A method for the simultaneous separation and direct determination
of oxalic acid (OA), tartaric acid (TA), malic acid (MA), vitamin C (VC), citric acid (CA), and succinic acid (SA) in Fructus mume using reversed-phase high-performance liquid chromatography with a UV detector in an acidic medium is presented in this study. In the experiment, the optimization of chromatographic conditions (i.e., the pH and flow rate of the mobile phase, the absorption wavelength, and temperature of column) that affect the separation degree and peak shape of organic acids has been obtained. The linear ranges are found to be 0.054.7 g for OA (r = 0.9999), 0.1110.5 g for TA (r = 0.9999), 0.11411.4 g for MA (r = 0.9999), 0.0333.30 g for VC (r = 0.9999), 0.15515.5 g for CA (r = 0.9998), and 0.19419.4 g for SA (r = 0.9996). For OA, TA, MA, VC, CA, and SA, the even recovery (n = 3) of six effective components are 100.9%, 99.97%, 101.2%, 102.1%, 101.1%, and 100.7%, respectively, and the largest relative standard deviation (n = 11) for the six components is less than 1.7%. The detection limits are 0.01 g for OA, TA, and VC; 0.05 g for MA; 0.03 g for CA; and 0.1 g for SA. In a single chromatographic run, OA, TA, MA, VC, CA, and SA can be determined in less than 7 min. The method can be used for the purpose of routine analysis and the quality control of a botanic (Fructus mume) containing these effective components. Introduction Fructus mume is dry and nearly a mature fruit of Prunus mume (Sieb.) Sieb. et Zucc. It is picked in summer when it has nearly grown to maturity. As a medico-material, it must become black by putting it in a container that must be separated from air after it has dried in a lower temperature (1). It is a useful medico- botanic that is recorded in the Chinese pharmacopoeia. The Chinese book, Bencao Gangmu, has recorded it as a drug that can cure diarrhea (2). Previously, scientists considered that there were small amounts of tartaric acid (TA), malic acid (MA), and succinic acid (SA) in Fructus mume besides the large amounts of citric acid (CA). Presently, results of modern pharmacology research indicate that the antibacterial, antifiber, and round- worm-killing actions of Fructus mume are related to its acid con- stituents such as TA, MA, CA, and SA (3,4). A number of methods have been reported for the determination of organic acids in a solution by fluorescence, chemilumines- cence (58), potential titration (9), high-performance liquid chro- matography (HPLC) (10), capillary isotachophoresis (11), capillary zone electrophoresis with electrochemical detection (12), and thin-layer chromatography (13). Many of these were for the determination of single organic acids (i.e., chemilumines- cence), and some of them were for the determination of total or mixed organic acids (i.e., potential titration and fluorescence). The normal-phase HPLC method coupled with an electrochem- ical detector offers advantages of high resolution with low detec- tion limits. It has therefore been used as a very sensitive method for the determination of a few organic acids in solution in recent years. Although octodecylsilyl (ODS) columns (normal phase) have been frequently used for the separation of the different organic acids, an important limitation factor with such columns is that an organic solvent is required, which often leads to com- plicated electrochemical behaviors when using an electrochem- ical detector. Besides this limitation, an electrochemical detector is not so popular in general analytical laboratories because it is too expensive and responds only to a limited number of compo- nents. In this work, a new method was developed for simultaneous and direct determination above multiple organic acids in the botanic Fructus mume by reversed-phase (RP) HPLC with UV detection. This method offers four advantages. First, a commer- cial HPLC instrument can be directly applied to the simultaneous Abstract Simultaneous and Direct Determination of Oxalic Acid, Tartaric Acid, Malic Acid, Vitamin C, Citric Acid, and Succinic Acid in Fructus mume by Reversed-Phase High-Performance Liquid Chromatography Chen Zhanguo* and Lu Jiuru Department of Chemistry, School of Chemistry and Material Science, Shaanxi Normal University, Xian, 710062, P.R. China Reproduction (photocopying) of editorial content of this journal is prohibited without publishers permission. Journal of Chromatographic Science, Vol. 40, January 2002 * Author to whom