Article pubs.acs.org/jpr
Detection and Identification of Heme c‑Modified Peptides by Histidine Affinity Chromatography, High-Performance Liquid Chromatography−Mass Spectrometry, and Database Searching Eric D. Merkley, Brian J. Anderson, Jea Park, Sara M. Belchik, Liang Shi, Matthew E. Monroe, Richard D. Smith, and Mary S. Lipton* Pacific Northwest National Laboratories, Richland, Washington 99352, United States S Supporting Information *
ABSTRACT: Multiheme c-type cytochromes (proteins with covalently attached heme c moieties) play important roles in extracellular metal respiration in dissimilatory metal-reducing bacteria. Liquid chromatography−tandem mass spectrometry (LC−MS/MS) characterization of c-type cytochromes is hindered by the presence of multiple heme groups, since the heme c modified peptides are typically not observed or, if observed, not identified. Using a recently reported histidine affinity chromatography (HAC) procedure, we enriched heme c tryptic peptides from purified bovine heart cytochrome c, two bacterial decaheme cytochromes, and subjected these samples to LC−MS/MS analysis. Enriched bovine cytochrome c samples yielded 3- to 6-fold more confident peptide−spectrum matches to heme c containing peptides than unenriched digests. In unenriched digests of the decaheme cytochrome MtoA from Sideroxydans lithotrophicus ES-1, heme c peptides for 4 of the 10 expected sites were observed by LC−MS/MS; following HAC fractionation, peptides covering 9 out of 10 sites were obtained. Heme c peptide spiked into E. coli lysates at mass ratios as low as 1 × 10−4 was detected with good signal-to-noise after HAC and LC−MS/MS analysis. In addition to HAC, we have developed a proteomics database search strategy that takes into account the unique physicochemical properties of heme c peptides. The results suggest that accounting for the double thioether link between heme c and peptide, and the use of the labile heme fragment as a reporter ion, can improve database searching results. The combination of affinity chromatography and heme-specific informatics yielded increases in the number of peptide−spectrum matches of 20−100-fold for bovine cytochrome c. KEYWORDS: C-type cytochromes, heme c, peptides, mass spectrometry, histidine affinity chromatography, enrichment, database searching, reporter ion, isotopic envelope
■
INTRODUCTION C-type cytochromes are proteins that have been covalently modified by the reaction of the two vinyl groups of heme b with the two cysteine sulfhydryl groups of a conserved CXXCH sequence motif (Figure 1A). This modification occurs through the action of the cytochrome c maturation machinery (CCM),1,2 and the resulting double-thioether-linked heme group is referred to as heme c. The genomes of many bacteria (such as Geobacter and Shewanella species) encode multiple multiheme c-type cytochromes.3−5 These multiheme cytochromes function in extracellular metal respiration by transporting electrons from cellular metabolism across the cell envelope to terminal electron acceptors external to the bacterial cell, such as oxidized iron, uranium, and other heavy metal and radiological contaminants.6−9 Such bacteria and their c-type cytochromes have been studied extensively because of their potential use in bioremediation,10,11 microbial fuel cells,12,13 and electrosynthesis of valuable biomaterials.14 Several proteomics studies have analyzed the expression of ctype cytochromes under various conditions.15−19 A shared © 2012 American Chemical Society
feature of these studies is that the cytochrome-rich fractions (cell envelope or extracellular polymeric substance) were purified and explicitly analyzed to efficiently detect cytochromes. Analyses of large-scale proteomics data sets have typically suggested that c-type cytochromes, particularly the heme c peptides, are under-represented.20,21 We surmised that direct detection of the heme-c-modified peptides would improve detection and expression profiling of c-type cytochromes and verify their post-translational modification. Compared to unmodified peptides and other posttranslational modifications, heme c peptides have a unique set of physicochemical properties, including the attachment of the heme group via covalent bonds to two cysteine residues (Figure 1A), the isotopic signature of iron, the lability of the heme group in the gas phase, and the fixed charge of the heme group in the oxidized state. Previous mass spectrometric studies of heme c peptides have detailed several of these features.22,23 Received: August 21, 2012 Published: October 19, 2012 6147
