"in"; 4 ' .r {I ' 'z q n .. . '2 " u -.'-l\ . . .~. law-ta 3-); n s u a u: '33 e s: c g on. '1 - , .- é‘ ‘P¢f‘;§l ‘1 3 £3 ‘ 'I? z‘;::}::‘k 514 8(ka ‘6‘ 1“ L- L3 " ' ‘.' b A‘.u:3 .. LIBRAfiY 3mm; Michigan State University .s-Tr A GEL DIFFUSION METHOD FOR THE QUANTITATIVE ASSAY OF DUCK HEPATITIS VIRUS By Michael G. Gabridge A.THESIS Submitted to Michigan State University in partial fulfillment of the requirements _for the degree of MASTER OF SCIENCE Department of Microbiology and Public Health 1966 Dedicated to my wife 11 ACKNOWLEDGEMENTS The author would like to express his gratitude to Dr. John P. Newman, Professor of Microbiology and Public Health, for the guidance, patience and consideration extended over the past several years. His fine example served as a constant guide during the research and preparation of this manuscript. I would also like to acknowledge the kind and helpful assistance of Mrs. C. M. Newth and Mrs. M. G. Shue, and express my appreciation to the staff of the Diagnostic Microbiology Laboratory of the M. S. U. Veterinary Clinic for their aid and advice. Last, but not least, a special thanks to my family for their words of encouragement and sincere interest in this project. ill TABLE INTRODUCTION . . . . . . . . LITERATURE REVIEW. . . . . . Duck Hepatitis. . . . . Duck Hepatitis Virus. . Quantitative Techniques Gel Diffusion . . . . . MATERIALS AND METHODS. . Virus . . . . . . . Virus Purification. Antisera. . . . . . .Diffusion Agar. . . Ouchterlony Plates. Diffusion Block and Grading Precipitate Reference Titration Electron Microscopy RESULTS. 0 O O O O O O 0 OF CONTENTS _Procedure . . . Lines . . . . . Diffusion Plate Tests Optimal Agar Length Heactant State. . . .Temperature Effects Single Diffusion. . Reference Titration Experimental Titration. O I O O O 0 iv Page \o\O\OUI«F'UNN to n: to A) id F4 &a F‘ Id h‘ id #4 l4 F‘ r4 ox c» c- h) \o w) (b O\ Kn c— u) r4 ta c> c: Page Statistical Analysis . . . . . . . . . . . . . . 30 DISCUSSION. . . . . . . . . . . . . . . . . . . . . . 33 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 40 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . 41 LIST OF TABLES Table Page 1. Results from diffusion tests with different lengths of agar strips . . . . . . . . . . . . 20 2. Effects of the reactant's physical state on the Precipitation pattern. 0 o o c c c o o o o 21 3. Comparison of results obtained with various incubation temperatures. 0 o o o o o o e o o o 23 4. Results from diffusion test with antiserwm incorporated in the agar o e o o o o o o o o o 25 5. Final composition of reactants and results from diffusion block titration . . . . . . . . 27 6. Tabulation of the results from replicate titration experiment 0 o o o o o o o o o e o o 31 vi LIST OF FIGURES Figures 1. Sketch of plastic diffusion block . . . . . . 2. Diagram of one reaction well. . . . . . . . . 3. Diffusion of DHV L278 and normal allantoic flu1d againSt antiserum B o o c o o c o o o c 4. Diffusion of DHV L278 and normal allantoic fluid against adsorbed-virus antiserum . . . 5. Illustration of the tendency for two lines or precipitate to appear as one o o o o o o o 6. Electron micrograph of shadowed preparation of purified duck hepatitis virus. . . . . . . 7. Relationship of virus concentration in reservoir to distance of leading edge of precipitate................. 8. Effect of virus concentration on velocity of precipitate migration . . . . . . . . . . . . 9. Histogram of frequency versus distance of the leading edge of precipitate from the antigen Well. 0 O O O O O O O O O O O O O O O O O O 0 vii Page 12 12 17 17 18 18 29 29 32 INTRODUCTION Because of their minute size and unique preperties, viruses can only be counted in an indirect manner. Methods range from negative staining in electron microscopy to hemagglutination and cytopathogenicity, but nearly all have drawbacks for routine diagnostic and small-scale research studies. .When the virus is non-cytopathic in cell culture and does not agglutinate red blood cells, one must often resort to the inoculation of embryonating chicken eggs with a series of virus dilutions and the determination of titer with the aid of statistics. This is time consuming, expensive, and subject to error from a variety of sources. .The purpose of this study was to find a new method for quantifying such viruses, using duck hepatitis virus as the model system. The new procedure involved virus precipitation in gel with antiserum, with the position of the precipitate proportional to the virus titer. Using a standard curve, the distance in millimeters from the point of origin could be transposed to the number of particles or infectious units per amount of inoculum. The ratio of reactant volume of reaction area was the inverse of that used in older methods, assuming that this would result in increased sensitivity. LITERATURE REVIEW Duck Hepatitis Duck hepatitis is a highly contagious disease specific for ducks. The morbidity and mortality rate is usually highest in ducklings less than three weeks of age. It was first recognized in New Ibrk state by Levine and Fabricant (1950), and has been reported in England (ASplin and MoLauchlin, 195h), Canada (MacPherson and Avery, 1956), Germany (Reuss, 1959), and Egypt (Hagen and Bruner, 1961). Clinically, the disease is manifested by a slit-like appearance of the eyes, squatting, general dullness and listlessness, convulsions, and opisthotonus (Hanson and Alberts, 1956). Death often occurs within one to two hours after the onset of symptoms (Fabricant gt_gl., 1957), and the mortality rate may be as high as 95 per cent. The most characteristic pathologic alteration is an enlarged, mottled liver (Hanson and Alberta, 1956) caused by hemorrhage and edema, and often accompanied by a Jaundice- like discoloration (Fabrioant gt’gl., 1957). .The spleen, heart and brain, though not outwardly affected, have been implicated since they contain detectable amounts of virus (Hwang and Dougherty, l96u). Microscopically, the lesions resemble those associated with human hepatitis and include proliferation of the bile ducts, necrosis of hepatic cells, and paravascular infiltra- tion with granulocytes and plasma cells (Hanson, 1958). 2 Duck Hepatitis Virus This RNA virus is a sphere with a diameter of 20-40 mu (Reuse, 1959) and is ether stable (Pollard and Starr, 1960). Virus particles found in hepatic cells of infected ducklings may be dispersed in the cytoplasm, packed into crystals, or contained within membranes (Richter g§_al., 1964). -At 25°C, 37°C, and 56°C there is a definite reduction in the ID50 after 5 weeks, 48 hours, and 90 min, respectively (Hanson 22.21., 1961). The virus remains viable after treat- ment with ammonium sulfate or 30 per cent methanol for 24 hrs, and viral infectivity is stable from pH 4.8 to 7.8.