Elemental labelling combined with liquid chromatography inductively coupled plasma mass spectrometry for quantification of biomolecules: a review.

This article reviews novel quantification concepts where elemental labelling is combined with flow injection inductively coupled plasma mass spectrometry (FI-ICP-MS) or liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS), and employed for quantification of biomolecules such as proteins, peptides and related molecules in challenging sample matrices. In the first sections an overview on general aspects of biomolecule quantification, as well as of labelling will be presented emphasizing the potential, which lies in such methodological approaches. In this context, ICP-MS as detector provides high sensitivity, selectivity and robustness in biological samples and offers the capability for multiplexing and isotope dilution mass spectrometry (IDMS). Fundamental methodology of elemental labelling will be highlighted and analytical, as well as biomedical applications will be presented. A special focus will lie on established applications underlining benefits and bottlenecks of such approaches for the implementation in real life analysis. Key research made in this field will be summarized and a perspective for future developments including sophisticated and innovative applications will given.

Introduction

The development of quantitative methods was and still is undoubted the key issue in bio-analysis. Today, analytical chemistry in biology cannot be thought without the concept of “omics” technologies, which are defined by their global and comprehensive nature. Hence, the ideal “omics” methods should provide not only comprehensive qualitative data, but also comprehensive data on a quantitative basis. Evidently, there is still a gap between the idea of global analysis – providing both qualitative and quantitative data – and technological realization, because not all vital questions can be answered appropriately. This is basically due to technological limitations (e.g. availability of standards, absolute quantification in samples with heavy matrix), and exactly this is the driving force for a challenging and fast developing research.

In 2004, with reference to the field of study the most recent “omics” technique, metallomics was introduced [1]. A comprehensive review on the concept and methodology of metallomics was published by Mounicou et al. [2]. The potential of quantitative analysis – on an absolute basis – was denoted as key advantage from the beginning. One of the core technologies in metallomics is liquid chromatography inductively coupled mass spectrometry (LC–ICP-MS). In several studies [139,3] the combination of chromatography with complementary molecular and elemental MS detection showed to be advantageous circumventing restrictions of each technique applied separately. ICP-MS excels other detection methods through the possibility of species unspecific calibration, the feasibility of IDMS strategies and superior multiplexing capabilities [4-6]. As a matter of fact, the method provides only elemental and no molecular information, but combined with other techniques such as molecular MS [3], heteroelement tagging [7] or elemental labelling [8], it provides a powerful platform for biomolecule quantification. In metallomics, native chromatographic separations are often required to avoid metal loss and structural changes (e.g. influences on composition of metalloproteins or transmetallation of elemental labels). These separation methods comprise size exclusion chromatography (SEC) [9-14,126] or ion exchange chromatography (IEC) [15-18], which are highly compatible to ICP-MS detection and were implemented in one or multidimensional separation schemes [19]. Recent developments of elemental speciation analysis are summarized by Harrington et al. [20].

The first quantitative metallomic studies addressed quantification via heteroelements (e.g. S, Se, P, or even metals in metalloproteins) present in biomolecules and amenable to ICP-MS analysis [7,9,13,21-23]. A review about mass spectrometry in bio-inorganic analytical chemistry was published by Lobinski et al. [24]. However, such approaches showed limitations regarding sensitivity and selectivity. As a consequence, in analogy, to quantitative proteomics, labelling techniques found their application in metallomics. The introduction of labelling techniques using stable isotopes (e.g. isotope coded affinity tag – ICAT [25]) was a key to the success of quantitative proteomics. In metallomics the use of elemental labels (e.g. element coded affinity tag – ECAT [26]) was propagated as promising alternative offering capital sensitivity and selectivity.

Generally, labelling techniques can be performed on the level of direct labelling of target molecules [8,26] or utilizing immuno-chemistry by labelling antibodies for the selective and sensitive detection of antigens [27-29]. The latter concept has been primarily implemented for quantitative studies by laser ablation ICP-MS (LA-ICP-MS) [21,30]. Recently; the strategy was applied for imaging purposes of biomarkers on cancer tissues [31].

One of the major challenges for the implementation of the concept of labelling is the fact that the labelling procedure itself often creates by-products or fragments, e.g. intermediate products generated at the coordination step [32], diastereomere products resulting from maleimide linking [33] or antibody fragments emerging after labelling due to the applied conditions [29]. Thus, this implicates heterogeneous labelling products and characterization has to be considered, especially when further applying the labelled compounds, e.g. in an immunoassay. In this context, LC–ICP-MS analysis for speciation of elemental labelled compounds is an important tool for quality control of labelling products (e.g. antibodies) providing essential background knowledge and allowing improvement of labelling procedures.

Once the labelling procedure has been established and the labelled products have been characterized they can be employed in further work. In most cases ICP-MS is highly selective regarding the detection of the elemental label and the samples can be directly analysed for the total element concentration without further clean-up or pre-concentration [2,34]. Hence, one section of this review will cover total element analysis in the context of elemental labels highlighting the advantages of flow injection combined with ICP-MS (FI-ICP-MS).

In many cases the development and application of LC–ICP-MS based assays is beneficial in terms of selectivity, sample consumption, analyte recovery and time efforts. Hence, the recent progress regarding this field plays a central role in this review, especially in connection with immunological reactions. Important methodological aspects of biomolecule labelling will be discussed and elemental labelling procedures, as well as different applications will be covered to emphasize the potential of elemental labelling combined with ICP-MS detection in bio-analytical studies.

General aspects of biomolecule quantification

The growing demand for quantification strategies in biological and biomedical applications led to a big wave of analytical developments. Therefore a multitude of well-established bioassays based on competitive enzyme-linked immuno sorbent assay (ELISA) [35] and immunoassays using fluorescence [35] – and radiolabelling [36] are available. These bio-analytical tool sets are complemented by MS approaches employing LC or FI combined with tandem MS [37-41]. Additionally, fluorescence detection (FLD) based strategies either in combination with chromatographic separation or with gel separations [42-44] are of importance. FLD features a high degree of sensitivity and selectivity, as well as multiplexing capabilities, which are somewhat limited by the availability of fluorescent dyes showing no spectral overlapping [45].

