Rods trigger light adaptive retinomotor movements in all spectral cone types of a teleost fish

Vision Res. Vol. 29, No. 4, pp. 389-3%,

1989

Copyright 0

Printed in Great Britain. All rights resewed

0042-6989/89 S3.W + 0.00 1989 Pcrgamon Press pk

RODS TRIGGER LIGHT ADAPTIVE RETINOMOTOR MOVEMENTS IN ALL SPECTRAL CONE TYPES OF A TELEOST FISH M. KIRSCH, H.-J. WAGNER* and R. H. DOUGLAS’ Institut fiir Anatomie und Zellbiologie der Universitaet Marburg, D-3550 Marburg, FRG and ‘City University, London, U.K. (Received 21 April 1988; in revised form 4 July 1988)

AI&met-Action spectra for light induced cone contraction are described for the two spectral cone types of a dichromatic (red/green) cichlid species (Aequidens pulcher). Criterion response thresholds (50% of maximal response amplitude) were determmed for seven wavelengths. After correcting for absorption by cornea and lens, the resulting action spectra were compared with the absorption spectra of the rod and cone visual pigments. We find (i) that the action spectra of red and green sensitive cones are almost identical and (ii) matched most closely the absorption spectrum of the rod visual pigment. We therefore conclude that light adaptive cone contraction is triggered by light absorption in rods, which in the dark adapted state arc located next to the external limiting membrane and are therefore in an optimal position to capture the incident light during early dawn. Possible mechanisms of signal transfer from rods to cones are discussed. Retinomotor movements

Action spectra

Cones

INTRODUCTION

In response to changes in environmental illumination rods, cones and pigment epithelium granules in many lower vertebrates undergo positional changes called retinomotor movements. These movements are most prominent at the beginning and the end of the daily light phase, but may well, in a graded manner, occur when individuals move between differently illuminated areas according to their activity pattern. As most species in question lack pupillary movements, retinomotor responses have been interpreted as an alternative mechanism to regulate the amount of light reaching the photoreceptor outer segments (Walls, 1942). At the same time the rearrangement of photoreceptors, while serving to position receptors appropriately for scotopic and photopic vision, should optimize the utilization of the available space at the level of the external limiting membrane (Walls, 1942; Bumside and Nagle, 1983). Furthermore, retinomotor movements serve to protect the light sensitive rods from direct exposure to intense daylight thus preventing photopigment bleaching through a shielding effect by the pigment epithelium (Douglas, 1982). *To whom correspondence addressed.

and reprint requests should be

Retina

Teleosts

Previous work has demonstrated that two factors contribute to the control of retinomotor movements. Firstly, efferent projections from the CNS may provide a pathway by which central inputs are conveyed to the retina, possibly accounting for the endogenous, rhythmic aspects of retinomotor movements (Ali, 1975; Douglas and Wagner, 1982; Zucker and Dowling, 1987). Secondly, another site of control of these movements resides in the retina itself. Reactions of all active elements are restricted to stimulated areas during in viva experiments (Easter and Macy, 1978) as well as in isolated retinal preparations where efferent control via nerve fibres or blood supply has been eliminated (Kirsch and Wagner, 1986). These results suggest that the initiation of retinomotor movements takes place at the individual photoreceptor level. The cellular basis of retinomotor changes has been studied extensively (Bumside and Nagle, 1983). Several retinal transmitters are involved in their regulation and modulation. Thus, dopamine has been attributed a central role in stimulating light adaptive movements of all elements. At the intracellular level c-AMP and Ca*+ act as second messengers regulating the contractile and force producing elements actin, myosin and microtubules (Pore110 and Burnside, 1984; Dearry and Bumside, 1984, 1986).

390

M.

