Received: 24 February 2017 | Accepted: 20 June 2017 DOI: 10.1002/mas.21542

REVIEW ARTICLE

Standardization approaches in absolute quantitative with

Francisco Calderón-Celis | Jorge Ruiz Encinar | Alfredo Sanz-Medel

Department of Physical and Analytical Chemistry, University of Oviedo, Oviedo, Mass spectrometry-based approaches have enabled important breakthroughs in Spain in the last decades. This development is reflected in the better

Correspondence quantitative assessment of protein levels as well as to understand post-translational Jorge Ruiz Encinar, Department of Physical modifications and protein complexes and networks. Nowadays, the focus of quantitative and Analytical Chemistry, University of Oviedo, Julián Clavería 8, 33006 Oviedo, proteomics shifted from the relative determination of proteins (ie, differential expression Spain. between two or more cellular states) to absolute quantity determination, required for a Email: [email protected] more-thorough characterization of biological models and comprehension of the proteome Funding information dynamism, as well as for the search and validation of novel protein biomarkers. However, Spanish Ministry of Economy and Competitiveness, Grant numbers: CTQ2016- the physico-chemical environment of the analyte species affects strongly the ionization 79412-P, BES-2014-068032 efficiency in most mass spectrometry (MS) types, which thereby require the use of specially designed standardization approaches to provide absolute quantifications. Most common of such approaches nowadays include (i) the use of stable isotope-labeled peptide standards, isotopologues to the target proteotypic peptides expected after tryptic digestion of the target protein; (ii) use of stable isotope-labeled protein standards to compensate for sample preparation, sample loss, and proteolysis steps; (iii) isobaric reagents, which after fragmentation in the MS/MS analysis provide a final detectable mass shift, can be used to tag both analyte and standard samples; (iv) label-free approaches in which the absolute quantitative data are not obtained through the use of any kind of labeling, but from computational normalization of the raw data and adequate standards; (v) elemental mass spectrometry-based workflows able to provide directly absolute quantification of peptides/proteins that contain an ICP-detectable element. A critical insight from the Analytical Chemistry perspective of the different standardization approaches and their combinations used so far for absolute quantitative MS-based (molecular and elemental) proteomics is provided in this review.

KEYWORDS absolute quantification, isotope labeling, mass spectrometry, quantitative proteomics

1 | INTRODUCTION

Abbreviations: AF4, asymmetric flow field flow fractionation; CE, capillary electrophoresis; A main scientific challenge is the understanding of an organism. The ESI, electrospray ionization; HPLC, high-performance liquid chromatography; IDMS, isotope 1,2 dilution mass spectrometry; ICP-MS, inductively coupled plasma mass spectrometry; sequencing of human genome, implied a huge step towards this goal, MALDI, matrix-assisted laser dissociation/ionization; MRM, multiple reaction monitoring; as it led to the identification and mapping of human genes that carry PTM, post-translational modification.; MS, mass spectrometry; SIL, stable-isotope labeled; SRM, selected reaction monitoring. genetic information. Nevertheless, comprehensive study of biological

Mass Spec Rev. 2018;37:715–737. wileyonlinelibrary.com/journal/mas © 2017 Wiley Periodicals, Inc. | 715 716 | CALDERÓN-CELIS ET AL. processes cannot be achieved just from the genomics perspective. techniques with radioactive, luminescence, or fluorescence detec- Different biomolecules play important roles in such processes, among tion,12,13 and immunoblotting techniques. Immunoassays-based tech- which proteins stand out clearly.3 Understanding the proteome is niques have been at the lead in proteomics studies in the last decades, therefore sine qua non for a full comprehension and clear view of the especially since the development of the major immunoassay-based processes that occur in an organism. methods used for targeted quantitative proteomics, the enzyme-linked The proteome is, however, much more complex than the genome immunosorbent assay (ELISA),14 and the development of the Western due to the so-called dynamism of the proteome. Although proteins Blotting.15 ELISA’s excellent sensitivity and throughput to quantify are gene-translation products, there is no 1-to-1 correspondence proteins even at low pg · mL−1 levels allows for a rather straightfor- because, whereas there are around 20 000 genes in the human genome, ward validation of biomarker candidates if high quality antibodies are the number of proteoforms might amount up to 1 000 000.4 Once available. Moreover, the possibility to study PTMs, protein weight, or proteins are translated in the ribosomes from mRNA, they can undergo location of antibody epitopes provided by Western Blotting have made modifications, named post-translational modifications (PTMs), whose them both “gold standards” in targeted protein quantification. extent varies with individual proteins, regulatory mechanisms within the Advances in mass spectrometry (MS), genomic sequencing, and cell, and environmental factors. On the other hand, proteins functions bioinformatics have revolutionized proteomics studies in molecular and are not based solely on their individual roles or actions, but also from cell biology research in the last decades. MS-based proteomic approaches protein-protein interaction networks, which regulate the function and can routinely identify thousands of known and novel proteins and/or eventual biological activity of the protein components of the network.5 protein PTMs events contained in biological samples.16,17 Additionally, In addition, the diversity in the abundance range of proteins is another MS might provide higher-quality quantitative data because typically, MS with proteomic studies. As an example, human protein levels can vary experiments can use several peptides of the protein as standards, and the more than 10 orders of magnitude in serum, from the tens of mg · mL−1 robustness of such peptide assays can be independently verified in each of albumins to the low pg · mL−1 interleukins.6 laboratory, study, and sample type. In contrast, the analytical performance This complexity of the proteome implies that it cannot be understood of immunoassays-based approaches cannot be properly tested because just from known genomic information. Protein locations or modifications the antibodies might not be sufficiently characterized, and the specificity (which are directly responsible for the particular regulation and activity of and absence of cross-reactivity in the response signal cannot be always a given protein), are not visible and barely predictable at the DNA/RNA assured. Moreover, the observed lack of reproducibility in measurements level. That is, proteomics studies are indispensable to identify the different with different sets of reagents and the unavailability of reference samples proteins, to study their regulation mechanisms and possible PTMs, to have been also highlighted.18,19 Still, MS has not completely replaced understand protein-protein interactions and, last but not least, to immunoassays in quantitative proteomics, most notably in clinical analysis quantitatively determine protein abundances and protein levels variations and validation, and therefore immunoassay-based techniques remain an in healthy and disease status. In this regard, the eventual construction and important tool in this framework. This situation is the result of insufficient elucidation of the human proteome map,7,8 resulted in a remarkable quality assurance of MS-based approaches, which lack controls and breakthrough towards functional proteomics understanding. calibrators for quantitative results validation.20,21 Protein identification and subsequent generation of protein catalogs Even so, MS is the leading trend in proteomics studies and is called have significantly aided one to understand protein biological activity. to lead in the near future studies of the proteome and functional Nevertheless, this sort of information is not sufficient, and the capability characterization of biologically important genes. In this sense, assess not only protein-level variations between two states, but protein molecular-based methods with soft-ionization sources, such as ESI absolute quantities is necessary in order to properly address a (electrospray ionization) and MALDI (matrix-assisted laser desorption/ comprehensive characterization of proteome dynamism.9 The capability ionization) are the reference approaches today. The attainable high to determine protein and PTM quantities might help to assess concentra- sensitivity, resolution, and acquisition rates make ESI-MS an excep- tion changes of relevant proteins during biogenesis, in the modifications of tionally useful tool to analyze complex proteomic samples, particularly cellular states, as well as to reveal biomarker proteins responsible for the with high-performance liquid chromatography (HPLC) separation, to appearance of several diseases or even potential targets for drug enable simultaneous and multiplex quantitative detection in a single development to cure such diseases. Moreover, such quantitative informa- measurement of target peptides.22,23 In contrast, MALDI-based MS tion is required for research of biological systems as a whole (systems offers some unique features in proteomics studies, particularly the biology), hence head towards development of mathematical models and dissociation of LC separation from sample introduction before MS algorithms to describe and simulate complex biological processes carried analysis, to allow separate sample preparation and mass spectrometric out through proteins (eg, signaling networks within the cell). analysis. Moreover, MALDI enables a non-destructive sample analysis, the predominant generation of single-charge ions and, very impor- tantly today, the ability to perform targeted protein imaging 1.1 | The role of mass spectrometry in quantitative analysis.24,25 The high sensitivity of the technique, combined with proteomics the possibility of direct solid analysis on tissue samples, makes MALDI- Traditionally, quantitative proteomics approaches included electro- MS a powerful technique for protein distribution profiling and phoresis separations with UV detection,10,11 immunochemistry biomarkers discovery with differential imaging approaches.26,27,28 CALDERÓN-CELIS ET AL. | 717

In parallel with the more-common molecular MS approaches, interaction models for such biological systems. In order to build up elemental mass spectrometry with inductively coupled plasma (ICP) these models, experimental data and absolute quantitative information ionization is also important for quantitative proteomic analysis, due to on a set of protein concentrations, their changes, their degradation its capacity to determine “heteroatoms” (any element in biologicals rates, or PTM levels might be of utmost importance.32 except C, H, N, and O) and isotopes present or artificially included In this vein, some approaches adopted for relative quantifications (labels) in a given biomolecule. ICP-MS robustness for ionization, which might also be used for absolute quantification,33 provided that there can be translated into quantification without the need for specific are reference standards available for all analytes or their amount added standards, is at the basis to explain why ICP-MS is a rising to the samples for the quantitative analysis is known. Such general complementary approach to the more widely established molecular- analytical approaches exploit the well-established concept in analytical based approaches.29,30 chemistry of “internal standardization,” in which a known amount of a differentiated, but as similar as possible to the analyte (usually isotope- labeled), well-characterized standard is added to the sample in such a 1.2 | The challenge of absolute quantification way that the measured response ratio between the target analyte and Most quantitative proteomics approaches, especially in large-scale the standard is used to quantify it in absolute terms. Figure 1 schemes analysis, compare a proteome in a sample under study related to a the timeline of such internal standardization in most common reference state.31 That is, they are relative quantifications that provide molecular MS-based absolute quantitative proteomics approaches. fold changes between target samples (eg, abundance ratios of the Most current absolute quantitative proteomics approaches suffer sought protein in both samples), or relative differential expression or from the hindrance that the analyte is a different chemical species than regulation of such proteins in at least two compared biological states. the standard at the beginning of the experiment. The soundness to Of course, the information provided by such inter-study sample translate quantities experimentally obtained (especially after time- comparisons in the determination of protein contents is only relative consuming sample preparation procedures) into actual quantity of the unless a standardized reference sample is used. In contrast, an peptide/protein in the sample is questionable.34 The reliability of the “absolute” quantification to provide actual determination of the total eventual quantification can be greatly affected by protein losses or amount or concentration of the individual protein(s) of interest in a degradation during sample preparation and/or the completeness of single proteome is the focus of present proteomic analysis. Absolute protein digestion if a bottom-up approach is used. In other words, it is quantification turns up as a more-accurate and -reproducible tool for still possible that the absolute protein concentration derived from systems biology applications (Box 1), especially to establish predictive those standard-base MS strategies does not accurately reflect the actual protein level originally present in the biological sample. Therefore, absolute protein quantification is not yet a universal quantification method. As a matter of fact, a large and certified BOX 1. Potential areas of interest in biological absolute quantification peptide/protein set of standards does not exist research benefited from absolute quantitative pro- yet. Their availability would be the first requisite to apply quality- tein data. assurance protocols in quantitative proteomics. In fact, absolute quantifications can only be carried out on a limited number of known 1. Evaluation of protein expression, degradation rates, and target proteins and peptides, and are mainly limited by the need for variability, and definition of PTMs levels in model synthetic and appropriately characterized peptide/protein standard in biological systems; each case. 2. Study of cellular organization and dynamics; 3. Determination of protein stoichiometries and ratio variability within protein complexes; 2 | NON-ISOTOPE LABELED 4. Study of changes of protein levels and activation status in STANDARDIZATION APPROACHES protein networks; 5. Validation of potential new biomarkers and their eventual MS-based quantitative proteomics today comprises “bottom-up” clinical utility; workflows, in which proteins are enzymatically digested into peptides 6. Study of potential protein therapeutic agents stability (in prior to their MS study, and “top-down” workflows, in which intact biological medium, or upon storage); proteins are directly analyzed with MS. 7. Investigation of enzymes kinetics; In targeted “bottom-up” approaches, proteins are mostly identi- 8. Development of optimum drug delivery; fied from a match of masses of observed hydrolyzed peptides and the 9. Pharmacokinetic studies on the behaviors of compounds corresponding MS/MS fragments with theoretical values in a database. during drug development. Peptide quantification is performed by comparison of the MS peak intensities of the different isotope-labeled (eg, heavy vs light) peptides. In the specific case of targeted absolute proteomics workflows, previously obtained information on the analyte is used to generate 718 | CALDERÓN-CELIS ET AL.

