5
I. Photochem. Photobiol. B: Biol., 13 (1992) 5-17
Invited Review (New Trends in Photobiology)
Retinal Isidoro
photoisomerase:
role in invertebrate
visual cells
M. Pepe and Carlo Cugnoli
Istituto di Cibemetica e Biojisica de1 C.N.R via Dodecaneso 33, Genova 16146 (Italy)
Abstract In invertebrate visual cells, the rhodopsin content is maintained at a high level by the fast process of photoregeneration during daylight. Rhodopsin is converted by photoabsorption to metarhodopsin, which is reconverted to rhodopsin by light. In addition, rhodopsin is regenerated by a slow process of renewal which takes days to complete and involves the biosynthesis of opsin. It is well known that rhodopsin can be formed from opsin only when ll-cis-retinal is present; this requires the existence of an isomerizing enzyme which is capable of transforming all-tram-retinal, released from the degradation of metarhodopsin, into the ll-cis-retinal isomer. In some invertebrate visual systems, experiments on rhodopsin regeneration have been interpreted by assuming that the isomerization reaction is a light-dependent process involving a retinal-protein complex. Two retinal photoisomerases which have been well characterized, i.e. bee photoisomerase and cephalopod retinochrome, are reviewed here. Their properties are compared in order to determine their physiological role, which is likely to be in the renewal of visual pigment rhodopsin. To conclude, a visual pigment cycle is proposed in which rhodopsin regeneration follows two light-dependent pathways. This greatly simplifies the rhodopsin regeneration scheme for invertebrate visual systems.
Keywords: Retinal photoisomerase, cycle.
retinochrome,
rhodopsin renewal, invertebrate
visual
1. Visual pigment regeneration In photoreceptor cells of both vertebrates and invertebrates, the visual pigment rhodopsin is made up of the apoprotein opsin which is bound to the chromophoric group retinal (vitamin A aldehyde) via a Schiff-base linkage. On absorption of a photon, the rhodopsin chromophore is isomerized from 11-c&retinal to all-puns-retinal. This is the first step of the visual process which triggers a dark reaction sequence that eventually leads to the initiation of the electrical signal on the plasma membrane of the photoreceptor.
loll-1344/92/$5.00
0 1992 - Elsevier Sequoia. AU rights reserved
6
In order to restore the rhodopsin content, a process of regeneration takes place. In vertebrates, the all-tram-retinal released by the hydrolysis of rhodopsin to opsin is isomerized back to 11-c&retinal within a few minutes by a dark regeneration process [l-5], the mechanism of which is still poorly understood. Retinol isomerase activity has recently been detected in the pigment epithelium of frogs and rats in which allfruns-retinol is converted to 11-&s-retinol in the dark [6]. However, the enzyme responsible for this activity has not yet been isolated and characterized. Furthermore, it should act in unison with an alcohol dehydrogenase in order to produce 11-&retinal for rhodopsin regeneration. Another process which restores the visual pigment content is the biosynthesis of rhodopsin. A newly synthesized molecule of opsin binds with ll-cb-retinal forming rhodopsin which is incorporated into the newly synthesized disc membrane [7]. The biosynthesis of rhodopsin is part of a continuous renewal of photoreceptor membranes which takes days to complete. However, in the visual system of invertebrates, rhodopsin is not hydrolysed by photon absorption, but is converted to a stable metarhodopsin, which can be reconverted to rhodopsin by light [S, 91. This fast process, called photoregeneration, plays an essential role in maintaining the rhodopsin content, and thus the receptor sensitivity, at a high level [Xl-121. Visual pigments of invertebrates are also regenerated by a slow process of membrane renewal quite similar to that occurring in vertebrates [13-161. Studies on the visual system of the blowfly have been of fundamental importance in clarifying the possible pathways of rhodopsin regeneration in invertebrates. During continuous illumination, a photoequilibrium is reached between rhodopsin and metarhodopsin: the ratio between the concentrations of rhodopsin and metarhodopsin in the fly compound eye depends on the spectral composition of the environmental light and the respective extinction coefficients of the two pigments. The kinetics of the photoreactions are dependent on the light intensity [lo, 121. The alternation between the two states rhodopsin and metarhodopsin, which is caused by photoabsorption, is not of unlimited duration: in the ommatidia there is continuous renewal which implies degradation and resynthesis of molecules. Rhodopsin degrades with a mean lifetime of about 5 days, whereas metarhodopsin decays 60 times more rapidly (mean lifetime of about 20 h) [17]. When the animals are kept in alternating light and darkness, the rhodopsin content remains constant because of photoregeneration and biosynthesis processes. In contrast, when the animals are kept in the dark for days, the rhodopsin content decays exponentially, and this is also the case when all-puns-retinal is injected into their eyes. When the flies are subsequently exposed to light, the rhodopsin content rapidly returns to normal values. The same effect, with the same kinetics, is obtained by injecting 11-c&retinal into the eyes of flies kept in darkness. These experiments clearly demonstrate that retinal isomerization in the fly ommatidia is a light-dependent process. The maximum effect is reached when the animals are kept in light of the blue-violet range (about 440 nm). This may seem strange at first glance: blue-violet light is scarcely absorbed by free retinal, whereas UV light is very effective in the isomerization of retinal, forming a mixture of different isomers such as 13-c&, 9-c& and 11-&r-retinal. However, all-tram-retinal released in the cell by the degradation of metarhodopsin cannot remain free because of its high reactivity, so that possible binding with a protein may have the twofold effect of shifting its absorbance to the visible range and making highly stereospecific photoisomerization possible. Indeed, retinal can bind to an amino group of a protein residue via a protonated Schiff-base linkage, thereby shifting its absorbance maximum from 380 nm
7 to about 440 nm. Moreover, the protein structure around the binding site may favour the photoconversbn of all-tram-retinal into 11-&r-retinal, which is the only isomer that can combine with opsin to give rhodopsin. 2. Retinal photoisomerase
from honey-bee
In order to explain the results from the experiments on the blowfly, Schwemer [17] postulated the existence of a retinal-protein complex capable of photoisomerase activity. However, a pigment had been isolated in the honey-bee compound eye some years earlier, and this pigment seems to have the required characteristics. The story began in Geneva at the University Medical School when Pepe was collaborating with Baumann and Perrelet on a project which aimed to identify and isolate visual pigment rhodopsin in the honey-bee retina. After injecting tritiated retinol into the haemolymph of live drones (honey-bee male) and then waiting for 6 h, it was found that the radioactivity was bound to a water-soluble light-sensitive protein [18]. The chromophore, introduced as retinol, was converted into retinal and bound to the protein via a Schiff-base linkage. Since all rhodopsins studied at that time were insoluble in aqueous media, the authors hypothesized that this retinal-binding protein was a soluble cytoplasmic precursor of rhodopsin. A water-soluble pigment was extracted from honey-bee heads by Goldsmith in 1958 1191. On exposure to light, the pigment was bleached, leading to a maximum absorbance decrease at about 450 nm. Goldsmith [19] first suggested that this pigment could be the visual pigment of the drone because its absorbance maximum matched the sensitivity maximum of the most common photoreceptor in the drone retina. However, later on, the unusual property of water solubility and the fact that most of the animals used for the preparation of this extract were worker bees (dominant receptor at 535 nm) led to some doubt about the nature of this pigment [20]. In 1980, Pepe and Cugnoli [21] extracted and purified a sufficient quantity of water-soluble pigment from heads of honey-bee workers to permit initial characterization. It was found that the pigment was made up of a protein with a molecular weight of about 27 000, which was bound to a chromophore identified as a mixture of all-transretinal and 11-c&retinal. The ratio of all-puns-retinal to 11-&-retinal decreased significantly when the pigment was irradiated with light at wavelengths of greater than 400 nm, indicating that all-trans-retinal was converted by light to 11&s-retinal. The spectrophotometric absorbance change of the pigment following irradiation revealed a maximum at 440 run. Therefore, these findings could no longer be interpreted in favour of a soluble precursor of rhodopsin which, in contrast, should have had an absorbance maximum at 535 nm and a chromophore conversion from 11-&-retinal to all-trans-retinal. During the same year, Pepe joined Schwemer’s group at Ruhr University, Bochum in order to characterize further the properties of bee pigment and to verify the possibility that this pigment could have photoisomerase activity. 2.1. Protein purification and molecular propertks Retinal photoisomerase was extracted from homogenized honey-bee heads in Trisglycine buffer, incubated with tritiated retinol, isolated by preparative electrophoresis and further purified on a diethylaminoethyl (DEAR) coltmrn [21, 221. The molecular weight of the protein was determined to be 50 000 by gel filtration and 27 000 by sodium dodecylsulphate-polyacrylamide gel electrophoresis, suggesting that the protein consists of a dimer [21, 231.
