For the Detection of Microbial Antigens ROBERT H

JOURNAL OF CLINICAL MICROBIOLOGY, Mar. 1984, p. 356-360 Vol. 19, No. 3 0095-1137/84/030356-05$02.00/0 Copyright © 1984, American Society for Microbiology Enzyme Immunoassays in Which Biotinillated 1-Lactamase is Used for the Detection of Microbial Antigens ROBERT H. YOLKEN* AND SIOK-BI WEE Eudowood Division of Infectious Diseases, Department ofPediatrics, School of Medicine, The Johns Hopkins Ilospital, Baltimore, Maryland 21205 Received 24 June 1983/Accepted 23 November 1983

The performance characteristics of enzyme immunoassays are determined to a great extent by the enzyme-substrate system utilized for the immunoassay. Beta-lactamases (penicillin amido-beta-lactamhy- drolase EC 3.5.2.6) offer a number of advantages which might make them useful in immunoassay systems. We linked beta-lactamase from Bacillus cereus with biotin and used the biotinillated enzyme to devise immunoassay systems for the detection of a number of microbial antigens. An assay system in which antibodies to the polyribitol phosphate antigen of Haemophilus influenzae type b were used was capable of detecting between 0.4 and 1.6 ng of that antigen. Similarly, an assay in which antibodies to the common antigens of adenoviruses and biotin-linked beta-lactamase were used was capable of detecting between 1 and 10 50% tissue culture infective doses of a strain of enteric-type adenovirus. When applied to the detection of rotavirus, a similar system in which biotinillated beta-lactamase was used was capable of detecting small amounts of antigen in a standard rotavirus preparation. This assay could also detect virus in 36 of 37 stool specimens from children with rotavirus gastroenteritis. The positive specimens could easily be distinguished from negative ones by the naked eye, and a permanent record of the qualitative results could be obtained by the use of a standard office photocopying machine. Beta-lactamases have promise for use in practical enzyme immunoassay systems, especially in situations in which expensive colorimetric instrumentation is not available.

Enzyme immunoassays (EIAs) are attaining widespread (Biosearch Research Chemicals, San Rafael, Calif.) by the use in the detection of microbial antigens in clinical speci- method of Bayer et al. (3) with a biotin-beta-lactamase ratio mens. EIA systems utilize enzyme substrate reactions to of 2 to 1 (mg/mg). The enzyme was reacted with the biotin measure the binding of antigen to specific antibody (16, 18, ester for 3 h at 25°C, and the unreacted biotin ester was 24). The performance characteristics of EIAs are thus deter- removed by extensive dialysis. The biotinillated enzyme was mined to a great extent by the reaction kinetics of the stored at -20°C until use. enzyme substrate system used (20, 21). Enzymes which are Avidin DH was obtained from Vector Laboratories, Sun- currently in widespread use include alkaline phosphatase nyvale, Calif. Avidin from streptomyces was obtained from (24), horseradish peroxidase (1), and beta-galactosidase (19). Bethesda Research Laboratories, Gaithersburg, Md. Avidin These enzymes are used largely because they can be linked from egg white was obtained from Sigma. Biotinillated to immunoglobulin molecules to yield stable antibody-en- horseradish peroxidase was obtained from Vector Labora- zyme conjugates which retain a substantial portion of the tories. Goat antibody to rotavirus was obtained from Refer- enzymatic and antigen-binding activities of the original re- ence Reagents Branch, National Institute of Allergy and agents (2, 12). Beta-lactamases represent another group of Infectious Diseases, Bethesda, Md. Goat and rabbit antibod- enzymes which can be utilized in EIA systems (5, 7, 9). ies to adenovirus were obtained from MA Bioproducts, These enzymes, which hydrolyze beta-lactams to the corre- Walkersville, Md. Burro antibody to polyribitol phosphate sponding open-ring penicillinoic acids, have the advantage of of Haemophilus influenzae type b (PRP) was obtained from widespread availability and favorable reaction kinetics and John Robbins, Bethesda, Md. Chicken antibody to rota- have the availability of substrates without mutagenic or virus, isolated from the yolks of immunized hens, was carcinogenic potential (4, 13, 17). In addition, beta-lacta- obtained from Marc