Nonisotopic Reagents for a Cost-Effective Increase in Sample

Article pubs.acs.org/ac

Nonisotopic Reagents for a Cost-Effective Increase in Sample Throughput of Targeted Quantitative Proteomics Mary Joan Castillo,† Adam J. McShane,† Min Cai,† Yuanyuan Shen,† Lei Wang,† and Xudong Yao*,†,‡ †

Department of Chemistry and ‡Institute for Systems Genomics, University of Connecticut, Storrs, Connecticut 06269, United States

Downloaded by GEORGETOWN UNIV on August 27, 2015 | http://pubs.acs.org Publication Date (Web): August 26, 2015 | doi: 10.1021/acs.analchem.5b01727

S Supporting Information *

ABSTRACT: The new technology of ultrathroughput MS (uMS) transforms the intrinsic capability of analyte multiplexing in mass spectrometry (MS) to sample multiplexing. Core technological advantages of uMS rely on the decoupled use of isotopic quantitation reference and nonisotopic mass coding of samples. These advantages include: (1) high sample-throughput potential, (2) utilization of minimal amounts of expensive stable isotopes for the quantitation reference, and (3) unleashing of the opensource exploration of the chemical structure diversity of nonisotopic reagents to significantly enhance the MS detectability of analytes. A particular uMS method, ultrathroughput multiple reaction monitoring (uMRM), is reported for one-experiment quantitation of a surrogate peptide (SVILLGR) of prostate specific antigen (PSA) in multiple serum samples. Following derivatization of the pair of spiked, isotopic reference (SVILLGR*) and endogenous, native peptide in each sample, all samples were pooled for a step of simultaneous enrichment and cleanup of derivatized peptide pairs using immobilized antibody. The MS analysis of the pooled sample reported the quantity and sample origin of the surrogate peptide. Several analyses with different sample throughput were presented, with the highest being 15-in-1. Screening of nonisotopic reagents used combinatorial libraries of peptidyl compounds, and the reagent selection was based on the derivatization effectiveness and the capability of MS signal enhancement for the peptide. The precision, accuracy, and linearity of the uMRM MS technology were found to be comparable with standard isotope dilution MRM MS.

R

or selected reaction monitoring (SRM) MS is the method of choice and has high analyte-multiplexing capability for quantifying target molecules. Broad utilization of MS for protein biomarker validation is, however, limited by the low sample throughput of current methods and generally undercompetitiveness of the method sensitivity compared to that for immunoassays. Strategies that address these disadvantages will unleash the full potential of contemporary MS for fast development of protein biomarkers. MS, combined with front-end separation techniques, is intrinsically capable of examining large numbers of analytes in a single experiment. These analytes can be either different molecules in a particular sample (analyte multiplexing) or common molecules that carry sample-specific mass tags

ecent advances in mass spectrometry (MS) open an intriguing route for the rapid development and application of disease-relevant proteins as biomarkers. Validation of protein biomarkers involves the rapid analysis of candidates in a clinically significant, large data set of patient−control samples.1,2 Therefore, the validation requires high samplethroughput strategies. This is currently accomplished by immunoassays in high-throughput formats. However, these assays rely on the slow and costly development of highly specific antibodies, and the resulting assays are necessarily nonflexible and only applicable to particular protein targets. More recently, MS-based methods have been rapidly developed as versatile and relatively low-cost approaches for quantifying proteins.3 MS quantitation of proteins using either derivatization-free or derivatization-based approaches has become increasingly important in the analysis of complex biological samples.4 In the protein biomarker pipeline, candidates identified through comparative proteomics are subsequently targeted for validation. Multiple reaction monitoring (MRM) © XXXX American Chemical Society

Received: April 28, 2015 Accepted: August 20, 2015

A

DOI: 10.1021/acs.analchem.5b01727 Anal. Chem. XXXX, XXX, XXX−XXX

Article


Analytical Chemistry

a particular demonstration for quantifying prostate specific antigen (PSA) in multiple serum samples; this particular workflow is termed as ultrathroughput MRM MS (uMRM MS).17 The use of “ultrathroughput” was introduced to distinguish uMRM MS from existing strategies that increase the sample throughput via speeding the analysis of each sample. The new technology provides a potential, economical path to surpass the current sample-throughput potential afforded by isotopic reagents for biomarker applications.

