A Tutorial Review of Labeling Methods in Mass Spectrometry-Based Quantitative Proteomics

Abstract

Recent advancements in mass spectrometry (MS) have revolutionized quantitative proteomics, with multiplex isotope labeling emerging as a key strategy for enhancing accuracy, precision, and throughput. This tutorial review offers a comprehensive overview of multiplex isotope labeling techniques, including precursor-based, mass defect-based, reporter ion-based, and hybrid labeling methods. It details their fundamental principles, advantages, and inherent limitations along with strategies to mitigate the limitation of ratio-distortion. This review will also cover the applications and latest progress in these labeling techniques across various domains, including cancer biomarker discovery, neuroproteomics, post-translational modification analysis, cross-linking MS, and single-cell proteomics. This Review aims to provide guidance for researchers on selecting appropriate methods for their specific goals while also highlighting the potential future directions in this rapidly evolving field.

1. Introduction

Mass spectrometry (MS) has been a cornerstone in the field of proteomics, offering unparalleled insights into complex protein information within biological systems. Over the past decades, MS-based proteomics has facilitated accurate genome-wide quantification of protein changes across multiple biological states in a single experiment.^1,2 This evolution is largely attributed to the advancements in labeling techniques, which have markedly enhanced the precision, accuracy, and throughput of quantitative proteomic studies.^3,4 Consequently, these labeling techniques have broadened the scope and capabilities of proteomic research, catalyzing significant progress in key applications such as protein structural elucidation,⁵⁻⁷ biomarker discovery,⁸⁻¹² and drug development.¹³⁻¹⁶

In the MS-based proteomics field, the predominant methodology employed is the bottom-up strategy.^2,3,17 This involves initially breaking down proteins into peptides through proteolytic digestion (Figure 1). This workflow begins with protein extraction, denaturation, reduction, and alkylation from biological samples to prepare them for enzymatic digestion. Enzymatic digestion is typically performed with trypsin, which cleaves proteins at the C-terminus of lysine and arginine residues with high specificity.¹⁸ Following digestion, peptides undergo rigorous cleanup processes, such as solid-phase extraction, to remove contaminants and, optionally, enrichment for targeted modification studies. The refined peptide mixture is separated via high-performance liquid chromatography (HPLC), primarily reversed-phase HPLC (RPLC) based on hydrophobicity, and then introduced into the mass spectrometer using an electrospray ionization (ESI) source.¹⁷ In MS, the mass-to-charge ratios and signal intensities of ionized peptides and their corresponding fragmentation products are measured. Bioinformatics tools process these fragmentation spectrum data against theoretical spectra generated from in silico digestion of protein databases, with strict statistical analysis to control the false discovery rate (FDR) and ensure result reliability.^19,20 Subsequently, the identified peptides in the sample are mapped to their source proteins, facilitating the identification of these related proteins.

Figure 1.

Conceptual workflow of bottom-up proteomics. Initially, proteins are extracted from biological samples and subsequently digested into peptides. Through chemical labeling strategies, stable isotopes are incorporated into peptides. The peptides are then analyzed using an LC-MS/MS system. The MS/MS mass spectra are analyzed via database search algorithms to identify peptides matching. Subsequently, the identified peptides in the sample are mapped to their source proteins, facilitating the accurate identification and quantification of proteins from complex mixtures. Figure created with BioRender.com.²³⁴

Mass spectrometry is not intrinsically quantitative.² The peak height or area in a mass spectrum does not accurately reflect a peptide’s abundance in the sample due to variations in peptide ionization efficiency and detectability.²¹ Obtaining quantitative information necessitates significant additional effort. Initially, label-free quantitation methods in MS-based proteomics were developed. It relied on matching peaks across different LC-MS runs and comparing the intensities of the peptide ions or spectrum counts.²²⁻²⁵ A significant challenge with this approach is run-to-run variability, which stems from differences in chromatographic separation, ionization efficiency, and instrument performance, all contributing to inaccurate quantification. Label-free methods also face the issue of missing values, with a substantial portion of peptides undetected across replicates, complicating result comparisons across experimental runs.²⁶

To address these limitations, stable isotope labeling techniques were introduced.²⁷ These techniques involve the incorporation of stable isotopes into different samples, combining them for sample preparation and analysis, then distinguishing them in the MS based on their m/z difference. Isotopic labeling allows for the direct comparison of different samples within the same MS run, thereby reducing variability and improving quantitative accuracy.^25,28

Labeling methods in quantitative proteomics can be categorized into various subcategories based on different criteria. These include quantitative ions (precursor ion-based quantification, reporter ion-based quantification, and hybrid methods), resolution requirements for distinguishing mass differences (high resolution required at precursor ion level, reporter ion level, or not required), and the methods of isotopic incorporation (chemical, enzymatic, or metabolic). This review aims to provide an overview of different labeling methods, focusing on their fundamental principles, advantages, and limitations. The review will also cover the applications and recent advancements of these labeling strategies, guiding researchers to choose suitable methods for their research goals and highlighting potential future directions in this rapidly evolving field.

2. Quantitative Labeling Methods

2.1. Precursor Ion-Based Quantification

Precursor ion-based quantitation stands as one of the earliest and most fundamental methods in proteomics.²⁹ These methods differentiate between light and heavy labeled samples by a mass difference introduced by stable isotopes. This mass difference is normally greater than 4 Da to minimize isotopic envelope interference.³⁰ Relative quantitation of these samples is achieved through comparing the extraction ion chromatograms from peptide precursor ions in MS1 spectra (Figure 2A).²⁵ Depending on the method for introducing stable isotope labels, precursor ion quantitation techniques can be categorized as metabolic, enzymatic, and chemical labeling.

Figure 2.

Graphical representation of quantitative labeling approaches in mass spectrometry. (A) Precursor ion-based quantification. Light and heavy labeled sample precursor ions are distinguished by a mass difference larger than 4 Da, Relative quantitation is achieved through the extracted ion chromatograms (XIC) for each labeled species. (B) Mass defect-based precursor ion quantification. This method exploits subtle mass difference introduced to each labeled species, reducing spectrum complexity, and enhancing multiplexing capability. (C) Reporter ion-based quantification. This method utilizes isobaric tags for labeling peptides from distinct samples then mixed them for MS/MS analysis. Labeled peptides will only generate a single peak in MS1 spectrum. Quantification is based on the intensity of their fragmentation-generated reporter ions. (D) Hybrid quantification. This method combines the principles of both light/heavy and isobaric tagging to enhance multiplexing capabilities.

Metabolic labeling involves culturing cells or feeding animals with isotopically labeled amino acids, facilitating label incorporation at the cellular or protein level. This approach reduces variations introduced in sample preparation at an early stage, thereby enhancing the quantitative accuracy. Consequently, metabolic labeling was considered the gold standard in proteomics quantification.³¹

¹⁵N labeling is one of the earliest developed metabolic labeling methods.³² Its application in quantitation is limited due to the variable number of ¹⁵N labels in peptides, leading to a complex isotopic pattern in the resulting spectra. Amino acid stable isotope labeling in cell culture (SILAC) is a more widely used method.³³ Stable isotopic labeled amino acids, typically lysine and arginine, are incorporated into cell culture media.³³⁻³⁵ After trypsin digestion, all peptides (except protein C-terminal ones) will have a predictable mass shift, simplifying data interpretation.³⁶ SILAC is extensively used in studies of cell-derived biological processes,³⁷⁻⁴¹ including single-cell analysis,⁴² and is adaptable to multicellular organisms by feeding labeled amino acids.⁴³⁻⁴⁶

Enzymatic labeling, with ¹⁸O labeling as the primary method, involves enzymatic digestion of proteins in ¹⁸O labeled water. This process will incorporate two ¹⁸O atoms into the C-terminal carboxyl groups of digested peptides.⁴⁷⁻⁴⁹ Although it has limited multiplexing capabilities of 2-plex, this method is suitable for a variety of proteomic samples and complements other labeling techniques.²⁹

Chemical labeling introduces stable isotope labels by attaching various tags to the functional groups of peptides in vitro. The chemical labeling methods can be applied across a wide variety of samples including cells, tissues, and various bodily fluids.

