Regeneration of immobilized antibodies on fiber optic probes

Biosensors & Bioelectronics 9 (1994) 585- 592 Regeneration of immobilized antibodies on fiber optic probes Daya Wijesuriya Geo-Centers, Inc., Fort Washington, Kristen Breslin, George Anderson, Center for BioMolecular MD, USA Lisa Shriver-Lake & Frances S. Ligler* Science and Engineering, Code 6900, Naval Research Laboratory, DC 20375-5348, USA Tel: [l] (202) 767 1681. Fax: [l] (202) 404 8938 (Received 26 November Washington, 1993; revised version received 17 May 1994; accepted 6 June 1994) zyxwvutsrqponmlkjihgfedcb Abstract: The regeneration of antibodies covalently immobilized to an optical fibre surface was investigated by dissociation of the antibody-antigen complex with three different solvents: (a) an acidic solution (0.1 M glycine hydrochloride in 50% (v/v) ethylene glycol, pH l-75), (b) a basic solution (0.05 M tetraethylamine in 50% (v/v) ethylene glycol, pH 11.0) and (c) 50% (v/v) ethanol in PBS. The fibres coated with polyclonal rabbit anti-goat antibody against a large protein retained 70% and 65% of the original signal after five consecutive regenerations with acidic and basic solvent systems, respectively. The fibres coated with monoclonal mouse anti-trinitrobenzene antibody specific for a small organic molecule, retained over 90% of the original signal when regenerated with basic and ethanol solutions. This study evaluated regeneration and reuse of antibody-coated fibre optic biosensors as a means of reducing routine laboratory analysis costs and time. Keywords: fibre optic biosensor, regeneration, INTRODUCTION Fibre optic biosensors utilizing the evanescent wave are being developed for the detection of pathogenic organisms, drugs, toxins, explosives, and environmental pollutants (Xie et al., 1990; Arnold et al., 1988; Wise et al., 1991; Ogert et al., 1992; Ligler et al., 1993). In most of these assays, antibodies specific to an antigen of interest are covalently immobilized on the core surface * To whom correspondence 09565663/94/$07.00@ should be addressed. 1994 Elsevier Science Ltd. antibodies of fused silica fibres. Depending on the size of the antigen, competitive or sandwich immunoassays are conducted at the fibre surface. A fluorescence signal is generated in real time upon binding of a complex that includes a fluorophore and the antigen to the immobilized antibody. Although successful, previously described fibreoptic immunoassays used one fibre per test. Not only was the cost a disadvantage, but fibre-tofibre variation affected the reproducibility of quantitative analyses. Regeneration of the antibody-coated surface for multiple uses by removing antibody-bound antigen is an alternative approach 585 D. W ijesuriy a et al. which would reduce these problems. Antibodyantigen (Ab-Ag) complexes are formed by several non-covalent interactions (i.e. electrostatic, Van der Waals and hydrogen bonding). To achieve effective dissociation of the antigen from the immobilized antibodies, the strength of these interactions needs to be reduced. Changes in pH, ionic strength and antigen solubility in the washing solution may accomplish this goal. Blanchard et al., (1990) studied the durability and regeneration of antibodies immobilized on commercial immunosorbents by monitoring antibody-antigen dissociation. Antibodies were covalently immobilized to glass, polystyrene beads, microtiter plates and polyvinylidene difluoride. It was reported that solutions consisting of either O-01 M hydrochloric acid, 10% propionic acid, 50% ethylene glycol, or 10% sodium dodecyl sulphate in 6 M urea could be used to dissociate only 2-26% of IgG antigen bound to rabbit antihuman IgG. Recent efforts on developing regenerable immunosensors employed optical fibres with antibodies immobilized at the distal end. The regenerable fibre optic immunosensor was originally developed for detecting human serum albumin using dansylated F(ab’) anti-albumin antibody fragments (Bright et al., 1990) and extended to detection of small molecular weight haptens (Betts et al., 1991). In both reports, much attention was focused on optimizing or selecting reagents for both rapid dissociation of antigenantibody complexes and on maintaining the stability of the immobilized antibody. For each fibre probe, between 60% and 90% of the antigen binding activity remained after 10 cycles of washing and re-exposure