KIRSCH et al

If retinomotor movements are under local control light absorption by the migrating cells themselves is the most likely trigger for these reactions. This hypothesis is best tested by establishing action spectra of the various components involved. For light adaptive pigment granule movements in frog (Liebman et al., 1969) and trout (Ali and Crouzy, 1968) action spectra have been found to closely resemble the spectral sensitivity of rods. Attempts to determine the spectral sensitivity of cone retinomotor movements have yielded conflicting results (amphibians: Grigonis and Fite, 1983; teleosts: Weidemann, 1966; Douglas and Wagner, 1984). We have therefore determined action spectra for light induced cone contractions in the retina of the dichromatic cichlid fish, Aequidens p&her, which has red-sensitive double cones with a I-max of 617 nm and green-sensitive single cones (I-max 544 nm) in addition to a rod pigment absorbing maximally at 501 nm (Levine and MacNichol, 1979). We show that the action spectrum for cone contraction is very similar in the two cone types and is most closely matched by the absorption spectrum of the rod system, suggesting that cone contraction is triggered through light absorption by the rods. MATERIAL

AND METHODS

Adult Aequidens p&her (Cichlidae, Perciformes, Teleostei; bodylengths 7-l 0 cm), bred in the laboratory were used in all experiments. Fish were brought up and reared under a 12 hr light/dark cycle at 22°C up to an age of approx. 9 months, before they were used. Experiments were performed according to a very strict schedule developed on the basis of results from pilot experiments. Due to the endogenous nature of retinomotor movements we found it necessary to perform all experiments during the normal dark phase of the animals. Furthermore, it was impossible to elicit cone contraction in the middle of the dark phase; therefore, all experiments were started exactly 2 hr before the beginning of the light phase, at which time there was a high and reproducible disposition to light adapt. It also proved essential to place the fish in the adapting tank the evening before the experiment, since mostly light adapted retinae and erratic results were obtained when fish were caught and transferred to the adapting aquarium just prior to the experiment. We suppose stress and the concomi-

tant catecholamine release to be the cause of this effect, and therefore took care to exclude any manipulation disturbing the animals before and during the experiment. Fish were exposed to light in a glass aquarium (40 x 20 x 12 cm) placed in the light beam of a projector, which was equipped with holders for neutral density, interference and heat filters. The front side of the aquarium was covered with a diffusing screen while all remaining sides were frosted white. Narrow-band interference filters (wavelengths: 453, 485, 513, 534. 623. 644, 660 nm, half-bandwidth of 5-10 nm, Schott) and neutral density filters (ND &6 in 0.3 log steps, Schott) were used to regulate wavelength and intensity. As a rule 4-6 different intensities were tested per wavelength in an attempt to cover the range of response amplitudes from fully light to fully dark adapted. Illumination levels (p W/cm2) behind the front side of the aquarium were measured at all wavelength/intensity combinations with a calibrated photodiode (TIL 77 or B 3). Light stimulation was started by a timer and lasted for 75 min, at the end of which the fish were decaptitated, their eyes enucleated, hemisected and placed in ice-cooled fixative (1% paraformaldehyde, 2.5% glutaraldehyde and 3% sucrose in 0.066 M phosphate buffer, pH 7.4) under infrared illumination. Six retinae were obtained by exposing fish either in pairs or alone for each wavelength/intensity combination. After fixation the rim of the eyecups was marked with an incision at the choroideal fissure to ensure correct positioning for subsequent sectioning. Tissue processing was carried out according to conventional EM protocols with postfixation in osmic acid, en bloc staining with uranyl acetate and Epon embedding. Section were stained with Richardson stain (Richardson er al., 1960). Evaluation of cone ellipsoid position was carried out on 1 pm, tangential sections from a region about 300 pm dorsal to the optic nerve (Fig. lc-e). Due to the curvature of the eyecup tangential sections ran through various retinal layers revealing a square mosaic of equal double cones and central single cones typical of cichlids (Wagner, 1978). Cone indices, as a measure of the adaptational state and degree, were defined as the relative distance (b) of ellipsoids from the external limiting membrane (ELM) as a fraction of the total distance (a) between the ELM and Bruch’s membrane (C, = b/a x 100). In pilot experiments, evaluation was performed in radial