(sample multiplexing). In the latter case, mass tagging is achieved via differentially labeling analytes with stable isotopes in a sample-specific manner; the isotope incorporation can be metabolic, enzymatic, or chemical.4 The relative MS intensity of the differentially tagged molecules allows for quantitation. In the protein biomarker regime, the analytes are peptides that result from proteolytic treatment of human proteome samples, e.g., sera, and these peptides are used as quantitation surrogates. Sample-specific tagging of common peptides in different samples can be achieved by chemical derivatization of peptides using isotopic reagents. There are collections of commercial and research-grade isotopic reagents in the toolbox of MSbased quantitative proteomics. Quantitation of the isotopelabeled peptides are carried out by single-stage MS analyses using mass-difference tagging reagents or using isobaric tagging chemicals for tandem MS (the so-called MS2 or MS3) experiments to monitor quantitation reporter ions produced from differentially tagged peptides. Multiplexing samples for a single experiment, or increasing the sample throughput for MS analysis, is continuously being pursued. Utilization of the mass defect of peptide fragment ions5 for multiplexed proteomic quantitation becomes feasible using isotopologues of commercial reagents with contemporary high-resolution mass spectrometers. Small differences in the mass defect of quantitation reporter ions can now be readily resolved.6 This approach has allowed multiplexing of 10−12 samples.7−10 A common challenge for the isobaric mass tagging strategy, however, originates from the coselection of precursor ions that have close retention time and mass-to-charge ratios. This coselection leads to compressed ratios of the reporter ions,11 because all peptides in a particular sample are derivatized with the same reagent. Methods of MS3-based quantitation12 and gas-phase purification13 have been proposed to address this issue, but at the expense of relatively reduced sensitivity.14 Further increases in the sample throughput of MS analysis can exploit a modular principle for workflow design, through the combined use of the isotopic quantitation reference and peptide derivatization. Hyperplexed proteomic quantitation of 18 samples, using isotopic reagents, has been presented.15,16 Also reported was a proof-of-concept demonstration of one-experiment quantitation of 25 samples using nonisotopic reagents for peptide derivatization.17 Sequential labeling of peptides with dual sets of isotopic reagents provides yet another route to increase the sample throughput for MS analysis, allowing for multiplexing of 12−16 samples.18,19 However, it is expensive to use isotopic reagents for multiplexing analysis of large numbers of samples. The cost of analysis surges more than proportionally as the sample number escalates. The structural diversity of isotopic reagents that can be economically synthesized is also limited. Furthermore, quantitation of low-abundance proteins in sera requires large sample volumes in order to enable MS analysis with desired quantitation limits; it becomes cost-prohibitive to use isotopic reagents to derivatize peptides in the digests of human sera in large volumes (e.g., 100 μL) and large numbers (100s to 1000s). Following the principle of modular design, we propose a novel MS methodology for quantifying proteins in many (N) proteome samples in a single (1) experiment (thus, N-in-1 analysis; broadly termed as ultrathroughput MS or uMS). A key technological feature is the use of structurally diverse, nonisotopic reagents to provide the needed sample-throughput potential for MS quantitation of surrogate peptides. We report

■

EXPERIMENTAL SECTION Dimethylacetamide (DMAc), methanol (MeOH), 4-(4,6dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM), N-methylmorpholine (NMM), 1,4-dimethylpiperazine (DMPipZ), urea, ammonium acetate (NH4OAc), iodoacetamide (IAA), and dimethyl pimelimidate (DMP) were purchased from Sigma-Aldrich (St. Louis, MO). Trifluoroacetic acid (TFA), acetonitrile (ACN), formic acid (FA), and DL-dithiothreitol (DTT) were purchased from Fisher (Pittsburgh, PA). Sequence-grade trypsin was purchased from Roche (Indianapolis, IN). Standard synthetic peptides were from Peptide 2.0 (Chantilly, VA), AnaSpec (Fremont, CA), and Thermo Fisher (Pittsburgh, PA). Female serum was from BioChemed (Winchester, VA). Human prostate specific antigen was purchased from Lee Biosolutions (St. Louis, MO). Anti-SVI antibody CON-1 was developed and purchased from Epitomics (Burlingame, CA). Dynabeads Protein G was from Life Technologies (Grand Island, NY). Direct-Q3 water system (Millipore, Billerica, MA) was used to purify deionized water. Samples were dried using a SpeedVac (Savant, Farmingdale, NY), vacuum oven (Fisher Scientific, Pittsburgh, PA), or lyophilizer (Labconco, Kansas City, MO). Incubation of samples was performed on a Hula Mixer (Invitrogen, Grand Island, NY) or a Thermomixer R (Eppendorf, Hauppauge, NY). Activation of Peptidyl Reagents and Derivatization of Signature Peptides for High-Signal Yield Screening. Peptidyl reagents were synthesized via standard solid-phase peptide synthesis and N-terminal capping to generate CapGrAAn...AA2AA1 (Supporting Information). Each activated peptidyl reagent was prepared by adding 1 μmol/11 μL DMTMM in DMAc, 9 μL of NMM/DMAc [1:80 (v/v)], and 20 μL of DMAc in a tube containing a dried aliquot of 2.5 μmol of peptidyl reagent. The resulting solution was incubated for 1 h at room temperature (RT). Immediately, 30 μL of the activated reagent was transferred into another tube with 5 μL of DMPipZ/DMAc [1:10 (v/v)]. A 12 μL aliquot from the resulting solution was added to the lyophilized standard peptide aliquot. The standard peptide mixture contained 1 nmol each of CFTR signature peptide NSILTETLHR (CFTR01) and PSA signature peptides SVILLGR (SVI), HSQPWQVLVASR (HSQ), and LSEPAELTDAVK. In addition to the main signature peptides, the standard mixture consisted of synthetic peptides YGGFLR, LSEAVTLK, IVGGWEK, NSILTETLR, HSTETLR, and SVIGGR. The reaction was incubated overnight at RT, quenched with 12 μL of ice-cooled 20% FA for 30 min, and then diluted to 75 μL with H2O. A 5 μL aliquot was transferred into another tube containing 85 μL of H2O, 5 μL of SVI*, and 5 μL of HSQ* (isotopic HSQ, not analyzed) solution. The samples were then analyzed using LC-TOF-MS and LC-MRM-MS (Supporting Information). The same protocol was performed to screen reagents for CFTR01. B


