Novel strategies for enhancing shotgun lipidomics for comprehensive analysis of cellular lipidomes

Abstract

Shotgun lipidomics is one of the most powerful tools in analysis of cellular lipidomes in lipidomics, which directly analyzes lipids from lipid extracts of diverse biological samples with high accuracy/precision. However, despite its great advances in high throughput analysis of cellular lipidomes, low coverage of poorly ionized lipids, especially those species in very low abundance, and some types of isomers within complex lipid extracts by shotgun lipidomics remains a huge challenge. In the past few years, many strategies have been developed to enhance shotgun lipidomics for comprehensive analysis of lipid species. Chemical derivatization represents one of the most attractive and effective strategies, already receiving considerable attention. This review focuses on novel advanced derivatization strategies for enhancing shotgun lipidomics. It is anticipated that with the development of enhanced strategies, shotgun lipidomics can make greater contributions to biological and biomedical research.

1. Introduction

1.1. Lipidomics and approaches for lipidomics

Lipids play many essential roles in organism, such as constituting cellular membranes, serving as an optimal matrix for transmembrane proteins, providing fuel storage and/or supplement, and regulating cell growth and survival as signaling [1]. Moreover, with the development of genomics, proteomics, and molecular biology, more and more vital functions of lipids are being discovered, including anti-diabetes and anti-inflammatory effects, regulation of mitochondrial ATP release, etc. [2, 3]. Accumulated evidence demonstrates that aberrant metabolism of lipids is involved in the pathogenesis of many major human diseases at the era of post-industrialization (e.g., obesity and diabetes, cardiovascular disease, cancer, neurodegenerative disorders, and autoimmune diseases) [4–8]. Therefore, determination of lipid homeostasis under different physiological and/or pathological conditions could assist us in understanding the underlying mechanism(s) of diseases, discovering potential biomarkers for early diagnosis, and searching novel drug targets.

However, cellular lipids are extremely complex and diverse. It has been predicted that tens to hundreds of thousands of possible lipids are present in individual cellular lipidomes at the levels of amol/mg to nmol/mg of protein [9, 10]. Moreover, the family of lipids is continually expanding with the discovery of new lipid species [2, 11, 12]. Furthermore, the content and composition of cellular lipids are highly dynamic. In addition to the distribution and diversities of lipid molecular species from species, cell types, cellular organelles, subcellular membrane, leaflets of membrane bilayers, and membrane microdomains [13], are also dynamically changing with environmental conditions and pathological perturbation [14, 15]. Therefore, identification and quantification of all lipids are tremendously challenging. This challenge promoted the emerging of lipidomics in early 2000 as an independent disciplinary field to investigate all lipids in a large scale and at the levels of intact molecular species [16, 17].

Although numerous other modern analytical technologies (e.g., nuclear magnetic resonance, fluorescence spectroscopy, high performance liquid chromatography (HPLC), microfluidic technology) have been developed and used in lipidomics research, particularly in the early stage of lipidomics development [18], mass spectrometry (MS)based approaches, which possess very high sensitivity and specificity and have greatly accelerated the progress of lipidomics, are the most powerful and remarkable ones [1, 19]. Generally, depending on whether the lipid solution delivered to the ion source chamber is under a constant concentration condition during the period of lipid analysis, these MS-based lipidomics approaches can be classified into two major categories: 1) LC-MS based lipidomics, in which the concentration of lipid solution delivered to the ion source is constantly changing after HPLC separation, and 2) direct infusion-based lipidomics, where the lipid solution is directly delivered into the ion source continuously and its concentration is kept constant. This review focuses on the latter one, which is also termed shotgun lipidomics [19, 20].

The principles of shotgun lipidomics are the maximal utilization of unique chemical and physical properties of lipid classes, subclasses, and individual molecular species to identify and quantify lipid species directly from organic extracts of biological samples at a constant concentration of lipid solution [20]. In contrast to the LC-MS based approaches, shotgun lipidomics technology can minimize/eliminate a few factors, which influence identification and quantification of individual lipid species, such as alterations in concentration, chromatographic anomalies, and ion-pairing alterations [21]. Additionally, constant infusion concentration allows one to have virtually unlimited time to improve mass spectral signal/noise ratios, to perform detailed tandem MS analyses in different modes (e.g., precursor-ion scanning (PIS), neutral-loss scanning (NLS), and product ion analysis), to conduct multi-stage analyses, and to ramp different instrumental variables (collision energy, collision gas pressure, ion mobility parameters, etc.) [21]. Therefore, these unique features of shotgun lipidomics enable it to become one of the widely used and the most powerful approaches in lipidomics, particularly for high-throughput analysis of lipids [21, 22].

