Anal. Chem. 2000, 72, 3443-3448
Thermometric Sensing of Peroxide in Organic Media. Application To Monitor the Stability of RBP-Retinol-HRP Complex Kumaran Ramanathan,* Birgitta R. Jo 1 nsson, and Bengt Danielsson
Department of Pure and Applied Biochemistry, Box 124 Centre for Chemistry and Chemical Engineering, Lund University, S-221 00 Lund, Sweden
The stability of horseradish peroxidase (HRP) in aqueous and organic solvents is applied to develop a simple thermometric procedure to detect the binding of retinoic acid-HRP conjugate to retinol binding protein (RBP). Butanone peroxide (BP) in organic phase and hydrogen peroxide in aqueous phase is detected thermometrically on a HRP column, immobilized by cross-linking with glutaraldehyde on controlled pore glass (CPG). Acetone, acetonitrile, methanol, and 2-butanol are used for detection of BP, in the flow injection analysis (FIA) mode. A linear range between 1 and 50 mM BP is obtained in all the organic solvents with a precision of 5-7% (CV%). The magnitude and nature of the thermometric response is significantly different in each organic solvent. The stability of HRP in the organic phase is used to study the stability of a retinoic acid-HRP conjugate bound to immobilized RBP. The response of HRP (to 20 mM BP) in the retinoic acid-HRP conjugate is used as an indicator of the stability of the RBP-retinoic acid-HRP complex, after challenges with various organic/aqueous solvents. Both immobilized HRP and RBP are stable at least for 6 months. The effect of o-phenylene diamine on the thermometric response of HRP is also investigated. A scheme for the design of a thermometric retinol (vitamin A) biosensor is proposed. Biosensing in an organic milieu1-3 and especially sensing of peroxides is a topic of continuing interest. The importance of peroxide biosensors is described in several reports.4-6 Although sensing of peroxide in aqueous systems is well-known, a similar development in nonaqueous (organic) systems is still in its infancy. Predominantly peroxide biosensors employ horseradish peroxide7 (HRP) as the biorecognition element. The most popular substrates for HRP in the organic phase include hydrogen peroxide, cumene hydroperoxide, and butanone peroxide (BP). Hydrogen peroxide as a substrate is extensively used in aqueous systems, while the (1) Dastoli, F. R.; Price, S. Arch. Biochem. Biophys. 1967, 122, 289-291. (2) Hall, G. F.; Turner, A. P. F. Anal. Lett. 1991, 24, 1375-1388. (3) Klibanov, A. M. CHEMTECH 1986, 354-359. (4) Schubert, F.; Saini, S.; Turner, A. P. F.; Scheller, F. Sens. Actuators, B 1992, 7, 408-411. (5) Wang, J.; Lin, Y.; Chen, L. A. Analyst (Cambridge, U.K.) 1993, 118, 277280. (6) Adeyoju, O.; Iwuoha, E. I.; Smyth, M. R. Talanta. 1994, 41, 1603-1608. (7) Lindgren, A.; Tanaka, M.; Ruzgas, T.; Gorton, L.; Gazaryan, I.; Ishimori, K.; Morishima, I. Electrochem. Commun. 1999, 1, 171-175. 10.1021/ac991368l CCC: $19.00 Published on Web 06/30/2000
© 2000 American Chemical Society
other peroxides are employed in the organic phase. The electrochemical detection of peroxide in organic phase is widely reported, while only a few reports deal with other sensing mechanisms.8 Most of the analytical techniques for peroxide determination, such as colorimetry, polarography, etc., are time-consuming and unsuitable for routine on-line analysis. In this regard, enzyme electrodes9-11 and thermal analysis have offered an alternative, relatively fast, and sensitive method for quantitative analysis of peroxides. In addition, HRP could be employed as a suitable enzyme label12 for monitoring specific reactions, e.g., binding of vitamin A (retinol) to retinol binding protein (RBP).13 A subclinical level of Vitamin A is reported to be an important clinical marker for xerophthalmia and keratomalacia, the leading cause of blindness in children. Reference ranges