The Isotope-Coded Affinity Tag (ICAT) was the first commercial reagent developed for proteomic quantification.^50,51 It features an iodoacetyl group for reacting with cysteine’s thiol group. However, this strategy loses information for all peptides without cysteine. Subsequently, the dimethylation reaction of N-termini and lysine residues (α- and ε-amine groups) was introduced to quantitative proteomics.⁵² This reaction involves the formation of a Schiff base with formaldehyde and subsequent reduction to dimethylamine with cyanoborohydride. The dimethylation reaction has high efficiency in mild conditions and produces no significant byproducts.⁵³ Furthermore, dimethylation improves the ionization efficiency of peptides since the formed tertiary amine exhibits enhanced ionization capabilities in electrospray.⁵⁴ By utilizing different isotopic forms of reagents, triplex dimethylation can be achieved with 4 Da mass difference.³⁰ Applying Lys-C enzyme for digestion, which results in each peptide having two reactive amine groups, allows for an increase in throughput to 5-plex labeling.⁵⁵

There are also other amine-reactive chemical tags like mTMT,⁵⁶ mTRAQ,⁵⁷ and iDiLeu⁵⁸ that can introduce mass differences at the precursor level. These tags are often variants of their isobaric version, differing in the number of heavy isotopes. We will introduce the structures of these tags in Section 2.3.

In summary, precursor ion-based quantification methods are realized by comparing peak area in MS1 spectra. Due to the specific m/z of each precursor ion, these methods offer high quantitative accuracy.²⁹ However, due to the complexity of proteomic samples, many coeluting peptides share similar precursor ion masses in MS1 spectra, and the S/N of MS1 peak is often lower than MS2. Thus, quantitation data for low-abundance species are less reliable. Additionally, the use of isotopes introduces multiple peaks for the same peptide, thus, increasing MS1 spectral complexity. This complexity can cause multiple MS/MS triggering for the same peptide, producing redundant MS2 spectra.³⁰ Also, to minimize isotopic envelope interference, these methods require a mass difference of >4 Da, limiting the throughput.

2.2. Mass Defect-Based Precursor Ion Quantification

Mass defect-based precursor ion quantification is a novel method that addresses the limitations of traditional MS1-based quantification methods, which typically require a mass difference of 4 Da. This approach utilizes the concept of mass defect, which means the neutron-binding energies of different stable isotopes result in small but distinguishable mass differences (commonly at the mDa level). Commonly used isotopes include ¹²C/¹³C (+3.3 mDa), ¹H/²H (+6.3 mDa), ¹⁶O/¹⁸O (+4.2 mDa), and ¹⁴N/¹⁵N (−3.0 mDa).⁵⁹ These tiny mass differences are indistinguishable in low-resolution but can be resolved using high-resolution instruments like Fourier Transform Ion Cyclotron Resonance (FT-ICR) and the Orbitrap instrument, thereby enabling multiplex quantification without increasing complexity of the MS1 spectrum (Figure 2B).

This mass defect principle was first incorporated into metabolic labeling. The Neutron Encoded (NeuCode) SILAC concept pioneered by Hebert et al.⁵⁹ employs two types of isotope-encoded lysine to culture yeast cells: one with six ¹³C atoms and two ¹⁵N atoms, and another with eight deuterium atoms. This labeling introduces a 36 mDa mass difference, which is distinguishable at MS resolutions exceeding 200 K (Figure 3). Merrill et al.³¹ expanded this to a 6-plex neutron-encoded SILAC experiment using additional lysine isotopologues, each with a 6 mDa difference, requiring a 960 K resolving power for accurate quantification.

Figure 3.

Illustration of NeuCode SILAC Strategy. (a) Theoretical mass of the possible isotopologues for a + 8 Da lysine amino acid. (b) Theoretical calculation depicting the percentage of peptides that are resolved when spaced 6, 12, 18, or 36 mDa for different resolving powers 15 thousand to 1 million. (a,b) Reproduced or adapted with permission from ref (59). Copyright 2013, Springer Nature. (c) MS1 scan collected with 30,000 resolving power of yeast LysC peptides and inset of a selected precursor having m/z at 827 (black trace). The signal recorded in a subsequent high resolution MS1 scan (480,000 resolving power) is shown in red–only at this high resolution is the quantitative data revealed. Presented below the MS1 scan is an MS/MS spectrum of the neutron encoded SILAC pair. Collisionally activated dissociation (CAD), also known as collision-induced dissociation (CID), was applied to generate MS/MS spectrum. Reproduced or adapted with permission from ref (235). Copyright 2018, Springer Nature.

Chemical labeling is another method for introducing mass defects. Hebert et al.⁶⁰ first reported mass defect-based chemical labeling, employing amine-reactive NeuCode tags composed of acetylated arginine- acetylated lysine- glycine. By incorporating six heavy isotopes in various configurations, they created a set of 4-plex labels with each label displaying a 12.6 mDa difference. However, the tags potentially hinder peptide identification rates due to the generation of noninformative tag-related product ions and arginine’s suppression of peptide chain cleavage. Thus, a more compact, less intrusive, mass defect-based tag would be more beneficial for effective peptide identification.

Our group has advanced amine-reactive chemical labels with mass defects. Frost et al.⁶¹ introduced the Dimethyl Pyrimidinyl Ornithine (DiPyrO) tags for quantitative proteomics. These tags are synthesized by reacting arginine with acetylacetone, yielding dimethyl pyrimidine derivatives. This process effectively reduces the level of peptide fragmentation suppression caused by arginine. Triplex and 6-plex quantifications are achievable at resolutions of 240 and 480 K, respectively, with an 8-plex variant proposed for future higher-resolution applications. Hao et al.⁶² explored duplex mass defect-based N,N-dimethyl leucine (mdDiLeu) tags for proteomics and amine-containing metabolomics quantification. Zhong et al.⁶³ further extended this approach to a 5-plex method by combining isotopic and mass defect labeling. With the compact structure of the mdDiLeu label, each mdDiLeu tag was synthesized efficiently in a single step with readily available materials. However, the incorporation of deuterium atoms into these tags might induce retention time shifts.

Mass defect tagging also stands out for its ability to generate distinct mass signatures, facilitating the selection of diagnostic peaks for accurate peptide identification and quantification. Tagging light and heavy mass defect labels in equal ratios can create distinct peak pairs with tiny mass difference for analytes, which is recognizable by software.⁶⁴ For example, Di et al.⁶⁵ utilized sialic acid chemical labeling to introduce unique mass defect patterns for enhanced detection and quantification of sialylated glycopeptides in complex samples. Similarly, Wang et al.⁶⁶ developed the mNeuCode (methyl-neutron-coding) method for targeted analysis of protein arginine dimethylation, using metabolic labeling to produce diagnostic mass defect peak pairs.

Overall, mass-defect methodology addresses the challenge of precursor interference, outperforming traditional MS1-based quantification methods in sampling depth and multiplexing capacity. However, it requires ultrahigh MS1-level resolutions (typically >200 K at m/z 200) to differentiate subtle mass differences on peptide precursors, necessitating advanced FT-ICR or Orbitrap instruments and longer cycle times. Multiple isotope incorporation is normally required to generate MS-resolvable mass difference. Thus, the mass defect’s multiplexing potential is limited by the availability of isotopologues.⁶¹ Also, achieving higher multiplexing will inevitably involve deuterium atoms, which could cause retention time shifts.

2.3. Reporter Ion-Based Quantification

Reporter ion-based quantification methods significantly improve throughput compared to precursor ion-based methods through the utilization of isobaric tags.^2,67 Isobaric tags have identical masses but differ in their isotopic composition. They are normally composed of three parts: a reporter group to generate reporter ions under fragmentation, a mass balancer group to ensure a consistent mass across different channels, and a reactive group that attaches the tag to peptides.

Peptides from different samples will be labeled with different plexes of isobaric tags and mixed for analysis. In MS, labeled peptides will generate only a single peak in MS1 analysis due to the identical chemical composition and mass of isobaric tags. However, during MS/MS fragmentation, the distinct reporter ions are released, allowing for the measurement of peptide abundance in each sample based on the intensity of the corresponding reporter ion. Compared with precursor ion-based quantification, reporter ion-based quantification only relies on the signal intensity of the reporter ions from tandem MS, without the need to construct extracted ion chromatograms (Figure 2C).

This method simplifies the MS1 spectra by generating multiplex labeled peptides as a single composite peak, enabling high-throughput quantification without an increase in spectral complexity. Moreover, by aggregating signals for the same species from all isobaric labeling channels, the MS1 signal is boosted due to the same m/z of the differently labeled peptides. Thus, the signal of peptide fragment ions in the MS/MS spectrum can also be enhanced, facilitating improved peptide identification and overall proteome coverage.