On the basis of the features and the mass spectrometers employed, at least three different platforms of shotgun lipidomics are developed and well documented in the literature, including tandem MS-based, high mass accuracy-based, and multi-dimensional MS-based shotgun lipidomics (MSMD-SL) [1, 21].

1.2. Limitations of conventional shotgun lipidomics

Individual lipids species of a polar lipid class possess a common head group. One or more characteristic fragment ion(s) can be yielded from the head group after collision-induced dissociation (CID). A tandem MS spectrum in the NLS or PIS mode can be performed to specifically detect all individual species of this class based on a characteristic fragment ion. In 1997, Brugger and co-workers developed a prototype of shotgun lipidomics to “isolate” the individual species of a class of interest through specific NLS or PIS with a triple quadrupole mass spectrometer [23]. After the double filtering process of MS/MS, the signal-to-noise ratios of spectra can be greatly enhanced. Notably, all individual species in a particular class can be detected in one MS/MS acquisition directly from lipid extracts of biological samples. Therefore, the tandem MSbased shotgun lipidomics approach, also termed conventional shotgun lipidomics, provides determination of the species of any targeted lipid class at the level of instrumentation sensitivity in a high-throughput fashion.

Because of the great advantages of conventional shotgun lipidomics, including simplicity, efficiency, ease of management, and less expensive instrumental requirement, this platform has been widely used for lipidomics profiling of biological samples, such as plant samples [24, 25]. However, a few concerns with this platform were recognized. Firstly, the detection with the so-called specific MS/MS scanning might not be entirely specific to the lipid class or the category of lipid classes of interest, whereas this nonspecificity might lead to introduction of some artifacts. For example, both the protonated monoisotopic precursor ions of choline glycerophospholipid (PC) and sphingomyelin (SM) species yield a “characteristic” phosphocholine product ion at m/z 184 in positive ion MS/MS. Secondly, the presence of isobaric (i.e., same nominal mass) lipid species of a class of interest within lipid extracts limits unambiguous lipid identification, such as PC(30:1) and aPC(31:1) (“a” indicates an alkyl ether linkage at the sn-1 position). Thirdly, the fatty acyl (FA) substituents of lipid species are not identified since the platform only targets to the class-specific head group fragments. For instance, it is difficult to distinguish 18:2–20:4 PC from 16:0–22:6 PC by this platform. Fourthly, some altered ionization conditions cannot be easily recognized during and after the experiments. Fifthly accurate quantification of the detected lipid species might not be as simple as expected because of the differential fragmentation mechanisms manifest in individual lipid species within each lipid class. Finally, if no or very low abundance characteristic fragment ion(s) are yielded from a head group after CID, this platform does not work for, as the key point to be successful with this platform is the specificity of the characteristic fragment.

2. Enhancement of shotgun lipidomics

2.1. High mass resolution MS-based shotgun lipidomics

Due to the rapid development of MS instrumentation, commercially available hybrid type mass spectrometers (e.g., quadrupole time-of-flight or quadrupole-Orbitrap) offer not only improved duty cycle that increases the detection sensitivity, but also very high mass resolution and mass accuracy (i.e., >10,000 mass resolution and <2–5 ppm mass accuracy) [26, 27]. Researchers can acquire a full-scan mass spectrum of a lipid sample of interest in the survey scan mode with the high mass resolution/accuracy and quickly conduct product ion analyses step by step within an entire mass region of interest using these instruments [28–31]. In the latter steps, the high mass resolution/accuracy inherent in these instruments also yields accurate measurement of the masses of fragment ions (0.1 amu or higher), which can greatly eliminate many possibilities of false-positive identification.