of retinol, tocopherol, and other carotenoids in serum for use in clinical and epidemiological studies have recently been published.14 An inverse relation of retinol levels with the development of cancer, cardiovascular disease, and cataracts is also established.15 The excretion of retinol in urine has also been associated to pathological conditions.16 All these and several other studies suggest the urgent need for the development of techniques for a retinol biosensor. In this regard both electrochemical sensing17 and thermometric sensing using the enzyme thermistor (ET) could be applied for designing such a biosensor. In addition, such a retinol biosensor must be designed to operate in an organic (nonaqueous) phase due to the insolubility of retinol in water. The ET is now well established as a thermometric biosensor system.18,19 Earlier we had reported the feasibility of employing (8) Flygare, L.; Danielsson, B. Ann. N.Y. Acad. Sci. 1988, 542, 485-488. (9) Wang, J.; Reviejo, A. J. Anal. Chem. 1993, 65, 845-847. (10) Wang, J.; Reviejo, A. J.; Mannino, S. Anal. Lett. 1992, 25, 1399-1409. (11) Mionetto, N.; Marty, J. L.; Karube, I. Biosens. Bioelectron. 1994, 9, 463470. (12) Dzgoev, A. B.; Gazaryan, I. G.; Lagrimini, L. M.; Ramanathan, K.; Danielsson, B. Anal. Chem. 1999, 71, 5258-5261. (13) Momeni, N.; Ramanathan, K.; Larsson, P. O.; Danielsson, B.; Bengmark, S.; Khayyami, M. Anal. Chim. Acta 1999, 387, 21-27. (14) Olmedilla, B.; Granado, F.; Martinez, E. G.; Blanco, I.; Hidalgo, E. R. Clin. Chem. 1997, 43, 1066-1071. (15) Kritchevsky, S. B.; Kritchevsky, D. Annu. Rev. Nutr. 1992, 12, 391-416. (16) Stephensen, C. B.; Alvarez, J. O.; Kohatsu, J.; Harmeier, R.; Kennedy, J. I., Jr.; Gammon, R. B., Jr. Am. J. Clin. Nutr. 1994, 60, 388-392. (17) Campanella, L.; Pacifici, F.; Sammartino, M. P.; Tomassetti, M. Bioelectrochem. Bioenerg. 1998, 47, 25-38. (18) Ramanathan, K.; Khayyami, M.; Danielsson, B. Methods in Biotechnology Enzyme and Microbial biosensor: Techniques and Protocols; Mulchandani, A., Rogers, K. R., Eds.; Humana Press, Inc.: Totowa, NJ, 1998,; pp 175186.
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000 3443
water/organic solvent mixtures with the ET setup for enzymeactivity measurements.20 In this report we demonstrate the ease of working with pure organic solvents for peroxide detection with the ET. These investigations are followed by application of the method to test the stability of a retinoic acid-HRP complex bound to an RBP column. Herein, HRP has been used as a “label” by conjugating it with retinoic acid (mimicking retinol) which binds to RBP. The stability of this complex is tested by challenges with different organic solvents of varying hydrophobicities and dielectric constants. The retention of the HRP thermometric response is used as an indicator for the integrity of the RBP-retinoic acidHRP complex. This is followed by release of the complex using an organic solvent mixture and recycling of the RBP column. A possibility of applying this approach for design of a thermometric retinol biosensor is proposed. EXPERIMENTAL SECTION Materials. Horseradish peroxidase (HRP, EC 1.11.1.7) type VI, retinol binding protein (RBP), bovine serum albumin (BSA), o-phenylene diamine (OPD), N-hydroxy succinimide (NHS), dicyclohexyl carbodiimide (DCC), and ethanolamine were from Sigma Chemical Corp., St. Louis, MO. Controlled pore glass (CPG, 67.53 m2 g-1, 50-nm pore size) was obtained from Pierce Chemical Co., Rockford, IL. Buffer salts (Na2HPO4, NaH2PO4, and KCl) and glutaraldehyde (25%) were from Merck, Darmstadt, Germany. All other reagents used were of analytical grade. All the organic solvents (from Merck) were double-distilled and dehydrated before use. Hydrogen peroxide (30%) was from Merck, and 2-butanone peroxide was from Fluka, Buchs, Switzerland. The ET Instrument. The instrumentation