to antigen, but the activity continued to decrease in a nearly linear fashion. Lu et al. (1992) developed a regenerable immunosensor based on a planar quartz waveguide to detect human IgG. According to this report, the dissociation of bound antigens was found to be more effective using aqueous diethylamine than using an acid solution of glycine hydrochloride (glycine-HCl), in terms of maintaining higher reassociation ability (over 97%). Using diethylamine, the antibody-coated surface was utilized up to eight times without any apparent loss of antigen binding activity. However, effective removal or dissociation of bound antigens from the antibodies on the surface involved a continuous wash with the regeneration 586 Biosensors & Bioelectronics solvent system for more than 30 min before reexposure to the same antigen solution. For routine analysis of a large number of samples, such time and reagent consuming regeneration steps are not acceptable. The work described here demonstrates the feasibility of regenerating antibodies immobilized in an evanescent wave fibre-optic biosensor. Model antibody-antigen pairs such as polyclonal rabbit anti-goat immunoglobulin G (rab antigIgG)/tetramethylrhodamine isothiocyanate-labelled goat IgG (TRITC-gIgG) and monoclonal mouse anti-trinitrobenzene (mouse anti-TNB)/ tetramethylrhodamine cadaverine-labelled TNB (TRC-TNB) were used for characterization. Efficiency of removing antigens bound to antibodies on the fibre surface was investigated using acidic and basic regenerating solvent systems. It was demonstrated that, in some cases, antibodycoated fibres can be regenerated using less severe denaturants if the solubility of bound antigen is increased. Finally, it was shown that quantitative dose/response relationships can be maintained after regeneration. MATERIALS AND METHODS zyxwvutsrqponmlkjihgfedcbaZYX Reagents Analytical grade reagents and distilled water were used for preparation of all solutions. Rabbit anti-goat immunoglobulin G (rab anti-gIgG) and tetramethylrhodamine isothiocyanate-labelled goat IgG (TRITC-gIgG) were purchased from Jackson Immunochemical Research, West Grove, PA. Mouse anti-trinitrobenzene immunoglobulin G (mouse anti-TNB IgG) was obtained from Organon Teknika-Bionetics Research, Rockville, MD. Tetramethylrhodamine cadaverine-labelled trinitrobenzene (TRC-TNB) was prepared by tetramethylrhodamine cadaverine mixing (Molecular Probes, Eugene, OR) dye with trinitrobenzene sulphonic acid (Pierce, Rockford, IL). Briefly, 100 ul of tetramethylrhodamine cadaverine (2 mg/ml) in borate buffered solution (BBS), pH 8-6 was mixed with 100 ul of trinitrobenzene sulphonic acid (3-5 mg/ml), in BBS, and the mixture was kept over night at room temperature. A silica column was run to separate TRC-TNB from free dye. Fractions of TRCTNB were eluted with chloroform (CHC&) and methanol (MEOH) (9: 1 v/v). Gamma-maleimidyl Biosensors & Bioelectronics Regeneration of immobilized antibodies on fibre optic probes butyryl succinimide (GMBS) and 3-mercaptopropyl trimethoxisilane were purchased from Fluka, Ronkonkoma, NY. Bovine serum albumin (BSA) and ethylene glycol were from Sigma, St Louis, MO. Glycine hydrochloride and ethanol were purchased from Aldrich, Milwaukee, WI and Warner Graham, Cockeysville, MD respectively. Tetraethylamine (TEA) was purchased from Fisher, Fair Lawn, NJ. Stock solutions of TRITCgIgG and TRC-TNB were prepared in PBS containing BSA (2 mg/ml) (PBS/BSA) and PBS/ BSA containing 10% (v/v) ethanol (PBS/B&A/ EtOH) respectively. Aqueous solvent systems used in the regeneration experiments had the following compositions and final pH: O-1 M glycine-HCl in 50% (v/v) ethylene glycol, pH l-75 (acidic); 0.05 M TEA in 50% (v/v) ethylene glycol, pH 11 (basic); 50% (v/v) ethanol in PBS. Preparation of optical fibres The fibre geometry (Anderson et al., 1993) and general preparation (Bhatia et al., 1989) have been described previously. Briefly, the plastic cladding was stripped from the distal 12 cm of the fibre (Quartz Products, Tuckerton, DE; 200 micron-diameter fused silica core) using a razor blade and the exposed core was tapered by slow immersion into hydrofluoric acid (33%). After tapering, the decladded portion was acid cleaned to generate the surface hydroxyl groups required for protein immobilization. The clean fibres were immersed in a 4% solution of 3-mercaptopropyltrimethoxysilane