There are two approaches of shotgun lipidomics based on high mass accuracy/resolution mass spectrometers: one is multi-PIS (NLS) high mass accuracy shotgun lipidomics [29–31]. In this strategy, the fragments corresponding to either FA constituents or head group moieties are recorded after CID, so any interesting PIS and/or NLS can be extracted from the data array of the product-ion mass spectra to identify a specific class of lipids using special software programs, such as LipidProfiler and LipidInspector [30, 31]. The other is data-dependent acquisition shotgun lipidomics. Identification and quantification of lipid species are achieved directly from product-ion mass spectra and full-scan mass spectra [32–34]. Software packages (i.e., LipidXplorer [35] and ALEX [36]) have also been developed to process full scan MS and product ion scan MS/MS data, respectively. Other variations of this shotgun lipidomics have also been developed, such as “top-down lipidomics”, “bottom-up shotgun lipidomics”, and “MS(All)” [32, 33, 37, 38]. The analysis of all high mass resolution MS-based shotgun lipidomics can be performed in the positive- and negative-ion modes in the presence of ammonium acetate in the infused solution [32].

By means of high mass accuracy spectrometers, this platform provides efficient and sensitive analysis of lipid species in a high-throughput fashion. Moreover, it can be performed in an untargeted manner to analyze any lipid species present in lipid extracts of biological samples if the dynamic range of the instrument is permitted and the software package covers all those species. Therefore, compared with conventional shotgun lipidomics, this platform has many advantages, including easy identification of isobaric lipid species of a class of interest and the FA substituents of individual lipid species. This platform has already been applied to many biological studies [27, 39–41].

Some concerns with this approach have also been recognized. It is better to add multiple (at least two) representative internal standards (ISs) for each lipid class, since quantification in the multi-PIS (NLS) strategy is based on tandem MS techniques [42]. Corrections for differential ionization responses of different species among a nonpolar lipid class should be considered when performing quantification studies of these species. Linear dynamic range of quantification largely depends on the used instrument under different experimental conditions. The mass overlap problem of some types of isomers, both of which have a similar/identical fragmentation pattern, cannot be resolved by this approach, thus needing other solutions, including chromatographic fractionation, differential or multistep lipid extraction, intrasource separation and selective ionization during MS analysis, or derivatization [38]. Finally, analysis of poorly ionized lipids, especially those in very low abundance, remains a huge challenge through this platform.

2.2. Multi-dimensional MS-based shotgun lipidomics

The third well-recognized shotgun lipidomics platform is MDMS-SL, which maximally exploits the unique chemical and physical properties inherent in discrete lipid classes or subclasses for analysis of lipids, even if the lipids are in very low abundance [19, 43, 44]. In this platform, differential hydrophobicity, stability, and reactivity of different lipid classes and subclasses are exploited during sample preparation (i.e., a multiplexed extraction approach) [45].

“Intrasource separation” is one of the major features of MDMS-SL. Briefly, the differential charge properties of different lipid classes, which are mainly determined by the head groups of polar lipid classes, are utilized to selectively ionize a certain category of lipid classes under multiplexed experimental conditions to separate many lipid classes in the ion source (i.e., intrasource separation) [46]. This separation method is similar to the electrophoretic separation of different compounds using their different isoelectric point (pI) values [20]. A workflow of the intrasource separation strategy used for global analysis of different lipid classes in MDMS-SL is illustrated in Fig. 1.

Fig. 1.

Schematic illustration of the workflow of MDMS-SL for analysis of cellular lipidomes directly from lipid extracts of biological samples [20, 22].

The concept of building blocks in lipid structures is fully exploited for identification of individual lipid species in MDMS-SL. It is well known that the majority of cellular lipid species are the combination of a few building blocks such as polar head groups, FA constituents, and backbones [20]. After CID, the fragmentation pattern of a lipid class largely relies on its intrinsic chemical structure and charge properties. Thus, the majority of lipid classes have a unique fragmentation pattern and the building blocks (corresponding to the fragment ions or the neutrally lost fragments) can be determined with two powerful tandem MS techniques (i.e., NLS and PIS) in a mass ramp format [43]. Individual lipid species, including isobaric and isomeric species, as well as regioisomers, can be effectively and thoroughly identified using these patterns of different lipid classes and subclasses [43, 47]. Accordingly, the informative fragmention(s) from either the head group or resulted from the neutral loss of the head group are used to identify the lipid class of interest, and PIS and/or NLS of FA chains can be used to definitely identify the individual molecular species present within the class. Mapping of these building blocks of a lipid class or a category of lipids constitute a two-dimensional mass spectrum [43]. Examples of MDMS-SL analysis of cellular lipidomes can be found elsewhere [21, 48].

3. Recent advanced strategies for enhancing shotgun lipidomics

In the past few years, many chemical derivatization strategies have been developed with efforts of researchers (Table 1). In the section, we summarized and discussed these strategies.

Table 1.

