ORIGINAL RESEARCH Luminescence 2000:15:305-309 Flow injection system with chemiluminometric detection for enzymatic determination of ascorbic acid Andrei F. Danet,’ Mihaela Badea? and Hassan Y. Aboul-Enein** “Univesity of Bucharest, Faculty of Chemist, PO Box 164, Bucharest 15, Romania chemical Research Institute, Department of Biotechnology, Spl. Independentel 202, PO Box 67, Bucharest 15, Romania Bioanalytical and Orug Laboratory, Biological and Medical Research Department (MBC-03), King Fatal Specialist Hospital and Research Centre, PO Box 3354, Rach, Saudi Arabia Received 10 October 1999: accepted 6 February 2000 ABSTRACT: A simple, selective and rapid method for determination of ascorbic acid from fruit juices was developed by combining a flow injection analysis (FLA) system with a chemiluminometric detector and a reactor with L-ascorbate oxidase immobilized on controlled pore glass, It was found that some reducing agents (eg ascorbic acid and mereaptoacetic acid) give chemiluminescence ‘with lurminol inthe presence of hexaeyanoferrate (III) in an alkaline solution. We used this new type of chemiluminescent reaction for the enaymatic determination of aseorbic acid, The background substraction method was used in order to avoid interference during ascorbic acid determination. Accordingly, two chemiluminometsic signals were registered for each determination, one signal corresponding to the sample that passed through the enzymatic reactor that decomposed the ascorbic acid completely, and the second signal corresponding to the sample that does not pass through the reactor. The difference between the two signals corresponds to ascorbie acid from the sample. The linear range of the method was 10-1000 pmoVL of ascorbic acid and the detection limit was '5 pmol/L. The throughput was four samplesh and RSD 3.13% (n = 10), This method was applied for determination of ascorbic in fruit juices. The results were compared with those found by the reference method, based on titrimetric determination with ‘ichlorophenolindaphenol, and the concordance was excellent. Copyright € 2000 John Wiley & S 6 Lid. KEYWORDS: chemiluminometric analysis: vitamin C, L-ascorbie acid: flow injection analysis: L-ascorbate oxidase INTRODUCTION Ascorbic acid (vitamin-C), found in citrus fruits and ‘vegetable products, is mainly used a cure for scurvy, drug, poisoning, liver disease, allergic reactions and athero- sclerosis. Various methods have been proposed for determination of ascorbic acid, such as titrimetric (1), fluorimetric (2), chromatographic (3-5), electroanalytical (6-5), spectrophotometric (10-11) and chemilumino- metric (12-14) methods. Each method has its advantages and disadvantages. In recent years, biosensors with high selectivity based on an electrode coated with ascorbate oxidase for ascorbic acid determination have appeared especially attractive. In these sensors, the consumption of dissolved oxygen takes place in an enzymatic reaction and is detected by an oxygen electrode (15). However, these sensors still suffer from stability problems. ‘Chemiluminescence (CL) has been successfully ap- sComrespondence 10: H. ¥. AboulEnein, Bioanalytical and Drug Laboratory, Biological and Medical Research Department (MBC-03), ing Faisal Specialist Hospital and Research Centre, PO Box 3354, Riyadh, Saul Arabia. E-mail enein@ shire edu.sa CContrctgrant sponsor: King Faisal Specialist Hospital and Research Gente, Riyadh, Saudi Arabia Copyright © 2000 John Wiley & Sons, Ld plied to determine ascorbic acid with excellent sensitivity over a wide linear dynamic range and with the use of simple instrumentation. The majority of papers describe ‘methods based on the reducing effect of ascorbic acid on iron (IT1) ion and measuring the iron (I1)-catalysed light ‘emission from luminol oxidation by