Over the past decades, various isobaric tag sets have been developed, each with unique chemical structures, as will be detailed. The mass defect concept has also been adopted into the reporter ion group to increase the multiplex capability. As the mass of the reporter ion is much lower than the intact peptide precursors, a substantially lower resolution power (typically >50K at m/z 200) is sufficient to resolve the reporter ion with several mDa level mass defect, compared with mass defect-based precursor ion quantification methods (typically >200 K at m/z 200 is required).^68,69

2.3.1. Commercially Available Reagents: TMT, TMTpro, and iTRAQ

In 2003, Thompson et al.⁷⁰ introduced the Tandem Mass Tag (TMT) concept. The initial design of TMT was a duplex tag system composed of amino acids, integrating an isotopically labeled amino acid tag, a mass normalization amino acid, a cleavage enhancement group (proline), and an N-hydroxy succinimide (NHS) ester as an amine-reactive group (Figure 4A). This design facilitated the concurrent acquisition of both peptide backbone and reporter ions through collision-induced dissociation (CID) or higher-energy collisional dissociation (HCD), enabling simultaneously relative quantification and identification in MS/MS spectrum.

Figure 4.

Chemical structures of different isobaric tags applied in quantitative proteomics. The red part of the structure indicates the reporter group; the blue part indicates the mass balancing group; and the black part indicates the amine reactive group. For each tag, the required resolution (at m/z = 200) to achieve baseline separation of the reporter ions is listed, providing information for the selection and application of these tags based on MS instrument resolutions. The dash lines represent HCD/CID cleavage site. For proteomics-based applications, HCD fragmentation is normally applied to generate reporter ions due to the low mass cutoff (or “one-third rule”) of CID activation for ion trap.⁴

Following the first version of duplex TMT, the isobaric tag for absolute and relative quantification (iTRAQ) was introduced with a 4-plex tag design.^71,72 This system utilized a more compact N-methyl piperazine reporter ion group and a carbonyl balance group (Figure 4B). iTRAQ was later expanded to an 8-plex configuration by incorporating additional stable isotopes and enlarging the balance group.⁷³ The mass of reporter ions was carefully designed to avoid peak overlap with peptide intrinsic fragmentation ions, such as the phenylalanine immonium ion.

Subsequent advancements led to a refined 6-plex TMT,⁷⁴ featuring a dimethylpiperidine reporter ion and a more streamlined balance group, while preserving the NHS moiety as the reactive group (Figure 4C). Compared with the first-generation of TMT, the newly developed TMT tag was more compact and was optimized for reporter ion formation. Subsequently, this 6-plex TMT was expanded to 11-plex labeling without changing its chemical structure.⁶⁸ This was achieved through the tiny mass difference of 6.3 mDa between the ¹³C and ¹⁵N isotopes, which could be differentiated at a resolution of 50K or higher (at m/z 200) for the reporter ions. Due to the availability of isotopologues, this version of TMT is limited to 11-plex while with the potential of reaching 18-plex.³

To meet the demand for higher multiplexing capacities, the TMTpro series was recently introduced.⁷⁵⁻⁷⁷ This series adopts isobutylproline for reporter ion generation and integrates two β-alanine residues as balance groups (Figure 4D). This novel structure allows for the incorporation of stable isotopes with less synthesis effort and lower cost. By incorporating nine stable heavy isotopes and utilizing the mass defect between ¹³C and ¹⁵N isotopes, up to 18-plex TMTpro labeling can be achieved. Compared to TMT 11-plex reagent, TMTpro exhibits higher hydrophobicity and requires lower fragmentation energy in MS/MS. Comparative studies have shown similar performance in peptide identification numbers and quantification accuracies between TMTpro 16-plex and TMT 11-plex reagents.⁷⁷

Among the commercially available isobaric tags, the TMT/TMTpro series has gained popularity, predominantly for two reasons: the high multiplexing capacity and the integration of TMT-centric templates by Thermo Fisher Scientific in the latest Orbitrap instruments and data processing software. This integration significantly streamlines the workflow, positioning TMT reagents as an increasingly accessible choice for researchers.

2.3.2. Cost-Effective Alternatives: DiLeu and IBT

While isobaric labeling methods like TMT and iTRAQ have gained popularity in quantitative proteomics, their widespread application has been hindered by the high cost of reagents, primarily due to the intricate and multistep synthesis processes.⁶⁹ Consequently, there is a demand for cost-effective isobaric tags.

N,N-Dimethyl leucine (DiLeu), developed in our laboratory, emerges as an efficient alternative. It was inspired by the generation of intense a1 ions post peptide dimethylation with leucine at the N-termini.^54,78−80 Initially conceptualized as a 4-plex tag by Xiang et al.,⁸⁰ DiLeu boasts a compact structure, comprising a reporter group, a carbonyl balance group, and an amine-reactive triazine ester group (Figure 4E). DiLeu’s advantages include the cost-effectiveness and high synthesis yield for various channels while maintaining high labeling efficiency and accurate protein quantitation. In comparison to TMT-labeled peptides, DiLeu tags also demonstrate improved peptide backbone cleavage and more intense reporter ion signals, facilitating the detection and quantitation of low-abundance peptide species.

Later, the multiplexing capacity of DiLeu was expanded from a 4-plex to a 12-plex system, utilizing mass defects from ¹⁵N, ¹³C, and D isotopes.⁶⁹ This expansion requires a resolution above 45K at m/z = 200 to achieve baseline separation of 12 reporter ions with a minimum spacing of 5.8mDa. Notably, seven of these channels necessitate only a single-step synthesis, while the remaining five channels require one additional ¹⁸O exchange step, enhancing the multiplex capability while preserving the synthesis-effectiveness. More recently, DiLeu’s multiplexing capability was escalated to a 21-plex through stepwise monomethylation, with a minimal spacing of 3 mDa between neighboring channels, demanding higher resolution (90K at m/z 200) for baseline resolve on FT-MS instruments.⁸¹

In its chemical composition, DiLeu employs 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium (DMTMM) triazine ester as the amine reactive group, necessitating an activation process prior to labeling. Stored in its carboxylic acid form, DiLeu has a longer shelf life and is free of hydrolysis degradation problem compared with NHS esters applied in TMT and iTRAQ.⁸² Both triazine ester and NHS ester can achieve complete labeling of peptide amine group within the labeling process. However, the DiLeu labeling process mandates additional purification step (like strong cation exchange) to remove excess activation reagent.⁸³

The compact structure of DiLeu requires the incorporation of deuterium atoms to extend the multiplexing capability. By placing deuterium atoms in the hydrophilic dimethylamine group, the deuterium shift effect is minimized.^30,52,69,81 It is noteworthy that 12-plex and 21-plex DiLeu do not have completely identical precursor masses but exhibit slight differences at the mDa level, as also observed with 18-plex TMTpro. This results in uncertain mass increments in MS1 for labeled peptides (normally around the average mass of all channels). Despite these variations, these labeling methods are still classified under isobaric labeling, as the subtle differences are not resolved in MS1 under low resolution to differentiate labeling channels. Also, these variations will not adversely affect data acquisition and protein identification workflows.^69,76,81

An alternative 8-plex version of DiLeu-based tags, not dependent on mass defect, increases multiplexing capacity by integrating larger balancer groups and more stable isotopes (Figure 4F).⁸⁴ The reporter ions of these isobaric tags are set 1 Da apart, permitting application on low-resolution instruments. However, a more significant retention time shift happens between channels due to deuterium placement in the balance group.

Apart from DiLeu, the Isobaric Tag (IBT) serves as another cost-effective alternative utilizing amino acids as building blocks. IBT accomplishes 10-plex labeling exclusively using ¹³C and ¹⁵N isotopes, thereby minimizing deuterium-induced retention time shifts.⁸⁵ The recently unveiled IBT-16plex introduces isopropylated–isobutylated glycine as the reporter group, complemented by glycine and β-alanine in the balance group (Figure 4G).⁸⁶ This configuration intentionally reduces reporter ion intensities to amplify b/y-ion production, thus, enhancing peptide identification. Comparative studies have revealed that IBT-16 plex substantially outperforms TMTpro-16plex in peptide and protein identification numbers in HeLa cell samples.⁸⁶

2.3.3. Summary

In summary, reporter ion-based quantification methods enable increased multiplexing capability while avoiding the problems of spectrum complexity in precursor-based quantification techniques. Although these methods demonstrate excellent throughput, they increasingly suffer from ratio distortion caused by the coisolation of precursor ions and subsequent cofragmentation of peptides. This issue adversely affects the accuracy and precision of these methods.^{2,3,25,29,83,87} We will discuss this challenge in section 3 and detail various methods developed to address this challenge.