The amino acid composition of the protein was also determined. Polar groups account for about 50% of the total, which explains its water solubibty. Aspartic and glutamic acid are also present in relatively high concentrations. The calculated isoelectric point gives a theoretical value of about 5.6. However, the experimental isoelectric point, determined by electrofocusing of the protein, is about 4.2, which suggests that some hydrophilic basic groups are inside the protein structure. 2.2. Spectrophotometric characteristics As shown by the absorbance spectra in Fig. 1, the protein from honey-bee can bind retinol as well as retinal. The absorbance maximum at 330 nm, shown by the protein sample incubated with retinol, is typical of free retinol in solution, indicating that the binding of the protein with retinol is presumably not covalent. The sample incubated with retina1 absorbs maximally at 440 nm, indicating the formation of a protonated SchitCbase linkage between retinal and protein. The binding site for retinol is presumably the same as that for retinal. This is suggested by the complete displacement of radioactive retinol bound to protein by the addition of an excess of retinal [23]. On changing the pH from 6.5 to 9.5 in the dark, the 440 run pigment is converted into an alkaline form which absorbs maximally at 365 nm, with a pK value of 8.4. This is the characteristic absorbance behaviour of protonated (440 nm) and unprotonated (365 nm) Schitf bases which form between retinal and amino groups. The stoichiometry of the binding between all-trans-retinal and protein was studied and the results indicate that one molecule of retinal can bind one molecule of protein (dimer) with a dissociation constant of K=2X 10m6 M [23].
-0.01
‘I
200
“““I
250
I”’
300
‘I
350
j
‘I
“““‘I
400
450
j ““““‘1 500
550
“I
600
”
650
Wavelength ( nm ) Fig. 1. Absorbance spectra of bee photoisomerase showing its ability to bind retinal as well as retinol. The protein was purified from a homogenate of bee heads incubated with all-tramretinal (a) or all-tranr-retinol (b) in 10 mM Tris-glycine buffer (pH 8.4) at room temperature (adapted from ref. 23).
9
The pigment is bleached on illumination. Figure 2 shows that there is a decrease in the absorbance at 440 nm accompanied by the formation of a photoproduct with an absorbance maximum at 370 nm. The pigment is also bleached in the dark in the presence of hydroxylamine, but not in the presence of cyanoborohydride. Bleaching in the presence of the latter only occurs in the light, suggesting that the binding site is buried in the interior structure of the protein. Light will cause a change in conformation and a consequent deprotonation of the Schiff-base linkage which will thereby become accessible to the reducing agent
P51. 2.3. Enzymic activity The extinction coefficient of the 440 nm pigment has been calculated to be approximately 47 000 M-’ cm-’ [25]. The photoproduct formed on irradiation (absorbance maximum at about 370 run; see Fig. 2) has an extinction coefficient of 24 000 M-l cm-‘. This value, which is considerably lower than that of the original pigment, suggests a conversion of the chromophore to a cis isomer. When analysed by high performance liquid chromatography (HPLC), the photoproduct was identified as llc&retinal (see Fig. 3). In a subsequent experiment aimed at demonstrating the enzymic activity of the bee pigment, all-tram-retinal was added in excess to a solution containing the protein, and the mixture was then continuously irradiated with light at a wavelength which is only absorbed by retinal bound to protein (528 nm). Figure 4 shows that all-nunsretinal is almost exclusively transformed into 11-&r-retinal. After about 2 h of irradiation at 528 nm, 11-c&retinal accounted for about 50% of the total. The formation of 13-
0.05
1
0.00
250
Fig. 2. Absorbance spectra of bee photoisomerase before (a) and after (b) 40 min irradiation with light of wavelength 490 nm. The inset shows the absorbance difference between curves a and b (adapted from ref. 24).
10
t
I
I
I
1
I
I
I
0
5
10
15
0
5
10
15
min
min
Fig. 3. HPLC analysis of the chromophore of bee photoisomerase before (a) and after (b) 40 min irradiation with monochromatic light of wavelength 490 nm (adapted from ref. 24).
60
40
20 -------¤ 0
0
50
100
150
200
250
300
350
time, min Fig. 4. Isomerization of all-tranr-retinal by continuous irradiation with 528 nm light at room temperature in the presence of 5.6X 10m6 M bee photoisomerase in 0.1 M phosphate buffer (pH 7). The initial concentration of all-&m-retinal was about three times the molar concentration of protein. The isomers are represented by the following symbols: (0) all-tram-retinal; (0) llc&retinal, (I) 13&s-retinal; (0) 9ckretinal (adapted from ref. 22).
11
&retinal, which was already present as a contaminant, and g-c&retinal remained under 10% of the total. The control samples which contained free retinal or retinal and heat-denatured protein were irradiated at the same wavelength as in the experiment of Fig. 4. However, the irradiation failed to produce any retinal photoisomerization. The reaction could be accelerated by using light of a shorter wavelength (500 nm), but many other retinal isomers were obtained as a result of photon absorption by free retinal. Since the photoisomerization of free retinal produces a mixture of isomers, with all-tram-retinal being the major component [26] (11-c&retinal is not the favoured product for energy and steric hindrance reasons), the result of retinal irradiation in the presence of bee photoisomerase is remarkable. The protein directs the isomerization of retinal by light almost exclusively towards the ll-cis-retinal formation. When 13-cis- or g-c&retinal was used to replace all-tram-retinal as the starting isomer, the same stereospecific photoisomerization was obtained in the presence of bee photoisomerase: ll-&retinal was always the favoured product [271.
2.4. Physiological role Although the photoisomerase was isolated from honey-bee heads, it is probable that it is located in the eyes, as it is very similar, if not identical, to the pigment found in the retina of honey-bee drones [21]. According to Goldsmith [19], photoisomerase is a major photopigment in the bee compound eye, accounting for up to 80% of the total photopigments, and has been localized in the primary pigment cells by antibody staining and electron microscopy [28]. Photoisomerase may have different functions. It may act as a carrier for the transport of retinol or retinal between the different compartments of the compound eye, as suggested by the water solubility of the protein. In addition to these functions, bee photoisomerase may be involved in the fundamental process of visual pigment regeneration. The dark degradation of metarhodopsin has been clearly demonstrated for Calliphora, but it is also very likely for honey-bees, as suggested by the decline of all-tram-retinal in honey-bee eyes during prolonged dark and orange adaptation [28]. All-frans-retinal released by the degradation of metarhodopsin could bind to photoisomerase with transformation into ll-cb-retinal on exposure to light in the visible range. Most of the ll-c&retinal could then be reduced to ll-cis-retinol by an alcohol dehydrogenase [28]. Alternatively, ll-cis-retinal could be directly transferred to opsin to regenerate rhodopsin. This complex function, which implies interactions between molecules, wasinvestigated by mixing a solution of photoisomerase, previously loaded with all-tians-retinal and subsequently irradiated, with a suspension of opsin membranes from bleached bovine rod outer segments (ROS). As a result, bovine rhodopsin was reconstituted, i.e. ll-cis-retinal, which is presumably still bound to protein after photoisomerization, can react with membrane opsin to yield rhodopsin
1241. The opsin membranes were prepared from vertebrate ROS due to the difficulty of obtaining these membranes from the honey-bee retina. However, it is difficult to imagine that photoisomerase would interact with the rhodopsin-metarhodopsin system of the bee rhabdomal membranes. Indeed, this system remains stable for hours and rhodopsin is continuously photoregenerated by the absorption of metarhodopsin. As other kinds of regeneration processes, such as dark regeneration, are probably excluded in the bee as in the blowfly [28], photoisomerase is probably involved in the slow
12
renewal of rhabdomal membranes, which therefore implies the biosynthesis of new molecules of rhodopsin. The ability of bee photoisomerase to bind retinal as well as retinol suggests that it could also have dehydrogenase activity. This hypothesis is further supported by the finding that a partial sequence of 30 amino acids, obtained at the University of Leeds, presents good homology with human prostaglandin dehydrogenase [29]. This latter protein, which has a dimer structure (molecular weight of monomer, 29 000), belongs to the family of short-chain alcohol dehydrogenases, which were first discovered in insects. However, experiments performed in the presence of protonated nicotinamide adenine dinucleotide phosphate (NADPH) and retinal in different conditions failed to detect any dehydrogenase activity associated with bee photoisomerase. 3. Retinal
photoisomerase
from cephalopod:
retinochrome