Summary of recent advanced derivatization strategies for enhancing shotgun lipidomics

Targeted lipid class	Methods	Reference
Lipids containing a hydroxyl group (such as DAG, MAG, NAE, oxysterol, oxidized FAs, and so on)	Dimethylglycine derivatization	[54, 96]
PPI	TMS-diazomethane reaction	[64]
BMP and PG	TMS-diazomethane derivatization	[79]
Double bond location(s) of FA chains	AMPP derivatization, Paternò–Büchi reaction Ozone reaction	[82, 84–86]
Aminophopholipids and plasmalogens	Sequential functional group selective derivatization	[38]

Abbreviation: DAG, diacylglycerol; MAG, monoacylglycerol; NAE, N-acylethanolamine; FA, fatty acyl or fatty acid; PPI, polyphosphoinositide; TMS-diazomethane, trimethylsilyl-diazomethane; BMP, bis(monoacylglycero) phosphate; PG, phosphatidylglycerol; AMPP, N-(4-amiomethylphenyl) pyridinium

3.1. Identification and quantification of lipid species containing a hydroxyl group

There are many kinds of lipid classes and species containing a hydroxyl group, such as diacylglycerol (DAG), monoacylglycerol (MAG), N-acylethanolamine (NAE), oxysterol, oxidized FAs (i.e., HETE), etc. Many of them are playing vital roles in cellular functions. For example, DAG species serve as intermediates in biosynthesis and degradation of triacylglycerols, glycerophospholipids, and glyceroglycolipids [49], and second messengers in many cellular processes [50]. Lots of research demonstrated that the aberrant metabolism of these lipids is associated with numerous disease states, including obesity, diabetes, brain jury, tumor, autoimmune diseases, etc. [51–53]. However, in addition to the presence in low abundance, the ionization efficiencies of these lipids are also relatively low because of their non-polar nature. Recently, following the same line of reasoning as aforementioned, our group has developed a new strategy to improve the analyses of these lipids through derivatization of the hydroxyl group with dimethylglycine (DMG) [54].

Taking the analysis of DAG species as an example, the representative product-ion ESI-MS spectra of the lithiated derivatives of DAG species with DMG are shown in Fig. 2. After CID, an abundant fragmentation ion (i.e., m/z 110) corresponding to lithiated DMG and a prominent product ion resulting from the neutral loss of 103 Da (i.e., DMG) were detected from all derivatives of lithiated DMG-DAG species. Therefore, these two fragment features can be employed to screen the presence of these lipids and quantify their mass levels (including isomers) [54].

Fig. 2.

Representative product-ion MS analyses of lithiated dymethylglycine (DMG)-diacylglycerol (DAG) species [54]. Product-ion analyses of 1,2-di16:0 (A), 1–16:0–2-18:1 DAG (B), 1,3-di16:0 (C), and 1–16:0–3-18:0 (D) DMG-DAG species were performed on a QqQ mass spectrometer (Thermo Fisher TSQ Vantage) with collision energy of 35 eV and collision gas pressure of 1.0 mTorr.

Compared with the product-ion mass spectrum of lithiated sn-1,2-DMG-DAG species, an additional moderate fragment ion corresponding to neutral loss of 87 Da (i.e., the loss of DMG as an aldehyde) existed in the product ion mass spectrum of sn-1,3DMG-DAG species (Fig. 2C). Thus, NLS 87 can be used to distinguish 1,3-DAG species from their 1,2-isomers [54]. The product-ion mass spectra of lithiated DMG-DAG species also displayed one or two pair(s) of specific fragment ions generated by the neutral loss(es) of FA chain(s) or lithium salt(s) (Fig. 2). Thus, the aliphatic chains and regioisomers of DAG could be determined according to the fragmentation features.

Typically, the intensities of the fragment ions corresponding to the loss of FA-Li salt from 1,2-DMG-DAG ions are much lower than those yielded from the loss of FA, while their intensities are almost the same in the product-ion mass spectra of lithiated 1,3-DMG-DAG species (Fig. 2). Based on this property, 1,2-DAG species could also readily be discriminated from its counterpart 1,3-isomers. Additionally, the intensity of the fragment ion resulted from the loss of sn-1 FA is higher than that of sn-2 FA due to the α-hydrogen atoms of FA substituent are more labile than that of DMG substituent [55]. Hence, this difference could be used to identify the FA position of 1, 2-DAG species. After quantification by using NLS87 for 1,3-DAG species in comparison to the IS, NLS103 is utilized to calculate the total DAG mass of the selected ion peak. The NLS spectra of naturally occurring FAs are used to determine the compositions of each isomer of the selected DAG ion peak [19, 54].

