Immunoreactivity of rhodopsin and opsin

Biochimica et BiophysicaActa, 802 (1984) 83-89

83

Elsevier

BBA 21886

I M M U N O R E A C T I V I T Y OF R H O D O P S I N A N D OPSIN JAMES J. PLANTNER a and E D W A R D L. KEAN a,b,*

a Lorand V. Johnson Laboratory for Research in Ophthalmology, Department of Surgery, Division of Ophthalmology, and b Department of Biochemistry, School of Medicine, Case Western Reserve University Cleveland, OH 44106 (U.S.A.) (Received March 26th, 1984)

Key words: lmmunoreactivity; Rhodopsin; Opsin

An examination by a radioimmunoassay of the relative affinity of opsin and rhodopsin for rabbit antibody raised against bovine rhodopsin revealed that opsin was the preferred antigen. About 10-fold greater amounts of rhodopsin than opsin were required to achieve 50% inhibition of binding of 12sI-labeled ligand in the RIA. Opsin was more reactive when examined in the light or dark, compared to rhodopsin incubated in the dark. Mixtures of opsin and rhodopsin (prepared by partial bleaching of rhodopsin or synthetic mixtures) exhibited increased reactivity with increasing mole fraction of opsin. This response was nonlinear, with small increases in opsin producing relatively large increases in reactivity. A partial fractionation of the antibody into two groups showing differential reactivities toward opsin and rhodopsin was achieved by affinity chromatography on opsin-Sepharose. However, with both groups, opsin was still the preferred antigen. Scatchard analysis of ~251-1abeled rhodopsin and opsin produced nonlinear plots, indicating the presence of multiple species of antibody. The affinities and binding capacities were similar for both labeled antigens. In competitive binding studies, the antibody showed a strong preference for either labeled ligand (rhodopsin or opsin) as compared to the unlabeled material. These latter observations indicate that altering rhodopsin either by bleaching or iodination produced changes in the relative immunoreactivity of the molecule.

Introduction

We have recently described a radioimmunoassay (RIA) for the visual pigment, rhodopsin, which is both rapid and sensitive [1,2]. Since all of analyses were routinely performed in the light, in reality, opsin was being measured. When native rhodopsin was assayed by performing all of the procedures in the dark, it was found that, while both opsin and rhodopsin were reactive in the assay and thus recognized by the antibody, opsin was the preferred ligand. The capacity of the antibody to distinguish between opsin and rhodopsin was investigated and is the topic of the present report. The relative immunogenicity of rhodopsin * To whom correspondence should be addressed. 0304-4165/84/$03.00 © 1984 Elsevier Science Publishers B.V.

and opsin to rabbit antibodies raised against both opsin and rhodopsin was investigated, as well as the response of these antigens to a sheep antibody obtained from another laboratory. About a 10-fold greater affinity of the antibody toward opsin as compared to rhodopsin was detected. Similar responses were obtained using the different antibody preparations. In addition, it was revealed that the antibody preparations could distinguish between radioiodinated and unlabeled antigens. Methods and Materials

Rhodopsin, All procedures involved in the preparation of rhodopsin, both labeled and unlabeled, were performed under dim red light (Wratten No. 1 filter). Rod outer segments were purified from

