Proteomics 2014, 14, 2427–2431
2427
DOI 10.1002/pmic.201400199
TECHNICAL BRIEF
PAS-cal: A repetitive peptide sequence calibration standard for MALDI mass spectrometry ¨ 4,5 , Arne Skerra3,6 Stefan K. Maier1,2,3 , Ksenia Bashkueva1 , Christoph Rosli 1,3 and Bernhard Kuster 1
¨ Munchen, ¨ Chair for Proteomics and Bioanalytics, Technische Universitat Freising, Germany ¨ Institute of Pathology, Research Unit Analytical Pathology, Helmholtz Zentrum Munchen, Neuherberg, Germany 3 Munich Center for Integrated Protein Science, CIPS-M, Munich, Germany 4 HI-STEM gGmbH, German Cancer Research Center DKFZ, Heidelberg, Germany 5 Junior Research Group Biomarker Discovery, German Cancer Research Center, Heidelberg, Germany 6 ¨ Biologische Chemie, Technische Universitat ¨ Munchen, ¨ Lehrstuhl fur Freising-Weihenstephan, Germany 2
Mass spectrometers equipped with matrix-assisted laser desorption/ionization (MALDI-MS) require frequent multipoint calibration to obtain good mass accuracy over a wide mass range and across large numbers of samples. In this study, we introduce a new synthetic peptide mass calibration standard termed PAS-cal tailored for MALDI-MS based bottom-up proteomics. This standard consists of 30 peptides between 8 and 37 amino acids long and each constructed to contain repetitive sequences of Pro, Ala and Ser as well as one C-terminal arginine residue. MALDI spectra thus cover a mass range between 750 and 3200 m/z in MS mode and between 100 and 3200 m/z in MS/MS mode. Our results show that multipoint calibration of MS spectra using PAS-cal peptides compares well to current commercial reagents for protein identification by PMF. Calibration of tandem mass spectra from LC-MALDI experiments using the longest peptide, PAS-cal37, resulted in smaller fragment ion mass errors, more matching fragment ions and more protein and peptide identifications compared to commercial standards, making the PAS-cal standard generically useful for bottom-up proteomics.
Received: May 7, 2014 Revised: July 18, 2014 Accepted: August 14, 2014
Keywords: Calibration / LC-MALDI / MALDI-TOF-MS / PMF / Protein identification / Technology
Additional supporting information may be found in the online version of this article at the publisher’s web-site
MALDI [1] is widely used in the proteomics field. One of its main applications is protein identification either by PMF [2] or LC-MALDI-MS/MS [3, 4]. For such applications, accurate mass determination is highly important and requires careful mass calibration. Especially in MALDI-TOF-MS, many Corespondence: Dr. Bernhard Kuster, Chair for Proteomics and ¨ Munchen, ¨ Bioanalytics, Technische Universitat Freising, Germany and Munich Center for Integrated Protein Science, CIPS-M, Munich, Germany E-mail: [email protected] Fax: +49-8161-715931 Abbreviations: ACTH, adrenocorticotropic hormone; Fmoc, fluorenylmethyloxycarbonyl; HCCA, ␣-cyano-4-hydroxycinnamic acid; Lys, lysine; ppm, parts per million; PSM, peptide spectrum match
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
factors (e.g. temperature, spot morphology [5], and position [6, 7]) influence mass accuracy, which leads to the need for frequent calibration. A mass calibration standard for bottomup proteomic applications is therefore desirable and should meet a number of requirements: (i) cover the mass range of (tryptic) peptides, (ii) provide many calibration points with small mass increments to enable fitting higher order calibration functions (as employed for TOF analyzers [8,9]), (iii) yield robust intensities in MS and MS/MS, and (iv) have a molecular composition that is not easily confused with the analytes under investigation. These criteria are not fully met by widely used peptide calibration standards that typically encompass few naturally occurring peptides or nonpeptidic polymers such as poly(propyl-glycol), PPG [9]. Here, we describe the Colour online: See the article online to view Figs. 1–3 in colour.
www.proteomics-journal.com
2428
S. K. Maier et al.
Figure 1. Arginine-terminated PAS-cal peptides. The PAS repeat unit used in this study is indicated by the dotted lines.