dx.doi.org/10.1021/pr3007914 | J. Proteome Res. 2012, 11, 6147−6158
Journal of Proteome Research
Article
Figure 1. (A) CAQCH + heme from the three-dimensional structure of bovine cytochrome c (PDB code 2b4z49), showing the covalent binding between cysteine residues in the CXXCH motif to the heme vinyl groups, as well as the coordination of Fe by the histidine residue in CXXCH. (B) Schematic of a heme c peptide binding to immobilized histidine showing coordination of iron by the imidazole nitrogen. (C) High-resolution mass spectrum of the major chromatographic peak resulting from histidine affinity enrichment of a bovine cytochrome c digest (averaged over the chromatographic peak). Multiple charge states (1+ through 4+) are visible for [CAQCHTVEK+ heme]. Also visible are the heme c fragment ion (due to in-source decay) at m/z 617.19, and the [M+2H]1+ and [M+2H]2+ ions of the unmodified peptide (again due to in-source decay) at m/z 1018.45 and m/z 509.73. All of these fragments coelute from the reverse-phase liquid chromatography column.
Purification of Multiheme c-Type Cytochromes
In this study, we took advantage of these unique properties in a two-pronged approach to improve detection of heme c peptides. First, we adapted the heme c tag protein affinity purification strategy described by Asher and Bren24 for use with tryptic peptides derived from c-type cytochromes. This enrichment procedure, which we term histidine affinity chromatography (HAC) uses Sepharose resin derivatized with histidine to bind heme iron atoms through the imidazole nitrogen of histidine (Figure 1B). Second, we developed a data analysis approach to filtering proteomics database search results that: (1) identifies the consequences of the unique properties of heme peptides on a typical proteomics workflow and (2) takes advantage of these properties to increases the number and confidence of identifications of heme c peptides from liquid chromatography-tandem mass spectrometry (LC−MS/MS) analyses. In addition to demonstrating enrichment of heme c peptides from both simple and complex mixtures, our results indicate that accounting for the double attachment of the heme group is important in identifying heme c peptides and that the singly charged heme fragment ion is a useful reporter for heme c peptides.
■
MtoA from the iron-oxidizing bacteria Sideroxydans lithotrophicus was overexpressed in Shewanella oneidensis MR-1 and purified as described.25 MtrF from S. oneidensis MR-1 (SO1780) was overexpressed and purified as previously described.26−29 These two decaheme c-type cytochromes are homologous, and both are implicated in extracellular electron transfer in their respective organisms.25−29 Proteolytic Digestion and Preparation of Peptide Samples
Bovine heart cytochrome c was dissolved in 9 M urea and heated at 60 °C for 60 min to denature. No reducing agents were added (1) because cytochrome c has no disulfide bonds, and (2) because of concerns that reduced heme iron would not be retained in histidine affinity chromatography, which later proved unfounded. The sample was diluted 8- to 10-fold with 100 mM ammonium bicarbonate solution (pH 8.0) and trypsin was added at a 1:20 ratio of trypsin to protein, and then digested overnight at 37 °C. MtrF is a peripheral outer membrane protein and was purified in Triton-X100 detergent,28 which interferes with trypsin activity. Triton-X100 was removed using a 0.5-mL detergent removal spin column (Pierce, Rockford, Illinois). The eluate from the spin column was then subjected to a second cycle of detergent removal in a new column. Removal of Triton-X100 was verified spectrophotometrically by the decrease in absorbance at 275 nm.30 The sample was denatured with 6 M urea at room temperature for 1 h, diluted 10- to 12fold with 100 mM ammonium bicarbonate (pH 8.0), and digested with trypsin at a 1:20 trypsin:protein ratio at 37 °C overnight. MtoA was treated similarly except that no detergent removal was necessary. (S. lithotrophicus MtoA, a periplasmic protein, was expressed as a recombinant protein in S. oneidensis
MATERIALS AND METHODS
Reagents L-Histidine, urea, and bovine heart cytochrome c were obtained from Sigma (St. Louis, MO); NHS-activated Sepharose 4 Fast Flow, from GE Healthcare (UK); and sequencing grade modified trypsin, from Promega (Madison, WI). All other chemicals were reagent grade from Sigma or Fisher.