in toluene for 30 min under nitrogen, then rinsed in toluene. Next, the silanized fibres were immersed in a 2 mM solution of the heterobifunctional crosslinker GMBS for 1 h and rinsed with PBS, pH 7.4. Finally, the fibres were incubated for 1 h in a solution containing 0.05 mg/ml polyclonal rab anti-gIgG or monoclonal mouse anti-TNB IgG in PBS and rinsed with PBS several times. Antibody-coated fibres were stored in PBS at 4°C. Immediately prior to use, nonspecific binding sites on the fibre were blocked with PBS/BSA or PBS/BSA/ EtOH for the gIgG or TNB assays, respectively. Fibre optic biosensor The fluorimeter portion of the fibre optic biosensor consists of a 50 mW argon-ion laser, an off-axis parabolic mirror and spherical lens. The laser beam (514 nm) passes through the off- axis parabolic mirror, and is focused by the spherical lens onto the proximal end of the fibre. A chopper and a lock-in amplifier are used for phase-sensitive detection. The collected fluorescence signal travels back up the fibre to the parabolic mirror where it is refocused through a bandpass filter (KV 550) onto a silicon photodiode. Data was collected using a laptop computer. A detailed description of this fibre optic fluorimeter is given elsewhere (Golden et al., 1992; Shriver-Lake et al., 1992). The antibody-coated region of an optical fibre was mounted in a glass capillary tube using Tconnectors at both ends (Golden et al., 1992). The distal end of the fibre extended beyond the end of the capillary tube so that none of the distal light could illuminate the solution. The fibre was immersed in PBS or PBS/ethanol (10%) until a stable baseline reading was obtained and this background signal was recorded. A peristaltic pump was used to circulate the labelled antigen and regenerating solutions over the antibody immobilized region of the fibre at a flow rate of 0.5 ml/min. Regeneration experiments Fibres coated with rab anti-gIgG were exposed to 500 ng/ml TRITC-gIgG and the fluorescence signal recorded after 1, 2 and 3 min. The excitation light source was shuttered off except during signal measurement and regeneration. The bound TRITC-gIgG (antigen) was removed from the antibody-coated fibre by washing with either the acidic or the basic solvent for 5 min. New background signal was measured in the presence of PBSBSA before re-exposure to antigen at the same concentration (500 ng/ml TRITC-gIgG). In this way, response of the sensor at 500 ng/ml TRITC-gIgG was recorded 6 times. For fibres coated with mouse anti-TNB IgG, the response was obtained with 1 ng/ml TRC-TNB, and the antibody-coated fibres were washed using the acidic solvent, the basic solvent, or the 50% ethanol solution as described above. Fibres coated with rab anti-gIgG were calibrated using a series of TRITC-gIgG standards with concentrations ranging from 400 to 1000 ng/ ml (in PBSBSA). Standard solutions were introduced in order of increasing concentration and fluorescence signals were recorded with each solution after 1, 2 and 3 min intervals. The same fibre was recalibrated after regeneration with 587 D. Wijesuriya et al. Biosensors & Bioelectronics the acidic solvent four consecutive times. For replicate measurements, fibre-to-fibre variation was accounted for by standardizing the signals to the highest initial response. RESULTS tions ethylene glycol was added to the acidic regenerating solutions. Percentages of ethylene glycol up to 50% were tested, and found to enhance dissociation significantly (not shown). Thus, for the rest of the experiments, an acidic solution of glycine-HCl (pH l-75) with 50% ethylene glycol was used. The solution was compared to a basic solution of TEA (pH ll*O), also containing 50% ethylene glycol. Antigen bound to the immobilized rab antigIgG was removed using acidic and basic solutions. Figure 2 shows the variation of responses (signal-background) at 500 ng/ml TRITC-gIgG after regeneration with the two different solvents. After five consecutive cycles of washing and reexposure to antigen, fibres regenerated with the An earlier observation which comes from the regeneration experiments was the importance of shuttering the excitation light during the assay. If the fibre was exposed continuously to the laser, there would be increasing amounts of photobleached antigen bound to the fibre. This led to an overestimate of elution efficiency. Thus, to obtain a true picture of regeneration, the laser light was blocked except when readings were taken. In a preliminary experiment, various 1 N acid solutions not containing ethylene glycol were 100 used to regenerate rabbit anti-gIgG fibres. The strongest acids, HCl and HzS04 more effectively 60 removed the bound antigen from the antibody 4) = 60 (Fig. 1). The less acidic solutions, formic acid and phosphoric acid were less successful. To g 40 examine the importance of hydrophobic interac- zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA : 100 - ‘iii g *_ dc 20 3 .