diDO-IPTL: A Peptide-Labeling Strategy for Precision Quantitative

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

diDO-IPTL: A peptide-labeling strategy for precision quantitative proteomics Jacob Waldbauer, Lichun Zhang, Adriana Rizzo, and Daniel Muratore Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b02752 • Publication Date (Web): 10 Oct 2017 Downloaded from http://pubs.acs.org on October 12, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

diDO-IPTL: A peptide-labeling strategy for precision quantitative proteomics Jacob Waldbauer*, Lichun Zhang, Adriana Rizzoa and Daniel Muratoreb Department of the Geophysical Sciences, University of Chicago, 5734 S Ellis Ave, Chicago, Illinois 60637, United States * Corresponding Author: [email protected]

a

Current address: Department of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, United States b Current address: Program in Quantitative Biosciences, Georgia Institute of Technology, Atlanta, Georgia 30332, United States

1 of 18

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

ABSTRACT: We present an analytical strategy, dimethylation-deuteration and oxygenexchange IPTL (diDO-IPTL), for high-precision, broad-coverage quantitative proteomics. The diDO-IPTL approach combines two advances in isobaric peptide terminal labeling (IPTL) methodology: first, a one-pot chemical labeling strategy for attaching isotopic tags to both the Nand C-termini of tryptic peptides, and second, a search engine (based on the Morpheus algorithm) optimized for identification and quantification of twinned peaks from peptide fragment ions in MS2 spectra. The diDO-IPTL labeling chemistry uses only high-purity, relatively inexpensive isotopic reagents (18O water and deuterated formaldehyde) and requires no postlabeling cleanup or isotopic impurity corrections. This strategy produces proteome-scale relative quantification results with high accuracy and precision, suitable for the detection of small protein abundance variations between complex biological samples. In a two-proteome mixture experiment, diDO-IPTL quantification discriminates 1.5-fold changes in abundance of over 1000 proteins with 88% accuracy. The diDO-IPTL methodology is a high-precision, economical approach to quantitative proteomics that is applicable to a wide variety of sample types. Chemical derivatization of peptides with isotopically-labeled tags has become a popular approach in quantitative proteomics. In vitro isotope labeling strategies are especially useful in biological systems that are not amenable to metabolic isotope incorporation, and for analysis of environmental or clinical proteome samples. Chemical isotopic labeling of peptides can be achieved by, for example, alkylation of cysteine thiols1,2 or peptide primary amines3,4, or enzymatically catalyzed oxygen isotope exchange at C-terminal carboxyl groups5,6; quantitation using these labels is generally done on the intact peptide ions at the MS1 level. Currently, the most widely used in vitro peptide isotopic labeling technique uses commercially produced, amine-reactive isobaric tags (TMT7 [Proteome Sciences/Thermo 2 of 18


Page 3 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scientific] and iTRAQ8 [SCIEX]). These tags contain a moiety that fragments under collisional-dissociation MSn analysis, producing a set of reporter ions in the MSn scan whose intensities are used to calculate relative abundances between multiple samples. Since their introduction, reporter-ion tagging reagents have been applied to a broad range of biological systems. Despite its many successes, reporter ion-based quantitative proteomics has some important caveats. Reporter ion-based tagging reagents are structurally complex, and their production requires sophisticated and expensive isotopic syntheses. Recently, N,N-dimethyl leucine (DiLeu) tagging reagents have been described whose synthesis is relatively simpler and cheaper9 but whether lab-synthesized or commercial, such fragmenting isobaric tags inevitably contain isotopic impurities and require batch-dependent correction factors be applied to raw reporter ion intensities in order to recover accurate quantitation data. Perhaps the most salient difficulty in reporter ion-based relative quantitation is the issue of interferences from co-eluting peptides and fragment ions, which require multistage MS acquisition and/or further empirical corrections be applied to generate accurate results and improve precision10-14. Hence, while reporter ion-based tagging methods are well-suited to many quantitative proteomics studies, alternatives that reduce susceptibility to interferences, complexity of mass spectrometry and quantitative corrections, and cost are desirable in many circumstances. Isobaric peptide terminal labeling (IPTL), first described by Koehler et al.15, differs from reporter ion-based labeling strategies, both in the chemistry used to achieve labeling and in how quantitative information is represented in peptide fragmentation spectra. In IPTL, peptides are labeled on both N- and C-termini, with peptides from one sample receiving a ‘heavy’ (neutron-enriched) tag on the N-terminus and a ‘light’ tag on the C-terminus, and those from another sample receiving the ‘light’ tag on the N-terminus and the ‘heavy’ tag on the C-terminus; two distinct isotopic tagging reagents are required. A key difference between IPTL labeling and other types of isobaric tagging is that, rather than relying on the fragmentation of the tag to produce a reporter ion that is detected in a separate region of the MSn spectrum from the main peptide fragment ions, each main peptide fragment (b- or y-ion) is itself a quantitative reporter ion. Hence, concerns about co-isolation of ions that interfere with reporters, or biases in sensitivity or resolution in different m/z ranges, are greatly diminished. 3 of 18