6148
dx.doi.org/10.1021/pr3007914 | J. Proteome Res. 2012, 11, 6147−6158
Journal of Proteome Research
Article
were dried in SpeedVac and reconstituted in 25 mM ammonium bicarbonate. Various wash buffer additives (SDS, Triton-X100, CHAPS, acetonitrile, and NaCl) were tested for their effect on histidineheme interactions. These tests were performed in batch mode with ∼200 μL of resin in the bottom of a microcentrifuge tube. The resin was equilibrated as above, washed first with 50 mM sodium phosphate (pH 7.0), then with the test solution, and finally with 300 mM imidazole in 50 mM sodium phosphate (pH 7.0). The fractions were analyzed by absorbance at 400 nm.
MR-1 and purified from the soluble fraction in the absence of detergent.25) To prepare tryptic digests of Escherichia coli K-12 MG1655, a 50-mL Luria broth culture was grown to late log phase. Cells were harvested by centrifugation, resuspended in 100 mM ammonium bicarbonate buffer, and lysed by bead-beating. The lysate was cleared by centrifugation, after which protein concentration was determined by the BCA assay. The soluble protein fraction was denatured by incubating at 60 °C for 30 min in the presence of 10 M urea and 5 mM dithiothreitol. The sample was diluted 10-fold with 100 mM ammonium bicarbonate buffer, and sequencing grade modified trypsin (Promega, Madison, WI) was then added in a 1:50 trypsin to protein ratio. Digestion was allowed to proceed for 18 h at 37 °C, after which the digest was desalted with a C18 Supelco Discovery solid-phase extraction cartridge and dried in a SpeedVac. Prior to histidine affinity chromatography, the sample was reconstituted in 50 mM sodium phosphate (pH 7.0) and passed through a 0.45 μm filter. E. coli tryptic peptides and the heme c peptide from bovine cytochrome c (preparation described below) were mixed in known mass ratios, and then subjected to histidine affinity chromatography and LC−MS/ MS analysis as described below.
LC−MS and LC−MS/MS Analysis
Peptide samples were analyzed by LC−MS using a previously described high pressure LC system32 with a 60 cm long, 75 μm inner diameter fused silica capillary (Polymicro Technologies, Phoenix, AZ) packed in-house with 3-μm Jupiter C18 stationary phase (Phenomenex, Torrance, CA). Mobile phase A consisted of 0.1% formic acid in water and mobile phase B, of 80% acetonitrile/20% water/0.1% formic acid. The LC column was coupled32 via electrospray ionization to the mass spectrometer. We used several different mass spectrometry platforms according to the needs of the particular experiment. The chromatographic conditions were same regardless of the mass spectrometer used. A Thermo Scientific LTQ Orbitrap mass spectrometer was used to analyze bovine cytochrome c and bacterial multiheme cytochrome samples using a “high/low” MS and MS/MS strategy, in which high-resolution precursor spectra were acquired with the Orbitrap mass analyzer of the hybrid mass spectrometer, and low-resolution MS/MS spectra were acquired with the ion trap. The high resolution MS data are necessary for identifying isotopic patterns characteristic of iron (see below). To evaluate the impact of higher-energy collisional dissociation (HCD), an additional cytochrome c sample was analyzed in triplicate using a Thermo Scientific LTQ Orbitrap Velos mass spectrometer using alternate CID/HCD MS/MS scans in a “high/high” strategy. Dynamic exclusion was applied in all MS/MS experiments. E. coli samples spiked with heme c peptide were analyzed using a Thermo Scientific Exactive Orbitrap mass spectrometer, which collected high-resolution MS data (without MS/MS) for analysis by the accurate mass and elution time (AMT) tag approach.33 These experiments provide an indication of the performance of HAC in complex mixtures as well as an indication of the feasibility of combining the HAC method with the AMT tag approach RAW data files for all LC−MS experiments described here may be downloaded from http://omics.pnnl.gov.
