidt, A.; Malmstroem, J.; Claassen, M.; Ori, A.; Szymborska, A.; Herzog, F.; Rinner, O.; Ellenberg, J.; Aebersold, R. Mol. Syst. Biol. 2011, 7, DOI: 10.1038/msb.2011.82 (28) Geiger, T.; Wehner, A.; Schaab, C.; Cox, J.; Mann, M. Mol. Cell. Proteomics 2012, 11 (3), DOI: 10.1074/mcp.M111.014050. (29) Wisniewski, J. R.; Ostasiewicz, P.; Dus, K.; Zielinska, D. F.; Gnad, F.; Mann, M. Mol. Syst. Biol.2012,8, DOI: 10.1038/ msb.2012.44. (30) Bhatia, V. N.; Perlman, D. H.; Costello, C. E.; McComb, M. E. Anal. Chem. 2009, 81 (23), 9819−982.

G


Ultrasensitive Proteome Profiling for 100 Living Cells by Direct Cell

Recommend Documents