Response enhancement for the evaporative light scattering detection for the analysis of lipid classes and molecular species

Choice of buffer in mobile phase can substantially alter peak areas in quantification of lipids by HPLC-ELSD

Evaporative light scattering detectors (ELSD) are commonly used with high-performance liquid chromatography (HPLC) to separate and quantify lipids, which are typically not easily detectable by more conventional methods such as UV-visible detectors. In many HPLC-ELSD methods to analyze lipids, a volatile buffer is included in the mobile phase to control the pH and facilitate separation between lipid species. Here, we report an unintended effect that buffer choice can have in HPLC-ELSD analysis of lipids - the identity and concentration of the buffer can substantially influence the resulting ELSD peak areas. To isolate this effect, we use a simple isocratic methanol mobile phase supplemented with different concentrations of commonly used buffers for ELSD analysis, and quantify the effect on peak width, peak shape, and peak area for seven different lipids (POPC, DOPE, cholesterol, sphingomyelin, DOTAP, DOPS, and lactose ceramide). We find that the ELSD peak areas for different lipids can change substantially depending on the mobile phase buffer composition, even in cases where the peak width and shape are unchanged. For a subset of analytes which are UV-active, we also demonstrate that the peak area quantified by UV remains unchanged under different buffer conditions, indicating that this effect is particular to ELSD quantification. We speculate that this ELSD-buffer effect may be the result of a variety of physical phenomenon, including: modification of aerosol droplet size, alteration of clustering of analytes during evaporation of the mobile phase, and mass-amplification or ion-pair effects, all of which could lead to differences in observed peak areas. Such effects would be expected to be molecule-specific, consistent with our data. We anticipate that this report will be useful for researchers designing and implementing HPLC-ELSD methods, especially of lipids.

Analytes’ Structure and Signal Response in Evaporating Light Scattering Detector (ELSD)

Background:

Working with an Evaporative Light Scattering Detector (ELSD), the target components are converted to a suspension of particles in a gas phase by a nebulizer and heated while the mobile phase is evaporated. Then, the incident light is directed at the remaining particles which are scattered and detected.

Methods:

The signal response of an ELS detector is studied through the correlation of the signal intensity of 65 compounds (at 30, 45 and 80°C) with their structural and physicochemical characteristics. Therefore, 67 physicochemical properties as well as structural features of the analytes were inserted as X variables and they were studied in correlation with their signal intensity (Y variable).

Results:

The collected data were statistically processed with the use of partial least squares method. The results proved that several properties were those that mainly affected the signal intensity either increasing or decreasing this response.

Conclusion:

The derived results proved that properties related to vapor pressure, size, density, melting and boiling point of the analytes were responsible for changes in the signal intensity. The light detected was also affected by properties relevant to the ability of a molecule to form hydrogen bonds (HBA and HBD) and its polarizability or refractivity, but at a lower extent.

Quantification of Lipids: Model, Reality, and Compromise

Lipids are key molecules in various biological processes, thus their quantification is a crucial point in a lot of studies and should be taken into account in lipidomics development. This family is complex and presents a very large diversity of structures, so analyzing and quantifying all this diversity is a real challenge. In this review, the different techniques to analyze lipids will be presented: from nuclear magnetic resonance (NMR) to mass spectrometry (with and without chromatography) including universal detectors. First of all, the state of the art of quantification, with the definitions of terms and protocol standardization, will be presented with quantitative lipidomics in mind, and then technical considerations and limitations of analytical chemistry's tools, such as NMR, mass spectrometry and universal detectors, will be discussed, particularly in terms of absolute quantification.

Principles of multiparametric optimization for phospholipidomics by 31P NMR spectroscopy

Phospholipids have long been known to be the principal constituents of the bilayer matrix of cell membranes. While the main function of cell membranes is to provide physical separation between intracellular and extracellular compartments, further biological and biochemical functions for phospholipids have been identified more recently, notably in cell signaling, cell recognition and cell-cell interaction, but also in cell growth, electrical insulation of neurons and many other processes. Therefore, accurate and efficient determination of tissue phospholipid composition is essential for our understanding of biological tissue function. ³¹P NMR spectroscopy is a quantitative and fast method for analyzing phospholipid extracts from biological samples without prior separation. However, the number of phospholipid classes and subclasses that can be quantified separately and reliably in ³¹P NMR spectra of tissue extracts is critically dependent on a variety of experimental conditions. Until recently, little attention has been paid to the optimization of phospholipid ³¹P NMR spectra. This review surveys the basic physicochemical properties that determine the quality of phospholipid spectra, and describes an optimization strategy based on this assessment. Notably, the following experimental parameters need to be controlled for systematic optimization: (1) extract concentration, (2) concentration of chelating agent, (3) pH value of the aqueous component of the solvent system, and (4) temperature of the NMR measurement. We conclude that a multiparametric optimization approach is crucial to obtaining highly predictable and reproducible ³¹P NMR spectra of phospholipids.

