Capillary Precipitin Test

APPLIED MICROBIOLOGY, Jan. 1968, p. 122-127 Vol. 16, No. 1 Copyright 1968 American Society for Microbiology Printed in U.S.A. Quantitative Aspects of the M Protein Capillary Precipitin Test JAY 0. COHEN AND LEO PINE National Communicable Disease Center, Atlanta, Georgia 30333 Received for publication 25 August 1967 A capillary procedure for quantitatively determining M protein is described. Capillaries were filled with measured amounts of serum and streptococcal extract. The capillaries were incubated, and then centrifuged to pack the precipitates. The relative sizes of the precipates were compared by a determination of the weights of their paper images (obtained by reflection from a microscope). Meaningful dilution curves were determined by this method. Variations of ph from 6 to 8 had little effect on the M protein precipitin test, and the test was not seriously affected by variations of the NaCl concentration from 0.85 to 4.67%. The addition of divalent ions (Ca++ and Mg++) did not influence the results. This test can be used to make quantitative comparisons of M protein preparations and to titrate type-specific antisera. In 1928, Lancefield reported the type-specific reaction of adsorbed antiserum with acid extracts of group A streptococci (4). The use of capillary pipettes to type group A streptococci by M precipitin reactions was described by Swift, Wilson, and Lancefield in 1943 (7). This procedure is suitable for qualitative M protein estimation. Quantitative determination of M protein has been done by dilution of the antigen for the qualitative precipitin test (5). A quantitative procedure based on the complete precipitation of M protein in centrifuge tubes and the analysis of total protein in the washed precipitates could be devised, but this would be quite wasteful of serum. Recently, Becker described a semiquantitative method for determination of M protein by use of a slide serum-gel diffusion technique (1). Extracts were diluted serially 1:2 and placed in peripheral wells around a central well containing homologous, type-specific M serum. Brock estimated M protein by a capillary method (2). He measured the length of precipitate that had settled to the bottom of the capillaries in fractions of millimeters under a dissecting microscope. The amount of M protein in the extract was roughly proportional to the length of the column of the precipitate. Recent interest in M protein for vaccines (3) and immunochemical studies indicates a need for methods for quantitative M protein determination. MATERIALS AND METHODS Capillaries. Typing capillaries, 0.7 to 1.0 mm in diameter and 152.4 mm in length, were used. Care was taken to avoid use of extremely wide- or narrowdiameter capillaries. Syringe. A glass and Teflon syringe (No. 7840, Cole-Parmer Instrument and Equipment Co., Chicago, Ill.) was used to fill the capillaries accurately. The plunger of the syringe was greased with highvacuum grease. The glass tip of the syringe was fitted with a small piece of rubber tubing, and a glass bead was inserted to close off the open end. A small hole was pierced in the rubber tubing with a 20-gauge needle. One end of a capillary was fitted into the hole in the rubber tubing. The capillary was filled, in order, with the serum, antigen, and buffer solution by counterclockwise turns of the vernier of the syringe. The amount of each reagent taken up could be measured quite accurately. The accuracy of the syringe was checked by the weighing of capillaries before and after they were filled with 0.01 ml of water. Weights did not vary by more than 2%. One volumetric unit on the syringe equaled 0.001 ml of fluid. Serum. Streptococcal M rabbit serum type 1 lot 4 and type 1 lot 1-C4 (Biological Reagents Section, National Communicable Disease Center), was used. Normal rabbit serum was used as a control. Antigen. Strain DS 2036-66 was obtained from the Streptococcus Unit, National Communicable Disease Center. Acid