2.4. Hybrid Quantification Methods

Recently, the pursuit of improved multiplexing capabilities and quantitative accuracy has led to the development of hybrid (or hyperplex) labeling methods. The fundamental concept is utilizing mass differences at the precursor level to enable the simultaneous analysis of multiple sets of isobaric labeled samples. Hybrid quantification approaches are typically achieved through two kinds of methods:

2.4.1. Combining Isotopic Labeling and Isobaric Labeling

For example, the combined precursor isotopic labeling and isobaric tagging (cPILOT) method utilizes pK_a differences between free amines on peptides.⁸⁸⁻⁹⁰ It selectively dimethylates the N-terminus amine group to introduce a mass difference, followed by lysine residue tagging with DiLeu or TMT isobaric reagents (Figure 5A). cPILOT methods can expand the multiplexing capability of original isobaric tags by two or three times. Another work achieved tandem labeling in one pot reaction combining mTRAQ and TMTpro,⁹¹ which minimizes the sample loss and reduces sample preparation times. However, these methods limited the protease to LysC to create double amine groups on peptide termini.⁸⁸ TAG-TMTpro method was reported as a more universal method.^92,93 It introduces mass differences by labeling tert-butyloxycarbonyl (Boc) protected alanine or glycine amino acid residues into peptide amine groups, followed by the deprotection and isobaric labeling on newly introduced amine groups in these amino acid residues. This method allowed for up to 54-plex quantification when combined with 18-plex TMTpro and up to 102-plex when combined with 18-plex TMTpro and 16-plex IBT isobaric tagging.^92,93

Figure 5.

(A) DiLeu cPILOT experimental workflow. Digested samples undergo stable isotope dimethyl labeling of peptide N-termini with either light or heavy dimethyl groups at low pH followed by labeling of lysine residues with 12-plex DiLeu isobaric tags at high pH. Pooled samples are analyzed by LC-MS/MS using CID MS2 acquisition of peak pairs for peptide sequence identification and HCD SPS-MS3 acquisition for accurate 24-plex quantification via DiLeu reporter ions. Reproduced or adapted with permission from ref (88). Copyright 2018, American Chemical Society. (B) Hyper plex single cell proteomics workflow combining SILAC and TMTpro labeling. First, the cells were cultured separately in regular media (light) or media containing ¹³C₆¹⁵N₂l-lysine and ¹³C₆¹⁵N₄l-arginine (heavy). After cell lysis, reduction, alkylation and digestion, each single cell sample was labeled with a different TMT reagent. The carrier and reference channels were added after quenching, heavy and light precursors were identified in MS1, selected for fragmentation, and analyzed by tandem MS. Reporter ions are produced for relative quantification. Reproduced or adapted with permission from ref (42). Copyright 2023, American Chemical Society.

Other strategies include combining SILAC with isobaric labeling. Recently, 2-plex SILAC combined with TMTpro has been applied to enhance the throughput of single-cell proteomics, enabling the analysis of up to 28 single cells with carrier and control channels in a single LC-MS run (Figure 5B).⁴²

2.4.2. Labeling Samples with Multiple Sets of Isobaric Tags

For instance, by mixing 11-plex TMT, 18-plex TMTpro, and 16-plex IBT labeled samples, up to 45-plex quantification can be achieved.⁹⁴ Other methods have also been reported recently with different tag combinations.^95,96 For further insights into these methods, we recommend reading the review article by Bowser et al.⁹⁷

One significant advantage of hybrid labeling methods is the substantial increase in throughput. This approach can save considerable instrument time and reduce sample preparation variability.⁹⁷ However, these strategies inherit issues from both precursor and reporter ion-based quantification methods, including increased spectral complexity at the precursor level, redundant MS2 spectra, and reporter ion ratio distortion problem.³ Also, when combining multiple sets of isobaric tagging, missing values of peptide identification may happen between different isobaric groups due to the ionization efficiency variation of isobaric tags, which will influence the reproducibility of proteins identified between groups.⁹³

3. Challenges in Reporter Ion-Based Quantitative Proteomics and Solutions

While isobaric labeling methods have become one of the most employed approaches for multiplex quantitative proteomics, they present notable challenges that may impact experimental outcomes. A significant problem is ratio distortion (or compression) due to coisolation and cofragmentation, which can severely affect the accuracy and precision of quantitative results.^2−4,83,87

In data-dependent acquisition (DDA)-based methods, a predefined isolation window selects ions for MS/MS fragmentation. Ideally, this isolation window isolates a single peptide precursor. However, due to the technological limitations of ion filters, the smallest achievable isolation window with ion filters, like quadrupole, is approximately 0.4 Th.⁹⁸ Any coeluting peptide ions within the mass isolation window will be coisolated and fragmented. The reporter ion ratios obtained in the MS/MS spectra will be distorted, as the reporter ions from contaminating precursors are indistinguishable from the target analytes (Figure 6A).

Figure 6.

(A) Illustration of ratio distortion problem in reporter-ion based quantification. PEP represents peptides from samples and INT represents coeluted interference peptides. Peptides sharing similar m/z values and retention times are coisolated and cofragmented in MS/MS, leading to quantification inaccuracies due to interference. This is represented by the merging of reporter ions in the MS2 spectrum, where the measured ratio is compressed compared with the theoretical ratio. (B) Illustration of SPS-MS3 method for reducing ratio distortion. The highest abundance peaks of ion trap MS/MS, which typically are b and y ions from the peptide of interest, are simultaneously isolated and further fragmented for an MS3 spectrum. Reporter ions in the MS3 spectrum are used for quantification. (C) and (D) Illustration of ion-mobility based strategies (TWIMS/TIMS and FAIMS, respectively) that provide extra gas-phase separation to separate coeluting peptide ions, utilizing additional separation dimensions such as drift tube (in TWIMS/TIMS) or compensation voltage (CV) changes (in FAIMS) to mitigate ratio distortion effects. (E) Schematic overview of the isobaric peptide termini labeling (IPTL) approach, where the red dot symbol above the heavy label symbol represents incorporated heavy isotope. PE1K represents peptides from sample 1, and IN1K represents interference peptides from the same sample. Similarly, PE2K and IN2K correspond to sample 2. The peptide-specific backbone ion pairs are employed for quantification. (F) Schematic overview of complementary ion-based methods, which leverages the peptide-specific complementary ion for quantification. These complementary ions provide accurate quantitative ratios, in contrast to reporter ions, which can be distorted due to cofragmentation interference.

Typically, the issue of ratio distortion can be evaluated using a dual proteomics model,^99,100 which involves labeling samples from different origins in various ratios (e.g., 5:2:1 for human-derived samples as study channel and 1:1:1 for yeast as interference) then mixing them for MS analysis. This approach allows for the assessment of the extent of ratio compression between different analytical methods by comparing the experimental ratios and theoretical ratios. More recently, an isobaric tag-based Triple Knockout (TKO) quality control platform has been developed and commercialized by Gygi and co-workers,^101,102 providing a new standard for optimizing MS data collection parameters to reduce ratio distortion.

Various strategies have been developed to mitigate ratio distortion in isobaric-labeling-based proteomics, which can be classified into three categories: (1) Modification of instrument acquisition settings; (2) application of an extra level of separation (Pre LC-fractionation; MS3; and Ion mobility); and (3) use of peptide-specific ions for quantification (Peptide fragment ion; Complementary ion). A comparative analysis of these methods is summarized in Table 1. This section aims to guide readers through the principles of each approach, assisting in the selection of the most suitable method for specific research needs.

Table 1. Comparative Analysis of Methods for Mitigating Ratio-Compression Problem.

methods	effect on mitigating ratio distortiona	special instrument required	influence on quantified peptides ID number	influence on analysis time and throughput	influence on data processing	other comments
modification of acquisition settings	+	no	N/A	N/A	N/A	easy to apply
sample fractionation	++	HPLC with fraction collector	increase	increase total analysis time	increase database search time	can be combined with other methods
MS3-based methods	+++	Orbitrap Tribrid MS	decrease due to extra duty cycle	N/A	N/A	gold standards for complex samples
ion-mobility based methods	+++	MS coupled with ion mobility spectrometry or FAIMS interface	increase	TIMS/TWIMS: ToF analyzers limit reporter ion to unit mass difference	need software compatible with IMS data interpretation	In-FAIMS fragmentation can cause false identification for glycopeptides samples
peptide fragment ion-based methods	++++	no	decrease	limited plex compared to normal isobaric labeling	need special software for b/y ion cluster identification/extraction	b/y ion clusters make MS/MS spectrum complex
complementary ion-based methods	++++	no	decrease (depending on the efficiency of complementary ion generation)	limited plex compared to normal isobaric labeling; reporter ion is limited to unit mass difference	need special software for complementary ion extraction and deconvolution	normally combined with narrow isolation windows to isolate monoisotopic peak

^aSymbols: +, partially effective; ++, good; +++, very good; ++++, excellent.

