ABSTRACT Light isomerizes the chromophore of rhodopsin, ll-cis retinal (formerly retinene), to the all-trans configuration. This introduces a succession of unstable intermediates--pre-lumirhodopsin, lumirhodopsin, metarhodopsin--in which all-trans retinal is still attached to the chromophoric site on opsin. Finally, retinal is hydrolyzed from opsin. The present experiments show that metarhodopsin exists in two tautomeric forms, metarhodopsins I and II, with),m ~ 478 and 380 m#. Metarhodopsin I appears first, then enters into equilibrium with metarhodopsin II. In this equilibrium, the proportion of metarhodopsin II is favored by higher temperature or pH, neutral salts, and glycerol The change from metarhodopsin I to II involves the binding of a proton by a group with pK 6.4 (imidazole?), and a large increase of entropy. Metarhodopsin II has been confused earlier with the final mixture of all-trans retinal and opsin (Xm, ~ 387 m/z), which it resembles in spectrum. These two products are, however, readily distinguished experimentally. On exposure to light, rhodopsin bleaches over transient intermediates to an eventual mixture of retinaP and opsin. If one begins with cattle rhodopsin, for example, at the temperature of liquid nitrogen (-195°C), the product of

We have recently described the synthesis of rhodopsin in a solution containing four components: vitamin A, the precursor of the rhodopsin chromophore; opsin, the protein of rhodopsin; and liver alcohol dehydrogenase and cozymase, the enzyme and coenzyme which oxidize vitamin A to retinene (Hubbard and Wald, 1951). 1 This experiment was first performed with a fish liver 0il concentrate as the source of vitamin A. When later we attempted to repeat it using crystalline vitamin A, almost no rhodopsin was formed. Such differences in behavior in two samples of what is ordinarily thought of as the same substance could have only one explanation. The vitamin A molecule exists in several different shapes, as different geometrical or cis-trans isomers. Liver oils are known to contain a mixture of such isomers, while ordinary crystalline vitamin A is a single isomer. Carotenoids in general are converted from any single cis-trans configuration to an equilibrium mixture of stereoisomers by irradiation with light in the presence of a trace of iodine. When crystalline vitamin A was treated in this way, it became as effective a precursor of rhodopsin as the liver oil concentrate. These experiments made it plain that the geometrical configuration of vitamin A is a dominant factor in the rhodopsin system. The present paper is concerned with the analysis of this relation.

Rhodopsin, the photosensitive pigment of rod vision, is a chromoprotein composed of the colorless protein, opsin, combined with the yellow carotenoid, retinene (C19H~CHO). The chemistry of rhodopsin has been explored for many years, but its molecular weight is still unknown. This is primarily due to the

A B STRACT Photoreceptor potentials were recorded extracellularly from the aspartate-treated, isolated retina of the skate (Raja oscellata and R. erinacea), and the effects of externally applied retinal were studied both electrophysiologically and spectrophotometrically. In the absence of applied retinal, strong light adaptation leads to an irreversible depletion of rhodopsin and a sustained elevation of receptor threshold. For example, after the bleaching of 60 % of the rhodopsin initially present in dark-adapted receptors, the threshold of the receptor response stabilizes at a level about 3 log units above the dark-adapted value. The application of l l-cis retinal to strongly light-adapted photoreceptors induces both a rapid, substantial lowering of receptor threshold and a shift of the entire intensityresponse curve toward greater sensitivity. Exogenous 11-cis retinal also promotes the formation of rhodopsin in bleached photoreceptors with a time-course similar to that of the sensitization measured electrophysiologically. All4rans and 13-cis retinal, when applied to strongly light-adapted receptors, fail to promote either an increase in receptor sensitivity or the formation of significant amounts of lightsensitive pigment within the receptors. However, 9-c/s retinal induces a substantial increase in receptor sensitivity and promotes the formation of isorhodopsin. These findings provide strong evidence that the regeneration of visual pigment in the photoreceptors directly regulates the process of photochemical dark adaptation.

