Enhanced Sensitivity in Proteomics Experiments Using FAIMS Coupled with a Hybrid Linear Ion Trap/Orbitrap Mass Spectrometer† Julian Saba,‡ Eric Bonneil,‡ Christelle Pomie`s,‡ Kevin Eng,‡ and Pierre Thibault*,‡,§ Institute for Research in Immunology and Cancer, Universite´ de Montre´al, P.O. Box 6128, Station Centre-ville, Montre´al, Canada H3C 3J7, and Department of Chemistry, Universite´ de Montre´al, P.O. Box 6128, Station Centre-ville, Montre´al, Canada H3C 3J7 Received December 22, 2008
We describe the use and application of high-field asymmetric waveform ion mobility spectrometry combined with nanoscale liquid chromatography mass spectrometry (nanoLC-FAIMS-MS) to improve the sensitivity and dynamic range of proteomics analyses on a hybrid LTQ-Orbitrap mass spectrometer. The ability of FAIMS to enrich multiply protonated peptides against background ions confers a marked advantage in proteomics analyses by decreasing the limits of detection to facilitate the identification of low-abundance peptide ions. These multiply charged ions are recorded into separate acquisition channels to enhance the overall population of detectable peptide ions from a single analysis. NanoLCFAIMS-MS experiments performed on peptides spiked into complex proteins digests provided more than 10-fold improvement in limits of detection compared to conventional nanoelectrospray mass spectrometry. This enhancement of sensitivity is reflected by a 55% increase in the number of assigned MS/MS spectra contributing to an overall improvement in protein identification and sequence coverage. The application of FAIMS in label-free quantitative proteomics is demonstrated for the identification of differentially abundant proteins from human U937 monocytic cells exposed to phorbol ester. Keywords: FAIMS • Orbitrap • Quantitative proteomics • Phosphorylation • U937 human monocytic cells
Introduction Mass spectrometry (MS) has emerged as a key technology for large-scale expression and identification of proteins from a wide range of cellular extracts. Different MS-based proteomics platforms have been used with or without gel electrophoresis to simplify the inherent complexity of protein extracts. While gel electrophoresis provide a convenient protein separation format, the quest for approaches providing higher throughput and reproducibility for relatively small amounts of cell extracts prompted numerous research groups to adopt gel-free methods. Complex protein mixtures are typically digested in solution using trypsin and/or Lys-C enzymes, and the corresponding proteolytic peptides are separated using two-dimensional liquid chromatography (2D-LC) with strong cation-exchange and reversed-phase chromatography prior to MS analysis.1,2 In view of the inherent complexity of tryptic digests, protein identification often rely on serial selection of precursor ions for tandem mass spectrometry (MS/MS) to generate peptide fragment ions that can be correlated with those predicted from non redundant protein databases.1-4 The success of any MS/ MS experiments for protein identification relies heavily on the † This paper is dedicated to Roger Guevremont for his pioneering work on FAIMS. * To whom correspondence should be addressed. Pierre Thibault. Phone: (514) 343-6910. Fax: (514) 343-6843. E-mail:
[email protected]. ‡ Institute for Research in Immunology and Cancer, Universite´ de Montre´al. § Department of Chemistry, Universite´ de Montre´al.