3.1. Modification of Instrument Acquisition Settings

Narrowing the precursor isolation window by reducing the MS/MS isolation window size can effectively lower the number of interfering ions within the window, thereby minimizing coisolation.¹⁰³ Thermo Fisher Scientific recommends setting a 0.7 Th narrow isolation window for analyzing TMT labeled samples, as recommended in their TMT application notes.^104,105

Additionally, triggering MS/MS at the LC peak apex, known as delayed fragmentation, can significantly decrease cofragmentation by a factor of 2.¹⁰³ These strategies are advantageous, because they generally do not necessitate extra sample processing steps and are straightforward to apply. However, it is important to note that these methods may only partially mitigate ratio distortion.¹⁰⁰

3.2. Pre-LC-MS Sample Fractionation

Prefraction methods such as high-pH reversed-phase LC,^106−108 strong cation exchange (SCX),¹⁰⁹ and Hydrophilic Interaction Liquid Chromatography (HILIC)^110,111 fractionation can effectively decrease sample complexity, thus reducing ratio suppression. However, this approach increases the total analysis time due to the additional fractionation of samples.

3.3. MS3-Based Methods

Ting et al.¹¹² pioneered the MS3 approach to tackle the challenge of ratio distortion utilizing Orbitrap tribrid mass spectrometers. This technique involved an additional fragmentation step: the most intense peptide fragment ion was further fragmented in MS3 to produce reporter ions. This extra level of separation significantly reduced interference ions, as coisolated peptides are less likely to generate intense peptide fragments ions. Building on this, the MultiNotch MS3 (or Synchronous Precursor Selection-MS3, SPS-MS3) method was introduced to address the sensitivity issues of the original MS3 technique.⁹⁹ It allowed for the systematic coisolation of multiple peptide fragment ions in the ion trap, significantly improving the method’s sensitivity (Figure 6B). Now SPS-MS3 has become the gold standard for analyzing isobaric labeled complex samples.¹¹³

The latest advancement in this method is the integration of real-time database searching (RTS) with the newest generation of tribrid Orbitrap instruments.^114,115 RTS adaptively triggers MS3 scans only when reliable peptide identifications occur in the ion trap, significantly increasing the scan efficiency and enhancing the depth of proteome analysis.

3.4. Ion-Mobility Based Methods

Ion mobility spectrometry (IMS) offers a novel approach to reduce coisolation. With its orthogonality to LC separation, IMS can provide complementary gas-phase separation to separate coeluting peptide ions and background ions. The bonus of IMS separation is it can enhance the depth of proteomic analysis without increasing total analysis time.¹¹⁶

Different IMS platforms have been applied in quantitative proteomics analysis (Figure 6C). Shliaha et al. combined traveling wave ion mobility separation (TWIMS) with narrowed quadrupole isolation to reduce precursor coisolation.¹¹⁶ Also, Ogata et al. used trapped ion mobility spectrometry (TIMS) to reduce ratio compression, achieving quantification results comparable with the SPS-MS3 method without compromising instrument sensitivity or speed.¹¹⁷

High-field asymmetric waveform ion mobility spectrometry (FAIMS) is the only commercial IMS compatible with Orbitrap mass analyzers and existing SPS-MS3 methods (Figure 6D).¹¹⁸ FAIMS has been shown to significantly enhance the dynamic range and accuracy of quantitative measurements without sacrificing protein identifications.^119−121 FAIMS-SPS-MS3 method was reported to have the highest quantitative accuracy followed by SPS-MS3, FAIMS-HRMS2, and HRMS2.^119,122 Recently, Fang et al. highlighted the utility of FAIMS in improving the accuracy of N-glycopeptide quantification, showcasing its versatility in post-translational modifications (PTM) analysis.¹²² However, Rangel-Angarita et al. also reported that in-FAIMS fragmentation can lead to false positive detection of glycopeptides, especially when high compensation voltage (CV) values was applied.¹²³

3.5. Peptide Fragment Ion-Based Methods

Ratio distortion in reporter ion-based methods stems from nondistinguishable reporter ions, leading to inaccuracies when peptide ions are cofragmented. Peptide backbone fragment ion-based methods, like Isobaric Peptide Termini Labeling (IPTL) introduced by Koehler et al.,^124−126 address this problem by applying peptide-specific fragment ions, ensuring accurate quantification even for cofragmented peptides. IPTL sequentially labels the Lys and N-terminal amine groups with stable isotopes, producing the same peptide mass in MS1 but distinct peptide backbone ions upon fragmentation (Figure 6E). Then precise quantification can be achieved by comparing the intensity of peptide backbone ions. IPTL can increase the confidence of quantification as each peptide fragment ion will carry quantification information.

Despite its advantages, IPTL has limitations of restricted multiplexing capacity compared with common reporter-ion based methods. Recent advancements, like multiplex pseudoisobaric dimethyl labeling (m-pIDL),¹²⁷ have expanded multiplexing capabilities by using large isolation window to isolate pseudoisobaric precursors. However, complex isotopic envelopes created in MS/MS still require advanced data processing. Software tools like IsobariQ¹²⁸ and ITMSQ¹²⁹ were also developed to enhance the identification rates of b/y ion clusters and assist the deconvolutions of quantification results.

Further innovations were developed to reduce the spectral complexity to increase the peptide identification rate. Zhou et al.¹³⁰ used the concept of mass defect to decrease the complexity with requiring high-resolution instruments. Tian et al.¹³¹ recently developed isobaric acetyl-isoleucine-proline-glycine (Ac-IPG) tags to reduce spectrum complexity by allowing b-ions to remain identical while y-ions isotopically labeled.

Other methods include specifically applying peptide a1 ions from the neutral loss of b1 ions for quantification. Zhang et al.¹³² recently established the 8-plex a1 ion-based proteome quantitation (APQ) method, utilizing isotopic labeled a1 ions for quantification. They further developed the deep-APQ method by sequentially acquiring high and low mass ranges for identification and quantification, offering deeper peptide coverage and higher quantitative accuracy.¹³³

3.6. Complementary Ion-Based Methods

Another strategy is based on complement reporter ion (or peptide-coupled reporter ion) clusters. This method leverages the unique fragments produced due to the loss of isobaric tag reporter ion and CO neutral loss (i.e., the intact peptide ions with the balancer group attached) (Figures 6F and 7B,C).^134−136 As those ions are specific to each precursor peptides and distinguishable in MS/MS, this strategy ensures accurate quantification even when peptides coisolate.³ A significant advantage of this approach over IPTL methods is that it generates the same b/y ions across different labels, thus not complicating database searches. Also, this approach does not necessitate higher-order MS scans, making it more accessible for laboratories without the latest MS equipment.¹³⁷

Figure 7.

Complementary ion-based quantification methods. (A) TMT- and TMTpro-tags comprise a reporter region (red), a balancer region (blue), and an amine-reactive NHS-ester moiety (green rectangle). The carboxyl group loss as CO during fragmentation is part of the balancer and highlighted in an ellipse (gray-blue). (B) When analyzing complex samples via shotgun proteomics, if MS2 reporter ions are used for quantification, the interfering peptides lead to a distortion of the measured ratios, as the source of reporter ions cannot be distinguished. However, because the masses of complementary ions are peptide-dependent and include the heavy isotope labels of the balancer region, they can be used for interference-free MS2 quantification. (C) During fragmentation of a TMTpro-modified peptide, the positively charged reporter ion is separated from the ion and a neutral CO molecule is lost. This leads to an ion where the balancer part is still attached to the peptide. Because the balancer region encodes the complementary heavy isotope labels of the reporter ion, the balancer-peptide conjugate is called the complementary ion. The charge state of the precursor ion is reduced by one. (A–C) Reproduced or adapted with permission from ref (137). Copyright 2021, American Chemical Society. (D) Left, coisolation of the natural isotope cluster in a standard isolation window centered on the precursor ion convolutes the relative abundance of peptide-coupled reporter ions. Right, an asymmetric narrow isolation window that reduce the signal from adjacent isotope peaks and enables direct quantification of complementary ions. Reproduced or adapted with permission from ref (138). Copyright 2018, Springer Nature.