FIGURE 1 Timeline of standard addition in the main absolute protein quantification strategies with molecular MS during a model quantitative proteomics analysis workflow, adapted from Malmström et al35 and Wang et al36

MS-based assays to detect and quantify corresponding predetermined Furthermore, ICP-MS-based approaches are currently gaining target peptides. Quantification of the target peptides is usually momentum as they can provide absolute protein quantification performed through the spike of samples with adequate peptide without the need for specific standards (stable-isotope labeled or standards (as “signature” or “proteotypic” peptides) that are chemically not) and any type of tagging. synthesized incorporating enriched stable isotopes, and comparing the MS-based quantitative methods that do not make use of stable- measured intensities of the light (endogenous) and the heavy isotope labeling are reviewed in the following sections. (standard) peptides.37 One of the major downsides of such methods, nonetheless, is that the primary stable isotope-labeled reagents 2.1 | Non-isotope labeled standardization approaches needed for the synthesis are not easy to obtain and are usually with molecular mass spectrometry expensive. Besides, the cost and time to synthesize, purify, and certify the stable isotope standard peptides for the desired absolute 2.1.1 | Standardization in quantitative label-free MS quantitation, as well as setting up spike-in experiments and standard curves, can be substantial. The high cost of the synthesized isotope Labeling methodologies require complex, time-consuming sample labeled standards continues to make such determinations of clinically preparation and reactions and can be relatively expensive. In contrast, relevant proteins comparatively expensive for routine analysis.38,39 label-free quantification is by nature the simplest and more-economi- In this context, the use of the so-called “label-free” approaches is cal approach. It is, in principle, applicable to any kind of sample, today a clear trend in proteomics, and have the potential to be at the including materials that cannot be labeled, and there is no virtual limit lead of the new developments in quantitative proteomics in the recent on the number of samples that can be compared/quantified. However, future.40 These approaches mostly employ computational resources despite the long-standing interest of protein quantification without for the quantification by a comparison of the intensity signals or the (isotopic) labels in the proteomics field, the accurate and robust average spectra matched to a peptide or protein obtained from MS proteome-wide quantification via label-free approaches remains a analysis.33 However, they might require internal standards for MS challenge. signal normalization, because at their current state they still lack the Label-free quantification is generally based on the comparison robustness and reproducibility required, and for the production of either of the mass signal intensities (spectrometric spectral counts) for absolute quantitative data. selected proteolytic peptides, or the number of peptide-sequencing Another alternative to the use of SIL standards is to tag directly the events observed for the samples/analytes to be analyzed.16,42,43 The tryptic peptides present in different samples with isobaric tags. These intensity values of each peptide in one controlled (reference) tags have a constant overall mass, so that the final mass of the same experiment can be compared with their respective intensities in one tryptic peptides present in different samples is still identical and will or more other experiments to yield relative quantitative informa- elute together in the LC-MS analysis after sample mixing.31,41 The tion.44,45 With appropriate computational tools, it is possible to infer isotopic composition of the component parts of such tags is different, the difference in protein abundance between two different biological so different MS/MS fragments are observed for each chromatographic states. The interest in label-free quantification is growing lately peak, which correspond to each sample analyzed. Hence, multiplex because, even though label-free methods suffer from compromised quantification can be carried out. Of course, absolute quantification is quantification accuracy (because there is less quantitative control in affordable if one of the samples analyzed corresponds to a standard the analytical procedure steps), the label-free algorithms can approxi- with known protein concentrations. mate the protein ratios between samples. Therefore, these methods CALDERÓN-CELIS ET AL. | 719

TABLE 1 Internal standardization in label-free approaches to control variations in the sample preparation, chromatographic, and MS performances Approach Description Ref. Retention time score (iRT) Empirical normalized values of a set of specific peptides retention times, that are consistent along 46 different LC configurations enabling prediction and scaling of peptides chromatographic behavior Reversed-phase liquid Synthetic protein standard that provides peptide calibrant points after digestion, to normalize 47 chromatography calibrants retention times, and evaluate and optimize HPLC performance (RePLiCal) Peptide trainer kits (PTK) Set of synthetic peptides to predict and correct for dynamic retention time drifts, re-scheduling the 48 time window that monitor target peptides in SRM analysis Halogenated peptides as internal Internal standards that contain halogens, taking advantage of their unique isotopic distribution and 49 standards (H-PINS) mass defect, to be used to evaluate label-free approaches reproducibility, and to assess labeling reaction efficiencies Pseudo Internal Standards (PISs) Proteins contained in the sample that do not show variation in expression levels among different 50 sample states, used as unlabeled internal standards. They are employed to correct for recovery during sampling process, as well as for matrix effects

can be appropriate when large-fold changes are expected and Alternatively, Winter and coworkers introduced a “label-free” approximate comparisons are enough.51 quantitative approach, which is based on the use of minimally Even though these label-free strategies are mostly limited to permuted peptide analogues (MIPA).57 MIPA introduces a positional relative quantification, absolute quantitative approaches have been also permutation of two related amino acids on a synthetic standard reported, such as emPAI (exponentially modified protein abundance peptide. Because the amino acid composition of the MIPA and the index), APEX (absolute protein expression), and iBAQ (intensity-based master peptide is the same, MS ionization efficiency will be very absolute quantification). emPAI uses an exponential expression that similar. Nevertheless, the fragments produced in the MS/MS analysis considers the number of observed and observable peptides per protein, will differ due to the differentiation in the peptides sequences which is proportional to the absolute protein abundance and might be produced by the permutation. MIPA-based absolute quantification can converted into absolute concentrations after measuring total protein be carried out from comparison of peak areas from MS or MS/MS amounts by BCA assay.52 APEX is an spectral count method combined spectra, depending on whether the targeted and MIPA-peptides elute with computational normalization, in which absolute protein abundance at different retention times or co-elute, respectively. is obtained from relating the protein abundance in the sample and the number of experimentally obtained and predicted peptide observations 2.1.2 | Isobaric tagging in molecular MS per protein.53 These label-free strategies hold great potential for quantitative proteomics, being able to estimate protein abundance During the past few years, several isobaric tagging strategies have levels with acceptable levels of accuracy,54,55 though they still present been developed and play important roles to compare different several practical limitations. High precise, reproducible, and accurate biological samples (differential proteomics) that generally resort to raw data are vital, as well as an exhaustive control and minimization of methods. The mass-labeling basis in these variations in the sample preparation, chromatographic, and MS procedures is directed by the reactive group of certain amino acids performances. These experimental variations, regardless of their source, of the protein with the tag. Consequently, such approaches are limited can be accounted for, provided that appropriate internal standards are to proteins or peptides that contain those certain amino acid residues. used.31,49 Table 1 summarized the use of such internal standards in The reporter section of the tagging reagents is different for each label-free approaches. sample/protein and provides different fragments in the MS/MS On the other hand, iBAQ transforms normalized protein intensi- analysis, whereas the “balance” section equilibrates any mass ties into estimated protein copy numbers, by means of relating the sum difference, so that every tag (reporter and balance section) has the of the MS intensities of all peptides matching to a certain protein, and same final mass. the number of theoretically observable tryptic peptides of the protein More than a decade ago, Ross and coworkers58 developed a (iBAQ intensity).56 In that work, iBAQ intensity was correlated to methodology based on this alternative mode of labeling named iTRAQ protein quantities by means of using a spike-in commercial protein set (isobaric tag for relative and absolute quantitation). The iTRAQ reagent standard, hence protein standards concentration was accurately is tagged to proteins or tryptic peptide sides at the N-terminus amino known. Those protein standards were used to build a calibration group and the side chain of Lys (Fig. 2A). The resulting isobaric-tagged curve (log-iBAQ intensity vs log-protein molar amount), so that the protein/peptides are mixed, and subjected to fractionations and to LC- linear regression could then be used to transform iBAQ intensities into MS/MS analysis for quantification. Due to the isobaric nature of the actual absolute quantities for all proteins identified in the sample. labeling reagents, the same peptide from each sample appears as a 720 | CALDERÓN-CELIS ET AL.

FIGURE 2 Isobaric tags for peptide/protein quantification. A) iTRAQ reagent, adapted from Ross et al.58 The reactive group is usually based on N-methylpiperazine, and can have up to four 13C and two 15N, the mass balance group is a carbonyl group with can have up to two 13C and two 18O, in such a way that the overall mass of the isobaric tags are the same. The reactive group is usually based on a NHS ester group that reacts with the peptides amines forming an amide bounding. B) reagent, adapted from Dayon et al.59 Similarly, to iTRAQ reagents, it comprises a reporter group (detected by MS) and a mass normalization group, of identical overall mass in the different reagents, and a reactive group. The reactive group is commonly an amine reactive group, although TMTs bearing carbonyl reactivity have been reported though hydrazide- and aminooxy-functionalized reactive groups60

single peak in the MS spectra. However, after the MS/MS fragmenta- isotopes, but in such a way that the mass of the isobaric tag is the same in tion, the tags will produce diagnostic, low-mass MS/MS signature ions all the TMT reagents (Fig. 2B). Dayon and coworkers have exploited the in such a way that the signal intensity ratios of these signature ions quantitative multiplexing capabilities of TMTs with the use of six-plex indicate relative proportions (ratios of downregulation or over- isobaric tandem mass tagging.66 Later on, McAllister and coworkers expression) of the corresponding peptides/proteins/samples.61 iTRAQ increased the multiplexing capability of the method with ion isotopo- technology enables relative comparison of differential expression logues with isobaric masses, in which a 15N was replaced by a 13C, which levels using different iTRAQ reagents (4- and 8-plex). This approach, therefore requires high-accuracy mass spectrometers to resolve the however, requires from high accuracy and reproducibility of the LC- induced mass shift66; 2 years later, the use of 10-plex TMT tags was MS/MS performance, and the use of an internal standard can be reported.67 As was the case with iTRAQ, proof-of-concept of the use of addressed in order to correct for instrumental variations.62 TMTs for absolute quantification was also developed in parallel. In the Moreover, absolute protein quantification can also be obtained work by Dayon and coworkers, potential protein biomarkers of brain- using the iTRAQ workflow.58 In this regard, Quaglia and coworkers damage disorders from cerebrospinal fluid were quantified in absolute evaluated the applicability of iTRAQ for absolute protein quantifica- terms by means of a calibration curve. To build up that curve, four tion. They combined the spiking of SIL peptides in a somatropin identical mixtures at different concentrations of TMT-labeled synthetic sample, with iTRAQ derivatization (after tryptic digestion), in order to proteotypic peptide standards of such proteins were measured by MS/ determine peptide and hence protein concentration from the area ratio MS, and therefore relate the reporter ion intensity and the known of the iTRAQ reagent reporter ion in both, tryptic and SIL peptides. To peptide concentration.68 Despite applicability and linearity of the do so, they used a calibration mixture of two synthetic SIL and non- approach was further validated with standard addition method using labeled somatropin peptides, which were treated likewise to the synthetic standard peptides, results showed variability among peptides sample, so that quantification could be addressed with both IDMS and from the same protein, which the authors attributed to different factors, iTRAQ approaches. This strategy enabled the validation with IDMS of such as the existence of PTMs, the tryptic digestion performance, or the the iTRAQ-based quantification of somatropin protein. Quantitative different stability and recovery among peptides. results with both approaches were equivalent, though iTRAQ offered In the use of isobaric tags, one of the main issues to take care of is lower precision than IDMS.63 Still, these strategies offer higher the potential errors or effects that might happen during sample accuracy than label-free approaches, which in contrast can provide preparation, and separations that can affect the quality and validity of higher proteome coverage and increased dynamic range. However, the MS/MS quantitative results. These effects can be corrected via iTRAQ is limited by inherent problems such as fold-change underesti- statistical approaches or normalization methods of the measured mation or isotopic impurities of the iTRAQ label,64 and by the need of experimental peptides intensities by resorting to constrained opti- efficient protein digestion, and reproducible and complete protein/ mizations and taking into account information about the labeling peptide derivatizacion.63 strategy.41 Together with iTRAQ, isobaric labeling with “ tags” (TMTs) holds great popularity. These tags, similarly 2.2 | Non-isotope labeled standardization approaches to iTRAQ reagents, consist on a peptide-reactive group (NHS ester) and with elemental mass spectrometry the isobaric tag, which comprises a reporter group (based on N-methylpiperazine) and mass-balance group.65 The reporter and ICP-MS analytical features for elemental trace analysis include potential mass-balance group masses are different through the use of different species-independent signal response of any element-containing CALDERÓN-CELIS ET AL. | 721

FIGURE 3 ICP-detectable heteroatom source and standardization timeline in the main elemental MS-based absolute quantitative proteomics strategies

species, excellent limits of detection (down to ng · L−1), as well as linear inorganic compounds that contain the measured heteroatom as dynamic range of up to 10 orders of magnitude. As a consequence, it is calibration standards.71 important as a complementary alternative to better-established ICP-MS detection of the non-metallic elements typically present in molecular MS-based quantitative approaches, particularly for absolute proteins (ie, S, P, Se) has been traditionally hampered and only very quantitative proteomics.69 In this regard, ICP-MS can be applied to recent instrumental advances, mostly based on a triple quadrupole quantify virtually all proteins by means of the direct determination of configuration for more efficient interference removal,72 have enabled naturally present ICP-MS-detectable heteroatoms (eg, metals, S, P, or their highly sensitive detection. ICP-MS-based studies in quantitative Se) or after appropriate element tagging to the target species,70 as proteomics had been specifically focused on analysis of the metallic schemes Fig. 3. and semi-metallic heteroatoms present in (metallo) proteins. Metal- loproteins contain coordinated metals (eg, Zn, Fe, Cu, etc) within the protein structure to provide them with important biological activities 2.2.1 | Direct quantification of peptide/proteins using (eg, enzymatic, storage, or transport). Consequently, ICP-MS-based elemental MS quantification of the metallic heteroatom can produce relevant Even though the four major elements present in proteins (C, H, O, N) information about its effect in the biological activities and functions cannot be directly measured with ICP-MS, there are elements of the proteins.70 Furthermore, the knowledge of the fixed metal/ (including metals, semimetals such as Se, or non-metals such as S or protein stoichiometry enables one to obtain information of protein P) naturally present in proteins whose quantitative measurement can quantities from the quantitative detection of the metallic heteroatom. be used for the absolute direct quantification of proteins and peptides. Even though ICP-MS-based quantification of metalloproteins is more Of course, complete species isolation is required because ICP-MS does based on the use of isotope dilution (see section 3.2), it is possible to not distinguish between the elemental signals coming from different determine the quantity of the metallic heteroatom bound to the co-eluting species. To achieve such protein determination, the protein by means of external calibrations of known concentrations of corresponding stoichiometry element:protein is required as well. metalloproteins.73,74 These studies are limited, nonetheless, by the Due to the virtual species-independent nature of the ICP-MS stability and strength of the protein-heteroatom union through the ionization,70 absolute quantifications can be approached with even analytical procedure and measurement, because metal-protein 722 | CALDERÓN-CELIS ET AL.