has been welldescribed in a recent publication.21 For work with the organic solvents, the O-rings were either made of viton or silicone rubber (for acetonitrile). Teflon tubings were used all along the fluid stream. In brief, an ET employs a sensitive thermistor (up to 0.0001 °C/mV) which responds to changes in the temperature of the solution in its vicinity. Organic solvents possessing a lower heat content provide a much more sensitive response compared with that of the aqueous phase.8 Immobilization. HRP, BSA, and RBP were cross-linked using glutaraldehyde. The technique of protein immobilization using glutaraldehyde has been described in detail elsewhere.22 In brief, 100 mg of amino silylated CPG was reacted with 1 mL of glutaraldehyde (25%) diluted 1:10 in phosphate buffer (PB), pH 6.0 at 4 °C overnight. The beads were washed thoroughly with water, and 500 units of HRP or 0.5 mg of RBP or 1 mg of BSA was added to the solution (200 mg of suspended beads) and allowed to react for 14 h at 4 °C, in the dark. The beads were sedimented by standing, and the supernatant containing the unreacted enzyme was used to assay the efficiency of immobilization. The unreacted sites were blocked with 0.1% BSA prepared in PB, following which they were treated with 2-3 mL of 0.2 M ethanolamine. (19) Ramanathan, K.; Rank, M.; Svitel, J.; Dzgoev, A.; Danielsson, B. Trends Biotechnol. 1999, 17, 499-505. (20) Danielsson, B.; Flygare, L. Sens. Actuators, B 1990, 1, 523-527. (21) Bin, X.; Ramanathan, K.; Danielsson, B. Advances in Biochemical Engineering/Biotechnology; Scheper, T., Ed.; Springer-Verlag: Berlin, 1999; pp 1-34. (22) Torabi, F.; Ramanathan, K.; Svanberg, K.; Khayyami, M.; Okamoto, Y.; Gorton, L.; Danielsson, B. Talanta. 1999, 50, 787-797.
3444 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
Figure 1. The reaction scheme for the coupling of retinoic acid to HRP in the presence of DCC and NHS (for details refer Experimental Section). The structure of retinol is also shown.
Flow Injection Analysis. The beads were packed in Teflon columns (4.5-mm-long and 2-mm i.d) and placed in the thermal probe of the ET unit. The PB, pH 6.0, was degassed and circulated into the ET unit using a peristaltic pump (Alitea) at 1 mL min-1 flow speed. The sensitivity of the ET was set to 10 m °C, and the baseline was allowed to stabilize before the measurement. Samples were injected through a 6-port valve (Rheodyne, USA) employing a sample loop of 100 µL. The thermal changes were recorded on a strip chart recorder (Kipp and Zonen, Delft, Holland), and the concentration of the peroxide versus the peak height was used for constructing the calibration graph. Depending on the nature of the experiment, various organic solvents (acetonitirile, acetone, 2-butanol, and methanol) were continuously circulated at 1 mL min-1 instead of the buffer. All solutions were degassed before use. Synthesis of the Retinoic Acid-HRP Conjugate. The method of synthesis (Figure 1) was modified starting with a procedure reported earlier.23 Retinoic acid, 0.1 mmol, was dissolved in 2.4 mL of dimethyl formamide (DMF). Similarly, 0.1 mmol of NHS and 0.1 mmol of dicyclohexyl carbodiimide (DCC) were dissolved in 0.1 mL of DMF separately. The three solutions were mixed in a round-bottomed flask, and the reaction was allowed to proceed for 20 h under a nitrogen atmosphere. The precipitate was sedimented by centrifugation at 1500 rpm for 5 min. The supernatant was mixed with 500 units of HRP dissolved in 1 mL of 0.1 M sodium borate buffer (pH 8.9). The reaction mixture was stirred for 20 h. The conjugate obtained was purified initially by dialysis against 0.1 M PB followed by a G-75 Sephadex column (40 × 1 cm) chromatography. The purity was tested (23) Zhou, H. R.; Cullum, M. E.; Gerlach, T.; Gage, D. A.; Zile, M. H. J. Nutr. Biochem. 1991, 2, 122-131.