= c - 0 100 60- = 60- E Q) 0 t $ ao- t e co zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 60 n. 40 20 - 20 oHCI p” 1.05 Acidic HzSO, HP04 1.12 Regenerating HCOOH 1.17 2.36 Solution Fig. I. Regeneration of the fibre optic probe was achieved with several O-1 N acids, including hydrochloric, sulphuric, phosphoric and formic acids. Rab anti-gZgG coated probes were incubated for 5 min with TRZTC-gZgG to obtain an initial maximum signal. The bound TRZTC-gZgG was eluted with one of the acid solutions for 5 min. Then the fibre activity was again tested. Results are shown as percent of the initial fluorescent signal after elution (open bar) and percent of the initial signal after a second incubation with TRZTC-gZgG (cross hatch bar). The mean signal + S. E. (standard error) (n = 3) is shown. 588 60 0 0 No. 'of R’e&era~ions5 Fig. 2. Regeneration of fibres coated with polyclonal rab anti-goat antibody. Fibres coated with rab antigZgG were incubated in 500 nglml TRZTC-gZgG for five min. In the upper panel, fibres were washed using 0.1 M glycine HCL in 50% (vlv) ethylene glycol, pH 1.75, washed with PBSIBSA and re-incubated in 500 ngl ml TRITC-gZgG. In the lower panel, fibres were washed using 0.05 M TEA in 50% (v/v) ethylene glycol, pH 11.0, prior to re-exposure to antigen. Responses are calculated as the percentage of the initial signal and are displayed as the mean + SE. (n = 3). Biosensors & Bioelectronics Regeneration of immobilized antibodies on fibre optic probes acidic solvent retained 70% of the original signal compared to 65% with basic solvent. The response did not consistently drop as the number of regenerations increased, so it was not possible to exactly predict the response. The decrease in response was not due to a decrease in total signal but rather to an increase in background signal level, indicating a significant amount of labelled antigen left on the fibre. Fibres regenerated with 50% ethanol solution yielded very poor responses even after three or four consecutive regenerations (data not shown). The situation was somewhat different with the regeneration of fibres coated with the mouse anti-TNB IgG antibody exposed to labelled antigens. Figure 3 shows the response profile of fibres coated with mouse anti-TNB IgG to 1 ngl ml TRC-TNB after five consecutive regenerations using acidic and basic solvent systems, upper and middle panels respectively. The acidic solvent did not regenerate the mouse anti-TNB IgGcoated fibres as well as the basic solvent. After five regenerations, only about 40% of the response remained, with the response decreasing significantly with each regeneration. Fibres treated with the basic regenerating solvent retained 85% of the signal after five regeneration cycles. The decrease in response from cycle to cycle after exposure to the basic solvent were small but consistent. For the fibres coated with mouse anti-TNB IgG, the best regeneration solvent was 50% ethanol as shown in Fig. 3, lower panel. The signal remained relatively constant through five regeneration cycles. Additionally, the 50% ethaNo. of Regenerations nol system showed the smallest increase in Fig. 3. Regeneration of fibres coated with monoclonal background over five regenerations suggesting anti-TNB antibody. Fibres coated with anti-TNB antithat it most efficiently removed bound antigens. body were incubated with l-0 nglml TRITC-TNB. In Figure 4 shows the variation of analytical dose/ the experiment depicted in the upper panel, fibres were response relationships of a representative fibre washed for 5 min using 0.1 M glycine HCl in 50% (vi coated with rab anti-gIgG upon multiple exposure v) ethylene glycol, pH 1.75 prior to re-exposure to to a series of TRITC-gIgG standard solutions, antigen. In the middle panel, fibres were washed using following regeneration