Detection, Identification, and Quantification of Selenoproteins in a

ARTICLE pubs.acs.org/ac

Detection, Identification, and Quantification of Selenoproteins in a Candidate Human Plasma Standard Reference Material Guillaume Ballihaut,† Lisa E. Kilpatrick,‡ and W. Clay Davis*,† †

Hollings Marine Laboratory, Analytical Chemistry Division, National Institute of Standards and Technology, 331 Fort Johnson Road, Charleston, South Carolina 29412, United States ‡ Chemical and Biochemical Reference Data Division, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, Maryland 20899-8320, United States

bS Supporting Information ABSTRACT: To understand the effect of Se supplementation on health, it is critical to accurately assess the Se status in the human body by measuring reliable biomarkers. The preferred biomarkers of the Se status are selenoprotein P (SelP) and glutathione peroxidase 3 (GPx3) along with selenoalbumin (SeAlb), but there is still a real need for reference methods and reference materials to validate their measurements. Therefore, this work presents a systematic approach to provide quality control data in selenoprotein measurements. This approach combines online isotope dilution affinity liquid chromatography (LC) coupled to inductively coupled plasma mass spectrometry (ICPMS), laser ablation ICPMS, and tandem mass spectrometry (MS/MS) to identify and quantify SelP, GPx3, and SeAlb in a human plasma reference material SRM 1950. Quantitative determinations of SelP, GPx3, and SeAlb were 50.2 ( 4.3, 23.6 ( 1.3, and 28.2 ( 2.6 ng g 1 as Se, respectively. The subsequent identification of the selenoproteins included nine SelP peptides, including two selenopeptides and nine GPx3 peptides, while albumin was identified with a protein coverage factor >95%. The structural elucidation of selenoproteins in the target Se affinity fractions in SRM 1950 provides information needed for method validation and quality control measurements of selenoproteins and therefore the selenium status in human plasma.

elenium (Se) is an essential micronutrient for human health.1 The nutritionally essential functions of Se are understood to be fulfilled by selenoproteins, which are proteins containing Se as the 21st naturally occurring amino acid selenocysteine (Sec, U). Twenty-five selenoproteins have been identified in humans with functions such as antioxidant protection and cellular redox balance, and collectively, selenoproteins are essential for life.2 The current Se recommended dietary allowance is estimated at ≈57 μg/day while the tolerable Se upper intake level is established at 400 μg/day.3 Therefore, Se intake must be tightly monitored and controlled to avoid severe Se deficiency associated with Keshan cardiomyopathy disease or Se toxicity. Selenium has recently gained a considerable interest due to epidemiologic data which showed that Se may have a protective effect in the prevention of cancer and other diseases.4 This has stimulated interest in testing selenium-supplementation effects on human health through human clinical trials.5,6 To understand the effect of Se intake on health, it is critical to accurately access the Se status in the human body and therefore to measure reliable biomarkers of the Se status. The National Institutes of Health defines a biomarker as “a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention.”7 Ideally, biomarkers are molecules quantifiable from blood plasma or another tissue of fluid easily accessible for measurements. Selenium supplementation studies in humans have identified three biomarkers of the Se status in blood: total selenium, selenoprotein

S

This article not subject to U.S. Copyright. Published 2011 by the American Chemical Society

P (SelP), and glutathione peroxidase 3 (GPx3).8 GPx3 and SelP plasma selenoprotein concentrations reflect the selenium intake required for the full expression of selenoproteins in the human body.9 Clinical trials have been conducted with the aim of determining the minimum Se intake needed to reach the full expression of selenoproteins, i.e., to optimize selenoproteins. A total intake of 49 μg/day optimized SelP concentration, while a lower Se intake of 35 μg/day optimized plasma GPx3 activity.9 The total plasma selenium concentration could not be optimized during selenium-supplementation trials and reflects a nonsaturable pool of selenium incorporated nonspecifically in plasma proteins in addition to the Se in plasma selenoproteins.9 Considering that selenoproteins represent the biologically active form of Se, plasma GPx3 and SelP selenoproteins are therefore considered as the most accurate selenium-status biomarkers and are preferred to estimate the Se human requirement. With the growing demand for accurate quantification methods for selenoprotein measurements in blood samples, appropriate plasma and serum reference materials are needed. The use of such reference materials is of the utmost importance in determining method accuracy and reproducibility as well as providing measurement quality control and assurance for unknown patient samples. Reference materials also support intra- and interlaboratory performance assessment and comparison. However, a few Received: August 11, 2011 Accepted: October 3, 2011 Published: October 18, 2011 8667