10.1021/pr801106a CCC: $40.75
2009 American Chemical Society
ability to detect the precursor ion and the quality of the subsequent product ion spectrum.5,6 At present, reliable peptide sequencing can be performed on low-femtomole levels of protein digests. However, the sequencing of peptide ions beyond these levels becomes increasingly challenging, due to the presence of interfering ions often masking peptides of interest.7 The enhancement of MS/MS sensitivity for peptide sequencing thus relies on technology developments enabling the selection of target ions while minimizing the contribution of undesired precursor ions of similar m/z values. Several key development of ion selection based on gas-phase mobility enabled the improvement of sensitivity and selectivity for mass spectrometry analyses. One such development is the use of high-field asymmetric waveform ion mobility spectrometry (FAIMS) to reduce chemical noise8 and enhance peptide detection.9-11 FAIMS is a gas-phase ion separation technique based on the differences in ion mobility at high versus low (K) electric fields.12,13 In FAIMS, ions are introduced between pair of electrodes (planar or cylindrical) to which is applied a high voltage asymmetric waveform. A carrier gas (typically helium and/or nitrogen) is used to displace ions across the electrodes. Under these experimental conditions, an ion will travel different distances based on the mobility ratio Kh/K during the two phases of the waveform, ultimately leading to its collision on one of the electrode. To prevent this collision, a small DC voltage called the compensation voltage (CV) is applied to either of the electrodes. Experimentally, the difference in ion Journal of Proteome Research 2009, 8, 3355–3366 3355 Published on Web 05/26/2009
research articles mobility is reflected in the value of the CV needed for an ion to be transmitted through the FAIMS device. The FAIMS unit thus acts as a filtering device, transmitting only those ions with an appropriate Kh /K ratio. Since ion mobility is charge dependent, FAIMS can be used to separate multiply charged peptide ions including conformers having different gas-phase mobility. MS-based proteomics experiments typically pose significant analytical challenges owing to the high degree of complexity of cellular proteomes not only in terms of scale and dynamic range of protein expression, but also in the diversity and stoichiometry of protein modifications.14 The challenges currently facing proteomics far exceed the current capabilities of modern MS instruments. The development of new high performance instruments to meet analytical challenges associated with proteomics led to the recent introduction of new instruments such as the hybrid LTQ-Orbitrap that comprises a linear ion trap (LTQ) coupled with an Orbitrap mass analyzer via a C-shaped ion storage trap.15,16 The LTQ-Orbitrap provides a number of performance advantages including its high mass accuracy (106 ions), and high dynamic range (>2000).17,18 While the LTQ-Orbitrap can detect ions over a linear dynamic range extending 3-4 orders of magnitude in ion intensity, it is limited by the number of ions that can be analyzed simultaneously as for other ion trapping instruments. Furthermore, the capability of detecting trace-level analytes is further limited by the inherent chemical noise and background ions of nanoLC-MS experiments. In this context, we were interested in evaluating the potential of FAIMS to enhance the sensitivity of the LTQ-Orbitrap for proteomics analyses. In the present work, we evaluated the analytical merits of FAIMS device coupled with a LTQ-Orbitrap for the identification of trace-level peptides present in complex protein digests. To date, no study has examined the interfacing of LTQ-Orbitrap to the FAIMS device, and to our knowledge, this represents the first account of its analytical potentials for proteomics experiments. We also evaluated the use and application of sequential CV stepping to enhance the overall population of detected peptide ions from nanoLC-MS experiments. Applications of FAIMS to proteomics research are further demonstrated for tryptic digests of simple protein mixtures and for digests of protein extracts of differentiated monocyte cell lines to probe changes in expression profiles upon stimulation with phorbol ester.