(ibid.). Chicken embryos are susceptible to infection with duck hepatitis virus, DHV (ASplln, 1958; Hwang and Daugherty, 1964). Characteristic lesions include a severe edema in the abdominal. region, general dwarfing, and a reddening of the skin, with death occurring two to six days after inoculation (Hanson and Alberts, 1956). .Pollard and Starr (1959) reported multiplication of the virus without cytopathic effects (OPE) in chicken embryo explants.t Trypsinized cells were not susceptible to the virus. This differs from the findings of Kheberle Efiwél° (1961) who successfully propagated DHV in monolayer cultures of chicken embryo liver cells diSpersed with trypsin. Fitzgerald 22.2L9 (1963) was the first to report a marked CPE and necrosis in duck embryo kidney cells after the 24th virus passage. Sazawa 32,51. (1963) tested a recently isolated strain of DHV for infectivity in duck kidney, chick kidney, chicken 4 embryo, and duck embryo cells. Multiplication occurred in the duck embryo cells and the CPE was indistinct and described as ”hazy”. No conspicuous CPE or cellular damage was reported by Hwang (1965) after twenty serial passages in duck embryo fibroblast cultures. Quantitative Techniques Methods for the quantitative assay of viruses may be grouped into three categories: quantal, enumerative, and graded. The quantal type of assay involves an "all or nothing" response in which a host exposed to the virus is classed as infected or non-infected. If the host is a test animal, infection may be recognized by visible or microsc0pic lesions, paralysis, or death. If embryonating eggs are used, there may be hemorrhage of subcutaneous tissue, congestion of the epidermis, dwarfing of the embryo, or similarrmanifestations (Cunningham, 1966). In cell culture systems, infection is evidenced by a CPE (Rubin and Temin, 1958), hemadsorption (White, 1963), hemagglutination (Fazekas de St. Groth and White, 1938), or a change in the pH of the medium (Huang, 1943). Enumerative assays involve some type of direct count and results are expressed in real rather than statistical units. Actual counts of viruses may be made with the electron micro- scope, but the number of physical particles does not always correlate with the number of infectious particles (Smadel SE.§£39 1939; Friedewald and Pickels, 1944; Issacs and Donald, 1955; Tyrell and Valentine, 1957). 5 The most widely used form of enumerative assay involves the counting of plaques or areas of necrosis in a monolayer cell culture (Dulbecco, 1952). This has been used with a _number of viruses, including Coxsackie virus (Hsiung and Melnick, 1955), foot and mouth disease virus (Bachrach, 1957), and polyoma virus (Dulbecco and Freeman, 1959). Dulbecco and Vogt (1953) proved mathematically the validity of assuming one viral particle produced one plaque, but there is an inordinate amount of variability in the method * and a departure from the expected distribution could easily occur (Armitage, 1957). (Graded assays are based on some measure of the degree of response, such as survival time, size of an infected organ, or size of a lesion (Harris, 1964). Such methods are normally used with oncogenic viruses (Rubin, 1960), but have also been used with Newcastle disease virus (Cunningham, 1966) and agents of the psittacosis group (Golub, 1948). Gel Diffusion Immunologic precipitation involves a reaction which can best be expressed by the Marrack-Heidelberger hypothesis (Cruickshank, 1963): bivalent antibody molecules link up adjacent antigen-antibody complexes, and if optimal prOportions exist, form a three-dimensional lattice network large enough to be visible. These combinations take place in a number of stages and are essentially reversible (Humphrey and White, 1964). . 1 Oudin (1946) first developed the technique of forming antigen-antibody precipitate(s) in gels using the single 6 diffusion tube test. Antigen was placed on top of an antiserum-agar mixture and diffusion resulted in a precipitate appearing in the latter. Elek (1948) and Ouchterlony (1948) modified this tech- nique to one of double diffusion where both antigen and antibody migrate toward each other through a gel before reacting. Precipitation is affected by the salt concentra- tion of the agar (Aladjem and Lieberman, 1952), the pH (Wilson and Pringle, 1954), and molecular weight of the antigen (Korngold and Van Leewen, 1957). The first demonstration of a virus-caused precipitation in agar was made by Jansen and Francis (1953) using influenza virus. Belyavin (1955) first suggested its diagnostic applica- tion. Viruses of mumps and Newcastle disease have also been studied with gel diffusion (Belyavin, 1957). Murty and Hanson (1961) did an analysis of DHV using a modified microgel diffusion method, and found three lines of precipitate when liver from infected duck embryos was diffused against unadsorbed hyperimmune rabbit serum. Only two of these lines were actually Specific for the DHV. .Precipitation and gel diffusion tests are generally used for qualitative analyses, but attempts have been made to establish quantitative applications. Heidelberger and Kendall (1935) were among the first to work in this area, and used the difference in optical density before and after precipitation as a parameter related to virus concentration. Adding different concentrations of antigen (Gussoni, 1964) or antibody(Bjork1und, 1952) and defining the inhibition 7 of visible precipitate as the endpoint in a titration may also be done. Attempts have been.made to deve10p a quantitative analysis based on the fact that the distance of a precipitate band from a well of reactant is commensurate with the concentra- tion of the substance being diffused (Preer, 1956; Korngold and Van Leewen, 1957; Aladjem g§_al., 1959). .Le Bouvier (1957) found the size and density of a band reflected both the absolute and relative concentrations of antigen or antibody. Such methods are continually being refined and modified, with a resultant increase in sensitivity. Feinberg (1957) used an adaptation of the original Ouchterlony technique and found that visible precipitate could be formed by as little as 0.05 ug of protein. A Studies utilizing double diffusion are generally considered more accurate than those using single diffusion since the former does not depend on how much material was in the reaction zone initially, but on how much can be fed into it (Crowle, 1961). Single diffusion tests are also more susceptible to secondary precipitation or non-specific multiple bands which give false results (Leuker and Crowle, 1963).i ~E1ek (1949) noted that in double diffusion with the migration of the two reactants taking place at an angle of 90 degrees to each other, conditions are produced for the interaction of antigen and antibody in all proportions. Expressions have been developed that relate the antigen's