van Regenmortel, Strasbourg, France. mases can be linked to biotin, thus allowing their use in A 2% stool filtrate from a child with gastroenteritis was recently described avidin-biotin systems (6, 23). This report used as a standard rotavirus antigen. This preparation con- describes the development of solid-phase ETAs in which tains rotavirus antigen detectable at a 1:640 dilution by a biotinillated antibodies, avidin, and biotin-labeled beta-lacta- standard indirect EIA (24). A standard isolate of enteric-type mase are used for the detection of viral and bacterial adenovirus was obtained from the stool of a child with antigens. adenovirus gastroenteritis and cultivated in an adenovirus- transformed cell line (Grahm's 293 cells) as previously MATERIALS AND METHODS described (15, 22). The preparation utilized as a standard Beta-lactamase from Bacillus cereus (type I) was obtained antigen contained approximately 1,000 50% tissue culture from Sigma Chemical Co., St. Louis, Mo. The enzyme was infective doses per ml of viable virus as determined by this biotinillated by reaction with N-hydroxysuccinimide ester cultivation system. PRP was obtained from Porter Ander- son, Rochester, N.Y. Rectal swab specimens were obtained from 83 children with acute gastroenteritis. The swabs were placed into 2 ml * Correspondirig author. of phosphate-buffered saline (PBS) and stored at -20°C until 356 VOL. 19, 1984 DETECTION OF MICROBIAL ANTIGENS 357 testing. A total of 37 specimens had rotavirus detectable by a beta-lactamase assays, the results were expressed in terms standard direct EIA (24), using peroxidase-labeled rabbit of percent activity by the following formula: percent activity antirotavirus, whereas the remaining 46 specimens were = (OD spec/OD max) x 100, where OD spec is the optical negative for rotavirus by this assay. density of the specimen measured at 450 nm and OD max is Performance of immunoassays. The optimal dilutions of the maximum optical density measurable in the reaction reagents for each assay were determined by checkerboard system, generally 1.2 optical density units. The assays for titration as previously described (24). Preliminary studies enteric adenovirus and PRP were performed in a manner directed at determining the optimal substrate conditions analogous to that described above, with the exception that were performed by reacting biotinillated beta-lactamase with the reagents for these antigens were substituted for the microtiter wells coated with egg white avidin. Immunoas- rotavirus antigens. In the case of the assays for PRP, burro says for the detection of the microbial antigens were per- anti-PRP was used to coat the solid phase, and biotinillated formed in a manner analogous to that of previously de- burro anti-PRP immunoglobulin G was used as the liquid- scribed enzyme immunoassays in which biotinillated phase antibody. In the case of the assays for adenovirus, antibodies, avidin, and biotinillated enzymes were used (23). goat antiadenovirus was used as the solid-phase antibody, The assay for rotavirus is described as follows. The wells of and biotinillated rabbit antiadenovirus was used as the microtiter plates were coated with goat antirotavirus serum liquid-phase antibody. For both assays, percent activity was or control nonimmune goat serum diluted 1:10,000 in 0.06 M calculated as described above. carbonate buffer (pH 9.6). After incubation overnight at 4°C, the microtiter plates were washed five times with PBS RESULTS containing 0.05% Tween 20 (PBST), and a sample of a Preliminary studies were directed at determining the opti- standard antigen, clinical specimen, or negative control was mal parameters of the assay in which biotinillated beta- added to duplicate or quadruplicate wells. After incubation lactamase was used. These studies were performed by for 1 h at 37°C, the plates were washed, and biotin-labeled binding avidin to the microtiter plates in place of a specific chicken antirotavirus diluted 1:200 in PBST containing 0.5% antimicrobial antibody. After incubation overnight at 4°C, gelatin was added. After incubation for 1 h at 37°C, the plates the plates were washed and processed with optimally diluted were washed, and avidin DH diluted 1:500 in PBS was biotinillated beta-lactamase as described above. Initial stud- added. After incubation for 