dx.doi.org/10.1021/ac2021147 | Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry human serum reference materials have been certified for their total Se content but not for their selenium-species content. In terms of the value assignment of small molecules and clinically relevant proteins in reference materials, isotope dilution mass spectrometry (IDMS) has been frequently used to maximize measurement accuracy and minimize measurement uncertainty.10 Application of species-specific IDMS to selenoprotein measurements is not a simple task due to the lack of suitably labeled or enriched intact selenoproteins. As a result, development of alternative procedures, including species-unspecific IDMS techniques, has been favored for the measurement of selenoproteins in human serum samples. The most advanced procedure consists of the efficient separation of Se species in human serum by double-column affinity liquid chromatography (dc-AF-LC), i.e., a heparin Sepharose column and a reactive blue 2 Sepharose column in tandem, with a postcolumn addition of an enriched Se spike for the quantification of Se compounds by isotope dilution inductively coupled plasma mass spectrometry (ID-ICPMS).11 If the Se mass balance is maintained (e.g., 100% column recovery) and the Se species eluted from the columns are unambiguously identified, this dc-AF-LC ID-ICPMS procedure can be used to assign selenoprotein values as Se in human blood. Unambiguous identification of Se species in human plasma, which aims to identify low-abundant selenium-containing proteins in complex human plasma matrix, requires the use of sensitive and specific methods. Separation of Se species from human serum samples by dc-AF-LC ICPMS typically results in three selenium-containing fractions corresponding to the nonretained fraction (F1), the heparin Sepharose retained fraction (F2), and the blue Sepharose retained fraction (F3). The only attempt to unambiguously identify selenoproteins in these fractions was carried out by Jitaru et al.; however, no selenoproteins were identified due to a lack of sensitivity of the mass spectrometry procedure used.12 In a previous investigation, Deagan et al. performed indirect identification of selenoproteins in these fractions and reported the presence of GPx3 in F1 through the detection of peroxidase activity and the detection of SelP in F2 by using a homemade SelP antibody.13 In F3, albumin was determined as a plasma selenium-containing protein based on a colorimetric method and an antiserum developed against human albumin.13 These promising results have never been reproduced, and selenoprotein identification has never been confirmed by a complementary sensitive and specific mass spectrometry based procedure. An error in protein identification can completely undermine the success of the quantitative measurement. Although albumin, GPx3 selenoprotein, and SelP selenoprotein have been unambiguously identified in the human plasma of healthy individuals by tandem mass spectrometry (MS/MS) analyses, the presence of other selenium-containing proteins in human plasma proteome has been also reported. A reduced set of human plasma proteins identified by the Human Proteome Organization with a confidence level higher than 95% included selenoprotein H and a protein similar to selenium-binding protein 1, which were identified by only one to three peptides, along with SelP and GPx3 identified with multiple peptides.14 Selenoprotein N was additionally identified by three distinct peptides in the characterization of the human plasma proteome by multidimensional liquid chromatography mass spectrometry analyses.15 Identification of new selenoproteins in the human plasma of healthy individuals is more than likely the result of an instrumental or sampling procedure bias, but cannot definitively be excluded. No

ARTICLE

specific bioassays have been developed for the identification of these potential plasma selenoproteins; consequently, mass spectrometry based procedures represent one of the best alternatives to confirm selenoprotein identities in human plasma and serum samples. The unambiguous identification of selenoproteins in selenium-containing fractions from dc-AF-LC still proves to be challenging and is critical to validate their quantitative measurements by ID-ICPMS in human plasma/serum. This work reports on the identification and quantification of selenium species in human plasma by online isotope dilution affinity liquid chromatography coupled to ICPMS with Se detection of the fractions by laser ablation ICPMS prior to analyses by LC/MS/MS as demonstrated through the analysis of candidate human plasma standard reference material SRM 1950 “Metabolites in Human Plasma”. Consequently, SRM 1950 is a new human plasma reference material with mass fraction values for its total Se and Se species content which will aid in the method validation and quality control measurements of selenium status in clinical samples.

’ EXPERIMENTAL SECTION Disclaimer. Certain commercial equipment, instruments, or materials are identified in this paper to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology (NIST), nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. Sample Identification. SRM 1950 Metabolites in Human Plasma consists of a fresh frozen human plasma pool collected from 100 healthy individuals (equal number of men and women) to represent the overall racial distribution of the U.S. population. Fourteen SRM 1950 vials (≈1 mL each) were used for this study. SRM 1598a Inorganic Constituents in Animal Serum and SRM 1577 Bovine Liver were used as control materials for the total Se measurements. Before analyses, SRM samples were removed from a 80 °C freezer, placed on the laboratory bench until the samples reached room temperature, and homogenized by gentle shaking. Reagents and Chemicals. SRM 3149 Selenium Standard Solution was used to calibrate the isotopically enriched 77Se spike (Trace Sciences International, Wilmington, DE) solution. Highpurity solvents were obtained from Burdick & Jackson (Muskegon, MI, U.S.A.). RapiGest was obtained from Waters Corporation (Milford, MA, U.S.A.). Trypsin (sequencing grade modified) was purchased from Promega (Madison, WI, U.S.A.). Ultrahigh purity nitric acid was obtained from Fisher Scientific (Suwanee, GA, U.S.A.). All other reagents were obtained from Sigma (St. Louis, MO, U.S.A.) unless specified otherwise. Analytical Procedures and Instrumentation. Total Selenium Quantification by Isotope Dilution Inductively Coupled Plasma Mass Spectrometry. Total Se measurements were made using a Thermo Electron X series II ICPMS (Bremen Germany) with a standard low-volume glass impact bead spray chamber (Peltier-cooled at +3 °C), concentric glass nebulizer, and operating in collision cell mode utilizing 8% H2 in 92% He as the collision gas. First, a 77Se isotopic spike solution was calibrated by reverse isotope dilution using SRM 3149 Selenium Standard Solution. Approximately 1.0 g from each unit of SRM 1950, 1 g of SRM 1598a, 0.25 g of SRM 1577, three aliquots of the standard 8668