It has been demonstrated that an identical approach could also be employed for identification and quantification of oxyterol, MAG, and NAE (including isomers) [54, 56]. In summary, this strategy enhances shotgun lipidomics for global analyses of lipid species containing a hydroxyl group (including their isomers), accelerating the recognition of underling mechanisms of relevant disease states.

3.2. Identification and quantification of polyphosphoinositide (PPI) species

PPI species are a class of cellular membrane lipids, deriving from phosphorylation /dephosphorylation of phosphatidylinositol (PI) and its phospho-derivatives at the position 3, 4, and/or 5 of the inositol ring in the lipid head group [57]. Accumulated studies demonstrated that PPI species or their products play vital roles in numerous cellular processes, including membrane trafficking, cell growth, survival, and motility [58, 59]. Therefore, identification and quantification of PPI species are very important, contributing to providing deeper insights into the biochemical mechanisms underpinning the diseases (e.g., obesity and diabetes) and discovering potential biomarkers for early diagnosis of these diseases.

However, analysis, specially accurate quantification, of PPI species was a big challenge for the following reasons: (1) they are highly polar so that it is more difficult to recover these lipids from biological samples compared to the majority of other membrane lipids; (2) their structures are very similar, such as PI(3)P, PI(4)P, PI(5)P, PI(3,4)P₂, PI(3,5)P₂, PI(4,5)P₂, and PI(3,4,5)P₃; and (3) their concentrations in biological systems are always very low. Although many analytical methods have been developed, all of them have different limitations [60–63]. For example, a strategy to enhance the analyses of these lipids through derivatization of the phosphate groups with trimethylsilyl-diazomethane (TMS-diazomethane) has been developed by Clark et al [62]. Through analysis of the methylated PPI species by LC-MS/MS, the approach could obviously improve the analysis sensitivity and give promising results for these lipid species. However, the approach is relatively time-consuming, and unable to identify and quantify all PPI species, which carry different FA chains and/or phosphate positions [62].

Recently, to achieve a comprehensive analysis of individual PPI isomers in the presence of IS, the approach of derivatization PPI species with TMS-diazomethane was further exploited by our group [64]. Taking the analysis of PIP₂ species as an example, the mass analyses of lithiated derivatives of PIP₂ classes demonstrated that individual PIP₂ classes have class-specific methylation patterns, that is, methylation of five sites was mainly in the classes of PI(3,4)P₂ and PI(4,5)P₂, while methylation of six groups in PI(3,5)P₂ was predominant. However, the ionization efficiencies of differently methylated ions of a class of PIP₂ species are essentially identical because the total ion counts summarizing all the methylated ions of the selected species were virtually identical to each other among the classes. This observation also suggested that the FA chains and the number of methylation sites as well as the site distribution have minimal impact on the ionization efficiency. Moreover, the lithiated species of PIP₂ class containing the same number of methyl groups yield an identical fragment ion (i.e., m/z 497.1, 511.1, 525.1, and 539.1) corresponding to lithiated polymethylphosphoionsitol (Fig. 3). Therefore, PIS analysis of these different methylated species in the positive-ion mode can be used to screen the presence of these lipids and quantify the total mass levels of the ions (including isomers) in the presence of an IS. Additionally, the class-specific methylation patterns of PIP₂ (as aforementioned) could be used for simulation of the mixtures of isomers, by which quantification of individual PIP₂ species can be achieved. There exist abundant fragment ions resulting from FA constituent(s) and an ion at m/z 125 corresponding to dimethylphosphate in product-ion mass spectra of PIP₂ species in the negative-ion mode. The features in the negative-ion mode can be exploited for identification of FA substituent(s) of individual PIP₂ species. It is noteworthy that IS for quantification of PIP₂ should be added at the earliest extraction step.

Fig. 3.

Representative scheme of the methylation reaction of PIP₂ with TMS-diazomethane and their resultant ions in MS and tandem MS analyses [64].

In summary, depending on the characterized features of MS/MS of methylated PIP₂ species in both positive- and negative-ion modes in combination with the classspecific methylation patterns of PIP₂, global analysis of individual PIP₂ species including phosphorylation isomers and FA chains are achieved. The procedures for analysis of other PPI species (i.e., PIP and PIP₃) are similar [64].