84 frozen, dark adapted bovine retinas (Hormel Co. Austin, MN) as described previously [3]. For use as a source of unlabeled antigen, rod outer segments were incubated overnight at 4 ° C with a 10-fold molar excess of ll-cis retinal (a gift from Hoffman-LaRoche Co., Nutley, N J) to regenerate rhodopsin from endogenous opsin. The rod outer segments, enriched in this manner (native plus regenerated), were washed by resuspension in 0.01 M Tris-acetate (pH 7.4) centrifuged at 40 000 × g for 20 min, and then extracted with 2% Emulphogene BC 720 (obtained as a gift from G A F Corp., New York) in 0.05 M Tris-HC1 (pH 7.0) (solution A). Dilutions were made with 0.05 M Tris-HC1 (pH 7.0)/0.3% Emulphogene BC 720. For the preparation of 125I-labeled rhodopsin, rod outer segments were extracted with solution A and native rhodopsin purified by calcium phosphate-Celite adsoption chromatography and Con A-Sepharose (Pharmacia, Uppsala) affinity chromatography (A278nm/h498 nm= 1.8-1.9) as described previously [3]. Rhodopsin was iodinated by the lactoperoxidase method as described previously [2], and purified by multiple gel filtrations [4]. Opsin was prepared by exposure of these preparations of rhodopsin to room light at 4°C. Rhodopsin concentrations were determined spectrally from the change in absorbance at 498 nm before and after bleaching in the presence of 0.1 M hydroxylamine using a molar extinction coefficient of 40 600 [51. Antisera. Three types of antibody preparation were examined: (1) rabbit antibody to rhodopsin; (2) rabbit antibody to opsin; (3) sheep antibody to opsin. The rabbit antisera were prepared as described previously [2], using Freund's complete adjuvant (Difco, Detroit, MI). The sheep preparation was obtained from Drs. Mark Zorn and David Papermaster (prepared as described by Papermaster et al. [6]). The anti-rhodopsin immunizations were performed under dim red light, while the anti-opsin immunizations were performed under normal room light. The rabbits were maintained under normal room lighting conditions (12 h light, 12 h dark) for periods from 4 - 9 months after inoculation. Antibody titrations, RIA. Antibody titrations were performed as described previously [1], using the soluble antibody procedure to determine im-

munological activity. Amounts of antibody from different preparations were appropriately diluted, such that the same amount of 125I-labeled opsin would be precipitated in each case. Plots of the RIA standard curves are computer fits of the data by a least-squares approximation. Addition of all rhodopsin solutions, incubations and washings were performed either under dim red light ('dark') or standard room lighting ('light'), as indicated.

Determination of affinity constant," binding capacity. Antibody binding studies were performed in which the amount of antiserum was held constant (0.05 ml of a 1 : 600 dilution), while the amount of 125I-labeled ligand (opsin or rhodopsin) was varied over a 600-fold range in concentration (from antibody excess to nearly saturating antigen excess). The data were analyzed by a Scatchard plot [7]. A computer fit of the data by use of the 'Ligand' program [8] (single ligand - two class binding curve) provided limiting values for high and low affinity antibody populations. Two methods were used to correct for the presence of nonspecific binding: (1) the amount of 1251 precipitated in matching tubes containing no immune serum was subtracted to provide the 'net' binding; (2) the 'ligand' program used to fit the data also provides for a statistically determined correction for nonspecific binding. Opsin-affinity column. Purified bovine opsin was coupled to aminohexyl-Sepharose 4B (Sigma, St. Louis, MO) using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide-HC1 as described by Papermaster et al. [9] All of the following procedures were performed in the light. Rabbit anti-rhodopsin in buffer B (0.15 M NaC1/5 mM phosphate (pH 7.4)/0.01% sodium azide) was incubated with the opsin-Sepharose for 2 h at room temperature. The suspension was then packed into a column and washed at 4 ° C with buffer B. The column ehiate was monitored for 278 nm absorbance and immunological activity, and the wash was continued until baseline values were reached. The column was then eluted with 3 M KSCN and monitored by 278 nm absorbance. Fractions of the KSCN eluate were pooled on the basis of A278, the pools dialysed against two changes of 20-fold volumes each of buffer B and assayed for immunological activity. Immunologically active material eluted with either buffer B or KSCN were pooled sep-

85

arately and concentrated using an (Lexington, MA) PM-10 membrane.