design, synthesis, and application of a mass calibration standard termed PAS-cal that addresses the above needs. The design was inspired by the PASylation technology introduced by Schlapschy et al. [10] in which recombinant therapeutic proteins are furnished with multiple sequential copies of an unstructured repeat of Pro, Ala and Ser residues, e.g. [ASPAAPSAPPAA]n . Based on this sequence, we synthesized two sets of 30 “PAS-cal” tryptic-like peptides ranging from 8 to 37 amino acids in length with either an arginine (Fig. 1) or lysine (Lys; not shown) residue at the C-terminus to aid ionization during the MALDI process [11]. The PAS-cal peptides were produced by solid-phase synthesis following the standard fluorenylmethyloxycarbonyl (Fmoc) strategy on a parallel peptide synthesizer (MultiPep, Intavis, Cologne, Germany) at 2 mol scale. For each peptide, proline, alanine, or serine (Intavis) were initially coupled to a solid support containing an immobilized arginine (Arg) or Lys residue (TentaGel S Trt-Arg(Pbf)Fmoc, TentaGel S Trt-Lys(Boc)Fmoc, Rapp Polymere, T¨ubingen, Germany) followed by the sequential coupling of further amino acids. To prevent the generation of erroneous sequences, free amino groups were capped by acetylation after each synthesis cycle and prior to the next Fmoc deprotection step. After completion of the synthesis scheme, peptides were released from the resin using 92.5% TFA, 5% triisopropylsilane, and 2.5% water and subsequently lyophilized. The complete synthesis method for the synthesizer is provided in the Supporting Information. For synthesis control as well as for PMF (see below), PAS-cal peptides and mixtures thereof were dissolved in water to 2 mol/mL (estimated from syn C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Proteomics 2014, 14, 2427–2431
thesis scale) and mixed 1:1 with ␣-cyano-4-hydroxycinnamic acid (HCCA, 10 mg/mL in 50% acetonitrile, ACN, 0.1% TFA, Bruker Daltonik, Bremen, Germany). Samples were spotted onto a stainless steel target using 0.5 L per spot, which were recrystallized using 0.1 L ethanol, resulting in a homogeneous matrix layer. MALDI-TOF spectra were acquired in positive-ion reflectron mode on an ultrafleXtreme MALDITOF/TOF instrument (Bruker Daltonik, operated by FlexControl 3.3) by summing up to 1000 laser shots over a mass range of 500–3500 m/z. The cubic enhanced method (FlexControl) was used for the calibration of mass spectra. While Lys-terminated PAS-cal peptides tended to form sodium and potassium adducts (data not shown), arginine containing peptides were dominantly observed as [M+H]+ and, as described before [11], yielded higher intensities. For the arginine containing set of PAS-cal peptides, we investigated two applications: (i) the mass calibration of MALDI-TOF spectra for protein identification by PMF and (ii) the calibration of MALDI-MS/MS spectra for protein identification via fragment ion based database searching. For the first application, 30 arginine terminated PAS-cal peptides were mixed to yield a balanced distribution of intensities in a MALDI mass spectrum (Fig. 2A) covering the range from 756.39 to 3140.62 m/z. The peak spacing corresponds to the residue masses of one of the three amino acids that make up the PAS-cal peptides, thus allowing multipoint calibration of MALDI-MS spectra. To evaluate the PAS-cal standard as an external calibrant for PMF, digests of BSA and bovine cytochrome C (n = 5) as well as a peptide calibration standard II (all from Bruker Daltonik) were analyzed by MALDI-TOF-MS where the measured spectra were either calibrated on the PAS-cal standard or on the commercial Bruker standard (both spotted nearby) using the cubic enhanced calibration option embedded in flexControl, which showed superior performance compared to the other embedded calibration methods (linear, quadratic, and linear correction, data not shown). For PMF, ultrafleXtreme mass spectra were converted into peak lists using flexAnalysis (version 3.3, Bruker Daltonik). Resulting xml files (see Supporting Information) were submitted to Mascot (2.4.1 Matrix Science, London, UK) and searched (using the and elements from the xml files) against the SwissProt database (version 57, taxonomy filter set to “other mammals”) considering carbamidomethylation of cysteines and oxidation of methionines as variable modifications. Peptide mass tolerance was set to 75 ppm (parts per million) and full trypsin specificity was required, allowing up to two missed cleavages. Mascot search results showed that the performance of both standards was comparable in terms of number of detected peptides and sequence coverage but PAS-cal calibration exhibited slight improvements in terms of mass error and protein score (Fig. 2B and C; note that the Mascot score is a log10 score, thus a difference of 10 score points translates into 10x better confidence). For the second application, the calibration of MS/MS spectra, very similar requirements apply with respect to the www.proteomics-journal.com
Proteomics 2014, 14, 2427–2431
2429
Figure 2. Evaluation of PAS-cal peptides for protein identification by PMF. (A) MALD-TOF-MS spectrum of the PAS-cal peptide mixture (30 peptides). The unannotated peaks in the lower mass range correspond to by-products of the peptide synthesis, which accumulate due to the combination of the single crude peptides. (B) PMF search results for BSA and cytochrome c tryptic digests. Vertical lines indicate averages of five measurements. (C) Relative mass errors for all measurements including a 4th order polynomial trend line (black line) along with the 95% confidence interval (grey lines).
