Validation of a Method for Determination of Phospholipase A 2 and Melittin in Bee Venom Preparations by Capillary Electrophoresis

Size: px

Start display at page:

Download "Validation of a Method for Determination of Phospholipase A 2 and Melittin in Bee Venom Preparations by Capillary Electrophoresis"

Cody Osborne
5 years ago
Views:

1 PACÁKOVÁ &S TULÍK:JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, DRUGS, COSMETICS, FORENSIC SCIENCES Validation of a Method for Determination of Phospholipase A 2 and Melittin in Bee Venom Preparations by Capillary Electrophoresis VĔRA PACÁKOVÁ and KAREL S TULÍK Charles University, Department of Analytical Chemistry, Albertov 2030, Prague 2, Czech Republic A method was validated for the determination of the 2 main components of bee venom, phospholipase A 2 and melittin, by capillary electrophesis (CE). Optimum resolution and selectivity were attained with a running electrolyte of 150 mm phosphoric acid, ph 1.8. The repeatability and day-to-day reproducibility of the migration times were better than 0.36 and 2.8%, respectively. The repeatability and day-to-day reproducibility of the normalized peak areas were better than 1.3 and 2.6%, respectively. The response of the UV detector at 190 nm was linear over < 2 concentration decades, from 0.05 to 1.5 mg/ml, with correlation coefficients of for phospholipase A 2 and for melittin. The limits of detection and quantitation were 4.5 and 15 µg/ml, respectively, for phospholipase A 2 and 1.6 and 6 µg/ml, respectively, for melittin. The reproducibility of the measurements with 2 different CE instruments was satisfactory; the mean concentration and relative standard deviation (RSD) values for phospholipase A 2 and melittin were 14.4% (RSD, 1.3%) and 51.4% (RSD, 1.1%), respectively, with instrument I; the corresponding values with instrument II were 14.5% (RSD, 2.8%) and 52.3% (RSD, 2.2%). The accuracy was estimated by comparison with a liquid chromatographic (LC) method. Differences between the CE and LC measurements were attributed to irreversible adsorption of the analytes on the LC column. The recoveries of phospholipase A 2 and melittin with the CE method were 98.8 and 101.7%, respectively. Received February 9, Accepted by JM October 15, Bee venom is one of the most important causes of allergic reactions in humans, and thus methods for the isolation of the venom components, their characterization, and their determination are needed for both diagnostic and therapeutic purposes. The composition of bee venom is known (1, 2). The principal allergen is the enzyme phospholipase A 2 (with a concentration of about 15% in dry venom). Another allergen is hyaluronidase (with a concentration of about 1 2%). The most abundant component is the polypeptide melittin (with a concentration of about 50%), which is poisonous and, therefore, must often be removed from allergenic preparations. A great variety of methods have been described for the characterization of bee venom. Biological effects of the whole venom or of its individual components, or typical reactions of proteins, peptides, and glycoproteins can be used for this purpose. For example, bee venom exhibits bacteriostatic and bactericidic properties (3, 4) and a weak hemolytic effect (5). Other properties that can be used to characterize bee venom involve the activity of hyaluronidase (6) and the formation of lysolecithin by phospholipase A 2 (7) that enhances hemolysis. Phospholipase A 2 also prolongs the coagulation of egg yolk (8) and enzymatically cleaves phospholipids with formation of carbon dioxide (9). Bee venom reacts as a typical protein. Many approaches have been taken to separate the components of bee venom, primarily liquid chromatography (LC) (10 16) and electrophoresis (16, 17). Size-exclusion LC is very well suited for separations of the proteins and polypeptides contained in bee venom. For example, bee venom components have been separated efficiently on a Separon HEMA-BIO 40 column with a mobile phase containing a denaturing agent, 0.2% trifluoroacetic acid (TFA) in 20% acetonitrile, which prevents nonspecific interactions of some components, e.g., melittin (16). However, it is difficult to separate compounds with similar molecular weights by using size-exclusion LC, but good separations can be obtained with reversed-phase LC (12, 14 16). UV photometric detection is satisfactory at a wavelength of 230 nm or lower, because diode array detection has demonstrated that bee venom components exhibit similar UV spectra (15). A strong interaction of basic melittin with silicabased materials has been observed, but very good separation and quantitation have been attained on a polymer-based reversed-phase HEMA-BIO 1000 C 18 column with a mobile phase of 0.25% aqueous TFA and an acetonitrile gradient (16). Capillary electrophoresis (CE) seems to be the most promising technique at present, because it permits highly ef-