spectrophotometrically for the enzyme peak (280 nm). The enzyme activity was also estimated on the basis of a colorimetric procedure.24 Stability Measurement of the RBP-Retinoic Acid-HRP Complex. To an RBP column described earlier, 100 µL of the conjugate was injected. This was followed by injection of 20 mM hydrogen peroxide to detect the binding of the conjugate to the RBP column. Once the binding was confirmed, different organic solvents were injected at least five times to release the conjugate from the RBP. The choice of the organic solvent mixtures, neat solvents, buffers, or salt solutions, was based on their ability to break specific interactions (between RBP and the conjugate) and release the conjugate. The successful release of the conjugate by a specific solvent or a mixture of solvents was tested by injection of 20 mM BP in a similar organic phase. The presence of a signal indicated the retention of RBP-retinoic acid-HRP complex, and the absence indicated the release. RESULTS AND DISCUSSION Recently, both hydrogen peroxide and butanone peroxide (BP) were measured in aqueous and organic milieu, respectively.25 With increasing interest for the use of organic solvents26 for measurement of peroxides, herein an attempt was made to employ a suitable organic phase (with a water content between 1 and 1.5% v/v) for thermometric measurement of peroxide. While in previous investigations27 hydrogen peroxide was employed for thermometric measurements with organic solvents, in the present instance, BP was used as a substrate for the immobilized HRP in the organic phase. The HRP immobilized on CPG retained almost 70% of its initial activity. There was no leaching of the HRP from the column during measurements. In general, about -100 kJ/ mol21 of enthalpy change was observed on reaction with peroxides. It was advantageous to employ BP as a substrate for HRP, as it was completely soluble in most organic solvents, compared with hydrogen peroxide, which had varying degrees of solubility in different organic solvents. The stability of HRP in such organic solvents was tested and exploited to detect the stability of an RBP-retinoic acid-HRP complex wherein HRP was used as a label. A schematic of the analytical procedure followed in the investigations is shown in Figure 2. Comparison of Hydrogen Peroxide Versus BP as an HRP Substrate. At the outset it was necessary to examine the efficiency of the thermometric response in the organic phase. It was observed that BP in acetone yielded a higher response compared with hydrogen peroxide in PB. Following this, both BP and hydrogen peroxide were examined in acetone. The profile (data not shown) demonstrated that BP responded well in acetone (with 1% aqueous phase) compared with hydrogen peroxide in a similar concentration regime. A similar result was reported earlier, wherein the thermometric response of hydrogen peroxide was higher in toluene than in buffer.20 This suggested that thermometric measurements are most sensitive in the organic phase as compared with in the aqueous phase. Also, between BP and hydrogen peroxide, BP provided a more intense signal compared (24) Bovaird, J. H.; Ngo, T. T.; Lenoff, H. M. Clin. Chem. 1982, 28, 2423-2426. (25) Wang, J.; Reveijo, A. J.; Agnes, L. Electroanalysis 1993, 5, 575-579. (26) Zaks, A.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 31923196. (27) Danielsson, B.; Flygare, L.; Velev, T. Anal. Lett. 1989, 22, 1417-1428.
Figure 2. A schematic of the overall process for design of a retinol biosensor (I) testing the response of peroxides in aqueous and organic phase using immobilized HRP, (II) using a retinoic acid-HRP conjugate to bind to an immobilized RBP column and testing its stability and recyclability using various organic solvent mixtures, and (III) a possible scheme for designing a retinol biosensor.
Figure 3. The response of HRP to BP between 10 and 100 mM: (a) in the absence and (b) presence of 0.5 mM OPD in the mobile phase (acetone with 1% H2O). Flow speed is 1 mL min-1.
with that of hydrogen peroxide. This could be attributed to the lower heat content of the organic solvents compared with that of water. In addition, the sensitivity of BP detection in acetone was much higher compared with hydrogen peroxide under similar conditions. Therefore, BP was used as a substrate for HRP for all measurements in the organic phase. This choice of the substrate was followed by investigations to test the regeneration of HRP using o-phenylene diamine (OPD). The Effect of OPD on the HRP Activity. o-Phenylene diamine (OPD) is normally used as a mediator for HRP and is known to regenerate the HRP activity in electrochemical investigations.6 However, using a thermal transducer, the presence of OPD (curve b) resulted in lowering the thermometric response in the organic phase (Figure 3). In the present instance the studies were carried out in acetone (curve a). The interaction of OPD with HRP was probably an endothermic process leading to some reabsorption of the heat liberated by the HRP-peroxide reaction. This resulted in a lowering of the thermometric signal. The findings suggest that the ET was sensitive to OPD-HRP interactions and could be exploited for future investigations. Interestingly, the linearity range of the BP detection was improved (y ) 0.2766x, R ) 0.99) up to 100 mM in the presence of OPD (curve b), possibly due to the regeneration of a fraction Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3445
Figure 4. The response of HRP to BP between 10 and 100 mM using (a) acetone, (b) methanol, (c) acetonitrile, and (d) 2-butanol, as the mobile phase. The organic phase contains 1% v/v H2O. Flow speed is 1 mL min-1.