Generally, the quantification task can be performed on a relative or absolute basis. For example, in proteomics, relative quantification was obtained by gel separation using densitometry or radiolabelling. These methods with low instrumental requirements were soon substituted by dedicated tools for relative protein quantification such as the differential gel electrophoresis (DIGE), a 2D gel electrophoresis (ELPHO) combined with fluorescent dyes [35]. The advent of mass spectrometry enabled the introduction of sophisticated labelling techniques, e.g. stable isotope labelling with amino acids in cell culture (SILAC [46]), ICAT [25] or isotope tags (iTRAQ [47,48]). For a more detailed discussion of the techniques we would like to refer the reader elsewhere [39,49,50]. In 2003, the protein absolute quantification (AQUA) technology [51] was introduced, which was the first time absolute quantification was addressed using isotope dilution mass spectrometry in proteomics. The role of isotope dilution strategies for accurate quantitative bio-analysis was comprehensively discussed by Bettmer et al. [52]. As a matter of fact, ICP-MS methods show great potential for absolute quantitative analysis with the additional possibility of implementation species unspecific calibration.

General aspects of labelling

In biochemistry the term labelling or tagging refers to a biotechnological inserted compound in order to alter the characteristics of a target molecule. In most cases those labels are employed for affinity-based enrichment of a substance of interest (e.g. FLAG or biotin tag) [8,53] or for selectivity enhancement of a detection method for an analyte (e.g. various fluorescein labels) [42,45,54]. In the case of metallomics, terminology differs between “label” and “tag” even though existing definitions are often conflicting. For example, IUPAC describes the term “label” as a marker, tag or indicator distinguishable by the observer but not by the system and used to identify a tracer [55], and therefore makes no difference between the term “label” and “tag”. In metallomics studies a label represents a compound, which is chemically attached to a biomolecule for its quantification, e.g. elemental labels such as lanthanides coordinated to macrocycles like DOTA [29,56,57], whereas a tag refers to a naturally present heteroelement amenable to ICP-MS detection, such as S, P, Se or other elements [9,22,23,58,59].

For labelling of biomolecules two basic strategies can be applied (Table 1). On the one hand pre-labelling of the analyte can be performed [46,47,57] by means of derivatization of the biomolecule prior to its insertion into a biological system and therefore altering the chemical properties of the target molecule. On the other hand post-labelling can be employed [28,29,31,60], where the biomolecule is labelled in the course of an analytical procedure (immunological reactions). Pre-labelling enables the analyst to track a molecule in a biological system, with the major drawback that the functionality of the target can be changed by the label itself. In contrast to that, post-labelling represents a two-step procedure: (1) is the labelling of an antibody against the target and (2) is the indirect detection of the biomolecule of interest via the prior labelled antibody. In this case, regardless if fluorescence or elemental labelled antibodies are applied it is a prerequisite that the labelling procedure does not affect antibody functionality [21,31,44]. In both cases the characterization of the labelled compound and knowledge about the stoichiometry of the reaction product is crucial for employment as quantification technique. The labelling itself can be obtained by exploiting the cellular metabolism, e.g. stable isotope labelling by amino acids in cell culture (SILAC) [19,46], enzymatic reactions, e.g. enzymatic digestion of a sample in H218O [61] or chemical reactions, e.g. Schiffsche base reaction [29]. However, enzymatic reactions can also be used for controlling the site specificity and stoichiometry of antibody labelling [62]. In some case labelling is obtained by DNA technology, e.g. site-specific conjugation of a cytotoxic drug to an antibody [63]. In other approaches chemical reactions are used for pre- or post labelling of biomolecules [8,47,64]. As can be readily observed in Table 1, depending on the kind of attached label different concepts for visualizing can be employed, e.g. fluorescence or chemiluminescence detection. The idea of ICP-MS as detector represents the youngest concept and gained prominence in more recent studies combined with elemental labelling strategies [29,33,65,66]. However, the accuracy of such strategies strongly depends on the stoichiometry of the reaction product. Therefore characterization of the labelled compounds by complementary ESI or MALDI-MS measurements is essential. Applications of the above mentioned labelling techniques are, e.g. medical diagnostics and pharmaceuticals, imaging or quantitative bio-analysis. The present review will underline the potential, which lies in elemental labelling approaches combined with LC based speciation and ICP-MS detection for absolute quantification of biomolecules in heavy matrix and will give an overview of pioneering applications.

Table 1

Overview of labelling approaches, detection techniques and application field.

	Pre-labelling	Post-labelling
Labelling approach	Metabolic (e.g. SILAC), enzymatic (e.g. digestion with H218O), genetic (e.g. bromdesoxyuridin labelling), chemical (e.g. ICAT, iTRAQ, ECAT)	Chemical (e.g. derivatization of antibodies with elemental labels or fluorescence tag), enzymatic (e.g. modification of antibodies via bacterial transglutaminase), genetic (e.g. site specific conjugation or fluorlabelling of an antibody)
Detection	Fluorescence, chemiluminescence, photometry, magnetic resonance imaging, mass spectrometry	Immunological reactions combined with, e.g. fluorescence, chemiluminescence, photometry, mass spectrometry (e.g. LC–ICP-MS, LA–ICP-MS)
Application	Diagnostics (e.g. MRI contrast agents) and pharmaceutics (e.g. radio pharmaceuticals), quantitative bio-analysis (e.g. heteroelement-tagging, derivatization of amino acid residues), imaging (e.g. fluorescence microscopy)	Histology (e.g. FACS), immunoassays (e.g. ELISA), imaging (immunostaining), quantitative bio-analytics (e.g. lanthanide labelled antibodies), fluorescence based assays

Elemental labelling in biomedical, therapeutic and diagnostic applications

Currently, the major application area of elemental labels is their biomedical use as diagnostic tools in MRI or as radio-pharmaceuticals [67-70]. Examples for commercially available paramagnetic agents are PARACEST or LIPOCEST [68]. Those formulations consist of a macrocycle like DOTA or NOTA combined with lanthanides like 169Tm, 157Gd or 140Ce [68,69]. Radio-pharmaceutical formulations consist of chelators like DOTA, DTPA or NOTA combined with radioactive 99Tc, 111In, 68Ga or 90Y. They are injected to the patient and serve as local radiation therapy drugs in cancer treatment [71-73].

Elemental labelling in analytical chemistry

An ideal labelling method for quantitative analysis should produce homogenous, well-characterized and completely labelled biomolecules with a well-controlled quantitative labelling reaction. The labelling procedure and the label itself should not affect biomolecule functionality (in the case of pre-labelling) or the reactivity of an antibody (in the case of indirect, immunological post-labelling). Additionally, multiplexing should be possible generating easy distinguishable stable analytes and analytical figures of merit should not be affected negatively by the label, e.g. decrease of ionization efficiency, resolution, robustness, repeatability or chromatographic efficiency. Moreover, a method for labelling should be easy to handle with a less tedious sample preparation and it should be applicable for every kind of sample [52,74]. Under realistic conditions not all of those characteristics can be fulfilled and as a matter of fact labelling provides no ultimate solution to all sensitivity and standardization problems, but it can be considered as another powerful tool in the toolbox of quantitative analysis. Internal standardization concepts are needed for accurate quantification to compensate for bias produced, e.g. by multi-step separation procedures. In this context, ICP-MS is a versatile tool for detection of elemental labels due to its sensitivity and element selectivity [34].