Retention behaviour of polyunsaturated fatty acid methyl esters on porous graphitic carbon

Retention with porous graphitic carbon was investigated with 25 structures of fatty acid methyl esters (FAMEs) with two different mobile phases: CH(3)CN:CHCl(3) 60:40 (v/v) and CH(3)OH:CHCl(3) 60:40 (v/v) with both 0.1% triethylamine (TEA) and an equimolar amount of HCOOH. Preliminary results showed that the use of TEA/HCOOH led to the response increase of saturated FAMEs with evaporative light scattering detection. No increase was observed for unsaturated one. These modifiers may slightly reduce the retention of FAMEs but did not significantly modify the separation factor with porous graphitic carbon. Thermodynamic parameters were calculated for each structure using Van't Hoff plot measured over the temperature range from 10 to 50 degrees C, with the both mobile phase conditions. All the studied compounds were found linked by the same retention mechanism on porous graphitic carbon. Quantitative in silico analysis of the retention using a molecular mechanics calculation demonstrated a good correlation between the retention factors and the molecular interaction energy values (r>0.93). Especially the Van der Waals energy was predominant, and the contribution of electrostatic energy was negligible for the quantitative analysis of the retention. The results indicate that Van der Waals force, hydrophobic interaction, is predominant for the retention of FAMEs on this packing material. The relative retention for highly unsaturated homologues can be changed by the selection of the weak solvent CH(3)CN or CH(3)OH. Then isomers differing only in the position of the carbon double bond on the alkyl chain can be separated and their behaviour is summarised as the closer the carbon double bonds to the FAME polar head, the more the retention decreases. Finally, the more important the number of carbon double bonds in the alkyl chain is, the smaller the retention is.

Evaporative light scattering detection: trends in its analytical uses

Evaporative light scattering detection (ELSD) is widely recognized as a universal tool for liquid and supercritical chromatographies. In addition, this detection technique is fully compatible with continuous-flow systems. In fact, the combination of continuous non-chromatographic techniques and ELSD affords the design of simple, reliable systems for extracting qualitative information. This paper reviews instrumental innovations regarding the miniaturization of evaporative light scattering detectors and their uses in micro and capillary liquid chromatography; also, it discusses their increasingly important role in the development of vanguard configurations for sample screening and the determination of total indices without the need for chromatographic separation. Moreover, it compares them with other types of chromatographic detectors in terms of performance. Finally, the potential of ELSD for solving real-life analytical problems arising from the need to meet (bio)chemical information needs is illustrated with various selected applications.

Application of a xenon arc lamp as a light source for evaporative light scattering detection

The standard tungsten-halogen light source used in a commercial evaporative light scattering detector (ELSD) was replaced with a 180 W xenon arc lamp. The xenon arc lamp possesses a broader spectrum in the UV region than the halogen source. The influence of the UV transmittance of five selected solvents was studied with a size-exclusion chromatography column. This solvent parameter was not observed to influence the ELSD response between the two light source settings. With the solvents studied, better sensitivity was obtained with the xenon arc lamp than the halogen lamp. This high-energy source was applied to ceramide III analysis with an octadecyl-grafted silica column and methanol:tetrahydrofuran 97:3 as the mobile phase, and the sensitivity of the quantification of ceramide III increased 16-fold for injected amounts of 14 approximately 140 ng. The molecular species in a sample of naturally occurring ceramides was analyzed using two C18 columns at 40 degrees C and gradient elution from 100% acetonitrile to 100% isopropanol in 30 min. The increased ELSD sensitivity achieved when using the xenon arc lamp allowed both the minor and major ceramide species to be observed, in contrast to the results achieved when the halogen lamp was used, where the increased photomultiplier voltage needed to observed the signals from the minor species caused the signals from the major ceramide species to occur above the detector response window.