of the HRP involved in the catalytic process. Thereby, the presence of OPD influenced both the intensity of the thermometric response and the linearity of the HRP response. However, due to very high HRP loading in the column, it was considered unessential to add OPD in the circulating buffer/organic phase for the regeneration of the HRP activity. Calibration and Analytical Characteristics. After confirmation of the HRP activity in acetone, its activity in a variety of organic phases (Figure 4) such as acetone (curve a), methanol (curve b), acetonitrile (curve c), and 2-butanol (curve d) was measured. It was observed that the increase in response was in the order of 2-butanol < acetonitrile < methanol < acetone. In all these instances, the linearity was found to be between 1 and 50 mM BP, providing a good working range, although, in all cases nonlinearity set in thereafter. Both the sensitivity and linearity were appreciable in the case of acetone (curve a). The reproducibility for a CPG immobilized HRP column was excellent with (7% (CV%). The adsorption or dilution effects could explain the scatter in the data at lower concentrations. However, at the most sensitive operation, 0.1 mM could be detected on such HRP columns. Detection limits of up to 2.5 µM were reported earlier using composite electrodes in an organic phase.25 However, in the case of ET under ideal conditions of instrument stability, a sensitivity of 10 µM could be achieved (S/N ) 3). The response time was found to be about 2.5 min, allowing a sample throughput of at least 15 per hour. The cycle time was about 5 min, with negligible dispersion. The Effect of Varying Water Content. The influence of varying amounts of the aqueous phase was also tested with methanol as the solvent. The response of HRP using methanol with 1, 10, and 20% water was tested. The methanol with 1% water was found to provide the most favorable response compared with that with 20% water. This decrease in sensitivity on increasing the percent water is as yet unclear. Earlier reports27 have demonstrated the opposite trend on increasing the percent water. The increased response with methanol (1% water) may perhaps be attributed to the lower 3446 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
heat content of methanol. In addition, the interaction of the organic medium with the “essential”28 water molecules on the HRP surface played an important role in the retention of the HRP activity. This suggested that the nature and the intensity of the response vary depending on the water content in the organic phase and the method of detection. Nevertheless, the survival of substantial activity in methanol indicated the usefulness of HRP in organicphase thermometric sensing. The Nature of the Thermometric Signal. A notable feature in the present investigations was the nature of the response, i.e., the thermometric response peaks were much sharper and more well-defined in the organic phase compared with the aqueous phase. An estimate of the full width at half maximum (FWHM in millimeters) of the thermometric response peak showed that in the organic phase (acetonitrile, 0.5; acetone, 0.3; methanol, 0.5; butanol, 0.5) the FWHM was almost 3 times lower than that in the aqueous phase (buffer, 3.0). This indicated well-defined reaction kinetics at the enzyme active site in the presence of the organic phase compared with that in the presence of the aqueous phase.20 On the basis of these findings, it could be stated that the thermometric technique was suitable for studies with pure organic phase and screening of HRP activity in various organic solvents such as acetone, methanol, acetonitrile, and 2-butanol. In most cases, the thermometric response is a combination of an increased enthalpy change and a decreased heat capacity of the organic solvent. The Effect of Solvent Hydrophobicity. The enzymatic response varies with the change in the hydrophobicity of the solvent. The hydrophobicity could be defined in terms of the log P values29 (where P is the partition coefficient of the given solvent between octanol and water). This has been well documented in an earlier report.30 In our case, it was observed that the intensity of the response varies as acetone > methanol > acetonitrile > 2-butanol. Their corresponding log P values are -0.24, -0.77, -0.34, 12.5, respectively. A similar effect was demonstrated earlier,30 wherein these were attributed to several complex effects. It was suggested that there was a positive correlation between 1/η of the solvent and the sensor performances in methanol, acetone, acetonitrile, and 2-butanol (wherein is the dielectric constant and η is the absolute viscosity of the solvent). With the exception of 2-butanol, the 1/η values followed the same trend as the log P value. In an earlier report20 we had demonstrated that the activity of HRP in toluene and with 5 and 10% diethyl ether was 10.6, 13.9, and 8.8 times higher than the activity in 0.1 M PB, pH 7.0. In addition, the enzyme activity lost during the operation in the organic solvent could be fully restored by reconditioning with PB. Following the understanding of the behavior of immobilized HRP with BP, using a thermometric detection scheme, the use of this concept was extended to monitoring the stability of a purified complex of retinoic acid and HRP on a column with immobilized retinol binding protein (RBP). The details on the (28) Borzeix, F.; Monot, F.; Vandecasteele, J.-P. Enzyme Microb. Technol. 1992, 14, 791-797. (29) Laane, C.; Boeren, S.; Vos, K.; Veeger, C. Biotechnol. Bioeng. 1987, 30, 81-87. (30) Iwuoha, E. I.; Smyth, M. R.; Lyons, M. E. G. Biosens. Bioelectron. 1997, 12, 53-75.