When rhodopsin is bleached by light, its prosthetic group, retinene, is broken away from its original sites of attachment on the protein, opsin. This process can be expected to uncover new groups on opsin, at the least those groups which had been involved in holding retinene, and perhaps others, associated with some general loosening of structure of the protein. Indeed it has been recognized that the bleaching of rhodopsin possesses some of the characteristics of a protein denaturation (Wald, 1935-36; Mirsky, 1936; Wald and Brown, 1951-52). Only two changes in the protein have heretofore been measured. Broda and Victor (1940) reported that on bleaching, the isoelectric point of frog rhodopsin shifts from 4.47 to 4.57. Wald and Brown (1951-52) found that the bleaching of rhodopsin (frog, cattle, squid) exposes 2 to 3 sulfhydryl groups for each retinene molecule liberated. This last observation implied that at sufficiently alkaline pH to ionize sulfhydryl groups, the bleaching of rhodopsin should liberate hydrogen ions. This

In the course of a study of acid-base changes on bleaching rhodopsin, we found that at pH more acid than 5.5 or more alkaline than 7 the exposure of rhodopsin to light is followed by the denaturation of its protein moiety, opsin (Radding and Wald, 1955-56). In the present paper we examine the stability of rhodopsin and opsin as functions of age and pH. We have used two criteria of stability: maintenance of the absorption spectrum of rhodopsin or of its maximal extinction at 500 m/z; and retention of its ability to regenerate after exposure to light. Aging of Rhodopsin Rhodopsin solutions can be kept at 3°C. for as long as 6 months with no appreciable change in absorption spectrum in the visible region (the a-band)? The capacity to regenerate after exposure to fight, however, declines throughout this period. To measure the regenerability, an aliquot of cattle rhodopsin solution is diluted with 2 to 3 volumes of phosphate buffer, pH 6.4. It is exposed to the

The effects have been examined of chymotrypsin, pepsin, trypsin, and pancreatic lipase on cattle rhodopsin in digitonin solution. The digestion of rhodopsin by chymotrypsin was measured by the hydrolysis of peptide bonds (formol titration), changes in pH, and bleaching. The digestion proceeds in two stages: an initial rapid hydrolysis which exposes about 30 amino groups per molecule, without bleaching; superimposed on a slower hydrolysis which exposes about 50 additional amino groups, with proportionate bleaching. The chymotryptic action begins at pH about 6.0 and increases logarithmically in rate to pH 9.2. Trypsin and pepsin also bleach rhodopsin in solution. A preparation of pancreatic lipase bleached it slightly, but no more than could be explained by contamination with proteases. In digitonin solution each rhodopsin molecule is associated in a micelle with about 200 molecules of digitonin; yet the latter do not appear to hinder enzyme action. It is suggested that the digitonin sheath is sufficiently fluid to be penetrated on collision with an enzyme molecule; and that once together the enzyme and substrate are held together by intermolecular attractive forces, and by the "cage effect " of bombardment by surrounding solvent molecules. The two stages of chymotryptic digestion of rhodopsin may correspond to an initial rapid fragmentation, such as has been observed with many proteinases and substrates; superimposed upon a slower digestion of the fragments. Since the first phase involves no bleaching, this may mean that rhodopsin can be broken into considerably smaller fragments without loss of optical properties. Willibald Kiihne, Professor of Physiology at Heidelberg, was at once the founder of retinal biochemistry and of enzymatic histochemistry. We owe to him the term enzymz and the first characterization of trypsin. All these interests came together in Kiihne's application of tryptic digestion to the analysis of retinal tissues (Kithne, 1878; cf. also Chittenden, 1882), in the course of which he observed that retinas, outer segments of rods, or fragmented rods can be digested with pancreatic trypsin without destroying rhodopsin.