Experimental Section Chemicals. Solvents for chromatographic analysis were all HPLC grade (Fisher Scientific and in-house Milli-Q water). TCEP (Tris[2-carboxyethyl] phosphine) was purchased from Pierce Biotechnology, Inc. (Rockford, IL). Modified porcine trypsin (sequencing grade) was purchased from Promega (Madison, WI). Glu-fibrinopeptide B, ammonium bicarbonate and urea were obtained from Sigma (St Louis, MO). Formic acid was obtained from EM Science (Mississauga, ON, Canada). Capillary HPLC columns for nano LC-MS were packed in-house using Jupiter C18 (3 µm) particles from Phenomenex, and fused silica tubing (Polymicro Technologies). Standard Protein Digest. All protein digests were purchased from Michrom Bioresources (Auburn, CA). Protein standard solutions were made by mixing protein digests of human lactotransferrin (LTF), Escherichia coli glycerokinase (GK), bovine serum albumin (BSA), bovine glutamate dehydrogenase 3356
Journal of Proteome Research • Vol. 8, No. 7, 2009
Saba et al. (GLUD), bovine lactoperoxidase (LPO), bovine catalase (CAT), rabbit aldolase (ALD) and yeast alcohol dehydrogenase (ADH) to a concentration of 40 fmol/µL. U937 Cell Culture, Protein Extraction, and Tryptic Digests. The human monocyte-like U937 hystiocytic lymphoma cells (ATCC, Manassas, VA) were cultivated in RPMI-1640 formulation supplemented with 10% fetal bovine serum (Hyclone, UT) and 1% Pen-Strep (Gibco-BRL, Grand Island, NY) at 37 °C in a 5% CO2 atmosphere. For stimulation experiments with phorbol ester, freshly washed cells were seeded in a 150cm2 dish at a density of 1 million cells/mL in 25 mL of culture medium. Concentrated phorbol 12-myristate 13-acetate (PMA) stock solution (Sigma) in dimethyl sulfoxide (Sigma) was added directly into the dish to a final concentration of 150 nM. Cells were left in the presence of the PMA for 1 h before protein extraction was carried out. Monocyte cells subjected or not to PMA were washed in phosphate buffered saline (Hyclone, UT) before being resuspended into isotonic lysis buffer supplemented with protease and phosphatase inhibitors (Sigma). Cells were placed in a metallic dounce homogenizer (Wheaton) and, after lysis, centrifuged at 13 500 rpm and 4 °C for 10 min to pellet unbroken cells and nucleus. The supernatant was collected and quantitated using micro BCA array (Pierce). Protein extracts (200 µg) were prepared in 50 mM ammonium bicarbonate, pH 8.1. Trypsin was added (enzyme/ protein 1/25), and samples were incubated overnight at 37 °C. TCEP was added to a final concentration of 10 mM, and the samples were left at room temperature for 1 h with gentle mixing. The samples were lyophilized after digestion and resolubilized in formic acid 0.2%/acetonitrile 5% prior to mass spectrometry analysis. FAIMS Interface and Integration of CV Stepping. The FAIMS interface was coupled with an LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific) via a nanoelectropray source. The FAIMS device used in these experiments is the commercial device available from Thermo Fisher Scientific (San Jose, CA). The outer electrode has an inner diameter of 18 mm and the inner electrode has a diameter of 13 mm. The gap between the inner and outer electrodes is 2.5 mm. The diameters of the entrance and exit apertures are 2.49 and 2.00 mm, respectively. The FAIMS electrodes were operated at a dispersion voltage (DV) of -5000 V with a 3.5 L/min flow of 50/50 helium/nitrogen (as specified by the manufacturer). The curtain voltage of the FAIMS interface was set to 1 kV and the electrospray voltage was adjusted to 3.6 kV (1.6 kV without FAIMS). The stable operating range of the desolvation flow rate is within (0.2 L/min of the optimal rate of 3.5 L/min. No detrimental effect on pumping efficiency of the instrument was noticed due to the additional gas load. The Ion Max source of the LTQ-Orbitrap was modified for nanoelectrospray ionization with an ADPC-IMS adapter from New Objective (Woburn, MA). The ion source was oriented at 45° angle relative to the inlet of the FAIMS electrodes at a distance of approximately 5 mm to obtain optimum spray. For infusion experiments, samples were introduced at a flow rate of 600 nL/min into the nanoelectrospray source using a Harvard Apparatus Model 1100 syringe pump (Holliston, MA). Flow rates were nominally set using the syringe pump and were checked for accuracy using a nanoflow meter from UpChurch Scientific (Oak Harbor, WA). The conditions for FAIMS-MS experiments were optimized by infusing GluFib at 100 fmol/µL under conventional nanoelectrospray-MS set up and recording signal intensity. This was
research articles