30 min at 37°C, the plates were ies revealed that a starch-iodine solution containing twice as again washed, and biotinillated beta-lactamase diluted much 12 and KI as had been previously described (13) 1:1,000 in PBST was added. After another incubation for 30 allowed a greater range of linear response and easier visual min at 37°C, the plates were washed, and starch-iodine readings in microtiter plates. Utilizing this starch-iodine substrate, modified from previously described solutions (13), solution (described above) and penicillin G at a concentra- was added. The modified substrate consisted of (milligrams per milliliter) starch, 2; penicillin G, 0.03; I2, 0.06; and KI, 1.6. The amount of decoloration of the starch-iodine sub- 140 strate due to the conversion of penicillin to penicillinoic acid was measured at a wavelength of 570 nm in a microplate 1.26 colorimeter (Dynatech MR-580). In addition, the reactions were scored with the naked eye and were recorded by 1.12 placing the microtiter plate on a standard office photocopier .98 (Xerox). To allow for the comparison of different enzyme substrate systems, the results of the assays with standard r)- .84 antigens were expressed in terms of percent activity by the a 70 following formula: percent activity = 1 - [(OD spec/OD 0 blank) x 100], where OD spec is the optical density of the .56 specimen measured at 570 nm and OD blank is the mean optical density of corresponding wells to which no antigen 42 was added. A dilution of standard antigen was considered to .28 be positive if it yielded a percent inhibition which was two standard deviations greater than the mean inhibition of .14 negative control antigens. In the case of clinical specimens, specimens were also reacted in wells coated with nonim- 0 mune serum in place of the antirotavirus antiserum to 500 50 5 .5 .05 .005 Neg control for nonspecific binding to the solid-phase antibody Ctr (24). For clinical specimens, a specific activity was calculat- Solid phase Avidin (ng/well) ed by subtracting the optical density in the wells coated with FIG. 1. Reaction of biotinillated beta-lactamase with solid-phase specific antiserum from that measured in the wells coated avidin. Avidin from egg white was bound to the wells of microtiter with nonimmune antiserum. A specimen was considered to plates at the indicated dilutions. After incubation for 14 h at 4°C, the be positive if it yielded a specific activity that was two plates were washed, and biotin-labeled beta-lactamase was added. standard deviations greater than the mean specific activity of After reaction for 1 h at 37°C, the plates were washed, and the four known amount of enzyme bound to the solid phase was quantitated by the negative specimens. addition of starch-iodine solution the indicated concen- with horseradish were containing Immunoassays peroxidase per- trations of penicillin. The amount of color remaining in the wells was formed in a manner analogous to that described above, quantitated by the measurement of optical density at a wavelength except that biotinillated horseradish peroxidase was substi- of 570 nm [OD(570)] in a microplate colorimeter. Negative control tuted for biotinillated beta-lactamase, and ortho-phenylene- indicates the mean and standard deviation of quadruplicate determi- diamine-H202 substrate (24) was used in place of the starch- nations performed in microtiter wells to which no avidin had been iodine-penicillin substrate. To allow for comparison with the added. 358 YOLKEN AND WEE J. CLIN. MICROBIOL. tion of 0.03 mg/ml, we noted specific decoloration in wells to 100 which as little as 0.05 ng of avidin had been initially added. The amount of decoloration was proportional to the amount 80 of avidin added to the wells in the range of 0.05 to 50 ng per well. The use of substrate solutions containing higher con- 60 centrations of penicillin resulted in assays which could also detect 0.05 ng of avidin. However, such substrates demon- strated lower ranges of linearity due to substantial decolora- 40 tion in blank wells to which no avidin had been added (Fig. a. 1). For the biotinillated beta-lactamase utilized, the starch- 20 iodine-penicillin substrate was substantially more sensitive than substrates utilizing chromagenic cephalosporins such as 0 nitrocefin (10) or Padac (14) (data not shown). 