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry calibration mixture, and four method blanks (Milli-Q water) were accurately weighed into a cleaned quartz microwave digestion vessel and combined with an accurately weighed aliquot (≈0.5 g) of isotopically enriched 77Se solution spike. Three grams of high-purity HNO3 and 2 g of Milli-Q water were added to each vessel under a class 100 clean bench. The vessels were then sealed and placed into a Perkin-Elmer Multiwave (Shelton, CT) microwave digestion unit for digestion. After cooling to room temperature, the contents of each vessel were transferred to precleaned 50 mL autosampler jars and were diluted with 1% butanol in Milli-Q water to approximately 50 mL to perform sample measurements within the working analytical range of the ICPMS instrument. Total Se concentration was determined by ID-ICPMS using the 77Se/80Se isotope ratio. All data were corrected for detector dead-time and instrument mass bias. The uncertainties were combined according to the International Organization for Standardization (ISO) and NIST guidelines.16,17 The major uncertainty components included the standard deviation of the mean of the selenium measurements (eight replicate measurements), the standard deviation of the mean of the spike calibration mixes (n = 4), the standard deviation of the mean of the blank concentration measurements (n = 4), the estimated standard uncertainty of weighing, the estimated standard uncertainty for the instrument mass discrimination, the estimated standard uncertainty for the instrument dead-time correction, the estimated standard uncertainty for the instrument background correction, and the standard uncertainty of primary calibrant SRM 3149. Online Isotope Dilution Affinity Chromatography Inductively Coupled Plasma Mass Spectrometry. A dual-pump Dionex ICS-3000 (Sunnyvale, CA) was used throughout the analysis. Separation of intact selenoproteins was carried out by means of double-column affinity chromatography (via a manual column switching valve) consisting of heparin Sepharose (HEP) (GE Healthcare, Piscataway, NJ) and blue Sepharose (BLUE) (GE Healthcare) stationary phases. The system setup and the packing procedure used are described in details elsewhere.18 The affinity chromatography columns were packed in house into PEEK columns (4 mm  50 mm). The outputs of both pumps of the IC system were joined via a PEEK tee piece and then coupled to the X7 ICPMS (Winsford, U.K.) nebulizer by PEEK tubing. The mobile phase composition chromatographic method details are listed in Supporting Information Table S-1. Se measurements after affinity chromatography were made using a Thermo Electron X7 ICPMS with a standard low-volume glass impact bead spray chamber (Peltier-cooled at +3 °C), concentric glass nebulizer, operating in collision cell mode utilizing 8% H2 in 92% He as the collision gas. Prior to analysis of the SRM samples, the mass flow of the delivered spike solution was calibrated by the collection of Se spike solution eluting from pump B at a flow rate of 0.150 mL min 1. The selenoprotein mass fraction was determined from analysis from the six units of SRM 1950. Injections (100 μL) of undiluted SRM 1950 plasma were loaded directly onto the HEP and BLUE columns with the constant introduction of the 77Se isotopic spike at a flow rate 0.150 mL min 1. GPx3 has no affinity with either column and thus directly elutes off the columns. The manual switching valve was switched at 5 min isolating the BLUE column from the flow stream, while the LC gradient is switched from the loading solvent (50 mmol L 1 ammonium acetate) to the elution solvent (1.5 mol L 1 ammonium acetate) allowing for the elution of the SelP from the HEP column. The manual

ARTICLE

switching valve is then switched back at 9 min allowing the SeAlb to elute from the BLUE column. The concentration of Se was determined by monitoring the transient signals of 77Se, 78Se, and 80Se and was calculated using the isotope dilution equation for online ID-ICPMS measurements using the 77Se/80Se ratio (due to improved signal-to-noise).19 Six injections of Milli-Q water (blank) were run concurrently with the analytical samples. There was no detectable Se in the blanks, and therefore no blank correction was applied. The uncertainties were combined according to ISO and NIST guidelines.16,17 The major uncertainty components included the standard deviation of the mean of the selenium measurements (six replicate measurements), the standard deviation of the mean of the spike calibration mixes (n = 4), the estimated standard uncertainty of weighing measurements, the estimated standard uncertainty for the instrument mass discrimination, the estimated standard uncertainty for the instrument dead-time correction, and the standard uncertainty of primary calibrant SRM 3149. Protein Identification. Collection of Selenium-Containing Fractions. Four successive 100 μL injections of SRM 1950 were introduced onto the affinity columns, and the selenium fractions were collected with an ISCO Foxy fraction collector. Seleniumcontaining fractions collected at the same retention time were pooled into three large fractions which were desalted and concentrated in a spin concentrator (3 kDa molecular weight cutoff, Millipore) to a final volume of 200 μL for the first Se fraction (F1), 400 μL for second Se fraction (F2), and 400 μL for the third Se fraction (F3). Detection of Selenoprotein on Gel Blots by Laser Ablation Inductively Coupled Plasma Mass Spectrometry. Instrumentation and procedures for electrophoresis, electroblotting and laser ablation ICPMS (LA-ICPMS) analyses were used as described in detail elsewhere.20 Briefly, gel electrophoresis of proteins was performed in denaturing conditions before protein transfer onto a poly(vinylidene fluoride) (PVDF) membrane by electroblotting. LA-ICPMS analyses were then performed in wet plasma conditions for maximum Se sensitivity by humidifying the plasma with a solution of 1% butanol. Tryptic Digestion of Proteins. Tryptic digests of SRM 1950 were prepared in solution as described in details elsewhere.20 To prepare tryptic peptide mixtures from proteins in seleniumcontaining fractions, protein concentration in F1, F2, and F3 was determined by a Bradford assay (Pierce, Rockford, IL, U.S.A.) on a UV vis spectrophotometer (Spectramax plus, Molecular Devices, Sunnyvale, CA, U.S.A.). Tryptic digests of SRM 1950 were prepared in solution as follows: ≈100 μg of proteins in selenium-containing fractions 1 and 2 and ≈200 μg of proteins in selenium-containing fractions 3 were reduced and denatured at 60 °C for 30 min following with addition of RapiGest and dithiothreitol in 50 mM Tris buffer (pH 8.0) at a final concentration of 0.2% (w/v) and 15 mmol L 1, respectively. The mixture was incubated again at 37 °C for 30 min. Next, the proteins were alkylated by adding iodoacetamide to a concentration of 25 mmol L 1, and the mixture was incubated at room temperature for 60 min in the dark with vortex mixing. The reaction was quenched by the addition of dithiothreitol (DTT) at a concentration of 50 mM and incubation of the mixture at room temperature for 30 min with vortex mixing. Digestion with trypsin was carried out at 37 °C for 20 h with a trypsin-toprotein ratio of 1:50, by weight. The digestion was halted by the addition of trifluoroacetic acid (TFA) to a concentration of 0.5% by volume and incubation of the sample at 37 °C for 60 min. 8669