diffusion coefficient to the slope of the line (Neff, 1957; 8 Becker, 1959) and the tangent of the angle (Allison and Humphrey, 1959) formed in such a system. MATERIALS AND METHODS 1132s The duck hepatitis virus used in this study was obtained from Dr. E. Dougherty of the Duck Research Laboratory at Eastport, Long Island. It was contained in allantoic fluid from infected chicken embryos and labelled L275; L2 signi- fied the virus was adapted to growth in chicken embryos and ”75' indicated the number of passages since the original isolation. Upon receipt it was stored at -70°C. Prior to use, the allantoic fluid was thawed and centrifuged at 500g for 20 min. The sediment was discarded and the supernatant fluid was kept at 4°C. Virus Purification The anion resin used for the virus purification was the chloride form of Dowex l-8X (200-400 mesh). To insure purity of the stock resin, it was washed in 3N sodium hydroxide for 4 hrs, rinsed once in distilled water, washed in 3N sodium chloride for 3 hrs, and given a final rinse in distilled, deionized water. The virus suspension was combined with the washed resin in a ratio of 231 so that the final volume was 20-30 ml (Mallman, 1961). The resulting mixture was placed in a 50 m1 centrifuge tube, shaken intermittently for l5.min,-centrifuged at 20,000g for 15 min, and the supernatant fluid removed. The resin pellet was resuspended in 10 m1 of 10% aqueous sodium phCSphate, shaken, and centrifuged in the same manner. 9 10 The resulting supernatant fluid, now referred to as "purified DHV”, was stored at 4°C. Its virus content was confirmed with electron micrographs and Ouchterlony tests. Antisera Several antisera were prepared using rabbits. Antiserum A was obtained after five intraveneous injec- tions of DHV L278. Amount of inoculum was 1.0 m1, administered at four-day intervals. The serum was collected one week after the fifth injection. Antiserum B was produced in reSponse to a stable suSpen- sion of 0.5 m1 of Freund's Complete Adjuvant and 0.5 m1 of DHV L278 injected subcutaneously into several places on the back of a rabbit. This was repeated in three weeks and the serum collected seven days later. Adsorbed antiserum B was obtained by combining normal allantoic fluid with antiserum in a ratio of 1:3, centrifuging, and discarding the sediment. A series of subcutaneous injections similar to those just described were used in the production of antiserum C, but the virus had been purified on the anion exchange resin. Diffusion Agar The agar used in the Ouchterlony plates and diffusion block was similar to that suggested by Campbell g§_gl. (1963). The trypan blue was included to increase the ease with which lines would photograph, and the pH was buffered to 7.8-8.0 to enhance precipitation. The composition of the agar was as follows: 11 Ionagar #2 * 0.425 gm Saline ' 46.8 ml Borate Buffer ** 2.5 m1 Merthiolate (1%) 1.0 ml Trypan Blue (1%) 0.2 ml The Ionagar, saline, and buffer were mixed, and auto- claved at 121°C for 15 min. The merthiolate and trypan blue were added after cooling and the mixture was held at 50°C until it was allowed to solidify in the plate or reaction well. * Available from Oxo Limited, London. , ** Borate Buffer: Boric Acid 6.184 gm Borax 9.536 gm NaCl 4.383 gm Distilled water to 1 liter. Final pH = 8.5 Ouchterlony Plates Ouchterlony plates were prepared from 100 x 15 mm plastic petri dishes filled with 20 m1 of diffusion agar. An ink diagram with the desired arrangement of wells was placed under the dish and holes were cut with a cork borer previously cleaned with alcohol and flamed. The wells, approximately 2 cm apart, were initially filled with 0.2 m1 of liquid. Plates were held at room temperature and the wells refilled after 24 hrs. Photographs were made using al35 mm camera with indirect fluorescent lighting. Diffusion Block and Procedure The apparatus used in the diffusion study was a specially constructed 12 x 4 x 3/4 inch block of plexiglass. The block contained eight separate reaction wells, each 1/2 inch in depth (Figures 1 and 2). A reaction well consisted of two 12 Figure 1. Sketch of plastic diffusion block. Figure 2. Diagram of one reaction well (measurements in inches). 13 circular reservoirs (3/4 inch diam.) for antigen and antibody connected by a 2 x 1/8 inch slot which contained agar and served as the reaction site. Before introducing the agar, the slot was isolated from the reservoirs by a short piece of masking tape applied to the latters' inside surfaces. A plastic 1.0 m1 serological pipette was used to deliver 1.0 m1 of diffusion agar into the slot. Once the agar solidified, the tape was removed and the agar strip cut to the desired length. This was done by plac- ing the block over a ruled paper and making certain that the final agar strip would be equidistant from both reservoirs. After the excess agar was removed, the reservoirs were swabbed with 70% alcohol and the block was placed under an ultraviolet lamp until dry. Reservoirs were filled with 1.0 m1 of the proper dilution of either antiserum or virus. After all solutions had been introduced into the appropriate wells, the block was covered with a 12 x 4 inch piece of flexible plastic film (previously cleansed with 70% alcohol) on top of which was placed a 12 x 4 x 1/8 inch section of plexiglass. radi eci itate nes Some degree of the relative intensity of precipitate lines was afforded by establishing an arbitrary scale as follows: 0 No visible reaction l4 - Line only visible when agar strip was removed from slot and examined from side + Light precipitate; barely visible in block; hard to measure since it has no definite boundries ++ Moderate precipitate; easily visible in block; easy to measure +++ Heavy precipitate; very easy to discern and measure This method is somewhat subjective but proved adequate since the results were only used for comparative purposes. Reference Titration Serial ten-fold dilutions of the virus were prepared using 4.5 m1 of sterile nutrient broth as the diluent. A sterile 1 m1 serological pipette was used to deliver 0.5 m1 of virus to the first tube of broth. Using a clean, sterile 1 m1 pipette, the mixture was aspirated and expelled 15 times prior to transfering 0.5 ml to the next tube-of broth. This pipette was discarded and the process repeated in all of the remaining tubes. ‘ Five 7-day old chicken embryos were each.inoculated with 0.2 ml of a virus dilution via the allantoic cavity (Cunningham, 1963). After injecting five embryos,-the last 0.1 m1 of the dilution in the syringe was put into a tube of brain-heart infusion broth and incubated 48 hrs to check for bacterial contamination. The eggs were incubated at 99.5°F and candled daily for 7 days. Death of embryos within the first 24 hrs was 15 attributed to trauma or contamination, and these embryos were not included