Article



Scheme 1. Workflow for the uMS Technologya

a

(a) General design; (b) detailed sample preparation steps for the workflow of N-in-1 uMRM MS quantifying PSA in multiple nondepleted serum samples. In contrast to isotopic reagents that exploit the principle of integrated design, the quantitation module is separated from the sample-coding module in the N-in-1 uMS workflow (left-hand box). The quantitation module refers to the addition of isotopic standard as quantitation reference, and the sample-coding module refers to the peptide derivatization. This partition between the addition of internal quantitation reference and derivatization-based sample coding presents methodological advantages and enables large-scale applications of MS methods for protein biomarker validation.

for 1 h at RT. Immediately, 8 μL of DMPipZ/DMAc [1:10 (v/ v)] was added into each activated reagent and mixed. The activated peptidyl reagent solution was then added to the lyophilized digest and incubated overnight at RT. The reaction was quenched with 58 μL of ice-cooled 20% FA for 30 min and concentrated in vacuo for 30 min. Each sample was then diluted to 1.5 mL with H2O and subsequently desalted. Eluted samples were combined and concentrated in vacuo and then by lyophilizer, until the total volume was about 300 μL. Single peptide preparation was performed on the pooled sample using 0.75 mg of beads (as below). The sample was reconstituted using 0.1% TFA and incubated at RT for 30 min. Finally, the sample was centrifuged at 13.2 × 1000 rpm for 30 min at 4 °C and analyzed in nanoLC-MRM MS (Supporting Information). Affinity-Based Peptide Preparation. The polyclonal antiSVI CON-1 antipeptide antibody (10 μg) was immobilized on Dynabeads Protein G (0.38 mg) as per the manufacturer’s procedure for 30 min. After binding, cross-linking was performed by adding 10 mM DMP and incubated for 30 min. The beads were then washed for unbound antibody and stored at 4 °C. For the affinity-based preparation of N-in-1 digests, 0.75 mg of Dynabeads Protein G with cross-linked antiSVI was used for each sample pool. The sample was ensured to have a pH of 8 (using PBS buffer-2% (v/v) MeOH) and incubated for 2 h. Nonspecific binders were then washed off, and derivatized SVI and SVI* bearing sample-specific tags were eluted with 5% acetic acid. The samples were diluted with H2O,

Preparation of N-in-1 uMS Samples. For recovery and precision (9-in-1) analysis, digests of 100 μL of nondepleted female serum with 300 ng/mL PSA protein were spiked with SVI* (theoretically 1:1 SVI/SVI*) and analyzed by MS to determine the recovery of SVI after digestion. See Supporting Information for digestion of serum samples. Aliquots were then prepared and spiked with a corrected SVI* amount to obtain a 1:1 ratio, followed by sample-specific derivatization and single affinity-based preparation of the derivatized peptides. For linearity (5-in-1) analysis, digests of 1/10 diluted female serum were spiked with SVI and SVI* to prepare a calibration curve with concentrations containing 1, 2.5, 10, 30, and 60 ng/mL SVI spiked with an equivalent of 8 ng/mL SVI*. Samplespecific derivatization and single affinity-based preparation of the derivatized peptides were then performed. For quantitation validation (15-in-1) analysis, digests of female serum were spiked with SVI and SVI* (the ratio was 4:1), reconstituted, and diluted 10 times. Concentrations containing 1, 2.5, 5, 7.5, 10, 20, 40, and 60 ng/mL SVI were prepared. For each concentration level, two different peptidyl reagents were used for sample-specific derivatization, followed by single affinitybased enrichment. Sample-Specific Derivatization. Each sample was assigned a unique peptidyl reagent for sample-specific derivatization. The activated reagent was prepared by adding 1 μmol/11 μL DMTMM in DMAc, 9 μL of NMM/DMAc [1:80 (v/v)], and 20 μL of DMAc in a tube containing dried aliquot of 2.5 μmol of peptidyl reagent. The resulting solution was incubated C


