Conductivity detection is a universal detection technique often encountered in electrophoretic separation systems, especially in modern chip-electrophoresis based devices. On the other hand, it is sparsely combined with another contemporary trend of enhancing limits of detection by means of various preconcentration strategies. This can be attributed to the fact that a preconcentration experimental setup usually brings about disturbances in a conductivity baseline. Sweeping with a neutral sweeping agent seems a good candidate for overcoming this problem. A neutral sweeping agent does not hinder the conductivity detection while a charged analyte may preconcentrate on its boundary due to a decrease in its effective mobility. This study investigates such sweeping systems theoretically, by means of computer simulations, and experimentally. A formula is provided for the reliable estimation of the preconcentration factor. Additionally, it is demonstrated that the conductivity signal can significantly benefit from slowing down the analyte and thus the overall signal enhancement can easily overweight amplification caused solely by the sweeping process. The overall enhancement factor can be deduced a priori from the linearized theory of electrophoresis implemented in the PeakMaster freeware. Sweeping by neutral cyclodextrin is demonstrated on an amplification of a conductivity signal of flurbiprofen in a real drug sample. Finally, a possible formation of unexpected system peaks in systems with a neutral sweeping agent is revealed by the computer simulation and confirmed experimentally.
We introduce CEval software (downloadable for free at echmet.natur.cuni.cz) that was developed for quicker and easier electrophoregram evaluation and further data processing in (affinity) capillary electrophoresis. This software allows for automatic peak detection and evaluation of common peak parameters, such as its migration time, area, width etc. Additionally, the software includes a nonlinear regression engine that performs peak fitting with the Haarhoff-van der Linde (HVL) function, including automated initial guess of the HVL function parameters. HVL is a fundamental peak-shape function in electrophoresis, based on which the correct effective mobility of the analyte represented by the peak is evaluated. Effective mobilities of an analyte at various concentrations of a selector can be further stored and plotted in an affinity CE mode. Consequently, the mobility of the free analyte, μA, mobility of the analyte-selector complex, μAS, and the apparent complexation constant, K('), are first guessed automatically from the linearized data plots and subsequently estimated by the means of nonlinear regression. An option that allows two complexation dependencies to be fitted at once is especially convenient for enantioseparations. Statistical processing of these data is also included, which allowed us to: i) express the 95% confidence intervals for the μA, μAS and K(') least-squares estimates, ii) do hypothesis testing on the estimated parameters for the first time. We demonstrate the benefits of the CEval software by inspecting complexation of tryptophan methyl ester with two cyclodextrins, neutral heptakis(2,6-di-O-methyl)-β-CD and charged heptakis(6-O-sulfo)-β-CD.
A new, fast, selective, and reliable capillary electrophoresis method has been developed for analysis of selected phosphoesters (phosphoserine, phosphoethanolamine, phosphoglycerol) and phosphate. The method is based on separation of specific phosphate containing headgroups (phosphoesters) which are cleaved from the glycerol skeleton of a phospholipid by a regioselective enzyme (phospholipase C). Analysis of intact phospholipids with the same polar headgroup but different fatty acids shows that fatty acid composition has a high impact on separation of phospholipids, so analysis of separated polar headgroups, which avoids this influence, represents a much more suitable approach for phospholipid class research. Optimization of method parameters results in running buffers of relatively narrow pH interval (pH about 10) where all phosphoesters are separated. Further method validation has shown that direct UV detection has a sufficient detection limit for all analytes to perform suitable analyses of cell membrane lipids. The optimized method was tested on the lysate of cell membrane of Bacillus subtilis, where all analytes were determined.
Although the classical formula of peak resolution was derived to characterize the extent of separation only for Gaussian peaks of equal areas, it is often used even when the peaks follow non-Gaussian distributions and/or have unequal areas. This practice can result in misleading information about the extent of separation in terms of the severity of peak overlap. We propose here the use of the equivalent peak resolution value, a term based on relative peak overlap, to characterize the extent of separation that had been achieved. The definition of equivalent peak resolution is not constrained either by the form(s) of the concentration distribution function(s) of the peaks (Gaussian or non-Gaussian) or the relative area of the peaks. The equivalent peak resolution value and the classically defined peak resolution value are numerically identical when the separated peaks are Gaussian and have identical areas and SDs. Using our new freeware program, Resolution Analyzer, one can calculate both the classically defined and the equivalent peak resolution values. With the help of this tool, we demonstrate here that the classical peak resolution values mischaracterize the extent of peak overlap even when the peaks are Gaussian but have different areas. We show that under ideal conditions of the separation process, the relative peak overlap value is easily accessible by fitting the overall peak profile as the sum of two Gaussian functions. The applicability of the new approach is demonstrated on real separations.