correspondence should be addressed. Journal of Chromatographic Science, Vol. 40, January 2002 36 direct determination of the six important biological components (six organic acids) in Fructus mume without any modification. Second, an RP C18 column is applied for the separation of these organic acids, which allows an aqueous solution to be used as the mobile phase. Third, an expensive electrochemical detector becomes unnecessary. Fourth, extracted solutions of Fructus mume do not need to be treated with any special chemical pro- cess before analysis. With this method, six important organic acids (OA, TA, MA, VC, CA, and SA) can be efficiently separated using an RP column with a mobile phase of a dilute aqueous solu- tion of (NH 4 )H 2 PO 4 and can be successfully determined with a UV detector. No simultaneous and direct determination of OA, TA, MA, VC, CA, and SA in a Fructus mume study has been previously reported. Experiment Instrument A Shimadzu (Kyoto, Japan) LC-6A HPLC system, SPD-6AV detector, CR-3A chromatography data-processing apparatus, and a Shimadzu UV-260 UV spectrophotometer were used. Reagents and material OA (C 2 H 2 O 4 ), TA (C 4 H 6 O 6 ), VC (C 6 H 8 O 6 ), CA (C 6 H 9 O 7 ), SA (C 4 H 6 O 4 ), (NH 4 )H 2 PO 4 , and H 3 PO 4 reagents were purchased from TianJin Chemical Reagent Company (TianJin, China). L-MA (C 4 H 6 O 4 ) was obtained from the Test and Research Institute of Medicines and Chemical Reagents and Living Things Products (Beijing, P.R. China). Fructus mume was yielded in Fujian and Sichua of China and appraised by Prof. Quanhong Liu, who works at the College of Life Science in Shaanxi Normal University (Xian, P.R. China). All chemicals were of analytical-reagent grade or higher and were used without further purification. All solid regents had been dried until constant weight before applying. The standard stock solutions were prepared by dissolving 0.0117 g of OA, 0.1645 g of TA, 0.0317 g of MA, 0.0413 g of VC, 0.0646 g of CA, and 0.0486 g of SA in a small amount of distilled water, respectively, and these solutions were then transferred into six volumetric bottles (50.00 mL). Some of the volume mobile phase was added into them until the mark was reached. The mixed standard stock solution was prepared by transferring the standard stock solution of 5.00 mL OA, 0.80 mL TA, 4.50 mL MA, 1.00 mL VC, 3.00 mL CA, and 5.00 mL SA into the same volu- metric bottle (25.00 mL). Some of the volume mobile phase was added into it until the mark was reached. The concentrations of each constituent in this mixed standard stock solution were 46.58 g/mL for OA, 105.28 g/mL for TA, 114.00 g/mL for MA, 33.04 g/mL for VC, 155.04 g/mL for CA, and 194.40 g/mL for SA. The mobile phase of 0.5% (NH 4 )H 2 PO 4 (w/v) was prepared by dissolving solid (NH 4 )H 2 PO 4 in distilled water. Then, it was adjusted to pH 2.80 with a solution of 1.0-mol/L H 3 PO 4 , and it was filtered under vacuum through a 0.45-m membrane (Shanghai, China) when used. Experiment method The chromatographic column was an LC-18 column (150- 4.6-mm i.d, 5-m particle size). The mobile phase was a 0.5% (NH 4 )H 2 PO 4 (w/v) aqueous solution, and its pH was 2.80. The optimum flow rate of the mobile phase was set at 1.0 mL/min, and the absorbance units full scale was 0.16. The detected wavelength was at 214 nm. The extraction solution of the sample Fructus mume was filtered under vacuum through the 0.45-m membrane. The injection volume of the sample solution was 5.0 L. Organic acid peaks were identified by comparing the retention time in the sample solution with that of the standard solution. The contents of OA, TA, MA, VC, CA, and SA in the sample solution were quantitated by comparing peak areas in the sample solution with that of known standards. Results and Discussion Optimization of chromatographic conditions The key work of this study was to achieve a sep- aration of the six organic acids by using RP-HPLC with an aqueous solution as the mobile phase. In order to achieve baseline separation, a series of experiments were performed. Optimization of mobile phase Because the effective constituents of Fructus mume are organic acid species, a mobile phase of Figure 2. Effect of mobile phase pH on the separation degree: (1) between OA and TA, (2) between TA and MA, (3) between MA and VC, (4) between VC and CA, and (5) between CA and SA. Figure 1. Effect of mobile phase concentration on the retention time of six constituents: