A method of capillary electrophoresis frontal analysis (CEFA) is developed for the first time to study the binding of ketoprofen to human serum albumin (HSA) and compared with high performance liquid chromatography frontal analysis (LCFA). The separation is performed in an uncoated fused-silica capillary (60-cm × 75-μm i.d., 50-cm effective length) with a phosphate buffer (pH 7.4, ionic strength of 0.17M) as the running buffer. The applied voltage is 13 kV and the detection is set at 254 nm. A trapezoidal peak of the unbound ketoprofen appears after HSA elution in the electropherogram. The plateau height of the peak is employed to determine the unbound concentration of ketoprofen in the HSA equilibrated sample solution. The CEFA method provides the advantages of small sample injection volume and rapidity and the disadvantage of low sensitivity compared with LCFA. CEFA is applicable to the binding parameter estimation of ketoprofen to the secondary binding site; an association c:onstant (K2) of 0.24 × 106M-1 and the number for the binding site per molecule HSA of 2.54 is estimated. In contrast, LCFA measures parameters for both primary and secondary sites, which are 1.05 ×106M-l and 0.94 for K1 and n1, respectively, and 0.12 × 106M-1 and 3.16 for K2 and n2, respectively. It is found that ketoprofen binds mainly at the primary site at a molecular ratio of ketoprofen versus HSA lower than 0.75, and the binding at the secondary site occurs at a higher ratio.

The EOF pump was successfully used as a means of introducing samples into a capillary system. An improved sample injection device has been developed using a TeflonTM union(TU) to link the two capillaries together. The capillary applied high voltage served as the EOF pump to pull the liquid inside while the other one served the purposes of isolating the electric field. Using the bias degree(BD) and SD of bias(SDB), it was possible to quantitatively determine the sampling bias and evaluate the effectiveness of injection approaches in bias elimination. Several related factors were evaluated, and it was found that TU approach could fully eliminate the bias under the optimal conditions. The fracture did not affect the efficiency, leak, or dilute the sample significantly. This approach was effective under both normal and reverse EOF situations and adapted to real samples. Finally, a TU method using grounded injection electrode was proposed and shown to be suitable for samples with low conductivity and high injection voltage.

We have used capillary electrophoresis in the frontal analysis mode(CE- FA) to determine the unbound concentration of clozapine in human serum albumin(HSA), human plasma, rabbit serum and plasma sample. The unbound clozapine concentration was directly measured from the height of the frontal peak. Samples were injected directly into an uncoated fused silica capillary (0.65 m (LC)×75 μm i. d.; LE= 0.35 m) and separation was accomplished within 11 min without extensive sample pretreatment. The most suitable running buffer to separate unbound clozapine peak from the other peaks due to endogenous substances was found to contain 1 mmol 1-1 EDTA, 0.5 mol 1-1 glycine, and 67 mmol 1-1 phosphate with pH 7.4. The concentrations of unbound clozapine agreed well with those determined by the conventional ultrafiltration method. The methodology is validated and good correlation and precision are obtained. It was found that clozapine is strongly bound to protein in human plasma, rabbit plasma, and serum, while hardly bound to HSA. The present method enables the determination of the unbound drug concentration in multiple equilibrium system with less than ultra- micro injection volume, and would be hence especially useful for the protein binding study in biological samples that are only available in minute quantities.

目的 研究格列美脲与人血清白蛋白（human serum albumin, HSA）结合作用。方法 采用高效液相色谱—迎头分析法（high performance liquid chromatography- frontal analysis, HPLC- FA）。色谱柱为内表面反相（internal- surface reversed- phase, ISRP）硅胶柱Pinkerton GFF Ⅱ- S5- 80(150 mmx 4.6 mm ID, 5 μm)；流动相为67 mmol·L-1磷酸盐等渗缓冲液(pH 7.4, I= 0.17 mol·L-1)；流速0.2 mL·min-1；检测波长230 nm；进样量900 μL；柱温37℃。结果应用非线性回归参数估算求得HSA分子上两类结合位点结合的格列美脲分子数及相应的结合平衡常数：n1= 1和K1= 5.1(μmol·L-1)-1；n2= 7和K2= 0.017(μmol·L-1)-1。结论 HPLC- FA法操作简便，适用于研究格列美脲与HSA的结合作用。

采用毛细管电泳—迎头分析模式体外实验测定了人血清白蛋白溶液、人血浆、兔血清和血浆样品溶液中游离氯氮平的浓度。根据前沿峰的峰高直接测定了样品溶液中的游离氯氮平浓度，并与传统超滤法进行了比较。通过考察施加的电压和运行缓冲液的组成对氯氮平电泳峰平台形成的影响确定了最优分离条件。讨论了氯氮平蛋白结合作用的选择性。