10.4 Working with Lipids · Biochemistry 2027

Lipid analysis starts with the question and the sample design

The level of the analysis must be settled first: fatty acid composition, lipid class or molecular species
Fatty acid ratios are addressed by FAME-GC, but class identity and sn position cannot be answered from composition alone
That needs molecular-species LC-MS/MS or position-specific enzyme treatment
Detection, identification and quantification are distinct stages
A doubling of relative signal is not a doubling of absolute concentration
Hydrolysis by lipases and oxidation by oxygen, light and heat begin as soon as the sample is taken
Records should include sample mass, tissue type, storage time, freeze-thaw cycles, container material and when the internal standard was added
Technical replicates assess instrument and preparation reproducibility while biological replicates capture individual variation
Reporting technical replicates as biological replicates overstates statistical confidence

Figure 10.4-1Overview of the lipid analysis workflow from question and sample design to interpretation

Lipid extraction needs organic solvents

Most lipids are poorly soluble in water because of a large non-polar portion
Amphipathic lipids such as phospholipids and glycolipids interact with both water and organic solvents
With an immiscible organic phase the partition depends on polarity, ionisation, hydrogen bonding and solvent composition
Releasing membrane lipids bound to protein needs a combination of solvents with different properties
The partition coefficient K as organic over aqueous concentration is a model for defined conditions, and acidic or basic lipids change charge with pH
Recovery describes how much of the target lipid in the original sample was recovered
Folch uses chloroform and methanol with the lipid-rich layer below
Bligh-Dyer adjusts the solvent ratio for sample water
The MTBE method of Matyash places the lipid layer on top, which favours automated collection
An internal standard added before extraction corrects extraction losses, though it cannot correct every lipid unless recovery behaviour matches
Example: recovering 8.2 of 10.0 nmol of a labelled lipid is 82 percent for that standard

Figure 10.4-2Technical and biological replicates measure different sources of variation; coordinates are hypothetical

Adsorption chromatography separates lipids by polarity

Chromatography exploits different distribution between a stationary and a mobile phase
On silica in normal-phase TLC more polar components interact strongly and move less
Actual distance and order depend on mobile phase composition, hydrogen bonding, ionic effects and temperature
TLC spots a mixture on a thin adsorbent layer and develops it with a rising mobile phase
Separated spots are compared with standards under the same conditions by distance and detection response
An equal Rf alone does not establish chemical structure
Rf equals spot distance divided by solvent front distance, dimensionless between zero and one
A diffuse front or streaked spot increases uncertainty in the measurement
HPLC pumps a liquid mobile phase through a column of small particles
Normal phase uses a polar stationary phase and non-polar mobile phase, while reversed phase uses hydrophobic C18-type material
In reversed phase, chain length and unsaturation separate molecular species within a class
Detection is a separate issue, since ordinary fatty acids lack strong chromophores and give weak UV signal
Light scattering, charged aerosol detection or mass spectrometry reveal different lipid sets and change the detection limit and quantitation range

Figure 10.4-3Lipids distribute differently between water and organic layers depending on functional groups and solvent

Figure 10.4-4Position of the organic layer in chloroform-based and MTBE-based extraction

Figure 10.4-5Migration of spots on silica TLC; order can change with mobile phase polarity

Figure 10.4-6Simulated TLC with a 6.0 cm solvent front and spot distances of 1.3, 3.0 and 4.3 cm

Figure 10.4-7Basic logic of normal-phase and reversed-phase separation

Gas chromatography analyses volatile lipid derivatives

Gas chromatography needs the analyte to be volatile and stable under heating and carrier gas flow
Fatty acids are awkward to analyse directly because the carboxyl group causes strong intermolecular interaction
Converting them to fatty acid methyl esters with methanol makes the analysis tractable
For a free fatty acid, R-COOH plus methanol gives R-COOCH3 plus water
Transesterification of a triacylglycerol gives glycerol and three FAMEs, though in practice several steps are involved
Derivatisation mostly destroys the original lipid class and sn position information
GC-FID uses flame ionisation for quantitation of most organic compounds
GC-MS adds mass spectra of separated components to support identification
Co-eluting compounds, double bond position and cis-trans isomers require careful column and standard selection
Peak area relates to amount, but response factors differ between compounds, so uncorrected areas are only relative signals
Example: 32 percent saturated and 68 percent unsaturated describes composition only, not which triacylglycerol contained them