FIGURE 4 Scheme of an “absolute” quantitative proteomics workflow based on combining elemental and molecular MS detection after identical separation (HPLC) conditions, adapted from Calderón-Celis et al.75 ICP-MS: the sample containing the sought analyte is spiked with a species-unspecific standard containing the detected heteroatom. After LC separation, heteroatom absolute content of the standard and the analyte(s) is accomplished by on-line isotope dilution analysis (mass flow equation, MFs, Fig. 7C) by relating the peak areas of the standard (of known concentration added to the sample to correct for injection errors and simplify IDA computations) and the analytes. ESI-MS: the detection of the analyte species provides their identification and/or sequence characterization, either with MS and MS/MS analysis (ie, top-down workflow), or with accurate MS analysis together with database information. This provides the stoichiometry heteroatom:protein molecule, hence the absolute amount of the corresponding protein can be finally obtained from the element content determined complex instability (especially during the LC separation step) can ruin of phosphorous in proteins is most relevant today to investigate one of the quantitative results.76 Consequently, the possibility to precisely the most important PTM, phosphorylation. Of course, as with sulfur, detect heteroatoms covalently joined to the protein structure is of high phosphorylation quantification requires the knowledge of stoichiom- interest. etry phosphorous/protein. Nonetheless, phosphorous is not as In this sense, one of the most-promising direct targets for such common as sulfur in protein structures. This fact is important because heteroatom-based quantitative studies is sulfur. Because it is present it enables one to use phosphorous elemental determination for direct in cysteine and methionine residues, sulfur is statistically present in detection (screening) of phosphorylated of proteins in complex almost all proteins. In fact, sulfur is not only a potential candidate for biological systems.80 Notably, such approaches also show great absolute quantification of proteins via ICP-MS, but for special studies potential to assess quantitatively the many procedures commonly of phosphorylation degrees and metalloproteins stoichiometries.77,78 used for purification, preconcentration, and/or isolation of peptides in In any case, application of direct sulfur measurement to eventual proteomics studies, as already demonstrated specifically for phos- species quantification is limited by the need to separate with LC the phopeptides enrichment strategies.81 typically very abundant sulfur-containing analytes with quantitative After prior developments in the ICP-MS analysis of phosphorous recoveries from the column. This is especially critical in the case of for the determination of the phosphorylation degree in proteins,82 and intact protein analysis. Even so, the continuous advances in HPLC the absolute and site-specific quantification of protein phosphoryla- together with triple quadrupole ICP-MS technology head towards the tion,81,83 in 2009 Zinn and coworkers proposed a method based on the possibility of absolute quantification of both peptides72 and intact use of phosphorous-based absolutely quantified standards peptides proteins.75 In fact, if correction of the ICP-MS sensitivity along the LC (PASTA) for the absolute quantification of phosphorous-containing gradient is achieved, absolute quantification of several intact proteins peptides (phosphorylated peptides).84 Such synthetic peptides, which in relatively simple mixtures using a single S-containing generic were also stable-isotope labeled, could be quantified in absolute terms standard, spiked to the sample prior to the LC-ICP-MS analysis, by calibration with a P-containing generic standard. Once their becomes possible. In this line, hybrid LC-MS configuration with both concentration is certified, these PASTA peptides can be used for elemental and molecular MS detection has the potential to in parallel quantitative purposes in LC-ESI-MS analyses, after their dephosphor- identify and quantify intact proteins directly (Fig. 4).79 The potential ylation with protein phosphatase. applicability of such approach depends on the availability of robust In this line, the recombination of proteins expressed in a platforms to achieve pre-MS separation of each and every protein selenomethionine-containing media led to exchange of methionine species. amino acids by selenomethionines to generate standards named RSIQ Another important elemental target in quantitative proteomics (recombinant isotope-labeled and selenium-quantified). These stand- studies is phosphorous. Unlike sulfur, phosphorous is monoisotopic, ards can also carry an isotopically labeled amino acid (ie, 13C) for that is, only a single stable isotopic form, 31P. The direct measurement subsequent use in quantitative proteomics with molecular MS CALDERÓN-CELIS ET AL. | 723 approaches. From the fact that such RSIQ protein standards carry a ToF-MS analyzer (CyTOF). Thereupon, further understanding of selenium-containing amino acids, quantification with ICP-MS mea- multiple features of cellular systems and multi-parametric biological surement of the selenium enables absolute determinations of the functions is possible by broadening the experimental toolbox for selenoprotein as well.85 systems biology investigations.88 In this line, “metal-coded affinity tags” (MeCAT) use a chelate complex that contains rare-earth elements as metal tags.89 This 2.2.2 | Direct protein tagging for ICP-MS based complex also contains a reactive group for covalent labeling to the quantification amino acids (Fig. 5B). The protein or peptide quantity is determined Whereas the measurement of naturally present heteroatoms enables directly by the assessment of the metal element signal with ICP-MS. direct quantification of proteins, phospho- and selenoproteins and Further, Bergmann and coworkers carried out “top-down” and almost any metalloprotein with ICP-MS,70 the use of non-present “bottom-up” MeCAT quantifications90 applying MeCAT labeling to heteroatoms as labels offers also some unique features. Artificial peptides and intact proteins. Metal quantification with FIA-ICP-MS of elemental labeling enables the quantification of proteins or peptides the lanthanide label amount enabled final absolute protein concentra- that otherwise could not be quantified with ICP-MS when the peptide tion evaluation. or protein does not contain ICP-detectable heteroatoms, or the Depending on the nature of the quantitative study, a more- required sensitivity cannot be achieved by the direct measurement of specific tag could be required, as is the case of mercury tags after their naturally present heteroatoms. In this sense, the chemical labeling of bonding to thiol groups of the sought amino acid residues (cysteine). the sought peptide/protein molecule with bifunctional reagents that There are some examples of this particular approach that use tags such contain the heteroatom (Fig. 5A), mostly rare-earth elements, can as small mercury compounds91 or p-hydroxymethilbenzoate92 enhance the ICP-MS sensitivity due to low ionization potential and lack (Fig. 5C). Concerns about toxicity, together with the limited of interferences. Great multiplexing capabilities due to the number of applicability of the approach (selective reactivity to just cysteine), potential heteroatoms (and isotopes) are also available.86,87 The make this tagging a rather academic solution. potential of this type of labeling and its multiplexing capabilities with Moreover, many peptides might not have a S-reactive group prone ICP-MS detection is clearly shown by the development of a novel to react with a more standard chelating complex (eg, MeCAT and “mass cytometry” technique, as an alternative to conventional flow mercury labeling). Therefore, a wide range of target sites for elemental cytometry. This technique consists of the labeling of cell components labeling is desirable. This is the particular case with tyrosine (Tyr)- with recognition molecules (mostly antibodies) that carry rare earth containing peptides that can selectively react with bis(pyridine) elements (and isotopes) that enable quantitative analysis with ICP and iodonium tetrafluoroborate (IPy2BF4), leading to the bioconjugation

FIGURE 5 Chemical structures of labeling reagents in ICP-MS approaches. A) Scheme of bifunctional reagents, which consist on a chelating group that bonds the ICP-MS detected heteroelement (mostly metals), a bridge or main body of the reagent, and a recognition group that reacts to the sought analyte. B) Schematic representation of MeCAT reagent, adapted from Ahrends et al.89 It contains a DOTA macrocycle for metal chelating, a spacer that connects the macrocycle and the maleimido reactive group that specifically reacts to thiol groups. The spacer might contain a biotin group for purification. C) Mercury tags: p-chloromercurybenzoate, p-hydroxymercurybenzoate, methyl-mercury, ethyl-mercury, adapted from Tholey and Schaumlöffel.86 D) Iodination reaction using bis-pyridinium iodinium tetrafluoroborate as reactive agent, adapted from Pereira Navaza et al93 724 | CALDERÓN-CELIS ET AL. of two iodine atoms to the metapositions of the activated aromatic those NPs exhibit photoluminescence properties as well. The use of ring. Therefore, iodine labeling can be used to quantify peptides/ these kind of labels for quantification requires the corresponding proteins that contain Tyr amino acids with the ICP-MS measurement bioconjugation reaction between NP and antibody to be well- of iodine (Fig. 5D). Again, a simple iodine-containing standard, such as characterized and under control. Regarding quantification, typical iodobenzoic acid, can be used as a generic standard for the Tyr- immunoassay calibration curves are typically carried out. However, if containing peptides/proteins previously labeled.93 Interestingly, io- the elemental composition of the nanoparticle and the stoichiometry dine ICP-MS detection is much more sensitive than phosphorous or nanoparticle/antibody are known, then the approach has the sulfur although, in turn, it is less sensitive than lanthanides. potential to provide quantification without the need for specific standards because the elemental determination could be directly translated into protein amounts. Still, as in many cases for 2.2.3 | Antibody tagging in immunoassay approaches immunoassays, the selectivity remains an important issue and with ICP-MS detection cross-reactivity as well as unspecific interactions are important An alternative to heteroatom bioconjugation with the peptide/protein concerns to take care of. of interest is the use of bifunctional reagents, which contain the In addition, if the elemental sensitivity achieved with nanoparticles heteroatom bound to an adequate protein-recognition moiety, such as is not enough, this kind of label enables one to resort to the so-called a specific antibody (see Fig. 3). “amplification” procedures. After the immunoassay has been success- We have already commented that quantification of proteins fully carried out, the nanoparticle can act as seed, and an elemental based on immunoassays have been a major method due to their layer can be deposited upon it, so that at the end a much bigger good sensitivity, selectivity, and throughput, especially the ELISA nanoparticle is measured. It is worth stressing that the direct use of method. Immunoassays are usually combined with luminescence such bigger nanoparticles is not as effective because the recognition techniques to provide fast and simple detection. However, it has capabilities of the antibody could be compromised. In this manner, the been demonstrated that ICP-MS detection in immunoassays can measured analyte contains a much higher number of elemental atoms overcome some limitations of luminescence detection, such as to amplify enormously the ICP-MS signal and so enhance significantly matrix effects; and therefore they do not require extra sample- the detection limits beyond the fg mL−1 level.101 preparation steps. Enhanced sensitivity is provided as well.94 Still, immunoassay-based approaches entail negative issues, inherent to the use of antibodies, that concern cross-reactivity, reagent 3 | STABLE-ISOTOPE LABELED antibodies saturation, or lack of repeatability among measurements STANDARDIZATION APPROACHES and sensitivity.18 Combination of immunoassays with ICP-MS detection was first Ever since the introduction of the concept of stable-isotope labeling introduced as an excellent and sensitive alternative for quantitative for the quantification of peptides,102 it has become a roaring trend in proteomics.95 In brief, an antibody that reacts specifically to the targeted quantitative proteomics. Typically, the synthetic standard target protein is used. This antibody is tagged with an elemental peptides are synthesized with an incorporated stable heavy isotope label, which can be measured with ICP-MS. This label can be a label (mostly 2H, 13C, 15N, 18O) on one or more selected amino acids in metal ion, such as europium,95 a polymer chain ligand that contains their sequences. These synthetic reference standard peptides, or lanthanides96,97 or magnetic nanoparticles that contain the proteins, mimic their endogenous natural counterparts (the analyte), in detected heteroatom,98 among others. These approaches combine such a way that they share the same physicochemical properties, the high-selectivity characteristics of recognition reactions of including chromatographic co-elution, ionization efficiency, and immunoassays with the highly sensitive detection of metal fragmentation pattern. However, whereas identical chromatographic elements in ICP-MS, besides the capability to detect different retention time can be guaranteed with 13C, 15N, or 18O labeling,103 for elemental isotopes to provide multiplexing capabilities.87 In this 2H labeling the actual retention time might be notoriously changed to vein, Bendall and coworkers have reported the labeling of around likely affect quality of the quantitative results, because conditions at 30 (epitope-specific) antibodies with different elements and the ESI source will not be identical for both isotopologues.104 isotopes for multiplexed cell analysis by ICP-MS-based mass Such stable-isotope labeled (SIL) standards only differ from the cytometry analysis.88 analyte in their isotopic compositions. Therefore, there is a well-known In addition, the possibility to use metallic nanoparticles (NPs) as defined mass difference so that they can be unambiguously antibody labels instead of single elements allows enhanced distinguished and easily detected with MS. Because the isotope label sensitivity for protein quantification because they might be formed is perceived by the mass spectrometer as a shift in the m/z ratio, and by thousands of atoms.30,99 Numerous types of NPs with very they are introduced into the sample at a known concentration, the different and unique physicochemical properties have been used, relative ratios of the resulting peptide pairs are a direct measure of the including gold nanoparticles (AuNPs), silver nanoparticles (AgNPs), absolute concentration of the target peptides/proteins in such sample. silver deposited upon AuNPs, lead nanoparticles (PbNPs), nano- Of course, accuracy of the measurement depends also on the ability to clusters (NCs), and quantum dots (QDs).100 Interestingly, many of discriminate from possible interfering signals105 in the MS and MSMS CALDERÓN-CELIS ET AL. | 725 spectra. One way to achieve this discrimination is to resort to high Of course, proteotypic peptides should be previously synthesized mass resolution analyzers or, alternatively, to reduce the sample and certified. Generation of peptide standards is commonly addressed complexity through sample fractionation, or via computational by stepwise chemical synthesis; solid-phase peptide synthesis (SPPS) is means.106 the method of reference (Fig. 6A).110 This approach produces peptides through the assembly of the amino acids onto a solid polymer support [eg, cross-linked poly(dimethyl acrylamide] or polystyrene resins). In 3.1 | Stable-isotope labeled standardization the case of stable-isotope enriched peptide standards, they can be approaches with molecular mass spectrometry generated by using the corresponding isotope labeled amino acids, usually enriched in the C-terminal in Lys or Arg, with 15N and 13C.107 Targeted absolute quantitative proteomics with HPLC-MS make use of However, in the course of the chemical synthesis, some peptides can such specific isotope-labeled peptides or proteins standards, which are suffer from artificial modifications over time, and several residue- or spiked into the sample at known concentration for each targeted sequence-dependent side reactions can also occur during the SPPS peptide or protein to be determined. process, such as the formation of diketopiperazine or aspartimide,111 which results in potential errors in the final absolute measurements. For these reasons, synthesis strategies based on biological, rather than 3.1.1 | Peptide-based approaches chemical, approaches are popular. These strategies are based on the in The vast majority of these strategies resort to synthetic isotope- vitro expression of an amino acid sequence from a specific gene in a labeled standards, analogous to proteotypic peptides of the species of biological media. In the case of isotopically enriched standards, the interest. The main approaches are AQUA (Absolute Quantification), in growing cell culture media is enriched in the desired isotopically which quantification is accomplished by means of spiking the sample enriched amino acids, usually arginine and lysine (metabolic labeling).33 with the adequate heavy synthetic peptide,107 and QconCAT Even though peptides can be directly synthesized with metabolic (Quantification Concatemer), which makes use of a synthetized approaches, the synthesis of proteins (as will be seen later) using their isotopically labeled protein built by concatemerization of several specific encoding gene is more common, as they can be digested after proteolytic peptides of the target protein(s) under study.108 the synthesis in order to produce the target proteotypic peptides. Typically, AQUA methodology consists of addition of the proper Once generated the stable isotope-labeled synthetic peptide synthetic SIL peptide, physically and chemically identical to the standards, the concentration of such standard has to be correctly endogenous part, and of known concentration, into the experimental certified in terms of purity in order to obtained accurate quantitative sample to obtain absolute quantitative data.38,102,107 Quantification is results. The most-common and -extended approach for peptide/ carried out with MS detection and “selected” (SRM) or “multiple” protein certification is amino acid analysis (AAA), which provides (MRM) reaction monitoring on the basis of the measured intensity ratio quantitative information on the amino acid content of peptides or of both peptides, natural and labeled (easily detected as pairs of signals proteins as well as their amino acid composition after a hydrolysis step separated by a known and constant Δm). Because the concentration of (mostly acid) to break the peptide bonds of the peptides or proteins. the labeled peptide added to the sample is known, the amount of the AAA presents a series of issues to take into account, such as (i) the acid natural peptide can be determined from the signal ratio. However, one hydrolysis, which can produce chemical modifications or degradations important issue is the initial estimation of the concentration of the in the amino acids when too rough conditions are employed, obligating labeled standard to be used, because the protein(s) of interest might be to the control and optimization of this procedure112; (ii) the low present in a broad range of concentration levels in the sample. This efficiency of the acid hydrolysis for long amino acid sequences, which limitation can be overcome with a preliminary MRM measurement.38 hampers the applicability of AAA to intact proteins quantification; (iii) Moreover, SIL peptides cannot be used in the initial isolation step of the high amount of starting materials; (iv) its lack of specificity, because the protein from the sample, and thereby the reproducibility assurance AAA detects aminoacids independently of the original peptide; and (v) of this sample preparation step is a must. the requirement for highly purified peptides, which usually implies Typical requisites for such internal standard SIL peptides are as purity assessment.113,114 An alternative approach to amino acid follows: (i) they should be proteotypic in order to avoid peptides with analysis is chemiluminescent nitrogen detection (CLN). CLN enables the same sequence from different proteins; (ii) efficiently ionizable in the direct determination of the peptide content even in crude mixtures order to provide a good detection by MS; and (iii) peptides with PTMs in contrast to AAA. The CLN approach consists of the burning of the or that might be subjected to polymorphism should be avoided. The compounds to achieve combustion of the nitrogen. Nitric oxide species election of the proteotypic peptides to be used as standards in formed react with ozone in a specific chemiluminescent reaction and quantitative proteomics is usually approached in an empirical mode, emit light in the red and infrared region of spectrum. The intensity and makes use of those peptides that have been previously sequenced measured with a photomultiplier tube is directly proportional to the in a shotgun experiment, because chromatographic and ionization amount of nitrogen in the sample.115 Nevertheless, the methodology properties can be controlled. Nevertheless, if such information is not lacks robustness and selectivity, because any compound in the sample available, then bioinformatics prediction programs can be used as well that contains nitrogen would contribute to the signal emitted, including to choose potential tryptic peptides.109 those containing ammonia, hydrazine, or nitrogen oxides. 726 | CALDERÓN-CELIS ET AL.

