Vol. 10, No. 3

JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1979, p. 334-338 0095-1 137/79/09-0334/05$02.00/0

Quantitative Assay of Soluble Beta-Hemolytic Streptococcal Antigens via an Immunochemical Turbidimetric Method with a Spectrophotometer STANLEY S. LEVINSON

Brookline Hospital, Brookline, Massachusetts 02146 Received for publication 19 June 1979

Soluble, group A, B, C, and G beta-hemolytic streptococcal antigens were successfully identified in a prototype spectrophotometric system by an immunochemical turbidimetric assay. Any spectrophotometric system which can take a zero reading followed by a second reading 2 or more min later can be used for the assay. Maximum absorbance was obtained near a wavelength of 340 nm. A wide range of linearity between antigen concentration and absorbance was observed at some antibody dilutions, resulting in a simple assay which can be used to quantitate amounts of antigen in solution. Minimal cross-reactions that present no problem in interpretation were observed. Simulated emergency samples were solubilized and assayed for group A and B bacteria within 3 h of colony recognition. Reproducibility of the absorbance resulting from the antibody-antigen reaction was great, with low coefficients of variation over a period of 50 days. The simplicity of the assay solutions, requiring only antisera and a buffer, and the accessibility to high levels of quality control are among the greatest assets of the technique to clinical laboratories.

Serological grouping of beta-hemolytic streptococci is performed by a variety of techniques, including the precipitin test (7), the fluorescent technique (10), and agglutination procedures (2, 3, 5, 8). These techniques require final identification by visual interpretations. Although all can provide excellent results when good-quality commercial antiserum is available, poor antisera can cause weak reactions or cross-reactions and nonspecific precipitation, leading to incorrect interpretations. Incorrect interpretations may also stem from the use of previously good antiserum which has lost activity during storage, from inadequate solubilization of streptococcal antigen, and from antigen or antiserum excess. These problems are difficult to eliminate when the sole criterion upon which the quality assessment relies is positive or negative results identified by visual interpretation. Use of instruments to monitor the precipitin reactions could allow more quantitative measurement of standards and blanks, thereby providing better criteria for standardizing both solubilizing enzyme and antiserum production by commercial sources and improved quality assessment in clinical laboratories for routine solubilization and assay. Automation of instrumental analysis could eliminate many of the tedious technical manipulations associated with microscopic examinations, capillary tube filling, and

rocking and interpretating slides. Automation could thereby provide results as rapidly as, but more efficiently and with better controls than, most other techniques. In this report, data are presented from a prototype system which indicate that a spectrophotometer can be used to quantitatively monitor the precipitin reaction between antiserum and soluble bacterial antigen. The technique uses an immunochemical turbidimetric assay similar to that used for measuring amounts of serum proteins by automated spectrophotometric equipment (1, 4). Group A, B, C, and G beta-hemolytic streptococci were examined. The data indicate that within 3 h after colony recognition, group A and B antigens could be solubilized and identified. All groups were identified within 2 h after overnight incubation.

MATERILS AND METHODS Preparation of bacteria. Beta-hemolytic streptococci were obtained from random outpatient specimens. After recognition on blood agar plates, colonies were transferred by a sterile loop to Todd-Hewitt (TH) broth. Cells were grown at 35°C either to a density of nearly 0.5 absorbance unit when measured in a cuvette (75 by 12 mm) at 500 nm with a Coleman Junior II spectrophotometer (The Perkin-Elmer Corp. Norwalk, Conn.) against a T-H blank or for periods of time indicated below. A 2.5-ml amount of T-H broth containing the cells was centrifuged for 15 min, the 334