extracts were made according to the Lancefield method reported by Moody et al. (6). Four standard extracts were prepared for this study. The standard extracts were kept frozen at -65 C. Capillary precipitin test. To prepare a series of dilutions of antigen, 0.008 ml of serum was drawn into the capillaries. This was followed by the drawing of 0.001 to 0.008 ml of antigen into the same capillaries. Whenever the antigen column was smaller than the serum column, 0.005 N phosphate buffer (ph 7), or other suitable buffer, was drawn up to bring the antigen level equal in volume to that of 122

VOL. 16, 1968 M PROTEIN CAPILLARY PRECIPITIN TEST 123 the serum. When dilutions of serum were tested against a constant amount of antigen, buffer, then serum, then 0.008 ml of antigen were drawn up the capillary. To remove the capillaries from their attachment to the syringe, the left thumb was placed on the bottom of the capillary and the capillary was gently removed from the hole in the rubber tubing with the right hand. The right index finger was then placed over the open tip of the capillary. The capillary was inverted, and pressure of the index finger was eased to allow the column of fluid to move to the center of the capillary. The capillary was then pressed down into the plasticine of the typing rack. Capillaries were incubated for 2 hr at 37 C, and then for 41 to 44 hr at 4 to 6 C. After incubation, the capillaries were removed from the plasticine and placed inside rubber adapters in 40-ml centrifuge cups. The capillaries were centrifuged for 20 min at 2,400 rev/min in an International centrifuge (model CS, head 240). Centrifugation caused the precipitate to pack tightly against the plasticine plug. After centrifugation, the capillaries were placed on a 50 X 75 mm slide. The open-end half of each capillary was broken off and discarded. The halves with the plasticine and precipitate were taped to the slide in groups of five or six with masking tape. The slide was placed on the stage of a trinocular microscope. The image of each capillary was projected through the 10 X objective, then through the photographic tube of the microscope. The image was projected through a 15 X ocular lens and then through a prism to a screen which was 11.4 cm from the prism. The image was cast upon a sheet of note-pad paper clipped to the screen. The outline of each image of precipitate was drawn by pencil on the paper. The paper outlines were cut out and weighed on an analytical single-pan balance with an accuracy of 40.01 mg. The amount of precipitate formed in each capillary was represented by the weight of its image cut from note paper of similar weights. At first, note-pad paper sheets were chosen at random. Later, to reduce sources of error, sheets which weighed 1.52 to 1.58 g were selected. RESULTS The development of a quantitative capillary test for M protein requires the ability to obtain uniform results in replicate extracts of a streptococcal culture. An 18-hr Todd Hewitt broth culture of strain DS 2036-66 was used to obtain five replicate samples of 20 ml each. The optical density of the culture was 0.58, which is equivalent to 0.19 mg of dry weight per ml or 3.8 mg of dry weight per sample. The tubes were centrifuged, and the supernatant fluids were aspirated. Exactly 0.25 ml of 0.2 N HCI in 0.85% saline, plus one drop of meta-cresol purple, was added to each tube. The tubes were placed in a boiling-water bath for 10 min to extract the M protein. When they were cool, the tubes were centrifuged for 50 min at 2,200 rev/min. The TABLE 1. M protein values for replicate extracts of a culture of streptococci group A type la Extract Wt (g) of paper images of the precipitates Mean capillariesb of6brne Rangec 1 0.21359 0.19544-0.24190 2 0.21554 0.19850-0.22500 3 0.21095 0.20550-0.22525 4 0.21794 0.21058-0.22905 5 0.20656 0.18935-0.21935 a Serum used was type 1 lot 4. The coefficient of variation was 5.6%. bthe overall mean was 0.21299 g. c