in the final results. Embryos found dead after this period were examined for edema in the abdominal region, dwarfing, and hemorrhaging to ascertain viral infection. The titer was then determined using the method of Reed and Muench (1938). Electron Microscogy Electron micrographs were taken to insure the purity of the virus adsorbed to the resin. The purified virus suspen- sion was sprayed with a pyrex nebulizer onto #2200 EFFA c0pper grids covered with formvar. After drying, the Specimens were shadowcast with palladium at an angle of 60 degrees. These were examined in an RCA, type EMU-2 electron microscope, and exposures were made on 2 x 2 photographic plates. Negative plates were made from these and the images were enlarged 4-5 times prior to printing on Kodabromide photographic paper. RESULTS Diffusion Plate Tests To test the various antisera for the capacity to precip- itate with the virus, several Ouchterlony plates were pre- pared. Figure 3 depicts such a plate in which antiserum B was tested against both DHV and normal allantoic fluid. Two lines of heavy precipitate are prominent, with a third line barely visible behind the peak under the middle well. Testing antiserum C in a similar manner served as a check on the relative purity of the resin-adsorbed virus suspension. One heavy line of precipitate along with two very light lines were formed (Fig. 4). One of the light lines was triangular in shape, while the other was much less distinct and its exact shape could not be easily determined. 'Antiserum A repeatedly failed to give any evidence of precipitate when diffused against virus used as the antigen. vA tendency for two lines of precipitate to lie almost in the same position was frequently observed.r It was noted for both antiserum B and C, and is illustrated in figure 5. The fainter of the two heavy lines appears single at the top well, but separates into two lines cpposite the lower wells. When diluted antisera were placed in the wells opposite a constant amount of virus, a continuousline encircling the center well was formed. All of the dilutions tested (1:1 to 1:10) gave some sort of visible reaction, but those higher than 1:2 were very indistinct. 16 l7 Figure 3. Diffusion of DHV L278 and normal allantoic fluid against antiserum B. l and 3 = DHV; 2 - allantoic fluid; 8 - antiserum B Figure 4. Diffusion of DHV L 78 and normal allantoic fluid against adsorbed-virus ant serum. l,3,5 - DHV; 2,4,6 - allantoic fluid; C = antiserum C l8 Figure 5. Illustration of the tendency for two lines of precipitate to appear as one. l;3,5 - DHV; 2,4,6 - allantoic fluid; a = antiserum B Figure 6. Electron micrograph of shadowed preparation of purified duck hepatitis virus (X 48,000). 19 The homogeneity of the purified virus suspension and antiserum C was also checked using electron microsc0py. Microdroplets of the virus particles can easily be distin- guished in micrographs of the adsorbed-virus suspension (Fig. 6) used to immunize the rabbit, while normal allantoic fluid failed to yield such patterns. Optimal Agar Length When different lengths of agar strips were used in the diffusion block, it was found that the shorter ones (é‘, 1', 1%“) gave the largest number of bands (Table 1). The 2 inch strip failed to display any signs of precipitate after six days, while the three shorter ones all had some‘evidence of reaction. The visibility or degree of reaction (see Materials and Methods) was best in the % inch strip, and the.l% inch strip had the least distinct lines. The distance given in this and remaining tables represents the distance of the lead- ing edge of the precipitate band in mm from the antigen well. Reactant State Changes in the precipitation pattern were observed when the physical state of the reactants was altered (Table 2). Having the reactant in the form of a gel was achieved by mixing the liquid reactant with a Specified volume of melted diffusion agar prior to placing it in the well. Table 2 also contains results from an analogous experiment where the reactants were in the liquid form but diluted with the appro- priate amounts of saline so the two sets of data might be compared. A very uneven, complex pattern of bands was 20 How «cpcpm pndpomom ooam a.c3oa 0 sentences .aeonapsa mama sen assesses o o o o send N as n.3m I was sea .c A Send ma as o.mH + was cm .9 as o.wH ++ ma: ow .m N coma H as 0.: + was w: .e as o.m +++ man mm .o as :.NH ++ was on .n as n.~a +++ was mm .m 3 send m necaa4 mencm cnoN Amado my “made my on no mo masseuse ssaaanaaa> ease senses nausea .madapm name no mnpwnca pacacmuae and: name» Scamsmuae Scam muHSmcm .H Sande 21 Table 2. Effects of the reactants' physical state on the precipitation pattern. LIQUID GEL tine lands Distance Pattern Bands Distance Pattern . (hrs) (m) (M) n o o [:1 o o [:3 24 I - a. 4.0 |:]:::] 1 a. 4.0 [:[::3 36 2 a. 2.5 III: | 3 a. 3.0 III | b. 4.0 b. 4.0 c. 4.5 48 3 a. 2.5 4 a. 2.5 ‘ b. 4.0 []I[::] b. 4.3 []l:[:] c. 6.0 c. 4.5 d. 8.0 60 5 a. 1.0 4 a. 2.0 [Ell | b. 2.0 II1[:J . b. 4.4 c. 3.0 c. 4.5 d. 5.5 d. 8.0 e. 7.0 72 S a. 1.0 5 a. 1.5 b. 2.0 “I![:] b. 2.5 fll[]:] c. 3.0 c. 3.0 d. 3.5 d. 5.0 e. 7.0 e. 8.0 22 obtained when a gel-liquid interface existed, whereas a gel- gel interface resulted in a more orderly arrangement of bands. The overlapping of lines in the former became less of a hin- drance in measuring as time increased Since the migration tended to separate the bands. When the reactants were in the liquid form, 72 hrs was the optimal time for measuring, but measurements could be made any time after 36 hrs when they were in the gel form. Liquids in the reactant reservoirs were also undesirable since they were easily spilled when the block was removed from the incubator to be inspected. The number of individual bands observed at the specified time intervals remained comparable in both cases, as did the distance between the two bands farthest apart.. Once the first band appeared, precipitate formed in both directions with the migration toward the antibody well being the more prominent. There was a linear relationship between the time and the distance of the band closest to the antibody well when the gel phase was used. The line was noted to move approximately 40.5 mm per 12 hr period. Temperature Effects The effects of temperature of the diffusion patterns involved both the number and the visibility of the precipitate bands (Table 3). At 4°C, only one or two lines appeared, depending.on the length of reaction strip, whereas 4 or 5 appeared at temperatures of 21° and 37°C. The distinctness or visibility also increased with an increase in temperature. As in the previous experiments, the shortest agar strip had 23 mus QOH .Q as o.na + as o.ma ++ mas om .c Sosa H as o.w + m9: ow .c as as + 2e as .e seam as m.m ++ my: mm .o as o.m + was on .9 as o.m ++ man on .6 send