Monoclonal antibody to rat uterine peroxidase and its use in identification of the peroxidase as being of eosinophil origin

Biochimica et Biophysica Acta, 802 (1984) 399-406

399

Elsevier

BBA 21926

M O N O C L O N A L ANTIBODY T O RAT UTERINE PEROXIDASE AND ITS USE IN IDENTIFICATION OF T H E PEROXIDASE AS BEING OF E O S I N O P H I L ORIGIN H U G H S. KEEPING * and C. R I C H A R D LYTTLE **

Department of Obstetrics and Gynecology, Division of Reproductive Biology, University of Pennsvh,ania, 36th and Hamilton Walk, Room 306 G3, Philadelphia, PA 19104 (U.S.A.) (Received August 29th, 1984)

Key words." Eosinophil; Peroxidase origin; Monoclonal antiboclv," (Rat uterus)

Peroxidase was purified from uteri of estrogen-treated rats by calcium chloride extraction, affinity chromatography on concanavalin A-Sepharose and hydrophobic interaction chromatography on phenyl-Sepharose. An overall purification of greater than 1700-fold was achieved with a final recovery of 27%. Monoclonal antibodies to peroxidase were subsequently prepared by immunization of male C 5 7 B L / 1 0 J mice with the highly purified peroxidase from rat uterus. Spleen and lymph node cells from the mice were fused with S p 2 / 0 - A g 14 mouse myeloma cells. The resultant hybrid cells were screened for production of antibody using a solid-phase, double antibody radioimmunoassay. The mature rat spleen, shown previously to be abundant in eosinophils, contains high peroxidase activity. Spleen peroxidase purified by the same procedure as the uterine enzyme cross-reacted with a monoclonal antibody, designated IgG-107B, used in all subsequent studies. Peroxidase extracted from isolated rat eosinophils also cross-reacted with the antibody and yielded identical titers as the spleen and uterine peroxidases. Spleen, uterine and horse eosinophil peroxidase had the same apparent molecular weight, 57 000, as determined by sodium dodecyl sulfate-urea polyacrylamide gel electrophoresis. Following electrophoretic transfer to nitrocellulose, spleen, uterine and eosinophil peroxidase reacted with monoclonal antibody, using an immunoblotting technique. These results provide biochemical and immunological evidence that the majority of the calcium chloride-extractable peroxidase activity from the uteri of estrogen-treated rats is derived from infiltrating eosinophils.

Introduction The large increase in the activity of peroxidase in the uterus after treatment of ovariectomized, immature, or mature rats with estrogen, as initially reported by Lucas et al. [1], has led to the suggestion this enzyme may be a specific marker protein in tissues responsive to estrogen [2,3]. However, in addition to the induced peroxidase in the luminal and glandular epithelium, an exogenous per* Present address: Department of Medicine, University of Pittsburgh, Pittsburgh, PA 15213, U.S.A. ** To whom correspondence should be addressed. 0304-4165/84/$03.00 ~ 1984 Elsevier Science Publishers B.V.

oxidase is present in the stromal and myometrial layers of the uterus brought by eosinophils after estrogen treatment [4]. The uterine and eosinophil enzymes are similar with respect to kinetic, chromatographic and spectral properties [5,6]. King et al. [7] have concluded that the majority of peroxidase activity in calcium chloride extracts of the uterus originates in the eosinophils, as demonstrated indirectly by a correlation of histochemical and biochemical assays. Olsen and Little [8] have recently obtained rabbit antibodies to rat uterine peroxidase that had been purified to greater than 95% homogeneity [9]. These antibodies strongly inhibited rat uterine and eosinophil peroxidase

400 activity, while rat mveloperoxidase activity was unaffected. However, conventional antibodies isolated from sera have some limitations, since they may consit of several different specificities to different antigenic sites. Therefore, in order to circumvent these problems, we have in the present study, prepared a mouse monoclonal antibody to rat uterine peroxidase. This antibody cross-reacts with rat eosinophil peroxidase and peroxidase isolated from rat spleen, but does not cross-react with bovine lactoperoxidase, as determined by two different methods. We conclude that calcium chloride-extractable peroxidase from the rat uterus is derived from invading eosinophils after estrogen treatment. A preliminary report of portions of this work was presented at the 64th Annual Endocrine Society Meeting, June 1982 (Proc. 64th Annu. Endoc. Soc. Meeting, p. 303, Abstr. 893). Materials and Methods

