Visual Neuroscience (1992), 8, 243-249. Printed in the USA. Copyright © 1992 Cambridge University Press 0952-5238/92 $5.00 + .00
Light-evoked contraction of red absorbing cones in the Xenopus retina is maximally sensitive to green light
JOSEPH C. BESHARSE1 AND PAUL WITKOVSKY2 'Department of Anatomy and Cell Biology, University of Kansas School of Medicine, Kansas City departments of Ophthalmology and Physiology and Biophysics, New York University School of Medicine, New York (RECEIVED July 16, 1991; ACCEPTED August 21, 1991)
Abstract To test the hypothesis that light-evoked cone contraction in eye cups from Xenopus laevis is controlled through a direct mechanism initiated by the cone's own photopigment, we conducted spectral-sensitivity experiments. We estimate that initiation of contraction of red absorbing cones (611 nm) is 1.5 log units more sensitive to green (533 nm) than red (650 nm) light stimuli. The difference is comparable to that predicted from the spectral-sensitivity function of the green absorbing, principal rod (523 nm). Furthermore, 480-nm and 580-nm stimuli which are absorbed nearly equally by the principal rod have indistinguishable effects on cone contraction. We also found that light blockade of nighttime cone elongation is much more sensitive to green than to red light stimuli. Our observations are inconsistent with the hypothesis tested, and suggest that light-regulated cone motility is controlled through an indirect mechanism initiated primarily by the green absorbing, principal rod. Keywords: Photoreceptors, Retinomotor movement, Dopamine, Spectral sensitivity
a signal to bring cones into play. In agreement with the latter idea, our work on Xenopus eyecups (Pierce & Besharse, 1987) suggested that dim white light was sufficient to maintain cones in their light-adapted state. This finding prompted us to ask which class or classes of photoreceptor triggered cone contraction. The experiments were facilitated by prior microspectrophotometric observations on photopigments in the Xenopus retina (Witkovsky et al., 1981a). Those observations indicated the existence of a class of red absorbing (612 nm) cones along with principal, green absorbing (523 nm) and blue absorbing (445 nm) rods. Although other cone classes may be present in Xenopus (Rohlich et al., 1989), the dominant class absorbs maximally in the red. We measured the light sensitivity of cone motility at different wavelengths and conclude that such motility is highly sensitive to green light. Our data suggest that lightadaptive cone motility in Xenopus laevis is controlled through a polysynaptic pathway activated by the principal, green absorbing rod. These observations were presented previously in abstract form (Besharse & Witkovsky, 1988).
Introduction Rhythmic events in the retina, such as photosensitive membrane turnover, retinomotor motility, and melatonin synthesis, are controlled through interactions of a circadian clock and the daily light-dark cycle (reviewed in Besharse et al., 1988). Retinomotor movements have been particularly useful as a model for elucidation of the mechanisms and circuitry that control rhythmic events. In both Xenopus laevis (Pierce & Besharse, 1985) and green sunfish retinas (Dearry & Burnside, 1986), dopamine contributes to the regulation of light-evoked cone contraction. This implies that processing of photic information through retinal circuits can influence light responses of photoreceptors. Although the details of the retinal circuitry have not been fully elucidated, the dopaminergic cells in Xenopus lack a prominent plexus in the outer retina (Schiitte & Witkovsky, 1991) and photoreceptors are directly responsive to dopamine (Dearry & Burnside, 1986). These observations suggest that dopamine-mediated feedback onto photoreceptors operates through a paracrine mechanism. At first glance these findings may seem counter-intuitive. Since photoreceptors are light-sensitive, the simplest mechanism for control of their light-adaptive motility would be a direct one involving the cells own photopigment and transduction machinery. On the other hand, the rod-to-cone system conversion that occurs around dawn suggests that functioning rods may initiate
Materials and methods We used eyecups from postmetamorphic Xenopus laevis (3-5 cm in length) obtained from Nasco, Inc. (Fort Atkinson, WI), which were maintained at 24-26°C on a 12-h light/ 12-h dark schedule for at least 2 weeks prior to the experiments. Eyecups were prepared using previously established procedures (Besharse et al., 1980; Pierce & Besharse, 1985) in light during the 1-h period just prior to the time of normal light offset, then incubated in
Reprint requests to: Joseph C. Besharse, Department of Anatomy and Cell Biology, University of Kansas Medical Center, 39th and Rainbow Blvd., Kansas City, KS 66103, USA.