m as m.mH + was oma .c send a as o.m + was oNH .o ooam as 0.: + mm: mm .9 as 3.0 ++ was Nd .c mead m as m.m + was mwd .n Sosa H a no: as n.m n my: wwa .n as m.m + was SSH .c send m Manama Hnwd on no 3833 SEER as as; fiesta assessments .mecdn opcpHaHooaa mo hpdpacau use audacsv can so chapcncaacp no 9009mm .m manna 24 the greatest number of bands. Bands in the % inch strip held at 4°C presented a problem in measuring Since they were curved, with one side of the curve almost parallel to the surface of the agar when viewed from the side. A t0p view only revealed diffuse areas of gray with very few distinct lines. In this case, the measure- ments taken were those where the band met the bottom of the agar strip. With the higher temperatures, the bands were essentially vertical and could be measured without removing the strip from the block. Single Diffusion One set of reaction wells was prepared with the anti- serum incorporated directly in the agar (Table 4). This was a form of single diffusion since the antibody concentration remained essentially constant compared to the gradient established when the virus diffused into the agar strip. Two types of Single diffusion were tested: one with equal volumes of antiserum and agar, and the other with twice as much agar. The results were similar for both, but the latter developed an extra line within oneawide band after 48 hrs. . . Broad areas of precipitate were formed when only the virus was allowed to migrate, while single and distinct bands usually resulted from double diffusion.. The latter was more desirable becauSe of the ease in measuring. One should note that in the Single diffusion experiment it was necessary to have the virus contained in a liquid 25 +++ + +1 + 9:. a a. 3.23.... H. a a. e i i .2 a s. 35...... I 1 ++ ++ 9:. a. a. 3:3...» .lilni .H' E .2. e. .. 38...... + + + . + 2;. z a. 3:33.. _ 3. H. I_il b an. an... an as soc-Lean} roe. . re... a .2... . so... a "523 .32 d. :— 32... 39:3 29.; 29.; "3.25... Ce? .5 c. ecu-L033... 3.9.3.»...- 5.) :3 8.3:? 8.; 3.3.1 .c .30... 26 since the use of a gel would result in antiserum diffusing into the antigen reservoir, thereby creating double diffusion. Reference Titration In order to establish a standard curve for the virus titration using precipitation, it was necessary to determine the quantity of virus in the allantoic fluid mixtures. This was accomplished using the method of Reed and Muench (1938). Five embryos were used with each dilution,.and all embryos found dead after 24 hrs had noticable hemorrhage, dwarfing, and edema. With L279, dilutions of 10'“ or lower killed all of the embryos, while those of 10"7 and higher did not affect any. The titer was 105’5 embryo lethal doses per 0.2 ml, or 106'2 ELD5 in 1.0 ml. The titer for L278 was 10°°5 ELD50 0 per 1.0 ml. The same results were obtained using the Spearman- Karber method (Harris, 1964), and with the Pizzi (1950) formula the standard error of the log ELDSO was i 0.109. Experimental Titration When various concentrations of virus were placed in the wells of the diffusion block and allowed to react against a constant amount of antiserum, precipitate formed at various points along the agar strip (Table 5). The more concentrated the virus, the farther the bands were from the antigen well. In the higher concentrations, several bands were evident, while the diffusion of less concentrated virus resulted in only a single band. 27 seed m .eumsoq comm .QSoa m .asnomdpad 3539a «63. same as 33.; 6.0 o mm.o m.o «nod w mm.o u m.o m.o m n.e o : m.o muoa a m.o : m.o m.o a m.a o.m mm.o m.o ages a mN.o u m.o m.o e m.oH n.a n m.o Huoa a n.o u m.o n.o m o.ma n.ma nm.o ,m.o 66H x mm.o u m.o m.o a m.ra n.ea u m.o cos a m.c n m.o n.o m o o u m.o cos a m.o m.o n.o n m e o m.o m.o n n m.o m.o a has «ms was «a .Hs. .Hs- .Hae a as. .Hs. Ass. .Hs. Ass. masseuse osaasm name means sunfish news sanomausa Has: ..aoapcnpap Mooan Scamsmudc scan nuances can .mpscpccoa no soapdmoasoo Honda .n panda 28 To insure that none of the bands used in the calculations were due to allantoic fluid, a control was established in which normal allantoic fluid was added to agar in a ratio of 1:1 and tested against the various antisera. The only visible reaction after 5 days was a very faint (+) line approximately 1 mm wide and located 5.5 mm from the antibody well. This was observed with both antiserum.B and C and was narrower in the latter. Since the leading edge of.the precip- itate was always more than 7.0 mm from the antigen well, non-specific bands should not interfere in a titration. When the virus was diffused, the distance of the leading edge of precipitate was checked at various time intervals. When plotted in a semi-log fashion against the titer, a linear relationship was observed (Fig. 7). All of.the lines tended to converge in one area which represents the "equiv- alence point”. This is where the antigen and antibody first meet and the dilution at which they are in Optimal proportions. The area to the right of this point is the region of antigen excess, as evidenced by the band moving toward the antibody well. The area between the top and bottom lines in this graph simultaneously represents a distance and a time interval, and is thus equivalent to velocity. By measuring the vertical distance between two lines and then dividing by the time difference, a velocity value can be determined. When these are calculated for several concentrations and then plotted against titer, another straight line relationship is observed (Fig. 8). i DISTANCE (mm) VELOCITY (mm/60 hrs) 21 18 15 12 29 . /A120 // i ‘ A 0 O 72 155 A 156 TITER (ELD /ml) Figure 7. Relationship of végus concentration in reservoir to distance of leading edge of precipitate. I . 105 156 TITER (ELD o/m1) Figure 8. Effect of virus cgncentration on velocity of precipitate migration. 30 Statistical Apglysis To determine the relative precision of the block titra- tion method, one virus sample was tested 15 times. The antigen was a 10-1 dilution of L279 with a final titer of 105°2 ELDSo/ml. Tests were run in a % inch agar strip incubated at 37°C for 75 hours prior to measuring (Table 6). A standard millimeter rule was used to determine the distance of the leading edge of precipitate from the antigen well, measured to the nearest 0.2 mm. Results are graphically represented in Figure 9. The mean of the 15 tests was 9.57 mm, while the mode and median were 9.4 mm. Sample standard deviation was 0.4 mm (Parratt, 1961). 