A photosensitive protein, retinochrome, containing all-tram-retinal as its prosthetic group, has been discovered in cephalopods [30-331. Retinochrome is most frequently found in association with membranes of the inner segments of visual cells and was therefore extracted in aqueous solutions containing 2% digitonin. As demonstrated by the reactions with hydroxylamine and sodium borohydride in the dark, retinal is bound to retinochrome via a &hi&base linkage which is accessible to the aqueous environment. Consequently, the absorbance spectrum is dependent on the hydrogen ion concentration. When the pH of the acid extract (pH 5) is adjusted to pH 10, the absorbance maximum at 490 nm decreases and that at 370 nm increases. This behaviour is similar to that of bee photoisomerase whose absorbance maximum changes from 440 to 365 nm when the pH is varied from 6.5 to 9. Furthermore, the absorbance maximum of squid retinochrome is at about 495 nm at pH 6.5 (see Fig. 5). When irradiated with orange light it is bleached giving a photoproduct (metaretinochrome) whose absorbance maximum also depends on pH (460 nm at pH 6.5 or 370 nm at pH 9; see Fig. 5). When this photoproduct is thermally denatured in order to detach the chromophore, and incubated in the dark with cattle opsin, rhodopsin is formed, which indicates that the chromophore of retinochrome is changed by light from all-trans-retinal to 11-&s-retinal [31]. The formation of squid rhodopsin was later demonstrated using rhabdomal membrane preparations containing opsin and metaretinochrome [34]. When retinochrome is irradiated at - 190 “C, it is converted to lumiretinochrome (with an absorbance maximum at 475 nm) and its chromophore is converted from alltrans-retinal to ll-cis-retinal. Lumiretinochrome is changed to metaretinochrome when warmed above 20 “C in the dark. Metaretinochrome and lumiretinochrome can be reconverted to retinochrome by irradiation [35]. Metaretinochrome, when incubated in the dark, can apparently regenerate retinochrome. Such spontaneous regeneration is very slow, lasting for 14 h at 28 “C, and reaches a value of about 70%. Faster regeneration is obtained in the dark when all-puns-retinal is added to metaretinochrome, as it quickly displaces the 11-cis chromophore of metaretinochrome in order to form retinochrome [32]. 4. Comparison
between
bee photoisomerase
and retinochrome
There is a striking similarity between retinochrome and bee photoisomerase. The two proteins can combine with various geometrical isomers of retinal, such as all-
13
pH 5.1 - pH 10.1
-0.1
-
-0.2
-
-0.3
~~~“~~“““““‘~“~‘~“~“~“““‘~~ 300 400 500 Wavelength
600
( nm )
Fig. 5. (a) Absorbance spectra of squid retinochrome in digitonin solution (pH 6.5) at room temperature before (full line) and after (broken line) irradiation with orange light. (b) Difference absorbance of squid retinochrome after irradiation due to pH change from 5.1 to 10.1 (adapted from ref. 33). tranr-retinal,
13-ci.vretinal and 94vretina1, and can catalyse their photoisomerization almost exclusively to ll-ck-retinal [27, 361. This suggests that there is a similar active site for both of these enzymes, i.e. a particular structure of the protein near the binding site which favours the photoproduction of 11-&-retinal. When retinochrome bleached by orange light is further exposed to light in the near-W spectrum (360 nm), it shows a net regeneration [33]; this photoregeneration, which reconverts ll-&-retinal to all-Puns-retinal, is also observed for the bee photoisomerase [29]. It is therefore probable that ll-ck-retinal produced by bee photoisomerase, which is insoluble in aqueous medium, will remain attached to the protein, in order to be transported to where it is needed to regenerate rhodopsin. The essential difference between retinochrome, bee photoisomerase and visual pigment rhodopsin lies in the stereoisomeric form of the chromophore, which is alltrakr-retinal in the two photoisomerases and 11-c&retinal in rhodopsin. Following irradiation of photoisomerase, all-tram-retinal is converted into 1 l&r-retinal, the reverse of the reaction that occurs in rhodopsin. This strongly suggests that retinal photoisomerase and rhodopsin support reciprocal regeneration: photoisomerase acts as a donor of ll&-retinal to opsin for the formation of rhodopsin and, in order to complete the cycle,
14
the degraded metarhodopsin releases all-truns-retinal, which regenerates photoisomerase, and so on. Although cephalopod retinochrome and bee photoisomerase share these and other similarities, such as a low molecular weight of the apoprotein (24 000 for retinochrome and 27 000 for bee photoisomerase), they differ in their solubility properties. Retinochrome is bound to membranes and detergents, such as digitonin, are required to dissolve it. In contrast, bee photoisomerase is water soluble. Therefore rhodopsin regeneration by retinochrome poses some problems since rhodopsin is located in the rhabdomal microvilli of the outer segment of the cephalopod visual cells, whereas retinochrome is mainly stored in the inner segments associated with the membranes of myeloid bodies [37, 381. It has been suggested that some molecules of retinochrome can move from inner to outer segments during light adaptation of squids [37]. Nevertheless, rhabdomal microvilli membranes are so far away from myeloid body membranes that a hypothesis for some suitable shuttle of retinal between these structures has been proposed. This hypothesis was suggested by the discovery of a water-soluble retinal-binding protein in the squid retina [39-41]. However, in honey-bee retina, such an intermediary between the isomerizing enzyme and rhodopsin may not be necessary, as the water-soluble molecule of photoisomerase can directly act as a carrier for 11-k-retinal.
5. Conclusions
and future
trends
The discovery of photoisomerase activity in different visual systems, such as those of bees, flies and cephalopods, leads to the simplified scheme shown in Fig. 6, in which the regeneration of rhodopsin occurs via two pathways only: photoregeneration and renewal, both of which are dependent on light.
hul
Rhodopsin
newly
tletarhodoptin
_-
\
.,
degraded opsin
synthetized opsin
f 11-cis
retinal
_
all-trans
retinal
photoisonerase lRenewal[ Fig. 6. Visual pigment cycle showing the two light-dependent pathways of rhodopsin regeneration. Photoreactions are represented by wavy arrows. The pathways of conversion of retinal to retinol are omitted for simplicity.
15
In future experiments on rhodopsin regeneration, it may be useful to examine the possibility of extending the simple cycle of Fig. 6 to other invertebrate visual systems. The fly Calliphora erythrocephala and the honey-bee Apk mellifera can form visual pigment only when 1 l-&-retinal is present, supplied by a blue-violet-absorbing isomerase [28,42]. No alternative pathways of dark regeneration of rhodopsin from metarhodopsin or biosynthesis of metarhodopsin or rhodopsin from all-pans-retinal, are observed in these visual systems. A dark biosynthesis of rhodopsin can occur, but only when ll&s-retinal is made available by the light-dependent process of isomerization. Dark regeneration of rhodopsin is not observed in cephalopods. Photoregeneration from metarhodopsin to rhodopsin is thought to be the only fast regeneration pathway [8, 91, while the slow regeneration process is supported by the photoisomerase activity of retinochrome. Metaretinochrome, or a soluble intermediary, will provide ll-cisretinal for newly synthesized opsin in order to regenerate rhodopsin during the renewal of photoreceptor membranes [43]. Evidence that a retinochrome-like photoisomerase exists in the retina of a marine gastropod Conomulex luhuanus was reported by Ozaki et al. [44]. While rhodopsin, isolated in the rhabdomal microvillus membranes, has a maximum absorbance at 474 nm, retinochrome, located in the photicvesicles of the visual cell body, absorbs maximally at 510 nm. When irradiated with orange light, retinochrome is bleached to give metaretinochrome changing its chromophore from all-tram-retinal to 11-c&retinal. Other evidence for the existence of photoisomerase activity has been found in the eyes of the arthropod crayfish [45]. In this visual system, the slow regeneration of rhodopsin after irradiation with orange light (which converts rhodopsin to metarhodopsin) is accelerated by the action of blue light. The possible extension of the scheme shown in Fig. 6, to other invertebrate visual systems would greatly simplify visual pigment regeneration. However, such an extension presents difficulties, especially in those systems in which observations have been interpreted by the dark regeneration of rhodopsin [46-48]. For instance, in the fly Drosophila melanogaster, where the visual pigment chromophore is 3-hydroxyretinal, it has been found that 11-cis-3-hydroxyretinal is formed even in the dark [49], excluding the presence of photoisomerase activity in this visual system. Further research is needed in other visual systems in order to evaluate the possibility that the process of rhodopsin regeneration in invertebrates consists of two main pathways: photoregeneration for prompt recovery, and biosynthesis of rhodopsin via photoisomerase activity for slower renewal.
Acknowledgments
This work was supported by P. F. Ingegneria Genetica, C.N.R. We wish to thank Miss Paolina Paperina for typing the manuscript and Dr. Clive Prestt for English correction.
References 1 R. Hubbard, Retinene isomerase, J. Gen. Physiol., 41 (1956) 935-962. 2 C. Baumann, The regeneration and renewal of visual pigment in vertebrates, in H. J. A. Dartnall (ed.), Han&ook of Sensoryphysiology, Vol. VII/l, Springer, Berlin, 1972, pp. 395-416.