Figure 10.4-8Conceptual flow from complex lipid to FAME derivatives analysed by GC-FID or GC-MS

Figure 10.4-9Simulated GC chromatogram of a FAME sample; retention times and peak sizes are teaching values

Specific hydrolysis helps infer lipid structure

Hydrolysis is water-dependent cleavage of a covalent bond
Ester bonds can be cleaved chemically or by lipases and esterases
Complete hydrolysis of a triacylglycerol gives one glycerol and three fatty acids
Ceramide with its amide bond needs a different specificity such as ceramidase
Archaeal ether lipids resist ordinary ester hydrolysis
Some lipases act preferentially at sn-1 or sn-3
Analysing the monoacylglycerol and fatty acids produced allows partial inference of original position
Acyl migration or incomplete selectivity changes the interpretation
Confirming structure requires controls, defined reaction times, standards and independent methods
PLA1 cleaves sn-1, PLA2 cleaves sn-2, PLC cleaves between glycerol and phosphate and PLD between phosphate and head group
Example: DAG plus phosphocholine from phosphatidylcholine indicates PLC, while PLD gives phosphatidic acid and choline and PLA2 gives lysophospholipid and fatty acid

Figure 10.4-10Hydrolysis of a triacylglycerol and the relation to its main products

Figure 10.4-11Representative cleavage sites of phospholipases A1, A2, C and D in a glycerophospholipid

Mass spectrometry: strong structural clues and clear limits

Mass spectrometry measures m/z of gas-phase ions rather than the neutral molecular mass
LC-MS often uses electrospray ionisation, giving M+H and M+Na in positive mode and M-H in negative mode
One molecule can give several peaks, so adducts and charge states must be identified
A neutral mass of 500.0000 Da gives about 501.0073 for M+H, 522.9892 for M+Na and 498.9927 for M-H
MS/MS selects a precursor ion, fragments it and records product ions
Diagnostic fragments of phospholipid head groups and acyl fragments report class and chain information
Not every structural isomer is resolved under the same conditions
Accurate mass narrows composition, but sn position, double bond position, geometry and stereochemistry stay unresolved without more evidence
Structure-specific reactions, ion mobility, ozonolysis, derivatisation and standard comparison may be needed
Structural information increases from PC 34:1 to PC 16:0_18:1 to PC 16:0/18:1 to PC 16:0/18:1(9Z)
The rule is to name only the level actually distinguished

Figure 10.4-12Representative positive and negative ion forms of a neutral molecule M with mass differences

Figure 10.4-13A precursor ion selected in MS1 is fragmented in MS/MS and the product ion pattern examined

Figure 10.4-14Change in structural resolution from PC 34:1 through PC 16:0_18:1 and PC 16:0/18:1 to position and geometry

Lipidomics catalogues and explores lipids and their functions

The lipidome is the set of lipid molecules found in a given subject and state
Lipidomics identifies, quantifies and compares them systematically and links them to biology
The number of lipid classes and isomers is enormous and each technique favours different lipids, so no single experiment measures everything
LIPID MAPS uses eight categories: fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, sterol lipids, prenol lipids, saccharolipids and polyketides
Targeted analysis selectively detects and quantifies defined lipids while untargeted analysis searches broadly for differences
Not every peak in an untargeted run is an identified lipid, and processing algorithms change the output
A difference between groups does not make that lipid the cause of disease
Diet, age, temperature, cell number, sampling time and instrument batch can all shift the signal
Reproducibility, independent validation and agreement with synthesising or degrading enzymes are needed before mechanistic claims