Amicon

Results

Use of different anti-opsin and anti-rhodopsin sera

Reactivity of rhodopsin and opsin RIA standard curves were prepared by incubating rhodopsin or opsin with rabbit anti-rhodopsin either in the dark or light, respectively, as described in Methods and Materials. As can be seen in Fig. 1, although both rhodopsin and opsin reacted with the antibody, greater reactivity was obtained with opsin as compared to rhodopsin. Thus, about 10-fold greater amounts of rhodopsin than opsin were required to achieve 50% inhibition of binding of the respective t25I-labeled antigen. When examined in seven different preparations of rod outer segments, the range varied from 8- to 20-fold, probably due to variation in the extent of opsin regeneration in the rod outer segments. As is also apparent in Fig. 1, the lines are not parallel, indicating that, although both opsin and rhodop-

I°° 1 60

sin were reactive with the antibody, they are not antigenically identical [7].

Antibody titrations were performed in the light with three anti-sera: rabbit anti-rhodopsin; rabbit anti-opsin, and sheep anti-opsin, to determine the relative amounts of antibody present. RIA standard curves were prepared in the light and dark as described in Methods and Materials with each of the antisera, using these conditions. As seen in Table I, while the absolute amount of unlabeled rhodopsin or opsin necessary to achieve 50% inhibition of binding of the respective labeled antigen to the antibody varied somewhat in each case, the relative amount of unlabeled rhodopsin necessary was 5- to 10-fold greater as compared to opsin with all three antisera. It is of interest that similar responses were obtained using antibody raised by the injection of purified rhodopsin under dim red light conditions as when the injections were carried out in the light, i.e., using opsin as the antigen.

Reactivity of mixtures of rhodopsin and opsin

"--.."lP,,• ! LIGHT NN~ DARK""--.• "-..... ~ ~

4O

"'O"'""'"B,.

Since small amounts of opsin have a large influence on enhancing reactivity, it was essential that the rhodopsin preparation used be as completely native as possible in order for the difference in light vs. dark reactivity to be detectable.

2O

0

TABLE 1

0.1

0.2 '

'

'0'.5"'"

t.o

2.0 '

'

p MOLES

'

5 i0.... io.o ~ 200',

Fig. 1. RIA of rhodopsin and opsin. Standard curves were prepared by adding increasing a m o u n t s of either unlabeled rhodopsin or unlabeled opsin (as indicated on the abscissa) to a constant volume of diluted rabbit anti-rhodopsin in the dark. The unlabeled rhodopsin was prepared by incubating purified rod outer segments with l l - c i s retinal followed by washing and detergent extraction. All subsequent steps were performed either under dim red light (dark, rhodopsin (O . . . . . . O)) or under room illumination (light, opsin (O O)). After a 1 h incubation, excess ~251-1abeled rhodopsin or opsin was added to each tube, respectively. Following an additional 1 h incubation, S. aureus cells were added, and the tubes were incubated for a final 1 h. The radioactivity in the pellet was measured following washing and centrifugation. The plots are computer fits of the data by a least-squares approximation.

REACTIVITY OF THE UNLABELED ANTIGENS R H O D O P S I N A N D OPSIN W I T H V A R I O U S A N T I S E R A Ratios represent the molar ratios of the a m o u n t of unlabeled r h o d o p s i n / o p s i n necessary to achieve 50% inhibition of binding of 125I-labeled antigen. Opsin and rhodopsin values represent the a m o u n t s of unlabeled antigen necessary to achieve 50% inhibition of binding of 125I-labeled rhodopsin in the RIA. Incubations were performed in the light for opsin and the dark for rhodopsin. Antiserum

Opsin (pmol)

Rhodopsin (pmol)

Ratio

Rabbit anti-rhodopsin Rabbit anti-opsin Sheep anti-opsin

1.0 1.5 3.1

9.5 a 15 15

9.5 10 4.8

a Analysis seen in Fig. 1.