VOL. 10, 1979

335

SOLUBLE BETA-HEMOLYTIC STREPTOCOCCAL ANTIGENS

supernatant was discarded, and the pellet was used to prepare streptococcal antigens. Preparation of antigens. Streptococcal antigens were solubilized by incubating the suspended pellet in 0.25 ml of Streptomyces albus enzyme at 450C for 1 h (9). S. albus enzyme was purchased from Difco Laboratories, Detroit, Mich. (lot 651004), and reconstituted according to the instructions of the manufacturer. After the incubation, the mixture was centrifuged for 15 min, and the tan-colored supernatant was used to determine group antigens. When not used immediately, the supernatant extracts were stored at -20°C. Assay of antigens. Depending on the experiment, 5 to 200 pl of extract was added to buffer in a test tube. The solution was warmed at 37°C, and 25 to 100 Ad of known antiserum was added to the solution. The final volume of 1.5 ml was mixed by inversion. The mixture was immediately inoculated into a spectrophotometric system, which included a Perkin-Elmer spectrophotometer and a Coleman 47 BCD printer (all from The Perkin-Elmer Corp.). The absorbance was monitored at 15-s intervals for 15 min at 340 nm in a 1-cm light path at 37°C. The buffer contained, per liter, 40 g of polyethylene glycol, 21 g of sodium fluoride, and 29 g of sodium chloride, as previously described (4), as well as 0.05 mol of tris(hydroxymethyl)aminomethane per liter adjusted to a pH of 7.4 with HCI. Antisera were purchased from Wellcome Research Laboratories, Beckenham, England (group A, lot K4006; group C, lot K3902; and group G, lot K4077) and from Difco Laboratories (group B, lot 649027). All groups were confirmed by the precipitation test method (7) performed on overnight bacterial growth.

RESULTS The results of all experiments are expressed as the absorbance at time t subtracted from the absorbance at zero time. The data shown in Fig. 1 through 4 and Table 1 were obtained by using antigens from bacteria which were grown in TH broth to a concentration of 0.5 absorbance unit per 2.5 ml as described above. Unless otherwise indicated, the data are from antisera against antigens for group A beta-hemolytic streptococci. Figure 1 shows the kinetic relationship at 15s intervals between group A antigen (solid circles) and antibody in the test system as manifested by absorbance (reflecting turbidity). Increasing the amount of antigen caused an increase in the absorbance. The slope of each curve represents the rate of reaction. The rate was never constant. It began rapidly and decreased toward an endpoint which was reached by about 10 min. Nevertheless, the absorbance at any time greater than 2 min substracted from the absorbance at zero time was linearly related to antigen concentration over a wide range of absorbances at an antiserum dilution of 1:15. The solid lines in Fig. 2 indicate this linear relationship for assays performed at t = 2, 5, and

0

0)2

2

4

8 lo 12 6 Time (minutes)

14

16

FIG. 1. Absorbance versus timne for antigen reaction with group A antisera. Symbols: *, antigen dilutions of 1: 15, 1:30, 1:60, 1:150, and 1:300 firom top to bottom, respectively, for group A; A\ and O, group B and Staphylococcus epidermidis antigen, respectively, in a 1:15 dilution. A 1:15 dilution of antiserum was used.

15 min. Equations describing the solid lines are given in the legend to Fig. 2. Shortening the timne

interval over which the reaction was measured reduced the sensitivity, although the reproducibility remained about the same. The relationships between antigen concentration and antibody concentration are shown in Fig. 3. The reaction could be monitored over a wide range of each. Each point represents the absorbance at t = 15 min. A wide range of linearity between antigen and absorbance occurred above an antiserum dilution of 1:20. Reproducibility of the absorbance resulting from the antigen-antibody reaction for two antigen concentrations over a period of 50 days is shown in Fig. 4. The small coefficients of variation (2.3 and 3.5%) indicated little loss of activity with tixne. This provided an excellent means for controlling the quality of the assay. Table 1 indicates that streptococci cross-react minimally with antisera for other groups. The ma0i.um cross-reaction was 0.005 absorbance unit, occurring with group G antigen and antisera against group B. An experiment was performined to determine the mionimal growth period of bacteria needed to provide enough antigen to secure a response

336

LEVINSON

J. CLIN. MICROBIOL.

U)

0.2

a)

L) c

0.2

.0 o In .0

0 0

150

60

30

15

Antigen

I_

-

300 150 100 75 0.2 0.4 0.6 0.8

60 50 1 1.2

FIG. 3. Absorbance versus antigen-antibody concentration for group A antigen and antiserum. Assay was 15 min. Symbols and antiserum dilutions were as follows: O, 1:15; A, 1:20; *, 1:30; A, 1:60.

