Chemicals, solvents, and synthetic lipid standards were purchased from Sigma-Aldrich St. Yeast extract and peptone were from BD Lyngby, Denmark. Lipid extract of bovine liver was purchased from Avanti Polar Lipids. Exponentially growing S. Cells were killed by adding perchloric acid to a final concentration of mM. In short, aliquots of lipid extracts or synthetic lipid standards were loaded in well plates, mixed with Samples were infused using a back pressure of 1. ITMS 3 data were acquired using max injection time of ms, automated gain control at 1e4 and 1 microscan.
MS ALL analysis of mouse plasma, mouse hippocampus and bovine liver was performed as previously described [ 1 , 2 ]. Lipid species are annotated as previously described [ 44 , 46 , 47 ]. PI SM ;2 [ 48 ]. SE PS — For triacylglycerols and cardiolipins the third and fourth acyl groups are appended analogously.
Proposal for a common nomenclature for fragment ions in mass spectra of lipids
The acyl groups are indicated in the order of i increasing carbon number and ii increasing double bond number. It is implemented using PHP and designed for retrieving lipid ionization and fragmentation information stored in the underlying ALEX lipid database. This version of LDA was adapted to support the herein described nomenclature for glycerolipids and glycerophospholipids. To establish a nomenclature for shorthand notation of lipid fragment ions we first undertook a study to identify commonalities in the fragmentation pathways of lipid molecules. To this end, we performed a comprehensive analysis of lipid fragmentation using structurally-defined lipid molecules from 47 different lipid classes, covering five lipid categories that are common to eukaryotic organisms S1 Table.
Using an Orbitrap Fusion mass spectrometer, these lipid molecules were fragmented in both negative and positive ion mode except for TAG and sterol lipids using high resolution MS 2 and MS 3 analysis with quadrupole-based CID and ion trap-based resonance-excitation CID [ 1 ]. As such, the recorded lipid fragmentation data is comparable to that of a broad range of instruments spanning low resolution triple quadrupole and ion trap machines to high resolution hybrid quadrupole time-of-flight, ion trap- and quadrupole-Orbitrap mass spectrometers.
To systematically annotate detected fragment ions across the five categories of lipids and the different analytical conditions we devised a procedure featuring three consecutive steps Fig 2. The rationales for each of these steps and guideline for their implementation are described in full detail in S1 Text and summarized in the following sections. Fragment ion spectra with shorthand notation for representative lipid molecules spanning five different lipid categories are shown in Figs 3 and 4. Step 2: These fragments are then annotated using fragment type-specific annotation rules described in detail in S1 Text.
- Online Characteristic Differentiation In The Spectra Of Saturated Hydrocarbons .
- Managing Transitions: The Chinese Communist Party, United Front Work, Corporatism and Hegemony (East Asia (New York, N.Y.).).
- Carbon-carbon couplings;
- Android Application Development for Java Programmers.
- Application to crude oils using data mining techniques;
- Deceived (Star Wars: The Old Republic, Book 2)!
- Petroleum Standards.
The precursor ion is annotated at the lipid species level i. Note that each chemical reaction is mass-balanced i.
Each structure is represented with charge, monoisotopic mass, shorthand notation and fragment type. The annotations shown in boldface prioritized are based on annotation rules outlined in Fig 2 step 3.
Annotation shown in boldface is prioritized based on the guidelines outlined in Fig 2 step 3. Non-prioritized shorthand notation is occasionally omitted to avoid overly congested mass spectra. Generalizing lipid fragmentation in this manner highlights three fundamental concepts that are inherent to our nomenclature rules. Second, inspecting the structural attributes of lipid fragment structures shows that four types of fragments can be produced by CID: LCFs lipid class-selective fragments , MLFs molecular lipid species-specific fragments , DBFs double bond location-specific fragments and intermediate molecular lipid species-selective fragments iMLFs.
In brief, LCFs encompass common structures that are released from all lipid molecules belonging to the same lipid class, they have identical mass, and they do not contain a hydrocarbon chain. MLFs are characterized by structures having only one hydrocarbon chain with variations in the number of carbon atoms, double bonds and hydroxyl groups.