1:100 1:400 1:1600 1:6400 I:25650 NeW The assay system in which solid-phase burro antibody to Ctr PRP, biotinillated rabbit anti-PRP antibody, avidin DH, and Dilution of Adenovirus biotinillated beta-lactamase were used could detect between FIG. 3. Reactivity of solid-phase immunoassays for adenovirus 0.4 and 1.6 ng of PRP in an EIA system for that antigen. The antigen with biotinillated beta-lactamase (U) and biotinillated horse- sensitivity of the assay in which biotinillated beta-lactamase radish peroxidase (0). The antigen used contained approximately was used was similar to the analogous assay in which 1,000 50% tissue culture infective doses of enteric-type adenovirus biotinillated horseradish peroxidase and peroxidase-ortho- (candidate type 39) by titration in an adenovirus-transformed cell phenylenediamine substrate were used, although at many line (Grahm's 293 cells). Percent activity was calculated as de- dilutions of antigen, greater activity was noted in the beta- scribed in the text. lactamase system (Fig. 2). Avidin from streptomyces (strep- tavidin) could be used in place of avidin DH as the bridge between biotinillated antibody and biotinillated beta-lacta- The assay for human rotavirus was also used to detect mase with a similar range of reactivity. However, unmodi- rotavirus antigen in 83 rectal swab specimens obtained from fied avidin from egg white could not be utilized in this way children with acute gastroenteritis. A total of 37 of these due to a high rate of nonspecific binding to the solid phase specimens were originally positive for rotavirus in a standard (data not shown). direct EIA system. A total of 36 (97%) of these specimens Assays in which microtiter wells coated with antibody to were positive in the rotavirus BIA in which biotinillated adenovirus, biotinillated chicken antibody, avidin DH, and beta-lactamase was used. The rotavirus-negative specimens biotinillated beta-lactamase were used could detect antigen were uniformly negative in the biotinillated beta-lactamase in a 1:400 dilution of an isolate of enteric adenovirus system (Fig. 5). The one false-negative specimen was also containing approximately 1,000 50% tissue culture infectious unreactive when biotinillated horseradish peroxidase was doses of this virus. This level of sensitivity was similar to substituted for biotinillated beta-lactamase, suggesting that that of an analogous assay in which biotinillated horseradish the antigen was not reactive with the biotinillated chicken peroxidase was used (Fig. 3). Similarly, an assay in which antibody used or had lost antigen activity during storage. solid-phase goat antibody to rotavirus, biotinillated chicken The positive specimens could be clearly picked out in the antibody, avidin, and biotin-labeled beta-lactamase were microtiter plate by the naked eye, and the results could be used could specifically detect antigen in a 1:1,600 dilution of recorded a standard office a standard rotavirus antigen obtained from a child with by using photocopier (Fig. 6). actute gastroenteritis. The level of sensitivity was similar to that of an EIA system in which biotinillated peroxidase was used (Fig. 4). 100

100 80

80 :F 60

* 60 41 0 40

I 40 20 20

0 0 25 6.2 1.5 4 .1 .025 Neg 1:100 1:400 1:1600 1:6400 Neg Ctr Ctr PRP (ng/welI) Dilutions of Rotavirus Antigen FIG. 2. Comparison of EIAs for PRP. *, Results with biotinillat- FIG. 4. Reactivity of EIAs for human rotavirus in which biotinil- ed beta-lactamase; 0, results with biotinillated horseradish peroxi- lated beta-lactamase and biotinillated horseradish peroxidase were dase. Percent activity was calculated from duplicate determinations used. The concentrations shown indicate dilutions of a standard as described in the text. The values depicted for negative control preparation of rotavirus obtained from a child with rotavirus gastro- indicate the mean and standard deviation of reactivity in quadrupli- enteritis as described in the text. Percent activity was calculated as cate wells to which buffer was added in place of antigen dilution. described in the text. VOL. 19, 1984 DETECTION OF MICROBIAL ANTIGENS 359