37 1.4

Antigen FIG. 2. Absorbance versus group A antigen concentration when reacted with specific antiserum. Antiserum dilution was 1:15. Symbols: 0, 15-min assay; 0, 5-min assay; A, 2-min assay. The solid lines are given by the following equations: y = (0.201 ± 0.005)x -0.015, SDy = 0.005, r = 0.996, n= 14 (0); y = (0.168 ± 0.004)x - 0.022, SDy = 0.004, r= 0.996, n = 14 (0); y= (0.121 ± 0.004)x-0.022, SDy =0.004, r= 0.994, n = 14 (A). Each point was performed in duplicate, with the range given by the bars. SDy, Standard deviation of y values about the line.

from the spectrophotometer. The experiment was designed to simulate severe emergency conditions, where only several streptococcal colonies were available on a plate containing mixed flora. Bacteria from six colonies were aseptically transferred to 2.5 ml of T-H broth. It is important to note that to avoid contamination, whole colonies were not taken; rather, the top surface of each colony was touched with the inoculating loop. Table 2 shows the results. In the assay, antigens obtained from these bacteria caused easily recognizable absorbance changes. The total time from colony recognition to group identification was the incubation time plus 2 h (1 h of solubilization with S. albus enzyme and 1 h for centrifugation and assay). The wide variation in absorbances seen with 1- and 2-h incubations was probably due to differences in the amount of inoculation. These differences became less

S

0.3F '\\

\a U) 0

.0

6

0.2

40

0.Q

@ I_

10

20

I-

I

I

30

40

50

Day

FIG. 4. Reproducibility of absorbance by group A antigen and antiserum reaction versus time in days. Antigen dilutions were as follows: (upper) 1:15, mean = 0.349, C, = 2.3%; (lower) 1:60, mean = 0.179, C,, = 3.5%. The antiserum dilution was 1:15. Each assay was performed at t = 15 min on different days. C,, Coefficient of variation.

with longer periods of incubation, as the numbers of bacteria began to plateau. Assay of the antigen-antibody reaction at wavelengths other than 340 nm indicated that the absorbance at 400 nm was approximately

VOL. 10, 1979

SOLUBLE BETA-HEMOLYTIC STREPTOCOCCAL ANTIGENS

TABLE 1. Assaya of streptococcal antigens by group-specific antisera Absorbance with following antigen: Antiserum C A B G A 0.282 -0.001 0.003 0.000 B 0.001 0.196 0.002 0.005 C 0.001 0.000 0.120 0.002 G 0.000 0.000 0.000 0.110 a Assay was for 15 min, using an antiserum dilution of 1:15 and an antigen dilution of 1:30.

TABLE 2. Grouping of streptococcal antigens extracted by S. albus enzyme after growth periods of 1, 2, 3, and 4 h in T-H brotha Absorbance Time A antigen and A antise- B antigen and B antiserum rum (h) Avg Range No. Avg Range No. 0.082 0.033-0.16 3 0.037 0.015-0.076 3 0.067 0.042-0.11 3 0.034 0.015-0.072 3 0.225 0.19-0.26 2 0.16 0.087-0.24 2 0.160 0.14-0.18 2 0.21 0.19-0.21 2 a Assay was for 15 min, using an antiserum dilution of 1:15 and an antigen dilution of 1:7.5. Bacteria were secured from the number of random patients indi1 2 3 4

cated.

61% of that at 340 nm, whereas at 500 nm it was 39%.