OA, 1; TA, 2; MA, 3; VC, 4; CA, 5; and SA, 6. Journal of Chromatographic Science, Vol. 40, January 2002 37 0.5% KH 2 PO 4 (w/v, pH 3.0) was used for the separation of TA and CA in the preliminary work. This condition was recommended by reference 14 for the separation of chlorogenic acid, caffeic acid, and isochlorogenic acid in plants. Under this condition, however, the six organic acids were not completely separated, especially OA, TA, and MA. A series of new mobile phases were then selected in order to obtain a better separation. When the solution of 0.5% KH 2 PO 4 was replaced by an aqueous solution of (NH 4 )H 2 PO 4 , the six organic acids were completely separated, especially OA, TA, and MA. In the experiment, by keeping the pH of the mobile phase at 2.8, the retention time of the six organic acids changed less signif- icantly when the aqueous concentration of (NH 4 )H 2 PO 4 changed from 0.1%, 0.25%, 0.5%, 0.75%, to 1.0% (shown in Figure 1). For convenience purposes in the experiment, we chose the concentra- tion of the mobile phase to be 0.5% (w/v). Therefore, 0.5% of (NH 4 )H 2 PO 4 was applied to the subsequent study. The effect of the pH value of the mobile phase on the retention time of the separated constituents was also examined. In this analysis, we observed that when other conditions did not change, it could greatly affect the separation degree of the six organic acids (taken from the separation degree between two peaks calcu- lated by reference 15) and thus cause the pH value of the mobile phase to change in small ranges. The smaller the pH value was in the mobile phase, then the smaller would be the separation degree. The greater the pH value was, then the greater the sepa- ration degree would be. However, when the pH value of the mobile phase was more than 3, the separation degrees between OA and TA, TA and MA, MA and VC, VC and CA, and CA and SA were reversely smaller. Figure 2 shows the effect of the mobile phase on the separation degree when pH was changed from 2.0, 2.4, 2.6, 2.8, 3.0, to 3.2. This situation was observed in the exper- iment. Although the reason was not completely clear to us, we considered that because the constituents were multiple carboxyl organic acids, when the pH of the mobile phase becomes large, the part of the carboxyl group in the molecule can form carboxy- late. The separated molecular polarity becomes stronger in this condition. The stronger the molecular polarity is, then the smaller would be the force of the molecule that is adsorbed on the column. Therefore, the retention time of the separated molecule would also be smaller, which is a possible reason when the pH of the mobile phase was more than 3 and the separation degrees between OA and TA, TA and MA, MA and VC, VC and CA, and CA and SA were reversely smaller. The experiment indicated that the suitable pH value was 2.8. Therefore, the pH value of 2.8 was applied to the subsequent study. Effect of wavelength Upon researching six organic acids by RP-HPLC, we found that the choice of wavelength was also an important factor. The stan- dard stock solution of six organic acids was investigated on the UV-260 UV spectrophotometer. The wavelength ranges for them were from 180 to 240 nm, but the maximum absorbed wave- length was approximately 190 nm (shown in Figure 3). However, the study met a problem of complexity. If the wavelength in ana- lyzing the sample was chosen in a lower range (i.e., 190 nm), the chromatogram would become more complex and the separation efficiency would not be satisfactory. This phenomenon was observed in this experiment. After many comparisons and consid- ering other factors that would affect the experiment, the suitable absorbed wavelength was finally chosen to be 214 nm. At this wavelength, not only did the peaks of the six organic acids in the extracted solution of Fructus mume appear, but also many of the other peaks disappeared. This was an important factor to ensure accurate determination of the six organic acids in Fructus mume. Linearity and detection limits A mixed standard stock solution was prepared as described Figure 4. RP-HPLC chromatogram of a standard stock solution: OA, 1; TA, 2; MA, 3; VC, 4; CA, 5; and SA, 6. Figure 3. UV spectrum of six chemical constituents: OA, 1; TA, 2; MA, 3; VC, 4; CA, 5; and SA, 6. Journal of Chromatographic Science, Vol. 40, January 2002 38 previously. In this mixed standard stock solution, the concentra- tion of the six