Search

Our study attempts to find an approach to distinguishing between the contribution to peak spreading in capillary zone electrophoresis (CZE) due to protein microheterogeneity and that due to interaction with the capillary wall, by analyzing correlations between observed peak spreading and peak asymmetry. The peak asymmetry was measured as ln[(tm-t1)/(t2-tm)] where tm, t1, and t2 are migration times at the mode of the peak and at the intersection of the peak width at half-height with the ascending and descending limbs, respectively. Two isoforms of recombinant green fluorescent protein (GFP-1 and GFP-2, 27 kDa molecular mass), glucose-6-phosphate dehydrogenase (GPD, 104 kDa), and the naturally fluorescent protein R-phycoerythrin (PHYCO, 240 kDa) were subjected to CZE in polyacrylamide-coated fused-silica capillaries of 50 and 100 microns diameters under varying conditions of protein concentration, field strength, and the initial zone length. Under conditions such that contributions to peak spreading from axial diffusion, thermal effects, and electrophoretic dispersion are negligible, the analysis of the interrelations between peak width and peak asymmetry was found to allow a conclusion as to the cause of peak spreading in CZE of protein. It appears that the peak width of GFP-2 originates mostly in protein microheterogeneity while that of GFP-1 is due to protein-capillary wall interactions. For PHYCO, both microheterogeneity and protein-capillary wall interactions contribute to peak spreading. GPD exhibits relatively little microheterogeneity or interaction with capillary walls. Thus, its peak width appears to be mostly affected by an extracolumn source of spreading such as the initial zone length.

A fast and simple method for the internal coating of capillaries in capillary zone electrophoresis (CZE) is that with epoxy-poly(dimethylacrylamide) (EPDMA). Duration of coating by that method is 30 min, compared with that of 24 h when using uncross-linked polyacrylamide (PA) under otherwise identical conditions. Under the conditions used for the CZE of proteins (pH 9.0, 2% polyethylene glycol), the capillary coating with EPDMA is stable for at least 50 consecutive runs as judged by the constancy of low electroosmotic flow, equalling the stability of coating achieved by PA. Protein mobilities and protein peak asymmetry (suggestive of reversible interaction with the capillary wall) are also found to be the same in EPDMA and PA coated capillaries. Differences between EPDMA and PA coating also exist: The former is unstable upon lowering the ionic strength of the buffer to 0.003, upon the addition of sodium dodecyl sulfate (SDS) to the buffer and in application to the hydrophobic analyte, polystyrene carboxylate.

To gain insight into the mechanisms of size-dependent separation of microparticles in capillary zone electrophoresis (CZE), sulfated polystyrene latex microspheres of 139, 189, 268, and 381 nm radius were subjected to CZE in Tris-borate buffers of various ionic strengths ranging from 0.0003 to 0.005, at electric field strengths of 100-500 V cm(-1). Size-dependent electrophoretic migration of polystyrene particles in CZE was shown to be an explicit function of kappaR, where kappa(-1) and rare the thickness of electric double layer (which can be derived from the ionic strength of the buffer) and particle radius, respectively. Particle mobility depends on kappaR in a manner consistent with that expected from the Overbeek-Booth electrokinetic theory, though a charged hairy layer on the surface of polystyrene latex particles complicates the quantitative prediction and optimization of size-dependent separation of such particles in CZE. However, the Overbeek-Booth theory remains a useful general guide for size-dependent separation of microparticles in CZE. In accordance with it, it could be shown that, for a given pair of polystyrene particles of different sizes, there exists an ionic strength which provides the optimal separation selectivity. Peak spreading was promoted by both an increasing electric field strength and a decreasing ionic strength. When the capillary is efficiently thermostated, the electrophoretic heterogeneity of polystyrene microspheres appears to be the major contributor to peak spreading. Yet, at both elevated electric field strengths (500 V/cm) and the highest ionic strength used (0.005), thermal effects in a capillary appear to contribute significantly to peak spreading or can even dominate it.