Enhanced Sensitivity in Proteomics Using FAIMS immediately followed by setting up the FAIMS apparatus and infusing the same concentration of GluFib. Under optimal conditions, the intensity of the doubly charged peptide ion at m/z 785.8 obtained with FAIMS should be within 20% of that observed with conventional nanoelectrospray.19,20 Optimal transmission of multiply charged peptide precursor ions from a tryptic digest of protein standards was determined by scanning the CV from -60 to 0 V at DV -5000, while sample is being directly infused. For optimal sensitivity, the temperature of the inner and outer electrodes was set to 70 and 90 °C, respectively. For CV stepping in nanoLC-FAIMS-MS experiments, the Xcalibur software allows changing the CV values in discrete increments from one scan to the next. Unless otherwise indicated, CV stepping experiments consist of stepping through 5 CV values at a rate of 1 s/CV step with 0.1 s interscan time while the corresponding data for each CV is acquired by the mass spectrometer. Mass Spectrometry. All MS analyses were performed using an LTQ-Orbitrap hybrid mass spectrometer with a nanoelectrospray ion source (ThermoFisher, San Jose, CA) coupled with an Eksigent nano-LC 2D pump (Dublin, CA) equipped with a Finnigan AS autosampler (Thermo Fisher, San Jose, CA). Protein digests were separated using a homemade 10 cm × 150 µm i.d. analytical column and a 4 mm length, 360 µm i.d. trap column packed with 3 µm C18 particles (Jupiter 300 Å, Phenomenex, Torrance, CA) and introduced into the mass spectrometer via 20 µm i.d. nanoelectrospray emitter. Tryptic digests (200-ng loading unless otherwise specified) were injected on the column, and elution was achieved using a linear gradient of 5-40% acetonitrile (0.2% formic acid) in 53 min and a flow rate of 600 nL/min. The conventional MS spectra (survey scan) were acquired in profile mode at a resolution of 60 000 at m/z 400. In all, MS/MS spectra were collected using data-dependent acquisition for multiply charged ions exceeding a threshold of 10 000 counts for nanoLC-MS and 4000 counts for nanoLCFAIMS-MS. These threshold settings correspond to twice the average intensity values for background ions observed in the densest region of the corresponding analyses. Database Searching. MS data were analyzed using the Xcalibur software (version 2.0 SR1). Peak lists were then generated using the Mascot distiller software (version 2.1.1, Matrix science) where MS processing was done using the LCQ_plus_zoom script. Database searches were performed using the search engine Mascot (version 2.1, Matrix Science, London, U.K.). Database searches were performed against a nonredundant IPI human database containing 66 921 entries21 (version 3.24, released May 2007) using Mascot (version 2.1, Matrix Science, London, U.K.) selecting human for U937 cell extract analysis. A Mascot search against a concatenated target/ decoy database consisting of a combined forward and reverse version of the IPI human database was performed to establish a cutoff score threshold of typically 30 for a false-positive rate of less than 2% (p < 0.02).22 The error window for experimental peptide mass values and fragment ion mass values were set to (0.02 and 0.5 Da, respectively. The number of allowed missed cleavage sites for trypsin was set to 1 and phosphorylation (STY), oxidation (M), deamidation (NQ) and carbamidomethylation (C) were all selected as variable modifications. No fixed modification was included in the search. Manual inspection of all MS/MS spectra for modified peptides was performed to validate assignments. Peptide Detection and Clustering. Raw data files (.raw) generated from the LTQ-Orbitrap acquisition software were
converted into text files representing all ions according to their corresponding m/z values, retention time, peak widths, intensityandchargestateusingin-housepeptidedetectionsoftware.23,24 Intensity values above a user-defined intensity threshold (10 000 or 4000 counts for non-FAIMS and FAIMS experiments, respectively) were considered for further analysis. Segmentation analyses were performed across different sample sets using hierarchical clustering with criteria based on their respective m/z, charge and time within user-defined tolerance ((0.01 m/z and (1 min). Normalization of retention time is then performed on the initial peptide cluster list using a dynamic and nonlinear correction. A moving-average time-window interpolation scheme is used to compute the time shifts for each peptide across the different data sets. For replicate LC-MS injections, this alignment confines the retention time distribution to less than (0.1 min (200 (Figure 2c) using FAIMS. Additional peptide ions were also visible on the same FAIMS-MS spectrum, although they remained undetected in nanoelectrospray (Figure 2b). Journal of Proteome Research • Vol. 8, No. 7, 2009 3359
research articles
Saba et al.