The maximal range was 0.18935 to 0.24190 g. supernatant fluid was carefully decanted from each tube into a 12-ml graduated centrifuge tube by use of a capillary pipette to give near-quantitative transfer. Each extract was neutralized with 0.22 to 0.23 ml of 0.2 N NaOH. The tubes were then centrifuged for 20 min at 2,200 rev/min. The clear supernatant fluids were used for subsequent tests. Replication of the five extracts was tested by the capillary precipitin test; this involved the drawing up of 0.008 ml of serum and 0.008 ml of extract into each capillary (Table 1). Although there was variation in each group of six assay capillaries, the averages for each group of capillaries are quite close to one another. Thus, five or six capillaries should be set up for each sample to obtain representative averages. An analysis of variance was performed to assess the variation among and within extracts. Compared to the variation within extracts, the variation among extracts was not statistically significant. That is, the mean values for replicate extracts of the same culture were not significantly different. The coefficient of variation (standard deviation/overall mean) was 5.6%. In Fig. 1, a standard curve is shown for dilutions of 0.001 to 0.008 ml of First Standard extract type 1 lot 4 serum. The curve was linear on a semilogarithmic plot for the first three concentrations; it then curved up at concentrations of 0.006 and 0.008 ml of antigen. Most points in this and subsequent figures represent the average of six values. Occasionally, a capillary from a group was broken. The Second Standard type 1 antigen was concentrated threefold by evaporation with a warmair jet. Several dilutions were made of the concentrated antigen (2/3, 1/3, and 1/6). Standard curves wert determined for each antigen dilution and for the full-strength concentrated antigen

124 COHEN AND PINE APPL. MICROBIOL. SO '& 200 oi 50-,, 150 :oc 0 120,O FIG. 1. Standard curve for IFirst Standard type I1 extract. Each capillary contained 0.008 ml of type I lot x 3 / 4 serum. FIG. 2. Standard curves for concentrated Second Standard type I extract and derived dilutions. Each capillary contained 0.008 ml of type I lot 4 serum. (Fig. 2 and 3). Figure 2 is a semilogarithmic plot. The various curves appear to be approximately parallel to each other, as indicated by the "eyefitted" straight lines. Equivalent concentrations of antigen tend to agree in image weights from one curve to another. For instance, 0.004 ml on the 1/6 curve equals 0.0024 ml of the 1/3 curve. This figure suggests that the shape of the curves with the same serum is independent of antigen concentration for a given type 1 M antigen. Values for the 0.008-ml concentration in the full-strength curve were lower than that for the 0.006-ml reading. This is seen best in the arithmetic plot in Fig. 3. The decrease in antigen precipitated at 0.008 ml of antigen as compared with 0.006 ml suggests that antigen excess was reached with the more concentrated antigens at the 0.008-ml level. Samples (1 ml) of the Fourth Standard antigen were adjusted to ph values of 6.0, 6.5, 7.0, 7.4, and 8.0 by dialysis for 4 hr against 0.01 M potassium phosphate buffer which contained 0.85% NaCl. After dialysis, the volumes were adjusted to be equivalent. Precipitation curves were determined for the extracts of different ph values (Fig. 4). Useful dilution curves were obtained at all the E va 90 0 0002 0004 0006 0008 AMOUNT OF EXTRACT (ml) FIG. 3. Arithmetic plot. Standard curves for concentrated Second Standard type 1 extract and derived dilutions. Each capillary contained 0.008 ml of type I lot 4 serum. 350 300 250 12 200 cn 0 _ 150 AMOUN' C,9 EXTRACT (,1R FIG. 4. Values ofm protein capillary precipitin tests at different ph values. Each capillary contained 0.008 ml oftype I lot 4 serum. The antigens used were samples of Fourth Standard type I extract dialyzed against buffered saline at ph values of6.0, 6.5, 7.0, 7.4, and 8.0.