.350 r Preliminary studies with the enzyme substrate system indicated that beta-lactamase substrates measuring the decoloration of starch-iodine solutions were most suitable for the detection of biotinillated beta-lactamase specifically bound to the wells of microtiter plates. This substrate system makes use of the fact that penicillinoic acid, the product of .300 the interaction of penicillin and beta-lactamase, oxidizes i elemental iodine to iodide ion, resulting in a decoloration of a 0 deeply blue starch-iodine solution (13). Although starch- iodine-penicillin substrates described previously could be used, we found that the modification of the substrate by the .250 p use of higher concentrations of iodine and lower concentrations of penicillin resulted in a more deeply blue solution which gave more definitive endpoints. The results could then be read with the naked eye, quantitated with a microplate colorimeter, or recorded with a standard office photocopying machine. In regard to quantitative determinations, the re- 8 .200 I- sults could be expressed in terms of the percent decrease in color compared with negative controls. However, in the case 4- of clinical specimens, a specific activity could also be calculated by subtracting the optical density of wells coated .2 with specific antiserum from that of wells coated with &.150 nonimmune serum. This determination allows for the distin- guishing of specific from nonspecific reactions in a manner analogous to that of other EIA systems (24). Utilizing this substrate, biotinillated beta-lactamase could be specifically detected bound to microtiter wells to which as little as 0.05 ng of avidin had been added. Since, under the .100 - conditions and protein concentrations used, more than 80% of the added proteins would be expected to bind to the microtiter plate (8, 11), this level represents the approximate theoretical sensitivity of avidin-biotin systems in which these reagents are used. When applied to microbial antigens, .050S assay systems in which biotinillated antibodies, avidin DH,

.025 <.010 A 000 *000 oo loo 00 0 .**Go Robvirus Negotive Rotovirus Psitive B FIG. 5. Results of testing of rectal swab specimens from children with gastroenteritis in the solid-phase EIA system in which biotinil- C was used. was calculated as lated beta-lactamase Specific activity D described in the text. Rotavirus presence was initially determined by a standard direct EIA with peroxidase-labeled antirotavirus antibody. The horizontal lines indicate a level of specific activity two E standard deviations above the mean specific activity of four specimens known not to contain rotavirus. F

DISCUSSION H The use of beta-lactamase as the enzymatic marker in EIA a beta- systems offers number of advantages. Although FIG. 6. Photocopy of a microtiter plate in which rectal swab lactamase can be directly linked to immunoglobulin by specimens were tested for rotavirus. The assays were performed reaction with cross-linking agents such as glutaraldehyde, utilizing biotin-linked beta-lactamase and starch-iodine-penicillin we elected to develop systems in which biotinillated antibod- substrate as described in the text. The wells in the odd numbered ies and biotinillated beta-lactamase linked with an avidin rows were coated with antirotavirus antibody, and the rows in the bridge were used. This method has the advantage that a even numbered wells were coated with nonimmune serum. Speci- single preparation of biotinillated enzyme can be used for the mens were added to duplicate wells coated with immune serum and coated with nonimmune serum. The was detection of a wide range of antigens, provided that biotinil- duplicate wells photocopy The prepared by placing the microtiter plate on the glass surface of the lated antibody to these antigens is available (6). high photocopier and using the standard settings. Clear wells (Al, A3, affinity of avidin for biotin ensures that the avidin-biotin Bl, B3, B5, B7, D5, D7, Fl, F3, Gi, G3) are ones coated with interactions will go to completion in a short period of time. specific antirotavirus antibody to which rotavirus-positive stools In addition, the fact that each molecule of avidin can react had been added. The corresponding wells in the even-numbered with up to four molecules of biotin offers the potential of rows (A2, A4, B2, etc.) are ones coated with nonimmune serum to magnification at the level of avidin-biotin interaction (3, 23). which the same specimens had been added. 360 YOLKEN AND WEE J. CLIN. MICROBIOL.