243
244 a defined culture medium (see Besharse & Dunis, 1983) containing 35 mM NaHCO 3 . Cultures were gassed with 5% CO 2 /95% O 2 (pH 7.4). To study light-evoked contraction of cones, eyecups were transferred to darkness at the time of normal light offset and incubated for 3 or 4 h to insure that cones were fully elongated. After dark treatment, cone contraction was initiated by exposure to white light (1.5 x 10~3 W/cm 2 /s) or light of different wavelengths and variable intensity for periods ranging 5-300 min (see below). To study blockade of cone elongation, eyecups were prepared as above and placed in darkness or in light of different wavelengths and variable intensity for periods ranging from 15-180 min. Eyecups were incubated in 35-mm Falcon culture dishes in groups of 4 or 5, each dish receiving a defined treatment. They were supported in surgical grade stainless-steel springs attached to the bottom of the culture dish as described previously (Pierce & Besharse, 1987) to insure that they faced the light source throughout the incubation. Light was delivered from an Oriel tungsten-halogen lamp (Oriel, Stratford, CT) via a mirror to a sealed lucite culture chamber. The illuminated field was sufficiently large (15-cm diameter) to permit simultaneous exposure of up to seven culture dishes. To measure effects of different intensities or wavelengths of light simultaneously, calibrated 2 x 2-in. neutral density (Oriel, Stratford, CT) filters were attached to the lid of each dish. The bottom and sides of dishes were masked with black tape to prevent light scatter. Light intensity was calibrated to account for absorption by the lid of the culture dish and medium (4 ml/dish). Intensities were measured (as W/cm 2 /s) incident at the bottom of each culture dish using a calibrated IL700 radiometer with an irradiance detector probe (International Light, Newburyport, MA). To study the effects of different wavelengths of light, we used band-pass filters (half-band width 10 nm; either from Balzers, Liechtenstein or Oriel) placed in a filter holder on the illuminator. Cone length, defined as the distance from the external limiting membrane to the proximal edge of the oil droplet (see Besharse et al., 1982), was measured in light-microscopic preparations. At the end of an experiment, eyecups were fixed under infrared illumination using an FJW Finder Scope or under the ambient illumination of the experiment. Fixation was for 90 min on ice in a mixture of 1.65% glutaraldehyde and 1% OsO 4 in 0.075 M cacodylate buffer (pH 7.4) followed by embedment in Polybed 812/Araldite mixture (Polysciences, Inc., Warrington, PA) and sectioning at 1 /tm. Sections were stained with Azure II or toluidine blue (Sigma Chemical Co., St. Louis, MO). Each experimental treatment included 4 or 5 eyecups from different animals. For each eyecup, lengths of 20 individual cones in the central retina within 1.0 mm of the optic nerve were measured. These measurements were used for frequencydistribution analysis of each treatment. In addition, average cone length in each eyecup was determined and these values were used in subsequent analysis of variance as described in Pierce and Besharse (1985, 1988). In this paper, cone length data are presented as a percent of the maximum movement (±S.E.M.) induced by light referenced to the dark control. Intensity response data were analyzed using a curve-fitting routine called Allfit based on a four-parameter form of the NakaRushton equation (DeLean et al., 1978). This microcomputerbased system was developed for analysis of dose-response data in pharmacology. It permitted simultaneous best-fit analysis of two or more intensity-response curves with independent estimation of the maximum, minimum, half-maximum, and slope of each curve. Standard error estimates for half maxima derived
J.C. Besharse and P.
Witkovsky
from the Allfit program are provided in the captions to Figs. 2, 4, and 8 as an index of the reliability of estimates. Results Light-evoked contraction is sensitive to green light If light absorption by cones controlled their contractile response, cone contraction would be maximally sensitive to red light because Xenopus cones absorb maximally at 611 nm (Witkovsky et al., 1981c). We therefore tested the effects of 30-min treatments of red (650 nm) and green light (533 nm). The choice of wavelengths was dictated by spectrophotometric analysis (Witkovsky et al., 1981a) indicating that the Xenopus retina is dominated by populations of red absorbing cones (611 nm), green absorbing principal rods (523 nm), and a small population of blue absorbing rods (445 nm). The choice of a 30-min light exposure was based on our observation using bright white light that cones contract maximally within 30 min (Fig. 1). Contrary to the hypothesis, cone contraction is more sensitive to green than red light (Fig. 2). From the curves, we estimate a 30-fold (1.5 log unit) greater sensitivity to green light. The half-maximal intensity in green light (2.28 x 109 quanta/ cm 2 /s) corresponds to an estimated absorption of 384 quanta/ rod/s, based on the dimensions of Xenopus rods and their estimated collecting area (Engbretson & Witkovsky, 1978). With alternative assumptions (i.e. different maxima, minima, or slope) about the properties of the curves in Fig. 2, the estimate of a 30-fold greater sensitivity to green light is unaltered. Our confidence in a sensitivity difference of at least this magnitude is further supported by similar differences in response to green and red light in two separate smaller scale experiments. The 1.5 log unit difference in sensitivity to red vs. green light is not statistically different from the 1.55 log unit difference
120
c o £ 3 O X LU
100 80 60
e> o O
40
c u
20
a>
a. -20 -5
5
15
25
35
45
55
65
Time (min) Fig. 1. The time course of light-evoked cone contraction. Eyecups were prepared in room light near the time of light offset and then maintained in darkness for 3 h to fully elongate cones. Eyecups were then exposed to white light (1 x 10~3 W/cm 2 /s) for periods ranging 0-60 min. Cone length is expressed as a percent of the maximum excursion referenced to the dark control. The values plotted are means (open circles) based on 5 eyecups per time point. Vertical bars extend one standard error above and below the mean.