Article

Analytical Chemistry centrifuged for 30 min at 4 °C, transferred to another tube to lyophilize, and stored prior to MS analysis.

diversity is left wide open. Virtually any nonisotopic, reactive chemical becomes a potential candidate for derivatizing peptides, and the selection of a particular structure is at the user’s choice. Many criteria can be used for the reagent selection, such as ease of synthesis, scalability, or commercial availability. The most significant criterion, however, is the capability of a reagent to improve the MS quantitation limit for a target peptide. While peptides cannot be amplified like nucleic acids via the polymerase chain reaction, the evidence that chemical derivatization of peptides can increase MS signals of the molecules has long been recognized.25−27 The practical utilization of this signal enhancement strategy, however, has been limited by costly synthesis of stable isotopic versions of signal-enhancing reagents. This is a disadvantageous consequence of the integrated reagent design. The use of reagent isotopologues for peptide derivatization is the current practice in MS-based quantitative proteomics. Nonisotopic reagents have been excluded so far, because different reagents have different separation and MS ionization properties which thus prevent accurate quantitation. Design and Preparation of In-House Peptidyl Reagent Library. N-Terminally capped short peptides were synthesized, with a general format of CapGr-AAn...AA2AA1 where n = 1, 2, 3, or 4 (Figure S1a and Supplementary Methods, for the reagent preparation). Only alanine, glycine, leucine, phenylalanine, and valine were used to make the reagents; the peptide length was limited to a maximum of four residues. The N-termini of peptidyl reagents were capped with seven different groups through either acylation [acetyl, propionyl, picolinoyl, and 4(diethylamino)benzoyl] or reductive dialkylation [dimethyl, pyrrolidin-1-yl (the product of glutaraldehyde), and isoindolin2-yl (the product of phthalaldehyde)], using similar procedures to previous work.5 The split-and-pool synthesis technique was used to generate small libraries of peptidyl reagents on the small scale, starting with 1 g of a mixture of trityl-Cl polystyrene resin preloaded with the first amino acid (equal moles); each library contained 6−12 reagents. Enhancement of MS Signal Yield through Peptide Derivatization. The reagent libraries were screened to develop reagents that could enhance the MS detection of SVI and another model peptide CFTR01 (NSILTETLHR; its isotopic reference was CFTR01*, NSILTET[L-13C615N]HR). Peptide CFTR01 is a surrogate peptide for cystic fibrosis transmembrane conductance regulator protein. Deficiency in plasma membrane expression and function of the protein results in diminished chloride ion transport, consequently developing into the pathological progression of cystic fibrosis.28 Peptidyl reagents were activated in situ using DMTMM/ NMM mediation (Scheme S1) forming triazine esters before derivatizing a synthetic peptide mixture (Figure S1b). The mixture of CFTR01 (1 nmol) and SVI (1 nmol), together with eight other synthetic peptides constituting a total of 20 nmol of peptidyl amine groups, was derivatized with 2.5 μmol (theoretical) of each of the reagent mixtures. The extent of derivatization was quantitatively determined by two parameters: (1) the percentage of chemical conversion (PCC) for peptides and (2) the signal yield for mass spectrometry (SYMS) that compares the MS signal for derivatized peptides with that for the corresponding nonderivatized peptides. Experimentally, isotopic counterparts SVI* and CFTR01* were spiked for SYMS measurements, at known concentrations and after quenching the derivatization (eq S1). A particular compound