DISCUSSION The results of these experiments indicate that this type of approach for identifying streptococci can be an asset to both clinical laboratories and manufacturers of items for use in such laboratories. Any spectrophotometric system which can measure absorbance near 340 nm and provide a zero reading followed by a second reading can be used. The linear portion of the absorbance-antigen concentration curve (Fig. 2) allows quantitative estimates of the amount of unknown group A antigen, using a simple two-point standard curve. Similar types of relationships have been found for group B, C, and G antigens, using an automated miniature centrifugal ana-

lyzer. These data introduce a means whereby manufacturers of S. albus enzyme and antiserum against soluble bacterial antigens can provide quantitative information relative to specific activity for their products. In clinical laboratories the quantitative and reproducible (Fig. 4) nature of the assay provides a simple means for improved quality control as compared with that using visually identifiable techniques, where a positive or negative re-

337

sponse is the sole control. In this assay, soluble antigen standards, which should fluctuate randomly within 2 standard deviations of an established mean, can be assayed along with the samples to assess the activity of the antisera, and S. albus enzyme quality may be evaluated daily in the same manner by including in the assay a control from a store of frozen cells digested along with the samples. Other advantages of the method include the following: (i) simplicity of assay solutions, requiring only a buffer and sample of each antiserum; (ii) no need for solidphase or other elaborate material coated with antibody; (iii) long-lived reagents without need for isotope; (iv) possible extension to other soluble bacterial antigens; (v) omission of need for neutralization step after solubilization of antigens with acids or bases because of the buffering capacity of the test mixture; (vi) easy adaptation to automated analyzers. The system described here was used because it can continuously monitor absorbance at many wavelengths, providing the experimental data important for the initial characterization of the method. Because it assays one sample at a time, it serves as a prototype system only. Automated analyzers such as the centrifugal type are ideally suited for this kind of analysis. They can assay 20 to 30 samples simultaneously. The data indicate that the rate of reaction is not constant (Fig. 1); therefore, simple kinetic assay is not applicable. Nevertheless, quantitative results from two-point measurements can be obtained from such analyzers in less than 5 min (Fig. 2). This rate of assay would provide the results of 100 samples in a maximum of 20 to 25 min, depending on the loading capacity of the particular analyzer. Some other methods which can provide results rapidly have been described previously (6, 8, 10, 11). These methods appear to be efficient when a few samples are being analyzed and may be preferred for emergency samples because of their rapidity, but all are laborious when compared with the possibility of analyzing large numbers of samples on an automated analyzer. This is because the former assays require either extensive microscopic examination (10), preparation of multiple capillary tubes, and adjustment of pH's (6) or careful examination of multiple slides to differentiate between complete coagglutination and nonspecific agglutination (8, 11). Once automated, the spectrophotometric assay requires only pipetting of reagents into the disk of the centrifugal analyzer, placing the disk in the instrument, and pressing a button. Samples and reagents can be pipetted into the next disk while the samples in the first disk are being measured. Thus, the

338

LEVINSON

spectrophotometric approach offers the possibility of great reductions in technical manipulations, as well as provides quality assurance levels that cannot be realized by the other techniques. Data secured from a centrifugal analyzer in our laboratory indicate that the solubilization step can be reduced to fewer than 20 min. Thus, solubilization, centrifugation, pipetting, and assay of 100 samples do not need to take more than 1.5 h. This provides an efficient alternative for routine identification by a single technician of multiple samples grown overnight and then tested the next morning. This approach may also be amenable to use with nephelometric instrumentation (13). Although in its present form the spectrophotometric identification does not appear to be as rapid from colony recognition to final identification as are some other techniques (6, 8, 10, 11), samples can be assayed for the important group A and B antigens in short periods of time after colony recognition (Table 2). The assay provides an excellent quantitative means for studying the optimal parameters for rapid solubilization by S. albus enzyme or other enzymes of whole bacteria in concentrations encountered in clinical laboratories. It is possible that further studies will provide a means by which the growth period of the bacteria can be reduced or the sensitivity of the assay can be increased or both, enabling a complete identification technique which can be completed in only 1 or 2 h. To achieve this end, extended studies using an automated analyzer with many strains are necessary for defining additional parameters such as the following. (i) Should fewer colonies and more extract be used? (ii) Could whole colonies be substituted for growth in T-H broth cultures? (iii) Can longer digestion under improved conditions of temperature and pH provide more antigen so that only one colony might be used? In the present study, cross-reactions between different antisera and antigen groups were negligible (Table 1). Clearly, a larger study is necessary (preferably on an automated system) for evaluating the contribution of cross-reaction of many known strains. The greatest drawback to widespread use of this method in bacteriology laboratories may be the lack of access to automated analyzers. With the advent of new methods using enzyme im-