However, the method also has its challenges. First, the complexity of the isotopic envelope of complementary ions needs additional data processing and deconvolution.¹³⁵ For peptide ions, when a standard isolation window is used, both monoisotopic peak and ¹³C isotopic peak will be coisolated and fragmented. This coisolation leads to 1 Da mass offset, resulting in peak overlap. This overlap complicates complement ion clusters, because the mass difference between each plex is also 1 Da (Figure 7D). Consequently, a deconvolution process is required to reveal the actual ratio. Recent works applied narrow (<0.5 Da) and/or asymmetric isolation window to specifically isolate the monoisotopic peak, thus reducing the need for deconvolution process and increasing the quantification accuracy (Figure 7D).^{113,136−138}

Second, the throughput of this method remains limited. Due to the present limitations in the instrument resolving power for large peptide-coupled ions, complementary ions require a 1 Da mass difference. Till now, the highest plex of complementary ion quantification achieved is 8-plex TMTproC reported by Wühr and co-workers (Figure 7).¹³⁷ They also proposed a transient super-resolution technique on the Orbitrap platform as a possible future solution to this problem.¹³⁹

Third, commercial TMT tags suffer from low formation efficiency of complementary ions.¹⁴⁰ Recent developments, such as the SulfOxide (SO)-tag¹⁴¹ and EASI tag,¹³⁸ utilize a sulfoxide-based tag for increased cleavage efficiency in MS/MS, thus enhancing the formation efficiency and yield of peptide-coupled reporter ions. Our group has recently introduced the 7-plex dimethylated leucine complementary ion (DiLeuC) tags.¹¹³ Compared with commercial TMT or TMTpro tags, DiLeuC tags demonstrate better complementary ion generation efficiency and are more cost-effective. Also, accurate proteome quantification at the single-cell level was achieved, indicating a promising future for this strategy.

4. Applications of Labeling Strategies in Quantitative Proteomics

4.1. Cancer Proteomics

Cancer cells exhibit abnormal protein expression patterns compared to normal cells.¹⁴² Proteomics techniques, such as MS and two-dimensional gel electrophoresis (2-DE),^143,144 are used to analyze these protein expressions to identify cancer-specific proteins (biomarkers) and explore the molecular mechanisms of tumor formation and development.^145,146 Post-translational modifications of proteins, such as phosphorylation, glycosylation, and ubiquitination, play a significant role in cancer progression.¹⁴⁷ Proteomics research focuses on these modifications to understand how they alter protein functions and signaling pathways in cancer cells, leading to uncontrolled growth and metastasis of cancer cells.¹⁴⁷

An important aspect of cancer proteomics is the identification of biomarkers. These biomarkers are proteins whose expression levels significantly change under carcinogenic conditions and can be used for early detection, prognosis, predicting treatment response, and monitoring disease progression.^145,146Table 2 presents representative cancer biomarkers identified using label-based strategies for protein quantitation.^148−158 For example, Gao et al. employed TMT technology to analyze hepatocellular carcinoma (HCC) associated with hepatitis B virus.¹⁴⁸ According to their phosphoproteomics methods, the phosphorylation of ALDOA was found to enhance glycolysis and proliferation in HCC cells with the CTNNB1 mutation, while PYCR2 and ADH1A were related to the metabolic reprogramming in HCC. These findings help researchers better understand and design effective treatment strategies for HCC.

Table 2. Representative Cancer Biomarkers Identified Using Labeling Proteomics Approaches.

cancer type	sample type	labeling strategy	potential biomarkers	feature of biomarker	author and reference
liver	clinical tissue	TMT	pyrroline-5-carboxylate reductase 2 (PYCR2), alcohol dehydrogenase 1A (ADH1A), phospho-aldolase A (ALDOA)	involved in HCC metabolic reprogramming and increased glycolysis and cell proliferation	Gao et al.148
prostate	LAPC4	SILAC	α(1,6)-fucosyltransferase (FUT8)	increased oncogenic activity and metastasis	Clark et al.149
	clinical serum	iTRAQ	CD59, haptoglobin, tetranectin	correlated with bone metastasis	Yan et al.150
pancreas	clinical serum	iTRAQ	apolipoprotein A-1 (APOA1), transferrin (TF)	correlated with the degree of histological differentiation	Lin et al.151
	clinical tissue	TMT	melanocyte inducing transcription factor (MITF), transcription factor binding to IGHM enhancer 3 (TFE3), transcription factor EB (TFEB)	increased activation of anabolic pathways, autophagy, and lysosomal catabolism	Perera et al.152
breast	clinical tissue	TMT	fatty acid-binding protein-7 (FABP7)	increased progression and metastasis	Asleh et al.153
	MCF7, MDA-MB-231	DiLeu	methylated pyruvate kinase M2 (PKM2)	increased cell proliferation, migration, and metastasis	Liu et al.154
ovarian	OV-90	SILAC	calcium-activated chloride channel 1 (CLCA1)	increased cell aggregation	Musrap et al.155
	clinical plasma	TMT	fibrinogen alpha chain (FGA), gelsolin (GSN)	correlated with tumorigenesis and metastasis	Zhang et al.156
lung	A549	TMT	threonine tyrosine kinase (TTK)	increased tumorigenesis	Chen et al.157
	CL1–5	SILAC	karyopherin alpha 2 (KPNA2)	increased cell migration	Wang et al.158

4.2. Neuroproteomics

Neuroproteomics is a specialized field within proteomics, focusing on the study of the proteome in the nervous system. This field involves comprehensive analysis of proteins in brain tissue, cerebrospinal fluid, and other neural components.¹⁵⁹ By comparing the proteomic characteristics of healthy and diseased neural tissues, researchers can identify abnormal protein expressions and modifications, thereby revealing the molecular mechanisms of neurological disorders such as Alzheimer’s disease,^160,161 Parkinson’s disease,¹⁶² schizophrenia,^163,164 and multiple sclerosis.^165,166

Isotope labeling techniques are widely used in neuroproteomics to address various biological questions. These techniques are primarily used to characterize the proteomes of neurological disorders, drug responses, or different regions of the brain.^167,168 For instance, isobaric labels such as DiLeu, iTRAQ, and TMT have been utilized to study dynamic protein changes in children with B-cell acute lymphoblastic leukemia during chemotherapy, investigate the proteomics of serum in Parkinson’s patients, and analyze circuit-specific proteomes and phosphoproteomes in the corticostriatal system of the mouse brain during development.^169−171 Additionally, numerous reports have focused on the studies of protein expression in the thalamus and cerebrospinal fluid of schizophrenia patients,¹⁶⁴ investigation of variations in protein expression levels in human brains after severe traumatic brain injury,¹⁷² and even exploring the mechanisms of memory formation in the hippocampus.^164,172,173 Overall, neuroproteomics provides deep insights into the complexity of the nervous system and its related diseases by revealing potential molecular linkages through the comparative quantification of proteome alterations.

4.3. Post-translational Modifications (PTMs)

Post-translational modifications refer to a series of chemical and biological changes after proteins are translated, including phosphorylation, glycosylation, acetylation, citrullination, and more.¹⁷⁴ These modifications can significantly alter the physical and chemical properties of proteins, thereby affecting their function and role within cells and biological systems.¹⁷⁵ PTMs play a critical role in a variety of cellular processes, such as signal transduction, DNA repair,¹⁷⁶ and cell cycle control.¹⁷⁷ Understanding PTMs is crucial for comprehensive proteomic analysis, as aberrant PTMs are often linked to pathogenesis, particularly in studying disease mechanisms, biomarker identification, and drug target discovery.^178,179

4.3.1. Phosphorylation

Phosphorylation is the process of adding a phosphate group to serine, threonine, or tyrosine residues in proteins. This modification significantly alters the charge and conformation of the protein and is one of the most common PTMs.¹⁸⁰ Phosphorylation plays a crucial role in regulating protein activity, signaling pathways, and various cellular processes, such as cell division and metabolism.¹⁸¹ In proteomics research, MS is the primary method for identifying and quantifying phosphorylation, and tandem mass spectrometry can precisely determine the specific sites of phosphorylation on peptides.¹⁸²

The stable isotope labeling of amino acids in cell culture (SILAC) technique, using [γ-¹⁸O₄] ATP labeling, is particularly used for specifically labeling phosphorylation sites. An additional advantage of this technique is the ability to identify and quantify multiple phosphorylation sites simultaneously through mass spectrometry.¹⁸³ Molden et al. demonstrated that SILAC has been used to label and reliably identify over 1,000 nuclear phosphorylation sites in a single experiment, measuring phosphorylation levels of nuclear proteins in different cell cycle states (asynchronous, G1/S, and M phase synchronized), and further identifying the most active phosphorylation sites under these conditions.¹⁸⁴

Isobaric labeling techniques, such as iTRAQ, TMT, and DiLeu, are also widely used in proteomics to study phosphorylation. These techniques can accurately quantify phosphorylation changes in multiple samples in a single experiment, providing important insights into cellular mechanisms and disease pathology.¹⁸⁵ Given the low abundance of phosphorylated peptides, workflows in phosphorylation studies often include enrichment techniques such as IMAC or TiO₂ to simplify samples and improve the detection rate of phosphorylated species.¹⁸⁶ Jiang et al. successfully identified 12,465 phosphorylated peptides and quantified 10,436 peptides in 10 samples treated with insulin or insulin-like growth factor 1 (IGF-1) in the lung cancer A549 cell line.¹⁸⁷ This comparative analysis provided detailed information, helping to outline different signaling pathways including mTOR (mechanistic target of Rapamycin), epidermal growth factor receptor (EGFR), and insulin signaling pathways.