Protein quantification by ICP-MS based techniques has been excellently reviewed by Wang et al. [50]. A comprehensive review focussing on elemental labelling for quantitative bio-analysis was published by Bettmer and co-workers, as well as by Prange and Pröfrock [52,74]. Moreover, a special issue about labelling of biopolymers was published in 2010 edited by Bettmer and Karst [75]. Methods using elemental loaded labels for enhancement of sensitivity and robustness of biomolecules quantification are gaining prominence and a number of key papers have been published [8,18,21,33,76-81].

Inductively coupled plasma mass spectrometry (ICP-MS) for biomolecule quantification

ICP-MS as an element specific detector was introduced 1980 by Houk and is widely employed as a highly sensitive and isotope specific detector for nearly all elements present in the periodic table [82]. The analytes can be detected simultaneously via ICP-MS in a very short time offering great potential for multiplexed applications. Other advantages of this technique are a wide linear dynamic range (up to nine orders of magnitude) and the applicability of IDMS strategies allowing species specific and species unspecific calibration [5,83]. Nowadays, ICP-MS has reached the status of a mature technique and in case that ICP-MS is combined to prior separation steps like LC, compound specific quantification can be performed [5,83,84]. First attempts are dealing with protein, drug or biomarker quantification via heteroelements-tagging (mainly S, P, I or Se) [9,23,59] in biomedical sciences. These developments triggered the introduction of elemental labelling concepts employing low abundant metals with lower ionization potentials.

Soon elemental labelling of biomolecules in combination with ICP-MS has been recognized as an emerging analytical research field [49,50,52,74]. Several studies addressing the evaluation and application of elemental labelling for ICP-MS linked immunoassays [6,21,28,60,77,85,86], for imaging by laser ablation ICP-MS (22,23,31) and other quantitative bio-analytical tasks [53,65,76,78,88] have been performed.

Methodology of elemental labelling

The basic principle behind labelling is the derivatization of the compound of interest or the antibody (which is used to detect the compound of interest) in order to enhance detection sensitivity and selectivity, as well as to reduce matrix effects [49]. Ideally, labelling derivatization steps should be straightforward and simple [89]. Furthermore, labelling efficiency and characteristics of the labelling procedure and the labelled compound (e.g. concerning reaction speed and output) determine the utility of this concept for further applications in bioassays. Another bottleneck for the employment of elemental labelling is the conservation of biomolecule functionality and the stability over the whole experiment regarding sample preparation, degradation and/or generation of intermediate products, storage or chromatographic separation [32,90,91]. In the following sections the methodology of elemental labelling is described focussing on the coordination of elemental labels, linker-label chemistry and application of elemental labelling techniques.

Chemistry of elemental labelling

The elemental labelling procedure comprises two steps (Fig. 1).

Fig. 1

Principle of elemental labelling of biomolecules. The scheme does not depict the correct charges of the compounds in the biological system.

The first step represents the coordination of an element to the internal cavity of a chelating motif resulting in a highly stable metal complex [90,92,93]. Chelating moieties can be divided in linear and cyclic chelators. Table 2 gives an overview on the molecular structure of chelating moieties applied in bio-analytical and medical sciences. A more detailed review of various kinds of metal complexes and organometallic derivatizing agents was published by Bomke et al. [89].

Table 2

Molecular structure of some applied chelating moieties.

Chelator	Chemical structure	References
DTPA (diethylentriamine pentaacetic acid)		[53,81,130]
EDTA (ethylendiamine tetraacetic acid)		[15,17,81,94,130]
DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid)		[29,30,44,57,64,80,93,109,138]
NOTA (1,4,7-triazacyclododecane-1,4,7-triacetic acid)		[56,71,73,91,94,123,131]
TETA (1,4,3,11-tetraazacyclotetradecane-1,4,8,11-tetraacetic acid)		[73,92,95,107]
Polyaza crown based reagents		[98,132]
Metallocene based reagents		[79,133]
Mercury labels		[134,135]

In the case of linear and cyclic chelators metal coordination is maintained via free carboxylic groups and nitrogen atoms forming a 6-, 7-, 8- or 9-fold coordinated complex with one or more water molecules attached to the complex [94,95]. The bonds formed between the chelator and the element are combinations of coordinative and ionic bonds. DOTA complexes combined with different elements (mainly 3-fold positively charged lanthanides) result in a complex with C4 symmetry axis, whereas NOTA shows C3 symmetry axis when combined with a metal [89,94,95]. Complex stability constants were evaluated in many publications and critically reviewed in the IUPAC technical report from Anderegg et al. [96]. Other studies regarding complex stability constants of elemental labels were published in Refs. [32,71,91,97-99]. For elemental labelling the most stable complexes are selected with log K values of about 20 at neutral conditions in order to avoid metal loss or metal exchange during labelling or in the biological system [68,69]. However, the polar nature of the elemental complexes often hampers reversed phase separation [89]. Moreover, the experimental conditions for measurement of complex stability are diverging and the obtained values often differ in one or two orders of magnitude, which can be deduced in the handbook of Martell and Hancock. The stability of elemental labels is mainly influenced by parameters like pH, concentration, temperature and ionic strength, as well as by the geometry of the chelator's internal cavity and the ionic radius of the element [100].

The second step of elemental labelling comprises the attachment of the metal complex to a biomolecule via a reactive linker. Table 3 shows the most prominent reactive groups.

Table 3

Prominent reactive groups and targets for conjugation of chelating moieties.

Reactive group	Target	References
Maleimlde	Thiol groups	[6,54,76,77,122,136]
Haloacetamide (e.g. iodo- or bromoacetamide)	Thtol groups	[26,103]
Isothiocyanato	Amlne groups	[21,29–31,44,57,64,71]
Azides	Carboxyl groups	[137]

Maleimide, iodoacetamide, bromoacetamide, isothiocyanato or azide groups are commonly used, but many other groups can also be employed [101] and are commercially available [www.piercenet.com [102]]. The majority of reactive groups of elemental labels target thiol or amine groups. In Fig. 2 the reaction mechanism of (A) the attachment of a maleimide reactive group to a thiol group of a protein and of (B) the attachment of an iosthiocyanato group to an amine group of a protein is depicted.