Light-evoked contractions in Xenopus retina
I
ursloi
120 100
u
80
K UJ 0)
•I
1 \
40
c a> o
20
t iX \ \
K \\ \ \i \
1 N
-
60
c o o
V >
K
01
0.
0
245
-
•
-co
i
i
1
5
5
1
|
8 7 Log Quanta
•
•
9
10 -2
cm
i
\l H
l
11
\ i
12
1
-1
sec
Fig. 2. Cone contraction in response to green (533 nm; closed circles) and red (650 nm; open circles) light stimuli of increasing quantal flux compared to a dark control (triangle; -oo). Data are plotted as the mean plus or minus the standard error as in Fig. 1. For some the standard error is plotted only in a single direction. Means are based on 5 eyecups per point except for green stimuli at 11.65 and 10.1 log quanta/cm2/s where JVwas 10 and 3 eyecups, respectively. The template curves are a best fit of the data using Allfit with the assumption that the data for both red and green stimuli share a common maximum, minimum, and slope. The curve fitting the green data is displaced 1.5 log units to the left on the X axis of that fitting the red data. The estimates of half maxima in log quanta/cm2/s with estimated standard error (DeLean et al., 1978) are 650 nm, 10.99 ± 0.74; and 533 nm, 9.49 ±0.49.
predicted from the spectral-sensitivity function for the 523-nm principal rod (see Fig. 4; Witkovsky et al., 19816). Thus, one interpretation of our data is that cone contraction is controlled indirectly through light absorption by the green absorbing, prin-
cipal rod. Nonetheless, we cannot rule out involvement of other photoreceptors, particularly blue absorbing rods. As a test of the hypothesis that green absorbing rods alone control lightevoked cone contraction, we tested the effects of light at 480 and 580 nm which are absorbed nearly equally by the 523-nm rod (see Witkovsky et al., 19816). If the blue absorbing (445 nm) rod were involved in controlling the response, greater blue sensitivity would be expected. To test this possibility, we conducted an experiment in which matched left-right pairs of eyecups were treated simultaneously at each wavelength over the range of expected sensitivity (Fig. 3). Our data show that blue and yellow test lights evoked similar contraction when matched for quantal flux. Thus, as shown in Fig. 4, our sensitivity observations fall close to those expected if the green absorbing principal rod alone controlled light-evoked cone contraction. Characterization of the light response Cone contraction is a slow process in which myoid shortening occurs at a rate of between 1 and 2 ^m/min (Fig. 1). The data in Figs. 2 and 3 were obtained by continuous light exposure throughout the period of cone contraction. To evaluate further the nature of the green light stimulus relevant for cone contraction, we varied the duration of the stimulus at saturating intensity (Fig. 5) and the duration of the exposure at subsaturating intensity (Fig. 6). In the first experiment, the total extent of cone contraction was measured at the end of a 30-min period from the start of the stimulus (533-nm green light at 2.3 x 1012 quanta/cm 2 /s). Bright light stimuli of 180 ms, 1.9 s, and 18 s had no effect on cone position measured at 30 min, whereas a 3-min stimulus caused cones to contract nearly 65% of the full
0
.
Excur;>ion
120
0)
5
100
_
\ \ \
0) (A
E
80
\
3
I -2.0 _
60
O a> o
o
E
_i
40
V
Per
\
-1.0
c
o c
'
«
ivity
^+—
c
c o
CO
m
00
in
m
o
20
Blockade of cone elongation by green light Cones normally elongate at light offset (dusk) and this effect is blocked by continued light exposure. We compared the efficacy of red and green light in blocking cone elongation in order to test the hypothesis that under these light-adapted conditions, maintenance of the contracted state would be mediated by the cones own red absorbing photopigment. We first determined that in darkness cones begin to elongate after a refractory period of about 1 h and are completely elongated within 3 h (Fig. 7). Thus, we chose a treatment period of 3 h during which eyecups were held in red or green light of varying intensity. As shown in Fig. 8, maintenance of contracted cones in the contracted
120 r c o '55
Would a subsaturating green light stimulus cause greater cone contraction if applied over a time interval greater than 30 min? Such a result would be predicted if light intensity was primarily affecting the rate of contraction. To answer this question, we evaluated cone position as a function of time in green light at intensities below (1 x 109 quanta/cmVs) and above (1 x 1010 quanta/cmVs) the half-saturating intensity for green light in Fig. 2. Retinas were evaluated at 30-min intervals up to 300 min at each intensity. At the lower intensity, there appeared to be an initial contraction of about 40% followed by a rebound. At the brighter intensity, cones contracted to just over 60% of the full excursion and remained at roughly that length for the duration of the experiment. All light-exposed eyecups differed significantly from the dark control but there was no significant difference among the light-treated tissues. Although the total number of quanta delivered was the same at 300 min in the lower intensity and 30 min at the higher intensity, the former were not significantly different from dark while the latter achieved a partially contracted state that was sustained for up to 300 min. These data are consistent with the view that light causes contraction to an average cone length that is intensitydependent and sustained.
120 r
Q.