Preparing the Histidine Resin
N-Hydroxysuccinimide (NHS)-activated Sepharose 4 Fast Flow resin (GE Healthcare) was derivatized with histidine according to the manufacturer’s recommendations, that is, a wash with 1 mM HCl followed by incubation for several hours with conjugation buffer of 10 mg/mL L-histidine, 0.2 M NaHCO3, 0.5 M NaCl (pH 8.3), then incubated with 0.1 M Tris (pH 8.5) to quench any remaining NHS groups. We also used an alternative procedure in which the histidine concentration in the conjugation buffer was increased to ∼40 mg/mL, the resin was incubated in the conjugation buffer overnight, and the Tris quench step omitted. (Note that histidine reacts readily with NHS esters to form an intermediate that is quickly hydrolyzed31). The two procedures produced resins with similar properties. Histidine Affinity Chromatography
HAC essentially followed the procedure of Asher and Bren.24 We used either spin columns (Pierce Handee-spin, resin volume ∼0.5 mL) or gravity columns (Bio-Rad, resin volume ∼1−2 mL) with flow rate enhanced by means of a squeeze bulb applied to the top of the column. In each case, the resin was equilibrated in 3−5 resin volumes binding buffer (50 mM sodium phosphate (pH 7.0)). After loading the sample onto the column, the column was washed with 3−5 volumes of binding buffer (or until absorbance at 280 nm = 0). Finally, specifically bound heme-c peptides were eluted with 1−3 bed volumes of elution buffer (300 mM imidazole, 50 mM sodium phosphate (pH 7.0)). After HAC, samples were desalted using either C18 OMIX microextraction pipet tips (Varian) or with C18 Supelco Discovery solid-phase extraction cartridges. Samples were assayed for peptide concentration using the BCA assay (Pierce, bovine serum albumin standards) and heme c content was measured by absorbance at ∼400 nm (any wavelength from 394 to 410 nm can be used). Note that the desalting step was necessary because imidazole in the elution buffer interferes with the BCA assay. The heme c peptide enrichment factor was calculated as (A400/concentration)enriched/(A400/concentration)starting material. Samples to be analyzed by LC−MS/MS
Data Processing
DeconMSn34 was used to convert Thermo-format “.RAW” files to “.dta” files searchable by Sequest. The data were searched using TurboSequest v27.1235 and the following settings: parent mass tolerance, 50 ppm; fragment mass tolerance, 0.5 Da; partially tryptic enzyme rules; and a dynamic modification of 615.1694 Da on cysteine (C) residues,22 Following Yang et al.,22 we calculated this modification mass by starting with the modification formula C34H32FeN4O4, subtracting the mass of a proton (1.0073 Da), and then adding the mass difference between 56Fe and 54Fe (1.9953 Da). The first adjustment accounted for the fact that the charge is carried by the oxidized iron of the heme group and not by a proton as assumed by Sequest. The second adjustment directed the search toward the most abundant peak of the isotopic distribution, which 6149
dx.doi.org/10.1021/pr3007914 | J. Proteome Res. 2012, 11, 6147−6158
Journal of Proteome Research
Article
corresponds to 56Fe/12C; for example, the peak labeled “56Fe/12C 817.32” in Figure 1C (inset). For bovine cytochrome c samples, the data were searched against a database of six standard proteins that included cytochrome c, as well as common trypsin and keratin contaminants. For MtoA, the data were searched against the MtoA sequence (Uniprot accession number D5CMQ0_SIDLE) plus trypsin and contaminants. For MtrF, the data were searched against the sequences of S. oneidensis MR-1 proteins MtrD, MtrE, and MtrF (Uniprot accession numbers Q8EG30_SHEON, Q8EG31_SHEON, and