Analytical Communications, January 1996, Vol33 (21-22) 21 Rapid Fluorescence Enzyme Linked lmmunosorbent Assay for Subtilisin Ioana Nitescuayb Frederick J. Rowell" and Robert H. Cummingb a,h North East Biotechnology Centre, School of Health Sciences, University of Sunderland, Sunderland, UK SR2 7EE University of Teesside, Middlesbrough, Cleveland, UK TSI 3BA North East Biotechnology Centre, School of Science and Technology, A simple and rapid ELISA with a fluorescent end-point has been developed for the measurement of subtilisin.The ELISA can be applied to the analysis of enzyme in solutions that have been eluted from filters following the capture of air samples in the workplace. The assay has adequate precision and shows no interference from other enzymes tested.A comparison has been made between this assay and a similar ELISA method using spectrophotometric detection. The use of fluorescence detection reduces the over-all assay time from 60 to 25 min and reduces the concentrations of antibody reagents required in the assay. The assay described is quicker and more sensitive than the current methods in use. During the production of many biologically active products, there is a chance that plant operators will be exposed to the product in the form of an aerosol or dust particles.Such exposure can lead to plant personnel developing allergies to the product. Subtilisin is an example of a typical serine protease enzyme used in the detergent industry. In common with other proteases it is a respiratory sensitizer and long- and short-term occupational exposure limits of 60 ng 117-3 for periods of 8 h and 15 min, respectively, have been set for workers who may inhale these proteases in the work place.2 At present, factory air is passed through a filter in a Galley sampler operating at 650 1 min-1 and the dust collected on the filter is eluted and assayed for the protease.There are two methods currently employed to monitor protease in the collected dust.Both of them are based on detection of the enzyme by measuring its catalytic activity .3,4 Measurements performed with a Casella Model A personal monitor (SKC, Blandford Forum, Dorset, UK), operating at 2 1 min-1, gave higher results than for the air samples monitored with Galley samplers, and owing to its lower sampling rate, half the samples collected were not detected by the enzymic method used.4 The use of the Casella device as a personal sampler offers the advantage of giving a better indication of individual worker exposure and hence development of more sensitive analytical methods are required to enable its use in this context.An extremely sensitive assay based on enzyme activity, which uses p-nitroanilide substrates has been proposed by the detergent industries, but the assay is not rapid and requires expensive synthetic substrate^.^ Ideally the assay should also be simple and sufficiently rapid to enable analysis of material eluted from filters to be .performed immediately at the site of sampling rather than the current post- sampling analysis which takes place at remote laboratories.We report on the use of fluorescence detection in a subtilisin- specific ELISA and its impact on the assay's performance compared with the use of a spectrophotometric end-point in this assay. Experimental Apparatus The assays were performed using black Dynatech (Billinghurst, West Sussex, UK) Immulon 4 microtitre plates. Absorbance measurements were recorded on a Dynatech MR 7000 microtitre- plate reader at 405 nm.Fluorescence measurements were recorded on a Dynatech Fluorolite 1000 fluorescence microtitre- plate reader at A,, = 340-360 nm and A, = 440-460 nm. Reagents All reagents were purchased from Sigma (Poole, Dorset, UK) unless otherwise stated. The buffer salts were of analytical- reagent grade (Merck, Poole, Dorset, UK). Polyclonal anti- bodies were generated in rabbits by using Subtilisin Carlsberg type 8 : bacterial.6 The antisera from rabbits A and B were mixed (1 + 1 v/v), affinity-purified and used for assay development .6 Coating buffer, pH 9.6.This was prepared from 2.93 g of sodium hydrogen carbonate, 2.5 g of sodium carbonate and 0.2 g of sodium azide, made up to 1 1 with distilled water.Substrate buffer (PNPP), pH 9.8. This was