with acidic solvent. The 0.05 M TEA zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJ in 50% (v lv ) ethylene glycol, pH II.0 percent signal (signal-background) was plotted prior to re-exposure to antigen. In the lower panel, the against concentration of antigen. Each standard wash solution was 50% (vlv) ethanol in PBS. The curve exhibited the same general dose-response mean + SE. of the signal expressed as a percentage of the initial response (n = 3) is shown. relationship, but the signal magnitude decreased with each successive regeneration cycle. The first order slopes of dose/response curves are 1.06 DISCUSSION (R* = 0.99) O-85 (R* = O-99), 0.66 (R* = 0.98) The choice of TEA or glycine-HCl in the solvent and O-50 (R* = 0.98) (uV/(ng/ml) for fresh fibre, after first regeneration, second regeneration, and used to regenerate antibody-based detection systems is based on the assumption that electrostatic third regeneration respectively. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Biosensors & Bioelectronics Regeneration of immobilized TNB IgG so effectively. Since TEA is less polar than glycine-HCl, the increase in solubility of TNB in TEA could also be partly responsible for the better regeneration of the mouse antiTNB IgG coated fibres in the basic solvent compared to the acidic solution. However, it is difficult to distinguish whether solubilization of antigen or weakening of antibody-antigen complexation or both caused the release of antigen from the fibre surface. The amount of analyte used in the direct binding assay for TRITC-gIgG (Fig. 4) ranged from 2.5 to 6.25 picomoles (1 ml of analyte solution circulated over the fibre at 2.5 nM to 6.25 nM) compared to 420 femtomoles of antibody immobilized on the fibre. This 6- to 15fold antigen excess is sufficient to result in probes which approach saturation. Thus, our results indicate that even if the immobilized antibodies on the fibre surface are nearly saturated with excess of antigens, effective regeneration can still be achieved. Others have found regeneration to work well when the immobilized antibody is in excess in the assay (Blanchard et al., 1990), while regeneration was less successful when the number of active binding sites control the dose/response character. This does not appear to be a severe limitation for antibodies immobilized on a fibre probe, which can be regenerated and reused for at least 3 cycles (Fig. 4). Smithrud et al. has recently reported that the complexation enthalpy of host-guest complexes is more favourable in water than in methanol due to specific solvent interactions in high polar solvents (Smithrud et al., 1991). Furthermore, the nature of the complexation force strengthened with increasing solvent polarity (Smithrud et 1990). This study was conducted using al., cyclophanes and benzene derivatives as the model receptors and ligands respectively. Many biotic and abiotic complexation processes of small molecules are characterized by thermodynamic characteristics similar to those measured for the benzene complex of cyclophane. Tight binding of apolar aromatic substrates in hydrophobic pockets of enzymes and antibodies is an enthalpydriven process (Bilton et al., 1979; Ross et al., 1981). Thus, weakening of the mouse anti-TNB IgG/TNB complex can be expected in less polar ethanol compared to water because of unfavourable enthalpy conditions. Although differences were evident in the dissociation of antigen from monoclonal mouse anti- antibodies on fibre optic probes TNB and polyclonal rab anti-goat antibodies, the reuse of antibody coated fibres proved to be highly feasible for qualitative measurements using either antibody type. The release of TRC-TNB from the mouse anti-TNB IgG coated fibre was sufficiently complete and reproducible to consider possibilities for calibrating the individual probes. Reusability for many tests would enable us to do both calibration and testing using the same fibre, thus avoiding the minor differences in response due to fibre-to-fibre variation. For analysis of unknown samples, one could produce a full standard curve, regenerate the fibre, and then test an unknown. An abbreviated approach would involve measuring one or two known concentrations, washing with solvent, testing the sample or samples, and comparing the known and unknown values to a calibration curve. The dissociation of antigen from