Fig. 2

Reaction mechanism of (A) attachment of a protein SH group to a maleimide reactive group and of (B) attachment of a protein NH2 group to an isothiocyanato reactive group via Schiffsche base reaction.

For the employment of thiol group labelling of, e.g. antibodies, a reductive step is required in order to reduce disulfide bonds to thiol groups which can be labelled. This reductive step has a negative impact on recovery and it should be considered that the structure of biomolecules of interest can be altered, e.g. fragmentation of antibody Fab and Fc regions [29]. Applying maleimide chemistry for labelling implies the generation of diastereomere products caused by the chiral centre created due to the bond between maleimide and thiol [33]. Concerning haloacetamides, relatively harsh labelling conditions are necessary (75 mM borate buffer pH 8.2, 20% acetone and 6 M urea for 2 h at room temperature) for formation of a bond between the halogen and the thiol group [103], which limits the generic applicability of this concept. Another possibility is a Schiffsche base reaction, which is the case for isothiocyanato reactive groups targeting primary amine groups of peptides or antibodies [104]. For employment of this labelling procedure it should be mentioned that for the Schiffsche base reaction a basic pH is required limiting the practical applicability of this strategy.

Resulting biomolecule-label conjugates need to be investigated in terms of conservation of biomolecule functionality, stability issues and completeness of labelling. Moreover, the efficiency (i.e. yield) of the employed labelling procedure regarding repeatability and reproducibility are essential, especially when quantitative analysis is performed. The efficiency of a labelling procedure reflects the output after elemental labelling and can be evaluated by the characterization of the labelling products at every step of the labelling process (e.g. after lanthanide coordination, reduction step or linking of the elemental label to an antibody). This is of special importance for quantitative bio-analysis because labelling generates more than one product, e.g. when employing MeCAT labelling with a maleimide linker [33] or when labelling of amine groups is performed via Schiffsche base reaction of isothiocyanato linkers [29]. As a consequence not only the proof of concept of an approach is important for the implementation of a new technology for real life bio-analytic tasks, but also the characterization of the output of an elemental labelling procedure, as well as the robustness and reproducibility of such procedures are crucial. Hence, characteristics of the biomolecule-label conjugates, i.e. stoichiometry, labelling yield, intermediate products or fragmentation should be assessed to draw conclusions concerning the quality of the labelling procedure and utility. In this context LC based analysis is promising, because of the high selectivity and robustness in separation of matrix components and the feasibility of characterization of elemental labelling products. But LC speciation remains also challenging due to the equal nature of the labelled and unlabelled compounds, which cannot be separated via molecular weight only, because of the low mass difference resulting from the attachment of the free label. This concern influences the applicability of elemental labelling techniques to global and comprehensive quantitative bio-analysis, which is also affected by factors such as reaction speed of a labelling procedure, handling, required tools, automation capabilities and user skills.

Analytical applications of elemental labelling strategies

The first application of elemental labelling in ICP-MS analysis concerned the set-up of immunoassays [27]. Recent exciting developments are in the field of LA-ICP-MS either for quantification or imaging [74,127]. Key research dealing with elemental labelling combined with LA-ICP-MS detection has been published by the group around Jakubowski. In first studies they performed elemental labelling of proteins (lysozyme and BSA) with Eu, Tb and Ho-SCN-Bn-DOTA and separated them via SDS-PAGE followed by LA-ICP-MS detection [21]. To provide a better understanding of the reaction chemistry ESI-MS measurements were carried out. In the same year [30] published an article about elemental labelling of antibodies against cytochromes P450 (CYPs) quantified via SDS-PAGE and LA-ICP-MS. For labelling of the monoclonal antibodies against CYP1A1 and CYP2E1, Eu-SCN-Bn-DOTA and iodine were employed and immuno-blots were carried out. Detection was performed concomitant and showed that antibodies maintained their antigen-binding properties after labelling and Eu, as well as iodine was bound to the antibodies [30]. Recent work of this group concerned multiparametric and simultaneous detection of 5 different cytochromes P450 (CYP1A1, CYP2B1, CYP2C11, CYP2E1 and CYP3A1) in rat liver microsomes applying the same concept but with an optimized labelling protocol for antibodies [29,64]. It should be taken into account, that no information about the exact labelling degree and possible changes of the antibodies (e.g. extent of fragmentation due to the labelling procedure itself) was available. Main influence factors on the accuracy are the extent of disulfide bond reduction (Are all bonds reduced to SH groups or are some protected from reduction by their tertiary structure?). This concern and the problem of suitable standards for matrix matched calibration via LA-ICP-MS limited the application of this concept. Another focus of this group lied on iodination of proteins, antibodies and proteomes and detection with LA-ICP-MS [59,104,105]. Further research was performed regarding other elemental labelling techniques, e.g. using maleimide chemistry [106] and employment of multiplexed immuno-histochemical detection of tumour markers in breast cancer tissues [31,59]. A suggestion of this group was the employment of labelled BSA as standard for future quantitative studies of tissue slices. Moreover, Müller et al. [86] performed immunoassays combined with LA-ICP-MS detection of gold cluster labelled antibodies. The research group of Lindscheid is focussing on different DOTA macrocycles functionalized with maleimide, iodoacetic acid or iodoacetamide groups combined with lanthanides for relative and absolute quantification of protein and peptides via ICP-MS and ESI-MS. The work of this group deals with quantification of commercially available protein and peptide standards. A major task would be the implementation of this approach for unknown proteins in samples with heavy matrix and the sound characterization of the labelled compounds (e.g. investigation of unknown charged intermediate products, which cannot be performed via ESI-MS). Other forward looking attempts of elemental labelling with and without ICP-MS detection for quantitative analysis of biomolecules were published in Refs. [6,28,53,77,78,88,107-110].

Currently, pioneering and innovative trends go in the direction of the analysis of elemental labelled biomolecules in liquid phase. These approaches demand a pre-separation of the biomolecules from matrix constituents or degradation products. Most frequently LC is applied as separation technique, because of the high robustness and selectivity, in combination with ICP-MS [34]. Moreover, other benefits of elemental labelling combined with LC–ICP-MS are capability of multiplexing, possibility of absolute quantification, e.g. via species unspecific IDMS and automation. Hence, the following chapter will centre on the recent developments of LC–ICP-MS based quantification of elemental labelled biomolecules in complex sample matrices.