Freund's adjuvants (complete and incomplete) were purchased from Miles (Elkhart, IN). Concanavalin A Sepharose 4B and phenyl Sepharose CL-4B were obtained from Pharmacia (Piscataway, N J). Hypoxanthine, aminopterin, thymidine, polymyxin B sulfate, molecular weight standards and lactoperoxidase were purchased from Sigma (St. Louis, MO). Female Sprague Dawley rats were obtained from King Animals (Madison, WI) and male C57BL/10J mice were purchased from Jackson Laboratory (Bar Harbor, ME). 12SI-labeled rabbit antimouse F(ab')2 was kindly provided by Dr, M. Cancro, Department of Pathology, University of Pennsylvania. Mouse myeloma cell line Sp2/0-Ag14 were kindly supplied by Dr. Thomas J. McKearn, Cytogen Inc. (Princeton, N J). Highly purified horse eosinophil peroxidase was generously supplied by Dr. Seymour J. Klebanoff, University of Washington (Seattle, WA). All other chemicals were reagent grade or better. Preparation of uterine CaCl 2 extracts. Mature female rats were injected subcutaneously with estradiol-17/3 (100/~g daily for 3 days) in 0.1 ml of 0.9% NaCI/50% ethanol. All animals were killed 20 h after the last injection. All subsequent procedures were performed at 4°C unless indicated otherwise. The uteri were dissected free of fat and

mesentery, blotted, weighed and cut into small pieces with scissors before homogenization (50 m g / m l in 10 mM Tris-HC1 (pH 7.2)) using a Polytron PT-10 (three 10-s bursts at speed setting 6). After removing samples for protein determination [10], the homogenate was centrifuged at 40000 × g for 30 min. Peroxidase enzyme was solubilized [2] by homogenizing the 40000 × g pellet in 10 mM Tris-HC1 (pH 7.2)/0.5 M CaCI2/10 mM phenylalanine. The homogenized pellet was centrifuged at 40000 x g for 30 rain to obtain a supernatant containing most of the peroxidase in the tissue. Isolation of eosinophils. Mature female rats received weekly intraperitoneal injection of polymyxin B sulfate (0.25 mg in 0.25 ml of 0.9% NaC1) after the method of Pincus [11]. Peritoneal eosinophils were obtained by the method of Jong et al. [12]. The percoll-purified eosinophil preparations contained 80 90% eosinophils, as determined by the method of Luna [13]. The remaining cells consisted of heterophils, monocytes, mast cells and a few fibroblasts. The cells were washed twice in Dulbecco's phosphate-buffered saline, then homogenized in 10 mM Tris-HC1 (pH 7.2)/0.5 M CaC12, before centrifugation at 40000 x g for 30 min to obtain the peroxidase activity. Measurement of peroxidase actwity. The peroxidase assay mixture (3.0 ml) comprised 13 mM guaiacol/0.33 mM H202/10 mM Tris-HC1 (pH 7.2)/0.5 M CAC12/0.01 1.0 ml sample. The linear increase in absorbance at 470 nm resulting from the oxidation of guaiacol was followed in a Gilford model 250 recording spectrophotometer at 25°C. 1 enzyme unit is defined as the amount of enzyme required to produce an increase of 1 absorbance unit per rain. Con A-Sepharose. Affinity chromatography was conducted at 4°C as previously described [14]. A total of 26730 units (1650 ml) of 0.5 M CaC12 uterine extract were pumped at 10 m l / h onto a 40 ml bed volume of Con A-Sepharose, which had been packed and washed with 10 mM Tris-HC1 (pH 7.2)/0.5 M CaC12/10 mM phenylalanine. After the sample volume was loaded, the column was washed with this same buffer until the absorbance at 280 nm had returned to a baseline value. Bound protein was eluted with 10 mM Tris-HCl (pH 7.2)/2.5 M CAC12/0.5 M c~-methyl-

