Influence of ignored and well-known zone distortions on the separation performance of proteins in capillary free zone electrophoresis with special reference to analysis in polyacrylamide-coated fused silica capillaries in various buffers. II. Experimental studies at acidic pH with on-line enrichment

The separation of acidic and basic model proteins was studied in capillary free zone electrophoresis in a polyacrylamide-coated, electroosmosis-free capillary at pH below their isoelectric points (pI) using various buffers at pH 2.7-4.8 with UV detection at 200 nm. The separation performance was significantly dependent on the coating quality, which may even differ within the same batch of capillaries. In addition, a washing step with 2 M HCl and the storage of the capillary in distilled water was essential for the performance. For high efficiency and resolution the choice of buffer constituents was extremely important which is discussed in quantitative terms in Part I. The most promising buffers were ammonium acetate and ammonium hydroxyacetate at pH 4 (ionic strengths: 0.12 and 0.15 M, respectively) with plate numbers up to 1,700,000 plates/m, corresponding to a zone width (2sigma) of only 1 mm in a capillary with 40 cm effective length, when the injected samples were dissolved in a 10-fold diluted background electrolyte (BGE), a zone even narrower than those obtained in polyacrylamide gel electrophoresis, the characteristic feature of which is remarkably thin zones. In the experiment giving this plate number, the calculated variance for longitudinal diffusion was larger than all the other calculated variances (those for the width of the starting zone, Joule heating, sedimentation and the curvature of the capillary). Interestingly, the effect of capillary curvature was significant. In addition, the sum of all other imaginable variances (corresponding to various types of slow on/off kinetics and hyper-sharp peaks) was in the most successful experiments only 28-50% of the variance for longitudinal diffusion. One hundred- to two hundred-fold dilution of the BGE improved the detection limits and provided high precision in both migration times and peak areas with ammonium hydroxyacetate and ammonium acetate as background electrolytes. However, that high dilution increased the variance 140-400% for these buffers, respectively, at least partly due to conductivity or pH differences between the sample and buffer zones (hyper-sharp peaks). Sedimentation of the enriched sample, a factor that has not previously been treated theoretically or experimentally, was probably another reason for our finding that peak heights did not increase when the sample was dissolved in a buffer diluted more than 200-fold, although pH changes and in some cases thermal expansion in the capillary also may contribute. Loss of protein may occur at the ionic strength 0.01 and lower due to precipitation. Limits of detection were in the range 4-17 pmol of proteins with ammonium acetate as BGE. No indication of denaturation of proteins at pH 4 was observed. However, the separation performance at pH 3 was not satisfactory and loss of proteins was observed, possibly indicating such problems. The protein mobilities decreased unexpectedly from pH 4 to 3--a further indication of conformation changes.