VOL. 16, 1968 M PROTEIN CAPILLARY PRECIPITIN TEST 125 i 5-5 a.1,l KEI - ",, 2..9c,. *IT.DIT VgC'2 OR C.CI2 0 0.002 0.004 0006 0008 AMOUNT OF EXTRACT (ml) FIG. 5. Values of M protein capillary precipitin tests at different NaCl concentrations. Each capillary contained 0.008 ml of type 1 lot 4 serum. The antigens used were samples of Fourth Standard type I extract dialyzed against different concentrations of saline buffered at ph 7.4. ph values tested, and the variations may be within the range of experimental error. The curves for ph 7.0 and 8.0 appeared to flatten out slightly for 0.006- and 0.008-ml levels. The optimal ph appeared to be 7.4. Here, the curve was elongated, with more separation between the various points. Another dialysis experiment was run with various NaCl concentrations. Samples of 0.9 ml were dialyzed for 4.25 hr against 4.25% NaCl, 8.5% NaCl, and 12% NaCl in 0.01 M potassium phosphate buffer (ph 7.4). The final concentrations are lower because of the 0.008 ml of serum in each capillary. In those capillaries with less than 0.008 ml of antigen, the fluid was brought up to volume with the appropriate buffered salt solution (Fig. 5). There was little practical difference between the salt concentrations tested except at the 6.42% level, where the curve became completely flat between 0.006 and 0.008 ml of antigen. The capillary M typing test was not affected significantly by variations in the salt concentration from 0.85 to 4.67%. Two 0.9-ml samples of the Fourth Standard type 1 extract were dialyzed against 0.02 M MgC12 and 0.02 M CaC12, respectively. The solutions also contained 0.85% NaCl and 0.01 M tris(hydroxymethyl)aminomethane (Tris buffer; ph 7.4). FIG. 6. Values of M protein capillary precipitin tests in the presence of divalent ions. Each capillary contained 0.008 ml of type I lot 4 serum. Antigens used were samples offourth Standard type I extract dialyzed against 0.02 m CaCl2 and 0.02 m MgC12 each in ph 7.4 Tris-buffered saline. A curve without either MgCl2 or CaCI2 is shown for comparison. FIG. 7...1, Standard curve for type I serum lot l-c-4. There was no difference between the MgCl2 and CaCl2 curves, and these divalent ions did not affect the M protein precipitation test (Fig. 6). Another lot of type 1 serum, lot 1-C-4, was tested by the quantitative precipitin test for its activity against the Fourth Standard extract. Two sets of dilutions were used. In the first set, amounts of serum from 0.001 to 0.008 ml were tested with 0.008 ml of the Fourth Standard extract. In the second set, the serum was kept constant at 0.008 ml, while amounts from 0.001 to 0.008 ml of Fourth Standard extract were tested. The diluent was phosphate-buffered (ph 7.4) 0.85% saline (Fig. 7). The reaction of lot 1-C-4 serum with the Fourth Standard antigen resulted in much less precipitate than that observed with type 1 lot 4 serum. The proportionate differences are greatest at the 0.001- and 0.002-ml levels. The results indicate that a weaker serum will precipitate less from an extract, and that the difference is most pronounced with weak antigens. In Fig. 7 the curves for the antigen dilution series and the serum dilution series are essentially 1.11- -X-

126 COHEN AND PINE APPL. MICROBIOL. superimposed. This indicates that even at the 0.001-ml serum level, antigen excess did not appear to influence the reaction. The curve for type 1 serum lot 4 is also shown for comparison. The two curves are slightly different in shape, but a portion of the serum 1-C-4 curve is parallel to the serum lot 4 curve. The quantitative capillary procedure used to estimate the amount of M protein can be standardized either in relation to an extract or preparation of M protein or in relation to a particular lot of typing serum. Since M protein preparations have been observed to deteriorate to some extent during storage, the serum might be the best standard. Thus, in Fig. 7, we could define one unit of M protein as that amount of M protein which, when it is reacted with 0.008 ml of type 1 lot 4 serum, gives a precipitate whose paper image in this system weighs 77 mg (as in the curve for the Fourth Standard M extract). An image which weighs 190 mg would represent 4 units of M protein. The Fourth Standard extract then contains 1 unit of M protein in 0.001 ml of extract, or 1,000 units of M protein per ml. DiSCUSSION It has been apparent since the introduction of the capillary precipitin test for M protein (7) that the test was quite