30
90
150
210
270
330
Time (min) Fig. 6. Light-evoked cone contraction as a function of time in green (533 nm) light stimuli of varying intensity. Closed square: darkness; open square: 4.5 x 100" quanta/cm 2 /s; open circles: 1 x 1010 quanta/cm 2 /s; and closed circles: 1 x 109 quanta/cm 2 /s. Stimulus intensity and times are matched so that the total number of quanta delivered for the closed circles at 300 min is equal to that delivered for the open circles at 30 min. Points are means based on 4 or 5 eyecups and are plotted as in Figs. 1-3. The points representing darkness (dark square) and bright green light (open square) were run in parallel separately along with each of the two curves at lower intensities so as to scale the responses.
100
cent Cone Excuirsion
-20 -30
1
80 60 40 20 0 -20 -30
0
30
60
90
120 150
180
Time (min)
excursion (Fig. 5). The 30-min stimulus caused a complete response. Thus, a bright green light stimulus must be integrated over a substantial period (between 18 and 180 s) to cause a significant contractile event. After this, however, substantial contraction will occur in continued darkness.
Fig. 7. Time course of cone elongation in eyecups incubated in darkness. Eyecups were prepared in room light near the time of normal light offset and were then fixed at times ranging from 0-180 min in darkness. Data points (solid circles) are means based on 4 or 5 eyecups per point. Data are plotted as in Figs. 1-3.
247
Light-evoked contractions in Xenopus retina 120 o 100 "3 3 80 x
UJ
• 60 o
- 4° c
£
20
Q.
0 5
6
7
8
9
10
11
12
13
14
Log Quanta cm" 2 sec" 1 Fig. 8. Blockade of cone elongation by green (533 nm; closed circles) and red (650 nm; open circles) light stimuli of increasing quantal flux compared to a dark control (triangle; —oo). Eyecups were prepared in room light near the time of normal light offset and were incubated for 3 h in darkness or in red (650 nm) or green light of variable quantal flux. Data are plotted as in Figs. 1-3. For some the standard error is plotted only in a single direction. Means are based on 3 or 4 eyecups per point except for data points for green light at quantal fluxes of 10.4, 9.5, 8.5, 8.4, and 7.1 which are means based on pooled data of replicate samples of 4 eyecups each (N = 8). The estimates of half maxima in log quanta/cm 2 /s with estimated standard error (DeLean et al., 1978) are 650 nm, 9.8 ± 0.95; and 533 nm, 6.7 ± 0.75.
state was substantially more sensitive to green than red light. Again the curves in Fig. 8 represent a best fit of the data using the curve-fitting approach of DeLean et al. (1978). The estimated difference in half-maximal effect of 3.1 log units is substantially larger than the 1.55 log units predicted from Witkovsky et al. (19816). In addition, the half-maximal intensity for green light estimated from the curve in Fig. 8 corresponds to the absorbance of only about 1 quantum/rod/s. Although substantial variability both within and among treatment groups places limits on our confidence in these estimates, our data suggest that blockade of cone elongation may be more sensitive to green light than the induction of contraction (see Fig. 2). Discussion Our principal finding is that cone contraction in Xenopus eyecups is much more sensitive to green than to red light. This is of interest because prior microspectrophotometric analysis revealed only red absorbing (612 nm) cones in Xenopus (Witkovsky et al., 1981a). If light regulated motility of cones via the cone's own photopigment and transduction machinery, such responses would be expected to be red sensitive. Thus, our findings indicate that light-regulated cone motility is controlled through an indirect mechanism involving some other photoreceptor. The only other known photoreceptor pigments in Xenopus are those of the principal, green absorbing rod (523 nm) and the blue sensitive rod (445 nm). Although we cannot rule out a contribution of the blue absorbing rods, our data favor the green absorbing rod because 480- and 580-nm stimuli matched for equal absorbance by the 523-nm rod were equally effective in initiating cone contraction.