16 3 S. Amer and M. Akhtar, Studies on the regeneration of rhodopsin from all-puns-retinal in isolated rat retinae, Nature, 245 (1973) 221-223. 4 C. D. B. Bridges, Vitamin A and the role of the pigment epithelium during bleaching and regeneration of rhodopsin in the frog eye, Exp. Eye Rex, 22 (1976) 435-455. 5 A. Knowles and H. J. A. Dartnall, The regeneration of visual pigments, in H. Davson (ed.), The Eye, Vol. 2B, Academic Press, London, 1977. 6 C. D. Bridges and R. A. Alvarez, The visual cycle operates via an isomerase acting on alltrans-retinol in the pigment epithelium, Science, 236 (1987) 1678-1680. 7 R. W. Young, Visual cells and the concept of renewal, Invest. Ophthalmol., I5 (1975) 700725. 8 R. Hubbard and R. C. C. St. George, The rhodopsin system of the squid, J, Gen. PhysioL, 41 (1958) 501-528. 9 P. K. Brown and P. S. Brown, Visual pigments of octopus and cuttlefish, Nature, 182 (1958) 1288-1290. 10 K. Hamdorf and J. Schwemer, Photoregeneration and the adaptation process in insect photoreceptors, in A. W. Snyder and R. Menzel (eds.), Photoreceptors Optics, Springer, Berlin, 1975, pp. 263-289. 11 K. Hamdorf, The physiology of invertebrate visual pigments, in H. Autum (ed.), Handbook of Sensor Physiology, Vol. VIU6A, Springer, Berlin, 1979, pp. 145-224. 12 P. Hillman, S. Hochstein and B. Minke, Transduction in invertebrate photoreceptors: role of pigment bistability, Physiol. Rev., 63 (1983) 668-772. 13 R. H. White, The effect of light deprivation upon the ultrastructure of the larval mosquito eye. III. Multivesicular bodies and protein uptake, J. w. Zool., 169 (1968) 261-278. 14 A. Perrelet, Protein synthesis in the visual cells of the honeybee drone as studied with electron microscope radioautography, J. Cell. Biol., 55 (1972) 595-605. 15 L. J. Goldman, S. N. Barnes and T. H. Goldsmith, Microspectrophotometry of rhodopsin and metarhodopsin in the moth Galleria, J. Gen. Physiol., 66 (1975) 383-404. 16 A. D. Blest, Photoreceptor membrane turnover in arthropods: comparative studies of breakdown processes and their implications, in T. P. Williams and B. N. Baker (eds.), The Effect of Constant Light on Visual Processes, Plenum, New York, London, 1980, pp. 217245. 17 J. Schwemer, Renewal of visual pigment in photoreceptors of the blowfly, J. Comp. Physiol. A, 154 (1984) 535-547. 18 I. M. Pepe, A. Perrelet and F. Baumann, Isolation by polyacrylamide gel electrophoresis of a light-sensitive vitamin A-protein complex from the retina of the honeybee drone, Vision Res., 16 (1976) 905-908. 19 T. H. Goldsmith, The visual system of the honeybee, Proc. Natl. Acad. Sci. USA, 44 (1958) 123-126. 20 T. H. Goldsmith, The natural history of invertebrate visual pigments, in H. J. A. Darnall (ed.), Handbook of Sensory Physiology. Photochemism of Vision, Vol. 7, Springer, New York, 1972, pp. 684-719. 21 I. M. Pepe and C. Cugnoli, Isolation and characterization of a water-soluble photopigment from honeybee compound eye, Vi&ionRes., 20 (1980) 97-102. 22 J. Schwemer, I. M. Pepe, R. Paulsen and C. Cugnoli, Light-induced trans-cis isomerization of retinal by a protein from honeybee retina, I. Comp. Physiol., 154 (1984) 549-554. 23 I. M. Pepe, C. Cugnoli, M. Peluso, L. Vergani and A. Boero, Structure of a protein catalyzing the formation of llcb-retinal in the visual cycle of invertebrate eyes, Cell Biophys., 10 (1987) 15-22. 24 I. M. Pepe, C. Cugnoli and J. Schwemer, Rhodopsin reconstitution in bleached rod outer segment membranes in the presence of a retinal-binding protein from the honeybee, FEBS Lett., 268 (1990) 177-179. 25 I. M. Pepe, J. Schwemer and R. Paulsen, Characteristics of retinal-binding proteins from the honeybee retina, &ion Res., 22 (1982) 775-781. 26 A. Kropf and R. Hubbard, The photoisomerization of retinal, Photochem. Photobiol., 22 (1970) 249-260.
17 27 C. Cugnoli, R. Mantovani, R. Fioravanti and I. Pepe, ll-&Retinal formation in the light catalyzed by a retinal-binding protein from the honeybee retina, FEBS L&t., 257 (1989) 63-67. 28 W. C. Smith and T. H. Goldsmith, The role of retinal photoisomerase in the visual cycle of the honeybee, J. Gen. Physiol., 97 (1991) 143-165. 29 I. M. Pepe, C. Cugnoli, J. N. Keen and J. B. C. Findlay, Retinal photoisomerase from honeybee compound eye: physiological role, Fifth Int. Conf: on Retinal Protein, Dourdan, France, Abstract 17. 30 T. Hara and R. Hara, New photosensitive pigment found in the retina of squid, Nature, 206 (1965) 1331-1334. 31 T. Hara and R. Hara, Rhodopsin and retinochrome in the squid retina, Nature, 214 (1967) 573-575. 32 T. Hara and R. Hara, Regeneration of squid retinochrome, Nature, 219 (1968) 450-454. 33 T. Hara and R. Hara, Cephalopod retinochrome, in H. J. H. Dartnall (ed.), Handbook of Sensory Physiology, Photochemistty of Vbion, Vol. VII/l, Springer, Berlin, 1972, pp. 720-746. 34 T. Seki, R. Hara and T. Hara, Reconstitution of squid rhodopsin and cattle rhodopsin by use of metaretinochrome in their respective membranes, Erp. Eye Res., 34 (1982) 609-621. 35 T. Hara and R. Hara, Cephalopod Retinochrome, Methods Enzymol., 81 (1982) 827-833. 36 T. Hara and R. Hara, Isomerization of retinal catalysed by retinochrome in the light, Nature, 242 (1973) 3W3. 37 T. Hara and R. Hara, Distribution of rhodopsin and retinochrome in the squid retina, J. Gen. Physiol., 67 (1976) 791-805. 38 K. Osaki, R. Hara and T. Hara, Dependency of absorption characteristics of retinochrome on pH and salts, Exp. Eye Res., 34 (1982) 499-508. 39 0. Osaki, A. Terakita, R. Hara and T. Hara, Isolation and characterization of a retinal binding protein from the squid retina, I&ion Res., 27 (1987) 1057-1070. 40 T. Hara, Visual photoreception of the squid: photopigment regeneration, in T. Hara (ed.), Molecular Physiology of Retinal Proteins, Yamada Science Foundation, Yamada, 1988, pp. 305-310. 41 A. Terakita, R. Hara and T. Ham, Retinal-binding protein as a shuttle for retinal in the rhodopsin-retinochrome system of the squid visual cells, I&on Res., 29 (1989) 639-652. 42 J. Schwemer, Pathways of visual pigment regeneration in fly photoreceptor cells, Biophys. Struct. Mech., 9 (1983) 287-298. 43 R. Hara and T. Hara, Squid m-retinochrome, Vision Res., 24 (1984) 1629-1640. 44 K. Ozaki, A. Terakita, R. Hara and T. Hara, Rhodopsin and retinochrome in the retina of a marine gastropod, Conomuler Luhnanus, Vision Res., 26 (1986) 691-705. 45 T. W. Cronin and T. H. Goldsmith, Dark regeneration of rhodopsin in crayfish photoreceptors, J. Gen. Physiol., 84 (1984) 63-81. 46 D. G. Stavenga, Dark regeneration of invertebrate visual pigments, in A. W. Snyder and R. Menzel (eds.), Photoreceptor Optics, Springer, Berlin, 1975, pp. 290-295. 47 G. D. Bernard, Dark processes following photoconversion of butterfly rhodopsin, Biophys. Struct. Mech., 9 (1983) 277-286. 48 T. H. Goldsmith and M. S. Bruno, Behaviour of rhodopsin and metarhodopsin in isolated rhabdoms of crabs and lobster, in H. Langer (ed.), Biochemistty and Physiology of Visual Pigments, Springer, Berlin, 1973, pp. 147-153. 49 T. Seki, S. Fujishita, M. Ito, N. Matsnoka, C. Kobayashi and K. Tsukida, A fly, Drosophila Melanogaster, forms ll-cb-3-hydroxyretinal in the dark, V&on Rex, 26 (1986) 255-258.