Fig. 1. Light-micrographs of semithin sections from Aequidens p&her retinae showing different adaptational states. Vertical sections: (a) and (b). Tangential sections: (exe). Fully dark adapted: (b) and (e); cone ellipsoids are positioned away from the external limiting membrane (ELM); pigment epithelial granules are concentrated in the basal parts of pigment epithelial cells. FUZZYlight adapted: (a) and (c); cone ellipsoids are aligned at the ELM; pigment granules have migrated towards the ELM. Intermediate adaptation: (d); cone ellipsoids occupy a wide range of space between ELM and Bruchs’s membrane (BM); pigment epithelium as in the dark adapted state. Scale bars: 20ym. DC (double cones):(+). SC (single cones) : (-).

391

393

Rods trigger light adaptive retinomotor movements

Neutral

dsnslty in stimulus beam

Fig. 2. Response/log I plots of double (0; DC) and single (A; SC) cone positions after adaptation to white light of different intensities (ND in stimulus beam). Cone position is normalized and expressed as percentage of the maximal displacement between light and dark adapted extremes. Values are means f 1 SD from a minimum of 200 ellipsoids from 4 retinae. (Radiant energy of unattenuated white light; 2.3 x 10-r mW/cm* at the cornea.)

malized to this response range. After plotting them on a probability scale with the ordinate scaled according to the Gauss-integral to linearize the otherwise sigmoid relationship (Fig. 2), regression lines were fitted to the data (Weber, 1967; Fig. 4). From such plots the light intensity eliciting 50% light adaptation could be determined. These thresholds were then converted to radiant energy (p W/cm2) incident on the cornea and corrected for unequal spectral absorption by the lens and cornea and finally expressed as log relative quanta1 sensitivities. The spectral absorption of the lens and cornea (Fig. 3) were determined using a Shimadzu u.v.-240 recording spectrophotometer fitted with an integrating sphere (for further details of the method see Douglas and McGuigan, 1988; in preparation). RESULTS

Figure 2 shows the reactions of red-sensitive double and green-sensitive single cones in response to white light stimulation of different intensities. Under these conditions, red- and green-sensitive cones behave identically, with single cones showing an insignificantly increased sensitivity compared to double cones. Taking criterion values of 10 and 90% of maximal displacement, the range of light intensities between dark and light adapted retinomotor positions spans about 2 log units. Between these values the extent of movement appears to be linearly proportional to the log of light intensity. However, although such an experiment is useful in that it indicates that the two spectral cone types do not migrate separately,

sections (Fig. la, b); in these preparations we found it difficult to clearly identify and obtain in sufficient numbers single cones, due to their size and distribution. Comparing the results from transverse sections with those from tangential sections showed that except for a broader distribution in tangential sections the mean C,-values in both types of evaluation were identical. The measurement of the relevant parameters was carried out semiautomatically with an image analyzing system (IPS, Kontron) where, because of the intense staining, ellipsoids could be automatically discriminated from other retinal structures. For single cones, discrimination using image analysis techniques was not always reliable; therefore their position and distribution was recorded manually with the aid of a drawing microscope. The drawings were then evaluated automatically. In this way several hundred double cones and more than 50 single cones were measured easily in a single section. As all identifiable ellipsoids in a section were measured the evaluation of a single, randomly chosen section per eye was sufficient to obtain enough double and single cone measurements. For each of the stimulation wavelengths I I I I I C,/log I curves were obtained with mean 700 300 600 600 400 C-values ranging from 9.3 f 2.35 (SEM; Wavelength (nm) n = 25) to 68.7 f 6.42 (SEM; n = 19), representFig. 3. Relative spectral transmission of Aequidens p&her ing the fully light and dark adapted state re- lens (....... ) and cornea (-). The transmission at 700 nm spectively. All C,-values from the different has been taken as 100% and each curve represents the wavelength/intensity combinations were noraverage of two lenses/corneas. 0