The established approach has been successfully applied to the analysis of PPI species present in brain cortex and liver samples of db/db mice. The comprehensive alterations of PPI species in this model are revealed for the first time. This application not only demonstrates the effectiveness of this approach, but also provides insights into the underlying mechanism(s) responsible for the diabetes-induced cognitive decline, as well as the association between diabetes and Alzheimer’s disease [64].

3.3. Analysis of isomeric bis(monoacylglycero)phosphate (BMP) and phosphatidylglycerol (PG) species

BMP is a class of negatively charged glycerophospholipids with an unusual sn-1/sn-1´ structural configuration. BMP is primarily enriched in late endosomal/lysosomal membranes [65] and plays important roles in glycosphingolipid degradation and cholesterol transport [66]. Elevated BMP is associated with many lysosomal storage diseases including mucopolysaccharidosis, Niemann-Pick disease type A/B/C, Gaucher disease, and Fabry disease [67–70]. PG, structural isomers of BMP, accounts for ~1–2% of phospholipids in most animal tissues. Like BMP, PG also involves in many vital cellular processes, such as serving as a precursor for the biosynthesis of cardiolipin [71], activating RNA synthesis [72] and a nuclear PKC [73], and inhibiting platelet-activating factor [74] and PC transferring [75].

However, although BMP and PG are structural isomers, both of them have an identical fragmentation pattern in tandem MS. It is virtually impractical to distinguish these two classes of lipids using shotgun lipidomics without prior extensive separation by HPLC. Therefore, mostly current methods for analysis of PG and BMP are based on HPLC-MS [76–78]. It is well known that HLPC-based studies on lipids are usually not comprehensive because of time restriction for totally identification of FA substituents. Therefore, a lack of effective approaches is hindering the extensive investigation of cellular function and metabolism of PG and BMP.

The derivative products of selected lipid species could generate new fragmentation patterns. On the basis of the principle and the experience of methylation reaction successfully used in analysis of PPI species, our group has further exploited the strategy to enhance shotgun lipidomics for quantitative analysis of isomeric BMP and PG species [79]. After methylation reaction with TMS-diazomethane, we demonstrated that the fragmentation patterns of ammoniated/lithiated Me-PG and Me-BMP species are very different. For instance, only one abundant ion resulting from the neutral loss of 203 Da (methylated glycerophosphate derivative) is present in the product ion mass spectra of all ammoniated Me-PG species examined, while this ion is very minimal (absent) in those of Me-BMP species, which only display one or two fragmentation ions corresponding to the losses of FA chain(s). These characters can be used to distinguish these isomers [79].

Specifically, individual PG and BMP species as well as their potential mixtures present in lipid extracts of biological samples are identified by separate product ion analysis of individual ions and their total contents are quantified in comparison to an IS by high mass accuracy/resolution MS. After a methylation reaction, NLS203 spectra of the methylated lipid extracts are acquired for identification and quantification of Me-PG species. Then, identification and quantification of individual Me-BMP species can be derived from these analyses [79]. The principle of the approach for analysis of isomeric PG and BMP species is illustrated in Fig. 4.

Fig. 4.

Schematic illustration of the methylation reaction with TMS-diazomethane for analysis of isomeric PG and BMP species (Panel A). Full-scan mass analysis (panel B) and neutral loss scan of 203 Da (panel C) after methylation [79].

The strategy has been successfully used for determining the amount of PG and BMP species in mouse liver and changes of their levels following high-fat diet (HFD) feeding (a diabetic state) for 9 months, revealing that the most of the quantified BMP species are significantly elevated after HFD and these BMP species contain a large amount of polyunsaturated FAs. Therefore, the developed strategy enable us to effectively and extensively investigate the metabolism and function of BMP under different states [79].

3.4. Identification of double bond location(s) of FA chains

In addition to being the essential structural building blocks of complex lipids, FAs play many significant roles in biological systems. Different chain lengths and numbers of double bonds of FAs lead to the complex family of FA species. Moreover, different locations of double bonds also form the isomers of an unsaturated FA (e.g., 18:1(n-7), 18:1(n-9), and 18:1(n-12) FA species). Their composition reflects the dietary history and FA biosynthesis because they come from different sources [14]. Therefore, the altered composition indicates the physiological/pathological responses of FA metabolism after a perturbation. Obtaining the information about their composition may lead us to understand the molecular mechanism(s) of the associated diseases [80]. However, the tandem MS approach, especially the mass spectrometer with relatively lower collision energy (e.g., less than 100 eV in QqQ type), isn’t effective in defining double bond locations of FAs directly from lipid extracts, because of the high bond dissociation energies associated with cleaving a double bond. Without characteristic fragment ions produced, it is difficult to identify and quantify double bond positional isomers using a tandem MS approach.