properties of calibration standards. One notable difference is that a single peptide which generates fragment ions covering the entire mass range may suffice to calibrate a tandem mass spectrum. PAS-cal peptides are ideally suited for this purpose as the C-terminal arginine residue promotes the formation of a strong y-ion series. The MALDI tandem mass spectrum of the longest PAS-cal peptide (PAS-cal37) is shown in Fig. 3A. Its appearance resembles that of the PAS-cal mixture because the singly charged fragment ions of PAS-cal37 show a similar amino acid spaced pattern as the protonated intact peptides from the peptide mixture. To evaluate the utility of PAS-cal37 (3140.61 Da, 50 calibration points) for the multipoint calibration of tandem mass spectra, we performed two LC-MALDI-MS/MS experiments using tryptic digests of Escherichia coli and compared the results with analogous duplicate experiments calibrated on the fragment ion spectra of adrenocorticotropic hormone (ACTH) fragment 18–39 (2464.19 Da, six calibration points), which is widely used as MS2 calibration standard. Escherichia coli peptides were separated by RP chromatography using an 80 min gradient (see C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Supporting Information for details). Eluting peptides were mixed online with HCCA (3 mg/mL HCCA (ProteoChem, Cheyenne, USA) in 70% ACN, 0.1% TFA) in water containing the standard peptides gonadoliberin, angiotensin I, neurotensin, and ACTH fragment 18–39, each at a concentration of 150 pmol/mL, and spotted onto an Opti-TOF LC/MALDI insert (AB SCIEX, Framingham, USA) using an automatic MALDI spotter (SunCollect, SunChrom, Friedrichsdorf, Germany). Twelve hundred spots were prepared per LC run, corresponding to 4 s fractions per spot. MALDI-MS and MS/MS measurements were carried out on a TOF/TOF 5800 System (AB SCIEX) controlled by TOF/TOF Series Explorer Software V4.1.0 (build 12). Per spot, 8 × 250 shots for MS spectra and up to 12 × 250 shots per MS/MS spectrum were summed up. MS spectra were internally calibrated using the four peptides spiked into the matrix solution. MS/MS spectra were externally calibrated using six calibration points from ACTH 18–39 or 50 calibration points of the 37 amino acid PAS-cal peptide (PAS-cal37). Tandem mass spectra from LC-MALDI-MS/MS experiments were processed by Protein
www.proteomics-journal.com
2430
S. K. Maier et al.
Proteomics 2014, 14, 2427–2431
Figure 3. Multipoint calibration of tandem mass spectra using PAS-cal37 and ACTH. (A) Tandem mass spectrum of PAS-cal37 exhibiting fragment ions covering the entire sequence. (B) Result summary of LC-MALDI-MS/MS measurements. Note that only those identifications were accepted that exceeded the Mascot peptide identity threshold of 30. (C) Distribution of absolute mass errors of matching fragment ions (see also Supporting Information Fig. 2). (D) Upper panel: number of matched fragments (within a tolerance of 0.15 Da) as a function of fragment ion mass (in bins of 50 Da). Middle panel: median mass error of matched fragments in ppm as a function of fragment ion mass. Lower panel: same as in (C) but in Da (see also Supporting Information Fig. 2).