FIGURE 6 Strategies of peptide/protein standards synthesis: A) Chemical synthesis based on solid phase supports (SPPS), adapted from Amblard et al.111 The amino acids are added in excess in order to assure maximal yield and in the desired order in each step of the synthesis. They are N-terminal protected, usually with tert-butyloxycarbonyl (t-Boc) or 9H-fluoren-9-ylmethoxycarbonyl (Fmoc) groups, in order to prevent self-coupling of the amino acids to compete with the peptide chain synthesis. The assembling process is usually controlled by monitoring with spectrophotometry the amount of Fmoc removed at each deprotection/coupling cycle,116 whereas the synthesis process can be controlled by measuring quantitatively the peptide chain or the unreacted free amino acids by the ninhydrin reaction.117 Deprotection step is usually carried out with TFA when BOC is the protecting group and with Piperidine/DMF when it is FMOC. Cleavage step is usually carried out with high concentration of TFA, together with phenol and TIPS. B) Metabolic synthesis through concatemerization. Firstly, the proteotypic quantification peptides that correspond to the target proteins are selected and synthesized together as a single protein through the corresponding synthetic designed gene. Gene expression can be performed in a medium that contains isotopically enriched lysine and arginine to produce isotope-labeled standards. C) Chemical ligation through chemoselective reaction for the union of peptides, adapted from Dawson and Kent.118 D) Metabolic synthesis with the PURE system, adapted from Xian et al.119 From the synthetic DNA precursor (previously amplified by PCR from synthetic oligonucleotides), the sought peptide or protein is synthesized with the PURE kit, which contains the required isolated enzymes, amino acids (which can be isotopically labeled), ribosomes, and RNAs. All recombinant protein components of the system are His-tagged, allowing purification of the synthesized untagged protein by removing the rest of the system’s components. E) FLEXIQuant synthesis strategy, adapted from Singh et al.120 The gene of interest is cloned into a wheat germ expression vector, and transcribed and translated in a wheat germ extract (WGE) in the presence of selected heavy isotope-labeled amino acids ([13C, 15N] arginine and lysine), which provides the in vitro expression of the full-length labeled protein standards

On the other hand, the proven advantageous features of standards can be directly assessed with calibrations via a single elemental MS (ICP-MS) for quantitative element detection of generic standard as simple as an inorganic or organic pure heteroatoms point to ICP-MS as an increasingly attractive alterna- compound.70 Certification of the peptide concentration through tive for protein and peptide standard certification. In this case, given its direct measurement cannot be done unless it contains a that the signal of the detected element is directly proportional to its heteroatom (any element different from C, O, H, N and F), such as absolute amount and independent of the chemical species where it is P, Se, or S. Because most peptides do not contain heteroatoms, their present, the certification of the concentration of the peptide certification is usually carried out via detection of a heteroatom CALDERÓN-CELIS ET AL. | 727 present in a tag chemically joined to the protein with strategies quantification of the target proteins as happens with QconCAT, but commented on section 2.2. also their natural flanking sequences. Therefore, the cleavage An important strength of AQUA is its selectivity and rapid efficiency of the PCS and the target protein is in principle matched development, even though it is not as sensitive as antibody-based to improve quantification accuracy. Consequently, errors in the immunoassays (ELISA) when is used directly without any enrichment digestion efficiency and existence of miscleavages can be better step. Yet, AQUA can be applied to a wide range of samples, from tissue corrected, because the adjacent aminoacids (which will be cut by the culture cells to human plasma and tissue; it is limited by instrumenta- enzyme) are included in both the natural and the standard peptide. tion capabilities; that is, the number of proteins that can be measured With the use of the PCS strategy, quantitative information regarding in a single LC-MS analysis, the duty cycle of the mass spectrometer, the stoichiometry of the components of a multi-protein system can be and the actual distribution of AQUA peptides across the LC elution obtained and what is more important, eventual absolute analytes profile. In 2004, a method named SISCAPA (Stable Isotope Standards quantification might be also provided. In any case, purification and and Capture by Anti-Peptide Antibodies), which combined MRM gene-construction processes are limited by larger PCS. For instance, in detection with antibody-based selective enrichment and purification the study of multi-protein complexes several PCS need to be used, and of peptides, was described.121 Briefly, the SIL peptides are spiked to the knowledge of the exact amount of each PCS added to the sample is the sample, which is subjected to enzymatic digestion. The proteolyzed required. sample is incubated in a medium that contains antibodies (affinity An important issue that affects not only AQUA and QconCAT, columns or conjugated to agarose or magnetic beads) that react but also any targeted quantitative proteomics approach, is the specifically to the sought peptides to enable their enrichment. Finally, selection of the peptide standard for quantification. Computational the enriched peptide is released and analyzed with LC-MS/MS.122 The approaches to predict peptides likely to be observed using MS are advantage of SISCAPA over common approaches is the possibility to clearly desirable. In this regard, a series of repositories that contain a combine the sensitivity and specificity of MRM analysis with the highly compendium of MS/MS proteomics data are available, such as selective enrichment capacity of antibodies to produce improved PeptideAtlas,130,131 SRMAtlas,132 or PRIDE (Protein Identifications detection limits.123 Database).133 PeptideAtlas is a database framework that contains In contrast to the AQUA approach, QconCAT technique, first peptides sequences and MS/MS data from proteomic experiments. described by Beynon and coworkers, is based on the use of artificial This resource is of public access, so that raw data from different proteins, which are concatemers of selected SIL proteotypic peptides scientists can be accepted into the repository. Such data are characteristic of the target proteins.108 This SIL standard protein is processed and reanalyzed, and finally gathered in a compendium usually synthesized through concatemerization, expressed from a library.134 In relation to PeptideAtlas, SRMAtlas is a database that plasmid synthetically prepared that contains the designed artificial contains specific information of proteomics assays based on the genes encoding the different target peptides, that are finally present in analysis of proteomes digests with tandem (SRM/MRM) MS, the product protein (formed as concatemer of peptides)124 (Fig. 6B). whereas PRIDE is a database that contains proteomics data on Nevertheless, there is an alternative approach named PolySIS protein/peptide identification as well as localization of PTMs. These (polyprotein stable-isotope labeled internal standard), which uses in public resources can facilitate the production of predictive vitro expression.125 After addition to the sample and tryptic digestion, algorithms for proteomic analysis, as well as the estimation of the the synthetic protein might provide a SIL standard peptide for each optimal proteotypic peptides when designing targeted proteomics natural proteotypic target peptide/protein, because they carry the experiments. To address these points in QconCAT, Eyers and 13 13 isotope label (usually [ C6] arginine and [ C6] lysine) from the coworkers have generated two complementary data sets of moment they were expressed. proteotypic peptides and developed CONSeQuence, a consensus One of the main concerns in QconCAT is to assure an identical prediction system built around four independent machine-learning digestion between the SIL designed QconCAT protein and the target algorithms to better understand peptide “detectability.”135 proteins.114 In those cases in which the protein sequence is not fully Because absolute quantification of proteins using labeled peptides preserved, and therefore the natural and the labeled proteins both is carried out through the measured intensity ratio between the natural have different amino acid sequences,126 the tryptic digestion and labeled peptides, and this ratio is related to the label peptide efficiency might be affected, especially if the differences in the concentration in order to determine the target protein concentration sequence surround the cleavage sites127 of the target peptides. The in the sample, spectral overlapping of both peptides must be avoided. importance to assure identical digestion conditions in order to obtain For that purpose, the isotopically labeled analog usually carries equimolarity of QconCAT peptides, has led to the inclusion of natural multiple tags. Nonetheless, this multi-isotopic labeling can lead to flanking sequences around the cleavage sites of the QconCAT protein isotopic effects and hence errors in the measurement of the intensity to better mimic the target protein.128 In 2007, the peptide-concatemer ratio; thereby methodological calibrations are needed to correlate the standards (PCS)129 strategy was developed for the accurate quantifi- intensity and molar ratios of both peptides.136 In this sense, González- cation of the stoichiometry of the components of the QconCAT Antuña and coworkers developed a strategy based on spectral protein with MS. In this strategy, these PCS standards do not only overlapping between the natural and isotopically labeled peptides, include the SIL tryptic peptides as potential internal standards for which are minimally labeled in 13C and named MOPs (mapped 728 | CALDERÓN-CELIS ET AL. overlapping peptides) to avoid isotopic effects. This strategy allowed biologically synthesized isotopically enriched proteins (peptides) for determination of the molar fractions of both peptide abundances quantification is SILAC (stable isotope-labeled amino acids for directly from the experimental mass spectra without need for quantification). In this approach, both forms of the proteins (light instrumental calibration.137 and heavy) are synthesized in two culture media, one of which contains a heavy form of a specific amino acid, usually [13C] Lys and [13C] Arg. Such SIL amino acid is incorporated into the proteins, which are likely 3.1.2 | Intact protein-based approaches synthesized under same chemical conditions than those unlabeled Whereas the use of SIL peptides as quantification standards in proteins in the “light” media with [12C] Lys or [12C] Arg. For further approaches such as AQUA and QconCAT is the main strategy today in quantitative purposes, SILAC synthetic proteins need the heavy form absolute quantitative shotgun proteomics, it has some limitations: of the amino acid to be completely incorporated into its sequence.141 incomplete coverage of the protein sequence, the need to characterize There are also other biological synthesis approaches described besides and control the proteolysis step efficiency, and the lack of control over SILAC, such as the PURE (protein synthesis using recombinant sample-preparation steps previous to digestion. Thereby, the natural elements) approach, which employs commercial kits for the cell-free alternative to overcome such drawbacks is the use of intact (full- in vitro protein expression (Fig. 6D)142,119]; or the FLEXIQuant (Full- length) SIL protein standards. Length Expressed Stable-Isotope-labeled Proteins for Quantification) The common workflow consists of the addition in known approach, in which the full-length labeled protein standards are concentrations of the synthetic SIL protein standard to the sample, expressed in vitro, after the gene of interest is cloned into a wheat germ at the earliest steps of the analytical process (eg, to the lysate, tissue, or expression vector, and transcribed and translated in a wheat germ biological fluid, that contain the sample proteins). After sample extract (WGE) (Fig. 6E).143 treatment (extraction, fractionation, etc), the protein extract is Likewise to the case of peptides, these synthesized SIL protein digested into peptides, followed by the LC-MS-MS/MS analysis. standards need to be characterized and certified in order to be used as Then, absolute quantities of the protein can be determined from the quantification standards and ensure quality results. Unsuitability of MS ratio of the light and heavy versions of the tryptic peptides. The AAA and limitations of CNL to intact protein quantification demand for main strategies that make use of SIL full-length proteins in absolute new strategies for full-length protein standards characterization and quantitative proteomics are PSAQ (Protein Standard Absolute certification. In this vein, proven capabilities of ICP-MS for the Quantification)39 and “Absolute SILAC.”138 Additionally, Wang and quantitative analysis of intact proteins75 place this methodology as a coworkers recently develop an strategy for absolute protein quantifi- feasible alternative, because lack of standards for protein certification cation based on the use of SILAC synthesized SIL proteins and non- result in the need for compound-independent approaches. In contrast labeled protein calibrators, named TAQSI (Targeted absolute quanti- to peptides, large number of amino acids in the protein sequence tative proteomics with SILAC internal standards).36 In this approach, entails the presence of at least one cysteine or methionine residue. absolute protein quantification is accomplished using the MS ratio of Consequently, certification of the SIL protein standard concentration heavy vs light proteolytic peptides of the proteins, and a calibration can be directly addressed through the ICP-MS detection of naturally curve built with the non-labeled protein calibrator set. present elements (principally S) and use of non-specific standards. As it is the case with peptide-based approaches, availability of the specific SIL protein standards is a limitation in protein-based 3.1.3 | Critical comparison of protein- vs peptide-based approaches for absolute proteomics. Full-length protein standard approaches cannot be synthesized with the SPPS stepwise procedure, because it is limited to small peptides (no more than 50-70 residues) as it is impeded As commented in the previous sections, both SIL peptide and protein by the accumulation of side-products from incomplete acylation and standards allow absolute quantitative proteomics. However, there are cleavage-related side reactions, as well as the time-consuming sources of error that affect final absolute protein quantitative results assembling of the amino acid chain. Thus, the synthesis of longer depending on the type of SIL standard used. Table 2 summarizes a peptides and proteins via chemical procedures has to be addressed by comparison of the main benefits and limitations of peptides or proteins coupling protected segments or, even more efficiently, by chemo- as standards for absolute quantification. selective ligation of non-protected purified sequences to enable a Quantitative approaches that measure directly intact proteins are faster assemblage of the peptide chain.139 This approach makes use of still limited today by the necessity of high mass-accuracy instruments reactive functionalities in unprotected ∼50-residue peptide segments (such as Fourier transform mass spectrometers, FTMS, or linear ion that form the target protein (Fig. 6C). The chemical ligation used to trap (LTQ-Orbitrap)), by the present lack of methods to efficiently bind them can consist, among others, on the formation of a thioester, fragment large proteins and, above all, by the difficulty to obtain an oxime, a directed disulfide, a thioether, a thiazolidine, or a peptide appropriate intact protein standards. Consequently, most quantitative bond in aqueous solution.118 proteomics approaches finally measure tryptic peptides, and make use Production of SIL proteins to be used as standards is more feasible of tandem-MS instruments.144,145 The corresponding SIL peptide through biological synthesis, and just requires the specific gene standards used for quantification, as shown in Fig. 1, can be directly encoding the protein to be available.124,139,140 The main approach to added to the sample (SIL peptides), or can be generated from a protein CALDERÓN-CELIS ET AL. | 729