Recently, great progress has been made in this field with the joint efforts of many researchers [81–86]. Our group exploited the charge-remote fragmentation nature in MS to achieve comprehensive analysis of lipids containing a carboxylic acid through derivatization with AMPP [82]. In fact, this method can also be used to identify the location of double bonds. Briefly, the locations of proximal double bond in AMPP-derivatized FAs are determined by diagnostic fragment ions arising from the markedly reduced 1,4-hydrogen elimination from the proximal olefinic carbons. In addition, fragmentation patterns resulting from allylic cleavages can also help the double bond position assignments [82]. In our study, the approach has been successfully employed to determine the compositions of three 18:1 FA isomers (i.e., 18:1(n-7), 18:1(n-9), and 18:1(n-12) FA species) in the mixture [82]. However, it is still difficult to identify the double bond positions of FAs present in complex lipids using this approach.

Blanksby and coworkers had developed the ozone-induced dissociation method for elucidation of double bond location [86]. However, the strategy requires special MS instruments that are not readily available, and the quantitation has rarely been successful in analysis of complex lipid mixture [86]. Recently, an alternative approach based on the coupling online photochemical Paternò–Büchi (PB) reaction with tandem MS has been exploited for pinpointing double bond locations in various kinds of lipid species (Fig. 5) [84], where acetone was used as the PB reagent for UV irradiation. Due to addition of acetone to intact unsaturated lipids, the reaction produced the ions with a mass shift of +58 Da compared to the corresponding ions previously obtained from these lipids. Additionally, after low-energy CID, the PB derivatives of unsaturated lipids yield abundant and characteristic paired fragment ions (∆m 26 Da) by cleavage at the original double bond locations. Therefore, these paired fragment ions, also termed the double bond diagnostic ions, can be used for determination of double bond location and quantification of these unsaturated lipid isomers [84]. This method has been employed for identification of double bond locations of 96 unsaturated FAs and glycerophospholipids present in rat brain tissue, revealing that about 50% of these determined lipids are mixtures of double bond location isomers for the first time [84]. It is anticipated that this approach could further broaden the scope of lipidomics research and provide us more information about cellular lipidomes.

Fig. 5.

Scheme of Paternò–Büchi reaction and formation of the diagnostic ions which locate the double bond positions from unsaturated lipids [84].

3.5. Sequential functional group selective derivatization of aminophospholipids and plasmalogens

As mentioned above, chemical derivatization is one of the most powerful strategies to enhance shotgun lipidomics for global analysis of selected lipid species utilizing the unique chemical properties of head groups of these lipid classes. Usually, every derivatization method aims at a subgroup or a few subgroups of these lipids which have a common functional group. In general, to avoid interference among lipids, these reactions are individually performed, which is time-consuming if multiple derivatization reactions should be conducted. Thus, an approach which could be used to analyze a few of different classes/subclasses of lipids at a time through a series of reactions is desirable.

Aminophospholipids (e.g., ethanolamine glycerophospholipid (PE) and phosphatidylserine (PS)) with other glycerophospholipids (e.g., PA, PC, and PG), and O-alkyl and O-alk-1´-enyl (plasmalogen) ether phospholipids may have the same ionic elemental compositions in ESI-MS analysis. To resolve the overlaps between them, many methods including intrasource separation [20, 21], acid hydrolysis [20], and derivatization techniques (e.g., Fmoc chloride [87], N-methylpiperazine acetic acid NHS ester [88], and 4-(dimethylamino)benzoic acid NHS ester derivatization [89]) have been developed based on their specific properties, but every approach can only resolve one kind of overlap. Reid group has developed a sequential functional group selective derivatization strategy for aminophospholipids and plasmalogen lipid species, maximally exploiting the features of every reaction and solving the two kinds of overlaps simultaneously [38].