Figure 10.4-15Lipidomics workflow from sampling and internal standards through extraction, separation, MS, annotation and statistics

Figure 10.4-16Relative signals measured in two hypothetical sample groups; difference size alone does not establish significance or causation

Quantitation: internal standards, calibration, recovery and matrix effects

In LC-MS peak area broadly depends on amount, but ionisation efficiency responds to structure, solvent, salts and matrix
Ion suppression means equal amounts can give unequal signals, so peak area does not equal absolute moles
The simplest correction is analyte area over internal standard area equals R times the mole ratio
Assuming R equals one is a teaching simplification, and real quantitation uses calibration standards
Example: analyte area 150,000, standard area 100,000, 2.00 nmol of standard and R of 1.00 give 3.00 nmol of analyte
Precision describes clustering of replicates and accuracy describes closeness to a reference value
The relative standard deviation is the sample standard deviation over the mean, times 100
The limits of detection and quantification mark the boundaries of detection and reliable quantitation and depend on the validation approach
A visible peak alone does not guarantee quantifiability

Figure 10.4-17Simulated example in which raw QC signal rises with analysis order while corrected variation is smaller

Quality control and statistical interpretation

A blank reveals background contamination from vessels, reagents and preparation
Pooled QC mixes portions of many samples and is injected repeatedly to monitor signal stability over the run
Internal standards, system suitability checks and randomised injection order are also core quality controls
A uniform rise in all peaks from instrument drift must not be read as activation of a pathway
Injection order coinciding with sample group can create a false group difference
Comparing hundreds of lipids at p below 0.05 produces many chance findings
Multiple-testing correction such as FDR should accompany effect size, confidence intervals, biological replicate number, distribution and detection limits
Statistical significance does not verify a mechanism, and absence of significance does not prove absence of effect
Example: PC 34:1 means of 100 and 140 give a 1.40-fold change and a log2 fold change near +0.485

Figure 10.4-18Overview of method choice depending on whether the question concerns class, fatty acid composition, intact species or double bond position

References

**Figure 10.4-1**Overview of the lipid analysis workflow from question and sample design to interpretation

**Figure 10.4-2**Technical and biological replicates measure different sources of variation; coordinates are hypothetical

**Figure 10.4-3**Lipids distribute differently between water and organic layers depending on functional groups and solvent

**Figure 10.4-4**Position of the organic layer in chloroform-based and MTBE-based extraction

**Figure 10.4-5**Migration of spots on silica TLC; order can change with mobile phase polarity

**Figure 10.4-6**Simulated TLC with a 6.0 cm solvent front and spot distances of 1.3, 3.0 and 4.3 cm

**Figure 10.4-7**Basic logic of normal-phase and reversed-phase separation

**Figure 10.4-8**Conceptual flow from complex lipid to FAME derivatives analysed by GC-FID or GC-MS

**Figure 10.4-9**Simulated GC chromatogram of a FAME sample; retention times and peak sizes are teaching values

**Figure 10.4-10**Hydrolysis of a triacylglycerol and the relation to its main products

**Figure 10.4-11**Representative cleavage sites of phospholipases A1, A2, C and D in a glycerophospholipid

**Figure 10.4-12**Representative positive and negative ion forms of a neutral molecule M with mass differences

**Figure 10.4-13**A precursor ion selected in MS1 is fragmented in MS/MS and the product ion pattern examined

**Figure 10.4-14**Change in structural resolution from PC 34:1 through PC 16:0_18:1 and PC 16:0/18:1 to position and geometry

**Figure 10.4-15**Lipidomics workflow from sampling and internal standards through extraction, separation, MS, annotation and statistics

**Figure 10.4-16**Relative signals measured in two hypothetical sample groups; difference size alone does not establish significance or causation

**Figure 10.4-17**Simulated example in which raw QC signal rises with analysis order while corrected variation is smaller

**Figure 10.4-18**Overview of method choice depending on whether the question concerns class, fatty acid composition, intact species or double bond position