immobilized antibody facilitates reuse of the antibody-coated substrate. In the case of antibodies immobilized in biosensors, reusability significantly increases the economic feasibility of using these relatively simple, fast systems for environmental and clinical analysis. We have demonstrated that antibodies immobilized on fibre optic probes can be freed of bound antigen and reused if the appropriate antibody type and solvent system are combined. ACKNOWLEDGEMENTS The authors wish to thank Dr Linda Tempelman, Lynn Cao and Joel Golden for their helpful suggestions and technical assistance. This work was supported by the Office of Naval Research, the Naval Medical Research and Development Command, and the US Army Medical Material Development Agency. The views expressed here are those of the authors and do not represent those of the US Navy or the Department of Defense. REFERENCES Anderson, G.P., Golden, J.P. & Ligler, F.S. (1993). A fibre optic biosensor: combination tapered fibres designed for improved signal acquisition. Biosensors & Bioelectronics, 8, 249- 256. Arnold, M.A. & Meyerhoff, M.E. (1988). Fibre optic sensors: fundamentals and applications. CRC Crit. Rev. Anal. Chem., 20, 149. 591 D. W ijesuriy a et al. Biosensors & Bioelectronics zyxwvutsrqponm after exposure to very low or very high pH represented the highest affinity fraction of the polyclonal population. However, the background 100 level of antibody-bound antigen continued to ii r 60 increase with increasing regeneration cycles. .m Affinity alone is not a sufficient explanation to 80 to account for the magnitude of this increase ‘; zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA observed. The extreme acidic and basic conditions t 40 could be denaturing an increasing proportion of t a the antibody-antigen complexes as the time of 20 the exposure to harsh conditions increases and preventing antigen dissociation. This possibility 0 suggests that elution times should be relatively 400 600 600 1000 short. Incomplete antigen dissociation is seen Trite-glgG hg/mll here as a decrease in the number of free antibody Fig. 4. Dose- response curves after antigen removal. sites available for antigen binding. This results M ultiple calibration curves were generated using a fibre in a decreased rate of binding and hence lower coated with rab anti- gIgG and TRITC- gIgG antigen as signals over background upon subsequent uses. described in methods. The solution used for the removal Treating the mouse anti-TNB IgG/TNB comof bound antigen was O- 1 M gly cine HCl in 50% (vlv) plex with the same acidic and basic solvents ethy lene gly col, pH 1.75. The percentage of the initial resulted in a much greater percentage of antigen maximum response + S.E. (n = 3) over the course of dissociation. The antibody has a moderate affinity 3 regenerations is shown. The response decreased with (approximately 10m6 M-l), and so the antigen each successive regeneration. may be easier to dissociate in the short period of time the complex is exposed to solvent. Furthermore, the small cyclic antigen may be less interactions are the primary attractive forces in prone to adsorb onto the immobilized antibody or the antigen-antibody bond. In addition to using fibre surface in the solvent than a large protein acidic and basic solutions for dissociating antiantigen. Some loss of antibody activity is apparent body-antigen complexes, it has been reported with continuing exposure to solvent, but the (Lu ef al., 1992) that organic solutions such as percent decrease is relatively predictable. It is ethylene glycol reduce Van der Waals and difficult to conclude whether the type of bonding hydrogen bonding interactions. Our results coninteractions involved with this particular antibodyfirmed this: addition of ethylene glycol was found antigen complex formation is more susceptible to improve the dissociation of bound antigen. to basic conditions than acidic conditions or Therefore in the work reported here, all the denaturation of the antibody is higher with regenerating solvents contained 50% ethylene extreme acidic conditions causing decline in glycol, except for 50% ethanol solution. However, response after several regenerations. However, the types and proportion of the non-covalent it is clear that mouse anti-TNB IgG/TNB complex interactions participating at the binding site may can be dissociated more