..:

M. KIRSCHet ul. 513nm

i

23456

23456

534nm

12345

23456

12345

23456

Netma Dsmlty In Stimulus Beam Fig. 4. Response/log I plots of cone positions after adaptation to monochromatic light for seven wavelengths at different intensities (ND in stimulus beam). Upper row: Double cones. Lower row: Single cones. Error bars have not been included in this graph as ordinate is not linearly scaled; they are within the same range as in Fig. 2 (SD: k S-15% of maximal displacement).

no information about absolute spectral sensitivity can be derived from it. Exposure to monochromatic lights of different intensities yielded response/log Zcurves similar to those shown in Fig. 2. They are illustrated in Fig. 4 for both, double (upper row) and single cones (lower row). For the determination of the 50% criterion response value, data were plotted on a probability scale (see Materials and Methods). It is evident that the curves for all wavelengths are essentially parallel indicative of a single photoreceptor mechanism regulating contraction of both cone types. It should be noted that the apparent differences in the slopes of the curves are a result of the non-linearly scaled ordinate, which exaggerates even small differences. When intensity values yielding the criterion response were transferred to log relative quanta1 sensitivity and plotted vs wavelength the action spectra shown in Fig. 5 were obtained. They are very similar for both double and single cones, showing maximal sensitivities at 5 13 nm and rapidly fall off at longer wavelengths. Data points for double and single cones are most often so close, that they could not be represented separately for every stimulation wavelength on the graph. The absorption spectra of the visual pigments were fitted to the double cone action spectrum using a least squares method. DISCUSSION The experiments presented here were designed to ascertain whether myoid contraction in spectrally different cone types can be elicited independently by stimulation with

monochromatic light, in an attempt to further characterize the site of control of retinomotor movements. There are two possible alternative mechanisms for bringing about light induced cone contraction; it could either be caused through light absorption by the cones themselves; alternatively light absorption by the rod system may be the trigger. In order to distinguish between these two alternatives the action spectra for both single and double cone light induced contraction must be compared to the absorption spectra of the various Aequidens p&her visual pigments.

I

I

500

600

Wavelength

4.

0,

700

(nm)

Fig. 5. Action spectra of light induced single (A) and double (0) cone contraction in Aequidens pulcher. For greater clarity, single cone data have been omitted wherever they were identical to the double cone sensitivity. These action spectra have been fitted to the visual pigment absorption spectra of rods (-; I-max 501 nm, optical density 0.54), single cones (0; I-max 544nm. optical density 0.3), and double cones (.......; I-max 617 nm, optical density 0.3) by a least squares method.