Pilot (Version 4.5, ABSCIEX) employing the Paragon Algorithm 4.5.00, 1654. The resulting peak lists (confer Supporting Information, in Mascot generic format) were searched against the SwissProt database (v57, taxonomy filter E. coli using Mascot version 2.4.1) with a mass tolerance for precursors of 75 ppm and 0.15 Da for fragment ions. Enzyme specificity was set to trypsin, allowing a maximum of two missed cleavages while carbamidomethylation of cysteines C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
was set as fixed modification. Search results were exported from Mascot and further analyzed using Microsoft Excel 2010 and Tableau (tableausoftware, Seattle, US). Peptide spectrum matches (PSMs) with a score lower than the Mascot identity score of 30 were categorically removed prior to subsequent analysis. As shown in Fig. 3B, PAS-cal37 calibration led to improvements in all metrics applied, including an increase of PSMs
www.proteomics-journal.com
2431
Proteomics 2014, 14, 2427–2431
by 18% with a concomitant 15% increase in peptide identifications and 13% increase in protein identifications. In addition, the average and median Mascot scores were also slightly improved in the PAS-cal37 calibrated experiments (again note log 10 space of the Mascot ion score so that even small score differences of three points correspond to a considerable effect in absolute space). Further analysis of PSMs with ion scores of equal or greater than the Mascot identity threshold (of 30) revealed two factors that are responsible for the observed improvements when using PAS-cal37 for the calibration of tandem mass spectra. First, more fragment ions were matched within the search tolerance (Fig. 3B and Supporting Information Fig. 1) and, second, the median fragment ion mass error was smaller compared to ACTH calibration (Fig. 3D and Supporting Information Fig. 2). Obviously, the number of matching fragments drives the identification score and the lower mass error of the fragments aids in discriminating against alternative PSMs for the same tandem mass spectrum. In summary, we have developed a synthetic peptide calibration standard tailored to the needs of bottom-up proteomics, termed PAS-cal. This standard comprises 30 tryptic peptides composed of proline, alanine, and serine (PAS), each in conjunction with a C-terminal arginine residue. PAS-cal offers a number of advantageous features such as flexible choice of mass range and number of calibrants, narrow mass spacing in MS and MS/MS, and no interference with known protein sequences. However, application of the PAS-cal peptide standard to MALDI imaging experiments (on tissue or upon trypsin digestion) is likely not feasible as many peptides useful for calibration would also interfere with the detection of the endogenous peptide and protein signals. While we have shown its utility for calibrating MALDI-TOF-MS and MS/MS spectra using HCCA, the standard should be equally suitable for other matrices (e.g. 2,5-dihydroxybenzoic acid) as well as for other mass analyzers such as ion trap and orbitrap instruments. The ease and low cost of synthesis make PAS-cal peptides an economically viable alternative to commercially available standards. The authors have declared no conflict of interest.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
References [1] Karas, M., Bachmann, D., Bahr, U., Hillenkamp, F., Matrixassisted ultraviolet laser desorption of non-volatile compounds. Int. J. Mass Spectrom. Ion Process 1987, 78, 53–68. [2] Pappin, D. J. C., Hojrup, P., Bleasby, A. J., Rapid identification of proteins by peptide-mass fingerprinting. Curr. Biol. 1993, 3, 327–332. [3] Mann, M., Wilm, M., Error-tolerant identification of peptides in sequence databases by peptide sequence tags. Anal. Chem. 1994, 66, 4390–4399. [4] Kamada, H., Fugmann, T., Neri, D., Roesli, C., Improved protein sequence coverage by on resin deglycosylation and cysteine modification for biomarker discovery. Proteomics 2009, 9, 783–787. [5] Garden, R. W., Sweedler, J. V., Heterogeneity within MALDI samples as revealed by mass spectrometric imaging. Anal. Chem. 2000, 72, 30–36. [6] Egelhofer, V., Gobom, J., Seitz, H., Giavalisco, P. et al., Protein identification by MALDI-TOF-MS peptide mapping: a new strategy. Anal. Chem. 2002, 74, 1760–1771. [7] Duncan, M. W., Roder, H., Hunsucker, S. W., Quantitative matrix-assisted laser desorption/ionization mass spectrometry. Brief Funct. Genomic. Proteomic. 2008, 7, 355– 370. [8] Bantscheff, M., Duempelfeld, B., Kuster, B., An improved two-step calibration method for matrix-assisted laser desorption/ionization time-of-flight mass spectra for proteomics. Rapid Commun. Mass Spectrom. 2002, 16, 1892– 1895. [9] Gobom, J., Mueller, M., Egelhofer, V., Theiss, D. et al., A calibration method that simplifies and improves accurate determination of peptide molecular masses by MALDI-TOF MS. Anal. Chem. 2002, 74, 3915–3923. ¨ [10] Schlapschy, M., Binder, U., Borger, C., Theobald, I. et al., PASylation: a biological alternative to PEGylation for extending the plasma half-life of pharmaceutically active proteins. Protein Eng. Des. Sel. 2013, 26, 489–501. [11] Krause, E., Wenschuh, H., Jungblut, P. R., The dominance of arginine-containing peptides in MALDI-derived tryptic mass fingerprints of proteins. Anal. Chem. 1999, 71, 4160–4165.