A major breakthrough in bio-analytical chemistry was the implementation of mass cytometry for phenotypic analysis of heterogeneous cell populations [77,111]. The analytical novelty lies in the employment of ICP-MS detection for single cell analysis, which can be used for determination of the response of cells to certain conditions by means of real time analysis of, e.g. biomarker expression of leukaemia cell lines. This strategy is challenging due to factors such as possible agglomeration of cells in the droplets or clogging problems.

Selected applications of FI- and LC–ICP-MS for quantification of elemental labelled biomolecules

In the following chapters we will demonstrate the advantages of FI and LC based quantification of elemental labelled biomolecules by discussing selected applications. We will not consider standard ICP-MS analysis or laser ablation based analysis of such samples and analytes – for these cases we want to draw the reader's attention to the magnitude of excellent reviews and textbooks available within the scientific ICP-MS community [5,20,35,87].

FI-ICP-MS based applications

The concept of flow analysis was developed independently by Ruzicka and Stewart in 1975 [112]. Flow analysis in general defines techniques for continuous flow analysis (CFA) and flow injection analysis (FIA). Basic principles underlying both strategies are the sample introduction into a carries stream, small bore conduits, sample transport to a flow-through detector and a detector response directly related to the analytes concentration [113]. FI-ICP-MS is performed using a defined and stable carrier flow in the range of low μL min−1 to 1 mL min−1 (ideally provided by a syringe pump or a pulsation free two-piston pump) in combination with an injector capable to precisely inject small sample volumes (low μL range) combined with an ICP-MS. The connection between the carrier flow and the ICP-MS nebulizer is simple and can be achieved with dedicated capillaries and connectors provided by all LC manufacturers. The approach is characterized by low sample consumption (typical injection volume is 1–25 μL), which can be beneficial when dealing with biological samples, where typically only low sample volumes are available [113,114]. The employment of FI-ICP-MS facilitates shortened cycle times because of reduced take-up times and shorter measurement duration enabling high throughput and automated analysis for fast screening of samples [114], which is essential for implementation of FI-ICP-MS in routine diagnostics in clinical applications. Additional advantages are a reduced sample preparation and the direct applicability of the instrumental set-up for speciation analysis carried out via employment of an LC column [57]. Even though pure standards can be quantified easily via FI-ICP-MS, this approach shows limitations in biological matrices. In this case sample complexity should be reduced, e.g. by matrix separation or alternative concepts before labelling can be applied. Additionally, if FI-ICP-MS is performed in combination with elemental labelling, it is essential that the labelling procedure creates a homologue product, i.e. no other component of the sample should carry the label. Otherwise off-line pre-separation or LC based speciation has to be carried out for isolation of the elemental labelled compound. Furthermore, it has to be stressed that FI-ICP-MS is limited by the fact that it can only provide information about the total element concentration in a sample, which is, as a matter of fact, identically with the information provided by conventional ICP-MS analysis utilizing standard sample introduction systems for liquid or solid samples. This implies that, for unambitious quantification of a selected analyte, the sample has to be pure in terms of the analyte, and that no other species of the target element (e.g. free metal) must be present in the sample. Consequently, simple FI based quantification affords highly selective labelling reactions and, as already mentioned fit-for-purpose pre-separation and clean-up steps.

Applications

Careri et al. [28] used 153Eu labelled antibodies for the development of an immunoassay targeting peanut allergens extracted from food samples. In this work an ELISA sandwich assay was performed where the elemental labelled antigen-antibody complex was digested with 7 M HNO3 and quantified via FI-ICP-MS. To check unspecific binding of the labelled antibodies, which is a limiting factor for the linear dynamic range of the assays at low allergen concentrations, protein-rich materials were analysed. Experimental conditions were optimized to obtain lowest unspecific antibody binding reactions and 153Eu background leading to LODs of 1.5 ng mL−1 for raw peanut protein extracts. This value is affected by the binding affinity of the antibody to the peanut allergen and by the non-specific background (e.g. produced by reagent impurities or unspecific binding). The LOD of the Eu labelled antibody itself was as low as 0.1 ng mL−1, emphasizing the potential of this approach for selective and sensitive quantitative analysis with improved linearity and reduced matrix effects. However, improvement of the labelling procedure to maintain low background levels and high antibody functionality is crucial for the sensitivity of this concept, which is an issue in ultra trace biomarker analysis.

Another FI-ICP-MS approach published by Takasaki et al. [115] employed quantum dots (QD) with a Cd, Te and Se bearing core for elemental labelling of adipose tissue-derived stem cells. Due to the small sample size (approximately 1 mg) an optimized microwave assisted acid digestion procedure was developed. For introduction of the small sample volume (20 μL) into the plasma a micro flow injection system equipped with a total consumption micro-nebulizer was constructed. The designed method was applied for the multi-element assessment (16 elements including Cd and Te) allowing quantitative analysis of the distribution of adipose tissue stem cells in mice organs [115]. In this pioneering approach, Takasaki et al. combined QD labelling (originally presented as fluorophore labelling concept) with FI-ICP-MS instead of FLD for quantification of QD derived metallic compounds (Cd, Te, Se) and other elements. The described strategy could compensate for the lack of available standards for measurement of the tissue samples via LA-ICP-MS. However, complementary use of LC–ICP-MS for speciation analysis, in combination with microwave assisted digestion and FI-ICP-MS could significantly improve the information content of results.

Kretschy et al. [57] employed an In or fluorescein labelled peptide drug in cell uptake studies. Sensitivity and precision of FI-ICP-SFMS combined with In labelling of the peptide Bβ15–42 were compared with FI-FLD measurements of the fluorescein labelled peptide Bβ15–42. The achieved absolute LOD for aqueous standards was 3 fmol for the FI-FLD method and 0.05 fmol for the FI-ICP-SFMS set-up. In this approach cellular samples were simply diluted 1:10 with 2% HNO3 prior to injection into the system. Total concentrations of the peptide Bβ15–42 could only be determined with the FI-ICP-SFMS set-up due to unspecific fluorescence emission when using FI-FLD. FI-ICP-SFMS results were comparable with those of a reference method, where samples were microwave digested and measured using IDMS and FI-ICP-SFMS. Different calibration strategies such as external calibration with and without internal standardization with Rh and species unspecific on-line IDMS with 113In were investigated for the FI-ICP-SFMS approach underpinning the potential of IDMS for enhancement of precision. Concerning short-term repeatability external calibration with internal standardization using Rh was superior with an RSD of 0.9% which demonstrates the potential of FI-ICP-SFMS compared with FI-FLD, where an RSD of 6.5% was observed. In all cases the biological variability of approximately 15% was representing a major influence factor on the total combined uncertainty, which should be considered when working with biological samples.