TABLE 2 Benefits and limitations of using either SIL-peptides or SIL-intact proteins as standards in common mass spectrometry-based quantification approaches Benefits Limitations • Applicability to complex samples (thousands of proteins per experiment) • Tryptic digestion efficiency (existence of miscleavages) • More flexibility and accessibility to chemical modifications, and easier • Ambiguity on the origin of redundant peptide sequence check out sequences Peptides • Better chromatographic performance and resolution (front-end • Extra steps required to determine suitable proteotypic separation) and quantitative analytical recoveries peptides if not available in data bases • Enhanced MS ionization (higher sensitivity) • Need for unaltered properties after labeling • Easier synthetic preparation, purification, characterization, and analysis • Reactive or labile amino acids are to be avoided • Later addition does not correct for initial protein losses • If peptides are low abundant or badly ionized, MS information might lead to inaccurate quantification • The standard can be added to the crude sample at first stages of the • Likely non-quantitative analytical recoveries process: full characterization, and correction for losses of protein or (fractionation and/or separation steps) tryptic digestion incomplete efficiency • Complete characterization of the analytical process • Limited dynamic range (high-abundant proteins might hinder analysis of low-abundant ones) Proteins • Possibility to measure at the protein or peptide level (after digestion) • Structure has to be identical to the endogenous protein, and properties need to remain unaltered after labeling • If protein is digested after spiked, there would be more potentially MS- • Complicated synthesis labeling, purification, and detectable SIL peptides than if a single SIL peptide is added, thus quantitative certification quantification is statistically more reliable

In any case, the final MS measurement is considered to be carried out at the tryptic peptide level standard that undergoes tryptic digestion together with the protein collection of peptides of suitable size for MS analysis. A 1:1 ratio of samples (SIL proteins). the peptide to precursor protein is typically assumed. Therefore, if the Due to their higher flexibility and accessibility to chemical digestion is not efficient, and thus the proteotypic peptide is not modifications, peptides are most efficiently labeled than proteins. Of quantitatively released, it might affect profoundly the precision and course, standard peptides are much easier to obtain from chemical accuracy of the final quantification. Consequently, control of synthesis as well. Additionally, the use of peptides instead of proteins denaturing conditions during in-solution hydrolysis is crucial for provides a better front-end separation and higher sensitivity.146 reliable absolute quantification.147 In addition, inherent problems However, there are several key aspects to take into account when associated with protein digestion such as miscleavages at the sites peptides are used as standards for absolute-quantitative proteomics: related to the target peptide must be always considered. Another mainly the election of the standard, the enzymatic digestion process, problem associated with protein digestion is the chemical modification and the analytical recovery of the target proteotypic peptide. of the amino acid side chains during the proteolysis. This modification The election of a representative peptide of the protein under might compromise the results by changing the molecular mass of the study, which normally has to be unique to such protein (the so-called fragments and consequently the amount of them eventually detected proteotypic peptide), is critical because some amino acid sequences with MS. Moreover, experimental protocols must be optimized toward are common to multiple proteins or protein isoforms. Therefore, they high proteolysis kinetic rates in order to avoid bias resulting from, for cannot be used either as conclusive evidence for the presence of the instance, peptide modification proceeding faster in the SIL internal sought protein or for its specific absolute quantification. Possible standard added than in the same stretch of sequence within the intact PTMs present in the sequence should also be known in order to be protein, released as a fragment.148 used as signature to probe the quantitative state of a PTM associated Although there are strategies that can simulate the digestion with a particular biological state or disease setting. In addition, procedure in the peptide standard, by means of incorporating the peptides with reactive or labile amino acid residues should be avoided peptide flanking amino acid sequence (PCS strategy), they cannot such as those that can suffer oxidation (Met, Cys, and Trp), peptide simulate the entire molecular environment of the protein. The use of bond instability (Asp-Pro and Asp-Gly), deamidation (Asp, Glu), or intact proteins as (labeled) internal standards offer here some relevant cyclization (N-terminal Glu).22 advantages in comparison to the use of peptides (Table 2) to evaluate Another critical consideration in SIL peptide-based approaches is and normalize digestion efficiency. Addition of a isotopically labeled the needed proteolytic/enzymatic digestion of the target protein. This version of the sought protein can correct most of the sources of error digestion step involves decomposition of intact proteins into a during digestion procedures, besides from its usefulness as potential 730 | CALDERÓN-CELIS ET AL. standard for quantification.149 Nevertheless, intact protein as stand- absolute quantification provided by the use of SIL reference peptides ards require the labeled protein must be identical to the endogenous with SRM, label-free quantification, and high-throughput proteome protein and properly and correctly folded. Otherwise, inconsistent sequencing with LC-MS/MS. Such a combination minimizes the costs tryptic digestion and cleavage efficiency can be produced.146,148 of isotopic labeling while maximizing coverage and accuracy. Moreover, it is commonly accepted in quantitative Analytical On the other hand, another approach based on the combination of Chemistry that the sooner the internal standard is added to the sample isotope labels and label-free quantification was developed by Wepf and the better, because more inevitable errors during sample preparation coworkers.153 They quantified bait proteins via isotope-labeled refer- steps would be corrected for. Therefore, a labeled standard should ence peptides corresponded toan affinity tag sequence andprey proteins ideally represent the full-length protein, because its use would permit a with label-free correlational quantification with the precursor ion signal complete survey and unbiased quantitative analysis.150,151 This intensities of proteotypic peptides generated in reciprocal purifications. strategy enables a much better control of the analytical process Of course, proteotypic peptides must be firstly identified to apply this because the standard is added at the beginning of the process, and method. Even though such proteotypic peptides can be computationally shares almost identical biochemical and physicochemical properties predicted or extracted from proteomic databases, they must be with the target protein (Fig. 1). Thereby, even prefractionation and experimentally validated. Furthermore, the applicability of SIL reference digestion steps errors, losses, or procedure inefficiencies before peptides in the affinity purification coupled to tandem MS (AP-MS/MS) analysis can be corrected. Moreover, direct analysis of intact proteins workflow is again limited because their synthesis for a large number of could lead to a more-reliable quantification of the proteins isoforms network components is challenging, time-consuming, and expensive.53 and PTMs, because traditional limitations due to the use of peptides for the proteome characterization, including incomplete or ambiguous 3.2 | Stable-isotope labeled standardization characterization of alternative splice forms, PTMs and protein approaches with elemental mass spectrometry cleavages, can be overcome.152 The ICP-MS capacity to simultaneously measure different isotopes of an element enables absolute quantification of peptide/protein species 3.1.4 | The potential to combine label-free strategies by the direct measurement of the isotope ratios of the target element and isotope labelling in the compound of interest after appropriate addition of an enriched Ideally, peptide’s MS signal intensity should be proportional to the isotope solution, whose concentration and isotopic abundances are absolute abundance of the peptide (and of the corresponding protein) accurately known. This quantification is based on the concept of in the sample. However, such an assumption is, in fact, erroneous elemental isotope dilution analysis (IDA), which consists of the because of effects such as variable sequence-dependent peptide alteration of the isotope composition of the sample by addition of a ionization efficiencies, suppression by co-elution with dominant known quantity of an isotopically enriched form of the same element species, missing observations derived from peak selection in MS/MS (Fig. 7).155 From the measurement of the altered isotope ratios of the analyses, etc. These limitations explain why stable isotope labelling is target element, quantification can be accomplished. Requirements for the most-extensive approach for absolute proteomics. However, eventual accurate quantification include the ICP-MS capacity to although modern MS/MS measurements are sensitive and highly measure accurately and precisely small variations in that elemental reproducible across laboratories and platforms, and they could isotope composition. Of course, the isotopes involved in the ratio must theoretically be extended to a full proteome, preparation of thousands be measured free of spectral interferences. In addition, isotope- of isotopically labeled peptides of known concentration is both a dilution methods require for equilibration of the natural and formidable and expensive task. In this way, new methodologies based isotopically enriched elements, which is easily achieved in the ICP on the use of MS/MS and isotope labeling strategies, in combination plasma. with computational label-free strategies, are popular in order to There is a wide range of applications in quantitative proteomics accomplish a more complete absolute quantification coverage of based on the elemental IDA concept, mostly classified into two proteomes.153,35 principal workflows (see Fig. 3). The first one makes use of an In this regard, the average-intensity method estimates a protein isotopically enriched standard in the same chemical form of the analyte abundance from the average mass spectrum peak intensity of its three (species-specific IDA). On the contrary, the standard containing the best-detected peptides.154 The spectral-counting approach takes into sought element enriched in a heavy isotope is generic without being account the observed counts on a number of MS/MS spectra copies necessary to be in the same chemical form than the analyte in the attributable to each protein43 and was improved latter by incorporat- second workflow (species-unspecific or online IDA). ing differential peptide ionization into the computation.53 Malmström and coworkers35 have combined these three concepts in a new 3.2.1 | Species-specific isotope dilution approaches approach that considers: (i) SRM measurements of a limited set of SIL internal reference standards; (ii) the average mass spectrum signal In the first case, as in targeted SIL quantitative proteomics with intensities of the top three peptides selected per protein; and (iii) molecular MS approaches, the labeled or synthetic isotopically weighed MS/MS spectral counts. Thereby, the authors combined the enriched form of the target species is spiked into the sample. If this CALDERÓN-CELIS ET AL. | 731