www.proteomics-journal.com
2427
DOI 10.1002/pmic.201400199
TECHNICAL BRIEF
PAS-cal: A repetitive peptide sequence calibration standard for MALDI mass spectrometry ¨ 4,5 , Arne Skerra3,6 Stefan K. Maier1,2,3 , Ksenia Bashkueva1 , Christoph Rosli 1,3 and Bernhard Kuster 1
¨ Munchen, ¨ Chair for Proteomics and Bioanalytics, Technische Universitat Freising, Germany ¨ Institute of Pathology, Research Unit Analytical Pathology, Helmholtz Zentrum Munchen, Neuherberg, Germany 3 Munich Center for Integrated Protein Science, CIPS-M, Munich, Germany 4 HI-STEM gGmbH, German Cancer Research Center DKFZ, Heidelberg, Germany 5 Junior Research Group Biomarker Discovery, German Cancer Research Center, Heidelberg, Germany 6 ¨ Biologische Chemie, Technische Universitat ¨ Munchen, ¨ Lehrstuhl fur Freising-Weihenstephan, Germany 2
Mass spectrometers equipped with matrix-assisted laser desorption/ionization (MALDI-MS) require frequent multipoint calibration to obtain good mass accuracy over a wide mass range and across large numbers of samples. In this study, we introduce a new synthetic peptide mass calibration standard termed PAS-cal tailored for MALDI-MS based bottom-up proteomics. This standard consists of 30 peptides between 8 and 37 amino acids long and each constructed to contain repetitive sequences of Pro, Ala and Ser as well as one C-terminal arginine residue. MALDI spectra thus cover a mass range between 750 and 3200 m/z in MS mode and between 100 and 3200 m/z in MS/MS mode. Our results show that multipoint calibration of MS spectra using PAS-cal peptides compares well to current commercial reagents for protein identification by PMF. Calibration of tandem mass spectra from LC-MALDI experiments using the longest peptide, PAS-cal37, resulted in smaller fragment ion mass errors, more matching fragment ions and more protein and peptide identifications compared to commercial standards, making the PAS-cal standard generically useful for bottom-up proteomics.
Received: May 7, 2014 Revised: July 18, 2014 Accepted: August 14, 2014
Keywords: Calibration / LC-MALDI / MALDI-TOF-MS / PMF / Protein identification / Technology
Additional supporting information may be found in the online version of this article at the publisher’s web-site
MALDI [1] is widely used in the proteomics field. One of its main applications is protein identification either by PMF [2] or LC-MALDI-MS/MS [3, 4]. For such applications, accurate mass determination is highly important and requires careful mass calibration. Especially in MALDI-TOF-MS, many Corespondence: Dr. Bernhard Kuster, Chair for Proteomics and ¨ Munchen, ¨ Bioanalytics, Technische Universitat Freising, Germany and Munich Center for Integrated Protein Science, CIPS-M, Munich, Germany E-mail: [email protected] Fax: +49-8161-715931 Abbreviations: ACTH, adrenocorticotropic hormone; Fmoc, fluorenylmethyloxycarbonyl; HCCA, ␣-cyano-4-hydroxycinnamic acid; Lys, lysine; ppm, parts per million; PSM, peptide spectrum match
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
factors (e.g. temperature, spot morphology [5], and position [6, 7]) influence mass accuracy, which leads to the need for frequent calibration. A mass calibration standard for bottomup proteomic applications is therefore desirable and should meet a number of requirements: (i) cover the mass range of (tryptic) peptides, (ii) provide many calibration points with small mass increments to enable fitting higher order calibration functions (as employed for TOF analyzers [8,9]), (iii) yield robust intensities in MS and MS/MS, and (iv) have a molecular composition that is not easily confused with the analytes under investigation. These criteria are not fully met by widely used peptide calibration standards that typically encompass few naturally occurring peptides or nonpeptidic polymers such as poly(propyl-glycol), PPG [9]. Here, we describe the Colour online: See the article online to view Figs. 1–3 in colour.
www.proteomics-journal.com
2428
S. K. Maier et al.
Figure 1. Arginine-terminated PAS-cal peptides. The PAS repeat unit used in this study is indicated by the dotted lines.