86

This was accomplished by incubating the purified rod outer segments from dark adapted Hormel retinas with l l-cis retinal prior to extraction (enriched rod outer segments). This incubation routinely increased the yield of A49s by 20-30%, indicating the presence of considerable amounts of opsin in the Hormel retinas. A portion of the detergent extract of enriched rod outer segments was bleached, and synthetic mixtures of rhodopsin and opsin were prepared having the same concentration of total ligand (rhodopsin + opsin), but varying in the mole fractions of opsin (or rhodopsin). RIA standard curves were then prepared by incubating with rabbit anti-rhodopsin in the' dark. As the mole fracton of opsin increased, so did the reactivity of the unlabeled antigen with the antibody (Table II), i.e., its ability to inhibit the binding of t25I-labeled rhodopsin to the antibody. This is reflected also by a decreased amount of antigen necessary to achieve 50% inhibition of binding with increasing mole fraction of opsin, as shown in Table II, as a 'reactivity factor'. The increase in reactivity, however, was not a linear function of the mole fraction of opsin. Small increases in opsin content produced large increases in reactivity. Similar results

T A B L E II REACTIVITY O F O P S I N + R H O D O P S 1 N T I O N O F T H E OPSIN C O N T E N T

AS A F U N C -

(A) Solutions containing the same concentration of total unlabeled antigen (opsin + rhodopsin) were prepared by mixing opsin and rhodopsin in the proportions listed. RIA performed in the dark. (B) A m o u n t of 125I-labeled rhodopsin bound at a constant amount of total unlabeled antigen (opsin + rhodopsin, 1.5 pmol). 100% bound control (i.e., no unlabeled antigen) contained 13 500 cpm. (C) Calculated by taking the reciprocal ratio of the amount of total unlabeled antigen ( o p s i n + rhodopsin) necessary to achieve 50% inhibition of binding of 125I-labeled rhodopsin as compared to the value obtained with 100% opsin (0.76 pmol). (A) Mole fraction of opsin

(B) 1251-rhodopsin b o u n d (cpm)

(C) Reactivity factor

0 0.2 0.4 0.6 0.8 1.0

10000 7 300 6 000 5 400 4900 4 500

0.12 0.42 0.63 0.76 0.88 1.00

were obtained when a solution of rhodopsin was partially bleached to varying extents and analyzed as above (data not shown).

Affinity chromatography of rabbit anti-rhodopsin Rabbit anti-rhodopsin was fractionated on opsin-Sepharose as described in Methods and Materials. Total recovery of antibody activity, based on that applied, however, was only 6 and 21% for the buffer B (unbound) and KSCN (bound) fractions, respectively. Subsequent elution of the column with 0.02 M acetic acid did not release additional antibody activity. The low recoveries may have been due to nonreversible binding of the antibody to the column, or reflected instability or losses from the subsequent steps, which involved dilution, dialysis and concentration. The following observations indicated that the latter procedures may have contributed greatly to the low recoveries. Control experiments were performed by diluting whole antiserum with either buffer B or 3 M KSCN to the extent experienced during chromatography, followed by conditions of storage at 4°C, dialysis and concentration that were applied to the chromatographed samples. The recovery of antibody activity after these treatments was 19 and 33% for the buffer B and 3 M KSCN controls, respectively. Antibody titrations were performed in the light to determine the relative amount of antibody in each pool. The buffer B eluate, KSCN eluate and the unfractionated rabbit anti-rhodopsin were diluted such that the same amount of 125I-labeled opsin would be precipitated in each case. RIA standard curves were then prepared for each sample of antibody in the light and dark. As can be seen in Table III, reactivity was greater with opsin than with rhodopsin for both fractions from the column as evaluated by the amount of unlabeled antigen necessary to achieve 50% inhibition of binding of the respective 125I-labeled antigen. When compared to the reactivity of the unfractionated material, i.e., rabbit anti-rhodopsin, the buffer B fraction (antibody not bound to opsinSepharose) was more reactive toward rhodopsin and less toward opsin. The KSCN fraction (antibody bound to opsin-Sepharose), however, showed less reactivity toward rhodopsin, and about the same reactivity toward opsin. Thus, the opsin-

87 T A B L E III REACTIVITY T O W A R D T H E U N L A B E L E D A N T I G E N S OPSIN A N D R H O D O P S I N O F S E R U M A F T E R F R A C T I O N A T I O N ON OPSIN-SEPHAROSE

24

20

Ratios represent the molar ratios of amount of unlabeled r h o d o p s i n / o p s i n necessary to achieve 50% inhibition of binding of 12SI-labeled antigen. Opsin and rhodopsin values represent the a m o u n t s of unlabeled antigen necessary to achieve 50% inhibition of binding of 1251-labeled rhodopsin in the R1A. Incubations were performed in the light for opsin, and in the dark for rhodopsin.