FIGURE 7 A) Concept of isotope dilution, adapted from Rodríguez-González et al.156 A sample with natural isotopic abundance of the target compound or element is mixed with a spike isotopically enriched in a non-abundant isotope in a known isotopic composition. The final mixture is analyzed, the concentration of the sample can be calculated from the abundances of the measured isotopes. B) Isotope dilution equation in species-specific IDA approaches that express the moles of target element in a sample (s), as a function of the moles of element in the isotopically enriched spike (sp), the experimental ratios of two isotopes (1 and 2) of the element in a mixture of sample and spike (Rm), and the element 1 2 1 2 isotopic abundances in the sample, A S and A S (ie, IUPAC isotopic abundances), and in the spike, A Sp and A Sp (therefore requiring previous characterization). C) Mass flow equation for species-unspecific IDA approaches. The sample’s mass flow (MFs) is expressed as function of the element concentration in the spike (Csp), the density and flow of the spike (dsp and fsp), the atomic weights of the element in the sample and 1 2 156 the spike (Aws and Awsp), Rm, A S,A Sp,and the isotope ratios in the sample and spike (Rs and Rsp), adapted from Rodríguez-González et al

step is done at the beginning of the process, then errors, losses, or continuous addition, commonly after the on-column separation, of a matrix effects that occur along all the analytical processes, including generic species isotopically enriched in the detected heteroelement, in sample preparation, digestion, degradation, chromatographic separa- a known concentration and isotopic composition. Isotopic equilibra- tion, and ICP-MS measurement can be corrected for. In fact, the target tion required for the IDA computation between the element- species and its corresponding isotopically labeled standard will arrive containing target species separated during chromatography and the to the ICP source simultaneously to allow for the required elemental generic isotopically enriched element-containing species added and isotopic equilibration. In brief, the elemental detection of the different mixed on-line with the chromatographic eluent is achieved again in the isotopes of the sought element gives an isotopic ratio. Taking into ICP plasma. Thereupon, peptide/protein absolute quantification can account that the elemental concentration of the added species- be carried out with on-line continuous monitoring (element mass/time) specific spike is known as well as the isotope abundances of natural of the isotope ratio, which can be transformed in a mass-flow and isotope-labeled versions of the compound (or measured chromatogram with the mass flow equation, derived from the isotope- previously), determination of the elemental concentration can be dilution equation. Integration of the chromatographic peaks of the accomplished with this simple elemental isotope ratio measurement different (previously separated) species allows one to calculate directly with the so-called “isotope dilution equation”.69,155 Such elemental the mass of the element.156 This common application of the concept of concentrations accurately computed can be translated into biomole- on-line IDA162,163 can be used for absolute quantification of peptides/ cule concentration given the known stoichiometry heteroatom/ proteins. Again, it can be applied to elements naturally present, such as biomolecule. This approach can be used in quantitative proteomics iron164 or sulfur,165,166 or intentionally added after labeling with measuring isotope ratios in elements either naturally present in the usually lanthanide-containing species such as europium167 or lute- peptides/proteins, or tagged to them, usually with reagent molecules. tium.168 This controlled labeling enables quantification of any peptide Hence, absolute quantification has been reported for either amino species, unlike direct quantification with naturally present elements, in acids,157,158 peptides,159 and even full-length proteins.160,161 which only peptides or proteins that contain that target element can be quantified. Given the virtually species-independent nature of ICP ionization, and that the spike is normally added after the chro- 3.2.2 | Species-unspecific isotope dilution approaches matographic separation, this spike does not require the same chemical On the other hand, quantitative proteomics with species-unspecific form than that of the studied species. Consequently, several different isotope dilution using ICP-MS detection is based on use of an online peptide/protein species can be quantified in one analysis and without 732 | CALDERÓN-CELIS ET AL. the need for specific (labelled or not) standards. Nevertheless, the highly accurate MS/MS instrumentation (currently considered as class identity of the species is finally necessary in order to translate the I -low risk- instrumentation by the FDA). However, there is still quantified amount of element in each chromatographic peak into insufficient quality control in MS-based approaches to hinder their compound quantity, usually accomplished with parallel molecular mass definitive settlement in legislated clinical protocols (eg, diagnosis or spectrometry approaches. In that sense, Fig. 4 shows a model disease monitoring), mainly due to the limited availability of properly workflow that described the synergic combination of elemental MS, validated quantification standards and (certified) reference materials which provides quantitative information, and molecular MS, enabling that could be used as controls.20,21 identity and sequence characterization of the species, as has already The use of SIL proteotypic peptide standards remains as the been tested for the absolute quantitative profiling of an assayed approach of reference in absolute quantitative proteomics with MS. proteome.75 This methodology represents a robust and accurate way to determine In brief, these species-unspecific IDA-based ICP-MS approaches peptide/protein absolute quantities in complex samples and can be hold great potential for the determination of biomolecules for which used even for very low-abundant analytes when combined with no specific standards are available or when the synthesis of such enrichment/purification methods like SISCAPA. This kind of quantita- standards is too complex or expensive. In fact, ICP-MS capabilities can tive approach requires, however, selection of the standard with MS or be of high relevance in quality control and proteomics standard MS/MS analysis prior to the quantitative assay. In this sense, several certification. Absolute quantitative data can become directly traceable open resources, such as databases and peptides libraries, have to the generic element-containing certified standard added to provide emerged recently that entail a significant aid in such task to facilitate consistency of the results and allow for quality-assurance protocols. Of the design of quantification assays and protocols in proteomics. course, such ICP-MS-based approaches are not intended for Moreover, the possibility to use full-length intact proteins as SIL comprehensive absolute quantification of the whole set of proteins/ standards prevents some of the limitations with peptides; for example, peptides in complex samples, but can be effectively applied to simpler the need to characterize and control the proteolysis step efficiency, samples such as the SIL peptide/protein standards produced (see and the lack of control over sample preparation steps previous to section 3.1) that, once accurately certified, could also be used as digestion. Yet, the use of either SIL peptides or proteins entails the use internal standards for massive absolute quantifications with LC-MS/ of stable isotope labeling and thus their use is restricted to availability. MS. Furthermore, the scope of those ICP-MS-based approaches has Alternatively, the use of isobaric tags as labels in quantitative been recently extended to moderately complex samples after proteomics is acknowledged mainly due to its potential multiplexing demonstration of their potential to simultaneously identify and capabilities for mostly relative but also absolute quantification of absolute quantify the proteins of the venoms of elapid snakes. peptides and proteins. Absolute quantification provided led to the precise determination of Synthetic production of quantitative standards (labeled or not) can the concentration of the different toxins (ie, 25-30) in the venom be a simple task because there are a large number of established without the need for any specific standard.79 protocols for peptide (and even protein) synthesis. However, the limited availability of standards for each specific peptide/protein target and robust strategies for their proper characterization are 4 | CONCLUDING REMARKS AND FUTURE obvious limitations. The outstanding parallel growth of label-free PROSPECTS approaches, which are based on computational treatment of the MS data for quantification instead of the use of specific peptide/protein Immunoassay-based techniques for absolute quantitative proteomics standards, turns them into potential competitors for extensive have been stablished as the methods of reference for decades, mainly quantitative proteomics in the near future. It should be stressed due to their sensitivity, selectivity, and experimental simplicity with that, although label-free quantifications are still plagued today with already-stablished protocols. These approaches are limited not only to analytical drawbacks especially in terms of reproducibility and specific antibodies availability but also to protein standards of the accuracy, it is reasonable to believe that their successful future is corresponding targets required to build the immune calibration curves. engaged to the specific developments of more-efficient bioinformatics Therefore, the development of new protocols for quantitative study of tools. new target proteins, such as biomarker potential candidates, can Finally, the certification of the quantification standards is become a laborious and long-lasting task. In this sense, MS has commonly approached with established but cumbersome and emerged as a powerful technique able to quantitatively characterize up prone-to-error techniques. In this regard, the use of ICP-MS can to thousands of proteins in one experiment. Absolute MS-based provide a simple, fast, and accurate method to certify the mass purity approaches have thus been described as the techniques to fill the of such standards. Moreover, the recent synergistic combination of existing gap in the process from biomarker discovery to validation, in ICP-MS with molecular mass spectrometry has provided absolute which there is a lack of quantitative immunoassay-based protocols.169 quantification of intact proteins even in relatively complex samples In fact, the potential multiplex for large-scale protein quantification without the need for any species-unspecific standards. However, its could eventually lead to its implementation in clinical validation of quantitative capacity is constrained by the resolving power of coupled biomarkers, which has been further reinforced by the development of separation techniques. A major challenge ahead is to develop a robust CALDERÓN-CELIS ET AL. | 733 platform to achieve pre-MS separation of each and every protein 19. Aebersold R, Burlingame AL, Bradshaw RA. Western blots versus species in the sample. On the other hand, the combination of the selected reaction monitoring assays: time to turn the tables? Mol Cell Proteomics. 2013;12:2381–2382. advantages of ICP-MS detection with the selectivity of element- 20. Regnier FE, Skates SJ, Mesri M, et al. Protein-based multiplex assays: tagged antibodies provides a most-promising tool to tackle the mock presubmissions to the US food and drug administration. Clin absolute, sensitive, and robust determination of a few targeted Chem. 2010;56:165–171. proteins (eg, panel of biomarkers required for diagnosis and/or 21. Boja ES, Fehniger TE, Baker MS, Marko-Varga G, Rodriguez H. Analytical validation considerations of multiplex mass-spectrometry- prognosis of a given disease). based proteomic platforms for measuring protein biomarkers. J Proteome Res. 2014;13:5325–5332. ACKNOWLEDGMENTS 22. Pan S, Aebersold R, Chen R, et al. Mass spectrometry based targeted protein quantification: methods and applications. J Proteome Res. Authors acknowledge financial support of Spanish Ministry of 2009;8:787–797. Economy and Competitiveness through CTQ2016-79412-P and 23. Hawkridge AM. Practical considerations and current limitations in quantitative mass spectrometry-based proteomics. In: Eyers CE, BES-2014-068032 (F.C.C). We thank Prof. Dominic M. Desiderio Gaskell S, eds. Quantitative Proteomics. Cambridge, UK: Royal Society (University of Tennessee Center for Health Science) and several of Chemistry; 2014. 3–25. anonymous reviewers for constructive comments that improved the 24. Hortin GL. The MALDI-TOF mass spectrometric view of the plasma – original manuscript. proteome and peptidome. Clin Chem. 2006;52:1223 1237. 25. Gessel MM, Norris JL, Caprioli RM. MALDI imaging mass spectrom- etry: spatial molecular analysis to enable a new age of discovery. – REFERENCES J Proteomics. 2014;107:71 82. 26. Cornett DS, Reyzer ML, Chaurand P, Caprioli RM. MALDI imaging 1. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis mass spectrometry: molecular snapshots of biochemical systems. of the human genome. Nature. 2001;409:860–921. Nat Methods. 2007;4:828–833. 2. Venter JC, Adams MD, Myers EW, et al. The sequence of the human 27. Franck J, Arafah K, Elayed M, et al. MALDI imaging mass genome. Science. 2001;291:1304–1351. spectrometry: state of the art technology in clinical proteomics. 3. Liebler DC, Yates JR. Introduction to Proteomics. Tools For the New Mol Cell Proteomics. 2009;8:2023–2033. Biology. Totowa, NJ:Humana Press; 2012. 28. Hajduk J, Matysiak J, Kokot ZJ. Challenges in biomarker discovery 4. Muñoz J, Heck AJR. From the human genome to the human with MALDI-TOF MS. Clin Chim Acta. 2016;458:84–98. proteome. Angew Chemie Int Ed. 2014;53:10864–10866. 29. Łobiński R, Schaumlöffel D, Szpunar J. Mass spectrometry in bioinorganic 5. Jeong H, Mason SP, Barabási A-L, Oltvai ZN. Lethality and centrality analytical chemistry. Mass Spectrom Rev. 2006;25:255–289. in protein networks. Nature. 2001;411:41–42. 30. Sanz-Medel A, Montes-Bayón M, del Rosario Fernández de la Campa 6. Anderson NL, Anderson NG. The human plasma proteome: history, M, Ruiz Encinar J, Bettmer J. Elemental mass spectrometry for character, and diagnostic prospects. MolCellProteomics. 2002;1:845–867. quantitative proteomics. Anal Bioanal Chem. 2008;390:3–16. 7. Kim M-S, Pinto SM, Getnet D, et al. A draft map of the human 31. Ong S-E, Mann M. Mass spectrometry-based proteomics turns proteome. Nature. 2014;509:575–581. quantitative. Nat Chem Biol. 2005;1:252–262. 8. Wilhelm M, Schlegl J, Hahne H, et al. Mass-spectrometry-based draft 32. Brönstrup M. Absolute quantification strategies in proteomics based of the human proteome. Nature. 2014;509:582–587. on mass spectrometry. Expert Rev Proteomics. 2004;1:503–512. 9. Hoofnagle AN, Aebersold R, Anderson NL, Felsenfeld A, Liebler DC. 33. Bantscheff M, Lemeer S, Savitski MM, Kuster B. Quantitative mass Painting a moving picture: large-scale proteomics efforts and their spectrometry in proteomics: critical review update from 2007 to the potential for changing patient care. Clin Chem. 2011;57:1357–1360. present. Anal Bioanal Chem. 2012;404:939–965. 10. O’Farrell PH. High resolution two-dimensional electrophoresis of 34. Zhang Y, Fonslow BR, Shan B, Baek M, Yates JR. Protein analysis by proteins. J Biol Chem. 1975;250:4007–4021. shotgun/bottom-up proteomics. Chem Rev. 2013;113:2343–2394. 11. Von Eggeling F, Gawriljuk A, Fiedler W, et al. Fluorescent dual colour 35. Malmström J, Beck M, Schmidt A, Lange V, Deutsch EW, Aebersold 2D-protein gel electrophoresis for rapid detection of differences in R. Proteome-wide cellular protein concentrations of the human protein pattern with standard image analysis software. Int J Mol Med. pathogen Leptospira interrogans. Nature. 2009;460:762–765. 2001;8:373–377. 36. Wang X, Liang Y, Liu L, Shi J, Zhu H-J. Targeted absolute quantitative 12. Yalow RS, Berson SA. Immunoassay of endogenous plasma insulin in proteomics with SILAC internal standards and unlabeled full-length man. J Clin Invest. 1960;39:1157–1175. protein calibrators (TAQSI). Rapid Commun Mass Spectrom. 2016;30: 13. MacBeath G. Protein microarrays and proteomics. Nat Genet. 553–561. 2002;32:526–532. 37. Schmidt C, Urlaub H. Absolute quantification of proteins using St and 14. Engvall E, Perlmann P. Enzyme-linked immunosorbent assay (ELISA) ard peptides and multiple reaction monitoring. In: Marcus K, ed. Quantitative assay of immunoglobulin G. Immunochemistry. 1971;8: Quantitative Methods in Proteomics. Totowa, NJ: Humana Press; 871–874. 2012. 249–265. 15. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins 38. Kirkpatrick DS, Gerber SA, Gygi SP. The absolute quantification from polyacrylamide gels to nitrocellulose sheets: procedure and strategy: a general procedure for the quantification of proteins and some applications. Proc Natl Acad Sci. 1979;76:4350–4354. post-translational modifications. Methods. 2005;35:265–273. 16. Aebersold R, Mann M. Mass spectrometry-based proteomics. 39. Brun V, Dupuis A, Adrait A, et al. Isotope-labeled protein standards: Nature. 2003;422:198–207. toward absolute quantitative proteomics. Mol Cell Proteomics. 17. Domon B, Aebersold R. Mass spectrometry and protein analysis. 2007;6:2139–2149. Science. 2006;312:212–217. 40. Cox J, Hein MY, Luber Ca, Paron I, Nagaraj N, Mann M. Accurate 18. Hoofnagle AN, Wener MH. The fundamental flaws of immunoassays proteome-wide label-free quantification by delayed normalization and potential solutions using tandem mass spectrometry. J Immunol and maximal peptide ratio extraction, Termed MaxLFQ. Mol Cell Methods. 2009;347:3–11. Proteomics. 2014;13:2513–2526. 734 | CALDERÓN-CELIS ET AL.