Cone motility also exhibits green sensitivity when light is used to block elongation of cones at the time of normal light offset (dusk). We conducted these experiments because in the light-evoked contraction experiments cones are initially elongated and, therefore, maximally screened from light by the rods. Thus, a cone-based mechanism could be obscured by the conditions of the experiment. Since the blockade experiments begin with light-adapted eyecups, we were able to determine if light absorption by cones would keep them from elongating under starting conditions in which they were optimally positioned for photon absorption. We found that when placed in darkness, cones remain contracted during a refractory period of about 1 h and then slowly elongate. An additional feature is that blockade of elongation was considerably more sensitive to green light than was initiation of cone contraction. The estimated half-maximal intensity for blockade of contraction by green light (about 1 quantum/rod/s) is near the threshold for the aspartate isolated, slow PHI response in superfused Xenopus eyecups (Witkovsky et al., 19816). The difference in sensitivity of the two responses may be related to the fact that the motile machinery necessary for elongation (microtubule based) is different from that mediating contraction (actin based) and that the two processes may be differentially sensitive to the second-messenger systems regulating motility (Porrello & Burnside, 1984; Gilson et al., 1986; Burnside & Dearry, 1986). In nature, high sensitivity to light and a refractory period prior to initiation of elongation would insure that cones remain in their light-adapted position during transient periods of darkness and during any conditions of dim illumination. Another conclusion suggested by our experiments is that light stimuli leading to cone contraction must be integrated over a period of minutes to cause a major contractile response. Furthermore, a given ambient light intensity appears to control the extent of cone contraction. We compared the temporal effects of subsaturating green light stimuli above and below the halfmaximal intensity necessary to cause complete contraction within 30 min. These stimuli were applied over periods of 30 to 300 min. Thus, a stimulus near the threshold for a significant contractile response at 30 min had no additional effect when extended for a 10-fold longer period. Likewise, a 10-fold brighter stimulus that caused a 65% contraction at 30 min did not cause significant additional contraction when extended to 300 min. These results suggest that light of intermediate intensities can cause contraction to intermediate positions that are sustained over extended periods of time. We believe that the indirect mechanism for light-regulated cone motility implied by our spectral-sensitivity experiments is likely to involve dopamine as a light signal. A substantial body of data in Xenopus (reviewed in Besharse et al., 1988) and green sunfish (reviewed in Dearry & Burnside, 1989a; Witkovsky & Dearry, 1991) suggests that dopamine is a principal effector of light-adaptive retinomotor activity. These data indicate that dopamine and D2-selective agonists cause light-adaptive motility and that D2-selective antagonists block the effects of dopamine in light. In Xenopus, dopamine antagonists block both lightevoked cone contraction and, in the light, induce dark-adaptive elongation, suggesting a tonic role for dopamine in maintenance of the light-adapted state (Pierce & Besharse, 1985). These findings form the basis for a model for control of cone motility of Xenopus (Fig. 9). Although many details of the model remain to be worked out, such a model may have broader applicability in regulation of other events in the outer retina. For example,
J.C. Besharse and P. Witkovsky
248
"ON" Bipolar
GABAergic Amacrine
f~\
Y
Dopaminergic Interplexiform
Fig. 9. A diagrammatic model illustrating the retinal circuitry thought to be involved in the regulation of retinomotor cone contraction in Xenopus. Light absorbed by rod photoreceptors (this study) initiates signaling through an indirect pathway that involves "ON"-bipolar neurons as evidenced by the sensitivity of light-evoked contraction to L-APB (Besharse, 1991). Light causes release of dopamine (Boatright et al., 1989) which directly affects photoreceptors (Muresan & Besharse, 1991) through a paracrine pathway (Pierce & Besharse, 1985). Cone movement is tonically influenced by GABAergic neurons that act through the regulation of dopamine release (Pierce & Besharse, 1988).
dopamine appears to play an important role in control of lightadaptation of horizontal cell responses (Witkovsky et al., 1988, 1989; Witkovsky & Shi, 1990) and in the control of light-evoked disc shedding (Pierce & Besharse, 1986; Besharse et al., 1988). The fundamental features of the model are that light absorbed by the principal rods controls cone motility (see above) through an indirect mechanism involving dopamine release from inner retinal neurons. Dopamine is the major catecholamine in Xenopus and appears to be released from interplexiform-like cells with limited distal ramifications (Schiitte & Witkovsky, 1991). The D2-like dopamine receptors are located on cone cell bodies as evidenced by dopamine-induced contractility of isolated cones (Dearry & Burnside, 1986), and by localization of binding to those sites with a D2-antagonist carrying a fluorescent probe (Muresan & Besharse, 1991). Because cones do not appear to receive contacts from dopaminergic cells, our model suggests that dopamine acts as a paracrine effector by diffusing from the inner retina to binding sites on the cones. This part of the model is supported by release studies of endog-