•

1

I

I

8

O;

IS

iI

oo Antiserum

Opsin (pmol)

Rhodopsin (pmol)

Ratio

Rabbit anti-rhodopsin a Buffer B-eluate b KSCN-eluate b

0.91 1.6 1.1

8.2 5.2 9.4

9.0 3.2 8.5

Initial material. b Pooled fractions of rabbit anti-rhodopsin serum eluted from opsin-Sepharose column with either buffer B or 3 M KSCN. a

Sepharose column was capable of partial fractionation of the antiserum on the basis of relative affinity for opsin. Nevertheless, the antibody which bound to the opsin affinity column was still capable of reacting with rhodopsin.

Scatchard analysis; affinity constants," binding capacity In order to investigate the affinities and binding capacities of the antibody, 125I-labeled opsin and rhodopsin were used as ligands. The capacity of these molecules to bind to a constant amount of rabbit antibody was measured over a wide range in concentration of 125I-labeled antigens (antib o d y / a n t i g e n = 12:0.02). This is in contrast to the experiments described above, in which the relative binding of unlabeled ligand was measured by the RIA. Scatchard plots of the binding data in the form of b o u n d / f r e e vs. bound were prepared (Fig. 2). Nonlinear plots were obtained with both opsin and rhodopsin, indicating a heterologous population of antibodies [7], as might be expected for whole serum. The curves were fit by computer analysis [8], using two classes of binding sites as a model. This simplification allowed an estimation of the parameters for the limiting classes of highand low-affinity sites. The values for the affinity constants, K t and K 2, and the respective binding

A

Oz

1'o

2!o

~io

410

50

60

Bound (pmoles)

Fig. 2. Scatchard analysis of 125I-labeled rhodopsin and opsin binding to rabbit anti-rhodopsin. Antibody binding curves were prepared in the dark or in the light. The a m o u n t of antiserum was held constant (0.05 ml of a 1:600 dilution) while the a m o u n t of 125I-labeled antigen (rhodopsin (A) or opsin (O)) was varied from 0.15 to 85 pmol.

capacities, n I and n 2, are provided in Table IV. The values which were obtained ( 1 0 7 - 1 0 9 M -1) indicated high affinity of the antibody for both labeled antigens [10]. The antibody concentration used was similar in magnitude to that of the high-affinity dissociation constant ( l / K 1 ) , and about 1% that of the low-affinity dissociation constant (1/K2). Unlike the results from the RIA analyses using the unlabeled antigens, the T A B L E IV S C A T C H A R D ANALYSIS O F 125I-LABELED R H O D O P SIN A N D O P S I N B I N D I N G T O R A B B I T A N T I RHODOPSIN A computer fit of the plots shown in Fig. 2 by a nonlinear approximation of a two-class binding curve provided the limiting values for high- and low-affinity antibody populations [8]. The data are presented as the mean_+S.D. 36 experimental points were collected for rhodopsin, and 47 for opsin. Parameter

Rhodopsin

Opsin

Affinity constant ( l / m o l ) K 1 (higher affinity) (1.0+0.1)-10 9 K 2 (lower affinity) (5.6 _ 1.8). 10 7

(6.6 + 0.6). 10 s (7.3 + 2.0)- 10 6

Binding capacity (mol ligand/l serum) n I (higher affinity) (1.5 + 0.2)-10- 5 n 2 (lower affinity) (3.5 + 0.4). 1 0 - 5