and biotinillated beta-lactamase were used could detect as de Gruyter, Inc., Berlin. little as 0.4 ng of purified PRP. Similarly, small amounts of 8. Lehtonen, 0. P., and M. K. Viljanen. 1980. Antigen attachment standard preparations of rotavirus and adenovirus prepara- in ELISA. J. Immunol. Methods 34:61-70. tions could be specifically detected in the EIA systems. The 9. Mitchell, G. F., R. R. Premier, E. G. Garcia, J. G. Hurrell, detection limits were equivalent to those found when biotin- H. M. Chandler, K. M. Cruise, F. P. Tapales, and W. U. Tiu. 1983. Hybridoma antibody-based competitive ELISA in Schis- illated peroxidase was used in place of the biotinillated beta- tosoma japonicum infection. Am. J. Trop. Med. Hyg. 32:114- lactamase. The assay system for rotavirus was also used to 117. detect antigen in 36 of 37 rectal swab specimens obtained 10. O'Callaghan, C. H., A. Morris, S. M. Kirby, and A. H. Shingler. from children who had gastroenteritis and were originally 1972. Novel method for detection of 3-lactamases by using a positive for rotavirus antigen in a direct EIA. No false- chromogenic cephalosporin substrate. Antimicrob. Agents Che- positive reactions were noted. The decoloration in the wells mother. 1:283-288. coated with specific antirotavirus antibody to which positive 11. Pesce, A. J., D. J. Ford, M. Gaizutis, and V. E. Pollak. 1977. specimens had been added could be easily seen with the Binding of protein to polystyrene in solid-phase immunoassays. naked eye. In addition, the results could be recorded use Biochim. Biophys. Acta 492:399-407. by 12. Pesce, A. J., R. R. Modesto, D. J. Ford, K. Sethi, D. N. Clyne, of an office photocopier, thus providing a permanent record and V. E. Pollak. 1976. Preparation and analysis of peroxidase of the results (Fig. 6). The assay system in which beta- antibody and alkaline phosphatase antibody conjugates, p. 7-23. lactamase is used might thus be particularly useful for EIAs In G. Feldmann, P. Druett, J. Bignon, and S. Avrameas (ed.), in locations removed from central laboratories, since sophis- Immunoenzymatic techniques. North-Holland Publishing Co., ticated colorimetric instrumentation would not be available. Amsterdam. The further development of EIAs in which beta-lactamase is 13. Ross, G. W., and C. H. O'Callaghan. 1975. 1-Lactamase assays. used might expand the scope of EIAs and allow for their Methods Enzymol. 113:68-79. performance under a wide range of investigative and clinical 14. Schindle, P., and G. Huber. 1980. Use of PADAC, a novel conditions. chromagenic beta-lactamase substrate, for the detection of beta- lactamase producing organisms and assay of beta-lactamase ACKNOWLEDGMENTS inhibitors, p. 169-176. In U. Bradbeck (ed.), Enzyme inhibitors. Verlag Chemie GmbH, Weinheim, Federal Republic of Germa- This work was supported by National Institutes of Health con- ny. tract NO1-AI-22680 and Public Health Service grant 1-ROl-AI- 15. Takiff, H. E., S. E. Straus, and C. E. Garon. Propagation and in 17604 from the National Institute of Allergy and Infectious Diseases, vitro studies of previously non-cultivatable enteral adenoviruses by the Thrasher Research Fund, Salt Lake City, Utah, and by the in 293 cells. Lancet ii:832. Lauren Jill Roth memorial fund. 16. Voller, A., A. Bartlett, and D. E. Bidwell. 