R-phycoerythrin (PHYCO, Mr 240 000), glucose-6-phosphate dehydrogenase (GPD, Mr 104 000) and two charge isomers of recombinant green fluorescent protein (GFP-1 and GFP-2, Mr 27 000) were subjected to capillary zone electrophoresis (CZE) in capillaries of 50, 100 and 150 microm inner diameter at various sample concentrations, electric field strengths, and lengths of the initial zone with the purpose of testing the hypothesis that protein - capillary wall interactions rather than thermal effects are predominantly responsible for the peak spreading of proteins in CZE. The efficiency of CZE was expressed in terms of the number of theoretical plates, N, or the plate height corrected by subtracting the contribution from initial zone length, H'. The latter has the advantage of solely reflecting contributions to the separation efficiency arising from intracolumn peak spreading in capillaries of different diameters. The separation efficiency measured varied widely, by two orders of magnitude, for these proteins under identical conditions, with GPD exhibiting the highest and PHYCO the lowest values of N. H' was found to be independent of sample concentrations within the concentration ranges studied, 1-100 microg/mL for PHYCO and 100-1000 microg/mL for GPD, while exhibiting a decrease with sample concentration for GFP, especially in 150 microm diameter capillaries, within the concentration range 1-100 microg/mL. H'was also found to be independent of electric field strength up to 300-400 V/cm for PHYCO and GFP. In all experiments, the CZE of proteins in 100 microm diameter capillaries provided a higher or, at least, equal efficiency, compared to that in 50 or 150 microm diameter capillaries. It may be concluded that the protein - capillary wall interactions and protein microheterogeneity are the dominant sources of peak spreading and their specific combinations are thought to be responsible for the wide variation in separation efficiency between proteins in CZE observed under identical conditions.

Capillary zone electrophoresis (CZE) was conducted in buffered solutions of polyacrylamide (PA) and polyethylene glycol (PEG) to find the degree and the manner in which separation and resolution of submicron-sized rigid spherical polystyrene sulfate and carboxylate particles were affected by the presence of those polymers. In resolving pairs of representative particles, maximal resolution was observed at or near the entanglement threshold concentration, c*, of the polymer. The value of that maximum represents a several-fold increase in resolution. Since c* can be calculated from intrinsic viscosity, and the latter from the molecular weight of the polymer (and some constants available in the literature), optimally resolving polymer conditions become predictable. The maximum can also be experimentally determined by measuring intrinsic viscosity and calculating c*, or by either systematically varying the concentration of a polymer of constant molecular weight or by varying the molecular weight of a polymer at constant concentration. An optimally resolving field strength is superimposed on the maximally resolving condition of polymer concentration and weight.

The retardation of three "spherical" proteins with Stokes' radii of 2.0, 2.4, and 3.0 nm (35-104 kDa) was studied in capillary zone electrophoresis (CZE), using semidilute solutions of polyethylene glycol (PEG), linear polyacrylamide (PA), and polyvinyl alcohol (PVA). The purpose was to test the models predicting that the ratio of particle radius, R, to the mesh size of polymer network (the correlation or screening length of a semidilute polymer solution), xi, directly governs the size-dependent retardation in the form: mu/muo = exp (-R/xi). Here xi = kc-0.75, where c is polymer concentration and the numerical factor kcan be calculated based on polymer molecular weight. In application to polymers in a "good solvent" (PA and PEG in the aqueous buffer) and to proteins of 2.4 and 3.0 nm radius, that relation between relative mobility and R/xi was found to be obeyed for PA, while for PEG the value of k derived from retardation experiments significantly exceeded that which was theoretically calculated. Thus, the retardation appears to be polymer-specific, rather than universal, even for polymers in a "good solvent". It is suggested that, in that case, retardation of proteins of R > 2 nm be quantitatively described in the form mu/muo = exp[-p(R/xi], where p is a parameter depending on monomer type and/or polymer polydispersity. For PVA, the logarithm of mu/muo was found to be linearly related to c (in line with the prediction that the aqueous buffer is a "poor solvent" for this polymer) and to be near-independent of R.