J. CLIN. MICROBIOL.

munoassay for measuring aminoglycosides (12) and other antibiotics by spectrophotometric means, automated analyzers may, in the future, become more common in microbiology laboratories. ACKNOWLEDGMENTS I thank Jane Kissling of the Massachusetts General Hospital for supplying some of the bacteria used in this study, Mindly Isaacs for typing the manuscript, Concetta Legero for technical assistance, and Kathy Levinson of North Shore Community College for helping to review and revise the manuscript.

LITERATURE CITED 1. Blom, M., and N. Hjorne. 1975. Immunochemical determination of serum albumin with a centrifugal analyzer. Clin. Chem. (Winston-Salem, N.C.) 21:195-198. 2. Christensen, P., G. Kahlmeter, S. Jonsson, and G. Kronvall. 1973. New method for the serological grouping of streptococci with specific antibodies adsorbed to protein A-containing staphylococci. Infect. Immun. 7: 881-884. 3. Edwards, E. A., and G. L. Larson. 1974. New method of grouping beta-hemolytic streptococci directly on sheep blood agar plates by coagglutination of specifically sensitized protein A-containing staphylococci. Appl. Microbiol. 28:972-976. 4. Finley, P. R., R. J. Williams, and J. M. Byers III. 1976. Immunochemical determination of human immunoglobulins with a centrifugal analyzer. Clin. Chem. (Winston-Salem, N.C.) 22:1037-1041. 5. Forsgren, A., and J. Sjoquist. 1966. Protein A from S. aureus. I. pseudo-immune reaction with human y-globulin. J. Immunol. 97:822-827. 6. Kholy, A. E., R. Facklam, G. Sabri, and J. Rotta. 1978. Serological identification of group A streptococci from throat scrapings before culture. J. Clin. Microbiol. 8:725-728. 7. Lancefield, R. C. 1933. A serological differentiation of human and other groups of hemolytic streptococci. J. Exp. Med. 57:571-595. 8. Lue, A. Y., I. P. Howit, and P. D. Ellner. 1978. Rapid grouping of beta-hemolytic streptococci by latex agglutination. J. Clin. Microbiol. 8:326-328. 9. Maxted, W. R. 1948. Preparation of streptococcal extracts for Lancefield grouping. Lancet ii:255-256. 10. Moody, M. D., E. C. Ellis, and E. Updyke. 1958. Staining bacterial smears with fluorescent antibody. IV. Grouping streptococci with fluorescent antibody. J. Bacteriol. 75:553-560. 11. Rosner, R. 1977. Laboratory evaluation of a rapid fourhour serological grouping of groups A, B, C, and G betastreptococci by the Phadebact Streptococcus Test. J. Clin. Microbiol. 6:23-26. 12. Standefer, J. C., and G. C. Saunders. 1978. Enzyme immunoassay for gentamycin. Clin. Chem. (WinstonSalem, N.C.) 24:1903-1907. 13. Sternberg, J. C., R. J. Anderson, R. C. Meyer, and J. E. Lillig. 1977. A rate nephelometer for immuno-precipitin measurement of specific serum proteins. Clin. Chem. (Winston-Salem, N.C.) 23:1155.

Quantitative assay of soluble beta-hemolytic streptococcal antigens via an immunochemical turbidimetric method with a spectrophotometer.

Vol. 10, No. 3 JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1979, p. 334-338 0095-1 137/79/09-0334/05$02.00/0 Quantitative Assay of Soluble Beta-Hemolyti...
757KB Sizes 0 Downloads 0 Views