Depending on the lipid class, these hydrocarbon chains can be classified as FA, alkanol and alkenol i. The second step in the procedure implements specific rules for shorthand notation of both charged and neutral fragment structures i. These rules are listed in full detail in the S1 Text , and exemplified in Figs 3 and 4 and S3 Fig showing MS n spectra and mass-balanced chemical reactions of representative molecules from five different lipid categories.
Second, structures of LCFs should be annotated by the lipid class abbreviation and its nominal mass in parentheses e. Third, MLFs should be denoted by the class of HCA, its original number of carbon atoms, double bonds and potential hydroxyl groups, and followed by in parentheses specification of any chemical modification listed in accordance to Hill notation [ 52 ] e.
Fourth, iMLFs should be denoted by the class of HCA, its original number of carbon atoms, double bonds and potential hydroxyl groups, and followed by in parentheses any chemical modifications e. Fifth, DBFs should be denoted by the class of HCA, its original number of carbon atoms, number of double bonds and locations of double bonds, followed by in parentheses any chemical modification e. Hence, for spectral annotation it is advisable to prioritize the use of nomenclature based on fragment type and whether the fragment structure s is charged or neutral. For MLFs and iMLFs, the decision tree-based routine also prioritizes whether to use shorthand notation based on nomenclature for charged structures or the composite of neutral losses.
Moreover, this framework also provides an avenue for automatically and consistently curating lipid fragmentation information in databases and harnessing the information to support high confidence lipid identification. Furthermore, comparing our nomenclature to that of LipidBlast [ 54 ] shows that this software uses only the positional descriptors sn-1, sn-2 and sn-3 to denote MLFs e.
Attached proton test (APT)
This might be considered adequate for analysis of synthetic lipid standards where the name and the structure of the lipid molecule are known. However, this nomenclature format will produce misleading spectral annotations and false-positive lipid identifications when used for analyzing complex biological samples where both positional- and structural lipid isomers are present [ 29 , 38 , 41 ].
Currently, the ALEX lipid calculator provides ionization information for over 25, lipid species from more than 89 lipid classes at the MS 1 level. Furthermore, the database also features curated MS 2 and MS 3 fragmentation information for more than , molecular lipid species covering 49 different lipid classes S1 Table.
To our knowledge, this is currently the most comprehensive freely available resource with curated information on lipid ionization and fragmentation. We note that the spectral information in the ALEX lipid calculator also features additional metadata that for all lipid molecules and fragment ions specifies lipid category, lipid class, adduct ion, charge and chemical formula.
Journal list menu
The application is freely available at www. To exemplify how the nomenclature for shorthand notation of lipid fragment ions facilitates confident lipid identification we manually shortlisted a set of low abundance lipid molecules for which fragment ion intensity is expected to be of poorer quality as compared to fragment ions derived from more abundant lipid molecules.
These lipid molecules were detected by MS ALL analysis of mouse plasma [ 2 ], mouse hippocampus [ 1 ] and bovine liver. Detection of these structure-specific fragment ions, and the intact lipid molecule by FTMS 1 , demonstrated specific detection of ACar in mouse plasma. At first hand, this highly unsaturated PS molecule was somewhat puzzling and difficult to reconcile with lipid metabolic pathways in mammalian cells.
This information univocally demonstrates that the mouse hippocampus lipidome includes the highly polyunsaturated and low abundance molecular glycerophospholipid species PS — corresponding to 0. Of note, our data is corroborated by previous report indicating the presence of PS — PS in mouse brain [ 56 ] and raises the mechanistic questions as to how it is synthesized and what its molecular functions are? The precursor ion matching deprotonated PS is highlighted in boldface.
To exemplify the use how the fragment nomenclature supports confident lipid identification using MS 3 fragmentation we selected a low abundance TAG species detected in bovine liver. Taken together, these fragment ions confidently identify the low abundance molecular lipid species TAG —— in the background of the much more abundant isomeric species TAG —— As a proof of concept we subsequently embedded our nomenclature rules in LDA Lipid Data Analyzer [ 49 ], a software supporting automated high confidence lipid identification and quantification [ 50 ].