■

RESULTS AND DISCUSSION Modular Design of uMS and Workflow for N-in-1 uMRM Quantitation of PSA in Nondepleted Sera. In contrast to isotopic reagents that exploit the principle of integrated design, the N-in-1 uMS workflow is signified by the decoupled use of isotopic quantitation reference and nonisotopic reagents for mass codings (Scheme 1a). The quantitation module refers to the addition of isotopic standard as quantitation reference, and the sample-coding module refers to the peptide derivatization. This partition between the quantitation reference and mass coding brings significant methodological advantages. The most outstanding advantage is the cost savings. In the technology of uMRM MS, the quantitation reference is the isotopically labeled counterpart of a surrogate peptide target; samples are spiked with a common quantitation reference that carries stable isotopes (the right-hand box of Scheme 1a). It is important to note that the isotopic quantitation reference is added only at comparable concentrations (e.g., at the equivalence to picograms or low nanograms of proteins per milliliter; a minute cost concern) with those of the endogenous peptide surrogate. Thus, only low amounts of isotopic quantitation reference (and thus decreased use of expensive stable isotopes) are needed. This is highly cost-effective. If stable isotopes were added to samples through peptide derivatization, the quantity of isotopes used would be many orders of magnitude higher, considering that high abundance proteins (and thus their resulting peptides) are many orders of magnitude higher in moles20 and these peptides also consume the expensive reagents. In this study, the unique peptide SVILLGR, denoted as SVI, is the quantitation surrogate for model biomarker PSA. An isotopically labeled counterpart, denoted as SVI* (SVIL[L-13C615N]GR), was custom-made as the common quantitation reference (Scheme 1b). Relative to the frequently used PSA surrogate, peptide LSEPAELTDAVK,21,22 the MS signal of SVILLGR was found to be about 40%. Further cost savings come from the use of inexpensive, nonisotopic reagents for sample-specific mass coding to improve the sample throughput. This removes the cost barrier involved in using integrated, isotopic reagents for proteomic quantitation. This economical advertence augments when serum samples of large volumes (e.g., 100 μL) are analyzed. To compete with highly sensitive immunoassays,23,24 the use of relatively large amounts of samples for quantitation with yet-tobe-more-sensitive MS instruments currently is a practical solution. For instance, if serum volume was increased from 10 to 100 μL, then the limit of quantitation could be increased by a maximal factor of 10. However, it is not economically feasible to use stable isotopic reagents to derivatize peptides from large volumes of sera. Total protein content in 100 μL of serum is around 8 mg, and their peptides would use tens of milligrams of reagents. A commercial set of isotopic reagents with six labeling states and 4 mg for each reagent can cost over $2,500 (as of May 2015); each is only sufficient for derivatizing peptides from 10 μL of nondepleted serum. Another important advantage is the open-source development of chemicals for derivatizing peptides to achieve enhanced MS detection. Without the need for considering stable isotopes in the reagent design, the door to utilizing the chemical D


Article


chromatography for the further reagent validation. First, analysis with derivatized PSA protein digests (2−184 ng/mL) with four different peptidyl reagents was performed. The resulting uMRM data for the surrogate SVI was found to be linear (Figure S4). Validation in biological samples utilized N-in-1 uMRM MS of digests of PSA-spiked sera. In a validation experiment for 16 reagent candidates, 16 aliquots of 100 μL of nondepleted sera were spiked with PSA at a concentration of 300 ng/mL and digested by trypsin. To each of the resulting digests, the same amount (5 pmol, experimentally determined in order to obtain similar MS signal intensities for both SVI and SVI*) of SVI* was added. Following the uMRM analysis workflow (Scheme 1b), 9 reagents resulted in strong MRM signals for the derivatized SVI and SVI* (Figure 2). This experiment was also


in the reagent library passed the initial screening, when its PCC was >90% and SYMS >100%. These two criteria were set to ensure that not only a reagent could effectively derivatize peptides (measured by PCC), but also importantly, the added steps of sample derivatization and cleanup did not lead to a detectability loss (measured by SYMS). Passing reagents had varying residue sequences and capping groups. They produced a wide range of SYMS values for the derivatized peptides (Figure 1a), up to 100-fold for CFTR01

Figure 2. 9-in-1 uMRM MS analysis. Ion chromatograms for quantitation of PSA in nondepleted sera. Overlapping peaks correspond to SVI and SVI* derivatized by nine peptidyl reagents listed. Dashed lines correspond to SVI while solid lines correspond to SVI* (isotopic quantitation reference). Figure 1. Screening for peptidyl reagents producing high signal yields. (a) Peptidyl reagents producing varying SYMS; (b) SYMS against aliphatic index (AI) for acetylated (Ac) and dimethylated (Dim) reagents.

designed to assess the quantitation precision for the uMRM workflow. In the derivatization of 9 successful reagents, the average ratio of CapGr-AAn...AA2AA1-SVI to CapGr-AAn...AA2AA1-SVI* was calculated to be 1.70 and the coefficient of variation (CV) to be 11.5% (Table S1). This precision was comparable to our previous proof-of-concept study (the CV was 9.2%) on a simple protein digest.17 MS analysis of serum digests is prone to complex matrix effects. Thus, the effect of derivatization reagents with different chemical structures on the chromatographic elution was carefully considered. A particular pair of derivatized surrogate and reference peptides coelute (Figure 2), thus experiencing the same matrix effect for ionization of both of the peptides. For some reagents wherein multiple peaks were observed in chromatograms, the total peak area was used for quantitation. The multiple peaks (Figure 2) could be attributed to the presence of isomers in reagents, which were in-house synthesized at low milligram scales. Nevertheless, this complication did not prevent accurate quantitation of SVI, because the signal for derivatized SVI was normalized by the concurrently derivatized reference, SVI*. This is yet another advantage of the decoupled used of isotopic reference and nonisotopic mass coding reagents, compared to integrated,

and 50-fold for SVI (data not shown). To investigate if there is a structural determinant controlling the MS signal enhancement, SYMS values for derivatized SVI peptides were plotted against their aliphatic index (AI) values29 (the acetyl capping group on the derivatizing reagents was not included in the AI calculation). A general decrease in SYMS was observed with the AI increase (Figure 1b and Figure S2). However, a similar effect was not apparent with reagents capped with a dimethyl group (Figure 1b). Dimethylated reagents were generally among the compounds with higher SYMS values. Among successful reagents derivatizing SVI (180 reagent candidates) and CFTR01 (102 reagent candidates), a common preference on the residue sequence of reagents was not obvious for the MS signal enhancement (Figure S3). A representative tandem MS spectrum of derivatized CFTR01 was shown in Figure S1c. Evaluation of Elution-Specific Matrix Effect on Surrogate-Reference Pairs of Derivatized Peptides. Upon selection of reagents with high SYMS values, individual reagents were prepared and purified by high performance liquid E