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry The sample was then centrifuged for 10 min at 14 000gn at 4 °C. Supernatant was transferred into a new vial, dried with a centrivap concentrator (Labconco, Kansas City, MO, U.S.A.), and resolubilized in 40 μL of 1% formic acid. Peptides were purified on ZipTip μC18 according to the manufacturer’s instructions (Millipore Corporation, Bedford, MA) and subjected to linear ion trap mass spectrometry (LTQ-MS) analyses. Tryptic digestion of proteins in selenium-containing bands was performed as described in details elsewhere.20 Peptides were purified on ZipTip μC18 according to the manufacturer’s instructions and subjected to LTQ-MS analyses. NanoLC-Chip LTQ-MS/MS Analyses. An Eksigent cHiPLC NanoFlex LC system was coupled to a linear ion trap (LTQ, Thermo Fischer Scientific). The Eksigent LC system was used for the separation of peptides and consisted of an autosampler, binary gradient pump, and column compartment. For LC separations, a LC-chip was used (Nano cHiPLC column 75 μm  15 cm ChromXP C18-CL 3 μm 120 Å, Eksigent technologies) with a two-step gradient chromatographic method from 2 25.8% acetonitrile/0.1% formic acid over 75 min and from 25.8 90% acetonitrile/0.1% formic acid over 10 min at a flow rate of 400 nL/min. The column effluent was directed into an LTQ ion trap MS, and MS/MS spectra were collected using Xcalibur software (version 2.0.7, Thermo Scientific). For each tryptic digest, replicate data-dependent MS/MS analyses were performed with the following settings: centroid mode with either (1) a mass range of 300 2000 or (2) a mass list of GPx3 or SelP peptide precursor ions, followed by eight MS/MS events in the linear ion trap. To prevent repetitive analysis, dynamic exclusion was enabled with a repeat count of 1 and an exclusion duration of 60 s (LTQ) (exclusion list size of 150) for the LTQ analyses. MS scan functions and HPLC solvent gradients were controlled by the Xcalibur data system (ThermoFisher, San Jose, CA). Data Analyses. A first type of analysis was performed using NIST MSPepSearch. Raw files (generated by Xcalibur) were converted into Mascot Generic format (MGF) with ReAdW4Mascot software, an extension of ReAdW.exe (Patrick Pedroli, Institute for Systems Biology). MGF files were submitted for query using NIST MSPepSearch search engine21 against a human protein reference library22 (released on January 14, 2010), which includes 345 489 MS/MS spectra of 196 315 peptides for a total proteome sequence coverage of 21%, 1414 HSA spectra for a total HSA protein sequence coverage of 96.06%, 38 SelP spectra for nine SelP tryptic peptides and a total SelP protein sequence coverage of 41.47%, and 45 GPx3 spectra for eight GPx3 tryptic peptides and a total GPx3 protein sequence coverage of 62.39%. The following types of modifications were included in the reference library utilized with NIST MSPepSearch: cyclization of N-terminal glutamine, oxidation of methionine, and carbamidomethylation of cysteine residues. Peptide identifications were considered to be correct for matches with a MS PepSearch value >450, which is an empirical approximation of a 1% false discovery rate threshold on the peptide spectrum match level. Experimental MS/MS spectra of peptides were compared to MS/MS spectra of peptides compiled in the NIST Peptide Mass Spectral Human library by using MS Search 2.0 2008 software.23 Another type of analysis was performed using SEQUEST. Raw data were converted into DTA files (SEQUEST, Thermo Finnigan) which were submitted for query against an IPI FASTA human protein database (released 10/08/2010, 89 378 entries, http://www.ebi.ac.uk/IPI/IPIhuman.html) using BIOWORKS software (version 3.3.1, Thermo Scientific). The following

ARTICLE

parameters were used during the database search with TurboSEQUEST: variable oxidation of methionine and variable carbamidomethylation of cysteine and selenocysteine residues. Peptide identifications were considered to be correct for matches with Xcorr scores above 1.5, 2.5, and 3.0 for +1, +2, and +3 peptide ions, respectively, and for peptide probability scores below 0.1. Peptide identifications found above these thresholds were verified manually.