Q8EG32_SHEON), plus trypsin and keratin contaminants. To estimate the false discovery rate of peptide identifications, the entire S. lithotrophicus ES1 (http://www.ebi.ac.uk/ena/data/ view/ADE12722) or S. oneidensis MR1 (http://www.ebi.ac.uk/ ena/data/view/AE014299) genomes were used. The Sequest search results were filtered according to the criteria used by Washburn, et al.36 (ΔCn2 ≥ 0.1; singly charged, fully tryptic and Xcorr ≥ 1.9; doubly charged, fully or partially tryptic and 2.2 ≤ Xcorr ≤ 3.0, or Xcorr ≥ 3.0 regardless of tryptic state; triply charged, fully or partially tryptic and Xcorr ≥ 3.75). Decon2LS37 was used to deisotope the data, and the “_isos.csv” results file was used to find deisotoped precursor masses in each scan. MASIC38 was used to find the precursor (full MS) scan corresponding to the chromatographic peak apex for each peptide selected for MS/MS. A second MASIC analysis was used to find reporter ions in MS/MS spectra. The Sequest synopsis files (containing at least the top 2 hits for each scan), Decon2LS “isos.csv” files, and the MASIC “SICstats.txt” and “ReporterIons.txt” files were loaded into separate tables of a Microsoft SQL Server database for subsequent analysis. Potential peptide−spectrum matches (PSMs) to heme c peptides are referred to as “heme hits”. Structured query language (SQL) queries were written to identify heme hits for each of the four classes described below. (The complete SQL code is included in Supporting Information.) Note that all queries included a selection for the presence of the CXXCH motif in the identified peptide.
■
to 5%, 6 M urea, and 1 M NaCl. Specific binding was abrogated by SDS at 0.1% and by acetonitrile (Supporting Information Figure S1). In the enriched sample (Figure 1C) we observed singly-, doubly-, triply-, and quadruply charged [CAQCHTVEK +heme], as well as the free heme c group as a singly charged fragment ion (m/z 617.19). The singly- and doubly charged ions of CAQCHTVEK without the heme group were also present at m/z 1018.45 and m/z 509.73, respectively. Both the heme fragment and the heme-free peptide ions coeluted with all the charge states of the intact heme peptide, indicating that the fragmentation occurred in the mass spectrometer and not during sample preparation or chromatography. Figure 2 shows results from an LC−MS (LTQ Orbitrap Velos) analysis of an HAC enriched tryptic digest of
RESULTS
We have explored two aspects of the problem of identifying heme peptides: chromatographic enrichment, and filtering of database searching results. We will first describe the use of HAC to enrich heme c peptides from increasingly complex samples: a digest of purified bovine cytochrome c (105 residues), a digest of a 355-residue, decaheme c-type cytochrome, and finally, the entire E. coli proteome with an exogenous heme c peptide spiked in at known ratios. We will then describe the database search strategy and its results for the bovine cytochrome c and decaheme cytochrome samples.
Figure 2. Enrichment of the heme c-containing peptide [CAQCHTVEK+heme] from a tryptic digest of bovine cytochrome c, analyzed by LC−Orbitrap-MS. (A) Total digest; (B) wash fraction from histidine affinity chromatography; (C) eluent fraction (i.e., from washing the column with 300 mM imidazole) from histidine affinity chromatography. Black lines, total ion chromatogram; red lines, sum of extracted ion chromatograms from heme-related species ([M+heme]1+, m/z 1634.50−1643.75; [M+heme+H]2+, m/z 816.30−819.90; [M+heme +2H]3+, m/z 544.50−546.60; [M+heme+3H]4+, m/z 408.60−410.70; [heme]1+, m/z 614.14−618.25; [M+H]1+, m/z 1018.40−1020.50; [M +H]2+, m/z 509.68−510.78).