41. Maes E, Hadiwikarta WW, Mertens I, Baggerman G, Hooyberghs J, 61. Wiese S, Reidegeld KA, Meyer HE, Warscheid B. Protein labeling by Valkenborg D. CONSTANd: a normalization method for isobaric iTRAQ: a new tool for quantitative mass spectrometry in proteome labeled spectra by constrained optimization. Mol Cell Proteomics. research. Proteomics. 2007;7:340–350. 2016;15:2779–2790. 62. Fuller HR, Morris GE. Quantitative proteomics using iTRAQ labeling 42. Chelius D, Bondarenko PV. Quantitative profiling of proteins in and mass spectrometry. In: Leung H-C, ed. Integrative Proteomics. complex mixtures using liquid chromatography and mass spectrom- Rijeka, Croatia: InTech; 2012. etry. J Proteome Res. 2002;1:317–323. 63. Quaglia M, Pritchard C, Hall Z, O’Connor G. Amine-reactive isobaric 43. Liu H, Sadygov RG, Yates JR. A model for random sampling and tagging reagents: requirements for absolute quantification of estimation of relative protein abundance in shotgun proteomics. Anal proteins and peptides. Anal Biochem. 2008;379:164–169. Chem. 2004;76:4193–4201. 64. Evans C, Noirel J, Ow SY, et al. An insight into iTRAQ: where do we 44. Neilson KA, Ali NA, Muralidharan S, et al. Less label, more free: stand now? Anal Bioanal Chem. 2012;404:1011–1027. approaches in label-free quantitative mass spectrometry. Proteomics. 65. Thompson A, Schäfer J, Kuhn K, et al. Tandem mass tags: a novel 2011;11:535–553. quantification strategy for comparative analysis of complex protein 45. Schilling B, Rardin MJ, MacLean BX, et al. Platform-independent and mixtures by MS/MS. Anal Chem. 2003;75:1895–1904. label-free quantitation of proteomic data using MS1 extracted ion 66. McAlister GC, Huttlin EL, Haas W, et al. Increasing the multiplexing chromatograms in skyline: application to protein acetylation and capacity of TMTs using reporter ion isotopologues with isobaric phosphorylation. Mol Cell Proteomics. 2012;11:202–214. masses. Anal Chem. 2012;84:7469–7478. 46. Escher C, Reiter L, MacLean B, et al. Using iRT, a normalized retention 67. Werner T, Sweetman G, Savitski MF, Mathieson T, Bantscheff M, time for more targeted measurement of peptides. Proteomics. 2012; Savitski MM. Ion coalescence of neutron encoded TMT 10-Plex 12:1111–1121. reporter ions. Anal Chem. 2014;86:3594–3601. 47. Holman SW, McLean L, Eyers CE. RePLiCal: a QconCAT protein for 68. Dayon L, Turck N, Scherl A, Hochstrasser DF, Burkhard PR, Sanchez retention time standardization in proteomics studies. J Proteome Res. J-C. From relative to absolute quantification of tryptic peptides with 2016;15:1090–1102. tandem mass tags: application to cerebrospinal fluid. Chim Int J Chem. 48. Gallien S, Peterman S, Kiyonami R, et al. Highly multiplexed targeted 2010;64:132–135. proteomics using precise control of peptide retention time. 69. Wang M, Feng W-Y, Zhao Y-L, Chai Z-F. ICP-MS-based strategies for Proteomics. 2012;12:1122–1133. protein quantification. Mass Spectrom Rev. 2010;29:326–348. 49. Mirzaei H, Brusniak M-Y, Mueller LN, Letarte S, Watts JD, Aebersold 70. Sanz-Medel A, Montes-Bayón M, Bettmer J, Fernández-Sanchez ML, R. Halogenated peptides as internal standards (H-PINS): introduction Ruiz Encinar J. ICP-MS for absolute quantification of proteins for of an MS-based internal standard set for liquid chromatography- heteroatom-tagged, targeted proteomics. TrAC Trends Anal Chem. mass spectrometry. Mol Cell Proteomics. 2009;8:1934–1946. 2012;40:52–63. 50. Tabata T, Sato T, Kuromitsu J, Oda Y. Pseudo internal standard 71. Svantesson E, Pettersson J, Markides KE. The use of inorganic approach for label-free quantitative proteomics. Anal Chem. 2007; elemental standards in the quantification of proteins and biomolec- 79:8440–8445. ular compounds by inductively coupled plasma spectrometry. J Anal 51. Staunton L, Clancy T, Tonry C, et al. In: Eyers CE, Gaskell S, eds. At Spectrom. 2002;17:491–496. Quantitative Proteomics. Cambridge: Royal Society of Chemistry; 72. Diez Fernández S, Sugishama N, Ruiz Encinar J, Sanz-Medel A. Triple 2014. quad ICPMS (ICPQQQ) as a new tool for absolute quantitative 52. Ishihama Y, Oda Y, Tabata T, et al. Exponentially modified protein proteomics and phosphoproteomics. Anal Chem. 2012;84: abundance index (emPAI) for estimation of absolute protein amount 5851–5857. in proteomics by the number of sequenced peptides per protein. Mol 73. Hare DJ, Grubman A, Ryan TM, et al. Profiling the iron, copper and Cell Proteomics. 2005;4:1265–1272. zinc content in primary neuron and astrocyte cultures by rapid online 53. Lu P, Vogel C, Wang R, Yao X, Marcotte EM. Absolute protein quantitative size exclusion chromatography-inductively coupled expression profiling estimates the relative contributions of tran- plasma-mass spectrometry. Metallomics. 2013;5:1656–1662. scriptional and translational regulation. Nat Biotechnol. 2007;25: 74. Hare DJ, Roberts BR, McColl G. Profiling changes to natively-bound 117–124. metals during Caenorhabditis elegans development. RSC Adv. 54. Arike L, Valgepea K, Peil L, Nahku R, Adamberg K, Vilu R. Comparison 2016;6:113689–113693. and applications of label-free absolute proteome quantification 75. Calderón-Celis F, Diez-Fernández S, Costa-Fernández JM, Ruiz methods on Escherichia coli. J Proteomics. 2012;75:5437–5448. Encinar J, Calvete JJ, Sanz-Medel A. Elemental mass spectrometry 55. Fabre B, Lambour T, Bouyssié D, et al. Comparison of label-free for absolute intact protein quantification without protein-specific quantification methods for the determination of protein complexes standards: application to snake venomics. Anal Chem. 2016;88: subunits stoichiometry. EuPA Open Proteomics. 2014;4:82–86. 9699–9706. 56. Schwanhäusser B, Busse D, Li N, et al. Global quantification of 76. Sarmiento-González A, Ruiz Encinar J, Cantarero-RoldánAM, mammalian gene expression control. Nature. 2011;473:337–342. Marchante-Gayón JM, Sanz-Medel A. HPLC−ICPMS and stable 57. Winter D, Seidler J, Kugelstadt D, Derrer B, Kappes B, Lehmann WD. isotope-labeled approaches to assess quantitatively Ti(IV) uptake Minimally permutated peptide analogs as internal standards for by transferrin in human blood serum. Anal Chem. 2008;80: relative and absolute quantification of peptides and proteins. 8702–8711. Proteomics. 2010;10:1510–1514. 77. Wind M, Wegener A, Eisenmenger A, Kellner R, Lehmann WD. Sulfur 58. Ross PL, Huang YN, Marchese JN, et al. Multiplexed protein as the key element for quantitative protein analysis by capillary liquid quantitation in saccharomyces cerevisiae using amine-reactive chromatography coupled to element mass spectrometrysulfur as the isobaric tagging reagents. Mol Cell Proteomics. 2004;3:1154–1169. key element for quantitative protein analysis by capillary liquid 59. Dayon L, Hainard A, Licker V, et al. Relative quantification of proteins chromatography coupled to element mass spect. Angew Chemie Int in human cerebrospinal fluids by MS/MS using 6-Plex isobaric tags. Ed. 2003;42:3425–3427. Anal Chem. 2008;80:2921–2931. 78. Szpunar J. Advances in analytical methodology for bioinorganic 60. Hahne H, Neubert P, Kuhn K, et al. Carbonyl-reactive tandem mass speciation analysis: metallomics, metalloproteomics and hetero- tags for the proteome-wide quantification of N-linked glycans. Anal atom-tagged proteomics and metabolomics. Analyst. 2005;130: Chem. 2012;84:3716–3724. 442–465. CALDERÓN-CELIS ET AL. | 735

79. Calderón-Celis F, Cid-Barrio L, Ruiz Encinar J, Sanz-Medel A, Calvete 98. Peng H, Chen B, He M, Zhang Y, Hu B. Magnetic quantitative JJ. Absolute venomics: absolute quantification of intact venom immunoanalysis of carcinoembryonic antigen by ICP-MS with proteins through elemental mass spectrometry. J Proteomics. mercury labels. J Anal At Spectrom. 2011;26:1217–1223. 2017;164:33–42. 99. Liu R, Wu P, Yang L, Hou X, Lv Y. Inductively coupled plasma mass 80. Maes E, Tirez K, Baggerman G, et al. The use of elemental mass spectrometry-based immunoassay: a review. Mass Spectrom Rev. spectrometry in phosphoproteomic applications. Mass Spectrom Rev. 2014;33:373–393. 2016;35:350–360. 100. López-Serrano A, Muñoz Olivas R, Sanz Landaluze J, Cámara C. 81. Pereira Navaza A, Ruiz Encinar J, Carrascal M, Abián J, Sanz-Medel A. Nanoparticles: a global vision. Characterization, separation, and Absolute and site-specific quantification of protein phosphorylation quantification methods. Potential environmental and health impact. using integrated elemental and molecular mass spectrometry: its Anal Methods. 2014;6:38–56. potential to assess phosphopeptide enrichment procedures. Anal 101. Garcia-Cortes M, Ruiz Encinar J, Costa-Fernandez JM, Sanz-Medel Chem. 2008;80:1777–1787. A. Highly sensitive nanoparticle-based immunoassays with elemen- 82. Wind M, Wesch H, Lehmann WD. Protein phosphorylation degree: tal detection: application to Prostate-Specific Antigen quantification. determination by capillary liquid chromatography and inductively Biosens Bioelectron. 2016;85:128–134. coupled plasma mass spectrometry. Anal Chem. 2001;73: 102. Desiderio DM, Kai M. Preparation of stable isotope-incorporated 3006–3010. peptide internal standards for field desorption mass spectrometry 83. Pereira Navaza A, Ruiz Encinar J, Sanz-Medel A. Absolute and quantification of peptides in biologic tissue. Biol Mass Spectrom. accurate quantification of protein phosphorylation by using an 1983;10:471–479. elemental phosphorus standard and element mass spectrometry. 103. Zhang R, Regnier FE. Minimizing resolution of isotopically coded Angew Chemie Int Ed. 2007;46:569–571. peptides in comparative proteomics. J Proteome Res. 2002;1: 84. Zinn N, Hahn B, Pipkorn R, Schwarzer D, Lehmann WD. Phosphorus- 139–147. based absolutely quantified standard peptides for quantitative 104. Zhang R, Sioma CS, Wang S, Regnier FE. Fractionation of isotopically proteomics. J Proteome Res. 2009;8:4870–4875. labeled peptides in quantitative proteomics. Anal Chem. 2001;73: 85. Zinn N, Winter D, Lehmann WD. Recombinant isotope labeled and 5142–5149. selenium quantified proteins for absolute protein quantification. Anal 105. Mallick P, Kuster B. Proteomics: a pragmatic perspective. Nat Chem. 2010;82:2334–2340. Biotechnol. 2010;28:695–709. 86. Tholey A, Schaumlöffel D. Metal labeling for quantitative protein and 106. Zhang Y, Askenazi M, Jiang J, Luckey CJ, Griffin JD, Marto JA. proteome analysis using inductively-coupled plasma mass spectrom- A robust error model for iTRAQ quantification reveals divergent etry. TrAC Trends Anal Chem. 2010;29:399–408. signaling between oncogenic FLT3 mutants in acute myeloid 87. Sanz-Medel A. Heteroatom-tagged” quantification of proteins via leukemia. Mol Cell Proteomics. 2010;9:780–790. ICP-MS. Anal Bioanal Chem. 2016;408:5393–5395. 107. Gerber SA, Rush J, Stemman O, Kirschner MW, Gygi SP. Absolute 88. Bendall SC, Simonds EF, Qiu P, et al. Single-cell mass cytometry of quantification of proteins and phosphoproteins from cell lysates by differential immune and drug responses across a human hematopoi- tandem MS. Proc Natl Acad Sci. 2003;100:6940–6945. etic continuum. Science. 2011;332:687–696. 108. Beynon RJ, Doherty MK, Pratt JM, Gaskell SJ. Multiplexed absolute 89. Ahrends R, Pieper S, Kuhn A, et al. A metal-coded affinity tag quantification in proteomics using artificial QCAT proteins of approach to quantitative proteomics. Mol Cell Proteomics. 2007;6: concatenated signature peptides. Nat Methods. 2005;2:587–589. 1907–1916. 109. Mallick P, Schirle M, Chen SS, et al. Computational prediction of 90. Bergmann U, Ahrends R, Neumann B, Scheler C, Linscheid MW. proteotypic peptides for quantitative proteomics. Nat Biotechnol. Application of metal-coded affinity tags (MeCAT): absolute protein 2007;25:125–131. quantification with top-down and bottom-up workflows by metal- 110. Merrifield RB. Solid phase peptide synthesis. I. The synthesis of a coded tagging. Anal Chem. 2012;84:5268–5275. tetrapeptide. J Am Chem Soc. 1963;85:2149–2154. 91. Guo Y, Chen L, Yang L, Wang Q. Counting sulfhydryls and disulfide 111. Amblard M, Fehrentz J-A, Martinez J, Subra G. Methods and bonds in peptides and proteins using mercurial ions as an MS-tag. protocols of modern solid phase peptide synthesis. Mol Biotechnol. J Am Soc Mass Spectrom. 2008;19:1108–1113. 2006;33:239–254. 92. Kutscher DJ, Bettmer J. Absolute and relative protein quantification 112. Rutherfurd SM, Gilani GS. Amino acid analysis. Curr Protoc Protein Sci. with the use of isotopically labeled p-Hydroxymercuribenzoic acid 2009;56:11.9.1–11.9.37. and complementary MALDI-MS and ICPMS detection. Anal Chem. 113. Fountoulakis M, Lahm H-W. Hydrolysis and amino acid composition 2009;81:9172–9177. analysis of proteins. J Chromatogr A. 1998;826:109–134. 93. Pereira Navaza A, Ruiz Encinar J, Ballesteros A, González JM, Sanz- 114. Mirzaei H, McBee JK, Watts J, Aebersold R. Comparative evaluation Medel A. Capillary HPLC-ICPMS and tyrosine iodination for the of current peptide production platforms used in absolute quantifica- absolute quantification of peptides using generic standards. Anal tion in proteomics. Mol Cell Proteomics. 2008;7:813–823. Chem. 2009;81:5390–5399. 115. Fujinari EM, Manes JD, Bizanek R. Peptide content determination of 94. Montoro Bustos AR, Trapiella-Alfonso L, Ruiz Encinar J, Costa- crude synthetic peptides by reversed-phase liquid chromatography Fernández JM, Pereiro R, Sanz-Medel A. Elemental and molecular and nitrogen-specific detection with a chemiluminescent nitrogen detection for Quantum Dots-based immunoassays: a critical detector. J. Chromatogr. A. 1996;743:85–89. appraisal. Biosens Bioelectron. 2012;33:165–171. 116. Coin I, Beyermann M, Bienert M. Solid-phase peptide synthesis: from 95. Zhang C, Wu F, Zhang Y, Wang X, Zhang X. A novel combination of standard procedures to the synthesis of difficult sequences. Nat immunoreaction and ICP-MS as a hyphenated technique for the Protoc. 2007;2:3247–3256. determination of thyroid-stimulating hormone (TSH) in human 117. Sarin VK, Kent SBH, Tam JP, Merrifield RB. Quantitative monitoring serum. J Anal At Spectrom. 2001;16:1393–1396. of solid-phase peptide synthesis by the ninhydrin reaction. Anal 96. Lou X, Zhang G, Herrera I, et al. Polymer-based elemental tags for Biochem. 1981;117:147–157. sensitive bioassays. Angew Chemie Int Ed. 2007;46:6111–6114. 118. Dawson PE, Kent SBH. Synthesis of native proteins by chemical 97. Majonis D, Herrera I, Ornatsky O, et al. Synthesis of a functional ligation. Annu Rev Biochem. 2000;69:923–960. metal-chelating polymer and steps toward quantitative mass 119. Xian F, Zi J, Wang Q, et al. Peptide biosynthesis with stable isotope cytometry bioassays. Anal Chem. 2010;82:8961–8969. labeling from a cell-free expression system for targeted proteomics 736 | CALDERÓN-CELIS ET AL.