In addition to using multiple lipid fragment ions to support lipid identification and outputting quantitative information of lipids identified at the molecular lipid species-level this software also features a convenient user-interface for reviewing individual MS 2 spectra in which fragment ions can be automatically annotated Fig 7. To exemplify the possibility to automatically annotate detected lipid fragment ions we made use of a resource dataset featuring LC-MS 2 data on a lipid standard mixture containing 78 different synthetic standards, including PE — and DAG — Fragment ions are automatically annotated by LDA and collectively used to identify the molecular lipid species PE — Fragment ions are automatically annotated by LDA and collectively used to identify the molecular lipid species DAG — Collectively, these fragment ions unequivocally identify the precursor ion as the molecular lipid species PE — Fig 7D.
Collectively, these fragment ions unequivocally identify the precursor ion as the molecular lipid species DAG — Taken together, these examples demonstrate that the proposed nomenclature system not only facilities manual lipid identification as outlined in the previous section , but can also be used in conjunction with software-based routines to easily verify the fidelity of automated lipid identifications.
Such investigations can be performed by feeding cells or animals with a wide range of metabolic precursors labeled with 13 C, 2 H, 15 N and 18 O. Depending on the organism, stable isotope-labeled precursors can be incorporated into different structural attributes of a lipid molecule and when labeling with a cocktails of precursors the stable isotope-labeled precursors can be incorporated simultaneously into a single lipid molecule. This yields an additional dimension of lipid structural complexity that can be harnessed using high resolution MS ALL technology [ 1 ].
However, the increased lipid structural complexity also calls for implementation of a systematic nomenclature that can adequately denote fragment ions derived from molecular lipid species having distinct structural attributes labeled with stable isotopes. To support shorthand notation of fragment ions derived from stable-isotope labeled lipids we extended the rule set of the fragment nomenclature. Full details of these rules are provided in S1 Text. Note that non-annotated fragment ions derive from co-isolated lipids. This extension is exemplified in Fig 7 showing fragment ion spectra from four representative lipid molecules labeled with different configurations of heavy nuclei.
Of note, LCFs are denoted by the lipid class abbreviation followed first by specification of stable isotopes and then by the nominal mass in parentheses. Similarly, MLFs are denoted by HCA abbreviation followed first by specification of stable isotopes and then by any chemical modifications in parentheses.
In summary, the ability of the fragment ion nomenclature to consistently account for shorthand notation of also stable-isotope labeled lipid fragment ions demonstrates its generic nature and highlight that it can readily be extended to describe a wide range of fragment structures and accurately match these to structure-specific fragment ions detected by MS n analysis. We note that the ALEX lipid calculator at the present features 28 lipid classes labeled with stable isotopes that can be generated when feeding cells or animals with the commercially available metabolic precursors 2 H 9 -choline, 2 H 3 -methionine, 2 H 6 -inositol, 13 C 3 15 N-serine and their combinations S1 Table.
In this report we have outlined a generic framework for shorthand notation of lipid fragment ions. We have demonstrated that the nomenclature is able to systematically and consistently describe structural details of fragment ions released upon CID of unlabeled and stable isotope-labeled molecular lipid species encompassing 47 lipid classes and five different lipid categories.
chapter and author info
Furthermore, we have shown that the nomenclature can be computerized and made searchable in the online ALEX lipid calculator to support both manual and automated high confidence lipid identification in biological sample matrices. Notably, the fragment nomenclature framework also provides an avenue to develop new algorithms for automated high confidence lipid identification in high throughput lipidomics studies. This text-based information can, for example, be harnessed for counting the frequency of specific fragment types across large number of samples and also for implementing complementarily filters to secure high confidence lipid identification e.
Based on its systematic design and its ability to be easily computerized we deem that our proposed fragment nomenclature can become a valuable addition to the expanding palette of cheminformatics tools that are being developed to assist the characterization of lipid molecules in biological systems.