401 mannoside. The pooled eluted peroxidase activity (fractions 118-210) consisted of 8858 enzyme units (32% recovery). Hydrophobic interaction chromatography was performed at room temperature, since better recoveries of enzyme were obtained than at 4°C. 39 ml (8424 units) of the Con A-Sepharose purified peroxidase was loaded onto a 10 ml bed volume of phenyl-Sepharose previously packed and washed with 10 mM Tris-HC1 (pH 7.2)/0.5 M CaCI 2 until the absorbance at 280 nm had returned to a baseline value. The column was then washed with this same buffer containing 35% ethylene glycol until the protein absorbance and peroxidase activity had returned to baseline values. The amount of ethylene glycol was subsequently increased to 70% (and the fraction size was decreased from 2.6 ml to 1.3 ml) to elute the major peak of enzyme activity. Fractions 86-95 (2304 units, 27% recovery) were used to immunize the mice. The final preparation had a specific activity of 662 units/mg protein. Immunization. The peroxidase used for immunization of male C57BL/10J mice was purified from uteri of estrogen-treated rats. The phenyl-Sepharose-pooled enzyme activity peak was dialyzed against 10 mM Tris-HC1 in order to remove the CaC12 and ethylene glycol which proved to be toxic to the mice. After dialysis, the precipitated protein (50 /~g) was emulsified via sonification with an equal volume of Freund's complete adjuvant and injected intradermally at multiple sites. 16 days later, individual mice were given intraperitoneal injections daily for 3 days with 25/~g of protein in incomplete adjuvant. 1 day after the final injection, the mouse was bled and the spleen and lymph nodes were removed under sterile conditions. Monoclonal antibody preparation. Spleen and lymph node cells from an immunized mouse were isolated and fused according to standard methods [15]. Positive hybridoma cell lines were subsequently subcloned in semi-solid agarose as previously described [16]. The 107B clone, which was the source of the antibody used for the experiments described in this paper, was identified as being of the IgG class by methods described previously [17]. The cloned cells were expanded to T-75 flasks at which time the hybridoma cells were frozen [18] and the monoclonal antibody was iso-

lated from the culture medium by ammonium sulfate precipitation [19]. Double antibody radioimmunoassay. In the initial screening step, monoclonal antibodies to peroxidase were detected by a solid phase, indirect, double antibody radioimmunoassay. Aliquots (10 t~l) of highly purified peroxidase were added to 96-well polyvinyl microtiter plates, diluted with the 10 mM Tris-HC1 (pH 7.2, 100 t~l), and the plates were incubated with constant shaking for 30 min at 25°C. The plates were centrifuged at 1500 × g for 10 min in order to fix the protein to the bottom of the wells. The plates were then rinsed three times with tap water followed by one wash with phosphate-buffered saline containing 3% bovine serum albumin. Binding of peroxidase to the bottom of the plate was confirmed by assay of the enzyme in situ after addition of guaiacol and HzO 2. Aliquots (50 t~l) of hybridoma medium from each growing culture were added to the precoated wells and incubated for 2 h at 23°C. The wells were then washed four times with 200 t~l of phosphate-buffered saline containing 3% bovine serum albumin, after which 125I-labeled rabbit anti-mouse F(ab')2 (50 /~1, (2.5-5.0). 105 cpm) in this same buffer was added. After incubation for 2 h at 23°C, the wells were washed four times with tap water before determination of radioactivity. Blank binding was determined by substituting hybridoma culture medium instead of cell culture supernatants. All wells that gave values for 125|_ labeled F(ab')2 binding 3-fold or higher above the blank values (150-300 cpm/well) were considered positive and were subsequently expanded to larger plates.

Transfer of proteins to nitrocellulose and immunoblotting. Peroxidase protein, eluted from the phenyl-Sepharose column that was used for immunization, was separated by electrophoresis in the presence of 0.1% sodium dodecyl sulfate and 4 M urea as previously described [20]. The following proteins were used as molecular weight standards: rabbit muscle phosphorylase b (92500), bovine serum albumin (68 000), chicken albumin (43 000), bovine carbonic anhydrase (31000) and bovine B-lactalbumin (18400). After electrophoresis, a section of the slab gel was electrophoretically transferred to nitrocellulose as previously described [21]. After transfer, a portion of the nitro-