Rods trigger light adaptive retinomotor movements

395

1973; Mariani and Lasansky, 1984) and gap junctions between different photoreceptor types Honig (1977) nomograms for vitamin A2 based have been observed in fish (Witkovsky et al., visual pigments with optical densities of 0.54 for 1974). Thus there is a possible morphological the rod pigment (assuming a rod outer segment basis for such an interconnection. Studies using an in vitro pharmacological length of 30 pm and an effective end on density approach (Dearry and Burnside, 1986) further of 0.018 pm-‘) and 0.3 for the cone pigments point to a common, integrating transmitter sys(assuming an outer segment length of 20pm and an effective end on density of 0.015 pm -I). tem for all retinomotor movements. AccordThese absorption spectra can then be fitted to ingly, dopamine, the transmitter used by one the data points using a least squares method class of teleost interplexiform cells, regulates all (Fig. 5). This clearly shows that action spectra the observed changes via D2-receptors. This poses an interesting question regarding the for both single and double cone contraction purely local control described by Easter and are more closely matched by the absorption spectrum of the rod visual pigment (mean Macy (1978) and Kirsch and Wagner (1986). Their findings and the present observation of square deviation = 0.09) than by the absorption spectrum of either of the cone visual pigments an indirect, rod mediated pathway may be (mean square deviation = 0.77 for the green explained if one assumes a functional restriction of the action of dopamine to the retinal area cone pigment). The red visual pigment cannot be fitted precisely to the data points due to where it has been released. This would have uncertainties about its absorption at short to be accomplished by either direct synaptic wavelengths. Even so, it is obvious that it contacts between interplexiform cells and photocannot possibly explain the cone action receptors or by diffusion barriers and effective uptake mechanisms. Alternatively, studies with spectra. Thus light induced cone contraction cells (Dearry and in Aequidens p&her is triggered through the isolated photoreceptor absorption of light by the rod system. Our Burnside, 1986) showed that they were capable of responding appropriately to either light or findings would therefore support a functional alone, suggesting a interconnection between rods and cones in dopamine stimulation the regulation of retinomotor movements. The possible dual regulation of myoid contraction. tendency of data points at long wavelengths to Interestingly, a similar direct effect of light does not appear to be effective in our in vim be higher than the rod absorption spectrum could indicate some kind of cone contribution. experimental system; it is conceivable, that the However, the scatter of all data points with dominating rod mediated pathway can only be respect to the rod absorption spectrum renders shown in the intact retina, and that isolated this possibility rather speculative. photoreceptors exhibit a different behaviour. Several lines of evidence suggest that rods Our findings are at variance with the results may control all retinomotor changes in teleosts. of a previous study (Douglas and Wagner, Firstly, there are no retinomotor movements 1982), in which it was reported that cone in teleost larvae before metamorphosis, i.e. contraction in the monochromatic catfish is before rods are present (Blaxter and Staines, regulated by the red-sensitive cones directly, 1970). Secondly, RPE granule movements in thus providing evidence for an independent frog and trout exhibit action spectra which action. Methodically this study and ours indicate regulation by a rod pigment (Ali and are very similar. Thus, species differences Crouzy, 1968; Liebman et al., 1969). It may as reflected by a greatly different retinal therefore be speculated that there is a uniform morphology (Ali and Anctil, 1976; Wagner, trigger mechanism for the concerted light 1978) and a highly divergent ecological and adaptive action of all elements performing reti- photic environment (Ali and Anctil, 1976) may nomotor movements. Rods appear especially account for the contrasting results. suited for this triggering role as they are the first In summary, we have shown that rods trigger visual receptors to absorb the incoming light retinomotor light adaptation in both cone types while cones and the retinal pigment epithelium of the bichromatic teleost Aequidens pulcher. are arranged at the back of the eye. Turtle and Thus, all photomechanical changes in this urodele rods and cones have been shown to be species seem to be regulated through a common connected by basal processes (telodendria) aris- pathway, which would lead from rods via ing from photoreceptor terminals (Lasansky, bipolar cells to interplexiform cells, which in

The absorption spectra of these pigments have been reconstructed using the Ebrey and

M. KIRSCHet al.

396

turn reach their targets via dopamine released at synaptic sites or acting at distance. Acknowledgements-This

study was supported by a grant from the DFG (WA 348/9) to H.-J. Wagner. We thank Maritha Lippman and Michael Schneider for excellent technical assistance.

photomechanical

cone contraction

in the catfish retina.

Invest. Ophthal. visual. Sci. 25, 524538.

Easter S. S. and Macy A. (1978) Local nomotor activity in the fish retina. 937-942. Ebrey T. G. and Honig B. (1977) New pendent visual pigment nomograms.

control of retiVision Res.

18,

wavelength deVision Res. 17,

147-151.

Grigonis A. M. and Fite K. V. (1983) Photomechanical responses of visual receptors in the retina of the bullfrog REFERENCES Ali M. A. (1975) Retinomotor responses. In Vision in Fishes (Edited by Ali M. A.), pp. 313-355. Plenum Press, New York.