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 22

Despite attractive features as a proteomic quantification technique, several factors have limited the wider application of IPTL. One is the requirement for isotopic labeling of peptide C-termini, which are difficult to distinguish chemically from side-chain carboxyl groups of aspartate and glutamate residues. This necessitated the use of the LysC protease in many prior IPTL studies, as that enzyme yields peptides with chemically distinct ε-amino groups adjacent to their C-termini, which can be derivatized separately from N-terminal αamines using O-alkyl isourea reagents17. The quantitation by isobaric terminal labeling (QITL) technique18 demonstrated IPTL labeling of tryptic peptides using C-terminal oxygen isotope exchange and N-terminal reductive dimethylation, but also requires separate guanidination of lysine side-chains as well as multiple intermediate buffer-exchange and purification steps. Perhaps more critically, IPTL labeling produces peptide fragmentation spectra that look fundamentally different from those of unlabeled or TMT/iTRAQ-tagged peptides, and which violate the ‘one-spectrum-one-peptide’ assumption that underlies most peptide-spectrum matching algorithms. Each b- and y-ion fragment in an IPTL spectrum is present as a twin pair of peaks, one of which derives from each labeled variant. In essence, each IPTL fragmentation spectrum is a chimera of the two peptide isotopologues. As a result, few computational tools have emerged to recover peptide sequences and quantifications from IPTL spectra; two examples include IsobariQ19 and ITMSQ20, both of which rely on peptide identification results from standard database search algorithms (Mascot and SEQUEST, respectively) and so perform quantification subsequent to, rather than in conjunction with, identification. The Paired Ions Scoring Algorithm21 adapted the open-source peptidespectrum matching engine Morpheus22 to identify and quantify peptides from the twinned b- and y-ion peaks in IPTL fragmentation spectra, but that software is limited to analysis of LysC-derived, N- and C-terminal dimethylated peptides. Here we report a methodology for IPTL peptide labeling and data analysis that addresses many of the technical issues that have limited wider adoption of this technique, and offers distinct advantages over other isobaric tagging methods. Like the QITL method18, the labeling strategy is a combination of two well-established chemical tagging techniques in quantitative proteomics, reductive dimethylation and C-terminal enzymatic oxygen isotope exchange. We have further optimized this chemistry for yielding correctly-labeled IPTL 4 of 18


Page 5 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


peptides in a single-pot reaction, without time-consuming and loss-prone intermediate purification steps or the need to protect lysine side-chains by guanidination. This labeling scheme uses only simple isotopic substrates, no complex syntheses or impurity corrections are required, and the total labeling cost per sample is only a fraction of that of commercial isobaric tagging reagents. We present a peptide-spectrum matching and quantification engine, MorpheusFromAnotherPlace, that enables straightforward identification of peptides from IPTL spectra, and accurate quantification of their relative abundances. We term this quantitative proteomics strategy diDO-IPTL, for dimethylation-deuteration and oxygenexchange IPTL.

Methods Protein extraction and peptide preparation Proteins were extracted by heating bacterial cell pellets (Desulfovibrio vulgaris Hildenborough or Rhodopseudomonas palustris TIE1, ~2E9 cells) to 95 ˚C for 20 minutes in a denaturing and reducing extraction buffer (1% SDS, 10% glycerol, 10 mM dithiothreitol, 200 mM Tris, pH 8). Cysteines were alkylated by addition of 40 mM iodoacetamide and incubation in the dark for 30 minutes. Where not otherwise specified, all solid reagents were dissolved in LC/MS-grade water (Fisher Optima). Proteins were purified by a modified eFASP (enhanced filter-aided sample preparation) protocol23, using Vivacon 500 concentrators (30 kDa nominal cutoff, Sartorius). Proteins were digested with MS-grade trypsin (Thermo Pierce) at 37˚C overnight, and peptides were eluted from the concentrator and dried by vacuum centrifugation. Peptide aliquots were quantified with the Quantitative Colorometric Peptide Assay (Thermo Pierce). Peptide terminal isotope labeling For C-terminal oxygen isotope exchange, per sample to be labeled, 2 µg of freeze dried MSgrade trypsin (Thermo Pierce) is reconstituted in 40 µl of buffer containing acetic acid (glacial LC/MS-grade; Fisher Optima), N-methylmorpholine (99+%, Acros) and water, either H218O (98.5 atom%

18

O; Rotem Inc.) or H216O (99.99 atom%

16

O; Sigma) in a ratio of 1.0

acetic acid : 2.0 N-methylmorpholine : 97.0 water (vol/vol/vol). The pH of the trypsin exchange buffer is 7.4. Dried peptide samples (~10-20 µg) are redissolved in the 40 µl of trypsin solution and incubated at 37 ˚C overnight. We chose buffer components that are readily available in pure liquid form so that small batches of O isotope-exchange buffer could be prepared without the need to precisely weigh and dissolve solid reagents in 5 of 18