I. Photochem. Photobiol. B: Biol., 13 (1992) 5-17
Invited Review (New Trends in Photobiology)
Retinal Isidoro
photoisomerase:
role in invertebrate
visual cells
M. Pepe and Carlo Cugnoli
Istituto di Cibemetica e Biojisica de1 C.N.R via Dodecaneso 33, Genova 16146 (Italy)
Abstract In invertebrate visual cells, the rhodopsin content is maintained at a high level by the fast process of photoregeneration during daylight. Rhodopsin is converted by photoabsorption to metarhodopsin, which is reconverted to rhodopsin by light. In addition, rhodopsin is regenerated by a slow process of renewal which takes days to complete and involves the biosynthesis of opsin. It is well known that rhodopsin can be formed from opsin only when ll-cis-retinal is present; this requires the existence of an isomerizing enzyme which is capable of transforming all-tram-retinal, released from the degradation of metarhodopsin, into the ll-cis-retinal isomer. In some invertebrate visual systems, experiments on rhodopsin regeneration have been interpreted by assuming that the isomerization reaction is a light-dependent process involving a retinal-protein complex. Two retinal photoisomerases which have been well characterized, i.e. bee photoisomerase and cephalopod retinochrome, are reviewed here. Their properties are compared in order to determine their physiological role, which is likely to be in the renewal of visual pigment rhodopsin. To conclude, a visual pigment cycle is proposed in which rhodopsin regeneration follows two light-dependent pathways. This greatly simplifies the rhodopsin regeneration scheme for invertebrate visual systems.
Keywords: Retinal photoisomerase, cycle.
retinochrome,
rhodopsin renewal, invertebrate
visual
1. Visual pigment regeneration In photoreceptor cells of both vertebrates and invertebrates, the visual pigment rhodopsin is made up of the apoprotein opsin which is bound to the chromophoric group retinal (vitamin A aldehyde) via a Schiff-base linkage. On absorption of a photon, the rhodopsin chromophore is isomerized from 11-c&retinal to all-puns-retinal. This is the first step of the visual process which triggers a dark reaction sequence that eventually leads to the initiation of the electrical signal on the plasma membrane of the photoreceptor.
loll-1344/92/$5.00
0 1992 - Elsevier Sequoia. AU rights reserved
6
In order to restore the rhodopsin content, a process of regeneration takes place. In vertebrates, the all-tram-retinal released by the hydrolysis of rhodopsin to opsin is isomerized back to 11-c&retinal within a few minutes by a dark regeneration process [l-5], the mechanism of which is still poorly understood. Retinol isomerase activity has recently been detected in the pigment epithelium of frogs and rats in which allfruns-retinol is converted to 11-&s-retinol in the dark [6]. However, the enzyme responsible for this activity has not yet been isolated and characterized. Furthermore, it should act in unison with an alcohol dehydrogenase in order to produce 11-&retinal for rhodopsin regeneration. Another process which restores the visual pigment content is the biosynthesis of rhodopsin. A newly synthesized molecule of opsin binds with ll-cb-retinal forming rhodopsin which is incorporated into the newly synthesized disc membrane [7]. The biosynthesis of rhodopsin is part of a continuous renewal of photoreceptor membranes which takes days to complete. However, in the visual system of invertebrates, rhodopsin is not hydrolysed by photon absorption, but is converted to a stable metarhodopsin, which can be reconverted to rhodopsin by light [S, 91. This fast process, called photoregeneration, plays an essential role in maintaining the rhodopsin content, and thus the receptor sensitivity, at a high level [Xl-121. Visual pigments of invertebrates are also regenerated by a slow process of membrane renewal quite similar to that occurring in vertebrates [13-161. Studies on the visual system of the blowfly have been of fundamental importance in clarifying the possible pathways of rhodopsin regeneration in invertebrates. During continuous illumination, a photoequilibrium is reached between rhodopsin and metarhodopsin: the ratio between the concentrations of rhodopsin and metarhodopsin in the fly compound eye depends on the spectral composition of the environmental light and the respective extinction coefficients of the two pigments. The kinetics of the photoreactions are dependent on the light intensity [lo, 121. The alternation between the two states rhodopsin and metarhodopsin, which is caused by photoabsorption, is not of unlimited duration: in the ommatidia there is continuous renewal which implies degradation and resynthesis of molecules. Rhodopsin degrades with a mean lifetime of about 5 days, whereas metarhodopsin decays 60 times more rapidly (mean lifetime of about 20 h) [17]. When the animals are kept in alternating light and darkness, the rhodopsin content remains constant because of photoregeneration and biosynthesis processes. In contrast, when the animals are kept in the dark for days, the rhodopsin content decays exponentially, and this is also the case when all-puns-retinal is injected into their eyes. When the flies are subsequently exposed to light, the rhodopsin content rapidly returns to normal values. The same effect, with the same kinetics, is obtained by injecting 11-c&retinal into the eyes of flies kept in darkness. These experiments clearly demonstrate that retinal isomerization in the fly ommatidia is a light-dependent process. The maximum effect is reached when the animals are kept in light of the blue-violet range (about 440 nm). This may seem strange at first glance: blue-violet light is scarcely absorbed by free retinal, whereas UV light is very effective in the isomerization of retinal, forming a mixture of different isomers such as 13-c&, 9-c& and 11-&r-retinal. However, all-tram-retinal released in the cell by the degradation of metarhodopsin cannot remain free because of its high reactivity, so that possible binding with a protein may have the twofold effect of shifting its absorbance to the visible range and making highly stereospecific photoisomerization possible. Indeed, retinal can bind to an amino group of a protein residue via a protonated Schiff-base linkage, thereby shifting its absorbance maximum from 380 nm
7 to about 440 nm. Moreover, the protein structure around the binding site may favour the photoconversbn of all-tram-retinal into 11-&r-retinal, which is the only isomer that can combine with opsin to give rhodopsin. 2. Retinal photoisomerase
from honey-bee
In order to explain the results from the experiments on the blowfly, Schwemer [17] postulated the existence of a retinal-protein complex capable of photoisomerase activity. However, a pigment had been isolated in the honey-bee compound eye some years earlier, and this pigment seems to have the required characteristics. The story began in Geneva at the University Medical School when Pepe was collaborating with Baumann and Perrelet on a project which aimed to identify and isolate visual pigment rhodopsin in the honey-bee retina. After injecting tritiated retinol into the haemolymph of live drones (honey-bee male) and then waiting for 6 h, it was found that the radioactivity was bound to a water-soluble light-sensitive protein [18]. The chromophore, introduced as retinol, was converted into retinal and bound to the protein via a Schiff-base linkage. Since all rhodopsins studied at that time were insoluble in aqueous media, the authors hypothesized that this retinal-binding protein was a soluble cytoplasmic precursor of rhodopsin. A water-soluble pigment was extracted from honey-bee heads by Goldsmith in 1958 1191. On exposure to light, the pigment was bleached, leading to a maximum absorbance decrease at about 450 nm. Goldsmith [19] first suggested that this pigment could be the visual pigment of the drone because its absorbance maximum matched the sensitivity maximum of the most common photoreceptor in the drone retina. However, later on, the unusual property of water solubility and the fact that most of the animals used for the preparation of this extract were worker bees (dominant receptor at 535 nm) led to some doubt about the nature of this pigment [20]. In 1980, Pepe and Cugnoli [21] extracted and purified a sufficient quantity of water-soluble pigment from heads of honey-bee workers to permit initial characterization. It was found that the pigment was made up of a protein with a molecular weight of about 27 000, which was bound to a chromophore identified as a mixture of all-transretinal and 11-c&retinal. The ratio of all-puns-retinal to 11-&-retinal decreased significantly when the pigment was irradiated with light at wavelengths of greater than 400 nm, indicating that all-trans-retinal was converted by light to 11&s-retinal. The spectrophotometric absorbance change of the pigment following irradiation revealed a maximum at 440 run. Therefore, these findings could no longer be interpreted in favour of a soluble precursor of rhodopsin which, in contrast, should have had an absorbance maximum at 535 nm and a chromophore conversion from 11-&-retinal to all-trans-retinal. During the same year, Pepe joined Schwemer’s group at Ruhr University, Bochum in order to characterize further the properties of bee pigment and to verify the possibility that this pigment could have photoisomerase activity. 2.1. Protein purification and molecular propertks Retinal photoisomerase was extracted from homogenized honey-bee heads in Trisglycine buffer, incubated with tritiated retinol, isolated by preparative electrophoresis and further purified on a diethylaminoethyl (DEAR) coltmrn [21, 221. The molecular weight of the protein was determined to be 50 000 by gel filtration and 27 000 by sodium dodecylsulphate-polyacrylamide gel electrophoresis, suggesting that the protein consists of a dimer [21, 231.