(1.64-0.01).10 -5 (8.2+1.3)-10 -5

88 Scatchard analysis revealed that the binding capacity and affinity constants for iodinated rhodopsin and opsin at the higher-affinity sites were similar to one another. At the lower-affinity sites, the affinity constant for rhodopsin was nearly 10-fold greater than that for opsin, while the number of binding sites (binding capacity) at the lower-affinity site for opsin was about twice that for rhodopsin (Table IV). Reactivity of iodinated vs. unlabeled antigen While the RIA analyses showed that unlabeled opsin was preferable to unlabeled rhodopsin (i.e., a greater amount of rhodopsin than opsin was needed to achieve 50% inhibition of binding, Fig. 1, Table I) competitive binding studies of the following type revealed that the antibody preferred a labeled antigen to an unlabeled one irrespective of the conformational state. A constant amount of each 125I-labeled antigen was mixed with increasing amounts of its corresponding unlabeled material, and the mixture was incubated with the antibody. If the antibody showed equal affinity for both labeled and unlabeled antigens, simple isotope dilution would have resulted, and a 50% inhibition in binding of the labeled antigen would have occurred when equal molar amounts were present. However, it was observed that in order to obtain 50% inhibition of binding of the labeled antigen, a 57-fold molar excess of unlabeled rhodopsin, and a 14-fold molar excess of unlabeled opsin were required over the amount of labeled material presented to the antibody. As a further indication of the high affinity of the antibody for the labeled material, reversibility of its binding was measured. The antibody was preincubated with labeled opsin, followed by incubations using up to a 500-fold excess of unlabeled opsin. The complex was then precipitated by the addition of Staphylococcus aureus cells, washed and counted. In no case was there a significant decrease in the amount of opsin which was bound (data not shown). Discussion

Previously [11], when analyzed by the Ouchterlony double-diffusion method, it was demonstrated on a qualitative basis that both rhodopsin

and opsin would react with rabbit anti-bovine rhodopsin antibody. The present report, utilizing the RIA [1,2], confirmed and quantitated these observations. When analyzed by the RIA, it was seen that the antibody was some 8- to 20-fold more reactive with opsin as compared to rhodopsin. Using synthetic mixtures of opsin and rhodopsin, the reactivity was shown to increase with increasing opsin content. Small increases in opsin produced large increases in reactivity. In addition to demonstrating selectivity by the antibody, these observations revealed limitations which must be considered in attempting to analyze visual pigments by the RIA, using an antibody preparation having these characteristics. First, the rhodopsin used to generate the standard curve must contain no opsin. This condition is difficult to ensure or even evaluate. Second, since any opsin present in the samples will also react in the dark, the inhibition of the binding of ~25I-labeled rhodopsin which was measured would be a summation of that due to rhodopsin plus opsin. Due to the greater reactivity of opsin, its contribution would result in an overestimation of the rhodopsin content if rhodopsin were used as the standard. Thus, in practical terms, the assay should be performed in the light, where only opsin would be present. The present studies have examined differences in response of antibodies made against bovine rhodopsin or opsin as antigens. Of the several antibody preparations against the visual pigments that have been reported, with one exception [12] all have shown either an exclusive or preferential reactivity toward opsin as compared to rhodopsin. This was observed with both heterologous preparations [13-15] and with a monoclonal preparation [16]. In the present studies, similar responses were obtained using antibodies that were prepared by injecting either rhodopsin or opsin into rabbits, as well as the antibody raised in sheep, obtained from another laboratory. The influence of detergents on the antigenicity of visual pigments, as described previously [12-16], was not examined in the present studies. Affinity chromatography of the rabbit antirhodopsin on opsin-Sepharose produced only partial fractination of antibody activity. It is not clear whether this was due to incomplete resolution of