402

celllose sheet was stained for protein [22]. A duplicate nitrocellulose sheet was immersed in 100 ml of phosphate-buffered saline/3% bovine serum albumin (buffer 1) and incubated for 60 min at 37°C on a rocking platform. The sheet was then immersed in a fresh solution of buffer 1 containing monoclonal antibody diluted 1 : 50 and incubated for 2 h at 23°C before being washed with 100 ml of 10 mM Tris-HC1 (pH 7.4)/0.9% NaC1/0.2% SDS/0.5% Triton X-100/0.5% bovine serum albumin/0.01% NaN 3 for 20 min with five change~, [23]. The sheet was then immersed in buffer 1 containing 125I-labeled rabbit anti-mouse F(ab')2 (5-105 cpm/ml) and incubated for 2 h at 23°C before being washed as described above. The sheet was air dried, wrapped in plastic and exposed to Kodak X-OMAT RP-1 film at - 7 0 ° C for 60 h with a DuPont Cronex Hi-plus intensifying screen. Results and Discussion

Purification of peroxidase from the uterus A summary of the procedure used to isolate peroxidase from the uteri of estrogen-treated rats is outlined in Table I. Affinity chromatography on Con A-Sepharose resulted in a 1092-fold increase in specific activity from the initial homogenate with a 32% recovery relative to the CaC12 extract. The enzyme was further purified by hydrophobic interaction chromatography on phenyl-Sepharose. This subsequent step produced an overall purification of greater than 1700-fold with a final recovery of 27%. An assessment of the purity after this final procedure is shown by SDS-urea poly-

acrylamide gel electrophoresis (Fig. 1, lane 2). The major band of Coomassie-stained protein has an apparent M r of 57000, as determined by the procedure of Laemmli [20]. This value is larger than those reported by other investigators using the method of gel filtration: 40000 by McNabb and Jellinck [24] and Olsen and Little [9]; 50000 by Lyttle and DeSombre [2], but less than that (77 000) obtained by Wagai and Hosaya [25]. Although minor contaminant bands are present in the preparation, the uterine peroxidase was suitable for the production of monoclonal antibodies, since it was possible to select for the hybridoma cell line which made antibody to the M r 57 000 protein. Lactoperoxidase (Fig. 1, lane 1) appears quite heterogeneous, as reported previously [20] and may be due to artifacts of the electrophoretic procedure [27]. Nevertheless, the major protein band had an apparent M r of 78000, which closely agrees with the value (77500) given by Rombauts et al. [27]. The protein molecular weight standards also contained several minor bands, however, each standard was electrophoresed in separate lanes on the same gel in order to obtain an accurate estimation of their relative mobility (results not shown).

Preparation of monoclonal antibody Male C57BL/10J mice were immunized with the purified uterine peroxidase preparation. The spleen and lymph node cells of an immunized mouse were fused with Sp2/0-Agl4 myeloma cells. Approx. 1.5% (13/887) of the wells containing viable clusters of hybridoma cells produced an antibody that bound to the peroxidase preparation

TABLE I PEROXIDASE P U R I F I C A T I O N F R O M RAT U T E R U S Mature rats received 100 /~g of estradiol for 3 days before sacrifice 20 h after the last injection. Percentage recovery is based on peroxidase activity in the CaCI 2 extract, because solubilization increase total enzymic activity over that which can be assayed in the homogenate. Protein was determined by the method of Bradford [10]. Peroxidase activity was determined by guaiacol assay [2]. Procedure

Homogenate CaC12 extract Con A-Sepharose Phenyl-Sepharose

Protein

Peroxidase activity

(mg/ml)

U/ml

3.16

1.2

1.04 0.52 0.21

16.2 216 192

U/mg protein 0.38

15.6 416 662

% recov, -

100 32 27

purif. 1

41 1 092 1 742

403

1

2

3

4

1

2

3

92.5 68

43

31

18.4 F Fig. 1. SDS-urea polyacrylamide gel electrophoresis of purified peroxidase from rat uterus. Slab gels were prepared after the method of Laemmli [20], and the proteins were stained with Coomassie blue. Uterine and spleen peroxidase were purified by Con A-Sepharose and phenyl-Sepharose chromatography. 10 ~tg of protein of each peroxidase and of each molecular weight standard were applied to the gel. Lane 1, lactoperoxidase; lane 2, uterine peroxidase; lane 3, spleen peroxidase; lane 4, molecular weight standards. The numbers in the margin represent the Mr in kilodaltons of the standards. F, dye front.

from rat uterus. These 13 positive clones were transferred to plates containing irradiated Lewis rat thymocytes which served as a feeder layer and enhanced survival of the positive hybridomas, Seven of the 13 positive clones lost their ability to secrete antibody after 3 - 4 weeks in culture, a p h e n o m e n o n that has been previously described [28,29]. The remaining positive h y b r i d o m a cell lines were subcloned in semi-solid agarose by the m e t h o d of Kennett [16]. The 107B clone, which was the source of the antibody used in all subsequent experiments, was identified as being of the I g G class by methods described previusly [17]. Monoclonal antibody I g G - 1 0 7 B was isolated from the cell culture medium by a m m o n i u m sulfate precipitation [19] and then frozen in aliquots at - 7 0 ° C before use.