Rigid spherical particles in the size range of 5-200 nm diameter were subjected to capillary zone electrophoresis (CZE) in semidilute solutions of uncross-linked polyacrylamide of M(r) 5, 7 and 18 x 10(6) (PA-5, -7 and -18, respectively) of varying concentrations up to 1.6% and at field strengths varying from 68 to 270 V/cm. For all particles under study, the experimental Ferguson plots, log(mobility) vs. polymer concentration, permit a linear approximation. Their slope, the retardation coefficient KR = delta log (mobility)/delta (concentration), for particles smaller than 30 nm in diameter increased with particle size in PA-5 and -7 independently of electric field strength and polymer M(r). The KR of particles of 30 nm in diameter or more was found to be independent of particle size at the lowest field strength used but to decrease with it at the higher values of field strength. The decrease was parallel but shifted to higher values of retardation when the polymer M(r) increased from 5 to 7 x 10(6). With a decreasing ratio of average mesh size of the polymer network, zeta, to particle radius, R, the approach to "continuity" of the polymeric medium (zeta/R << 1) with both increasing particle size and polymer concentration does not result in the retardation behavior expected according to the macroscopic (bulk) viscosity of the solution. These experimental observations were hypothetically interpreted in terms of a transition to a retardation mechanism comprising the formation of a polymer depletion layer near the particle surface--polymer solution interface. Peak width exhibited an overall increase with PA-7 concentration for all particles studied. For particles of 30 nm in diameter or less, the increase was steepest when the radius of the particle was approximately commensurate with zeta at a given polymer concentration. For the largest particle, 205 nm in diameter, peak broadening with polymer concentration was found to correlate linearly with peak asymmetry. CZE of the particles in PA-18 solutions revealed abnormal behavior, with both mobility and peak width remaining near-constant up to a concentration of 0.08% and sharply declining at higher concentrations. The decline of relative mobility is the same-for the entire particle size range used, while peak width declines in direct relation to particle size.

An automated gel electrophoresis apparatus, recently available commercially, allows one to follow the band during electrophoresis in real time, and lends itself therefore to an evaluation of bandwidth as a function of migration time (the dispersion coefficient), resolution and band shape. These determinations assume the constancy of band area with migration time and at various gel concentrations. The purpose of the present study was to verify these assumptions. Representative proteins and sodium dodecyl sulfate (SDS)-proteins, either natively fluorescent or fluorescein carboxylate labeled, were found to exhibit band areas which approach constancy as a function of migration time in both agarose and polyacrylamide gel electrophoresis, provided that (i) the protein concentration under the band was low enough to obviate self-quenching of fluorescence; (ii) the separation of the protein of interest from contaminants had progressed sufficiently during the time at which band areas were measured; (iii) the baseline under the peak was sufficiently well defined. However, band areas decrease with increasing gel concentration. Protein peaks exhibited leading and trailing tails. The ratio of the combined tail area to total area appeared to be near-constant at varying migration times. However, that ratio increases with increasing gel concentration. The tail area does not appear to be an artifact of fluorometric detection since it is reproduced upon fluorimetric analysis of the protein eluted from gel slices after electrophoresis. However, it may be due to photochemical destruction under the conditions of repetitive fluorometric peak detection.

Peak width and peak asymmetry of rigid spherical particles in the size range of 3-100 nm radius (R) were measured in capillary zone electrophoresis (CZE), using buffered uncross-linked polyacrylamide of Mr 5.0 X 10(6). Polymer concentration-dependent spreading of peak width and peak asymmetry were found to parallel one another. The parallelism holds whether the particle size is within the "small" (R < 20 nm) or "large" (R > 20 nm) size ranges previously found to differ in the mechanism of particle size dependent retardation of electrophoretic migration (S. P. Radko and A. Chrambach, Electrophoresis 1996, 17, 1094-1102). In application to the "small" particle size range, the parallelism between band width and band asymmetry can be qualitatively interpreted to be consistent with the Giddings-Weiss mechanism (G. H. Weiss et al., Electrophoresis 1996, 17, 1325-1332) of electrophoresis in polymer-containing media which postulates a dependence of band width and band asymmetry on the equilibrium between "stationary" and "mobile" states of the particle.

A homogenate of sea urchin (Lytechinus pictus) eggs rich in exocytotic membrane vesicles (granules) was subjected to analytical and preparative electrophoresis in the commercial automated horizontal gel electrophoresis apparatus (HPGE-1000, LabIntelligence, Belmont, CA) capable of intermittent scanning of the migration path, using buffered solutions of polyvinylpyrrolidone (PVP). The nonfluorescent granules were detected by "fluorescence reduction", i.e., a decrease of fluorescence intensity due to the absorbance and/or light scattering properties of the particle. Granules migrated at linear migration rates in buffers ranging from 0 to 2.5% PVP. Two bands were observed and optimally separated in 1.5% PVP solution. As shown by sodium dodecyl sulfate (SDS)-polypeptide patterns, the material recovered from the bands was qualitatively indistinguishable from the two major fractions A and C of granules previously separated by free-flow electrophoresis in the absence of polymer. Ferguson plot analysis failed to provide the sizes of the granules in view of the narrow PVP concentration range available for mobility measurement and the unavailability of chemically homogeneous size standards.