hydrogen peroxide (14), However, it was found recently that reducing agents (eg uric acid, cortisone, ascorbic acid, and cysteine) give CL with luminol in the presence of a catalyst in an alkaline solution without the presence “of hydrogen peroxide (16). We have also reported a method for sulphydryl (SH-) compound determinatign based on this observation (17), There are several theories about this new CL reaction. ‘The chemilominescence spectra obtained from the reducing agents are similar to luminol. Therefore, the chemiluminescence species is assumed to be an electro- nically excited aminophthalate produced by the reaction ‘of luminol with the reducing agents in the presence of a catalyst (potassium ferricyanide) in an alkaline solution. A flow analysis system has been set up for determining the ascorbic acid from fruit juices using the above- mentioned reaction. A reactor with immobilized ascor- bate oxidase was used for selective decomposition of the ascorbic acid. The signal registered after the passage of306 (ORIGINAL RESEARCH the sample through the enzymatic reactor was compared seep ihat obtained for the sample that did not flow through the reactor. The difference between the *¥0 signals was dependent on the ascorbic acid concentaton Signals Nal signal obtained after the decomposition of the ascorbic acid is given by interfering or foreign Substances. The major advantage of this enzymatic srethod coupled with CL detection is its high selectiv’ty, Compared with other methods, the proposed method Cree potential advantages of simplicity and rapidity or ascorbic acid determination. EXPERIMENTAL Reagents All reagents were of analytical grade; bi-distilled water eis used for the preparation of solutions of all dilutions srtsyoek solution of ascorbic acid (Chimopar, Bucharest, Romania) (0.02 moV/L. in water) was stored jn, the fefrigerator. Working. standards were prepared daily from stock solution by appropriate dilution. Amin’ vremicedconuolled pore glass pearls (200-400 mesh) ‘Aldrich, Milwaukee, WD were used as immobilizing Support for ascorbate oxidase. Ascorbate oxides (EC Trp 3.3), with a reported activity of 143 Ulm solid, wae purchased from Sigma (St. Louis. MO) and <0 used Puttar any further purification. Glutaraldehyde 25% and potassium ferricyanide were from Sigma (St Lov MO). ree CL reagent, luminol (Aldrich, Milwaukee, WH), wes prepared in a 0.1 moV/L sodium carbonate soluron 1) Prepares ensure the necessary alkaline pH for the CL orasrion. Sodium phosphate and. tris(hydroxymethy)) reactionunane (Tvs) were used for the preparation of buffers, Enzyme reactor ‘Aminosilanized glass pearls were activated by soeking amine in a 10% glutaraldehyde solution prepared in Bi- Gisuiled water (PH 5.8-6.0), after which they Were sisthed with cooled water on a Gs filter (Schott, Jena- Germany). Immobilization of the ascorbate oxidase 9% citieved by placing 200 mg from the activated support serif in dni. 0.1 moV/L phosphate buffer solution (PH $6) with a content of 500U ascorbate oxidase. with setasional stirring at 42°C. The non-immobilized enzy re seca tnoved by washing with 100 mL. cooled 0.1 mol phosphate buffer solution, pH 5.6, and then with 500 mL 1 mol/L, NaCl "The ascorbate oxidase reactor was made by packing the immobilized enzyme in a PVC be (d.=2 Ti, Tength = 6 cm). After utilization, the reactor was washed ere NaCl 0.1 moVL, for 20min and was kept i refrigerator at 4°C. Ccopysight © 2000 John Wiley & Sons, Ld |A.F, Dine, M. Bade 2 HY. Abo: Ks{Fe(CN) ominol Figure 1. low analysis system with CL detection, Puss Pibstakde pumps: ER, enzymatic reactor Wi ‘scorbale eerae immobilized onto CPG; 1, injection valve guess Trmworway cock valve! C. flow cell, PMT. ‘Shotomultiptier tube: R,registrator; W, waste ‘The reactor with immobilized ascorbate oxidase was used for 3 weeks without any Joss in enzymatic activity sree 6 weeks the enzymatic activity of the immobilized arrpate oxidase was reduced to about 55% ofits inital activity. Flow manifold ‘The original conception of the device used for Cle direction of ascorbic acid determination is presented ese. The peristaltic pumps tubes were made of ‘Tygon G.d.