with absolute quantification. Mol Cell Proteomics. 2016;15: 139. Miranda LP, Alewood PF. Accelerated chemical synthesis of peptides 2819–2828. and small proteins. Proc Natl Acad Sci. 1999;96:1181–1186. 120. Singh S, Kirchner M, Steen JA, Steen H. A practical guide to the 140. Lebert D, Dupuis A, Garin J, Bruley C, Brun V. Production of use of FLEXIQuant method. In: Marcus K, ed. Quantitative Methods in stable isotope-labeled proteins for absolute quantitative proteomics. Proteomics. Totowa, NJ: Humana Press; 2012;295–319. In: Gevaert K, Vandekerckhove J, eds. Gel-Free Proteomics. Methods 121. Anderson NL, Anderson NG, Haines LR, Hardie DB, Olafson RW, and Protocols. Totowa, NJ: Humana Press; 2011. 93–115. Pearson TW. Mass spectrometric quantitation of peptides and 141. Ong S-E, Mann M. A practical recipe for stable isotope labeling by proteins using stable isotope standards and capture by anti-peptide amino acids in cell culture (SILAC). Nat Protoc. 2007;1:2650–2660. antibodies (SISCAPA). J Proteome Res. 2004;3:235–244. 142. Tuckey C, Asahara H, Zhou Y, Chong S. Protein synthesis using a 122. Whiteaker JR, Zhao L, Anderson L, Paulovich AG. An automated and reconstituted cell-free system. Curr Protoc Mol Biol. 2014;108: multiplexed method for high throughput peptide immunoaffinity 16.31.1–16.31.22. enrichment and multiple reaction monitoring mass spectrometry- 143. Singh S, Springer M, Steen J, Kirschner MW, Steen H. FLEXIQuant: a based quantification of protein biomarkers. Mol Cell Proteomics. novel tool for the absolute quantification of proteins, and the 2010;9:184–196. simultaneous identification and quantification of potentially modi- 123. Xu Q, Zhu M, Yang T, Xu F, Liu Y, Chen Y. Quantitative assessment of fied peptides. J Proteome Res. 2009;8:2201–2210. human serum transferrin receptor in breast cancer patients pre- and 144. Eng JK, McCormack AL, Yates JR. An approach to correlate tandem post-chemotherapy using peptide immunoaffinity enrichment cou- mass spectral data of peptides with amino acid sequences in a pled with targeted proteomics. Clin Chim Acta. 2015;448:118–123. protein database. J Am Soc Mass Spectrom. 1994;5:976–989. 124. Pratt JM, Simpson DM, Doherty MK, Rivers J, Gaskell SJ, Beynon RJ. 145. Perkins DN, Pappin DJC, Creasy DM, Cottrell JS. Probability-based Multiplexed absolute quantification for proteomics using protein identification by searching sequence databases using mass concatenated signature peptides encoded by QconCAT genes. Nat spectrometry data. Electrophoresis. 1999;20:3551–3567. Protoc. 2006;1:1029–1043. 146. Yates JR, Ruse CI, Nakorchevsky A. Proteomics by mass spectrome- 125. Anderson L, Hunter CL. Quantitative mass spectrometric multiple try: approaches, advances, and applications. Annu Rev Biomed Eng. reaction monitoring assays for major plasma proteins. Mol Cell 2009;11:49–79. Proteomics. 2006;5:573–588. 147. Schmidt C, Lenz C, Grote M, Lührmann R, Urlaub H. Determination of 126. Scott KB, Turko IV, Phinney KW. Quantitative performance of protein stoichiometry within protein complexes using absolute internal standard platforms for absolute protein quantification using quantification and multiple reaction monitoring. Anal Chem. multiple reaction monitoring-mass spectrometry. Anal Chem. 2010;82:2784–2796. 2015;87:4429–4435. 148. Arsene CG, Ohlendorf R, Burkitt W, et al. Protein quantification by 127. Keil B. Specificity of Proteolysis. Berlin, Heidelberg: Springer Verlag. isotope dilution mass spectrometry of proteolytic fragments: 1992. cleavage rate and accuracy. Anal Chem. 2008;80:4154–4160. 128. Achour B, Al-Majdoub ZM, Al Feteisi H, Elmorsi Y, Rostami- 149. Hoofnagle AN. Quantitative clinical proteomics by liquid chroma- Hodjegan A, Barber J. Ten years of QconCATs: application of tography-tandem mass spectrometry: assessing the platform. Clin multiplexed quantification to small medically relevant proteomes. Int Chem. 2010;56:161–164. J Mass Spectrom. 2015;391:93–104. 150. Zhou H, Ning Z, Starr AE, Abu-Farha M, Figeys D. Advancements in 129. Kito K, Ota K, Fujita T, Ito T. A synthetic protein approach toward top-down proteomics. Anal Chem. 2012;84:720–734. accurate mass spectrometric quantification of component stoichi- 151. Gregorich ZR, Ge Y. Top-down proteomics in health and disease: ometry of multiprotein complexes. J Proteome Res. 2007;6:792–800. challenges and opportunities. Proteomics. 2014;14:1195–1210. 130. Desiere F, Deutsch EW, King NL, et al. The PeptideAtlas project. 152. Tran JC, Zamdborg L, Ahlf DR, et al. Mapping intact protein isoforms Nucleic Acids Res. 2006;34:D655–D658. in discovery mode using top-down proteomics. Nature. 131. Desiere F, Deutsch EW, Nesvizhskii AI, et al. Integration with the 2011;480:254–258. human genome of peptide sequences obtained by high-thoroughput 153. Wepf A, Glatter T, Schmidt A, Aebersold R, Gstaiger M. Quantitative mass spectrometry. Genome Biol. 2004;6:R9.1–R9.12. interaction proteomics using mass spectrometry. Nat Methods. 2009; 132. Picotti P, Lam H, Campbell D, et al. A database of mass spectrometric 6:203–205. assays for the yeast proteome. Nat Methods. 2008;5:913–914. 154. Silva JC, Gorenstein MV, Li G-Z, Vissers JPC, Geromanos SJ. 133. Martens L, Hermjakob H, Jones P, et al. PRIDE: The proteomics Absolute quantification of proteins by LCMSE: a virtue of parallel MS identifications database. Proteomics 2005;5:3537–3545. acquisition. Mol Cell Proteomics. 2005;5:144–156. 134. Deutsch EW, Lam H, Aebersold R. PeptideAtlas: a resource for target 155. García Alonso JI, Rodríguez-González P. Isotope Dilution Mass selection for emerging targeted proteomics workflows. EMBO Rep. Spectrometry. Cambrige, UK: The Royal Society of Chemistry; 2013. 2008;9:429–434. 156. Rodríguez-González P, Marchante-Gayón JM, García Alonso JI, 135. Eyers CE, Lawless C, Wedge DC, Lau KW, Gaskell SJ, Hubbard SJ. Sanz-Medel A. Isotope dilution analysis for elemental speciation: a CONSeQuence: prediction of reference peptides for absolute tutorial review. Spectrochim Acta Part B At Spectrosc. 2005;60: quantitative proteomics using Consensus machine learning ap- 151–207. proaches. Mol Cell Proteomics. 2011;10:M110.003384. 157. Ruiz Encinar J, Schaumlöffel D, Ogra Y, Lobinski R. Determination of 136. Kuhn E, Wu J, Karl J, Liao H, Zolg W, Guild B. Quantification of C- selenomethionine and selenocysteine in human serum using reactive protein in the serum of patients with rheumatoid arthritis speciated isotope dilution-Capillary HPLC-inductively coupled using multiple reaction monitoring mass spectrometry and 13C- plasma collision cell mass spectrometry. Anal Chem. 2004;76: labeled peptide standards. Proteomics. 2004;4:1175–1186. 6635–6642. 137. González-Antuña A, Rodríguez-González P, Ohlendorf R, Henrion A, 158. Hermann G, Møller LH, Gammelgaard B, et al. In vivo synthesized 34 Delatour V, García Alonso JI. Determination of Cystatin C in human S enriched amino acid standards for species specific isotope dilution serum by isotope dilution mass spectrometry using mass overlapping of proteins. J Anal At Spectrom. 2016;31:1830–1835. peptides. J Proteomics. 2015;112:141–155. 159. Deitrich CL, Cuello-Nuñez S, Kmiotek D, et al. Accurate quantifica- 138. Hanke S, Besir H, Oesterhelt D, Mann M. Absolute SILAC for tion of selenoprotein P (SEPP1) in plasma using isotopically enriched accurate quantitation of proteins in complex mixtures down to the seleno-peptides and species-specific isotope dilution with HPLC attomole level. J Proteome Res. 2008;7:1118–1130. coupled to ICP-MS/MS. Anal Chem. 2016;88:6357–6365. CALDERÓN-CELIS ET AL. | 737

160. Harrington CF, Vidler DS, Watts MJ, Hall JF. Potential for using characterisation of 34S-labelled yeast. J Anal At Spectrom. 2010; isotopically altered metalloproteins in species-specific isotope 25:989. dilution analysis of proteins by HPLC coupled to inductively 166. Feng L, Zhang D, Wang J, Li H. Simultaneous quantification of coupled plasma mass spectrometry. Anal Chem. 2005;77: proteins in human serum via sulfur and iron using HPLC coupled to 4034–4041. post-column isotope dilution mass spectrometry. Anal Methods. 161. Brauckmann C, Frank C, Schulze D, Kaiser P, Stosch R, Swart C. 2014;6:7655–7662. Preparation and characterisation of an 57 Fe enriched haemoglobin 167. Patel P, Jones P, Handy R, Harrington C, Marshall P, Evans EH. spike material for species-specific isotope dilution mass spectrome- Isotopic labelling of peptides and isotope ratio analysis using LC- try. J Anal At Spectrom. 2016;31:1846–1857. ICP-MS: a preliminary study. Anal Bioanal Chem. 2008;390: 162. Lásztity A, Viczián M, Wang X, Barnes RM. Sample analysis by on-line 61–65. isotope dilution inductively coupled plasma mass spectrometry. 168. Rappel C, Schaumlöffel D. Absolute peptide quantification by J Anal At Spectrom. 1989;4:761–766. lutetium labeling and NanoHPLC−ICPMS with isotope dilution – 163. Rottmann L, Heumann KG. Development of an on-line isotope analysis. Anal Chem. 2009;81:385 393. dilution technique with HPLC/ICP-MS for the accurate determi- 169. Diamandis EP. Protein quantification by mass spectrometry: is it – nation of elemental species. Fresenius J Anal Chem. 1994;350: ready for prime time? Clin Chem. 2009;55:1427 1430. 221–227. 164. Grebe M, Pröfrock D, Kakuschke A, et al. Comparison of different methods for the absolute quantification of harbour seal transferrin How to cite this article: Calderón-Celis F, Encinar JR, Sanz- glycoforms using HPLC-ICP-MS. J Anal At Spectrom. 2012;27:440. Medel A. Standardization approaches in absolute 165. Giner Martínez-Sierra J, Moreno Sanz F, Herrero Espílez P, quantitative proteomics with mass spectrometry. Mass Spec Santamaria-Fernandez R, Marchante Gayón JM, García Alonso JI. Rev. 2018;37:715–737. https://doi.org/10.1002/mas.21542 Evaluation of different analytical strategies for the quantification of sulfur-containing biomolecules by HPLC-ICP-MS: application to the