Ali M. A. and Anctil M. (1976) Retinas of Fishes: an Atlas. Springer. Berlin. Ali M. A. and Crouzy R. (1968) Action spectrum and quanta1 thresholds of retinomotor responses in the brook trout, Salvelinus fontinalis (Mitchill). 2. Vergl. Physiol. 59, 8689.

Blaxter J. H. S. and Staines M. (1970) Pure-cone retina and retinomotor responses in larval teleosts. J. Mar. Biol. Ass. U.K. SO, 449460.

Bumside B. and Nagle B. (1983) Retinomotor movements of photoreceptors and retinal pigment epithelium: mechanisms and regulation. In Progress in Retinal Research (Edited by Osborne N. and Chader G.), pp. 67-109. Pergamon Press, New York. Dearry A. and Bumside B. (1984) Effects of extracellular Ca*+, K+ and Na+ on cone and retinal pigment epithelium retinomotor movements in isolated teleost retinas. J. gen. Physiol. 83, 589-61 I. Dearry A. and Burnside B. (1986) Dopaminergic regulation of cone retinomotor movement in isolated teleost retinas: I. Induction of cone contraction is mediated by D2-receptors. J. Neurochem. 46, 1006-1021. Douglas R. H. (1982) The function of photomechanical movements in the retina of the rainbow trout (S&no gairdneri). J. exp. Biol. 96, 389-403.

Douglas R. H. and McGuigan C. M. (1988) The spectral transmission of fresh water teleost ocular media: an interspecific comparison and a guide to potential ultraviolet sensitivity. (In preparation). Douglas R. H. and Wagner H.-J. (1982) Endogenous patterns of photomechanical movements in teleosts and their relation to activity rhythms. Cell Tissue Rex 266, 133-144. Douglas R. H. and Wagner H.-J. (1984) Action spectrum of

(Rana catesbeiana). Brain Behav. Evol. 22, 212-222.

Kirsch M. and Wagner H.-J. (1986) Zur Kontrolle der morphogenetischen Veraenderungen der Goldfischretina. Verh. Anat. Ges. 80, 821-822.

Lasansky A. (1973) Organization of the outer synaptic layer in the retina of the larval tiger salamander. Phi/. Trans. R. Sot. Land. B 265, 471489.

Levine J. S. and MacNichol E. F. (1979) Visual pigments in teleost fishes: effects of habitat, microhabitat and behaviour on visual system evolution. Sensory Processes 3, 95-131.

Liebman P. A., Carroll S. and Laties A. (1969) Spectral sensitivity of retinal screening pigment migration in the frog. Vision Res. 9, 377-384. Mariani A. P. and Lasansky A. (1984) Chemical synapses between turtle photoreceptors. Brain Res. 310, 351-354. Porello K. and Burnside B. (1984) Regulation of reactivated contraction in teleost retinal cone models by calcium and cyclic adenosine monophosphate. J. Cell. Biol. 98, 223&2238. Richardson K. C., Jarett L. and Finke E. H. (1960) Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol. 35, 313-323. Wagner H.-J. (1978) Cell types and connectivity patterns in mosaic retinas. Adv. Anat. EmbryoI. Cell Biol. 5513, l-8 I. Walls G. L. (1942) The Vertebrate Eye, pp. 780. Hafner, New York. Weber E. (1967) Grundriss der Biologischen Starisrik. Fischer, Stuttgart. Weidemann H. L. (1966) Der EinfluB quantengleicher Farblichter auf die Retinomotorik dreier Zapfentypen in der Netzhaut des Guppy (Lebistes reticularus Peters). PhD Thesis, Universitaet Goettingen. Witkovsky P., Shakib M. and Ripps H. (1974) Interreceptoral junctions in the teleost retina. Invest. Ophrhal. visual Sci. 13, 9961009.

Zucker C.L. and Dowling J. E. (1987) Centrifugal fibres synapse on dopaminergic interplexiform cells in the teleost retina. Nature Land. 330, 166-168.