= 102mm) and the other tubes were made Of polytetrafluoroetbylene (PTFE) (i.d.=0.5 mim). croke-channel pump, P) (Pharmacia Fine Chemicals) aoe paed for filling the injection loop with sample, whose wane was 50 HL. The injection was performed with 2 Sway valve model 5051, theodyne-type (Cotath CAs Tsar The three reagents, namely, carrier (bi-distiled Tuminol and potassium ferricyanide solutions, fespectvely, are driven by a four-channel pensalie pump P; (AUC-Bucharest). A laboratory-made Aeeciot ree ow analysis. spiral cell_were employed for ane curing the CL radiation, The flow cell was a spiral Trawemade cell made of PTFE tube (length 50 cm id {3 mm), The voltage applied to the photomultiphier «be (PMT) was 820 V. The signals were sraphicelly recorded ‘with a Eldim 621.02 (Germany) y-t recorder. wt two-way cock valve was used to select the pathway for the sample—a way through the ascorbate oxidase reactor, of a way that bypasses the enzymatic reactor. Procedures tis necessary for each sample determination to make (9 injections. The fist injection corresponds fos the sample vit does not pass through the ascorbate oxidase react Tee GE cignal obtained for this injection is due to Pe rete acid and other interfering compounds in the Luminescence 2000;15:305-309‘“Aemilorinom Hex (rian unity 8 thar The second injection corresponds {0 the sample that is passed through the enzymane Feactor where the ascorbic acid is decomposed. Once ascorbic acid ig the decomposed, the CL signal given will be due to the imerference of foreign Substances present in the Sample. The working temperature was 24he For activation of the enzyme. a G1 mol/L phosphate buffer, pH=5.6, was passed through the enzymatic ‘eactor for 5 min after each sample, RESULTS AND Discussion In order to optimize the Working parameters of the flow concn analysis device, we studied the influence of soleniantations of Iuminol and. potassiwns ferricyanide solutions, flow rate and sample DH. The influence of potassium ferricyanide, which is 9 catalyst for the CL reaction, is illustrated in Fig, 2 Potassium ferricyanide as a catalyst for the chemilumi. the ott reaction has been tested for a 2 mmol/L luminol, the ascorbic acid concentration being 0.5 mmol. Although the obtained signals are intense at concentra. pons found 15 mmoVL. it is not advisable to work at these concentrations because the Signal is not stable, A subserttition of 20 mmol. ferricyanide hhas been used subsequently. Effect of luminol concentration Une influence of luminol concentration fg shown in Fig.3, where one may observe that the magnitude of FIA signale 'S directly proportional to luminol Concentration of 0.5- Copyright © 2000 John Witey & Sons, Lut | | I Lumino mmout) Figure 3. Effects of tuminol concentration on FLA. Cyesaian rate 8 LOM. Cocrbic a= 0.5 mince total flow rate = 23 mL/min 2.0 mmol/L (prepared in 0.1 mol. NazCO3), after which wa ches @ plateau. A 2.0 mmol/L luminer concentration was subsequently employed, We Specify that a freshly MePared luminol solution must be uses This solution eis HS activity for a week when kept in a refrigerator. Effect of total flow rate Flow rate determines the ley, of the background and the inagnitude of the signal due to its influence on the mixing Osan cwteen sample and reagent, Samples. with 0.5 mmol/L ascorbic acid were injected and the total flow rate was varied between 0.8 and 5.6 mL/min (Fig D- At Tow flow rates, the signal decreased due to the FIA