2 550 PACÁKOVÁ &S TULÍK: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, 2000 ficient separations with short analysis times, small quantities of samples and chemicals, and thus low cost of analysis. Sample volumes as low as a few nanoliters allow multiple repetitions of the analyses and the use of expensive certified standards, which is especially beneficial in allergen analysis (18). When automated sample introduction is used, the peak areas are highly reproducible with a relative standard deviation (RSD) of <2% (16). The goal of the present study was validation of a method for analysis of bee venom by capillary zone electrophoresis. The method is based on a procedure published earlier (16) that has been used for more than 3 years in the pharmaceutical company SEVAC (Prague, Czech Republic) for production control of bee venom allergen preparations, for determination of allergen concentrations in the products, and in allergen stability tests. Experimental Allergen Samples and Their Preparation SEVAC kindly provided samples of raw bee venom and its allergic fraction, as well as all the allergen preparations. Melittin had been removed from the raw bee venom by gel chromatography. A solution of raw bee venom in 0.05M acetate buffer, ph 4.75 (1 mg/ml) was applied to a Sephadex G-50 (or G-75 or G-100) column (475 ml, cm id). The column was eluted with the same buffer at a flow rate of 0.4 ml/min. After elution of a half column volume, fractions were collected and their UV absorbance at 280 nm was measured. Protein fractions, containing phospholipase A 2 and hyaluronidase and eluting before the main peak corresponding to melittin, were collected and lyophilized. Enzyme activities were measured by comparing the samples with standard enzymes of known activities. Phospholipase A 2 activity was evaluated from the rate of production of lysolecithins and fatty acids, resulting from its reaction with egg yolk lecithins. The fatty acids were neutralized by NaOH, and the ph decrease over time was measured, because the enzyme activity was inversely proportional to the reaction time. Hyaluronidase activity was monitored after the cleavage of hyaluronic acid. Hyaluronidase diffused to agarose with the substrate, hyaluronic acid, which decomposed. The uncleaved hyaluronic acid was then precipitated with cetyl pyridine. The size of the rings produced at the site of the enzyme-substrate reaction was proportional to the enzyme activity. Standard substances of phospholipase A 2 and melittin were obtained from Sigma Chemical Co. (St. Louis, MO). The other chemicals were orthophosphoric acid and sodium hydroxide from Lachema (Brno, Czech Republic) and trifluoroacetic acid, Tris buffer, and sodium dodecylsulfate (SDS) provided by Romil Chemicals (Cambridge, UK). All chemicals were analytical grade and were used as received. Deionized water (Milli Q, Millipore Corp., Bedford, MA) was used. Standard solutions were prepared in deionized water at a concentration of 1 or 2 mg/ml and were appropriately diluted before use. Instrumentation The CE measurements were performed with an ATI Unicam instrument (Cambridge, UK) and a Unicam 4225 variable-wavelength UV photometric detector (system I), or with a P/AGE System 210 instrument (Beckman Instruments, Fullerton, CA) and a UV photometric detector (system II). Various fused-silica capillaries were used in system I, with total length (L T ), cm; distance to detector (L D ), cm; and inner diameter, 50 or 75 µm; at a voltage of 15, 20, or 30 kv. System II used a 57 cm 50 µm id capillary, with an L D of 50 cm, at a voltage of 15 kv. The measurements were performed at 25 C in both systems. The samples were introduced hydrodynamically in system I ( mbar for min, depending on analyte concentration), or electrokinetically in system II (5 kv for 10 s). The fully automated injection in system I is based on Boyle s law, which accurately predicts the pressure increase when a known volume of gas is compressed to a smaller known volume. A broad pressure range can be generated in this way and applied to sample injection or capillary washing. UV photometric detection was performed at 190 nm (system I) and at 214 nm (system II). Rinsing Procedure The following procedure was used for rinsing the CE capillaries: flushing with 0.1M NaOH for 5 min, then with deionized water for 10 min, and finally with running buffer solution for 30 min, always at 1000 mbar. This procedure was performed each morning before measurements were started and was repeated after every 20 sample injections. After each analysis, the capillary was washed with running buffer for 2 min at a pressure of 1000 mbar. Results and Discussion Method Development To characterize an allergenic preparation, it is most convenient to measure the concentration of a specific allergen (19, 20). This can be done with bee venom because the standard substances, phospholipase A 2 and melittin, isolated from bee venom, are commercially available. Three electrolyte systems were tested by Pacáková et al. (16): (1) an electrolyte whose ph was close to neutral solution, 20 mm sodium phosphate, ph 5.0; (2) a basic electrolyte containing an anionic surfactant as a pseudostationary phase, 20 mm Tris- 50 mm SDS, ph 9.0; and (3) and an acidic electrolyte, 150 mm phosphoric acid, ph 1.8. In electrolytes (1) and (2), basic proteins and peptides adsorbed on the capillary wall, and the migration times became gradually longer and the peaks became broader. The simplest electropherogram was obtained in the acidic electrolyte system, (3), in which the electroosmotic flow and analyte adsorption on the capillary walls were suppressed. However, the analysis time remained short because highly positively charged analytes migrated more rapidly to the cathode. An example of the separation and identi-