design, synthesis, and application of a mass calibration standard termed PAS-cal that addresses the above needs. The design was inspired by the PASylation technology introduced by Schlapschy et al. [10] in which recombinant therapeutic proteins are furnished with multiple sequential copies of an unstructured repeat of Pro, Ala and Ser residues, e.g. [ASPAAPSAPPAA]n . Based on this sequence, we synthesized two sets of 30 “PAS-cal” tryptic-like peptides ranging from 8 to 37 amino acids in length with either an arginine (Fig. 1) or lysine (Lys; not shown) residue at the C-terminus to aid ionization during the MALDI process [11]. The PAS-cal peptides were produced by solid-phase synthesis following the standard fluorenylmethyloxycarbonyl (Fmoc) strategy on a parallel peptide synthesizer (MultiPep, Intavis, Cologne, Germany) at 2 mol scale. For each peptide, proline, alanine, or serine (Intavis) were initially coupled to a solid support containing an immobilized arginine (Arg) or Lys residue (TentaGel S Trt-Arg(Pbf)Fmoc, TentaGel S Trt-Lys(Boc)Fmoc, Rapp Polymere, T¨ubingen, Germany) followed by the sequential coupling of further amino acids. To prevent the generation of erroneous sequences, free amino groups were capped by acetylation after each synthesis cycle and prior to the next Fmoc deprotection step. After completion of the synthesis scheme, peptides were released from the resin using 92.5% TFA, 5% triisopropylsilane, and 2.5% water and subsequently lyophilized. The complete synthesis method for the synthesizer is provided in the Supporting Information. For synthesis control as well as for PMF (see below), PAS-cal peptides and mixtures thereof were dissolved in water to 2 mol/mL (estimated from syn C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Proteomics 2014, 14, 2427–2431
thesis scale) and mixed 1:1 with ␣-cyano-4-hydroxycinnamic acid (HCCA, 10 mg/mL in 50% acetonitrile, ACN, 0.1% TFA, Bruker Daltonik, Bremen, Germany). Samples were spotted onto a stainless steel target using 0.5 L per spot, which were recrystallized using 0.1 L ethanol, resulting in a homogeneous matrix layer. MALDI-TOF spectra were acquired in positive-ion reflectron mode on an ultrafleXtreme MALDITOF/TOF instrument (Bruker Daltonik, operated by FlexControl 3.3) by summing up to 1000 laser shots over a mass range of 500–3500 m/z. The cubic enhanced method (FlexControl) was used for the calibration of mass spectra. While Lys-terminated PAS-cal peptides tended to form sodium and potassium adducts (data not shown), arginine containing peptides were dominantly observed as [M+H]+ and, as described before [11], yielded higher intensities. For the arginine containing set of PAS-cal peptides, we investigated two applications: (i) the mass calibration of MALDI-TOF spectra for protein identification by PMF and (ii) the calibration of MALDI-MS/MS spectra for protein identification via fragment ion based database searching. For the first application, 30 arginine terminated PAS-cal peptides were mixed to yield a balanced distribution of intensities in a MALDI mass spectrum (Fig. 2A) covering the range from 756.39 to 3140.62 m/z. The peak spacing corresponds to the residue masses of one of the three amino acids that make up the PAS-cal peptides, thus allowing multipoint calibration of MALDI-MS spectra. To evaluate the PAS-cal standard as an external calibrant for PMF, digests of BSA and bovine cytochrome C (n = 5) as well as a peptide calibration standard II (all from Bruker Daltonik) were analyzed by MALDI-TOF-MS where the measured spectra were either calibrated on the PAS-cal standard or on the commercial Bruker standard (both spotted nearby) using the cubic enhanced calibration option embedded in flexControl, which showed superior performance compared to the other embedded calibration methods (linear, quadratic, and linear correction, data not shown). For PMF, ultrafleXtreme mass spectra were converted into peak lists using flexAnalysis (version 3.3, Bruker Daltonik). Resulting xml files (see Supporting Information) were submitted to Mascot (2.4.1 Matrix Science, London, UK) and searched (using the and elements from the xml files) against the SwissProt database (version 57, taxonomy filter set to “other mammals”) considering carbamidomethylation of cysteines and oxidation of methionines as variable modifications. Peptide mass tolerance was set to 75 ppm (parts per million) and full trypsin specificity was required, allowing up to two missed cleavages. Mascot search results showed that the performance of both standards was comparable in terms of number of detected peptides and sequence coverage but PAS-cal calibration exhibited slight improvements in terms of mass error and protein score (Fig. 2B and C; note that the Mascot score is a log10 score, thus a difference of 10 score points translates into 10x better confidence). For the second application, the calibration of MS/MS spectra, very similar requirements apply with respect to the www.proteomics-journal.com
Proteomics 2014, 14, 2427–2431
2429
Figure 2. Evaluation of PAS-cal peptides for protein identification by PMF. (A) MALD-TOF-MS spectrum of the PAS-cal peptide mixture (30 peptides). The unannotated peaks in the lower mass range correspond to by-products of the peptide synthesis, which accumulate due to the combination of the single crude peptides. (B) PMF search results for BSA and cytochrome c tryptic digests. Vertical lines indicate averages of five measurements. (C) Relative mass errors for all measurements including a 4th order polynomial trend line (black line) along with the 95% confidence interval (grey lines).