Figure 3. Enhanced peptide detection using nanoLC-FAIMS-MS. Reconstructed ion chromatogram of a tryptic digest of eight protein standards (200 fmol each) from (a) nanoLC-MS and (b) nanoLC-FAIMS-MS at CV values of -22 V, -28 V, -34 V -40 V and -46 V. Elution conditions as indicated in Experimental Section.
In general, the transmission of peptide ions with respect to CV voltage is distributed over 4-6 V under the present conditions. Thus, the use of several CV steps is required to enhance the peak capacity of FAIMS in nanoLC-MS experiments, as previously described by other groups.20,26 In the present study, enhanced coverage of multiply charged ions can be achieved by cycling through 5 CV steps of -22, -28, -34, -40, and -46 V while collecting mass spectra in different channels. We next evaluated the integration of CV stepping for the nanoLC-MS analysis of a tryptic digest from 8 protein standards and compared the corresponding ion profiles with that of nanoelectrospray. The total ion chromatogram (TIC) of a tryptic digest (200 fmol each inj.) under conventional nanoelectrospray is presented in Figure 3a together with the number of unique ions for each charge state. The same analysis performed using FAIMS is shown in Figure 3b and highlights distinct ion profiles enabling the identification of unique ions across all CV steps. A comparison of the ion distribution for these two analyses revealed a 2-fold reduction in the number singly charged ions when using FAIMS. More importantly, we noted an increase of approximately 60% in the number of multiply charged ions with larger gains observed primarily for doubly- and triply charged ions. 3360
Journal of Proteome Research • Vol. 8, No. 7, 2009
The selectivity of FAIMS provides an inherent advantage in MS/MS experiments by enriching the proportion of multiply charged peptide ions while minimizing singly charged interference ions arising from column bleed, siloxane adducts or mobile phase ion clusters. Furthermore, the reduction of interfering ions also reduces the presence of confounding ions in MS/MS spectra when precursors of similar m/z values are fragmented together. This was evidenced when comparing the number of unique peptides identified in separate nanoLC-MS/ MS experiments of this simple protein digest where we noted an increase of 35% in peptide assignment using FAIMS (data not shown). Furthermore, we observed that common peptides (211 unique sequences) showed equal or higher Mascot scores in FAIMS compared to non-FAIMS experiments. Enhancement of Protein Identification and Application to Protein Expression Analyses Using NanoLC-FAIMS-MS. To evaluate the reproducibility of CV stepping in the larger context of protein identification, we compared the intensities of peptide ions in nanoLC-FAIMS-MS experiments to the corresponding analyses obtained using fix CV values. Replicate nanoLCFAIMS-MS analyses (n ) 3) were conducted on complex protein extract from human monoblastic U937 cells. A total of 6 CV values (-22, -26, -30, -34, -38, and -42 V) were selected
Enhanced Sensitivity in Proteomics Using FAIMS
Figure 4. Comparison of detected and identified peptides from human monocytes protein extract using nanoLC-MS/MS with and without FAIMS (200 ng inj.). (a) Distribution of peptide population as a function of ion intensity for a complex protein digest of human monoblastic U937 cells using nanoLC-MS/MS and nanoLCFAIMS MS/MS (CV stepping). (b) Distribution of identified peptides as a function of precursor ion intensities with and without FAIMS. Elution conditions as indicated in Experimental Section.