signals, fersoanide =20 mmol, Corry Luminescence 2000415:303-309308 ORIGINAL RESEARCH Be 3 . a oo 8 Figure teens of th sale -cirlaton ine hough the ascorbate oxidase reactor on the decomposition of ascorbic acid. Chemin = 2 mmolL: Chon friygrite = 20 mmol Conic acy = 0.5 mmol: flow rate through enzymtic reac tor 03 tL imin, ‘missing’ CL emission in the detector flow cell, For flow rates of 2.54.5 mL/min. approximately the same peak heights were obtained. A flow rate of 3 mL/min was therefore chosen as the optimal flow rate. Effect of sample pH The pl of the sample solution was found to be critical for CL measurements. Due to the fact that the chemilumi- nescent reaction requires an alkaline pH, sample pH <5 causes a drastic decrease of the CL signal. At pH >65 ascorbie acid is very unstable. Subsequently, we used samples made in bi-dstlled water (pH 5.8-6.), Influence of the sample re-circulation time through the enzymatic reactor For ascorbic acid determination in real samples, complete decomposition of the ascorbic acid is essential. Fig. 5 shows the influence of the sample recirculation time through the ascorbate oxidase reactor, For a recirculation time of 15min, the decomposition of ascorbic acid is ~98%. Therefore, a recicculation time of 15 min was, selected as an optimum recirculation time to ensure complete decomposition of ascorbic acid. Analytical performances The following are the optimized parameters for ascorbic acid determination: total flow rate of 3 mL/min (the ow rates of the three channels were the same); sample volume of 50 4L: luminol concentration of 2.0 mmol/L potassium ferricyanide concentration of 20 mmol/L. The Copysisht © 2000 John Wiley & Sons, Lid ALF. Dine, M, Badea and H. Y, Abou!-Eacin calibration graph is linear in the concentration range of 10-1000 umoVL ascorbié acid The equation for the calibration graph is: Tey, = 8.883 ¢ ~ 3.043, where Icy =the height of FIA signal (relative units) and ‘e=the concentration of ascorbic acid (mol/L). The correlation coefficient is r= 0.9998. The relative standard deviation for 10 replicate sample injections of 0.5 mmoV/L ascorbic acid is 3.13%. The detection limit is S HmOV/L ascorbic acid (SIN >2) with RSD = 11.3% (five replicates). The throughput is four samples/h. Ascorbic acid determination from fruit juices Natural fruits were bought freshly. Each fruit was squeezed to obtain its juice. The juices were centrfi- gated for 20min at 10,000 rpm until a clear liquid was obtained. These liquids were diluted with 0.05 mmoV/L. Phosphate buffer, pH = 6.5 For ascorbic analysis in natural juices the reactor with ascorbate oxidase was used for selective decomposition of the ascorbic acid. The signal registered after the passage of the sample through the enzymatic reactor (ecirculation time 15 min) was compared with that ‘obtained for the sample that not flow through the reactor. ‘The difference between the two signals was interpolated on a calibration graph obtained for ascorbic acid. The residual signal obtained after the decomposition of the ascorbic acid is given by the interfering foreign substances. Table 1 shows the ascorbic acid content of some fruit Juices analysed by the proposed method and compared with the standard titrimetric method. Although the sample matrix is complex, the results show that the concentration of ascorbic acid in various kinds of juices ‘Table I. Determination of ascorbic acid in fruit juices Ascorbic acid (jg/ml)® Juice “Added” Found® Recovery (@) Claimed Lemon ~ 763 = 781 40 802 975, 80 Bat 1012 Orange = 615 - 669 40 686, 1025 80 mm 102.5 Kiwi 720 - 7a 50 m 100 823 {Average of thee determinations. Added to the undilted samples {According to the proposed metho. * According tothe oficial method (based) inthe iri determing: sn with 2,6-dihlorophenolingopheno) (17). Luminescence 