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 22

excessive quantities of isotopically enriched water. The use of pure N-methylmorpholine and acetic acid also avoids isotopic dilution of the highly-enriched water with exchangeable oxygen from buffer components. For N-terminal dimethylation, 2 µl of 11.3 M monochloroacetic acid (Sigma) is added to each sample to lower the pH to 2.6. Then 2 µl of 16% formaldehyde is added — either CH2O (Thermo Pierce) to samples previously incubated with H218O or CD2O (CDN Isotopes) to those incubated with H216O. For reductive alkylation, 2 µl of 4.8 M sodium cyanoborohydride (Sigma) is added and the samples incubated 1 hour at 45 ˚C. The reaction is halted by addition of 2 µl of 5 M ammonium formate (LC/MS-grade; Fisher Optima), and then the samples are acidified by addition of 8 µl of 100% formic acid (LC/MSgrade; Fisher Optima). The samples are then analyzed by LC-MS without further purification (e.g. micro-SPE/ZipTip or vacuum drying). Final total volume of the labeled, MS-ready sample is approximately 56 µl. The full labeling protocol is also available at protocols.io (www.protocols.io/view/iptl-new-version-oct-19-2015-d2i8cd). We refer to peptides labeled with the combination of

16

O2 on the C-termini and dideuterated formaldehyde on the N-

termini as “16O labeled” and those with

18

O2 on the C-termini and unlabeled formaldehyde

on the N-termini as “18O labeled”. LC-MS/MS Labeled peptide samples (6 µl of reaction product, or 1-2 µg of peptide) were injected onto a trapping column (OptiPak C18, Optimize Technologies) and separated on a capillary C18 column (Thermo Acclaim PepMap 100 Å, 2 µm particles, 50 µm I.D. × 50 cm length) using a water-acetonitrile + 0.1% formic acid gradient (2-50% acetonitrile over 180 min) at 90 nl/min using a Dionex Ultimate 3000 nanoLC system. Peptides were ionized by a nanoelectrospray source (Proxeon Nanospray Flex) fitted with a metal-coated fused silica emitter (New Objective). Mass spectra were collected on an Orbitrap Elite mass spectrometer (Thermo) operating in a data-dependent acquisition (DDA) mode, with one high-resolution (120,000 m/∆m) MS1 parent ion full scan triggering 15 MS2 Rapid mode CID fragment ion scans of selected precursors; the duty cycle for spectral acquisition under these conditions was approximately 2 s. Only multiply-charged parent ions were selected for fragmentation, and dynamic exclusion was enabled with a duration of 25 s and an exclusion window of ±15 ppm.

6 of 18


Page 7 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Data deposition Raw data, peak list files (.mgf format), spectrum-level output from MorpheusFromAnotherPlace and postprocessed protein-level output have been deposited in the MassIVE repository under accession MSV000080457. Data analysis Raw Orbitrap MS data are converted to Mascot Generic Format (.mgf) by Proteome Discoverer Daemon (Thermo). Spectra in the resultant MGF files are matched to a sequence database and isotopologue abundance ratios quantified by a modified version of the opensource peptide-spectrum matching engine Morpheus22, using elements of the Paired Ion Scoring Algorithm21 previously published for analysis of LysC-derived, N- and C-terminal dimethylated IPTL data. We have dubbed our modified implementation of the Morpheus/PISA algorithms MorpheusFromAnotherPlace (MFAP; see SI text and Fig S-1); source code, executables, and parameter files are available at GitHub (github.com/WaldbauerLab). We have optimized the Morpheus scoring function for quantitative accuracy and precision as well as FDR (via target-decoy database searching) in analysis of IPTL spectra. The Morpheus scoring function implemented in MFAP is:

where M is the Morpheus score, I refers to intensity in the fragment ion spectrum, N refers to numbers of fragment ions in the spectrum, and the subscripts T, 16 and 18 refer to total ions and those matched to

16

O- and

18

O-versions of the peptide, respectively. Spectrum-

level FDR is calculated as the number of decoy (reversed) database hits divided by the number of target database hits above a given Morpheus score threshold. The abundance ratio between

16

O- and

18

O-labeled versions of a peptide is calculated as the

ratio of the summed intensities of fragment ions matched to each version (i.e., I18/I16). For database searching with MFAP, N-terminal dimethylation (both d4- and d0-) and C-terminal 18

O2 substitution, as well as Met oxidation, are included as dynamic modifications. Precursor

ion tolerance is set to ±25ppm, fragment ion tolerance to ±0.6Da, and the most-intense 200 peaks retained from each MS2 spectrum. MFAP produces spectrum-level output including peptide identifications, target-decoy-based false discovery rates (q-values), and log2transformed abundance ratios for each identified IPTL spectrum. PSM-level MorpheusFromAnotherPlace output is then postprocessed by an R script that finds 7 of 18



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 22

breakpoints in the q-value curve and suggests Morpheus score cutoff thresholds to the user (see SI text and Fig S-2; script also available at GitHub link above). In general, spectrum-level FDRs in our diDO-IPTL datasets are

diDO-IPTL: A Peptide-Labeling Strategy for Precision Quantitative

Recommend Documents