3 PACÁKOVÁ &S TULÍK:JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, fication of bee venom components in this electrolyte is given in Figure 1. Figure 1A depicts the separation of the raw bee venom components, including melittin, whose content was about 51.5%. Electropherograms of the phospholipase A 2 and melittin standard substances are shown in Figure 1, B and C, respectively. The electropherogram of a bee venom allergen preparation is shown in Figure 1D. Figure 1B shows that the phospholipase A 2 produced 2 main peaks with migration times of and min. The area of the second peak was used for quantitation. The peak area ratio of these 2 components is about 1:10, which corresponds to the concentrations of hyaluronidase and phospholipase A 2 in bee venom. The hyaluronidase and phospholipase A 2 were separated from the other components of bee venom by gel chromatography, and their absorbances were measured. The ratio of their molar absorbances was The presence of hyaluronidase in the standard phospholipase A 2 from Sigma Chemical Co. was confirmed by measurement of its enzymatic activity. The hyaluronidase standard substance, isolated from bee venom, is not commercially available. The efficiency of the separation was evaluated. The number of theoretical plates for melittin, calculated under the experimental conditions given in Figure 1, was , and the resolution of phospholipase A 2 and melittin was 1.3. Validation of the Method The recommendations (21, 22) for validation of the CE method were followed in this work. The type of CE instrument used in this work was evaluated by Kunkel et al. (22). Figure 1. Separation and identification of bee venom components in 150 mm phosphoric acid, ph 1.8. (A) raw bee venom (1.12 mg/ml), (B) phospholipase A 2 standard substance (1 mg/ml), (C) melittin standard substance (0.5 mg/ml), and (D) bee venom allergen preparation (0.3 mg/ml). Experimental conditions: System I; concentration of bee venom, 1.12 mg/ml; L T, 75 cm; L D, 60 cm; id, 50 µm; voltage, 20 kv; injection, 30 mbar, 6 s; and UV photometric detection at 200 nm.

4 552 PACÁKOVÁ &S TULÍK: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, 2000 Table 1. Repeatability a and day-to-day reproducibility b of migration times (t m ) and normalized peak areas (A n ) for phospholipase A 2 (P2) and melittin (M) c Type of measurement Component Concn, mg/ml t m, min RSD, % A n RSD, % Repeatability P M Day-to-day reproducibility P M a Six measurements. b Sixty measurements in 1 week. c Experimental conditions: System I; 150 mm phosphoric acid, ph 1.8; L T, 75 cm; L D, 60 cm; id, 50 µm; voltage, 20 kv; injection, 30 mbar, 0.1 s; and UV photometric detection at 190 nm. Selectivity There were no interferences from the matrix components present in the allergen preparations, such as the physiological solution (0.9% aqueous NaCl) and mannitol, a stabilizer added in excess (1:200) to bee venom preparations. Repeatability and Day-to-Day Reproducibility The repeatability of the migration times and peak areas of the main components of bee venom was determined first. Solutions of phospholipase A 2 and melittin standard substances at concentrations of 0.5 and 1.0 mg/ml were analyzed 6 times in electrolyte (3) within 2 h. The average migration times and normalized peak areas (peak areas divided by migration times), together with the standard deviations (SDs) and RSDs, were calculated. The same procedure was repeated in the course of 1 week (60 injections) to evaluate the day-to-day reproducibility. The results are summarized in Table 1, which shows that the repeatability and day-to-day reproducibility can be considered satisfactory. Compared with our previous measurement (16), the repeatability of migration times improved in this work. The previously published RSDs for phospholipase A 2 and melittin, 6.5 and 5.6%, respectively, decreased in this validation to 0.37 and 0.35%, respectively. This decrease can be attributed to the inclusion of the rinsing procedure in the CE method. The day-to-day reproducibility of the migration times attained values of 3.1% for phospholipase A 2 and 2.8% for melittin. The improvement in the peak area reproducibility apparently stems from normalization of the peak areas. Detection Linearity The maximum sensitivity of the UV detection was obtained at 190 nm (the UV response was 1.4 times higher than that at 200 nm). The degree of detection linearity was found from the dependence of the normalized peak areas, A n (A n is the normalized area in arbitrary units/min), on the concentration, c, of the phospholipase A 2 and melittin standard substances (in mg/ml). The results are given in Table 2. The parameters of the calibration dependence, A n =a+b c, were calculated, and the following values were obtained: a = (SD = 8.969), b = (SD = ), and correlation coefficient = for phospholipase A 2 ;a= (SD = 6.498), b = (SD = ), and correlation coefficient = for melittin. As pointed out by a reviewer of this paper, the plot of the response corrected for the intercept and divided by the concentration vs the concentration indicated that the concentration dependence of the response was not perfectly linear; however, the deviations from linearity were only significant for the lowest concentrations, at which the error attained a value of about 10%. The rather limited linear range of <2 concentration decades stems from the necessity of injecting small samples ( 10 nl), because the peaks become distorted Table 2. Dependence of normalized peak areas (A n ) for phospholipase A 2 and melittin on concentration a Concn, µg/ml Phospholipase A 2 A n Melittin RSD, RSD, % A n % a Experimental conditions: Same as in Table 1.