31 Table 6. Tabulation of the results from replicate titration experiment. Value Frequency Deviation (f)(dev.) (f)(dev.)2 (x1) (r) (a...) 9.0 l -0.57 0.57 0.3249 9.2 1 -0.37 0.37 0.1369 9.4 7 -0.17 1.19 0.2023 9.6 3 +0.03 0.09 0.0027 10.0 1 +0.43 0.43 0.1849 10.4 2 +0.83 1.3778 1.66 Mean = XXI/n = 1&306/15 = 9057 made 3 90” Median = 9.4 Sample Variance =21(X1-i)2/n-l = 2.2295/14 = 0.1593 Sample Standard Deviation = (sample variance)% = 0.4 Mean Deviation =21|dev.l /n = 0.29 95% Confidence Interval = 9.375239.77 FREQUENCY .40 .20 32 l 9.0 9.2 9.4 9.6 9.8 10.0 10.2 10.4 DISTANCE (mm) Figure 9. Histogram of frequency versus distance of the leading edge of precipitate from the antigenlwell. DISCUSSION The results from diffusing duck hepatitis virus against various antisera essentially confirmed the findings of Murty and Hanson (1961). Two lines of precipitate were Specific for the virus, with a third appearing occasionally. Since the third line lay on or immediately adjacent to one of the specific lines, this may actually have been atresolution of one line into two. However, it was noted that the intensity of this third line was inversely prOportionalxto the degree of purity of the antiserum, and since the majoritylof par- ticles in normal allantoic fluid are about the same Size as the virus (Nazerian, 1960), the third line most probably was due to embryonic impurities. Though the original gel precipitation study of DHV was done several years ago, no attempt has been made prior to this time to apply the data quantitatively. In all of the published works on DHV where a titer had been determined (Hanson and Alberta, 1956; Rao gp_gl., 1958; Pollard and Starr, 1959; Kaeberle 22.21., 1961; Hwang and Daugherty, 1962; Sazawa pp,§l., 1963; Fitzgerald and Hanson, 1966), the only method used was that of quantal titration in chicken embryos. Doing this for each phase of an experiment involves a very large number of embryonating eggs, an egg incubator, complete sets of virus dilutions, and daily candling for 6-7 days. The diffusion method suggested here only requires a plastic diffusion block, one set of dilutions, and antiserum. 33 34 The incubation time would be shortened to 5 days and the block could be kept in a standard laboratory incubator. The principle behind the apparatus used in this study is that diffusion results from a difference in concentrations. In any closed system, molecules will migrate from one area to another until they are in equilibrium, or their concentra- tions are equal. In an Ouchterlony plate, about 0.1 ml of a substance is placed in a depression in the agar.. If there is 20 ml of agar, diffusion will occur because the well concentration, Cw’ is greater than the concentration of the substance in the agar, C and the a' a’ final concentration, Cf, will be 0.1/20 x Cw’ or 5% of the At equilibrium, Ci = C original concentration in the well. A similar ratio would exist for most macro and micro plate tests. In a tube test with equal volumes of antiserum, agar, and antigen, Cf would be 33% of Cw. r~ In the diffusion block method, Cf = l/2.25 x Cw or 44%, since there was 1.0 ml in each well plus 0.25 ml of agar in the slot. These figures only represent approximations, Since as a precipitate forms, the molecules involved are no longer available for diffusion; more molecules are then fed into the area, with a pr0portionate decrease in Ca’ Comparatively, however, the concentration available for reaction is the highest in the block design and it should thus have the greatest sensitivity. This sensitivity could easily be increased by.changing the proportions in a block of the same general design. A much narrower reaction Slot could be made, and it could fill 35 itself with agar by capillary action. Smaller reservoirs would result in the use of less serum, one ofithe limiting factors in this series of experiments. Examining and recording the positions of lines could be done with the aid of a low power microsc0pe, which.could also be used to photograph the lines so that readings could be taken from an enlarged print. A diffusion titration in such a block would thus have the advantage of increased.sensitivity and decreased reactant volume and incubation time. Dilutions would cover a space of several logs, instead of 1/2...l/4... l/8...etc. as in the plate and tube tests. . To determine the titer, two alternatives are available. One can either measure the distance of the leading edge of precipitate and compare it to the standard curve for that specific time, or take measurements at two different times and calculate the velocity for comparison with another curve (Figures 7 and 8). Both velocity and distance exhibited a linear relationship to concentration as expected (Kabat and Mayer, 1964). The velocity curve, however, represents the average of several distance curves and should be more reliable. Standard curves such as these would have to be estab- lished for each antiserum. If enough serum were available, one curve would be sufficient for up to hundreds of tests. Of the antisera tested here, B was the most efficient with regards to distinctness of precipitate and antibody titer. Absorbing it with normal allantoic fluid minimized 36 the non-specific lines, but simultaneously exerted a diluting effect on the Specific portion of the antiserum resulting in decreased intensity. Since the non-Specific lines appear between those which are Specific, they Should not interfere in a test of this nature and absorption is not necessary. The antiserum prepared using virus purified on the anion exchange resin gave the same type of curve for titer versus distance, but the lines were not as distinct. Trying a number of different resins and using a more concentrated virus suSpension should give an antiserum equal to or better than those now in use. Several of the environmental variables for the diffusion block reaction have been tested. The optimalicombination involved double diffusion in a 2 inch agar strip with the ‘ reactants in the form of a gel. This should be allowed to react at 37°C for a period of 5 days. Using this information plus the data from antiserum B given in figures 7 and 8, a hypothetical titration would be conducted in the following manner. The sample would be combined directly with agar in a 1:1 ratio and placed in well #1. A measurable line would form if the titer in.the well was 10°'5 to 106’0 EDD 50' 10-2, 10-3, and 10°“ dilutions of the sample plus agar to The next three wells would have detect a titer greater than 10°. The next well would contain agar plus equal volumes of the sample and a known virus suspension having a titer of 105‘0. If the sample titer were 10°'°, the minimum titer of this mixture would be 5.5 x 10° and would be detectable. The remaining wells could then serve 37 as controls if they contained antiserum versus saline, virus vs. saline, and a known virus concentration vs. the antiserum used in the titration. The measurable limits of such a system would be approximately 10°'° to 1010 ELDSO per ml of sample. If one had an estimate of the titer or desired a more accurate determination after the initial results, a series of dilutions within one or two logs of the