prepared from 105 g of diethanolamine, 0.1 g of magnesium chloride made up to 1 1 with distilled water. Phosphate buffered saline with 0.05% Tween 20 (PBST), pH 7.4 (assay buffer). This was prepared from 36 g of sodium chloride, 2.15 g of potassium dihydrogen phosphate, 7.4 g of disodium hydrogen phosphate, 2.5 ml of Tween 20, and 0.5 g of sodium azide, made up to 5 1 with distilled water.TRZS buffer, pH 7.0. This was prepared from 0.05 mol 1-' TRIS-HC1, 0.01 mol -1 calcium chloride, made up to 1 1 with distilled water. Ethanolamine bufSer, pH 8.6. This was prepared from 12.2 g of ethanolamine, made up to 1 1 with distilled water. Optimization of the Fluorescence Assay Wells of the black Immulon 4 microtitre plates were coated by passive adsorption with solutions of subtilisin in coating buffer (pH 9.6) over the range 1-10 pg ml-1.After standing at 4 "C overnight, the plates were washed three times with distilled water, dried at 35 "C and immediately sealed and stored, at room temperature, in the dark until used. The optimum concentration was 10 pg ml-1. The optimum dilution of affinity-purified rabbit antiserum and goat anti-rabbit IgG labelled with alkaline phosphatase was determined by in- cubating dilutions of 1 : 100-1 : 1000 of the former and 1 : 100-1 : 2500 of the latter.Respective dilutions of 1 : 1000 and 1 : 2500 produced assays of the maximum sensitivity. ELISA for Subtilisin Using Fluorescence Detection Equal volumes of 1 : 1000 affinity-purified antiserum and 1 : 2500 goat anti-rabbit IgG labelled with alkaline phosphatase, both diluted in PBST were incubated for at least 1 h at room temperature prior to the assay.Aliquots of sample in PBST (50 1-11), or freshly prepared subtilisin standards ranging from 0.1 pg ml-l to 100 pg ml-l (prepared in 50 pl of PBST) were added to the wells of the coated plates, followed by the pre-incubated antibodies (100 pl).After incubation for 15 min, the wells were washed three times with distilled water. Finally, aliquots of 4-methylumbelliferyl phosphate (50 1-11 of 0.2 g 1-1 solution in substrate buffer, pH 9.8) were added to each well and the fluorescence was measured after 10 min. The standard curve for22 Analytical Communications, January 1996, Vol 33 fluorescence detection was steepest at 10 ng ml-1-5 pg ml-1 of subtilisin corresponding to a drop of 1500 fluoresence units.ELISA for Subtilisin Using Spectrophotometic Detection Equal volumes of 1 : 100 affinity-purified rabbit antiserum and 1 : 250 goat anti-rabbit IgG labelled with alkaline phosphatase, both diluted in PBST were pre-incubated for at least 1 h at room temperature prior to the assay.Aliquots of sample in PBST (50 pl) of freshly prepared standards of subtilisin ranging from 0.1 pg ml-l to 100 pg ml-1 (prepared in 50 p1 of PBST) were added to the wells of the coated plates followed by the pre-incubated antibodies (100 pl). After incubation for 30 min, wells were washed three times with excess amounts of distilled water. Finally aliquots of 4-nitrophenyl phosphate (50 pl of a 2 g 1-1 solution in substrate buffer) were added to each well and the absorbance measured at 405 nm, when a pronounced yellow colour developed at room temperature.The resulting standard curve for a development time of 30 min is shown in Fig. 1, which also shows the corresponding standard curve for the fluorescence assay. The steepest part of the former standard curve was between 10 ng ml-1 and 5 pg ml-1 of subtilisin, corresponding to a fall of 0.5 absorbance units.Precision and Sensitivity Within-assay precision (s,) for both assays was determined by performing six assays with a subtilisin sample of 10, 100 and 1000 ng ml-1 on one plate. Between-assay s, was assessed using the 10 ng ml-1 sample in six separate assays. Both ELISAs were performed with subtilisin standards and six sample blanks (assay buffer).The mean and standard deviation (s) of the blanks were used to determine the LOD of subtilisin from the standard curve at the 95% confidence level (i.e., the concentration of subtilisin corresponding to zero fluorescence or absorbance minus 2s of the zero). Specificity Stock solution of active subtilisin in PBST buffer (100 pg ml- I ) was inactivated by overnight incubation at 80 "C and