enous dopamine from Xenopus eyecups (Boatright et al., 1989) showing that extracellular levels of dopamine increase fourfold in light under the same conditions used in studies of cone movement. As measured by Witkovsky etal. (1991), dopamine would achieve concentrations in these experiments sufficient to activate D2-receptors. Other aspects of the model take into account the retinal mechanisms that control dopamine release. The idea that the light-signaling pathway includes sign-inverting "ON"-bipolar neurons is supported by the finding that the "ON"-bipolar agonist, L-2-amino-4-phosphonobutyrate (Slaughter & Miller, 1981) blocks light-evoked cone contraction (Besharse, 1991) and dopamine-mediated horizontal cell coupling (Dong & McReynolds, 1991). One possibility is that this pathway is essential to convey the light signal to inner retinal neurons. It is also clear that GABA plays a role. GABA and its agonists cause darkadaptive cone elongation that is blocked by dopamine, and the GABA antagonist, picrotoxin, causes cone contraction that is blocked by the dopamine antagonist, spiperone (Pierce & Besharse, 1988). The details of this analysis are fully consistent with the idea that GABAergic effects on cone motility reflect its well-established role in the tonic regulation of dopamine release (reviewed in Besharse et al., 1988). However, interpretation is clouded by the fact that GABA is found in both amacrine and horizontal cells and that a role for the latter through feedback synapses cannot be ruled out (Pierce & Besharse, 1988). Indirect control through a rod pathway that regulates dopamine release may be a widespread feature of retinomotor motility. It has recently been shown in a cichlid fish that movement of both red and green absorbing cones is highly sensitive to a green stimulus maximally absorbed by rods (Kirsch et al., 1989). Furthermore, the spectral sensitivity for melanin pigment migration in a ranid frog matches that of the green absorbing rod (Liebman et al., 1969), and motility of long wavelength cones is markedly green sensitive as well (Grigonis & Fite, 1983). A possible interpretation is that an indirect pathway such as that described above (see Fig. 9) coordinately controls multiple motility events in the outer retina. This interpretation is consistent with the finding in green sunfish that dopamine is involved in the regulation of motility of rods, cones, and melanin pigment (Dearry & Burnside, 1986, 19896), and that dopamine plays a role in the regulation of pigment movement in a ranid frog (Dearry et al., 1990). Nonetheless, it should be emphasized that, in green sunfish, motility of both rods and cones can be induced directly in isolated cells (Dearry & Burnside, 1986), and that light has been reported to initiate cone movement directly in intact catfish (Douglas & Wagner, 1984). The latter findings leave open the question of the relative importance of direct and indirect mechanisms among teleost species. In our view, the indirect, polysynaptic pathway outlined above offers at least three advantages when compared with a direct pathway. First, it provides a mechanism for control of cone motility by the highly sensitive rod system. Thus, low light stimuli, at either dawn or dusk, would tend to place cone photoreceptors in optimal anatomical position for photon capture. Second, it provides a mechanism for humoral regulation of events that occur on a slow time scale relative to that of visual signaling through the cone pathway. Third, it provides a mechanism for the regulation of the cytoskeletal machinery necessary for cone contraction through alterations in second-messenger systems (i.e. cAMP) that are not directly coupled to the cone's own phototransduction system (Besharse et al., 1982; Porrello & Burn-
Light-evoked contractions in Xenopus retina side, 1984; Gilson et al., 1986). Finally, it should be emphasized that cone contraction is an instructive model system that may provide general insight into the regulation of other rhythmic events in the retina.
249 DOUGLAS, R. & WAGNER, H.-J. (1984). Action spectrum of photomechanical cone contraction in the catfish retina. Investigative Ophthalmology and Visual Science 25, 534-538. ENGBRETSON, G.A. & WITKOVSKY, P. (1978). Rod sensitivity and visual pigment concentration in Xenopus. Journal of General Physiology 72, 801-819. GILSON, C.A., ACKLAND, N. & BURNSIDE, B. (1986). Regulation of re-
Acknowledgments We would like to thank Susan Stone, Greg Cahill, and Mary Pierce for helpful discussions and Gwen Spratt and Jim Geiser for technical assistance. This work was supported by National Institutes of Health Research Grants EY02414 (J.C.B.), EYO357O (P.W.), and Core Grant EY01842 and by a Senior Investigator Award (P.W.) from Research to Prevent Blindness, Inc.
activated elongation in lysed cell models of teleost retinal cones by cAMP and calcium. Journal of Cell Biology 102, 1047-1059. GRIGONIS, A.M. & FITE, K.V. (1983). Photomechanical responses of visual receptors in the retina of the bullfrog (Rana catesbeiana). Brain, Behavior, and Evolution 22, 212-222. KIRSCH, M., WAGNER, H.-J. & DOUGLAS, R.H. (1989). Rods trigger
light-adaptive retinomotor movements in all spectral cone types of a teleost fish. Vision Research 29, 389-396. LIEBMAN, P.A., CARROLL, S. & LATIES, A. (1969). Spectral sensitivity
References BESHARSE, J.C. (1992). The "ON"-bipolar agonist, L-2-amino-4phosphonobutyrate, blocks light-evoked cone contraction in Xenopus eye cups. Neurochemical Research 17, 75-80. BESHARSE, J.C. & DUNIS, D.A. (1983). Rod photoreceptor disc shedding in eye cups: relationship to bicarbonate and amino acids. Experimental Eye Research 36, 567-580. BESHARSE, J.C. & WITKOVSKY, P. (1988). Light-evoked contraction of red cones in Xenopus eye cups is highly sensitive to green light. Investigative Ophthalmology and Visual Science (Abstract Suppl.) 29, 107. BESHARSE, J . C , TERRILL, R.O. & DUNIS, D.A. (1980). Light-evoked disc
shedding by rod photoreceptors in vitro: relationship to medium bicarbonate concentration. Investigative Ophthalmology and Visual Science 19, 1512-1517. BESHARSE, J . C , DUNIS, D.A. & BURNSIDE, B. (1982). Effects of cyclic
adenosine 3',5'-monophosphate on photoreceptor disc shedding and retinomotor movement. Journal of General Physiology 79, 775-790. BESHARSE, J . C , IUVONE, P.M. & PIERCE, M.E. (1988). Regulation of
rhythmic photoreceptor metabolism: a role for post-receptoral neurons. Progress in Retinal Research 7, 21-61. BOATRIOHT, J.H., HOEL, M.J. & IUVONE, P.M. (1989). Stimulation of
endogenous dopamine release and metabolism in amphibian retina by light- and K + -evoked depolarization. Brain Research 482, 164-168. BURNSIDE, B. & DEARRY, A. (1986). Cell motility in the retina. In The Retina: A Model for Cell Biological Studies, Part I, ed. ADLER, D. & FARBER, D., pp. 151-206. New York: Academic Press. DEARRY, A. & BURNSIDE, B. (1986). Dopaminergic regulation of cone retinomotor movement in isolated teleost retinas: I. Induction of cone contraction is mediated by D2 receptors. Journal of Neurochemistry 46, 1006-1021. DEARRY, A. & BURNSIDE, B. (1989a). Regulation of cell motility in teleost retinal photoreceptors and pigment epithelium by dopaminergic D2 receptors. In Extracellular and Intracellular Messengers, ed.