Table 1. Challenging of the Retinoic Acid-HRP Complex Bound to the RBP with Various Organic/ Aqueous Solvents Followed by Injection of 20 mM Butanone Peroxide in the Same Solvent screening no. 1 2 3
Figure 5. The elution profile of retinoic acid-HRP conjugate from a Sephadex G-75 column (40 × 1 cm) using 0.1 M PB, pH 7.0 as the eluent. (A) The retinoic acid-HRP conjugate (confirmed by its binding to immobilized RBP) and (B) unreacted HRP (confirmed by spectrophotometric assay).
4 5 6 7
purification and characterization of the complex are described in the following sections. Purification of the Retinoic Acid-HRP Complex. After investigating the behavior of HRP in various organic solvents, a conjugate of the HRP with retinoic acid was synthesized (described in the Experimental Section). We explored the possibility of studying the binding of this retinoic acid-HRP conjugate with retinol binding protein (RBP). Both retinol and retinoic acid share the same skeletal structure (Figure 1), except for the presence of a carboxylic acid group in the latter. Figure 5 shows the purification of the complex using a Sephadex G-75 column. The fraction collected under the first peak (A) showed binding to a RBP column, while the fraction under the second peak (B) showed no such activity and was confirmed to be the unreacted HRP, on the basis of a colorimetric assay. The shoulder that accompanies peak A could not be assigned to any specific molecular complex. It may perhaps be a moiety having a molecular weight intermediate between that of HRP and that of the retinoic acid-HRP conjugate. The specificity of the conjugate (retinoic acid-HRP) was tested by injecting it into a column containing immobilized BSA and into another column containing immobilized RBP and following this with the injection of BP. Nonspecific binding of the conjugate to BSA was found to be negligible, while a thermometric peak was obtained for HRP (coupled to the conjugate) from the RBP column, following the injection of 10 mM BP. This indicates retention of the specificity by the RBP for retinoic acid, despite the immobilization, and also the ability to monitor such binding thermometrically using HRP as the enzyme label. Testing the Stability of the RBP-Retinoic Acid-HRP Complex. On an RBP column, 100 µL of retinoic-acid-HRP conjugate was injected. Thereafter it was washed with copious amounts of PB. Various aqueous/organic solvents (see Table 1) were injected (100 µL) at least five times, following which BP prepared in a similar solvent was injected. Intermittent checks were also made using injections of various buffers, e.g., 0.1 M glycine buffer (pH 2.2), followed by injection of hydrogen peroxide in a similar medium. The thermometric response from HRP indicated that the complex (retinoic acid-RBP) was unaffected (intact) by the injection of the various organic solvents by ions such as phosphate or glycine, or by the changes in pH. It also demonstrated the stability of HRP to such challenges with the organic phase. The
8 9 10 11 12
solvent system phosphate buffer NaCl glycine buffer β-mercaptoethanol ethanol ethanol + NH4SO4 ethylene glycol + ethanol DMF chloroform + methanol acetone + acetonitrile 2-butanol ethanol + hexane
concn
mean Th. peak ht. (mm)d
SD
CV (%)
aqueous 0.1 M, pH 7.0
6.0a
0.76
13.1
17.2a
1.03 1.15
6.0 18.5
0.1 M 0.1 M, pH 2.2 nonaqueous neatb
6.2a 15.0
0.79
5.26
neatb 50% + 0.1 M
14.7 16.5
1.03 0.80
7.00 4.84
neatb 1:1
16.0
0.78
4.87
neatb neatb 1:1
17.0 25.7
0.79 1.20
4.64 4.66
neatb 1:1
20.0
0.81
4.05
neatb
12.0 1.9c
0.77 0.41
6.41 21.5
neatb 1:1
a For 20 mM injection of H O . b With 1.5% water. c The reading after 2 2 correcting for the baseline value. d All readings are an average of five injections.