constituents were 46.58 g/mL for OA, 105.28 g/mL for TA, 114.00 g/mL for MA, 33.04 g/mL for VC, 155.04 g/mL for CA, and 194.40 g/mL for SA. The injection volume range varied from 0.5 to 100 L. The peak areas were assigned to the longitudinal coordinate (y), and the injection quantities (g) were assigned to the transverse coordinate (x). Using the opti- mized conditions as described previously, the detection limit (signal-to-noise equal to 3) was 0.01 g for OA, 0.01 g for TA, 0.05 g for MA, 0.01 g for VC, 0.03 g for CA, and 0.1 g for SA. The linear ranges were 0.054.7 g for OA, 0.1110.5 g for TA, 0.11411.4 g for MA, 0.0333.30 g for VC, 0.15515.5 g for CA, and 0.19419.4 g for SA. The regression coefficient (r) was 0.9999 for OA, 0.9999 for TA, 0.9999 for MA, 0.9999 for VC, 0.9998 for CA, and 0.9996 for SA. The relative standard deviation (RSD) was examined, and the RSD values (n = 11) were 0.25% for OA, 0.99% for TA, 1.68% for MA, 0.99% for VC, 0.61% for CA, and 1.23% for SA. The mixed standard stock solution was continuously injected three times every 3 h during a day. Every injected volume was 15 L and the result was averaged. The intraday RSD values (n = 5) were 0.45% for OA, 1.1% for TA, 1.69% for MA, 1.23% for VC, 0.86% for CA, and 1.80% for SA. Also, the solution was only continuously injected three times every day. The operation was the same as that previously described. The interday RSD values (n = 5) were 2.8% for OA, 2.1% for TA, 1.7% for MA, 67.8% for VC, 6.7% for CA, and 2.2% for SA. Obviously, the variation of VC in the determination was very great when its kept time was more than 15 h (for example, the RSD value (n = 5) was 67.8% during 96 h). After 48 h, VC had been decomposed to approximately 69.9%. This was because of the VC being a stronger reducer that can easily be oxidized by oxygen in air when the system pH is less than 4. Thus, one possi- bility is that VC can be determined timely (time < 15 h) in an anal- ysis. The results are summarized in Table I, and a chromatogram is shown in Figure 4. Determination of sample Cores of Fructus mume were removed. At 50C, the cored Fructus mume was sufficiently dried. After that, cored Fructus mume was ground to powder and the powder was filtered through a screen (40 hole/inch). The treated powder was dried again until constant weight at 50C. The powder (1.0000 g) was accurately weighed, and it was extracted with approximately 50 mL water for 2 h according to reference 11. The extracted solu- tion was cooled and filtered. The sediment was washed three times, and each time approximately 5 mL distilled water was used. The eluted and filtered solutions were mixed together and completely transferred into a volumetric bottle (100 mL), and some of the mobile phase was added to the bottle until the mark was reached. Thus, the analytical solution of the sample was obtained. The sample solution was injected into the chro- matographic column until it reached 5 L. Organic acid peaks were identified by comparing their retention times in the sample solution with that of the standard solution, and the contents were quantitated by comparing their peak areas in Table I. Linearity, Detection Limit, and RSD %Intraday %Interday %RSD RSD RSD Constituents Equations r Linear ranges Limits (n = 11) (n = 5) (n = 5) OA y = 314004x + 2041 0.9999 0.054.7 g 0.01 g 0.25 0.45 2.80 TA y = 93616x 1027 0.9999 0.1110.5 g 0.01 g 0.99 1.10 2.10 MA y = 23076x + 2028 0.9999 0.11411.4 g 0.05 g 1.68 1.69 1.70 VC y = 83833x + 5832 0.9999 0.0333.30 g 0.01 g 0.99 1.23 67.8 CA y = 43549x 33 0.9998 0.15515.5 g 0.03 g 0.61 0.86 6.70 SA y = 18779x 6209 0.9996 0.19419.4 g 0.10 g 1.23 1.80 2.20 Table II. Determination Results of Effective Components in the Botanic Fructus mume Sample %OA %TA %MA %VC %CA %SA %Total acid OA/TA/MA/VC/CA/SA Contents of components in Fructus mume (Fujian) 0.554 0.023 6.44 0.84 16.66 4.75 29.27 1:0.04:11.63:1.56:30.07:8.57 Contents of components in Fructus mume (Sichuan) 0.098 0.243 3.53 0.012 6.99 4.3 15.17 1:2.48:36.02:0.12:71.33:43.88 Figure 5. RP-HPLC chromatogram of the extracted solution of Fructus mume: OA, 1; TA, 2; MA, 3; VC, 4; unknown substances, 5 and 6; CA, 7; and SA, 8. Journal of Chromatographic Science, Vol. 40, January 2002 39 the sample solution with that of known standards. This process was repeated three times with each sample. The results are shown in Table II, and a chromatogram is shown in Figure 5. Interference For the previously