’ RESULTS AND DISCUSSION Total Se and Selenium-Species Quantification by ID-ICPMS in SRM 1950. The results of the ID concentration determinations

for total selenium and selenium species in SRM 1950 are given with expanded uncertainties at the 95% confidence interval. The uncertainty in the Se value is calculated as an expanded uncertainty according to the ISO guide.16 For the total Se measurements, the measured mass fraction of Se in SRMs 1577c and 1598a control materials were within the uncertainty range of the certified value listed on the respective certificates of analysis. The total measured selenium value of SRM 1950 (105 ( 3.8 ng g 1) is significantly higher than the total Se value certified in the serum-based BCR-637 (81 ( 7 ng g 1, IRMM, Geel, Belgium). This significant difference in Se content could be attributed to the different origin of blood donors—U.S. donors for the SRM 1950 and Danish donors for BCR-637—as levels of Se in foods produced in different countries vary, and Se intakes in the United States are among the highest recorded in the world.24 The quantification of eluted Se species as Se by online IDMS is based on the conversion of the raw data chromatograms (counts/s vs time) into mass flow chromatograms using the approach proposed by Rottmann and Heumann.19 In this method, the signals measured for each isotope are ratioed point by point (77Se/80Se) for the chromatographic run. These ratios can be converted into the flow of detected Se (ng min 1) via the well-known published isotope dilution equation,19 and finally, integration of the Se peaks can be carried out. A representative Se mass flow chromatogram obtained from the analysis of a SRM 1950 sample using this analytical setup is shown in Figure 1. The observed features of the Se mass flow chromatogram are in agreement with those reported in previous publications18 with three selenium-containing fractions corresponding to the nonretained fraction (F1), the heparin Sepharose retained fraction (F2), and the blue Sepharose retained fraction (F3). The determined value and expanded uncertainty based on a 95% confidence level for the mass fraction of Se in F1, F2, and F3 for SRM 1950 are 23.6 ( 1.3, 50.2 ( 4.3, and 28.2 ( 2.6 ng g 1, respectively. The online IDMS approach provides accurate quantification for the Se species eluting from the double affinity columns but is unable to correct possible Se loss during chromatographic elution, i.e., incomplete chromatographic recovery. However, the sum of the selenium species for 1950 (102.0 ( 5.8 ng g 1) agrees with the total selenium value measured independently by ID-ICPMS (105 ( 3.8 ng g 1) which demonstrates the complete mass fraction balance of total Se in the sample. The presence of inorganic selenium or small organic selenium compounds in human plasma/serum at any appreciable concentration is not likely and has only been reported to occur following the storage of human serum samples at elevated temperatures.25 However, in order to ensure that all Se in SRM 1950 is 8670

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry

ARTICLE

Table 1. Selenium-Containing Proteins Identified by Shotgun Proteomic Analyses of F1, F2, and F3 SeleniumContaining Fractions Collected from Double-Affinity LC of SRM 1950a selenium-containing proteins albumin

fraction 1

fraction 2

fraction 3

21 unique

19 unique peptides

123 unique

peptides selenoprotein P

peptides 1 unique peptide

glutathione peroxidase 3 others (SelH, SelN, ...) total protein number

123

230

104

a

Figure 1. Mass flow chromatogram SRM 1950—metabolites in human plasma.

protein-bound, a 600 μL aliquot of SRM 1950 was filtered with a 3 kDa molecular weight cutoff filter and the filtrate was analyzed by reversed-phase LC ICPMS. No small selenium-containing compounds or inorganic selenium were found in the filtrate at a detectable level which confirmed that Se in SRM 1950 is bound to proteins with a molecular weight higher than 3 kDa. Shotgun Proteomic Analyses of Selenium-Containing Fractions from Double-Affinity LC of SRM 1950. In order to determine which selenium-containing proteins and/or selenoproteins are present in human plasma, shotgun proteomic analyses of crude SRM 1950 was first attempted. A total of 119 proteins were identified, none of which was identified as a selenoprotein. This confirmed that selenoproteins are low-abundant proteins and more specific methods need to be used for their identification. Shotgun proteomic analyses of F1, F2, F3 selenium-containing fractions of SRM 1950 were then performed. A total of 660 μL of SRM 1950 was injected by increments of 100 μL injections into the dual-affinity heparin Sepharose and blue Sepharose columns, as Se was monitored by ICPMS. The three seleniumcontaining fractions F1, F2, and F3, respectively, corresponding to the nonretained fraction, the heparin Sepharose retained fraction, and the blue Sepharose retained fraction were detected, collected, and digested with trypsin. A shotgun proteomic approach based on a “data-dependent acquisition” (DDA) strategy was applied in order to identify proteins in the three selenium-containing fractions. The first round of nanoLC MS/MS experiments successively analyzed peptide mixtures from F1, F2, and F3 and was followed by two more rounds under the same conditions to obtain triplicate analyses of each sample. Following this procedure, the acquired MS/MS spectra were analyzed by a spectral searching approach using NIST MSPepSearch search engine with a human protein spectral library (released on January 14, 2010). Only peptides with a NIST MS Search score above 450, giving an estimated false discovery rate (FDR) of 1%, were considered. An average of 123, 230, and 104 proteins were, respectively, identified in the three fractions, but only two selenium-containing proteins, albumin and SelP, were identified. Table 1 reports the average number of peptide spectrum matches (PSM) and unique peptides identified for albumin, SelP, and other seleniumcontaining proteins in these three Se fractions. Albumin was identified in all three selenium-containing fractions, but the highest numbers of PSM and unique peptides for