properties of calibration standards. One notable difference is that a single peptide which generates fragment ions covering the entire mass range may suffice to calibrate a tandem mass spectrum. PAS-cal peptides are ideally suited for this purpose as the C-terminal arginine residue promotes the formation of a strong y-ion series. The MALDI tandem mass spectrum of the longest PAS-cal peptide (PAS-cal37) is shown in Fig. 3A. Its appearance resembles that of the PAS-cal mixture because the singly charged fragment ions of PAS-cal37 show a similar amino acid spaced pattern as the protonated intact peptides from the peptide mixture. To evaluate the utility of PAS-cal37 (3140.61 Da, 50 calibration points) for the multipoint calibration of tandem mass spectra, we performed two LC-MALDI-MS/MS experiments using tryptic digests of Escherichia coli and compared the results with analogous duplicate experiments calibrated on the fragment ion spectra of adrenocorticotropic hormone (ACTH) fragment 18–39 (2464.19 Da, six calibration points), which is widely used as MS2 calibration standard. Escherichia coli peptides were separated by RP chromatography using an 80 min gradient (see C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Supporting Information for details). Eluting peptides were mixed online with HCCA (3 mg/mL HCCA (ProteoChem, Cheyenne, USA) in 70% ACN, 0.1% TFA) in water containing the standard peptides gonadoliberin, angiotensin I, neurotensin, and ACTH fragment 18–39, each at a concentration of 150 pmol/mL, and spotted onto an Opti-TOF LC/MALDI insert (AB SCIEX, Framingham, USA) using an automatic MALDI spotter (SunCollect, SunChrom, Friedrichsdorf, Germany). Twelve hundred spots were prepared per LC run, corresponding to 4 s fractions per spot. MALDI-MS and MS/MS measurements were carried out on a TOF/TOF 5800 System (AB SCIEX) controlled by TOF/TOF Series Explorer Software V4.1.0 (build 12). Per spot, 8 × 250 shots for MS spectra and up to 12 × 250 shots per MS/MS spectrum were summed up. MS spectra were internally calibrated using the four peptides spiked into the matrix solution. MS/MS spectra were externally calibrated using six calibration points from ACTH 18–39 or 50 calibration points of the 37 amino acid PAS-cal peptide (PAS-cal37). Tandem mass spectra from LC-MALDI-MS/MS experiments were processed by Protein
www.proteomics-journal.com
2430
S. K. Maier et al.
Proteomics 2014, 14, 2427–2431
Figure 3. Multipoint calibration of tandem mass spectra using PAS-cal37 and ACTH. (A) Tandem mass spectrum of PAS-cal37 exhibiting fragment ions covering the entire sequence. (B) Result summary of LC-MALDI-MS/MS measurements. Note that only those identifications were accepted that exceeded the Mascot peptide identity threshold of 30. (C) Distribution of absolute mass errors of matching fragment ions (see also Supporting Information Fig. 2). (D) Upper panel: number of matched fragments (within a tolerance of 0.15 Da) as a function of fragment ion mass (in bins of 50 Da). Middle panel: median mass error of matched fragments in ppm as a function of fragment ion mass. Lower panel: same as in (C) but in Da (see also Supporting Information Fig. 2).