To sum up, the advantages of flow injection based ICP-MS lie in the fact that small sample volumes can be analyse in a very short time (>90 samples per hour) without tedious sample digestion. Additionally, FI-ICP-MS offers great multiplexing capabilities and enables different quantification strategies, which are a key issue for implementation of this concept in routine diagnostics. Bottlenecks concern sample characteristics such as purity, labelling degree and background levels of labels, as well as antibody selectivity and functionality.

Other attempts are utilizing gold nanoparticles for labelling of antibodies or proteins facilitating signal amplification and hence sensitive detection of various biomolecules in serum samples via FI-ICP-MS. Li et al. [116] used antibody affinity binding in combination with gold nanoparticle labelling and ICP-MS detection for the quantification of Escherichia coli O157:H7 pathogens enabling the detection of 500 E. coli O157:H7 cells in 1 mL of sample (500 colony forming units per mL sample). The assay showed high specificity in tests with non-pathogenic E. coli [116]. In the work of Liu and Yan [117] aptamer modified gold nano particles and antibody modified silver nano particles were employed for simultaneous determination of insulin and cytochrome c in human serum. The results indicate the potential of such concepts for massively multiplexed assays [117]. In the pioneering work of Jarujamrus et al. [118] the antigen chloramphenicol (CAP) was labelled with gold nano particles and detected via anti-CAP immobilized on solid supports. The immunoassay takes place in 96 well plates and shows good sensitivity, linearity and precision in competitive measurements of mixed samples containing CAP and CAP-BSA conjugates. This paper describes the feasibility of the applied strategy for trace analysis in biological samples [118].

LC–ICP-MS based applications

In comparison to FIA, LC separation combined with ICP-MS significantly enhances methodological selectivity and reduction of sample complexity. In the context of biomolecule separation, size exclusion chromatography, ion exchange chromatography and reversed phase chromatography are most frequently applied. All three methods have some specific disadvantages, but can also be combined in multidimensional setups for enhancement of chromatographic selectivity. SEC is often regarded as separation method running under “physiological” conditions, which is beneficial for analysis of compounds, which are unstable at basic or acidic pH or low buffer concentrations. As a drawback, SEC separations often suffer from low and non-repeatable column recoveries and low chromatographic efficiency. Ion chromatography often needs buffers with very high ionic strength (up to 1 M), which is mitigating long-term stability of the ICP-MS sensitivity. The high organic solvent content of eluents employed in normal and reversed phase chromatography is often limiting regarding stability and robustness, especially when acetonitrile is mandatory, as in the case of HILIC. In combination with elemental labelling, LC–ICP-MS based quantification shows advantages such as high sensitivity in challenging sample matrices (e.g. required for ultra trace analysis of biomarkers), high selectivity (e.g. via employment of immune-reactions) and capability of multiplexing and species unspecific IDMS. Limitations concern the characterization of the labelled compounds (e.g. stoichiometry or fragmentation) for sustainment of accuracy.

However, the state-of-the art regarding ICP-MS sample introduction (e.g. cooled spray chambers, oxygen addition) and LC miniaturization regarding both column and particle diameter allowing lower flow rates and higher separation efficiency, respectively, triggered the development of LC–ICP-MS methods, which are highly suitable for analysis of elemental labelled biomolecules.

Most of the LC–ICP-MS based studies focus on the labelling of peptides, proteins or antibodies. Researchers have early recognized that separation of the labelled analytes or conjugates from free, remaining labelling reagents and other impurities is a prerequisite for accurate quantification. Moreover, multi-compound methods, where the same functional group of different analytes is derivatized demand for chromatographic methods providing baseline separation for each analyte. In this context, Rappel and Schaumlöffel [65] utilized Lu-DTPAA (Lu-diethylenetriamine-N,N,N′,N″,N″-pentaacetic dianhydride) for derivatization and absolute quantification of labelled peptides via nano RP-ICP-MS detection combined with species unspecific IDMS. In a two-step procedure the amino groups of peptides were labelled with DTPAA followed by loading of the chelator with Lu. Determination of the peptides was carried out using a acclaim pep map 100 C18 nano-column (75 μm × 15 cm, 3 μm) operated at 300 nL min−1 combined with ICP-MS detection and post column addition of 176Lu yielding a recovery of nearly 100% and 4.9% precision. The achieved LOD was 179 fmol with an injection volume of approximately 10 nL. For implementation of this concept into routine analysis or diagnostics the applicability has to be verified in real life samples with different matrices, because this can significantly affect the recovery of the labelling procedure (e.g. unspecific interactions during labelling or separation of labelled and unlabelled compounds).

Another approach follows the basic idea of derivatization of a protein or peptide with a MeCAT reagent (lanthanides coordinated to DOTA-Bn macrocycle) via maleimide chemistry. After separation of the analytes via SDS-PAGE the protein bands were extracted and analysed utilizing FI-ICP-MS or LC–MS. The proof of concept study was published by Ahrends et al. [8] where peptides, BSA and α-lactalbumin were measured undigested with FI-ICP-MS or after proteolysis with LC–MS/MS. For both strategies a previous reduction of the sample using TCEP was required for generation of thiol groups from disulfide bonds. This reduction can affect the recovery and outcome of labelling due to fragmentation or generation of intermediate products. Hence, absolute quantification via FI-ICP-MS should be considered as challenging because total element concentrations are measured and the accuracy depends on background, clean up steps, labelling conditions and labelling degree. However, subsequent work investigated analytical robustness and suitability of this concept for biological applications [119]. Additionally, Pieper et al. [120] from the same group examined the fragmentation behaviour of MeCAT labelled BSA utilizing nano or capillary LC–MS for detection. Moreover, a iodoacetamide reactive group combined with the MeCAT lanthanide label was employed for quantification of digested BSA and β-lactalbumin via SDS-PAGE or nano-RP-ESI-MS/MS (applying a Zorbax 300 SB C18 column) revealing higher labelling efficiency compared to the maleimide functionalized MeCAT approach. Furthermore, this concept yields more homogenous products owing to the lack of diastereomere creation when employing iodoacetamide linking to thiol groups. A drawback of this method are the relatively harsh reaction conditions under which labelling is carried out limiting the applicability of this technique, e.g. for antibodies [103].