based on a preliminary evaluation of the ion population of this cell extract. A scatter plot of ion intensities on replicate nanoLCFAIMS-MS runs (n ) 3) indicated that 95% of reproducibly detected peptide ions showed RSD values lower than 35%, suggesting that CV stepping is reproducible over 3 orders of magnitude (Supplementary Figure 2a). We then compared intensity values obtained using CV stepping to those acquired using fix CV conditions. A good correlation of ion intensities was obtained between these experiments with RSD values within 45% for 95% of the ion population (Supplementary Figure 2b). The reproducibility and sensitivity in ion intensities obtained here confers the advantage of identifying CV values for optimal transmission of specific ions for subsequent analyses or for targeted identification. For example, peptide maps from different cell extracts can be generated for each CV value to identify differentially abundant proteins. An inclusion list of peptide ions can then be generated for subsequent nanoLC-FAIMS-MS/MS for each target CV value. The use of a complex protein extract from human monoblastic U937 cells also provides an opportunity to evaluate the overall improvement in protein identification for proteomics application when using FAIMS. We first determined the distribution of peptide ions according to ion intensity in nanoLC-MS experiment with and without FAIMS (Figure 4). A total of 12 236 peptide ions were detected in FAIMS compared to 4926 without FAIMS (Figure 4a). The analysis conducted with FAIMS resulted in a higher proportion of low-intensity ions that remained undetected using conventional nanoelectrospray. Under conventional nanoLC-MS conditions, the intensity threshold is typically set to 10 000 counts for MS/MS experiments to minimize background noise contribution. Below this
research articles intensity level, peptide sequencing is difficult and can yield ambiguous protein identification. However, in nanoLC-FAIMSMS/MS analysis, the intensity threshold for MS/MS sequencing can be lowered to 4000 counts. Successful peptide sequencing from MS/MS data relies on both the ability to detect precursor ions and the quality of the corresponding product ion spectra. The selective transmission of multiply protonated peptide ions and the minimization of undesired background ions by FAIMS can result in better quality MS/MS spectra especially for lowabundance peptides.27 This is illustrated in Figure 4b that compares the distribution of identified peptides along with the intensity of the precursor ions with and without FAIMS. NanoLC-FAIMS-MS/MS experiments enabled the sequencing of 3034 unique peptide ions, whereas a total of 1958 unique peptide ions were identified when the corresponding analyses were performed without FAIMS. Larger gains in identification were observed for low-abundance peptide ions where FAIMS enabled the assignment of 485 peptide ions below intensity levels of 10 000 counts compared to less than 20 without FAIMS. Several examples of the enhanced peptide detection using FAIMS are shown as Supplementary Figures 3-6 for lowabundance peptides. It is noteworthy that Canterbury et al., reported a 10-fold decrease in ion intensity when using FAIMS compared to conventional electrospray.26 This observation is due to the fact that ions are spread through out the analytical region of the electrodes without any confinement in the lateral dimension. In the present conditions, we observed up to 2-fold decrease in ion intensity primarily for abundant peptide ions. However, we also noted the opposite situation for lowabundance ions where higher absolute intensities were observed using FAIMS resulting in an extended dynamic range of ion detection. The ability to enrich the proportion of multiply charged peptide ions while minimizing the contribution of undesired background ions using FAIMS also provided a marked advantage for the generation of higher quality MS/MS spectra. This is illustrated in Figure 5a for 2409 peptides where more than 75% of matched sequences using FAIMS showed equal or higher Mascot scores than in non-FAIMS experiments. Improvement in the number of assigned MS/MS spectra was also evidenced by an overall increase of 26.1% in sequence coverage for common proteins using FAIMS compared to conventional nanoelectrospray (Figure 5b). Similarly, we noted that the distribution of peptide per protein was generally higher for FAIMS across the entire range of identified proteins (Figure 5c). We attributed the improvement in Mascot scores to higher quality of MS/MS spectra and the lower number of spurious fragment ions from coselected precursor ions. Despite the high resolution (up to 100 000) provided by the LTQ-Orbitrap, ambiguous peptide identification can arise due to the inherent sample complexity. To evaluate the application of FAIMS in minimizing the occurrence of mixed MS/MS spectra, we spiked various amounts of cytochrome C tryptic digests (1 to 100 fmol/ µL) into a protein digest from U937 cell extract. A narrow region of the extracted mass spectrum (m/z 583-588) taken from nanoLC-MS analysis of a 10 fmol spike digest with and without FAIMS is shown in Figure 6a,b. At this concentration, the doubly protonated peptide ion can be unambiguously identified from the FAIMS analysis (Figure 6b), whereas the same ion overlaps with the third isotope ion (m/z 584.85) of a more abundant peptide ion when nanoelectrospray is used (Figure 6a). The MS/MS spectrum of this ion acquired without FAIMS showed abundant fragment ions from the coeluting peptide Journal of Proteome Research • Vol. 8, No. 7, 2009 3361