estimate would be prepared and the findings from all of them used to find the average. I In the preparation of the standard curves it should be noted that any acceptable system could be used for the reference titration. The method of Reed and Muench and that of Spearman-Karber were used here because they are the most widely used. The former, however, has been the object of a number of critical reviews such as that of Finney (1959), who observed that ”no satisfactory measure of precision is available" and doubts that their basic assumptions are all valid. Basing the standard curves on a more accure reference system would eliminate some of these objections. .This would be feasible even if the better method were more difficult since it need be done only once for a large number of titra- tions. In the statistical analysis of replicatehtitrations, results tended to cluster about a single point. Results and calculations are given in table 6. It will be noted that the 38 mean deviation is the average deviation before squaring, and is thus less sensitive to large fluctuations than the commonly used standard deviation. A small mean deviation, such as the calculated 0.29, indicates measurements are closely grouped and the distribution has a sharp peak (Parrat, 1961). Two measurements (both 10.4 mm) were questionable since they were so far from the average value (9.5714 mm) and their pattern of lines was much less distinct than the others. This could have been caused by putting too much virus in the reaction well, or having too little agar in the slot. In either case, the two values lie outside of the 20- range where P = .95, so there is at least a 95% probability that a normal measurement will not be equal to or greater than 10.4. These two values may also be rejected on the basis of the 2.5d (Skoog and West, 1966), ”Q" (Dean and Dixon, 1951), and 't' (Moore pp,§l., 1951) tests for the significance of outly- ing values. If these two measurements were omitted, the remainder of the data would have the following parameters: Mean = 9.45 variance 2 0.07 ‘Standard Deviation a 0.27 Mean Deviation = 0.20 These values are more precise than those given previously and indicate that the deviation could be much lower than expected. One way to obtain more accurate calculations would be to increase the number of observations substantially. 39 There are some tests reported to have a higher sensitivity than that in the block test, but these use nitrogen determina- tions (Heidelberg and Kendall, 1935; Rappaport, 1957) or cptical density readings (Schmidt, 1957). They have disadvan- tages in time, equipment expense, complex purifications and chemical analyses, and the fact that non-Specific precipita- tion will give erroneous results. The apparatus and method presented in this study con- stitute an attempt to avoid some of these difficulties while achieving high sensitivity with some degree of precision. Its application should be especially pertinent when used in conjunction with viruses for which the quantitative methods are quite limited, such as with duck hepatitis virus. l. 3. SUMMARY Two antigenic components specific for duck hepatitis virus, strain L278, were detected using gel diffusion. Duck hepatitis virus was adsorbed to an ion exchange resin in an attempt at purification. Presence of the virus was confirmed with precipitation tests and electron microscopy. Apparatus necessary for a new form of gel diffusion was designed and constructed. Several variables were tested for the reactions occurring in the apparatus, including temperature, time, length of reaction well, type of diffusion, and physical state of the reactants. A new method for the quantification of duck hepatitis virus was established in which the position of precip- itate in a reaction well was pr0portional to the virus titer. Its principle and reliability were discussed, and suggestions offered for possible improvements. 40 BIBLIOGRAPHY Aladjem, F. and Lieberman, M. 1952. The antigen-antibody reaction. I. The influence of sodium chloride concentra- tion on the quantitative precipitin reaction. J. Immunol., 62: 117-130. Aladjem, F., Jaross, R., Paldinu, R., and Lackner, J. 1959. The antigen-antibody reaction (III). J. Immunol., 83: 221-231. Allison, A. C. and Humphrey, J. H. 1959. Estimation of the size of antigens by gel diffusion methods. Nature, 181: 1590-1592. Armitage, P. 1957. Studies in the variability of pock counts. J. of Hygiene, 55: 564-581. Asplin, F. D. and McLauchin, J. D. 1954. Duck hepatitis virus. Vet. Record,‘§§: 456-458. Asplin, F. D. 1958. An attenuated strain of duck hepatitis virus. Vet. Record, 19: 1226-1230. Bachrach, H. L., Callis, J., Hess, W., and Patty,,R. 1957. A plaque assay for foot and mouth disease virus and kinetics of virus reproduction. Virology, 4: 224-236. Becker, E. L. and Neff, J. C. 1959. Antigen-antibody reactions in agar. J. Immunol., 83: 571-581. Belyavin, G. 1955. The direct floculation of influenza virus. Lancet, 1, 698-700. . 1957. The influenza virus floculation reaction as a method of antigenic typing. J. of Hygiene, 55: 281-289. 41 42 Bjorklund, B. 1952. Specific inhibition of precipitation as an aid in antigen analysis with gel diffusion method. Proc. Soc. Exptl. Biol. & Med., 12: 319-324. Le Bouvier, G. L. 1957. Poliovirus precipitins - a study by means of diffusion in agar. J. Exptl. Med., 828: 661-675. Campbell, D. M., Garvey, J. S., Gremer, N. E. 1963. ”Methods in immunology," 1st ed., W. A. Benjamin, Inc., New Ibrk. Cruickshank, R. 1963. "Modern trends in immunology," vol. I, Butterworth and Co., London. Cunningham, C. H. 1966. ”A laboratory guide in virology," 6th ed., Burgess Publishing Co., Minneapolis. Dixon, W. J. and Massey, F. J. 1957. “Introduction to statistical analysis,“ 2nd ed., McGraw-Hill Book Co., New Ibrk. Dulbecco, R. 1952. Production of plaques in monolayer tissue culture by Single particles of an animal virus. Proc. Nat. Acad. Sci., 28: 747-752. Dulbecco, R. and Vogt, M. 1953. Plaque formation and isola- tion of pure lines with poliomyelitis virus. J. Exptl. Med., 22: 167-182. Dulbecco, R. and Freeman, G. 1959. Plaque production by polyoma virus. Virology, 8: 396-397. Elek, S. D. 1948. The recognition of toxigenic bacterial strains in vitro. Brit. Med. J., 1, 493-496. ____. 1949. The serological analysis of mixed floculating systems by means of diffusion gradients. Brit. J. Exptl. Paths, lg: “’84. 