5 PACÁKOVÁ &S TULÍK:JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, Table 3. Determination a of the allergenic component, phospholipase A 2, in an allergen preparation with a declared content of 500 µg/ampoule Sample Phospholipase A 2 found, µg Mean, µg RSD, % a Experimental conditions: Same as for Table 1. Each sample (the contents of 1 ampoule) was dissolved in 500 µl deionized water. with larger sample volumes. The limit of detection (LOD) and the limit of quantitation (LOQ) were determined as the concentrations at which peak height corresponded to 3 and 10 times the SD of the peak-to-peak noise value, respectively. The values obtained for phospholipase A 2 and melittin were 4.5 and 1.6 µg/ml (LOD) and 15 and 6 µg/ml (LOQ), respectively. An example of the analysis of an allergen preparation with a declared amount of 500 µg allergens/ampoule, to which 20 mg mannitol was added, is given in Table 3. Three samples were analyzed. The contents of each ampoule were dissolved in 500 µl water and analyzed in system I. The contents of each ampoule were analyzed 6 times, and the above calibration plot was used for the calculations. Reproducibility The reproducibility of the procedure was obtained from repeated analyses of raw bee venom preparations. Six samples were analyzed in 2 different instrumental systems, I and II, with different capillary lengths, different sample introduction modes, and detection at different wavelengths. The samples were dissolved in deionized water at a concentration of 1 mg/ml, and the phospholipase A 2 and melittin standard substances were used with the absolute calibration method as described above. The results are given in Table 4, which shows that the procedure exhibits acceptable reproducibility. The higher RSDs for system II were probably caused by the fact that instrument II was older and consequently less sophisticated. Accuracy The accuracy of the procedure was ascertained from a comparison with the determination by a different method. In our previous paper (16), we reported the results of an LC determination, i.e., 17.1% for phospholipase A 2 and 42.9% for melittin, which correspond to the CE measurement values of 15.8% for phospholipase A 2 and 47.3% for melittin. The somewhat lower LC value for melittin can be explained by the irreversible adsorption of melittin on the stationary phase used. Recovery Standard solution volumes of 100 µl for phospholipase A 2 and 300 µl for melittin, with an analyte concentration of 1 mg/ml, were added to 600 µl placebo (composition: 0.9% NaCl, 0.02% human serum albumin, and 20 mg mannitol) and analyzed 6 times in system I. The amounts found were 98.5 µg phospholipase A 2 and 305 µg melittin, which correspond to recoveries of 98.8% and 101.7%, respectively. Ruggedness The ruggedness of the procedure, i.e., its reproducibility obtained in various laboratories with different instruments and chemicals, could not be fully tested, because the method is new and so far has been used only in our laboratory in analyses for SEVAC. Nevertheless, some idea about its ruggedness can be obtained from the reproducibility measurements (Table 4),because the differences in the instrumentation and the experimental conditions (systems I and II) are large. On this basis, it can be assumed that the procedure is sufficiently rugged to be used routinely for the intended purpose. Stability Testing The electrolyte used, 150 mm phosphoric acid, ph 1.8, was found to be stable for >3 years, if kept in a refrigerator at 5 C. The standard substances of melittin and phospholipase Table 4. Reproducibility of the determination of bee venom components, phospholipase A 2 (P2) and melittin (M), in systems I a and II b Sample System I System II P2, % M, % P2, % M, % Mean, % RSD, % a System I: same experimental conditions as for Table 1. b System II: L T, 75 cm; L D, 57 cm; id, 50 µm; voltage, 15 kv; injection electrokinetically, 5 kv for 10 s; UV photometric detection at 214 nm; concentration of bee venom, 1.4 mg/ml.