Article



sample complexity, (1) the chemical complexity caused by peptide derivatization; the antibody reagent maximally removes interfering chemicals (e.g., large molar excess of reagents) from the derivatization mixtures. The chemical interference deleterious to MS analysis compromises the limit of quantitation for derivatization-based quantitative proteomics. (2) Derivatized nontarget peptides are also removed in the uMRM workflow, similar to affinity-based MRM MS. The other is to enrich targeted peptides, in order to compensate for the analyte dilution caused by pooling multiple samples. The uMRM workflow pools N numbers of samples together for a single MS experiment, which in turn results in Ntimes dilution of peptide targets. For instance, when the samples were combined in the 15-in-1 uMRM MS analysis discussed below (Figure 3), analytes in each sample were

isotopic derivatization reagents. It is interesting to note that, due to their inherent hydrophobicity differences, derivatized SVI peptides have different retention times, spreading over the reversed-phase gradient. The elution order for the derivatized peptides was more or less predictable when retention time calculators were used (Table S2). In comparison, peptides derivatized with isotopic reagents (excluding certain deuterium labeled) of a given chemical structure coelute. The high quantitation precision is another advantageous consequence of the modular design of the uMRM MS technology. The decoupled use of isotopic quantitation references and peptide mass coding allows for the passage of authentic quantitative information in original samples.15−17 The common quantitation reference SVI* was added before the derivatization, and this practice mitigated quantitation problems associated with quantitative proteomics that are based on isotopic derivatization of peptides.4 In such a multiplexing experiment, each sample is separately derivatized with one reagent from a set of isotopologues, which brings in an added source for variations in accuracy and precision. It is unavoidable to have incomplete derivatization and side reactions, but the target SVI and the quantitation reference SVI* would experience the same degree of the derivatization imperfection. Additionally, the uMRM technology can also be used in pair with isotopic proteins as quantitation references, which can further eliminate quantitation variations originated from protein-level sample preparations.28 Another investigation examined the practicality of using reagent candidates for the uMRM MS quantitation of PSA at concentrations that are clinically relevant.21,30 This investigation used aliquots of a master digest, and each aliquot had an absolute amount of SVI that was equivalent to 1 ng of PSA and a relative concentration of a mixture of 1 ng/mL PSA in 1/ 10 nondepleted female serum (further details in the footnote for Figure S5). Seven out of the 14 reagent candidates successfully quantified equivalents of 600 pg of PSA on column using a 4000 QTRAP (an earlier generation of triple quadrupole mass spectrometers), shown in Figure S5. This experiment allowed for the evaluation of the variability when reagents were changed. At a low concentration of SVI (equivalent to 1 ng of PSA) in the serum digest matrix, successful reagents were identified and quantitation of SVI using these reagents gave a CV of 12.0% (Table S3). Utilization Comparison of Antibody Reagents in Immunoassays, Affinity-Based MRM MS, and uMRM MS. The experimental realization of the signal-enhancing potential of derivatization reagents required a critical step of sample preparation: the cleanup and enrichment of derivatized SVI and SVI* by antibodies. After sample-specific derivatization of serum digests, all samples were pooled for a single, affinitybased preparation at the peptide level (Scheme 1b). Custommade polyclonal antibodies against the C-terminus of peptide SVI (noting that the peptide derivatization occurs at the Ntermini of the peptides) were immobilized on magnetic Protein G beads. In contrast, immunoassays use monoclonal antibodies with very high analyte specificity. Compared with common methods of affinity-based MRM MS like the technology of stable isotope standard capture by antipeptide antibody, they use one antibody reagent for each sample.1,2,20,31 Therefore, the use of a single antibody reagent in the uMRM workflow is less expensive and procedurally simpler. Two unique utilities of the affinity-based preparation step are essential to the uMRM MS technology. One is to reduce the