Hann et al. [13] performed immuno affinity assisted SEC-ICP-MS measurements of a In-DOTA labelled peptide Bβ15–42-antibody complex in cellular samples using the BioSuite125 UHR SEC column (4.6 mm × 300 mm, 4 μm particle size) operated at a 350 μL min−1 flow of 20 mM Tris–HCl buffer (pH 7). The developed method was further applied for determination of metallodrug-protein adducts in solution via retention time shift resulting from the formation of an antibody antigen complex (Ru labelled transferrin adducts of an anti-cancer drug bound by an anti-transferrin antibody). Improved detection limits of 1 pmol for transferrin adducts could be achieved by employment of a fully automated two-dimensional SEC-IC separation prior to ICP-MS detection. This approach underlines the potential of LC–ICP-MS for the separation of different antibody antigen complexes, which is crucial for the implementation of this concept for bioassays. Another work from this group in the year 2008 addresses the quantification of the free In DOTA labelled peptide Bβ15–42 in cellular samples via RP-ICP-MS using a Nucleosil C8 column. Accordingly, the unlabelled peptide was quantified via RP-ESI-TOF-MS employing the same instrumental conditions. Concerning the determination of the analyte in aqueous standards both methods revealed comparable sensitivity and limits of detection of about 3 fmol on column. In contrast to that, quantification of the peptide Bβ15–42 in cell lysates was only possible with the concept of elemental labelling combined with RP-ICP-MS detection due to severe matrix suppression in RP-ESI-TOF-MS [80].

In the study of Iwahata et al. [121] quantitative analysis of single fruit fly samples for determination of branched amino acids was accomplished using RP-ICP-CCMS and RP-ICP-SFMS both combined with pre-column derivatization of the amino acids with a metal tag reagent (Rub2m-O-Su). The employed set-up enabled amino acid detection by LC–ICP-MS and RP separation provided by increased hydrophobicity owing to derivatization. After extraction of the fruit fly sample and analyte derivatization, the investigated amino acids could be separated using an Inertsil ODS-3 RP column (1.0 mm × 150 mm, 3 μm) operated at 50 μL min−1 (injection volume 1 or 5 μL) in gradient mode 10 mM a sodium acetate buffer (pH 4.8)/ACN/H2O. 101Ru and 102Ru were calibrated in transient mode with a correlation coefficient of R2 = 0.999 using amino heptanoic acid as internal standard.

Esteban-Fernandez et al. [33] addressed absolute quantification of lysozyme and BSA on the peptide level using RP-ICP-MS (Polaris 3 C18-Ether, 150 mm × 1 mm, 5 μm; Varian), while complementary RP-ESI-MS of the digested and labelled proteins served as identification technique. For RP-ESI-MS 6 different synthetic standard peptides labelled with Eu-MeCAT-maleimide were employed as standards and were selected because of their different behaviours in reversed phase separations and good identification capabilities. For quantification of the Eu-MeCAT-maleimide labelled peptides resulting from BSA and lysozyme digestion via RP-ICP-MS species unspecific isotope dilution analysis (IDA) with post-column addition of a Eu2O3 spike was performed. Summarizing, this study was the first successful application of the MeCAT technique for cysteine labelling applied for absolute protein quantification and can serve as an example for the fruitful combination of organic and inorganic MS together with elemental labelling. Recently, a proof of concept study for dual labelling of thiol and amine groups of peptides via MeCAT-iodoacetamide and DOTA-NHS-ester combined with lanthanides was published by this group [66]. The method was optimized for cysteine containing standard peptides and applied for BSA and HSA, offering capabilities for tracking peptide/protein reactions or estimation of thiol and primary amine groups. In principle, the peptides were first immobilized on Sepharose 6B resins and labelled with the DOTA-NHS ester followed by coordination of lanthanides to the complex. Afterwards, the complexes were eluted and cysteine residues were labelled with a lanthanide loaded MeCAT-iodoacetamide. Detection was maintained parallel by the use of RP-ESI-MS (applying a Zorbax 300 SB C18 column) for determination of labelling completeness and yield, and RP-ICP-MS utilizing the same column for gathering of relative quantitative information (Fig. 3). A drawback of this application was the lack of accuracy of the quantitative determinations, which was suggested to be due to systematic errors.

Fig. 3

LC–ESI-FTICR-MS (A) and LC–ICP-MS (B) chromatograms of standard peptides p1 and p2 dually labelled in the amino and thiol with DOTA-NHS-Ho and MeCAT-IA-Tm, respectively. The extracted ion chromatograms (EIC) of the dual-labelled p1 and p2 are also presented in (A). The labelling positions are indicated by an asterisk (amino positions) or dot (thiol positions) in the sequence.

Zheng et al. [18] performed relative quantification of protein mixtures containing RNase, cytochrome c and lysozyme utilizing lanthanide labelling of the amine groups with Ce and Sm coordinated to DTPAA (diethylenetriamine-N,N,N′,N″,N″-pentaacetic dianhydride). For this study a cation exchange column (TSK GEL SP-5PW 7.5 mm × 75 mm, 70 μm) operated at 0.8 mL min−1 was coupled to an ICP-MS instrument (see Fig. 4). Gradient elution with 20 mM sodium phosphate buffer as eluent A (pH 6.5), and 500 mM ammonia chloride and 20 mM sodium phosphate buffer as eluent B (pH 6.5) was applied using an injection volume of 100 μL. 140Ce and 152Sm were used as analytes and signal intensities were compared. Multi element capabilities were emphasized since labelling with different lanthanides is possible providing the conditions for high throughput and top-down proteomics. Labelling was optimized with highest yields at a 10-fold molar excess of DTPAA compared to the proteins and a 2-fold molar excess of Ce and Sm compared to DTPAA-protein. In a first step the proteins were labelled with the chelator DTPAA (2 h, 37 °C) and in a second step the lanthanides were coordinated which was rapid and complete after 2 h at 37 °C. The presented approach revealed LODs between 0.2 and 7 pmol for the investigated proteins with good precision (RSD of 5%). No considerations about unspecific interactions due to the sequence of the labelling procedure were reported. In the work of Waentig et al. higher lanthanide background levels were observed when the lanthanide was attached after the chelator was linked to the biomolecule [104].

Fig. 4

HPLC/ICP-MS chromatograms of Ce labelled proteins (A) and Sm-labelled proteins (B).