The amino acid composition of the protein was also determined. Polar groups account for about 50% of the total, which explains its water solubibty. Aspartic and glutamic acid are also present in relatively high concentrations. The calculated isoelectric point gives a theoretical value of about 5.6. However, the experimental isoelectric point, determined by electrofocusing of the protein, is about 4.2, which suggests that some hydrophilic basic groups are inside the protein structure. 2.2. Spectrophotometric characteristics As shown by the absorbance spectra in Fig. 1, the protein from honey-bee can bind retinol as well as retinal. The absorbance maximum at 330 nm, shown by the protein sample incubated with retinol, is typical of free retinol in solution, indicating that the binding of the protein with retinol is presumably not covalent. The sample incubated with retina1 absorbs maximally at 440 nm, indicating the formation of a protonated SchitCbase linkage between retinal and protein. The binding site for retinol is presumably the same as that for retinal. This is suggested by the complete displacement of radioactive retinol bound to protein by the addition of an excess of retinal [23]. On changing the pH from 6.5 to 9.5 in the dark, the 440 run pigment is converted into an alkaline form which absorbs maximally at 365 nm, with a pK value of 8.4. This is the characteristic absorbance behaviour of protonated (440 nm) and unprotonated (365 nm) Schitf bases which form between retinal and amino groups. The stoichiometry of the binding between all-trans-retinal and protein was studied and the results indicate that one molecule of retinal can bind one molecule of protein (dimer) with a dissociation constant of K=2X 10m6 M [23].
-0.01
‘I
200
“““I
250
I”’
300
‘I
350
j
‘I
“““‘I
400
450
j ““““‘1 500
550
“I
600
”
650
Wavelength ( nm ) Fig. 1. Absorbance spectra of bee photoisomerase showing its ability to bind retinal as well as retinol. The protein was purified from a homogenate of bee heads incubated with all-tramretinal (a) or all-tranr-retinol (b) in 10 mM Tris-glycine buffer (pH 8.4) at room temperature (adapted from ref. 23).
9
The pigment is bleached on illumination. Figure 2 shows that there is a decrease in the absorbance at 440 nm accompanied by the formation of a photoproduct with an absorbance maximum at 370 nm. The pigment is also bleached in the dark in the presence of hydroxylamine, but not in the presence of cyanoborohydride. Bleaching in the presence of the latter only occurs in the light, suggesting that the binding site is buried in the interior structure of the protein. Light will cause a change in conformation and a consequent deprotonation of the Schiff-base linkage which will thereby become accessible to the reducing agent
P51. 2.3. Enzymic activity The extinction coefficient of the 440 nm pigment has been calculated to be approximately 47 000 M-’ cm-’ [25]. The photoproduct formed on irradiation (absorbance maximum at about 370 run; see Fig. 2) has an extinction coefficient of 24 000 M-l cm-‘. This value, which is considerably lower than that of the original pigment, suggests a conversion of the chromophore to a cis isomer. When analysed by high performance liquid chromatography (HPLC), the photoproduct was identified as llc&retinal (see Fig. 3). In a subsequent experiment aimed at demonstrating the enzymic activity of the bee pigment, all-tram-retinal was added in excess to a solution containing the protein, and the mixture was then continuously irradiated with light at a wavelength which is only absorbed by retinal bound to protein (528 nm). Figure 4 shows that all-nunsretinal is almost exclusively transformed into 11-&r-retinal. After about 2 h of irradiation at 528 nm, 11-c&retinal accounted for about 50% of the total. The formation of 13-
0.05
1
0.00
250
Fig. 2. Absorbance spectra of bee photoisomerase before (a) and after (b) 40 min irradiation with light of wavelength 490 nm. The inset shows the absorbance difference between curves a and b (adapted from ref. 24).
10
t
I
I
I
1
I
I
I
0
5
10
15
0
5
10
15
min
min
Fig. 3. HPLC analysis of the chromophore of bee photoisomerase before (a) and after (b) 40 min irradiation with monochromatic light of wavelength 490 nm (adapted from ref. 24).
60
40
20 -------¤ 0
0
50
100
150
200
250
300
350
time, min Fig. 4. Isomerization of all-tranr-retinal by continuous irradiation with 528 nm light at room temperature in the presence of 5.6X 10m6 M bee photoisomerase in 0.1 M phosphate buffer (pH 7). The initial concentration of all-&m-retinal was about three times the molar concentration of protein. The isomers are represented by the following symbols: (0) all-tram-retinal; (0) llc&retinal, (I) 13&s-retinal; (0) 9ckretinal (adapted from ref. 22).
11
&retinal, which was already present as a contaminant, and g-c&retinal remained under 10% of the total. The control samples which contained free retinal or retinal and heat-denatured protein were irradiated at the same wavelength as in the experiment of Fig. 4. However, the irradiation failed to produce any retinal photoisomerization. The reaction could be accelerated by using light of a shorter wavelength (500 nm), but many other retinal isomers were obtained as a result of photon absorption by free retinal. Since the photoisomerization of free retinal produces a mixture of isomers, with all-tram-retinal being the major component [26] (11-c&retinal is not the favoured product for energy and steric hindrance reasons), the result of retinal irradiation in the presence of bee photoisomerase is remarkable. The protein directs the isomerization of retinal by light almost exclusively towards the ll-cis-retinal formation. When 13-cis- or g-c&retinal was used to replace all-tram-retinal as the starting isomer, the same stereospecific photoisomerization was obtained in the presence of bee photoisomerase: ll-&retinal was always the favoured product [271.
2.4. Physiological role Although the photoisomerase was isolated from honey-bee heads, it is probable that it is located in the eyes, as it is very similar, if not identical, to the pigment found in the retina of honey-bee drones [21]. According to Goldsmith [19], photoisomerase is a major photopigment in the bee compound eye, accounting for up to 80% of the total photopigments, and has been localized in the primary pigment cells by antibody staining and electron microscopy [28]. Photoisomerase may have different functions. It may act as a carrier for the transport of retinol or retinal between the different compartments of the compound eye, as suggested by the water solubility of the protein. In addition to these functions, bee photoisomerase may be involved in the fundamental process of visual pigment regeneration. The dark degradation of metarhodopsin has been clearly demonstrated for Calliphora, but it is also very likely for honey-bees, as suggested by the decline of all-tram-retinal in honey-bee eyes during prolonged dark and orange adaptation [28]. All-frans-retinal released by the degradation of metarhodopsin could bind to photoisomerase with transformation into ll-cb-retinal on exposure to light in the visible range. Most of the ll-c&retinal could then be reduced to ll-cis-retinol by an alcohol dehydrogenase [28]. Alternatively, ll-cis-retinal could be directly transferred to opsin to regenerate rhodopsin. This complex function, which implies interactions between molecules, wasinvestigated by mixing a solution of photoisomerase, previously loaded with all-tians-retinal and subsequently irradiated, with a suspension of opsin membranes from bleached bovine rod outer segments (ROS). As a result, bovine rhodopsin was reconstituted, i.e. ll-cis-retinal, which is presumably still bound to protein after photoisomerization, can react with membrane opsin to yield rhodopsin
1241. The opsin membranes were prepared from vertebrate ROS due to the difficulty of obtaining these membranes from the honey-bee retina. However, it is difficult to imagine that photoisomerase would interact with the rhodopsin-metarhodopsin system of the bee rhabdomal membranes. Indeed, this system remains stable for hours and rhodopsin is continuously photoregenerated by the absorption of metarhodopsin. As other kinds of regeneration processes, such as dark regeneration, are probably excluded in the bee as in the blowfly [28], photoisomerase is probably involved in the slow
12
renewal of rhabdomal membranes, which therefore implies the biosynthesis of new molecules of rhodopsin. The ability of bee photoisomerase to bind retinal as well as retinol suggests that it could also have dehydrogenase activity. This hypothesis is further supported by the finding that a partial sequence of 30 amino acids, obtained at the University of Leeds, presents good homology with human prostaglandin dehydrogenase [29]. This latter protein, which has a dimer structure (molecular weight of monomer, 29 000), belongs to the family of short-chain alcohol dehydrogenases, which were first discovered in insects. However, experiments performed in the presence of protonated nicotinamide adenine dinucleotide phosphate (NADPH) and retinal in different conditions failed to detect any dehydrogenase activity associated with bee photoisomerase. 3. Retinal
photoisomerase
from cephalopod:
retinochrome