research articles
Saba et al. The enhanced protein identification observed using FAIMS was also evaluated in the context of differentiated cell model systems to identify early changes in protein expression. The use of FAIMS was advantageously exploited for monitoring protein expression arising from stimulation of U937 cells by phorbol 12-myristate 13-acetate (PMA). This tumor promoter is an exogenous analogue to diacylglyceride (DAG) that binds protein kinase C (PKC) and leads to its prolonged activation.28 Chemical stimulation of U937 cells by PMA initiate cellular differentiation into a more mature macrophage state.29 This differentiation process plays an important role in inflammation, autoimmune diseases, and in the formation of atherosclerotic plaques.30 Their phenotypic and functional characteristics are intimately linked to their stage of maturation. As part of this cellular reorganization, macrophages acquire their functions through a programmed expression of specific proteins and with remodeling of the cell membrane. Macrophages are morphologically different from monocytes by the appearance of cell surface villosities and a thick plasma membrane. Such cellular changes reflect their roles as central effectors and regulatory cells of the inflammatory response. To fulfill these functions, macrophages in their active state are able to produce a wide distribution of inflammation regulators (prostaglandins, leukotrienes, etc.) and microbicidal and cytotoxic agents (superoxide, hydroxyl radical, hydrogen peroxide), together with proteases, cytokines, and proteins involved in tissue reorganization (elastase, collagenase, regulatory growth, angiogenesis factors).
Figure 5. Improvement of sequence coverage and protein identification using nanoLC-MS/MS with and without FAIMS. (a) Comparison of Mascot score for common peptides identified in a protein digest of human monoblatic U937 cells. (b) Distribution of sequence coverage for the top 25 proteins based on the number of identified peptides. (c) Distribution of identified peptides per protein.
ion at m/z 583.85 corresponding to tryptic peptide T438-448 (AVAQALEVIPR) from T-complex protein 1 subunit γ (Figure 6c). In contrast, the doubly protonated peptide ion at m/z 584.81 is clearly resolved in the FAIMS analysis at a spike level of 10 fmol. The MS/MS spectrum of this ion contains prominent y-type fragment ion series enabling the identification of the tryptic peptide T29-39 (TGPNLHGLFGR) from cytochrome C (Figure 6d). It is noteworthy that a positive identification of this peptide by Mascot was obtained for a spike level of 10 fmol using FAIMS, whereas the corresponding assignment without FAIMS required a spike level of at least 100 fmol. These examples illustrate that confounding fragment ions from coselected precursors can give rise to mixed MS/MS spectra that cannot be readily assigned by current database search engines. This situation is more likely to occur for low-abundance peptide ions present in digests of increasing sample complexity. The additional peak capacity of FAIMS confers an inherent advantage in the analysis of complex tryptic digest by reducing the occurrence of conflicting ions in MS-MS experiments. 3362
Journal of Proteome Research • Vol. 8, No. 7, 2009
To probe changes effected by PMA stimulation, U937 monoblast cells (3 day cultures) were split in two to yield control and PMA-exposed (150 nM PMA, 1 h) cell extracts. Cytosol proteins were reduced, alkylated and quantitated prior to MS analyses. NanoLC-FAIMS-MS analyses were performed using CV stepping while interleaving injections across control and PMA sample replicates (n ) 3). These analyses enabled the profiling of peptide distribution for each CV value. We identified a total of 2761 and 3034 peptide ions for control and PMAstimulated cells corresponding to 444 and 507 unique proteins, respectively. It is noteworthy that the corresponding analysis performed without FAIMS enabled the identification of 1895 peptides (351 proteins) and 1958 peptides (362 proteins) for control and PMA-treated cells, respectively. Figure 7a shows a volcano plot distribution of log2 fold change versus log10 p-values for all 3804 unique peptides identified in both conditions. A total of 356 unique peptides showing reproducible changes in abundance in either control or PMA-treated cells are represented by the dotted areas depicting change in abundance ratio greater than 1.5-fold), and p-values