Specifically, first, the amino groups of aminophospholipids are derivatized with a ¹³C₁-S,S´-dimethylthiobutanoyl-N-hydroxysuccinimide ester (¹³C₁-DMBNHS) reagent, containing a “fixed charge” sulfonium ion. This reaction not only enhances the ionization efficiency of these lipid species, increasing the detection sensitivity, but also produces the ions with a mass shift, which helps to distinguish aminophospholipids (i.e., PE and PS) from other glycerophospholipids. After the DMBNHS derivatization reaction, the selective derivatization of the plasmalogen O-alk-1´-enyl double bond using iodine and methanol can be performed sequentially without additional sample handing or cleanup between reaction steps prior to MS analysis (Fig. 6). Therefore, the differentiation of plasmalogens and unsaturated O-alkyl ether-containing lipids can be readily achieved [38]. Both reactions are simple, fast, and can be batch-conducted in 96 well plates. Remarkably, these reactions can be processed to completion with nonspecific modification of other lipid classes [90, 91]. Additionally, the sequential chemical derivatizations can also enhance lipid structural characterization and relative quantification by shotgun lipidomics [38].

Fig. 6.

Schematic illustration of sequential chemical derivatizations for comprehensive analysis of aminophospholipids and plasmalogen lipids [38].

3.6. Others

In addition to the derivatization strategies aforementioned, many novel approaches have been developed to enhance shotgun lipidomics for global analysis of cellular lipidomes [92–95]. For instance, methylation of amine as well as phosphate functional groups with ¹³C-diazomethane has been demonstrated to significantly improve lipid ionization, resolves overlaps of some isobaric and isomeric lipids, and alters fragmentation behaviors of PE, PS and PC species, and thus simultaneously enhances comprehensive analyses of several phospholipid classes in lipid extracts of biological samples [93]. Single-phase methyl tert-butyl ether extraction and low temperature diacetyl derivatization are adopted to prevent α-hydroxy migration within MAG species to facilitate unambiguous characterization of their regiospecific sn-positional isomers [94]. Therefore, it is anticipated that more and more approaches would be explored in the future.

4. Summary and perspective

Although the lipidomics discipline only emerged around 2003, great advances have been made in the field with development of various analytical technologies. Shotgun lipidomics is one of the most powerful analytical approaches in global analysis of cellular lipidomes directly from lipid extracts of diverse biological samples. However, due to the limitations of shotgun lipidomics, analysis of some types of isomers and poorly ionized lipids, particularly those species in very low abundance, remains a huge challenge. Many novel strategies (e.g., chemical derivatization) have been developed to enhance shotgun lipidomics for comprehensive analysis of cellular lipidomes. Chemical derivatization is one of the most effective strategies, since it not only selectively enhances ionization efficiency of lipid species of interest, but also generates new fragmentation patterns, yielding abundant, informative, and characteristic fragments. These fragments can be exploited to identify and quantify isomeric lipids with identical or very similar fragmentation patterns. Therefore, derivatization strategies further increase the coverage of shotgun lipidomics for analysis of unionizable lipids with higher accuracy/precision.

Shotgun lipidomics has been successfully employed for investigations in a variety of pathophysiological conditions. The findings from the research strongly support that lipidomics not only identifies the alteration of lipid species, which may be potential biomarker for disease diagnosis and/or prognosis in different states, but also provides insights into the underpinning biochemical mechanism(s) responsible for the pathophysiological states. We believe that shotgun lipidomics can make great contributions to metabolism and translational research with development of novel enhancement strategies.

Highlights.

Review the novel strategies for enhancing shotgun lipidomics for comprehensive analysis of cellular lipidomes

Abbreviations:

AMPP

N-(4-amiomethylphenyl) pyridinium

BMP

bis(monoacylglycero) phosphate

CID

collision-induced dissociation

DAG

diacylglycerol

¹³C₁-DMBNHS

¹³C₁-S,S’-dimethylthiobutanoyl-N-hydroxysuccinimide ester

DMG

dimethylglycine

FA

fatty acyl or fatty acid

HPLC

high performance liquid chromatography

IS

internal standard

MAG

monoacylglycerol

MS

mass spectrometry

MSMD-SL

multidimensional MS-based shotgun lipidomics

NAE

N-acylethanolamine

NLS

neutral-loss scanning

PB

Paternò–Büchi

PC

choline glycerophospholipid

PE

ethanolamine glycerophospholipid

PG

phosphatidylglycerol

PI

phoshatidylinositol

PIS

precursorion scanning

PPI

polyphosphoinositide

PS

phosphatidylserine

SM

sphingomyelin

TMS

diazomethane trimethylsilyl-diazomethane