1976. Enzyme immunoassays for parasitic diseases. Trans. R. Soc. Trop. Med. Hyg. LITERATURE CITED 70:98-106. 17. Voogd, C. E., J. J. Van det Stel, and J. A. Jacobs. 1980. On the 1. Avrameas, S., and T. Ternynck. 1969. Peroxidase-labelled anti- mutagenic action of some enzyme immunoassay substrates. J. body and Fab conjugates with enhanced intracellular penetra- Immunol. Methods 36:55-61. tion (letter). Immunochemistry 8:1175-1179. 18. Waart, M. R. D., A. Snelting, J. Cichy, G. Wolters, and A. 2. Avrameas, S., T. Ternynck, and J. L. Guesdon. 1978. Coupling Schuurs. 1978. Enzyme-immunoassay in the diagnosis of hepati- of enzymes to antibodies and antigens. Scand. J. Immunol. tis with emphasis on the detection of "e" antigen (HbeAg). J. 8(Suppl. 7):7-23. Med. Virol. 3:43-49. 3. Bayer, E. A., M. Wilchek, and E. Skutelsky. 1976. Affinity cyto- 19. Wantanabe, N., Y. Niitsu, S. Ohtsuka, J. Koseki, Y. Kohgo, I. chemistry. The localization of lectin and antibody receptors on Urushizaki, K. Kato, and E. Ishikawa. 1979. Enzyme immunoas- erythrocytes via the avidin-biotin complex. FEBS Lett. 68:240- say for human ferritin. Clin. Chem. 25:80-82. 244. 20. Wisdom, G. B. 1976. Enzyme immunoassay. Clin. Chem. 4. Citri, N. 1971. Penicillinase and other 1-lactamases, p. 23-46. In 22:1243-1255. P. D. Boyer (ed.), The enzymes, vol. 4. Academic Press, Inc., 21. Yolken, R. H. 1982. Enzyme immunoassays for the detection of New York. infectious antigens in body fluids: current limitations and future 5. Geetha, P. B., A. A. Koshy, C. N. Dandawate, B. J. Shaikh, prospects. Rev. Infect. Dis. 4:36-68. S. N. Ghosh, and P. S. Borkar. 1982. Detection of hepatitis B 22. Yolken, R. H., F. Lawrence, F. Leister, H. E. Takiff, and S. E. surface antigen by ELISA technique using penicillinase (beta- Strauss. 1982. Gastroenteritis associated with enteric type ade- lactamase) labelled caprine antibody to HbsAg. Hind. Antibiot. novirus in hospitalized infants. J. Pediatr. 10:21-26. Bull. 24:34-37. 23. Yolken, R. H., F. J. Leister, L. S. Whitcomb, and M. Santo- 6. Guesdon, J. L., T. Ternynck, and S. Avrameas. 1979. The use of sham. 1983. Enzyme immunoassays for the detection of bacteri- avidin-biotin interaction in immunoenzymatic techniques. J. al antigens utilizing biotin-labelled antibody and peroxidase Histochem. Cytochem. 27:1131-1139. biotin-avidin complex. J. Immunol. Methods 56:319-327. 7. Joshi, U., V. Rashavan, G. Zemse, A. Sheth, P. S. Borkar, and S. 24. Yolken, R. H., P. J. Stopa, and C. C. Harris. 1980. Enzyme Ramachandran. 1978. Use of penicillinase in the estimation of immunoassay for the detection of rotavirus antigen and anti- human chorionic gondadotropin and human placental lactogen body, p. 692-699. In N. Rose and H. Friedman (ed.), Manual of by enzyme-linked immunoassays, p. 233-245. In S. B. Pal (ed.), clinical immunology, 2nd ed. American Society for Microbiolo- Enzyme labelled immunoassay of hormones and drugs. Walter gy, Washington, D.C.