Microanalytical systems for separations of stratum corneum ceramides

The small amount of lipids from human skin obtained with noninvasive sampling method led us to investigate microanalytical separation techniques. The lipid class analysis was performed with a micro polyvinyl alcohol-silica (PVA-Sil) column. The gradient elution was from heptane to acetone/butanol 90:10 v/v in 4%/min at 78 microL/min. In addition an evaporative light scattering detector (ELSD) was modified for micro-LC. All solvents contained 0.1% of triethylamine and formic acid in stoichiometric amount, which increased the ELSD response. In these conditions, the cholesterol eluted before free fatty acid, and squalene and triglycerides close to the dead volume. The various ceramide classes eluted following the order of the increased number of hydroxyl groups. The LOD for ceramides was 2.2 ng. The advantages of this method are the use of a normal stationary phase more reliable due to its chemical stability, its surface homogeneity and its development in microchromatography without chlorinated solvents which offers small LOD and the whole profile of lipids present in stratum corneum (SC). A method using a narrow-bore PVA-Sil column was used to collect ceramide fraction. Then the molecular species were analysed with a porous graphitic carbon column in capillary LC using a gradient from CH3OH/CHCl3 70:30 v/v to CHCl3 at 2%/min with a flow rate at 5 microL/min. The LOD obtained for ceramide was 1 ng. Both methods were assessed with SC samples obtained by rinsing a 5.7 cm2 area of the forearm with 25 mL of ethanol.

Phospholipid hydrolysis in a pharmaceutical emulsion assessed by physicochemical parameters and a new analytical method

The aim of this work was to develop a simple high-performance liquid chromatography (HPLC) technique with evaporative light scattering detection (ELSD) for the separation and quantification of the major phospholipid (PL) and lysophospholipid (LPL) classes contained in a pharmaceutical phospholipid-based emulsion. In the established method, phosphatidylcholine (PC), phosphatidylethanolamine (PE), sphingomyeline (SM), lysophosphatidylcholine (LPC) and lysophosphatidylethanolamine (LPE) were separated with a PVA-Sil stationary phase and a binary gradient from pure chloroform to methanol:water (94:6 v/v) at 3.4%/min. The ELSD detection was enhanced using 0.1% triethylamine and formic acid in each gradient mobile phases. Factors such as stationary phase and ELSD drift tube temperature were optimized, concluding in optimal temperatures of 25 degrees C for separation and 50 degrees C for evaporation. This HPLC-ELSD method was then applied to a PL-emulsion exposed to autoclaving and accelerated thermal conditions at 50 degrees C. Hydrolysis of PC and PE followed first-order kinetics, representing only 45% of the total lipid mass after 3 months. The chemical stability was correlated to commonly measured formulation physical and physico-chemical parameters such as droplet size, emulsion pH and zeta-potential.

Optimisation of the separation of four major neutral glycosphingolipids: application to a rapid and simple detection of urinary globotriaosylceramide in Fabry disease

A simple method for the separation of the four major neutral glycosphingolipids, present in all human tissue, was developed. This gradient normal phase-HPLC method utilises a polyvinyl alcohol bonded stationary phase and an evaporative light-scattering detection (ELSD). Screening pure solvents in a binary gradient elution mode allowed, in a first step, to assess the behaviour of the studied solutes and to select the solvents for further mobile phase optimisation. The proportion of the remaining solvents was defined to reach a maximal resolution. The reduction of the analysis time and the enhancement of the signal were obtained by optimising the gradient slope and the flow-rate. Optimal levels of triethylamine and formic acid (TEA-FA) for the enhancement of the evaporative light scattering detector response were established at 0.1% (v/v). Thus, the optimal conditions for the separation of the four glycosphingolipids was obtained with a gradient elution from a 100% chloroform to a 100% acetone:methanol (90:10 (v/v)) mobile phase at 0.2 ml min-1, using a 10% min-1 gradient slope. Finally, this method was applied to detect the excess of one of the neutral sphingolipids, namely globotriaosylceramide (Gb3) in the urine of patients affected with Fabry disease. A liquid-liquid extraction of the sediments obtained from an aliquot of only ten ml of urine proved sufficient to detect the excess of Gb3 present in both hemizygote and heterozygote patients. In all, the ability of our method to detect abnormal amounts of Gb3 in urinary sediments could allow the diagnosis of weakly symptomatic Fabry patients in large screening programs

Adaptation of an evaporative light-scattering detector to micro and capillary liquid chromatography and response assessment

A commercially available evaporative light-scattering detection (ELSD) system was adapted for micro and capillary LC. Therefore the various parameters involved in the droplet formation during the nebulization step in the ELSD system were studied. It was shown that the velocity term in the Nukiyama Tanasawa equation remains constant, leading to droplets of the same order of magnitude for narrow bore and capillary columns. Consequently, the ELSD modification was performed by decreasing the internal diameter of the effluent capillary tube in the nebulizer nozzle and by keeping its external diameter constant. Next, response curves for a conventional and the developed micro and capillary LC were compared as to investigate why a linear ELSD response is often obtained when used in micro or capillary LC. By splitting the flow rate post column, we showed that the nebulization process was not at the origin of the phenomenon. For ceramide III and tripalmitin, the response curves were found to be non-linear. However the curvature was less significant when the columns internal diameter decreased. Calculated particle size profiles for micro or capillary LC suggest that the particle entering the detection chamber are bigger than under conventional LC conditions. Last, triethylamine and formic acid were used to increase the response of the detector. The response enhancement, expected from previous studies, was established for the two lipids involved in this study.