Pilot (Version 4.5, ABSCIEX) employing the Paragon Algorithm 4.5.00, 1654. The resulting peak lists (confer Supporting Information, in Mascot generic format) were searched against the SwissProt database (v57, taxonomy filter E. coli using Mascot version 2.4.1) with a mass tolerance for precursors of 75 ppm and 0.15 Da for fragment ions. Enzyme specificity was set to trypsin, allowing a maximum of two missed cleavages while carbamidomethylation of cysteines C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
was set as fixed modification. Search results were exported from Mascot and further analyzed using Microsoft Excel 2010 and Tableau (tableausoftware, Seattle, US). Peptide spectrum matches (PSMs) with a score lower than the Mascot identity score of 30 were categorically removed prior to subsequent analysis. As shown in Fig. 3B, PAS-cal37 calibration led to improvements in all metrics applied, including an increase of PSMs
www.proteomics-journal.com
2431
Proteomics 2014, 14, 2427–2431
by 18% with a concomitant 15% increase in peptide identifications and 13% increase in protein identifications. In addition, the average and median Mascot scores were also slightly improved in the PAS-cal37 calibrated experiments (again note log 10 space of the Mascot ion score so that even small score differences of three points correspond to a considerable effect in absolute space). Further analysis of PSMs with ion scores of equal or greater than the Mascot identity threshold (of 30) revealed two factors that are responsible for the observed improvements when using PAS-cal37 for the calibration of tandem mass spectra. First, more fragment ions were matched within the search tolerance (Fig. 3B and Supporting Information Fig. 1) and, second, the median fragment ion mass error was smaller compared to ACTH calibration (Fig. 3D and Supporting Information Fig. 2). Obviously, the number of matching fragments drives the identification score and the lower mass error of the fragments aids in discriminating against alternative PSMs for the same tandem mass spectrum. In summary, we have developed a synthetic peptide calibration standard tailored to the needs of bottom-up proteomics, termed PAS-cal. This standard comprises 30 tryptic peptides composed of proline, alanine, and serine (PAS), each in conjunction with a C-terminal arginine residue. PAS-cal offers a number of advantageous features such as flexible choice of mass range and number of calibrants, narrow mass spacing in MS and MS/MS, and no interference with known protein sequences. However, application of the PAS-cal peptide standard to MALDI imaging experiments (on tissue or upon trypsin digestion) is likely not feasible as many peptides useful for calibration would also interfere with the detection of the endogenous peptide and protein signals. While we have shown its utility for calibrating MALDI-TOF-MS and MS/MS spectra using HCCA, the standard should be equally suitable for other matrices (e.g. 2,5-dihydroxybenzoic acid) as well as for other mass analyzers such as ion trap and orbitrap instruments. The ease and low cost of synthesis make PAS-cal peptides an economically viable alternative to commercially available standards. The authors have declared no conflict of interest.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
References [1] Karas, M., Bachmann, D., Bahr, U., Hillenkamp, F., Matrixassisted ultraviolet laser desorption of non-volatile compounds. Int. J. Mass Spectrom. Ion Process 1987, 78, 53–68. [2] Pappin, D. J. C., Hojrup, P., Bleasby, A. J., Rapid identification of proteins by peptide-mass fingerprinting. Curr. Biol. 1993, 3, 327–332. [3] Mann, M., Wilm, M., Error-tolerant identification of peptides in sequence databases by peptide sequence tags. Anal. Chem. 1994, 66, 4390–4399. [4] Kamada, H., Fugmann, T., Neri, D., Roesli, C., Improved protein sequence coverage by on resin deglycosylation and cysteine modification for biomarker discovery. Proteomics 2009, 9, 783–787. [5] Garden, R. W., Sweedler, J. V., Heterogeneity within MALDI samples as revealed by mass spectrometric imaging. Anal. Chem. 2000, 72, 30–36. [6] Egelhofer, V., Gobom, J., Seitz, H., Giavalisco, P. et al., Protein identification by MALDI-TOF-MS peptide mapping: a new strategy. Anal. Chem. 2002, 74, 1760–1771. [7] Duncan, M. W., Roder, H., Hunsucker, S. W., Quantitative matrix-assisted laser desorption/ionization mass spectrometry. Brief Funct. Genomic. Proteomic. 2008, 7, 355– 370. [8] Bantscheff, M., Duempelfeld, B., Kuster, B., An improved two-step calibration method for matrix-assisted laser desorption/ionization time-of-flight mass spectra for proteomics. Rapid Commun. Mass Spectrom. 2002, 16, 1892– 1895. [9] Gobom, J., Mueller, M., Egelhofer, V., Theiss, D. et al., A calibration method that simplifies and improves accurate determination of peptide molecular masses by MALDI-TOF MS. Anal. Chem. 2002, 74, 3915–3923. ¨ [10] Schlapschy, M., Binder, U., Borger, C., Theobald, I. et al., PASylation: a biological alternative to PEGylation for extending the plasma half-life of pharmaceutically active proteins. Protein Eng. Des. Sel. 2013, 26, 489–501. [11] Krause, E., Wenschuh, H., Jungblut, P. R., The dominance of arginine-containing peptides in MALDI-derived tryptic mass fingerprints of proteins. Anal. Chem. 1999, 71, 4160–4165.
www.proteomics-journal.com