6 554 PACÁKOVÁ &S TULÍK: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 3, 2000 A 2, when kept frozen, were stable for >4 years (their activity did not decrease). When they were dissolved in deionized water and kept in a freezer between analyses, solution stability was somewhat poorer, about 1 year. The same conclusions can be drawn from the bee venom allergen measurements. Lyophilized allergen preparations were stable for >4 years if kept at temperatures 20 C, and for about 1 year when stored at 37 C. However, dissolved samples kept at room temperature decomposed within 2 weeks; decreased enzyme activity and a loss of resolution in CE were observed. The analytical parameters obtained demonstrate that the method is readily applicable to monitoring the composition of bee venom. The procedure is routinely applied to the production control of bee venom allergen preparations and to their stability tests. The method has been found to be reliable, fast, and inexpensive, because the consumption of chemicals and standards is very small. The CE method did not exhibit any deterioration in performance during analysis of >100 samples, and replacement of the separation capillary was not necessary. Acknowledgments Part of this work was supported by grants UK 200/1996 and FRVS 139/1999. References (1) Müller, U.R. (1988) Insektenstichallergie, Klinik, Giagnostik und Therapie, Gustav Fischer Verlag, Stuttgart, Germany, pp (2) Habermann, E. (1972) Science 177, (3) Ortel, S., & Markwardt, F. (1955) Pharmazie 10, (4) Benton, A.W., Morse, R.A., & Kosikowski, F.V. (1963) Nature (London) 168, (5) Neumann, W., & Habermann, E. (1954) Arch. Exp. Pathol. Pharmakol. 222, (6) Habermann, E. (1957) Biochem. Z. 329,1 10 (7) Habermann, E., & Neumann, W. (1957) Biochem. Z. 328, (8) Habermann, E., & Hard, K.L. (1972) Anal. Biochem. 50, (9) Habermann, E. (1957) Biochem. Z. 328, (10) Shepherd, G.W., Elliott, W.B., & Arbesman, C.E. (1974) Prep. Biochem. 4,71 78 (11) Renck, B., & Einarsson, R. (1980) J. Chromatogr. 197, (12) Banks, B.E.C., Dempsey, C.E., Pearce, F.L., Vernon, C.A., & Wholley, T.E. (1981) Anal. Biochem. 116,48 57 (13) Einarsson, R. (1983) Acta Chem. Scand. Ser. B 37, (14) Einarsson, R., & Renck, B. (1984) Toxicon 22, (15) Räder, K., Wildfeuer, A., Wintersberger, F., Bossinger, P., & Mücke, H.W. (1987) J. Chromatogr. 408, (16) Pacáková, V., S tulík, K., Hau, P.T., Sýkora, B., & Vinš,I. (1995) J. Chromatogr. 700, (17) Einarsson, R., & Moberg, U. (1981) J. Chromatogr. 209, (18) Pacáková, V., S tulík. K., & Tichá, M. (1997) J. Chromatogr. B 699, (19) Yunginger, J.W. (1991) An. Allergy 66, (20) Platts-Mills, T.A.E., & Chapman, M.D. (1991) Allergy Clin. Immunol. 87, (21) Atria, K.D., & Rudd, D.R. (1995) Chromatographia 41, (22) Kunkel, A., Degenhart, M., Schirm, B., & Waetzig, H. (1997) J. Chromatogr. A 768,17 27

Journal of Chemical and Pharmaceutical Research

Available on line www.jocpr.com Journal of Chemical and Pharmaceutical Research ISSN No: 0975-7384 CODEN(USA): JCPRC5 J. Chem. Pharm. Res., 2011, 3(1):138-144 Simultaneous RP HPLC determination of Latanoprost