89

independent species, the lack of detergent during chromatography, or in fact reflected the presence of multiple species having varying reactivities toward opsin and rhodopsin. The RIA and Scatchard analyses strongly support the latter. Thus, the nonparallel lines generated for rhodopsin and opsin in RIA analysis (Fig. 1) indicate that, while both antigens are reactive, their relative affinities for the antibody differ [7]. In addition, Scatchard analysis produced curved plots for both opsin and rhodopsin, indicating the presence of heterogeneous antibody populations reactive toward both antigens. Competitive binding studies using labeled and unlabeled ligands revealed that the antibody prepared against rhodopsin had a greater affinity for radioiodinated rhodopsin and opsin as compared to the unlabeled material. Within the population of labeled ligands (representing about 1% of the total), however, Scatchard analysis indicated that, at the high-affinity sites, the reactivities were essentially the same for iodinated rhodopsin and iodinated opsin. In sum, the heterologous antibody preparation raised against bovine rhodopsin or opsin showed the following characteristics: (1) demonstrated a strong preference toward opsin as compared to rhodopsin; (2) had a greater affinity toward labeled antigens as compared to the unlabeled material; (3) did not distinguish between iodinated rhodopsin and iodinated opsin as concerns the high affinity sites. Thus, although opsin and rhodopsin have the same primary amino acid sequence, differences in relative reactivities of the antibody population in whole serum were detected not only to the large conformational changes which accompany the process of bleaching [17], but also to relatively minor structural changes which affect primarily one amino acid, tyrosine, as a result of the ir~troduction of 125I.

Acknowledgements The authors express their appreciation to Drs. Mark Zorn and David Papermaster, Department of Cell Biology, Yale University for the generous

gift of sheep anti-opsin antiserum; to Drs. Nathaniel Woodruff and Kenneth Neet of the Department of Biochemistry and to Dr. Tomuo Hoshiko for the Department of Physiology of this institution of aid in performing the computer analysis of the Scatchard plots. We are grateful to Dr. F.J.M. Daemen of the University of Nijmegen for providing us with copies of the manuscripts from his laboratory prior to their publication [13-15]. This work was supported in part by Public Health Service Research Grants EY 00393 and EY 03685 from the National Eye Institute and by the Ohio Lions Eye Research Foundation.

References 1 Plantner, J.J., Hara, S. and Kean, E.L. (1982) Exp. Eye Res. 35, 543-548 2 Lentrichia, B.B., Plantner, J.J. and Kean, E.L. (1980) Exp. Eye Res. 31, 1-8 3 Plantner, J.J. and Kean, E.L. (1976) J. Biol. Chem. 251, 1548-1552 4 Hara, S., Plantner, J.J. and Kean, E.L. (1983) Exp. Eye Res. 36 441-445 5 Wald, G. and Brown, P.K. (1953-54) J. Gen. Physiol. 37, 189 - 200 6 Papermaster, D.S., Schneider, B.G., Zorn, M.A. and Kraehenbuhl, J.P. (1978) J. Cell Biol. 77, 196-200 7 Chard, T. (1978) An Introduction to Radioimmunoassay and Related Techniques, pp. 332-328 and 394-395, North-Holland, Amsterdam 8 Munson, P.J. and Rodbard, D. (1980) Anal. Biochem. 107, 220-239 9 Papermaster, D.S., Converse, C.A. and Zorn, M. (1976) Exp. Eye Res. 23, 105-15 10 Werblin, T.P. and Siskind, G.W. (1972) Immunochem. 9, 987-1011 11 Plantner, J.J. and Kean, E.L. (1983) Biophys. Biochim. Acta, 744, 312-319 12 Peterson, R.J., Hargrave, P.A. McDowell, J.H., and Jackson, R.W. (1983) Vis. Res. 23, 267-273 13 Margry, R.J.C.F., Jacobs, C.W.M., DeGrip, W.J. and Daemen, F.J.M. (1982) Vision Res. 22, 1447-1449 14 Margry, R.J.C.F., Jacobs, C.W.M., Bonting, S.L., DeGrip, C.W.M. and Daemen, F.J.M. (1983) Biochim. Biophys. Acta 742, 463-470 15 Schalken, J.J., Margry, R.J.C.F., DeGrip, W.J. and Daemen, F.J.M. (1983) Biochim. Biophys. Acta 742, 471-476 16 Molday, R.S. and Mackenzie, D. (1983) Biochemistry 22, 653-660 17 Rafferty, C.N., Cassim, J.Y. and McConnell, D.G. (1977) Biophys. Str. Mechanism 2, 277-320