Tryptic digests of each fraction were analyzed by LC MS/MS, and average numbers of unique peptides identified for GPx3, SelP, albumin, and other selenium-containing proteins are reported.

albumin were found in F3. The number of unique albumin peptides identified in each replicate analysis of F3 was consistent with 123, 121, and 126 peptides identified in the three replicate analyses. Therefore, 95% of the albumin sequence was covered by the peptides identified in F3. The presence of albumin in F1 and F2 analyses may be attributed to a blue Sepharose affinity column leakage or to an oversaturation of the albumin binding sites in the affinity column. Contrary to the albumin which was identified by multiple peptides in all LC MS/MS analyses, SelP was determined in only one of the three LC MS/MS replicates analyses of F2 with a total of three unique peptides identified by only one PSM for each. Neither GPx3 nor any other selenoproteins (e.g., SelN, SelH, ...) was determined in any replicate LC MS/MS analyses of F1, F2, or F3. The spectral searching approach applied in this study used a human protein library with limited selenoprotein sequence coverage factors of 41.47% and 62.39% for SelP and GPx3 protein, respectively. Therefore, a sequence searching approach with Sequest search engine using a human protein library was attempted in order to identify SelP and GPx3 peptides that could not be identified during spectral searching analyses (not represented in spectral libraries). Sequence searching analyses of MS/MS spectra acquired by LC MS/MS analyses of F1, F2, or F3 proteins did not detect SelP and GPx3 or any other selenoproteins or selenium-containing proteins. In order to identify selenoproteins present in F1 and F2, improved purification techniques are needed. Sensitive Detection and Specific Identification of Human Plasma Selenoproteins. Identification of GPx3 and SelP Selenoproteins in Candidate Human Plasma SRM 1950 after DualAffinity LC ICPMS Analyses by LA-ICPMS Followed by MS/MS. In order to detect and identify low-abundant selenoproteins, a strategy based on laser ablation of proteins on gel blots with online detection of Se by ICPMS was recently developed.20 Accordingly, proteins in F1 and F2 were separated by gel electrophoresis, electroblotted onto a PVDF membrane, and Se was detected by triplicate LA-ICPMS analyses (Figure 2). A 23 kDa selenium-containing protein was detected in all triplicate analyses of F1. In F2, a selenium-containing protein at approximately 49 kDa was detected in all triplicate analyses. These results show the advantage of this LA-ICPMS strategy for the ultrasensitive detection of selenium-containing proteins at physiological concentrations. 8671

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry

ARTICLE

Figure 2. Detection and identification of selenoproteins in selenium-containing fractions F1 (diagram a) and F2 (diagram b) after double-affinity LC ICPMS analyses of SRM 1950. Left: Se detection in proteins electroblotted onto a PVDF membrane by LA-ICPMS analyses. Right: number of SelP or GPx3 peptide-spectrum matches (PSM) identified in the selenium-containing bands by LC MS/MS analyses. Peptides were identified by a full scan of ions in the m/z range of 300 2000 or by a scan of m/z of SelP or GPx3 peptide ions only.

Proteins in the selenium-containing bands from F1 and F2 were then identified by using a shotgun proteomic approach based on a DDA strategy. After in-gel tryptic digestions of the proteins, MS/MS analyses were performed. The DDA strategy applied consisted of detecting ions through a large 300 2000 m/z range and selecting the most abundant ions of the MS scan for fragmentation in a second round of MS analysis. Once MS/ MS spectra had been acquired, the next stage involved peptide spectrum matching (PSM) by spectral searching approach with NIST MSPepSearch against a human spectral protein database. Only peptides with a NIST MS Search score above 450, giving an estimated FDR of 1%, were considered. As a result, GPx3 was the only selenoprotein identified in the duplicate analyses of the 23 kDa selenium-containing band found in F1, with five GPx3 unique peptides identified represented by seven PSM (Figure 2a).

SelP was the only selenoprotein identified in the 49 kDa seleniumcontaining band found in F2. Eleven PSM corresponded to seven SelP unique peptides (Figure 2b). Although only a few PSM were attributed to GPx3 and SelP peptides, these shotgun proteomic experiments showed that GPx3 and SelP were the only selenoproteins in F1 and F2 with a GPx3 sequence protein coverage factor of 26.5% in F1 and a SelP sequence protein coverage factor of 17.7% in F2. In order to better characterize these human plasma SelP and GPx3, a modified DDA strategy was then utilized by limiting the ion detection in the first MS scan to predicted m/z for SelP and GPx3 peptides ions before their fragmentation in a second round of MS analysis. Reducing the m/z list scan allowed for a more rapid acquisition of GPx3 and SelP peptide ions per unit of time and resulted in a larger number of PSM for these peptides. 8672

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry

ARTICLE

Figure 3. Selenoprotein P AEENITESCQUR and ENLPSLCSUQGLR selenopeptides identified by sequence searching approach with the SEQUEST search engine and a human protein database including the alkylated cysteine and selenocysteine: (a) MS/MS spectrum assigned for AEENITESCQUR peptide; (b) MS/MS spectrum assigned for ENLPSLCSUQGLR peptide. The tables list the theoretical masses (charge = +1) of b and y fragment ions for AEENITESCQUR and ENLPSLCSUQGLR.