4.3.2. Glycosylation

Glycosylation, a key process in protein PTM, involves the addition of carbohydrate structures to proteins. This modification significantly affects protein structure and function, playing an important role in numerous biological processes and diseases.^188,189 Glycosylation is primarily divided into two types: N-linked glycosylation, where sugar chains are attached to the nitrogen atom of asparagine residues, and O-linked glycosylation, where sugar molecules are attached to the oxygen atom of serine or threonine residues.^190,191

Analyzing glycosylation in proteomics is a challenging task, primarily due to the diversity and complexity of glycan structures. Each glycosylation site may exhibit different glycan structures, thereby imparting a high degree of variability to the protein function. MS has become a key technology in glycoproteomics, enabling the identification and characterization of glycosylation sites and their attached glycan structures.¹⁹² However, the analysis becomes more complex due to the diversity of glycan types and the instability of the glycans. To overcome these challenges, researchers often employ enrichment strategies such as lectin affinity chromatography and HILIC to isolate glycopeptides from complex samples prior to MS analysis.^193,194 In addition to using HILIC to enrich low-abundance glycopeptides, Wang et al. developed the boost-DiLeu strategy, which enhances the quantification of low-abundance proteins and peptides by adding an additional boosting channel, enabling more comprehensive quantitative analysis of glycopeptides (Figure 8).¹⁹⁵ This method successfully quantified 1172 intact glycopeptides, including 164 glycoproteins and 18 glycopeptides, showing significant correlation with Alzheimer’s disease.

Figure 8.

Boost-DiLeu: enhanced isobaric N,N-dimethyl leucine (DiLeu) tagging strategy. (A) Illustration of a one-tube sample preparation strategy to amplify the signal with isobaric tags for quantitative glycoproteomic analysis. The boosting channel significantly increases the low-abundance glycopeptides. (B) The boost-DiLeu strategy workflow involves extracting proteins from biological samples, followed by enzymatic digestion and labeling using a one-tube sample preparation process. DiLeu 118d served as the boosting channel. After labeling, samples were combined for HILIC enrichment and high-pH (HpH) fractionation, followed by LC-MS/MS analysis. Reproduced or adapted with permission from ref (195). Copyright 2022, American Chemical Society.

4.3.3. Citrullination

Citullination, also known as deimination, is a PTM of proteins in which arginine residues in proteins are converted into citrulline. This transformation neutralizes the amino acid and increases its mass by 0.98 Da. Citrullination is catalyzed by a group of enzymes known as peptidylarginine deiminases (PADs) and has significant implications for normal physiological functions and various pathological conditions.¹⁹⁶ Notably, abnormal citrullination is associated with several diseases; for instance, in rheumatoid arthritis, antibodies against citrullinated proteins are one of the disease markers and contribute to its pathogenesis.^197,198 Citrullination is also related to other conditions such as multiple sclerosis,¹⁹⁹ Alzheimer’s disease,²⁰⁰ and cancers,²⁰¹ demonstrating its widespread impact on human health.

Identification strategies for protein citrullination include antibody- and MS-based techniques. Antibody methods identify citrulline residues and detect them through Western blotting or immunostaining procedures, but this approach struggles with low-abundance citrullinated proteins and cannot provide the specific location of citrullination sites.^202,203 MS-based detection relies on the mass difference between the citrulline and arginine residues. However, other PTMs, such as the deamidation of asparagine or glutamine, can also lead to similar mass changes, posing difficulties in data interpretation.

To overcome these challenges, Holm et al. developed a 2,3-butanedione tagging strategy, which specifically reacts with citrulline residues to produce a 50 Da increase in monoisotopic mass.²⁰⁴ Furthermore, Li et al. combined 2,3-butanedione tagging, biotin-thiol labeling, and isobaric labeling techniques to achieve enrichment, large-scale analysis, and multiplexed quantitative analysis of protein citrullination from MCF7 cell lines in response to various types of DNA damage responses (DDR).²⁰⁵ This research highlighted significant biological processes influenced by protein citrullination and identified enhanced citrullination in RNA-binding proteins and DNA repair proteins.

4.4. Cross-Linking MS

Cross-linking mass spectrometry (XL-MS) is a widely used technique in the fields of proteomics and structural biology, primarily for mapping protein–protein interactions and elucidating their spatial organization. A key advantage of XL-MS is its ability to analyze large protein complexes and capture dynamic interactions of proteins in their native state, which is often challenging for other structural biology techniques like X-ray crystallography, NMR spectroscopy, and cryo-electron microscopy.²⁰⁶

XL-MS relies on specially designed cross-linking reagents that create covalent bonds between neighboring amino acid residues within a protein or between different proteins. These reagents contain two reactive groups connected by a spacer arm. When proteins contact these reagents, they form cross-links between adjacent amino acids, thereby capturing and stabilizing the spatial interaction. In XL-MS, cross-linkers are designed to react with specific side chains of amino acids. The choice of cross-linker is crucial as it determines the types of interactions and distances that can be analyzed, thereby influencing the accuracy and specificity of the protein–protein interaction map obtained.^207−209

Quantitative cross-linking mass spectrometry (qXL-MS), combined with different labeling methods, has witnessed significant progress in studying protein–protein interactions and the dynamic changes in protein structure.²¹⁰ This approach achieves label-free quantitation by comparing the MS signal intensities of cross-linked peptides in different samples. Specifically, it involves extracting MS1 chromatographic peak areas and employing parallel reaction monitoring (PRM) for MS2-based quantitation.^211,212

In the SILAC quantification method, quantitation is achieved by comparing the relative abundance of light and heavy labeled cross-linked peptide pairs at the MS1 level.²¹³ Chemical labeling methods, such as TMT, are used for multiplexed isotopic labeling of cross-linked peptides in multiple samples simultaneously. To overcome measurement distortion issues due to coisolation of precursor ions, accurate quantitation of cross-links is performed at the MS3 level.²¹⁴

Recently, Chavez et al. developed the isotopic quantitative protein interaction reporter (iqPIR) technology, utilizing biotin affinity tags to enrich cross-links and incorporating stable isotopes selectively into the cross-linker for isotopic pairs of cross-linked peptides in multiple samples.²¹⁵ This method produces unique isotopic signatures in the MS2 spectra. Currently, Wippel et al. applied the 6-plex iqPIR strategy to MCF-7 cells treated with five different Hsp90 inhibitors, 1756 cross-linking sites were successfully identified, and 1650 of these sites were quantified, revealing the interaction network dynamics of specific drug categories (Figure 9).²¹⁶

Figure 9.

6-plex isobaric quantitative protein interaction reporter (iqPIR) experimental workflow. (A) Breast cancer cells (MCF-7) were treated with five different heat shock protein 90 (Hsp90) inhibitors, with experiments conducted in three separate replicates. The treated cells were incubated in vivo and subsequently cross-linked with 6-plex iqPIR. After labeling, the cells were lysed, and proteins were extracted. The labeled samples were then pooled together, followed by proteolytic digestion. (B) The combined sample was initially separated by size exclusion chromatography (SEC), followed by further fractionation using strong cation exchange (SCX). Enrichment of cross-linked peptides was then achieved through the binding of the iqPIR biotin to avidin beads. The sample underwent analysis by LC-MS. The search for and quantitation of cross-linked peptides were performed as described by Chavez et al. (C) MS1/MS2 spectra of cross-linked peptide. The MS1 spectrum shows the isotope envelope of the precursor ion, which includes all six isobaric labeled cross-linked peptides. The MS2 spectrum shows reporter ions from iqPIR and fragment ions from peptides A (b/y ions in green) and B (b/y ions in purple). Reproduced or adapted with permission from ref (216). Copyright 2022, American Chemical Society.