A photosensitive protein, retinochrome, containing all-tram-retinal as its prosthetic group, has been discovered in cephalopods [30-331. Retinochrome is most frequently found in association with membranes of the inner segments of visual cells and was therefore extracted in aqueous solutions containing 2% digitonin. As demonstrated by the reactions with hydroxylamine and sodium borohydride in the dark, retinal is bound to retinochrome via a &hi&base linkage which is accessible to the aqueous environment. Consequently, the absorbance spectrum is dependent on the hydrogen ion concentration. When the pH of the acid extract (pH 5) is adjusted to pH 10, the absorbance maximum at 490 nm decreases and that at 370 nm increases. This behaviour is similar to that of bee photoisomerase whose absorbance maximum changes from 440 to 365 nm when the pH is varied from 6.5 to 9. Furthermore, the absorbance maximum of squid retinochrome is at about 495 nm at pH 6.5 (see Fig. 5). When irradiated with orange light it is bleached giving a photoproduct (metaretinochrome) whose absorbance maximum also depends on pH (460 nm at pH 6.5 or 370 nm at pH 9; see Fig. 5). When this photoproduct is thermally denatured in order to detach the chromophore, and incubated in the dark with cattle opsin, rhodopsin is formed, which indicates that the chromophore of retinochrome is changed by light from all-trans-retinal to 11-&s-retinal [31]. The formation of squid rhodopsin was later demonstrated using rhabdomal membrane preparations containing opsin and metaretinochrome [34]. When retinochrome is irradiated at - 190 “C, it is converted to lumiretinochrome (with an absorbance maximum at 475 nm) and its chromophore is converted from alltrans-retinal to ll-cis-retinal. Lumiretinochrome is changed to metaretinochrome when warmed above 20 “C in the dark. Metaretinochrome and lumiretinochrome can be reconverted to retinochrome by irradiation [35]. Metaretinochrome, when incubated in the dark, can apparently regenerate retinochrome. Such spontaneous regeneration is very slow, lasting for 14 h at 28 “C, and reaches a value of about 70%. Faster regeneration is obtained in the dark when all-puns-retinal is added to metaretinochrome, as it quickly displaces the 11-cis chromophore of metaretinochrome in order to form retinochrome [32]. 4. Comparison
between
bee photoisomerase
and retinochrome
There is a striking similarity between retinochrome and bee photoisomerase. The two proteins can combine with various geometrical isomers of retinal, such as all-
13
pH 5.1 - pH 10.1
-0.1
-
-0.2
-
-0.3
~~~“~~“““““‘~“~‘~“~“~“““‘~~ 300 400 500 Wavelength
600
( nm )
Fig. 5. (a) Absorbance spectra of squid retinochrome in digitonin solution (pH 6.5) at room temperature before (full line) and after (broken line) irradiation with orange light. (b) Difference absorbance of squid retinochrome after irradiation due to pH change from 5.1 to 10.1 (adapted from ref. 33). tranr-retinal,
13-ci.vretinal and 94vretina1, and can catalyse their photoisomerization almost exclusively to ll-ck-retinal [27, 361. This suggests that there is a similar active site for both of these enzymes, i.e. a particular structure of the protein near the binding site which favours the photoproduction of 11-&-retinal. When retinochrome bleached by orange light is further exposed to light in the near-W spectrum (360 nm), it shows a net regeneration [33]; this photoregeneration, which reconverts ll-&-retinal to all-Puns-retinal, is also observed for the bee photoisomerase [29]. It is therefore probable that ll-ck-retinal produced by bee photoisomerase, which is insoluble in aqueous medium, will remain attached to the protein, in order to be transported to where it is needed to regenerate rhodopsin. The essential difference between retinochrome, bee photoisomerase and visual pigment rhodopsin lies in the stereoisomeric form of the chromophore, which is alltrakr-retinal in the two photoisomerases and 11-c&retinal in rhodopsin. Following irradiation of photoisomerase, all-tram-retinal is converted into 1 l&r-retinal, the reverse of the reaction that occurs in rhodopsin. This strongly suggests that retinal photoisomerase and rhodopsin support reciprocal regeneration: photoisomerase acts as a donor of ll&-retinal to opsin for the formation of rhodopsin and, in order to complete the cycle,
14
the degraded metarhodopsin releases all-truns-retinal, which regenerates photoisomerase, and so on. Although cephalopod retinochrome and bee photoisomerase share these and other similarities, such as a low molecular weight of the apoprotein (24 000 for retinochrome and 27 000 for bee photoisomerase), they differ in their solubility properties. Retinochrome is bound to membranes and detergents, such as digitonin, are required to dissolve it. In contrast, bee photoisomerase is water soluble. Therefore rhodopsin regeneration by retinochrome poses some problems since rhodopsin is located in the rhabdomal microvilli of the outer segment of the cephalopod visual cells, whereas retinochrome is mainly stored in the inner segments associated with the membranes of myeloid bodies [37, 381. It has been suggested that some molecules of retinochrome can move from inner to outer segments during light adaptation of squids [37]. Nevertheless, rhabdomal microvilli membranes are so far away from myeloid body membranes that a hypothesis for some suitable shuttle of retinal between these structures has been proposed. This hypothesis was suggested by the discovery of a water-soluble retinal-binding protein in the squid retina [39-41]. However, in honey-bee retina, such an intermediary between the isomerizing enzyme and rhodopsin may not be necessary, as the water-soluble molecule of photoisomerase can directly act as a carrier for 11-k-retinal.
5. Conclusions
and future
trends
The discovery of photoisomerase activity in different visual systems, such as those of bees, flies and cephalopods, leads to the simplified scheme shown in Fig. 6, in which the regeneration of rhodopsin occurs via two pathways only: photoregeneration and renewal, both of which are dependent on light.
hul
Rhodopsin
newly
tletarhodoptin
_-
\
.,
degraded opsin
synthetized opsin
f 11-cis
retinal
_
all-trans
retinal
photoisonerase lRenewal[ Fig. 6. Visual pigment cycle showing the two light-dependent pathways of rhodopsin regeneration. Photoreactions are represented by wavy arrows. The pathways of conversion of retinal to retinol are omitted for simplicity.
15
In future experiments on rhodopsin regeneration, it may be useful to examine the possibility of extending the simple cycle of Fig. 6 to other invertebrate visual systems. The fly Calliphora erythrocephala and the honey-bee Apk mellifera can form visual pigment only when 1 l-&-retinal is present, supplied by a blue-violet-absorbing isomerase [28,42]. No alternative pathways of dark regeneration of rhodopsin from metarhodopsin or biosynthesis of metarhodopsin or rhodopsin from all-pans-retinal, are observed in these visual systems. A dark biosynthesis of rhodopsin can occur, but only when ll&s-retinal is made available by the light-dependent process of isomerization. Dark regeneration of rhodopsin is not observed in cephalopods. Photoregeneration from metarhodopsin to rhodopsin is thought to be the only fast regeneration pathway [8, 91, while the slow regeneration process is supported by the photoisomerase activity of retinochrome. Metaretinochrome, or a soluble intermediary, will provide ll-cisretinal for newly synthesized opsin in order to regenerate rhodopsin during the renewal of photoreceptor membranes [43]. Evidence that a retinochrome-like photoisomerase exists in the retina of a marine gastropod Conomulex luhuanus was reported by Ozaki et al. [44]. While rhodopsin, isolated in the rhabdomal microvillus membranes, has a maximum absorbance at 474 nm, retinochrome, located in the photicvesicles of the visual cell body, absorbs maximally at 510 nm. When irradiated with orange light, retinochrome is bleached to give metaretinochrome changing its chromophore from all-tram-retinal to 11-c&retinal. Other evidence for the existence of photoisomerase activity has been found in the eyes of the arthropod crayfish [45]. In this visual system, the slow regeneration of rhodopsin after irradiation with orange light (which converts rhodopsin to metarhodopsin) is accelerated by the action of blue light. The possible extension of the scheme shown in Fig. 6, to other invertebrate visual systems would greatly simplify visual pigment regeneration. However, such an extension presents difficulties, especially in those systems in which observations have been interpreted by the dark regeneration of rhodopsin [46-48]. For instance, in the fly Drosophila melanogaster, where the visual pigment chromophore is 3-hydroxyretinal, it has been found that 11-cis-3-hydroxyretinal is formed even in the dark [49], excluding the presence of photoisomerase activity in this visual system. Further research is needed in other visual systems in order to evaluate the possibility that the process of rhodopsin regeneration in invertebrates consists of two main pathways: photoregeneration for prompt recovery, and biosynthesis of rhodopsin via photoisomerase activity for slower renewal.
Acknowledgments
This work was supported by P. F. Ingegneria Genetica, C.N.R. We wish to thank Miss Paolina Paperina for typing the manuscript and Dr. Clive Prestt for English correction.
References 1 R. Hubbard, Retinene isomerase, J. Gen. Physiol., 41 (1956) 935-962. 2 C. Baumann, The regeneration and renewal of visual pigment in vertebrates, in H. J. A. Dartnall (ed.), Han&ook of Sensoryphysiology, Vol. VII/l, Springer, Berlin, 1972, pp. 395-416.