The GPx3 and SelP peptides identified in the selenium-containing bands are reported in the tables of Figure 2. In the 23 kDa band from F1, a total of 360 PSM corresponding to 9 GPx3 unique peptides were identified (Supporting Information Figure S-1), resulting in a GPx3 sequence coverage factor of 43.3%. In the 49 kDa Se band from F2, 115 PSM for seven SelP unique peptides were found (Supporting Information Figure S-1). Although 10 times more PSM were identified for SelP, the resulting SelP sequence coverage factor was 17.7%. Unequivocal determination of the selenoproteins as the Se species of human plasma can be reached if selenium-containing peptides (or selenopeptides) of these proteins are identified. Unfortunately, no SelP or GPx3 selenopeptides have been observed or documented up to this point. As a result, MS/MS spectra of these selenopeptides are not in the current human spectral library used, which hindered the use of spectral searching approach to identify selenopeptides. Sequence searching approaches can be promising alternatives in the identification of peptides not previously observed, such as SelP selenopeptides. Therefore, a sequence searching experiment was performed on the MS/MS spectra acquired during the modified DDA. The SEQUEST search engine was used with a human protein database including variable modifications such as oxidation of methionine and alkylation of cysteine and selenocysteine residues. Using this approach, two SelP selenopeptides, AEENITESCQUR and ENLPSLCSUQGLR where U represents the selenocysteyl residue, were identified (Figure 3). The identification of these two selenopeptides yielded a final SelP sequence coverage

factor of 24.4% in F2 and demonstrates unequivocally that selenoprotein P is a selenium-containing protein in F2. MS/ MS spectra of these selenopeptides will be added to the human spectral library to facilitate future studies on selenoprotein P and its Se isoforms.

’ CONCLUSIONS In this study, human plasma SRM 1950 was determined for its total Se content and for its selenium-species content. A dc-AFLC ICPMS strategy was applied to separate the human plasma Se species in three selenium-containing fractions (F1, F2, F3), in which Se was determined by ID-ICPMS measurements. Unambiguous identification of GPx3 in F1, SelP in F2, and Alb in F3 was achieved, whereas no other selenoproteins or seleniumcontaining compounds were identified. No inorganic Se or small Se molecules were detected after filtration of SRM 1950, and complete selenium recovery from the columns was obtained, as the sum of the selenium species for 1950 determined by dc-AFLC ID-ICPMS corresponds to the total selenium value measured independently by ID-ICPMS. Consequently, SRM 1950 is the first human plasma SRM with a measured total Se value of 105.5 ( 2.3 ng g 1 and mass fractions values of 23.6 ( 1.3 ng g 1 as Se for GPx3, 50.2 ( 4.3 ng g 1 as Se for SelP, and 28.2 ( 2.6 ng g 1 as Se for Alb. Recent experimental studies on “the optimization of plasma selenium biomarkers for the assessment of the selenium nutritional requirement” revealed that when the total plasma selenium 8673

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674

Analytical Chemistry concentration is superior to 90 ng g 1, selenoproteins are generally fully expressed (or optimized).9 Therefore, the total Se concentration of 105.5 ( 2.3 ng g 1 in SRM 1950 suggests that selenoproteins were optimized in this pool of human plasma collected from 100 healthy individuals representing the overall racial distribution of the U.S. population. It has already been suggested that selenoproteins are optimized in human plasma of residents of the United States because their dietary Se intake are generally superior to 80 μg/d.26 Consequently, it can be hypothesized that optimized values for selenoproteins are those measured in SRM 1950, with 23.6 ( 1.3 ng g 1 as Se for GPx3 and 50.2 ( 4.3 ng g 1 as Se for SelP, whereas any additional selenium is incorporated in Alb. In this respect, similar Se values for GPx3 (23 ( 10 ng g 1) and SelP (49 ( 15 ng g 1) have been recently reported along with a total Se concentration above 90 ng g 1 in a study analyzing the serum samples of 399 healthy Greek individuals.27 In future studies, other methods should be applied in order to certify selenoprotein values in SRM 1950 and to determine SelP isoforms. In this respect, target proteomic analyses based on the quantification of SelP or GPx3 proteotypic peptides will be of the utmost interest when isotopically enriched SelP and GPx3 are available.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ REFERENCES