Fig. 2. Electrotransfer of peroxidase to ntrocellulose. Duplicate samples of the peroxidases in lanes 1-3 of Fig. 1 were electrophoretically transferred to nitrocellulose. The nitrocellulose sheet was stained after transfer with Coomassie blue according to the method of Burnette [22]. Lane 1, lactoperoxidase; lane 2, uterine peroxidase; lane 3, spleen peroxidase.

Specificity of monoclonal IgG-IO7B The specificity of monoclonal antibody IgG107B was determined by reacting the antibody with various peroxidases previously resolved by SDS-urea polyacrylamide gel electrophoresis and then subsequently transferred to nitrocellulose paper (Fig. 2). The autoradiogram of the imm u n o b l o t shows that the antibody is only b o u n d to the M r 57000 b a n d of rat and spleen peroxidase, while bovine lactoperoxidase is not recognized by the antibody. N o n e of the other protein bands in the partially purified spleen peroxidase or commercially obtained lactoperoxidase were recognized by the monoclonal antibody. W h e n the i m m u n o b l o t t i n g procedure is performed in the absence of the monoclonal antibody there is no binding of the t25I-labeled second antibody, dem-

404

onstrating that the radiolabeled second antibody is not nonspecifically bound to the peroxidase protein. An immunoblot was also performed with partially purified eosinophil peroxidase isolated from horse blood, as described by Jorg et al. [30]. The autoradiogram showed that the M r 57000 band of the eosinophilic enzyme cross-reacts with the monoclonal antibody. Since the cross-reacting band is the same M r as the peroxidase isolated from the rat uterus, these results demonstrate further that the peroxidase in the rat uterus is derived from eosinophils. Cross-reactivity of IgG-107B was also determined by the solid phase, indirect, double antibody radioimmunoassay. Equivalent antibody titers were obtained to the uterine and spleen enzymes, as well as peroxidase isolated from rat eosinophils, while bovine lactoperoxidase did not yield any discernable titer (Fig. 3). This finding is in agreement with Morrison et al. [31[, who found that antibodies to bovine lactoperoxidase did not cross-react bovine spleen and uterine peroxidase. The immature rat uterus has been shown previously to contain very little peroxidase activity [2] and is virtually devoid of eosinophils [7]. When a CaC12 extract of immature rat uteri is fractionated on phenyl-Sepharose, no measurable peroxidase activity is present in any of the fractions (Fig. 4A),

SPLEEN --ZUTERUS-I-EOSINOPHIL . ~ . LACTOPEROXIDASE -.I,2

%

-05

-12

- 1.8

-24

-30

-35

-42

Fig. 3. Mouse monoclonal IgG-IO7B antibody titer to peroxidase. Calcium chloride (0.5 M) extracts from rat spleen, uterus and eosinophils were loaded on phenyl-Sepharose. Aliquots (0.1 unit of peroxidase activity) from the peak of enzyme activity eluted off the column with 70% ethylene glycol were precipitated onto the microtiter plates. Bovine lactoperoxidase (0.1 unit) was also bound to the wells of the plates. Cell culture medium from hybridoma 107B was isolated by 50% a m m o n i u m sulfate precipitation, diluted in phosphatebuffered saline /3% bovine serum albumin and assayed for antibody binding by the double antibody radioimmunoassay.