of retinal screening pigment migration in the frog. Vision Research 9, 377-384. MURESAN, Z. & BESHARSE, J.C. (1991). Identification of dopamine D2 receptors in the Xenopus retina by fluorescence microscopy. Investigative Ophthalmology and Visual Science (Abstract Suppl.) 32, 1260. PIERCE, M.E. & BESHARSE, J.C. (1985). Circadian regulation of retinomotor movements 1. Interaction of melatonin and dopamine in the control of cone length. Journal of General Physiology 86, 671-689. PIERCE, M.E. & BESHARSE, J.C. (1986). Melatonin and dopamine interactions in the regulation of rhythmic photoreceptor metabolism. In Pineal and Retinal Relationships, ed. O'BRIEN, P.J. & KLEIN, D.C., pp. 219-237. New York: Academic Press. PIERCE, M.E. & BESHARSE, J.C. (1987). Melatonin and rhythmic photoreceptor metabolism: melatonin-induced cone elongation is blocked at high light intensity. Brain Research 405, 400-404. PIERCE, M.E. & BESHARSE, J.C. (1988). Circadian regulation of retinomotor movements: II. The role of GABA in the regulation of cone position. Journal of Comparative Neurology 270, 279-287. PORRELLO, K. & BURNSIDE, B. (1984). Regulation of reactivated contraction in teleost retinal cone models by calcium and cyclic adenosine monophosphate. Journal of Cell Biology 98, 2230-2238. ROHLICH, P., SZEL, A. & PAPERMASTER, D.S. (1989). Immunocytochem-
ical reactivity of Xenopus laevis retinal rods and cones with several monoclonal antibodies to visual pigments. Journal of Comparative Neurology 290, 105-117. SCHUTTE, M. & WITKOVSKY, P. (1991). Dopaminergic interplexiform cells and centrifugal fibres in the Xenopus retina. Journal ofNeurocylology 20, 195-207. SLAUGHTER, M.M. & MILLER, R.F. (1981). 2-Amino-4-phosphonobu-
tyric acid: a new pharmacological tool for retina research. Science 211, 182-184. WITKOVSKY, P. & SHI, X.-P. (1990). Slow light and dark adaptation of horizontal cells in the Xenopus retina: a role for endogenous dopamine. Visual Neuroscience 5, 405-413. WITKOVSKY, P. & DEARRY, A. (1991). Functional roles of dopamine in the vertebrate retina. Progress in Retinal Research (in press).
REDBURN, D. & PASANTES-MORALES, H., pp. 229-256. New York:
WITKOVSKY, P., LEVINE, J.S., ENGBRETSON, G.A., HASSIN, G. & MAC-
Alan R. Liss, Inc. DEARRY, A. & BURNSIDE, B. (19896). Light-induced dopamine release from teleost retinas acts as a light-adaptive signal to the retinal pigment epithelium. Journal of Neurochemistry 53, 870-878.
NICHOL, E.F. (1981o). A microspectrophotometric study of normal and artificial visual pigments in the photoreceptors of Xenopus laevis. Vision Research 21, 867-873.
DEARRY, A., EDELMAN, J.L., MILLER, S. & BURNSIDE, B. (1990). Do-
pamine induces light-adaptive retinomotor movements in bullfrog cones via D2 receptors and in retinal pigment epithelium via Dl receptors. Journal of Neurochemistry 54, 1367-1378. DELEAN, A., MUNSON, P.J. & RODBARD, D. (1978). Simultaneous anal-
ysis of families of sigmoid curves: application to bioassay, radioligand assay, and physiological dose-response curves. American Journal of Physiology 235, E97-E102. DONG, C.-J. & MCREYNOLDS, J.S. (1991). The relationship between light, dopamine release and horizontal cell coupling in the mudpuppy retina. Journal of Physiology 440, 291-309.
WITKOVSKY, P., YANG, E.-Y. & RIPPS, H. (19816). Properties of the blue
sensitive rod in the Xenopus retina. Vision Research 21, 875-883. WITKOVSKY, P., STONE, S. & BESHARSE, J.C. (1988). Dopamine modi-
fies the balance of rod and cone inputs to horizontal cells of the Xenopus retina. Brain Research 449, 332-336. WITKOVSKY, P., STONE, S. & TRANCHINA, D. (1989). Photoreceptor to
horizontal cell synaptic transfer in the Xenopus retina: modulation by dopamine ligands and a circuit model for interactions of rod and cone inputs. Journal of Neurophysiology 62, 864-881. WITKOVSKY, P., RICE, M. & NICHOLSON, C. (1991). High extracellular
levels of dopamine in Xenopus retina detected by high speed cyclic voltametry. Society of Neuroscience Abstracts 624, 1565.