ARTICLE

(13) Deagen, J. T.; Butler, J. A.; Zachara, B. A.; Whanger, P. D. Anal. Biochem. 1993, 208 (1), 176–181. (14) States, D. J.; Omenn, G. S.; Blackwell, T. W.; Fermin, D.; Eng, J.; Speicher, D. W.; Hanash, S. M. Nat. Biotechnol. 2006, 24 (3), 333–338. (15) Liu, X.; Valentine, S. J.; Plasencia, M. D.; Trimpin, S.; Naylor, S.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 2007, 18 (7), 1249–1264. (16) ISO. Guide to the Expression of Uncertainty in Measurement; International Organization for Standardization: Geneva, Switzerland, 1993. (17) Taylor, B. N.; Kuyatt, C. E. NIST Technical Note 1297: Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results; U.S. Government Printing Office: Washington, DC, 1994. (18) Jitaru, P.; Prete, M.; Cozzi, G.; Turetta, C.; Cairns, W.; Seraglia, R.; Traldi, P.; Cescon, P.; Barbante, C. J. Anal. At. Spectrom. 2008, 23 (3), 402–406. (19) Rottmann, L.; Heumann, K. G. Fresenius’ J. Anal. Chem. 1994, 350 (4 5), 221–227. (20) Ballihaut, G.; Kilpatrick, L. E.; Kilpatrick, E. L.; Davis, W. C. J. Anal. At. Spectrom. 2011, 26 (2), 383–394. (21) Stein, S.E., Rudnick, P.A., Eds. NIST Search Engines and Software. NIST MSPepSearch. National Institute of Standards and Technology: Gaithersburg, MD. http://peptide.nist.gov. (22) Stein, S.E., Rudnick, P.A., Eds. NIST Peptide Mass Spectral Libraries. Human Peptide Mass Spectral Reference Data, H. s., ion trap, Official Build Date: Feb. 4, 2009. National Institute of Standards and Technology, Gaithersburg, MD. http://peptide.nist.gov. (23) Stein, S.E., Rudnick, P.A., Eds. NIST Search Engines and Software. NIST MS Search 2.0 2008. National Institute of Standards and Technology, Gaithersburg, MD. http://peptide.nist.gov. (24) Reilly, C. Trends Food Sci. Technol. 1998, 9 (3), 114–118. (25) Palacios, O.; Lobinski, R. Talanta 2007, 71 (4), 1813–1816. (26) National Academy of Sciences, Institute of Medicine, Food and Nutrition Board. Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, and Carotenoids; National Academy Press: Washington, DC, 2000; pp 284 324. (27) Letsiou, S.; Lu, Y.; Nomikos, T.; Antonopoulou, S.; Panagiotakos, D.; Pitsavos, C.; Stefanadis, C.; Pergantis, S. A. Proteomics 2010, 10 (19), 3447–3457.

(1) Rayman, M. P. Lancet 2000, 356 (9225), 233–241. (2) Reeves, M. A.; Hoffmann, P. R. Cell. Mol. Life Sci. 2009, 66 (15), 2457–2478. (3) Rayman, M. P. Br. J. Nutr. 2008, 100 (2), 254–268. (4) Whanger, P. D. Br. J. Nutr. 2004, 91 (1), 11–28. (5) Reid, M. E.; Duffield-Lillico, A. J.; Sunga, A.; Fakih, M.; Alberts, D. S.; Marshall, J. R. Int. J. Cancer 2006, 118 (7), 1777–1781. (6) Lippman, S. M.; Klein, E. A.; Goodman, P. J.; Lucia, M. S.; Thompson, I. M.; Ford, L. G.; Parnes, H. L.; Minasian, L. M.; Gaziano, J. M.; Hartline, J. A.; Parsons, J. K.; Bearden, J. D., III; Crawford, E. D.; Goodman, G. E.; Claudio, J.; Winquist, E.; Cook, E. D.; Karp, D. D.; Walther, P.; Lieber, M. M.; Kristal, A. R.; Darke, A. K.; Arnold, K. B.; Ganz, P. A.; Santella, R. M.; Albanes, D.; Taylor, P. R.; Probstfield, J. L.; Jagpal, T. J.; Crowley, J. J.; Meyskens, F. L., Jr.; Baker, L. H.; Coltman, C. A., Jr. JAMA, J. Am. Med. Assoc. 2009, 301 (1), 39–51. (7) Atkinson, A. J., Jr.; Colburn, W. A.; DeGruttola, V. G.; DeMets, D. L.; Downing, G. J.; Hoth, D. F.; Oates, J. A.; Peck, C. C.; Schooley, R. T.; Spilker, B. A.; Woodcock, J.; Zeger, S. L. Clin. Pharmacol. Ther. 2001, 69 (3), 89–95. (8) Ashton, K.; Hooper, L.; Harvey, L. J.; Hurst, R.; Casgrain, A.; Fairweather-Tait, S. J. Am. J. Clin. Nutr. 2009, 89 (6), 2025S–2039S. (9) Xia, Y.; Hill, K. E.; Li, P.; Xu, J.; Zhou, D.; Motley, A. K.; Wang, L.; Byrne, D. W.; Burk, R. F. Am. J. Clin. Nutr. 2010, 92 (3), 525–531. (10) Bunk, D. M. Proteomics 2010, 10 (23), 4220–4225. (11) Hinojosa Reyes, L.; Marchante-Gayon, J. M.; García Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2003, 18 (10), 1210–1216. (12) Jitaru, P.; Roman, M.; Barbante, C.; Vaslin-Reimann, S.; Fisicaro, P. Accredit. Qual. Assur. 2010, 1–8. 8674

dx.doi.org/10.1021/ac2021147 |Anal. Chem. 2011, 83, 8667–8674