Of the two groups of light receptors found in the vertebrate retina--the rods and cones--there is every reason to be more interested functionally in the cones. These are the organs of daylight vision. We depend upon them principally for the appreciation of form, and perhaps entirely for the appreciation of color. Nevertheless, until very recently the rods provided almost all that is known of visual substances and processes. The reason is that rods contain much more visual pigment than cones. Rods when dark-adapted are brightly colored, either red, owing to rhodopsin; or purple, owing to porphyropsin. Cones, however, look colorless under all circumstances (cf. KRhne, 1879). Kilhne concluded that the cones lack light-sensitive pigments; but that is impossible. Without photosensitive pigments there can be no vision. For light to act in any fashion, it must be absorbed; and for visible light, this demands a pigment. If one fails to see such a pigment in cones, this can mean only that it is too dilute to be apparent. Precisely because the cones contain so little light-sensitive pigment, intense light is required to stimulate them. It is this that makes them the organs of daylight vision. Hence the first attempt to extract a photosensitive pigment from cones was made with the chicken retina, which contains a few rods among a large predominance of cones (Wald, 1937 b). From it an impure mixture of rhodopsin and the cone pigment was extracted. The cone pigment was identified by its special sensitivity to deep red light, which scarcely affects rhodopsin. An extract of dark-adapted chicken retinas exposed to red light of wave lengths longer than 650 m/z bleaches slightly, the extinction falling maximally at 560 to 575 m/~, depending upon the pH. This is the cone pigment. Judging from its spectral properties, it is violet in color. It was therefore called iodopsin (Greek ion = violet). After the bleaching in red light is completed, the residue, exposed to white light, bleaches further, the extinction now falling maximally at 505 to 510 m/z; this is rhodopsin. These observations were confirmed in detail

Rhodopsin is synthesized by the combination of opsin with s cis isomer of retinene, called neo-b; and bleaches to a mixture of opsin and allotrans retinene (Hubbard and Wald, 1952-53). The latter must be re-isomerized to neo-b before it can contribute again to rhodopsin synthesis. For vision to go on, therefore, all-trans retinene--or the all°trans vitamin A with which it is in equilibrium--must be continuously isomerized to neo-b (~z. Fig. 1). This is the process with which the present paper is concerned. When a light-adapted animal is placed in the dark, the rhodopsin concentration rises immediately in an essentially linear fashion, and levels off gradually, reaching the darkoadapted level in about 3 hours in the frog (Zewi, 1939) and in about 1 hour in man and the rabbit (Rushton e ~ a/., 1955). A supply of neo-b retinene is therefore available. What is its source? There is as yet no indication that neo-b vitamin A occurs outside the eye. 1 Large stores of it have, however, been found in the eyes of the lobster (Wald and Burg, 1955), and several workers in this laboratory have identified neo-b vitamin A in retinas and pigment layers of cattle and frogs. ~ It appears likely that the eye itself possesses a mechanism for producing neo-b from all-trans retinene or vitamin A. There h'ave been indications that all-trans retinene or vitamin A can, at

the typical stereospecificity of the vertebrate visual pigments. This is true for the pigment in the chloride-depleted, "blue-shifted " state as well as for the normal pigment with added chloride. While in the chloride-deficient state, pigment regeneration occurred with both l l-cis- and 9-cis-retinals and the regenerated photopigments were also in the blue-shifted, chloride-depleted state. As with the native pigment, these regenerated pigments were bathochromically shifted to their normal positions by the addition of chloride. Chloride-deficient opsin by itself also responded to chloride for the pigment regenerated with l l-cis-retinal from such chloride-treated opsin was in the normal 521-position. Regeneration was always rapid, reaching completion in <5 min, and was significantly faster than for cow rhodopsin regenerating under the same conditions. This rapid rate was found with or without chloride, with both ll-cis- and 9-cis-retinals and in the presence of the sulfhydryl poison, p-hydroxymercuribenzoate (PMB). Like the native chloridedeficient pigment, the regenerated chloride-depleted photopigments responded to PMB by a blue shift beyond the position of the chloride-deficient state. The addition of chloride to these "poisoned " regenerated pigments caused a bathochromic shift of such magnitude as to indicate a repair of both the PMB and chloride-deficient blue shift. In this discussion the possible implications of these results to phylogenetic considerations are considered as well as to some molecular properties of the 521-pigment.