Post-column addition as a method of controlling triacylglycerol response coefficient of an evaporative light scattering detector in liquid chromatography–evaporative light-scattering detection

The non-linear response is generally the main limitation to the general quantitative use of evaporative light-scattering detection (ELSD). In the particular case of triacylglycerol (TG) analysis, we present a preliminary paper dealing with the use of post-column additives as a means of monitoring the response of such a detector. As TG can form molecular association complexes (ligand-ligate associations) with either cholesterol, urea or silver nitrate, we report the influence of the concentration of each of these chemical compounds in the liquid phase directed towards the ELSD system. The results show that the response coefficient b of the calibration curve either decreases from 1.25-1.30 to 0.51 or increases from 1.25-1.30 to 1.78 according to the nature and concentration of post-column additive. The use of cholesterol as additive, at a discrete concentration, may lead to a linear response curve (b = 1), i.e. to the direct proportionality of ELSD response versus the TG concentration, making quantitative analysis of such solutes easier. On the other hand, to improve sensitivity, the addition of silver nitrate may be chosen for an increase in b value.

Isolation of ceramide fractions from skin sample by subcritical chromatography with packed silica and evaporative light scattering detection

Separative method of lipid classes from the stratum corneum was developed with packed silica and supercritical CO2 containing 10% of methanol at 15 degrees C, 15 MPa and 3 ml min(-1). The elution order of lipid classes was first esterified cholesterol, triglycerides, squalene co-eluted in a single peak, then free fatty acids, free cholesterol, ceramides and finally glycosylceramides. The ceramides were eluted in several fractions which depended on the number of hydroxyl groups in the molecule, i.e. more hydroxyl groups were contained in ceramides, more important was the retention. Moreover, the retention was not altered by the presence of carbon double bond and variation of the alkyl chain length. The ceramide response with the evaporative light scattering detector was improved by turning the influence of the solvent nature on the response to advantage. Therefore, addition of various solvents with or without triethylamine and formic acid were tested in post-column due to the incompatibility of such modifiers with silica stationary phase. Thereby the solvent conditions for the separation and the detection can be adjusted almost independently. The response was greatly increased by post-column addition of 1% (v/v) triethylamine and its equivalent amount of formic acid in dichloromethane introduced at 0.1 ml min(-1) into the mobile phase. This device had allowed the detection of 400 ng of ceramide with a S/N = 21, whereas no peak was observed in absence of the post-column addition. Finally, the method was applied to the treatment of skin sample which led to highly enriched ceramide fraction.

Eluotropic strength in non-aqueous liquid chromatography with porous graphitic carbon

Porous graphitic carbon is an attractive packing for the chromatographic analysis of highly hydrocarbonaceous compounds with non-aqueous mobile phase. An eluotropic-strength scale of 10 pure organic solvents was established using the methylene selectivity from the fatty acid methyl ester homologous series (chain length between 18 and 31 carbon atoms). Eight binary mobile phases combining a weak solvent: methanol or acetonitrile with a strong solvent: toluene, chloroform, dichloromethane or tetrahydrofuran at different volume fractions phi of strong solvents (ranging from 0.3 to 1.0) were tested and their eluotropic strengths were then compared with those of pure solvents. The curves of the eluotropic strength versus the volume fraction of the strong solvent followed two different trends: linear or curved. The knowledge of the pure solvent strength is not sufficient to predict the eluotropic strength of solvent in the mixture. Then modelling of the eluotropic strength for binary mobile phases was envisaged in order to provide a prediction tool. This model was assessed for the establishment of the composition of eight iso-eluotropic mobile phases. Good assessment was found except in the case of toluene with acetonitrile where the difference between the predicted and the real value was the highest.

Structure-retention diagrams of ceramides established for their identification

Molecular species analysis of ceramides was carried out using porous graphitic carbon with gradient elution: chloroform-methanol from 45:55 to 85:15 with a slope at 2.7%/min. These conditions gave a linear relationship between retention data and structure of ceramides. It was demonstrated that linearity occurred when a high slope value of linear gradient elution was used. Thereby the linear diagram was evolved by plotting the adjusted retention time against the total number of carbon atoms of ceramide molecules. Each line represents one ceramide class. Such a Structure-Retention Diagram describes ceramide retention and thus constitutes an identification method using only retention data. This Structure-Retention Diagram was assessed and compared to another obtained from octadesyl-grafted silica in terms of their reproducibility, precision and ability to provide ceramide identification. Better identification was obtained using the results from both Structure-Retention Diagrams. This approach with a two-dimensional separation system allowed to take advantage of the specificity of both identification models.