described chromatographic conditions, the smallest separation degree between two peaks was greater than 1.5 (according to the calculation procedure in reference 15). This conclusion indicated that the impurity in the extracted solution of the sample did not interfere in the determination of the six organic acids component (shown in Figure 5). Investigation of recovery In order to evaluate the accuracy of this method for the deter- mination of the six compounds in Fructus mume, recovery studies were carried out. The recoveries of six constituents were 99.8% to 102.5% for OA, 99.6% to 101.1% for TA, 99.8% to 103.2% for MA, 99.9% to 105.3% for VC, 99.8% to 102.3% for CA, and 100.1% to 101.5% for SA (shown in Table III). These results indicated that the accuracy of the method is higher for the deter- mination of OA, TA, MA, VC, CA, and SA in Fructus mume. Conclusion This study demonstrated that the use of RP-HPLC for the deter- mination of six organic acids in Fructus mume can be success- fully applied to the determination of these organic acids in other botanics. Although a UV conductor suffers poor detection limits for the determination of the six effective compounds in Fructus mume, it offers the advantages of simple procedures. Therefore, it is a suitable method for the routine quality-control analysis of botanics containing the organic acids. Acknowledgments The authors gratefully acknowledge Prof. Wan Xiuqin for providing the HPLC appa- ratus. References 1. Great Dictionary of Chinese Traditional and Herbal Drugs, 2nd ed. Shanghai People Press, Shanghai, 1977, Vol. 1, p 464. 2. Shi-Zhen Li. Bencao Gangmu, 2nd ed. People Health Press, Beijing, China, 1982, Vol. 1, p 1736. 3. Jing-Tang Sun. The action of Fructus mume for the roundworm and gallbladder. Chin. Trad. Herb. Drugs 18(4): 28 (1987). 4. Hong-Mei Shen, C. Qiao, and Z. Su. The research and development of Fructus mume in chemistry, pharmacology and clinical medicine. Chin. Trad. Pat. Med. 15(7): 35 (1993). 5. Jian-Xiu Du, Yin-Huan Li, and Jiu-Ru Lu. Determination of ascorbic acid by on-line cou- pling chemiluminescence. J. Shaanxi Normal Univ. 28(4): 8183 (2000). 6. R.L. Veuzey and T.A. Nieman. Chemiluminescent determination of clinically important organic reductants. Anal Chem. 51: 2092 (1979). 7. Man-Liang Feng and Jiu-Ru Lu. Determination of ascorbic acid by chemiluminescence method with copperlominol. Chin. J. Anal. Chem. 23(1): 7072 (1995). 8. Fu-Chang Wang, Wei Qin, and Zhu-Jun Zhang. Flow- injection/chemiluminescence sensor for the determination of ascorbic acid. Chin. J. Anal. Chem. 25(11): 125558 (1997). 9. Zhan-Guo Chen and Jiu-Ru Lu. Determination for free total content of organic acid by direct potential titration in compound Chinese herbal medicine containing Fructus mume. J. Shaanxi Normal Univ. 28(4): 7880 (2000). 10. Geng-Qing Yang, Yan Zhu, and Xiao-Ping Zhuang. Simultaneous determination of ascorbic acid, saccharine, benzoic acid and ascorbic acid in beverages by high performance liquid chromatog- raphy. Chin. J. Chromatogr. 7(5): 31821 (1989). 11. Hong-Mei Shen, C. Qiao, and C. Su. Quantitative dynamic analysis of organic acid in Fructus mume by isotachophoresis. Chin. Pharm. J. 30(3): 133 (1995). 12. Gang Chen, Xiang-Huan Ding, and Jan-Nong Ye. Determination of rutin and L-ascorbic acid in pharmaceutical preparations and fruit juices by capillary zone electrophoresis with electrochemical detec- tion. Chem. J. Chin. Univ. 21(9): 136468 (2000). 13. Guang-Zhong Wang, Yang Qiu, Jing-Bing Chen, Xiao-Chuan Ye, and Dan Li. Determination of succinic acid in Pheretima aspergillum (E. Perrier) by TLC-scanner. Chin. J. Pharm. Anal. 17(4): 274 (1997). 14. Wang Tianzhi, Li Yongmei, and Wang Zhixiao. Determination of three kinds of organic acid in honeysuckle flower by RP-HPLC. Chin. J. Pharm. Anal. 20(5): 29396 (2000). 15. Chinese Pharmacopoeia, 6th ed. Guangdong Science and Technology Press, Guangzhou, China, 1995, Vol. 1, p 38 and appendix. Manuscript accepted September 10, 2001. Table III. Recovery of Organic Acids Added to the Botanic Fructus mume Components Added (g) Recovered (g) %Recovery (n = 3) OA 0.233 0.239 102.5 0.466 0.468 100.5 2.330 2.325 99.8 TA 0.526 0.524 99.6 1.053 1.055 100.2 5.260 5.260 101.1 MA 0.570 0.588 103.2 1.140 1.146 100.5 5.700 5.689 99.8 VC 0.165 0.174 105.3 0.330 0.333 101.0 1.650 1.648 99.9 CA 0.775 0.774 99.8 1.550 1.569 101.2 7.750 7.928 102.3 SA 0.972 0.987 101.5 1.944 1.954 100.5 9.720 9.730 100.1