16 3 S. Amer and M. Akhtar, Studies on the regeneration of rhodopsin from all-puns-retinal in isolated rat retinae, Nature, 245 (1973) 221-223. 4 C. D. B. Bridges, Vitamin A and the role of the pigment epithelium during bleaching and regeneration of rhodopsin in the frog eye, Exp. Eye Rex, 22 (1976) 435-455. 5 A. Knowles and H. J. A. Dartnall, The regeneration of visual pigments, in H. Davson (ed.), The Eye, Vol. 2B, Academic Press, London, 1977. 6 C. D. Bridges and R. A. Alvarez, The visual cycle operates via an isomerase acting on alltrans-retinol in the pigment epithelium, Science, 236 (1987) 1678-1680. 7 R. W. Young, Visual cells and the concept of renewal, Invest. Ophthalmol., I5 (1975) 700725. 8 R. Hubbard and R. C. C. St. George, The rhodopsin system of the squid, J, Gen. PhysioL, 41 (1958) 501-528. 9 P. K. Brown and P. S. Brown, Visual pigments of octopus and cuttlefish, Nature, 182 (1958) 1288-1290. 10 K. Hamdorf and J. Schwemer, Photoregeneration and the adaptation process in insect photoreceptors, in A. W. Snyder and R. Menzel (eds.), Photoreceptors Optics, Springer, Berlin, 1975, pp. 263-289. 11 K. Hamdorf, The physiology of invertebrate visual pigments, in H. Autum (ed.), Handbook of Sensor Physiology, Vol. VIU6A, Springer, Berlin, 1979, pp. 145-224. 12 P. Hillman, S. Hochstein and B. Minke, Transduction in invertebrate photoreceptors: role of pigment bistability, Physiol. Rev., 63 (1983) 668-772. 13 R. H. White, The effect of light deprivation upon the ultrastructure of the larval mosquito eye. III. Multivesicular bodies and protein uptake, J. w. Zool., 169 (1968) 261-278. 14 A. Perrelet, Protein synthesis in the visual cells of the honeybee drone as studied with electron microscope radioautography, J. Cell. Biol., 55 (1972) 595-605. 15 L. J. Goldman, S. N. Barnes and T. H. Goldsmith, Microspectrophotometry of rhodopsin and metarhodopsin in the moth Galleria, J. Gen. Physiol., 66 (1975) 383-404. 16 A. D. Blest, Photoreceptor membrane turnover in arthropods: comparative studies of breakdown processes and their implications, in T. P. Williams and B. N. Baker (eds.), The Effect of Constant Light on Visual Processes, Plenum, New York, London, 1980, pp. 217245. 17 J. Schwemer, Renewal of visual pigment in photoreceptors of the blowfly, J. Comp. Physiol. A, 154 (1984) 535-547. 18 I. M. Pepe, A. Perrelet and F. Baumann, Isolation by polyacrylamide gel electrophoresis of a light-sensitive vitamin A-protein complex from the retina of the honeybee drone, Vision Res., 16 (1976) 905-908. 19 T. H. Goldsmith, The visual system of the honeybee, Proc. Natl. Acad. Sci. USA, 44 (1958) 123-126. 20 T. H. Goldsmith, The natural history of invertebrate visual pigments, in H. J. A. Darnall (ed.), Handbook of Sensory Physiology. Photochemism of Vision, Vol. 7, Springer, New York, 1972, pp. 684-719. 21 I. M. Pepe and C. Cugnoli, Isolation and characterization of a water-soluble photopigment from honeybee compound eye, Vi&ionRes., 20 (1980) 97-102. 22 J. Schwemer, I. M. Pepe, R. Paulsen and C. Cugnoli, Light-induced trans-cis isomerization of retinal by a protein from honeybee retina, I. Comp. Physiol., 154 (1984) 549-554. 23 I. M. Pepe, C. Cugnoli, M. Peluso, L. Vergani and A. Boero, Structure of a protein catalyzing the formation of llcb-retinal in the visual cycle of invertebrate eyes, Cell Biophys., 10 (1987) 15-22. 24 I. M. Pepe, C. Cugnoli and J. Schwemer, Rhodopsin reconstitution in bleached rod outer segment membranes in the presence of a retinal-binding protein from the honeybee, FEBS Lett., 268 (1990) 177-179. 25 I. M. Pepe, J. Schwemer and R. Paulsen, Characteristics of retinal-binding proteins from the honeybee retina, &ion Res., 22 (1982) 775-781. 26 A. Kropf and R. Hubbard, The photoisomerization of retinal, Photochem. Photobiol., 22 (1970) 249-260.
17 27 C. Cugnoli, R. Mantovani, R. Fioravanti and I. Pepe, ll-&Retinal formation in the light catalyzed by a retinal-binding protein from the honeybee retina, FEBS L&t., 257 (1989) 63-67. 28 W. C. Smith and T. H. Goldsmith, The role of retinal photoisomerase in the visual cycle of the honeybee, J. Gen. Physiol., 97 (1991) 143-165. 29 I. M. Pepe, C. Cugnoli, J. N. Keen and J. B. C. Findlay, Retinal photoisomerase from honeybee compound eye: physiological role, Fifth Int. Conf: on Retinal Protein, Dourdan, France, Abstract 17. 30 T. Hara and R. Hara, New photosensitive pigment found in the retina of squid, Nature, 206 (1965) 1331-1334. 31 T. Hara and R. Hara, Rhodopsin and retinochrome in the squid retina, Nature, 214 (1967) 573-575. 32 T. Hara and R. Hara, Regeneration of squid retinochrome, Nature, 219 (1968) 450-454. 33 T. Hara and R. Hara, Cephalopod retinochrome, in H. J. H. Dartnall (ed.), Handbook of Sensory Physiology, Photochemistty of Vbion, Vol. VII/l, Springer, Berlin, 1972, pp. 720-746. 34 T. Seki, R. Hara and T. Hara, Reconstitution of squid rhodopsin and cattle rhodopsin by use of metaretinochrome in their respective membranes, Erp. Eye Res., 34 (1982) 609-621. 35 T. Hara and R. Hara, Cephalopod Retinochrome, Methods Enzymol., 81 (1982) 827-833. 36 T. Hara and R. Hara, Isomerization of retinal catalysed by retinochrome in the light, Nature, 242 (1973) 3W3. 37 T. Hara and R. Hara, Distribution of rhodopsin and retinochrome in the squid retina, J. Gen. Physiol., 67 (1976) 791-805. 38 K. Osaki, R. Hara and T. Hara, Dependency of absorption characteristics of retinochrome on pH and salts, Exp. Eye Res., 34 (1982) 499-508. 39 0. Osaki, A. Terakita, R. Hara and T. Hara, Isolation and characterization of a retinal binding protein from the squid retina, I&ion Res., 27 (1987) 1057-1070. 40 T. Hara, Visual photoreception of the squid: photopigment regeneration, in T. Hara (ed.), Molecular Physiology of Retinal Proteins, Yamada Science Foundation, Yamada, 1988, pp. 305-310. 41 A. Terakita, R. Hara and T. Ham, Retinal-binding protein as a shuttle for retinal in the rhodopsin-retinochrome system of the squid visual cells, I&on Res., 29 (1989) 639-652. 42 J. Schwemer, Pathways of visual pigment regeneration in fly photoreceptor cells, Biophys. Struct. Mech., 9 (1983) 287-298. 43 R. Hara and T. Hara, Squid m-retinochrome, Vision Res., 24 (1984) 1629-1640. 44 K. Ozaki, A. Terakita, R. Hara and T. Hara, Rhodopsin and retinochrome in the retina of a marine gastropod, Conomuler Luhnanus, Vision Res., 26 (1986) 691-705. 45 T. W. Cronin and T. H. Goldsmith, Dark regeneration of rhodopsin in crayfish photoreceptors, J. Gen. Physiol., 84 (1984) 63-81. 46 D. G. Stavenga, Dark regeneration of invertebrate visual pigments, in A. W. Snyder and R. Menzel (eds.), Photoreceptor Optics, Springer, Berlin, 1975, pp. 290-295. 47 G. D. Bernard, Dark processes following photoconversion of butterfly rhodopsin, Biophys. Struct. Mech., 9 (1983) 277-286. 48 T. H. Goldsmith and M. S. Bruno, Behaviour of rhodopsin and metarhodopsin in isolated rhabdoms of crabs and lobster, in H. Langer (ed.), Biochemistty and Physiology of Visual Pigments, Springer, Berlin, 1973, pp. 147-153. 49 T. Seki, S. Fujishita, M. Ito, N. Matsnoka, C. Kobayashi and K. Tsukida, A fly, Drosophila Melanogaster, forms ll-cb-3-hydroxyretinal in the dark, V&on Rex, 26 (1986) 255-258.