43 Fabricant, J., Richard, C., and Levine, P. 1957. The pathology of duck hepatitis virus. Avian Dis., _]_.: 256-275. Feinberg, J. G. 1957. Identification, discrimination and quantification in Ouchterlony gel plates. Int. Arch. Allergy & App. Imm., 2;: 129-152. Finney, D. J. 1959. The design and analysis of an immunolog- ical assay. Acta Micro., 8: 341-368. Fitzgerald, J. E., Hanson, L., and Winegard, M. 1963. Cytopathic effects of duck hepatitis virus in duck embryo kidney cell culture. Proc. Soc. Exptl. Biol. 2 Med., ;;2: 814-816. Fitzgerald, J. E. and Hanson, L. 1966. Certain prOperties of a cell-culture modified duck hepatitis virus. Avian Dis., 22: 157-161. Friedewald, W. M. and Pickels, E. G. 1944. Centrifugation and ultrafiltration studies on allantoic fluid prepara- tions of influenza virus. J. Exptl. Med., 12: 301-317. Golub, 0. J. 1948. A single dilution method for the estima- tion of LDS titers of the psittacosis-DOV group of 0 viruses in chicken embryos. J. Immunol., 52: 71-82. de St. Groth, S. F. 1938. An improved assay for the infectivity of influenza viruses. J. Hygiene, 88: 151-162. Gussoni, C. 1964. Quantitative measurements of low concentra- tions of allergen by graded antiserum adsorption in gel diffusion. Int. Arch. Allergy & App. Immunol., 84: 1-16. 44 Hagen, W. A. and Bruner, D. W. 1961. "The infectious diseases of domestic animals," 4th ed., Comstock Publishing Assoc., Ithaca, New Ybrk. Harris, R. J. 1964. ”Techniques in experimental virology," Academic Press, London. Hanson, L. and Alberta, J. 1956. Virus hepatitis in ducklings. J. A. V. M. A., 288: 37-38. Hanson, L. E. 1958. Histological lesions in ducks with virus hepatitis. Am. J. Vet. Res., 22: 712-718. Hanson, L. E., Rhoades, H. C., and Schricker, R. L. 1961. PrOperties of duck hepatitis virus. Poultry Sci., 22: 1412-1413. Heidelberger, M. and Kendall, F. E. 1935. A quantitative theory of the precipitin reaction. J. Exptl._Med., 88: 467-483. Hsiung, G. D. and Melnick, J. L. 1955. Plaque formation with polio, coxsackie, and orphan viruses in bottle cultures of monkey epithelial cells. Virology, ;:L 533-535. Huang, C. H. 1943. ,A visible method for titration and neutralization of viruses on the basis of pH changes in tissue culture. Proc. Soc. Exptl. Biol. 2 Med., 52: 160-161. . Humphrey, J. H. and White, B. G. 1964. “Immunology for students of medicine," F. A. Davis Co., Philadelphia. Hwang, J. and Dougherty, E. 1962. Serial passage of duck hepatitis virus in chicken embryos. Avian Dis.,‘8: 435-440. 45 _____. 1964. Distribution and concentration of duck hepatitis virus in inoculated ducklings and chicken embryos. Avian Dis., 8: 264-268. Isaacs, A. and Donald, H. B. 1955. Particle counts of haemagglutinating viruses. J. Gen. Microbiol., l8: 241-247. Jensen, K. E. and Francis, T. 1953. Antigen-antibody precip- itates in solid medium with influenza virus. J. Immunol., 12: 321-325. Kabat, E. A. and Mayer, M. 1964. "Experimental immunology,” 2nd ed., Charles C. Thomas, Pub., Springfield, Illinois. Kaeberle, M. L., Drake, J. W., and Hanson, L. E. 1961. Cultivation of duck hepatitis virus in tissue culture. Proc. Soc. Exptl. Biol. & Med., 228: 757. Korngold, L. and van Leewen, G. 1957. The effect of anti- genic molecular weight on the curvature of precipitin lines in the Ouchterlony technique. J. Immunol., 18: 172-177. Leuker, D. C. and Crowle, A. J. 1963. Secondary precipita- tion in immuno-diffusion tests. Int. Arch. Allergy & App. Immunol., 22: 65-80. Levine, P. P. and Fabricant, J. 1950. A hitherto undescribed disease of ducks in N. America. Cornell Vet., 22: 71-86. MacPherson, L. W. and Avery, R. J. 1957. Duck hepatitis virus in Canada. Canad. J. Comp. Med., 2;: 27-31. Mallmann, V. H. 1960. The demonstration of a cell-attachment- and-growth factor and comparative studies of infectious 46 bronchitis virus before and after separation from the factor. Ph. D. thesis, Michigan State University, East Lansing, Mich. Moore, F. J., Cramer, F. R., and Knowles, R. G. 1951. "Statistics for medical students," The Blakistron Co., New Ibrk. Murty, D. and Hanson, L. 1961. A.modified microgel diffusion method and its application in the study of duck hepatitis virus. Am. J. Vet. Res., 88: 274-278. Nazerian, K. 1960. Electron microscoPy studies of the virus of infectious bronchitis and erythrocytes agglutinated by trypsin modified virus. M. S. thesis, Michigan State University, East Lansing, Mich. Neff, J. C. and Becker, E. L. 1957. Antigen-antibody reac- tions in agar (III). J. Immunol., 28: 5-10. Ouchterlony, 0. 1948. In vitro method for testing the toxin producing capacity of Diphtheria bacteria. Acta. Path. Micro. Scand., 82: 186-191. Oudin, J. 1946. Methods d'analyze immunochemique par precipitation Specifique en milieu galifie. Compt. Rend. Acad. Sci., 888: 115-116. Parratt, L. G. 1961. "Probability and experimental errors in science,” 1st ed., John Wiley 2 Sons, Inc., New York. Pizzi, M. 1950. Sampling variation of the fifty per cent endpoint determined by the Reed and Muench (Behrens) method. Human Biol., 88: 151-190. Pollard, M. and Starr, T. 1959. PrOpagation of duck hepatitis virus in tissue culture. Proc. Soc. Exptl. Biol. 2 Med., 22;: 521-524. 1+7 Pollard, n. and Starr, T. 1960. PrOpagation of duck hepatitis virus in tissue cultures prepared with collagenase. J. Bact., 82: 571-572. Preer, J. R. 1956. A quantitative study of a technique of double diffusion. J. Immunol., 11: 52-60.. Rao, S. B., Srinivasan, V. V., and Iyer, S. G. 1958. Studies on duck hepatitis virus. Indian Vet. J., 22: 534-539. Rappaport, I. 1957. The antigen-antibody reaction. J. Immunol., 18:. 247-255. Reed, L. and Muench, H. 1938. A simple method for estimating fifty per cent endpoints. Am. J. Hyg., 81: 493-397. Reuse, U. 1959. Virusbiologishe untersuchungen bei der entenhepatitis. Zentralblatt fur Veterenarmedizin,‘8: 209-248. Richter, W. R., Rdzok, E. J., and Moize, S. M. 1964. Electron microscopy of viruslike particles associated with duck hepatitis virus. Virology, 82:. 114-116. Rubin, H. and Temin, H. 1958. Infection with the rous sarcoma virus in vitro. Fed. Proc., 11: 994-1003. Rubin, H. 1960. An analysis of the assay of rous sarcoma cells in vitro by the infective center technique. Virology, 12: 29-49. Sazawa, R., Sugimori, T., Shimizu, T., and Watanabe, M. 1963. Isolation and behavior of duck hepatitis:virus in Japan. Tokyo Nat. Inst. An. Health Quart., 2: 71-76. Schmidt, W. C. 1957. The quantitative precipitin reaction of type 19 M protein antigen of group a streptococci and antistreptococcal rabbit sera. J. Immunol., 18: 178-184. 48 Skoog, D. A. and West, D. M. 1966. ”Fundamentals of analytical chemistry,“ 1st. ed., Holt, Rinehart, and Winston, New Yerk. Smadel, J. E., Rivers, T. M., and Pickels, E. G. 1939. Estimation of the purity of elementary bodies of vaccinia. J. Exptl. Med., 12: 379-385. Tyrell, D. A. and Valentine, R. C. 1957. The assay of influenza virus by hemagglutination and electron microscopy. J. Gen. Microbiol., 18: 668-675. White, D. 0. (ed.). 1963. ”A post graduate course in tissue culture," The Cell Culture Soc. of Victoria, Victoria, Australia. Wilson, M. W. and Pringle, B. H. 1954. Experimental studies of the agar plate precipitin test of ouchterlony. J. Immunol., 12: 232-243.