Histidine Affinity Chromatography of Tryptic Digests of Bovine Cytochrome c
Using the spin column method, the BCA assay to measure total protein concentration, and absorbance at 405 nm to track heme peptide concentration, we achieved 3.45 ± 0.3-fold enrichment of the heme-containing tryptic peptide from bovine heart cytochrome c (mean ± standard deviation of four experiments). A similar value was obtained using a gravity column and a second batch of resin. Note that the theoretical maximum enrichment calculated from the molecular weights of intact cytochrome c and its heme-containing tryptic peptide assuming complete digestion is 7.5-fold. The heme iron-histidine interaction was stable in Triton-X100 up to 5%, CHAPS up
cytochrome c. The signals of all the detected heme-c related peptides shown in Figure 1C were summed to construct the extracted ion chromatograms (red traces). The heme c peptide is readily detected in the digest (Figure 2A) but is not present in the HAC wash fraction (Figure 2B). By contrast, the signal 6150
dx.doi.org/10.1021/pr3007914 | J. Proteome Res. 2012, 11, 6147−6158
Journal of Proteome Research
Article
from the HAC eluent fraction (Figure 2C) consists of almost exclusively heme-c related species that elute around 43.7 min. These results indicate that the histidine-conjugated NHSSepharose resin is selective for heme-c-modified peptides. We analyzed two HAC-fractionated digests of cytochrome c, one using the LTQ Orbitrap with a “high/low” strategy, and one (in triplicate) using the LTQ Orbitrap Velos with a “high/ high” strategy and alternating CID/HCD MS/MS scans. Using the peak areas (derived from precursor scans only) as shown in Figure 2, we calculated the peptide [CAQCHTVEK+heme] to be 7.3(±2.8)-fold enriched in the eluent fraction compared to the starting digest (average ± standard deviate of all four analysis). This value is higher than that obtained by UV−vis measurements, most likely because the sensitivity of the BCA assay differs for the digest and the purified peptide. Our value is also close to the 7.5-fold theoretical maximum, which suggests that HAC captured a large majority of the heme peptide present. In terms of peptide identifications, we identified three times more MS/MS spectral matches for [CAQCHTVEK +heme] in the HAC eluent than in the starting digest (Table 2). Histidine Affinity Chromatography of Decaheme Cytochromes
We also tested the HAC procedure on tryptic digests of the decaheme cytochromes MtoA from S. lithotrophicus ES-1and MtrF from S. oneidensis MR-1. The total and selected ion chromatograms from LC-Orbitrap-MS analyses for several representative tryptic peptides from MtoA are shown in Figure 3. Note that the HAC eluent consists almost entirely of heme peptides, indicating that many heme peptides are completely bound to the HAC resin and specifically eluted by imidazole ([TPNCQTCHGESANHLK+heme], green trace; [ATQTEVCYTCHK+heme], teal trace; [ATQTEVCYTCHKER+heme], gray trace). However, other peptides do not bind at all and are present only in the wash fractions ([ISTHEPIEGKVVCSDCHPHGSAGPK+heme], magenta trace; [GPFLFAHQPVTEDCTNCHMPHGSNIAPLLK+heme], brown trace). This finding is most likely not due to column overload because several heme c peptides are present only in the eluent. Further, during HAC of bovine cytochrome c peptides at similar loading, the heme peptides can be observed to form a narrow red band at the top of the column that spreads very little during the wash, and does not migrate until imidazole is added, which suggests that the binding occurs in the top few percent of the resin bed (Supporting Information Figure S2). Importantly, the number of histidine residues in a peptide is negatively correlated with the degree of retention on the HAC column (Supporting Information Figure S3), suggesting that peptide-derived histidine residues coordinate the heme iron and prevent binding by the stationary-phase imidazole groups. Overall, the results indicate that heme c containing peptides bind to HAC resin.