effectively with the basic vary for each antibody-antigen pair. We chose solution than the rab anti-goat IgG/gIgG complex. to examine two different types of antibodyThe affect of the antigen solubility on the antigen pairs which have been used extensively dissociation of antigen from immobilized antibody for device optimization and assay developments. was further explored using the mouse anti-TNB One pair includes a polyclonal antibody (rab antiIgG coated fibres. While methanol and ethanol gIgG) recognizing a large, hydrophillic protein solutions have been widely used to release (gIgG) antigen and the other includes a monoantigens from affinity columns, short exposures clonal antibody (mouse anti-TNB IgG) and a of the rab anti-goat antibody-protein antigen small, somewhat hydrophobic antigen (TNB). complexes to the ethanol solution removed less Neither the acidic nor the basic solvent system antigen than the acidic and basic solvents. Howwas capable of 100% removal of antigens bound ever, the antigen TRC-TNB dissolves very well to the immobilized rab anti-gIgG antibodies on in ethanol, which may explain why this wash the fibre surface. One would first assume that regenerated the fibres coated with mouse antithe antibodies that did not release antigen even I 590 I I 1 D. Wijesuriya et al. Betts, T.A., Catena, G.C, Huang, J., Litwiler, KS., Zhang, J., Zagrobelny, J. & Bright, F.V. (1991). Fibre-optic-based immunosensors for haptens. Anal. Chim. Acta, 246, 323-330. Bhatia, S.K., Shriver-Lake, L.C., Prior, K.J., Georger, R. & Ligler, J., Calvert, J.M., Bredehorst, F.S. (1989). Use of thiol-terminal silanes and heterobifunctional cross linkers for immobilization of antibodies on silica surfaces. Anal. Biochem., 178, 408-413. Biltonen, R.L. & Langerman, N. (1979). Microcalorimetry for biological chemistry: experimental design, data analysis and interpretation. Methods Enzymology, 61, 287-318. Blanchard, G.C., Taylor, C.G., Busey, B.R. & Williamson, J. (1990). Regeneration of immunosorbent surfaces used in clinical, industrial and environmental biosensors. J. Immunol. Meth., 130, 263-275. Bright, F.V., Betts, T.A. & Litwiler, K.S. (1990). Regenerable fibre-optic-based immunosensor. Anal. Chem., 62, 1065-1069. Golden, J.P., Shriver-Lake, L.C., Anderson, G.P., Thompson, R.B. & Ligler, F.S. (1992). Fluorometer and tapered fibre optic probes for sensing in the evanescent wave. Optical Eng., 31(7), 1458-1462. Golden, J.P., Anderson, G.P., Ogert, R.A., Breslin, K.A. & Ligler, F.S. (1992). An evanescent wave fibre optic biosensor: challenges for real world sensing. SPIE Proceedings, 1796, l-8. Ligler, F.S., Golden, J.P., Shriver-Lake, L.C., Ogert, R.A., Wijesuriya, D. & Anderson, G.P. (1993). Fibre-optic biosensor for the detection of hazardous materials. Immunomethods, 3, 122-127. 592 Biosensors & Bioelectronics Lu, B., Lu, C. & Wei, Y. (1992). A planar quartz waveguide immunosensor based on total internal reflection principle. Anal. Len, 25(l), l-10. Ogert, R.A., Brown, J.E., Singh, B.R., ShriverLake, L.C. & Ligler, F.S. (1992). Detection of Clostridium botulinum toxin A using a fibre opticbased biosensor. Anal. Biochem., 205, 306-312. Oroszlan, P., Thommen, C., Wehrli, M., Duveneck G. & Ehrat. M. (1993). Automated optical sensing system for biochemical assays: a Challenge for DLISA? Anal. Meth. and Znstr., l(l), 43-51. Ross, P.D. & Subramanian, S. (1981). Thermodynamics of protein association reactions: forces contributing to stability. Biochemistry, 20, 3096-3102. Shriver-Lake, L.C., Anderson, G.P., Golden, J.P. & Ligler, F.S. (1992). The effect of tapering the optical fibre on evanescent wave measurements. Anal. Lett., 25(7), 1183-1192. Smithrud, D.B. & Diederich, F. (1990). Strength of molecular complexation of apolar solutes in water and organic solvents is predictable by linear free energy relationship: a general model for solvation effects on apolar binding. J. Am. Chem. Sot., 112, 339-343. Smithrud, D.B., Wyman, T.B. & Diederich, F. (1991). Enthalpically driven cyclophane-arene inclusion complexation: Solvent-dependent calorimetric studies. J. Am. Chem. Sot., 113, 5420-5426. Wise, D.L. & Wingard, L.B. (eds.) (1991). Biosensors with Fibre Optics. Humana Press, Clifton, New Jersey, l-358. Xie, X., Suleiman, A.A. & Guibault, G.G. (1990). A urea fibre optic biosensor based on absorption measurement. Anal. Lett., 23(12), 2143-2153.