nor did IgG-107B bind to any of the proteins eluted from the column, as determined by the double antibody radioimmunoassay. Fractionation of uteri from estrogen-treated rats resulted in a peak of peroxidase activity being eluted by 70% ethylene glycol and a concomitant peak of antibody binding occurring with the peroxidase activity~ as shown in Fig. 4B. A CaC12 extract of immature rat spleen yielded two peaks of peroxidase activity, eluted with 35 and 70% ethylene glycol respectively (Fig. 4C). However, antibody binding was limited to the peak of peroxidase activity eluted with 70% ethylene glycol. These results show that peroxidase is present in calcium chloride extracts of the rat uterus only after estrogen treatment and that the monoclonal antibody recognizes this enzyme in the hormonally stimulated uterus alone. In addition to eosinophils, the rat uterus also contains some neutrophils or heterophils [32], which contain myeloperoxidase, an enzyme previously shown to be biochemically [33,34] and immunologically distinct from eosinophil peroxidase [35]. Since the peroxidase we isolated from the rat uterus does not have the same molecular weight or number of subunits as that reported for myeloperoxidase [33,34], we feel that the antibody was not raised to the neutrophil enzyme. The antibody only recognized the spleen peroxidase activity eluted with 70% ethylene glycol, the same concentration which eluted uterine peroxidase (Fig. 4B) and rat eosinophil peroxidase, as described in the legend of Fig. 3. Thus, the antibody shows partial specificity in that it does not cross-react with any peroxidase specifies isolated from the rat. As well as the calcium chloride-extractable peroxidase of uterine tissue, rat uterine luminal fluid also contains a peroxidase [36]. This enzyme has different physical properties in that it is soluble and has a higher molecular weight than the uterine tissue peroxidase [5,37,38]. The cellular origin of this uterine fluid peroxidase is presently undetermined. Ultrastructural studies of estrogen-inducible endometrial peroxidase clearly demonstrate that the enzyme is synthesized and secreted into the uterine lumen [39]. It is possible that the luminal fluid peroxidase originates in the luminal and glandular epithelium, while the peroxidase in the tissue itself is derived from the infiltrating

405

I0~ >

TIoC 5oo I

i

35% El+h~ly

8,

1.0

i

70% Eth~Gly

0.8

6

I t

I I I I t

r~_

4

~

z

H

0.6 f

I

+ t I

S' _ I0

20 3'0 4'0 gO FRACTIONNUMBER

A

3

0.4

•4

0.2

-3

5

Y~

0

6'0

TloC 5oo I2'

~6

I

I

I I I

8-

x~-

o~

70% Elh Gly

f

+

I I I

I I I

I

35% EthGly

j

g

-8

1.8

.7

1.4

'6

•I

.0

I I

,2.

2.4

J

k

20 4'o 6'o 80 B

Acknowledgements H

The authors would like to thank Dr. Thomas McKearn for the help and guidance with the hybridoma techniques. This work has been been supported by NIH Grant HD-14695, the Rockefeller Foundation (RF-788082) and the Mellon Foundation.

-s 3 6

0.8

.4

0.4

-3

0

-2

I

'

I00

120

References

FRACTION NUMBER TIoCsoo Eth ly

>-

• 2.4

8

'

-7

E

I0'

s

~--~j 8

o.

2.0

1.6 ~ - 6

"' ~ 6 . . . .

o-----o--~

1.2 ~=

}p,

2 0 "

7

-

~'~'¢ 6~ ~

eosinophils. In preliminary experiments, we did not see any cross reactivity of IgG-107B with rat uterine luminal perixodase (Keeping, H.S. and Lyttle, C.R., unpublished data). Thus, the present results provide, for the first time, direct evidence that uterine peroxidase is identical to the enzyme found in eosinophils, based on it's similar molecular weight and cross-reactivity with eosinophil peroxidase using a monoclonal antibody. We conclude, therefore, that uterine tissue peroxidase is immunologically identical to eosinophil peroxidase and that the eosinophil enzyme accounts for the majority of the peroxidase activity in calcium chloride extracts of the estrogen-stimulated rat uterus.

20

40

c

FRACTIONNUMBER

,6o

,~ -5

0.4

3

0

.2

,~o

Fig. 4. Binding of monoclonal antibody lgG-107B to 0.5 M calcium chloride tissue extracts separated on phenyl-Sepharose: comparison between (A) immature rat uteri, (B) uteri from estrogen-treated rats and (C) immature rat spleen. Aliquots (10 t~l) of all fractions collected from the phenyl-Sepharose column were precipitated onto microtiter plates. Monoclonal antibody IgG-107B was added to each well and assayed for antibody binding by the solid phase, double antibody radioimmunoassay. Peroxidase activity of each fraction was measured by the guaiacol assay [2] and total protein in each fraction was estimated by absorbance at 280 nm.