2000;15:305-308‘Chemiluminometic analysis of ascorbic acid ‘are in good agreement for both methods (all data are within the 95% confidence level). CONCLUSIONS ‘The determination of ascorbic acid by the proposed flow injection method with chemiluminometric detection has advantages over conventional methods in that it is faster and simpler. By using the ascorbate oxidase reactor for ascorbic acid decomposition, the fruit juices can be analysed with only simple pretreatment (centrifugation) and with great selectivity. The dynamic range and the precision of the proposed method are similar ta those provided by non-enzymatic flow injection methods with spectrophotometric (10) or chemiluminometric detection 3). Acknowledgements One of the authors (HLY.A-E) wishes to thank the Administration of King Faisal Specialist Hospital and Research Centre for its support of the Bioanalytical and. Drug Development Laboratory research program. REFERENCES 1. Verma KK, Jain 8. Sahasrabuddhey B, Gupta K. Mishra S, Soli phase extraction cleanup for deierhining ascorbic acid and ‘ehydroascorbie acid by tration ‘with 2 6dichiorophenolindo phenol. . AOAC 1996: 79. 1236-1253, 2. Ensafi AA, Rezaei B. Flow injection analysis determination of ascorbic acd with spectoluorimetie detection, al, Let 1098: 31 333-342, 3. Aboui-Enein HY, ALDurabi 1A, Analysis of L- and p ascorbic Copyright © 2000 John Wiley & Sons, Lid ORIGINAL RESEARCH 309 acid in fruits and fruit inks by HPLC. Sem. Food Anal. 1999; 4 31-37 Moll N, Joly 1P. Determination of ascorbic acid in bees by high performance liquid chromatography with electrochemical detec. tion. J. Chromatogr. 1987; 408: 347-356, Lee HS, Costes GA. Measurement of total vitamin C activity in citrus produets by HPLC? a review J. Lig, Chrom Rel. Tectve 1999; 22° 2867-2387, Casella 1G, Elecro-oxidation of ascorbic acid on the dispersed platinum glassy-carbon electode and its amperometric determinae tion in low injection aralysis. Electroanalysis 1996: 8. 12-134, {Leon LE. Amperometrie flow-injection method forthe assy of ascorbic acid based on the photochemical reduction. of methylene blue. Tafonta 1996; 43: 1275-1279, Garcia Alonso J. de Ja Fuente Sincher C, Tascén Garcia ML, Varquer Barbado MD, Sinchez Batanero P. Electrosnaltical properties of a potassium hexaeyanoferte (il) modified poly Dytoe electrode. Application to vitamin C determination. nal err 1999; 32: 1961-1980, Friedrich ©, Sontag G, Pitiner F, Electrochemical FIA-system combined with an immobilized reactor for determination of ascorbic acid. Nutrition 1993: 17: 605-608, Molina-Diaz A, Ontega-Carmona I, Pascual-Reguera ML Indirect, spectrophotometrc determination of ascorbic aid with ferrocine by flow injection aalysis. Talanta 1998: 47: 531-536, ‘Tutem E, Ulkuseven B, Apak R. Spectrophotontetie determing: tion of ascorbic acid in pharmaceuteal products with the copper ‘thodamine reagent. Anal. Lett 1992.25; 171183, Alwanhan AA. Determination of ascorbic acid by flow injection with chemiluminescence detection. Analet 1995; 118, 633-612, Zhang Z. Qin W. Chemiluminesence flow sensor forthe {etermination of ascorbic acid with immobilized reagents. Tale 1986: 43: 19-124, Kubo H. Toriba'A. Chemiluminecence flow injection analysis of reducing agents based on the luminol reaction. inal Chim. Acta 1997: 383: 5-349, Uchiyama $. Chemically amplified curent response of vitamin C based on the cyclic reaction between Lascorbie seid and dehy doe ascorbic acid, using dthioheithol and ascorbate oxidase Tanta 1992: 39; 1289-1292. Danet AF, Badea M. Aboul-Encin HY. Determination of paroxon by inhibition of aceislehotinesterase and a new type of chemo luminescence reaction. Biospeciroscr in press. Romanian STAS. Determination of Tol and Active Ascorbic Acid, 182/15-79, 2184, 1990, Luminescence 2000%15:308-309 To