Figure 3. 15-in-1 uMRM MS analysis. (a) Ion chromatograms of derivatized SVI and SVI* at clinically relevant concentrations (dashed lines correspond to SVI while solid lines correspond to SVI*). Serum digests were spiked with SVI to mimic PSA (1 to 60 ng/mL) in a proof-of-concept N-in-1 linearity experiment. The samples were derivatized by the following reagents: Pr-ALA and Pr-ALG (1 ng/ mL), Pr-VLGG and Pr-LGG (2.5 ng/mL), Pr-GLGG and Pr-GLV (5 ng/mL), Pr-VLA and Pr-GGA (7.5 ng/mL), Dim-GLVV and Pr-GFA (10 ng/mL), Dim-FVAA and Pr-GFG (20 ng/mL), Dim-LGF and PicGGG (40 ng/mL), and Dim-GAG and Pic-GGA (60 ng/mL). DimGLVV derivatized SVI was not detected. (b) Low-concentration plot for SVI constructed by a single, 15-in-1 uMRM MS. F



Analytical Chemistry subsequently diluted by 15 times. The original sample volume ranged from 67 to 4000 μL, constituting a total volume of about 15 mL. Using a single step of affinity preparation, the derivatized SVI and SVI* were reconstituted in a final volume of 10 μL; therefore, the concentration of endogenous SVI for each sample was increased by 6.7 to 400 times. Linearity Investigation of the uMRM MS Quantitation. Linearity of quantitation was investigated through a 5-point curve with concentrations of 1 to 60 ng/mL SVI in 1/10 diluted serum digests spiked with a constant SVI* amount of 8 ng/mL, resulting in an R2 value of 0.9923 (Figure S6). To examine quantitation variations in analysis of a higher number of samples in this concentration range, a 15-in-1 uMRM MS analysis was performed (Figure 3a). SVI and SVI* (theoretically 1:1) were spiked at equivalent concentrations of 1 to 60 ng/mL SVI, which are clinically relevant. The average ratio of SVI to SVI* was found to be 0.92 (±0.09), while the overall CV was 10.3% for the replicate sample preparations. When plotted, an R2 of 0.9986 was obtained (Figure 3b), wherein two samples for each concentration level were derivatized by two different peptidyl reagents and averaged. The CV was determined for three replicate injections of two different reagents and ultimately averaged for all samples. Only one sample spiked at 10 ng/mL was included in the plot as the other derivatized sample was not detected (the initial experiment was designed to be 16-in-1 MS).

ACKNOWLEDGMENTS

■

REFERENCES

(1) Kuhn, E.; Whiteaker, J. R.; Mani, D. R.; Jackson, A. M.; Zhao, L.; Pope, M. E.; Smith, D.; Rivera, K. D.; Anderson, N. L.; Skates, S. J.; Pearson, T. W.; Paulovich, A. G.; Carr, S. A. Mol. Cell. Proteomics 2012, 11, M111.013854. (2) Kennedy, J. J.; Abbatiello, S. E.; Kim, K.; Yan, P.; Whiteaker, J. R.; Lin, C.; Kim, J. S.; Zhang, Y.; Wang, X.; Ivey, R. G.; Zhao, L.; Min, H.; Lee, Y.; Yu, M. H.; Yang, E. G.; Lee, C.; Wang, P.; Rodriguez, H.; Kim, Y.; Carr, S. A.; Paulovich, A. G. Nat. Methods 2014, 11, 149−155. (3) Carr, S. A.; Anderson, L. Clin. Chem. 2008, 54, 1749−1752. (4) Yao, X. Anal. Chem. 2011, 83, 4427. (5) Shi, Y.; Bajrami, B.; Yao, X. Anal. Chem. 2009, 81, 6438. (6) McAlister, G. C.; Huttlin, E. L.; Haas, W.; Ting, L.; Jedrychowski, M. P.; Rogers, J. C.; Kuhn, K.; Pike, I.; Grothe, R. A.; Blethrow, J. D.; Gygi, S. P. Anal. Chem. 2012, 84, 7469−7478. (7) Hebert, A. S.; Merrill, A. E.; Bailey, D. J.; Still, A. J.; Westphall, M. S.; Strieter, E. R.; Pagliarini, D. J.; Coon, J. J. Nat. Methods 2013, 10, 332. (8) Werner, T.; Sweetman, G.; Savitski, M. F. l.; Mathieson, T.; Bantscheff, M.; Savitski, M. M. Anal. Chem. 2014, 86, 3594. (9) Erickson, B. K.; Jedrychowski, M. P.; McAlister, G. C.; Everley, R. A.; Kunz, R. C.; Gygi, S. P. Anal. Chem. 2015, 87, 1241. (10) Frost, D. C.; Greer, T.; Li, L. Anal. Chem. 2015, 87, 1646−1654. (11) Gan, C. S.; Chong, P. K.; Pham, T. K.; Wright, P. C. J. Proteome Res. 2007, 6, 821. (12) Ting, L.; Rad, R.; Gygi, S. P.; Haas, W. Nat. Methods 2011, 8, 937. (13) Wenger, C. D.; Lee, M. V.; Hebert, A. S.; McAlister, G. C.; Phanstiel, D. H.; Westphall, M. S.; Coon, J. J. Nat. Methods 2011, 8, 933. (14) McAlister, G. C.; Nusinow, D. P.; Jedrychowski, M. P.; Wühr, M.; Huttlin, E. L.; Erickson, B. K.; Rad, R.; Haas, W.; Gygi, S. P. Anal. Chem. 2014, 86, 7150. (15) Dephoure, N.; Gygi, S. P. Sci. Signaling 2012, 5, rs2. (16) Everley, R. A.; Kunz, R. C.; McAllister, F. E.; Gygi, S. P. Anal. Chem. 2013, 85, 5340. (17) Yao, X.; Bajrami, B.; Shi, Y. Anal. Chem. 2010, 82, 794. (18) Gu, L.; Evans, A. R.; Robinson, R. A. S. J. Am. Soc. Mass Spectrom. 2015, 26, 615. (19) Evans, A. R.; Robinson, R. A. S. Proteomics 2013, 13, 3267. (20) Anderson, N. L.; Anderson, N. G.; Haines, L. R.; Hardie, D. B.; Olafson, R. W.; Pearson, T. W. J. Proteome Res. 2004, 3, 235. (21) Fortin, T.; Salvador, A.; Charrier, J. P.; Lenz, C.; Lacoux, X.; Morla, A.; Choquet-Kastylevsky, G.; Lemoine, J. Mol. Cell. Proteomics 2009, 8, 1006−1015. (22) Shi, T.; Fillmore, T. L.; Gao, Y.; Zhao, R.; He, J.; Schepmoes, A. A.; Nicora, C. D.; Wu, C.; Chambers, J. L.; Moore, R. J.; Kagan, J.; Srivastava, S.; Liu, A. Y.; Rodland, K. D.; Liu, T.; Camp, D. G., 2nd; Smith, R. D.; Qian, W. J. Anal. Chem. 2013, 85, 9196−9203. (23) Schiess, R.; Wollscheid, B.; Aebersold, R. Mol. Oncol. 2009, 3, 33. (24) Makawita, S.; Diamandis, E. P. Clin. Chem. 2010, 56, 212. (25) Regnier, F. E.; Julka, S. Proteomics 2006, 6, 3968. (26) Kulevich, S. E.; Frey, B. L.; Kreitinger, G.; Smith, L. M. Anal. Chem. 2010, 82, 10135. (27) Shuford, C. M.; Muddiman, D. C. Expert Rev. Proteomics 2011, 8, 317−323.