1 Lucas, F.V., Neufeld, H.A., Utterback, J.G., Martin, A.P. and Stotz, E. (1955) J. Biol. Chem. 214, 775-780 2 Lyttle, C.R. and DeSombre, E.R. (1977) Proc. Natl. Acad. Sci. USA 74, 3162-3166 3 Jellinck, P.H., Newcombe, A. and Keeping, H.S. (1979) Adv, Enzyme Regul, 17, 325-342 4 Brokelmann, J. and Fawcett, D.W. (1969) Biol. Reprod. 1, 59-71 5 Hosoya, T. and Saito, T. (1981) J. Biochem. (Tokyo) 89, 203-215 6 0 l s e n , R.L., Flatmark, T. and Little, C. (1982) Biochem. J. 82, 91-94 7 King, WJ.., Allen, T.C. and DeSombre, E.R. (1981) Biol. Reprod. 25, 859-870 8 0 l s e n , R.L and Little, C. (1982) Biochem. J. 207, 613-616 9 0 l s e n , R.L. and Little, C. (1982) Acta Chem. Scand. B35, 1-4 10 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254 11 Pincus, S.H. (1978) Blood 52, 127-133 12 Jong, E.C., Henderson, W.R. and Klebanoff, SJ. (1980) J. Immunol. 124, 1378-1382 13 Luna, L.G. (1968) in Manual of Histologic Staining Methods of the Armed Forces Institute of Pathology (Luna, L.G., ed.), pp. 111-112, McGraw-Hill, New York 14 DeSombre, E.R. and Lyttle, C.R. (1978) Cancer Res. 38, 4086-5090 15 McKearn, T.J. (1980) in Monoclonal Antibodies (Kennett,

406

16

17

18

19

20 21 22 23 24

R.H., McKearn, T.J. and Bechtol, K.B., eds.), pp. 368-369, Plenum, New York Kennett, R.H. (1980) in Monoclonal Antibodies (Kennett, R.H., McKearn, T.J. and Bechtol, K.B., eds.), pp. 375, Plenum, New York McKearn, T.J. (1980) in Monoclonal Antibodies (Kennett, R.H., McKearn, T.J. and Bechtol, K.B., eds.), pp. 388-390, Plenum, New York Kennett, R.H. (1980) in Monoclonal Antibodies (Kennett, R.H., McKearn, T.J. and Bechtol, K.B., eds.), pp. 375, Plenum, New York Jonak, Z.L. (1980) in Monoclonal Antibodies (Kennett, R.H., McKearn, T.J. and Bechtol, K.B., eds.), pp. 405-406, Plenum Press, New York Laemmli, U.K. (1970) Nature (Lond.) 227, 680-685 Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354 Burnette, W,N. (1981) Anal. Biochem. 112, 195-203 Buhler, J.M., Huet, J., Davies, K.E., Sentenac, A. and Fromageot, P. (1980) J. Biol. Chem. 225, 9949-9954 McNabb, T. and Jellinck, P.H. (1975) Biochem. J. 151, 275-279

25 Wagai, N. and Hosoya, T. (1982) J. Biochem. (Tokyo) 91, 1931 1942 26 Carlstrom, A. and Vesterberg, O. (1967) Acta Chem. Scand. 21, 271-278 27 Rombauts, W.A., Schroeder, W.A. and Morrison, M. (1967) Biochemistry 6, 2965-2977 28 Christmann, J.L. and Dahmus. M.E. (1981) J. Biol. Chem. 256, 11798-11803 29 Davis, M.M., Kim, S.K. and Hood, L. (1980) Cell 22, 1 2 30 Mormon, M., Allen, P.Z., Bright, J. and Jaysinghe, W. (1965) Arch. Biochem. Biophys. 111, 126 133 31 Rytomaa, T. (1960) Acta Pathol. Microbiol. Scand. 50, (Suppl. 140), 1 118 32 Desser. R.K., Himmelhoch, R., Evans, W.H., Januska. M., Mage, M. and Shelton, E. (1972) Arch. Biochem. Biophys. 148, 452-464 33 Smith, D.C. and Klebanoff. S.J. (1970) Biol. Repro& 3, 229-235 34 Hosoya, T., Sasaki, K. and Wagai, N. (1981) J. Biochem. (Tokyo) 89, 1453-1463 35 Anderson, W.A., Kang, Y.H. and DeSombre, E.R. (1975) J. Cell Biol. 64, 668-681