Yan et al. [122] performed a proof of concept study for absolute quantification of three digested or intact Eu-MMA-DOTA (Eu-1,4,7,10-tetraazacyclododecan-1,4,7-triacetic acid-10-maleimidoethylacetamide) labelled proteins (lysozyme, insulin and ribonuclease A) utilizing RP-ICP-MS detection combined with species unspecific IDMS. The labelling procedure was optimized with lysozyme as a model substance starting with reduction of the disulfide bonds with TCEP followed by conjugation of MMA-DOTA to the SH groups of lysozyme. Optimized conditions for this step were a 10-fold molar excess of MMA-DOTA compared to SH groups at pH 6.8–7.6 for 40 min at 47 °C. Afterwards the MMA-DOTA biomolecule complex was loaded with Eu3+ by chelation using a 5-fold molar excess of Eu with respect to DOTA at pH 5.8 which should avoid possible hydrophilic precipitation of Eu (no further details concerning pH and temperature during chelation). In this work the labelling procedure started with the attachment of the macrocycle to the biomolecule followed by the coordination of Eu to the conjugate. Other studies with this sequence of labelling [21,104], reported high background levels of free lanthanides (mainly due to their adhesive nature), which was not discussed in this work. Moreover, the use EDTA for integration of free Eu could also have an impact on labelling efficiency due to possible attachment of the COOH groups of EDTA to free NH2 groups (e.g. at lysine residues) of the labelled biomolecule conjugate. However, the lysozyme-MMA-DOTA-Eu complex was confirmed with ESI-MS at mass 19,712.0, whereas no signal could be observed for the lysozyme-MMA-DOTA conjugate. Accordingly, no loss of Eu during HPLC separation with 0.05% TFA in water and ACN was observed. which was assumed regarding to the literature log K values of 23.5 indicating high stability of the Eu-DOTA complex. The labelled peptides were quantified after tryptic digestion or as intact proteins using a Zorbax 300 SB C18 column (1.0 mm × 150 mm, 3.5 μm) operated at 0.05 mL min−1 coupled to an ICP-MS detector combined with species unspecific IDMS using 153Eu as a spike. LC separation was carried out utilizing gradient elution with 0.05% TFA in water/ACN and LODs between 0.819 and 1.638 fmol were achieved for the intact proteins. Moreover, complementary ESI-MS measurements were carried out, providing another dimension of separation. The results showed that not all of the expected labelled tryptic peptides from lysozyme and ribonuclease A could be detected, suspecting an incomplete digestion of the proteins. This hypothesis was further supported by the detection of a missed cleavage peptide from ribonuclease A. This implicates that intact proteins may be more feasible targets. 151Eu/153Eu ratios were monitored and IDMS resulted at 5.8% precision at the 10 pmol level and recovery of intact protein quantification was 97.9% in six different experiments. The covalent attachment of MMA-DOTA to the SH groups of the investigated proteins is selective and efficient under mild conditions. Harshness of labelling conditions should also be considered for preservation of the biomolecules functional activity, e.g. when labelling an antibody [6].

Determination of complex stability constants under selected chromatographic conditions is crucial for bioassays employing elemental labelling. Kretschy et al. [91] assessed complex stability of different chelating moieties such as DOTA, NOTA and DTPA combined with lanthanides in competitive experiments under acidic chromatographic conditions (pH 2.5). Lanthanide labelling was performed at pH 6.5 and at 5 °C for 15 min. Incubation of the lanthanide complex directly on the autosampler was carried out at 5 °C and 37 °C simulating storage as well as physiological conditions. Kinetic experiments were performed after coordination of the lanthanide to the macrocycle, by means of measurement of the samples at different time points. The simultaneous measurement of 11 different lanthanide complexes with LC–ICP-MS utilizing a mixed mode stationary phase (Asahipak GS 220 HQ, 4.6 mm × 250 mm, 6 μm) revealed that log K values were significantly lower (with max. log K of 5) during chromatographic separation at pH 2.5, compared to other studies performed at neutral conditions with log K values of about 20 [140,90,123]. The applied pH of 2.5 for chromatography is typical for RP separations or bioassays. Hence, changes of complex stability constants should be taken into account when performing quantitative analysis with lanthanide labelled biomolecules.

Conclusion

Critical aspects of elemental labelling strategies

Labelling may affect biomolecule properties

Both concepts of labelling involve changes of the target itself (pre-labelling) or of the labelled antibody employed for detection of the target (post labelling) presumably altering:

the functionality of the biomolecule,

the structure of the labelled compound,

the extend of fragmentation,

the generation of intermediate products or other by-products.

These effects need to be critically assessed when employing a newly developed labelling assay.

Experimental procedure affects recovery

Labelling is a multi-step procedure (e.g. clean-up steps, extraction or centrifugal filtration) entailing losses on every level of the applied strategy, which impairs the recovery. Internal standardization is crucial to ensure capital recovery or losses due to the labelling procedure have to be taken into account.

Quality control and determination of the labelling yield is critical

The separation of the labelled from the unlabelled target is challenging due to very similar characteristics and possible instability in LC separation, complicating the exact determination of the stoichiometry, the attached labels per molecule and the labelling sites. Besides that, the quantification of pure labelled biomolecules, on a native basis, is complex. Additionally, trans-metallation of the element coordinated to the chelating moiety of the label complex can occur, e.g. due to the acidic pH of applied chromatography or bioassay conditions. Moreover, the linking reaction used for attachment of the label to the biomolecule of interest often lacks in selectivity and requires sound knowledge about the 3D structure of the labelled compound (e.g. accessible thiol or amine groups, protein folding). Therefore, the detailed characterization of the labelled biomolecule and possible resulting fragments (e.g. after TCEP based reduction) or intermediate products (e.g. diastereomeres when employing maleimide chemistry) is essential for implementation of the concept of elemental labelling in bio-analysis. However, in most cases structural elucidation via molecular MS or NMR is needed.

No generic labelling procedure is available

The applied labelling conditions need to be optimized for every target individually. Even in the case of relatively similar antibodies (e.g. IgG's), significant differences of the tertiary and quaternary structure influence the labelling output.

Commercial availability of labelling reagents is critical

Not all required labelling reagents are commonly available, and exist in appropriate purity (e.g. elemental labels, macrocycles). Moreover, in-depth characterization of the labelled compound, e.g. 3D structure, sequence or accurate molecular weight is often deficient.

Concluding remarks

As a matter of fact elemental labelling in combination with ICP-MS (FIA and LC) is a promising tool for quantitative bioassays. However, the tool set is still in the state of fundamental studies. Basic research on labelling degree, labelling stability and conjugate stoichiometry will allow to fully exploit the potential of this techniques in biology. For a future breakthrough of the method an increase of sensitivity by significantly increasing the metal labelling degree is demanded. Promising fundamental studies are currently performed, e.g. quantum dots [124].

In all cases ICP-MS will be the key technology in the improvement of elemental labelling and most importantly in the establishment of high through put bioassays.