4.5. Single-Cell Proteomics

Single-cell mass spectrometry (scMS) has emerged as an increasingly important tool in biological research, marking a significant shift from traditional bulk cell analysis to in-depth study at the individual cell level.²¹⁷ Traditional proteomic methods usually rely on large numbers of cell samples, resulting in average protein expression profiles that can mask subtle differences between individual cells. These differences are crucial in areas such as cancer research,²¹⁸ immunology,²¹⁹ and developmental biology.²²⁰ Single-cell proteomics aims to reveal the complex protein composition of individual cells, providing a more refined view of cellular processes and cell-to-cell heterogeneity within cell populations.

The main challenge in single-cell proteomics is the detection and quantification of small amounts of proteins present in individual cells.^221,222 Label-free MS is a common approach for single-cell analysis. Combined with data-independent acquisition (DIA) technology, label-free approaches can quantify up to 2000 proteins across different stages of the cell cycle.²²³ However, its major limitation is the low throughput in sample processing. Due to the time-consuming analysis process, only a limited number of cells can be processed each day, which is impractical for studies that require analysis of a large number of different cells.^224,225 Another critical challenge in the label-free technique of single-cell proteomics is the issue of missing values. This issue primarily arises from suppression effects caused by high-abundance proteins, which complicates the accurate profiling of the entire proteome across single-cell samples. Consequently, this leads to variability in protein quantification across different samples, potentially resulting in incomplete proteome information that hampers the detection of changes in low-abundance proteins.²²¹

To overcome the limitations of label-free MS, an isotopically labeled multiplexed scMS approach has been developed. The introduction of isobaric tags for iTRAQ or TMT and labeled carrier proteomes significantly enhances the MS signal, thus improving protein group coverage and ion count per peptide.²²⁶ By labeling proteins from carrier samples, reference samples, and individual cells with TMT, the cell ratios of these labeled samples can be precisely controlled.²²⁷ Schoof et al. applied a 16-plex TMTpro quantitative method, enabling the quantification of approximately 1000 proteins from multiple single-cell samples in a single LC/MS-MS analysis.²²⁵ This approach provides a more efficient and in-depth analytical means for single-cell level proteomic research.

4.6. PlexDIA

Traditional mass spectrometry often faces challenges in simultaneously achieving high throughput,²²⁸ in-depth proteome analysis,²²⁹ and minimal missing value. Employing multiplex chemical labeling methods can improve the sample throughput, reduce the MS analysis time for each sample, and reduce the number of missing values. However, common multiplex isobaric tagging methods are not compatible with DIA acquisition methods.³

Derks et al. demonstrated that plexDIA, using 3-plex nonisobaric tags, enables the quantitative analysis of approximately 8000 proteins per labeled sample in 1-h gradients, significantly reducing data loss by more than 2-fold among different samples (Figure 10).²³⁰ This method effectively reduces the variability in protein composition between samples and across runs. In single-cell analysis, plexDIA can quantify about 1000 proteins per cell, achieving 98% data completeness in 5 min LC gradient. Therefore, plexDIA is particularly attractive for nanogram-level sample analysis, as it allows for precise and in-depth proteomic quantification analysis without the need for offline peptide separation.

Figure 10.

Experimental design for plexDIA analysis. (A) High confidence identified precursors from one label can be accurately transferred to other isotopologous precursors with FDR control, featuring identical retention times and known mass differences between different tags. The ratios of fragment ions were measured by calculating the most reliable isotopologous precursor against other similar precursors at their peak signal intensity, where interference is minimal. (B) Proteomes from various species and cell types were mixed in specific ratios, creating a benchmark with known protein ratios across a broad dynamic range. LF-DIA analysis utilized 500 ng from three distinct samples (A, B, and C), analyzed separately. In contrast, plexDIA combined these samples, each labeled with unique nonisobaric mass tags (mTRAQ), aiming to significantly increase quantitative data points without sacrificing accuracy. Reproduced or adapted with permission from ref (230). Copyright 2022, Springer Nature.

4.7. Real-Time Analytics

Real-time analytics in proteomics is a cutting-edge approach that is significantly enhancing the capabilities of mass spectrometry-based studies.²³¹ Traditional data-dependent acquisition (DDA) methods in LC-MS/MS often prioritize analyzing the most abundant or easily ionizable peptides, neglecting the less abundant ones. Additionally, DDA suffers from limitations in its dynamic range, resulting in a bias toward peptides that ionize more efficiently.

In contrast, a real-time search (RTS) in proteomics brings significant advantages by addressing these limitations. RTS rapidly matches the mass spectra against a database, identifying peptides and proteins.²³² The real-time approach also means that peptides are not selected for MS/MS based on abundance alone. Instead, RTS can be tailored to target specific peptides of interest, including those of lower abundance, providing a more comprehensive and representative view of the proteome. This also helps to reduce sample bias and ensures a broader dynamic range of detection.

Schweppe et al. developed an advanced platform named Orbiter for real-time selection (RTS) in multiplexed synchronous precursor selection (SPS)-MS3 quantitative proteomics.²³¹ This platform effectively eliminates inefficient and time-consuming MS3 scans caused by mismatches between MS2 and peptide spectra in traditional SPS-MS3 methods. Orbiter integrates RTS with error rate analysis in its online proteomics analysis pipeline, making it a rapid and efficient tool for identifying peptide spectral matches and quantifying proteins of interest. Orbiter’s RTS technology significantly enhances the speed and accuracy of proteomic analysis, capable of quantifying over 8000 proteins in half the time required by standard SPS-MS3 analysis, covering ten different proteomes. Recently, Yu et al. developed a multiplexing-based targeted pathway strategy with real-time analytics, termed GoDig.²³³ It leverages real-time analytics to pinpoint the position of target analytes presented in each sample. Without internal standard peptides, this method relies on proteomics data from past proteome-wide experiments into elution and spectral libraries. Real-time elution calibration enables the multiplexing of hundreds of target analytes into single assays. The necessary information for targeting lists derived from approximately 10,000 human proteins and 7000 mouse proteins that have been compiled as a resource. This resource is available through the GoDig assay builder Web site, facilitating the future development of pathway-specific measurement analysis (Figure 11). Overall, real-time analytics represents a significant advancement in proteomic analysis technology, improving speed and efficiency, which is particularly valuable in high-throughput analysis and large number sample processing.

Figure 11.

GoDig enables real-time profiling for pathway-specific measurement analysis. (A) GoDig offers multiple features that enable the multiplexed targeted proteomics, including (1) real-time elution calibration utilizing abundant MS1 peaks; (2) detection via coelution-triggered parallel reaction monitoring (PRM) scans for tracking elution; (3) identification through matching with a high-resolution spectral library; and (4) quantification using synchronous precursor selection (SPS). (B) Elution position prediction involves periodically capturing a few MS2 spectra to accurately ascertain the elution position. (C) After calibrating the elution window, GoDig initiates rapid PRM scans targeting specific peptides within the window to track their elution without MS1 detection. Upon detecting a target, GoDig captures a high-resolution MS2 scan for matching against the spectrum library and for selecting SPS ions, followed by quantification through MS3 analysis. Reprinted or adapted with permission under a Creative Commons CC-BY from ref (233). Copyright 2023, Springer Nature.

5. Conclusion

In recent years, advancements in labeling-based proteomics have been pivotal in transforming our understanding of complex biological systems. This review thoroughly discusses the fundamental principles, applications, advantages, and limitations of various labeling strategies. This review also highlights the evolution of quantitative labeling methods, from metabolic labeling techniques such as SILAC, which integrates stable isotopes into cellular proteins, to chemical labeling approaches such as TMT and DiLeu that allow for increased throughput. Each method possesses unique advantages and inherent limitations. It is important to select the appropriate labeling strategy according to different experimental factors, including the experiment goals (discovery-based or targeted analysis), the number of samples (requirement for throughput), sample types (enabling in vivo/in vitro labeling), sample complexity (to decide if methods are needed to mitigate coisolation) and the availability of LC-MS instrument types.

Labeling methods in MS-based quantitative proteomics have become indispensable tools in bioscience applications, facilitating research in fields including cancer biomarker discovery, neuroproteomics, post-translational modifications, and the study of protein–protein interactions and single-cell proteomics. As these techniques continue to evolve, they will undoubtedly reveal new insights into the molecular mechanisms underlying health and disease, facilitating the discovery of novel biomarkers and therapeutic targets. The continued advancement and application of labeling methods in proteomics research hold great promise for enhancing our understanding of biological complexity and driving forward the frontiers of biomedical science.

Author Contributions

^¶ These authors contributed equally: Z.W. and P.-K.L. CRediT: Zicong Wang investigation, methodology, visualization, writing-original draft; Peng-Kai Liu investigation, methodology, visualization, writing-original draft; Lingjun Li conceptualization, funding acquisition, investigation, project administration, resources, supervision, writing-review & editing.

The authors declare no competing financial interest.