In this paper we determine acid dissociation constants, limiting ionic mobilities, complexation constants with β-cyclodextrin or heptakis(2,3,6-tri-O-methyl)-β-cyclodextrin, and mobilities of resulting complexes of profens, using capillary zone electrophoresis and affinity capillary electrophoresis. Complexation parameters are determined for both neutral and fully charged forms of profens and further corrected for actual ionic strength and variable viscosity in order to obtain thermodynamic values of complexation constants. The accuracy of obtained complexation parameters is verified by multidimensional nonlinear regression of affinity capillary electrophoretic data, which provides the acid dissociation and complexation parameters within one set of measurements, and by NMR technique. A good agreement among all discussed methods was obtained. Determined complexation parameters were used as input parameters for simulations of electrophoretic separation of profens by Simul 5 Complex. An excellent agreement of experimental and simulated results was achieved in terms of positions, shapes, and amplitudes of analyte peaks, confirming the applicability of Simul 5 Complex to complex systems, and accuracy of obtained physical-chemical constants. Simultaneously, we were able to demonstrate the influence of electromigration dispersion on the separation efficiency, which is not possible using the common theoretical approaches, and predict the electromigration order reversals of profen peaks. We have shown that determined acid dissociation and complexation parameters in combination with tool Simul 5 Complex software can be used for optimization of separation conditions in capillary electrophoresis.
- MeSH
- antiflogistika nesteroidní chemie MeSH
- beta-cyklodextriny chemie MeSH
- elektroforéza kapilární metody MeSH
- flurbiprofen chemie MeSH
- ibuprofen chemie MeSH
- ketoprofen chemie MeSH
- koncentrace vodíkových iontů MeSH
- magnetická rezonanční spektroskopie MeSH
- naproxen chemie MeSH
- osmolární koncentrace MeSH
- počítačová simulace MeSH
- software MeSH
- termodynamika MeSH
- viskozita MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
We extended the linearized model of electromigration, which is used by PeakMaster, by calculation of nonlinear dispersion and diffusion of zones. The model results in the continuity equation for the shape function ϕ(x,t) of the zone: ϕ(t) = -(v(0) + v(EMD) ϕ)ϕ(x) + δϕ(xx) that contains linear (v(0)) and nonlinear migration (v(EMD)), diffusion (δ), and subscripts x and t stand for partial derivatives. It is valid for both analyte and system zones, and we present equations how to calculate characteristic zone parameters. We solved the continuity equation by Hopf-Cole transformation and applied it for two different initial conditions-the Dirac function resulting in the Haarhoff-van der Linde (HVL) function and the rectangular pulse function, which resulted in a function that we denote as the HVLR function. The nonlinear model was implemented in PeakMaster 5.3, which uses the HVLR function to predict the electropherogram for a given background electrolyte and a composition of the sample. HVLR function also enables to draw electropherograms with significantly wide injection zones, which was not possible before. The nonlinear model was tested by a comparison with a simulation by Simul 5, which solves the complete nonlinear model of electromigration numerically.
We introduce a computer implementation of the mathematical model of capillary zone electrophoresis described in the previous paper in this issue (Hruška et al., Electrophoresis 2012, 33), the program PeakMaster 5.3. The computer model calculates eigenmobilities, which are the eigenvalues of the Jacobian matrix of the electromigration system, and which are responsible for the presence of system eigenzones (system zones, system peaks). The model also calculates parameters of the background electrolyte: pH, conductivity, buffer capacity, ionic strength, etc., and parameters of the separated analytes: effective mobility, transfer ratio, molar conductivity detection response, and relative velocity slope. In addition to what was possible in the previous versions of PeakMaster, Version 5.3 can predict the shapes of the system peaks even for a complex injected sample profile, such as a rectangular plug. PeakMaster 5.3 can replace numerical simulation in many practically important configurations and the results are obtained in a very short time (within seconds). We demonstrate that the results obtained in real experiments agree well with those calculated by PeakMaster 5.3.