sensitive, and that it had quantitative aspects. When Group A streptococci are typed by this method, reactions are described in a range from + to + + +. Similar designations combined with serial dilutions have been used to estimate the strength of antisera or antigens. If one attempts to compare the M protein production of two strains of the same type, or of extracts of streptococci grown on two different media, qualitative and semiquantitative methods are not adequate. Our data indicate that identical volume replicate extracts of a streptococcal culture give reproducible quantitative results with the capillary M protein test. This enables valid comparisons to be made between two different streptococcal cultures of the same M type by use of a quantitative M protein precipitin test. One weak point in the present qualitative test is the strong, or + + +, reaction. Such designation covers a range of about 80- to 300-mg weights for paper images in the quantitative test. The results reported here indicate that most of the procedures in use today in the qualitative test are adequate for this purpose. Positive precipitation results were observed over a wide range of antigen and antibody concentrations. Variations in the ph from 6 to 8 had little influence on the results, and the test was not distorted when the salt concentration was raised from 0.85 % to 4.7%. The divalent ions Ca++ and Mg++ did not influence the results at the concentrations tested. The amount of precipitate formed is dependent on the strength of the serum as well as on the antigen concentration. Weaker sera will give less precipitation than will stronger sera with the same extract, and the difference is most critical at low antigen concentration. The accuracy and replication of quantitative M protein determinations depend upon volumetric accuracy of the procedures. Extractions of the same quantity of cells into a known volume of fluid must be made for valid comparison. The more care with which the reactants were measured, and with which the capillaries were incubated, centrifuged, and read, the more reproducible were the results. We found no evidence of antigen excess except in the case where an extract was concentrated threefold for use as an antigen. In some other experiments, potent extracts reacted without evidence of antigen excess. The tendency of the concentrated Second Standard extract to show antigen excess effects may be due to the concentration procedure. The results indicated that antigen excess was not a problem with strong unconcentrated extracts when tested with two different lots of type 1 antiserum. The quantitative capillary M protein test described here should be useful for immunochemical experiments with streptococci. For determination of M values for comparison of M preparations, it is recommended that a standard serum be used, because M protein solutions are known to be less stable than serum. Further refinements of the procedures may lead to still greater accuracy. The test as described here makes possible quantitative comparisons of M protein preparations and titration of type-specific antisera. LITERATURE CITED 1. BECKER, C. G. 1967. Enhancing effect of type specific antistreptococcal antibodies on the emergence of streptococci rich in M protein. Proc. Soc. Exptl. Biol. Med. 124:331-335. 2. BROCK, T. D. 1963. Effect of antibiotics and inhibitors on M protein synthesis. J. Bacteriol. 85:527-531. 3. Fox, E. N., M. K. WITTNER, AND A. DORFMAN. 1966. Antigenicity of the M proteins of Group A hemolytic streptococci. IlI. Antibody responses and cutaneous hypersensitivity in humans. J. Exptl. Med. 124:1135-1151. 4. LANCEFIELD, R. C. 1928. The antigenic complex of Streptococcus haemolyticus. 1. Demonstration of a type-specific substance in extracts of Streptococcus haemolyticus. J. Exptl. Med. 47:91-103. 5. LANCEFIELD, R. C., AND G. E. PERLMANN. 1952.

VOL. 16, 1968 M PROTEIN CAPILLA] RY PRECIPITIN TEST 127 Preparation and properties of type-specific M antigen isolated from a Group A, type 1 hemolytic streptococcus. J. Exptl. Med. 96:71-82. 6. MOODY, M. D., J. PADULA, D. LIZANA, AND C. T. HALL. 1965. Epidemiologic characterization of Group A streptococci by T-agglutination and M-precipitation tests in the public health laboratory. Health Lab. Sci. 2:149-162. 7. SWIFT, H. F., A. T. WILSON, AND R. C. LANCE- FIELD. 1943. Typing Group A hemolytic streptococci by M precipitin reactions in capillary pipettes. J. Exptl. Med. 78:127-133.