A human lymphocyte population undergoing apoptosis in vitro due to gamma-irradiation was fractionated by free-flow electrophoresis in triethanolamine--Na-acetate buffers, containing up to 50 mM NaCl, with pH 6.0, 7.2 and 8.5, made isotonic by addition of sucrose. As shown by a flow cytometric analysis of the eluate, the distribution of apoptotic lymphocytes is shifted to the range of higher electrophoretic mobilities relative to that of viable ones at pH 8.5, yielding cell fractions enriched in apoptotic cells by a factor of 3 to 5. The difference in rates of electrophoretic migration observed at a mildly alkaline pH but not at a neutral or mildly acidic one suggests that the surface of apoptotic lymphocytes is more acidic than that of viable ones.

Particle size-dependent retardation ("molecular sieving") in electrophoresis can be achieved in polymer solutions or gels. In the semidilute concentration range of polymer solutions, mobility of "rigid, spherical" particles can be predicted from their size (in the size range less than 20 nm radius), the screening length specific for the particular polymer and a constant that can be experimentally determined for a polymer. That constant could be universal for all hydrophilic uncharged polymers. Band spreading in polymer solutions is constant over a range of particle sizes and polymer concentrations and increases beyond that range. Presumably, the critical division between the two ranges occurs when the diameter of the particle exceeds the screening length of the polymer network. Resolution, defined as separation divided by the sum of bandwidths, can thus be predicted for a limited particle size and polymer concentration range. In gels, resolution can be estimated from the slopes and free mobility intercepts of the Ferguson plot. However, the previous computation of resolution needs to be revised in view of the fact that band spreading in gels proceeds in proportion to time, presumably through interaction with the polymer, and not as a function of the square root of time as would be the case if band spreading resulted from diffusion.

Subjecting particles in the size range of 3 to 1085 nm radius (R) to capillary electrophoresis in buffered solution of entangled uncrosslinked polyacrylamide (M(r) 5 x 10(6)), it was found that particle size-dependent retardation ("molecular sieving") becomes electric field- and particle size range-dependent once the particle size exceeds 15-20 nm in radius. The field strength dependence of the retardation coefficient [KR = d(log mobility)/ d(polymer concentration] and the positive or negative sign of dKR/dR suggest the existence of two different mechanisms of molecular sieving depending on the particle size range: particles with diameters less than the screening length (or blob size) of the polymer network are thought to penetrate into the available spaces within a discontinuous polymer network; particles with diameters larger than the screening length (or blob size) of the polymer network are thought to undergo size-dependent retardation by exerting shear stress against polymer chains, and displacing them, so as to cause local deformations in a continuous polymer network. A limit in the separating capacity of molecular sieving, due to a sharp increase in the rate of band widening with polymer concentration, was found when the value of the retardation coefficient exceeded 60 (mL/g).

Rat liver microsome components were separated by capillary zone electrophoresis in buffers containing substituted agarose, agarose crosslinked polyacrylamide, polyacrylamide, polyethylene glycol and dextran of various molecular weights. The best resolution of the components was obtained with polymers of 10-21 nm geometric mean radius. Both the crude and the purified preparations of microsomes exhibit a single major peak. In a Tris-borate-EDTA (TBE) buffer, containing polyacrylamide of 5 x 10(6) molecular weight, it has a retardation coefficient, KR, of 0.77 +/- 0.02. Translation of the KR value to geometric mean radii, R, on the basis of the standard curve applicable to polymer solutions, KR vs. R, with polystyrene carboxylate size standards in both dextran and polyacrylamide solution allows one to estimate a value of R as 13-16 nm for the major microsome component. The value is smaller than expected from electron microscopic measurements (100-250 nm), possibly due to the chemical and geometric differences between microsomes and the polystyrene particles used as size standards. The crude preparation also contains a minor component which is smaller and less charged than the major component. Another component, apparently very much larger than the major one, is seen in TBE buffer but not in a potassium-N-(2-hydroxyethyl)piperazine-2'-(2-ethanesulfonic acid) buffer and is therefore thought to be an artifact of interaction with borate. After a short incubation under conditions promoting delayed microsome fusion, i.e. in presence of GTP and Mg++ and in the absence of polyethylene glycol, the electrophoretic pattern changes dramatically: it now exhibits five unretarded, highly mobile and, therefore, presumably large components in addition to the two original retarded components of the microsome and a less highly charged species similar in KR to the smaller of the original two components.