decrease in the response, for instance, with a mixture of 1:1 acetone/acetonitrile (20 mm) compared with 1:1 chloroform/ methanol (26 mm) showed the extent to which the HRP was affected by the specific solvent or solvent mixtures. It was not our intention to compare the responses (Figure 2, I) obtained from HRP (in the conjugate) toward hydrogen peroxide (in aqueous phase) and BP (in organic phase), but only to demonstrate the retention of activity of HRP bound to retinoic acid in both these phases for the design of a retinol biosensor. However, after injection of a mixture of 1:1 ethanol/hexane, there was a very weak response (mean thermometric peak height of 1.9 mm) from the BP injection, due to the release of the conjugate from the RBP column. An earlier report16 also demonstrated that the retinol could be extracted from a retinol-RBP complex in urine using a mixture of methanol and hexane by twophase extraction for 30 s, which was in support of our present findings. The release of the retinoic acid-HRP conjugate was verified by colorimetric determination of the HRP in the eluate from the column. In addition, the binding and release could be cycled (Figure 2, II) and was tested at least three times on the RBP column. These experiments suggest that considerable information could be obtained on the stability of RBP-retinoic acid-HRP complexes by employing the thermometric approach. In addition, the technique was very adaptable for screening purposes with several organic solvents on immobilized HRP. The investigations point out that the retention of HRP activity in organic solvents could be used as a tool for monitoring the binding of retinoic acid-HRP conjugates to RBP. The conformational stability of RBP was sufficient in various organic solvents31 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3447
Figure 6. The principle for the detection of retinol (vitamin A). The binding of retinoic acid-HRP conjugate to immobilized RBP followed by thermometric detection could be employed. (a) immobilization of RBP to CPG, (b) binding of the retinoic-acid HRP complex to immobilized RBP, (c) testing the stability of the RBP-retinoic acid-HRP complex by challenging with various organic solvents, and (d) thermometric detection of the HRP label.
and retains its selectivity to retinoic acid (mimicing retinol) despite covalent cross-linking with glutaraldehyde. On the basis of this approach a competitive assay (Figure 2, III) could be designed for a retinol biosensor. Here a competition between labeled (HRP) and unlabeled retinol (Figure 6) could compete for the active site on RBP. The need for such a Vitamin A (retinol) sensor was indicated earlier,30 and our approach could be used for this purpose. Such a biosensor could also be operable in an organic phase. This aspect is to be pursued in future investigations. CONCLUSIONS The enzyme thermistor based thermometric sensing is a useful tool for the sensing of peroxides both in the aqueous and nearly pure organic phases. Enhanced sensitivity and detectability was obtained for peroxide sensing in the organic phase. The stability of the cross-linked HRP was appreciable and was employed for thermometric sensing in the flow injection mode. The reproducibility in all the solvents tested, expressed as a coefficient of variation, was between 5 and 7%. The nature of the response varied depending on the presence or absence of regenerating agents such as o-phenylene diamine, (31) Goodman, D. S.; Raz, A. J. Lipid Res. 1972, 13, 338-347.
3448
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
which is known to regenerate HRP in electrochemical systems. In the presence of the organic phase, the response was sharper, with minimum dispersion, and more intense. About 15 injections could be made per hour with an average sensing of 2.5-3 min. The system was successfully applied to detect the binding of a retinoic-HRP conjugate to an immobilized RBP column. Challenging the complex with a variety of neat organic solvents and solvent mixtures could monitor the stability of such binding. The information obtained from such investigations could be employed for the possible design of a thermometric retinol biosensor. The concept is more general and could be employed for design of a biosensor using other transduction principles. ACKNOWLEDGMENT The authors wish to thank Dr. B. Yomtov for fruitful discussions. K.R. wishes to thank SIDA/SAREC and the Swedish Institute for the financial support during the course of this work.
Received for review November 29, 1999. Accepted April 19, 2000. AC991368L