Figure 3. Enrichment of heme c-containing peptides (indicated in the legend) from a tryptic digest of Sideroxydans lithotrophicus ES-1 MtoA by histidine affinity chromatography, analyzed by LC−MS. (A) Input digest; (B) HAC column flow-through; (C) buffer wash; (D) eluent (300 mM imidazole). Peptides were identified by accurate mass, isotope pattern, and MS/MS spectra. Note that some peptides are not bound by the histidine affinity chromatography resin (e.g., pink and brown traces); these peptides have histidine residues in addition to the CXXCH motif.
Figure 4. HAC enrichment of the peptide [CAQCHTVEK+heme] from complex mixtures. HAC-purified [CAQCHTVEK+heme] was mixed with tryptic digests of the soluble fraction of E. coli cell lysate at the indicated mass ratios and subjected to HAC as described in the text, and analyzed by Exactive LC−MS experiments. Data are the peak areas from extracted ion chromatograms of the [CAQCHTVE+H +heme]2+ signal. Black circles, digests prior to enrichment; white circles, wash fraction from HAC experiments; red squares, HAC eluent fractions. (Inset) Close-up of the three lowest mass fraction data points, together with a linear fit.
Histidine Affinity Chromatography of Complex Mixtures
To evaluate the effectiveness of HAC chromatography in isolating heme c peptides from complex mixtures, we added the HAC-purified cytochrome c peptide [CAQCHTVEK+heme] to a tryptic digest of E. coli lysate in known ratios that ranged from 1 × 10−4 to 0.2 mass fraction cytochrome c peptide. These mixtures were subjected to HAC and analyzed by LC−MS (Figures 4 and S4, Supporting Information) using the Orbitrap Exactive mass spectrometer. No MS/MS data were collected, but the Exactive analysis nevertheless provides information
about the effectiveness of HAC, and the potential for combining HAC with the AMT tag approach.33 At mass fraction ≥0.01, the normalized extracted ion chromatogram peak area of [CAQCHTVEK+heme]2+ was nearly constant, 6151
dx.doi.org/10.1021/pr3007914 | J. Proteome Res. 2012, 11, 6147−6158
Journal of Proteome Research
Article
which suggests maximal recovery of the heme c peptide. Between mass fraction 5 × 10−3 and 1 × 10−4, there was a linear relationship between the ratio and the peak area (R2 = 0.996). At mass fraction ≥1 × 10−3, the [CAQCHTVEK +heme] peak is a major feature of the total ion chromatogram (Supporting Information Figure S4), and the average mass spectrum of this peak resembles that in Figure 1C, with little contamination from other peptides. At the lowest ratio of 1 × 10−4, the heme peptides are clearly detected (Supporting Information Figure S4) although the heme peptide peak is only a minor feature of the chromatogram. This lowest mass fraction corresponds to 0.4 ng of the heme c peptide injected onto the LC column, in a background of 4 μg total peptides. These results suggest that the HAC method will be applicable to detect c-type cytochromes in cell lysates, although some prefractionation may be necessary. In addition, the quality of the isotopic envelope (and thus the accurate mass determination) obtained for the cytochrome c peptide at the lowest ratio tested suggests that the AMT tag approach combined with HAC will be an effective strategy for identifying heme peptides from proteome-scale samples.
will be rejected. To overcome this problem, we considered the top two hits for every MS/MS spectrum. The identification is accepted provided both the first and second ranked PSMs identify the same peptide (with the modification on alternate sites), and provided the Xcorr and ΔCn2 scores of either of the two top PSMs meet the filtering criteria described above. Class II: Isotopic Distribution of Iron. The major isotopes of iron are 56Fe (0.917 mol fraction) and 54Fe (0.058 mol fraction).39 The main effect on the shape of the isotopic envelope is the presence of two small peaks (labeled “54Fe/12C” and “54Fe/13C” in the inset of Figure 1C) on the low m/z side of the isotopic distribution. Heavier iron isotopes have essentially no effect on the shape of the isotopic envelope. Our procedure uses the parent ion isotope distribution to confirm the presence of iron in the peptide. Class II hits represent heme peptides identified by the top two Sequest results, regardless of Xcorr or ΔCn2 score, whose precursor ion mass measurement error is