CONCLUSION The uMS technology increases the sample throughput for proteomic quantitation and can accelerate MS-based development of protein biomarkers. A 15-in-1 uMRM analysis is demonstrated using nonisotopic reagents for peptide derivatization. These reagents are economical and structurally customizable. The uMS workflows involve additional sample processing steps, compared to nonderivatization-based proteomics quantitation, in order to achieve signal enhancement. These steps include chemical derivatization (as for all other derivatization based quantitative proteomics methods) and affinity enrichment (as for affinity MRM MS methods). As the method requires antipeptide antibodies for sample preparation, the potential of uMS can be greatly realized for high sample throughput studies, such as the validation of already verified biomarker candidates. The cost savings for the new uMS technology arises from the decoupled use of expensive stable isotopes for quantitation reference and large numbers of potential, nonisotopic reagents for sample-specific mass coding. The uMS principle can be further explored to develop new methods for nontargeted quantitative proteomics. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b01727. Additional figures, tables, equations, and methods (PDF)

■

■

This work is supported by the National Institutes of Health/ National Cancer Institute−the Innovative Molecular Analysis Technologies program, NCI-IMAT (1R21CA155536), and the Cystic Fibrosis Foundation (YAO07XX0). The authors would like to thank Dr. Jacob Kagan for helpful discussion during the course of the study. They also thank Dr. Bekim Bajrami, Dr. Pamela Ann Diego-Limpin, and Alexander Gomes for initial efforts in the project.

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. G


Article



(28) McShane, A. J.; Bajrami, B.; Ramos, A. A.; Diego-Limpin, P. A.; Farrokhi, V.; Coutermarsh, B. A.; Stanton, B. A.; Jensen, T.; Riordan, J. R.; Wetmore, D.; Joseloff, E.; Yao, X. J. Proteome Res. 2014, 13, 4676− 4685. (29) Ikai, A. J. Biochem. 1980, 88, 1895−1898. (30) Kulasingam, V.; Smith, C. R.; Batruch, I.; Buckler, A.; Jeffery, D. A.; Diamandis, E. P. J. Proteome Res. 2008, 7, 640. (31) Whiteaker, J. R.; Zhao, L.; Zhang, H. Y.; Feng, L.-C.; Piening, B. D.; Anderson, L.; Paulovich, A. G. Anal. Biochem. 2007, 362, 44.

H


Nonisotopic Reagents for a Cost-Effective Increase in Sample

Recommend Documents