then used in the rapid fluorescence and spectrophotometric assays as described above.The assay was also performed in the presence of esperase, collagenase and a-amylase, over the concentration range 10.0 ng ml-1-100 pg ml-1. The assay was also run with subtilisin standards in an assay buffer containing 1% non- enzyme detergent.Results and Discussion Precision and Sensitivity For the fluorescence assay, the within-assay s, was 6.67,7.92 and 5.95% for 10,100 and 1000 ng ml- 1 subtilisin, respectively. The between-assay s, was 6.43% for 10 ng ml-1 subtilisin. For the spectrophotometric assay, the within-assay s, was 3.35, 1.8 and 7.42% for 10, 100 and 1000 ng ml-1 subtilisin, respectively.The between-assay s, was 13% for 10 ng ml- *. Spe ciji c ity No significant interference was detected with inactivated subtilisin in either the fluorescent or the spectrophotometric assays. A possible explanation is that proteases are subject to autodegradation and molecules which are partially unfolded due to heat treatment are vulnerable to autolysis.7 No significant cross-reaction was observed for esperase, collagenase and amylase in either assay.The presence of 0.1% non-enzyme detergent in the assay buffer did not significantly change the shape of the subtilisin standard curve for either assay. Conclusions Although both assays displayed the same LOD, a high background signal is observed in the fluorescence assay due to The LOD of subtilisin was 1 pg ml-I for both assays.100 I 90 80 70 6 60 2 50 .p 40 (I) 30 20 10 0 h - 0 0.001 0.1 10 1000 100000 Su bti I is in con cent rat ionhg ml-' Fig. 1 trophotometric (.) detection. Comparison of rapid ELISA using fluorescence (0) and spec- the instability of the substrate (4-methylumbellifery1 phos- phate). Thus, the percentage displacement from the blank signal is reduced in comparison with spectrophotometric detection.The sensitivity of the fluorescence assay at 50% signal displacement was better than the sensitivity of the spectrophoto- metric assay (see Fig. 1). The over-all time for the assay is improved by using fluorescence detection instead of spec- trophotometric detection from 60 to 25 min. The current chromogenic assays,s,' which take 20 min per sample are therefore considerably slower than either of these ELISA methods, especially since use of 96-well microtitre plates enables batch processing of up to 40 samples in duplicate per plate.The amount of affinity-purified antibodies required to prepare the single reagent for the fluorescence assay is one tenth of the quantity used in the spectrophotometric assay, making the assay more cost-effective to operate.Similar results in terms of precision and specificity were obtained. Heat-inactivated sub- tilisin did not cross-react in either assay. The relative potencies of active and inactive forms of subtilisin as respiratory sensitizing agents are currently unknown. The sensitivity and precision of both ELISA methods combined with their speed commend them for use in monitoring the workplace atmosphere as part of an occupational health and safety monitoring programme.The authors thank the DTI Chemical and Biotechnological Division for financial support, and Dynatech Laboratories for the use of the Fluorolite fluorescence plate reader. References Koochaki, Z., Cumming, R. H., Rowell, F. J., and Stewart, I. W., Process Biochem., 1995, 30, 589. Health and Safety Executive, Occupational Exposure Limits, 1994, EH 40/94, HMSO, London. Dunn, E., and Brotherton, R., Analyst, 1971, 96, 159. Bruce, C. F., Dunn, E., Brotherton, R., Davies, D. R., Hall, F., and Potts, M., Ann. Occup. Hyg., 1978, 21, 1. Rothgeb, T. M., Goodlander, B. D., Garrison, P. H., and Smith, L. A., J. Am. Oil Chem. Soc., 1988,65, 806. Rowell, F. J., Cumming, R. H., and Nitescu, I., Anal. Chim. Actu, 1995, in the press. Gallagher, T., Bryan, P., and Gilliland, G., Proteins: Structure, Function, and Genetics, 1993, 16, 205. The Standing Committee on Enzymatic Washing Products, Fifth Report, 1991, The Soap and Detergent Industry Association, Cumming, R. H., Rowell, F. J., and Nitescu, I., Proceedings of the Eighth Conference of the Aerosol Society, July 1994, York. p. 1-65. Paper 5107452E Received November 14, I995 Accepted December 13, 1995