Light-evoked contraction of red absorbing cones in the Xenopus retina is maximally sensitive to green light
JOSEPH C. BESHARSE1 AND PAUL WITKOVSKY2 'Department of Anatomy and Cell Biology, University of Kansas School of Medicine, Kansas City departments of Ophthalmology and Physiology and Biophysics, New York University School of Medicine, New York (RECEIVED July 16, 1991; ACCEPTED August 21, 1991)
Abstract To test the hypothesis that light-evoked cone contraction in eye cups from Xenopus laevis is controlled through a direct mechanism initiated by the cone's own photopigment, we conducted spectral-sensitivity experiments. We estimate that initiation of contraction of red absorbing cones (611 nm) is 1.5 log units more sensitive to green (533 nm) than red (650 nm) light stimuli. The difference is comparable to that predicted from the spectral-sensitivity function of the green absorbing, principal rod (523 nm). Furthermore, 480-nm and 580-nm stimuli which are absorbed nearly equally by the principal rod have indistinguishable effects on cone contraction. We also found that light blockade of nighttime cone elongation is much more sensitive to green than to red light stimuli. Our observations are inconsistent with the hypothesis tested, and suggest that light-regulated cone motility is controlled through an indirect mechanism initiated primarily by the green absorbing, principal rod. Keywords: Photoreceptors, Retinomotor movement, Dopamine, Spectral sensitivity
a signal to bring cones into play. In agreement with the latter idea, our work on Xenopus eyecups (Pierce & Besharse, 1987) suggested that dim white light was sufficient to maintain cones in their light-adapted state. This finding prompted us to ask which class or classes of photoreceptor triggered cone contraction. The experiments were facilitated by prior microspectrophotometric observations on photopigments in the Xenopus retina (Witkovsky et al., 1981a). Those observations indicated the existence of a class of red absorbing (612 nm) cones along with principal, green absorbing (523 nm) and blue absorbing (445 nm) rods. Although other cone classes may be present in Xenopus (Rohlich et al., 1989), the dominant class absorbs maximally in the red. We measured the light sensitivity of cone motility at different wavelengths and conclude that such motility is highly sensitive to green light. Our data suggest that lightadaptive cone motility in Xenopus laevis is controlled through a polysynaptic pathway activated by the principal, green absorbing rod. These observations were presented previously in abstract form (Besharse & Witkovsky, 1988).
Introduction Rhythmic events in the retina, such as photosensitive membrane turnover, retinomotor motility, and melatonin synthesis, are controlled through interactions of a circadian clock and the daily light-dark cycle (reviewed in Besharse et al., 1988). Retinomotor movements have been particularly useful as a model for elucidation of the mechanisms and circuitry that control rhythmic events. In both Xenopus laevis (Pierce & Besharse, 1985) and green sunfish retinas (Dearry & Burnside, 1986), dopamine contributes to the regulation of light-evoked cone contraction. This implies that processing of photic information through retinal circuits can influence light responses of photoreceptors. Although the details of the retinal circuitry have not been fully elucidated, the dopaminergic cells in Xenopus lack a prominent plexus in the outer retina (Schiitte & Witkovsky, 1991) and photoreceptors are directly responsive to dopamine (Dearry & Burnside, 1986). These observations suggest that dopamine-mediated feedback onto photoreceptors operates through a paracrine mechanism. At first glance these findings may seem counter-intuitive. Since photoreceptors are light-sensitive, the simplest mechanism for control of their light-adaptive motility would be a direct one involving the cells own photopigment and transduction machinery. On the other hand, the rod-to-cone system conversion that occurs around dawn suggests that functioning rods may initiate
Materials and methods We used eyecups from postmetamorphic Xenopus laevis (3-5 cm in length) obtained from Nasco, Inc. (Fort Atkinson, WI), which were maintained at 24-26°C on a 12-h light/ 12-h dark schedule for at least 2 weeks prior to the experiments. Eyecups were prepared using previously established procedures (Besharse et al., 1980; Pierce & Besharse, 1985) in light during the 1-h period just prior to the time of normal light offset, then incubated in
Reprint requests to: Joseph C. Besharse, Department of Anatomy and Cell Biology, University of Kansas Medical Center, 39th and Rainbow Blvd., Kansas City, KS 66103, USA.
243
244 a defined culture medium (see Besharse & Dunis, 1983) containing 35 mM NaHCO 3 . Cultures were gassed with 5% CO 2 /95% O 2 (pH 7.4). To study light-evoked contraction of cones, eyecups were transferred to darkness at the time of normal light offset and incubated for 3 or 4 h to insure that cones were fully elongated. After dark treatment, cone contraction was initiated by exposure to white light (1.5 x 10~3 W/cm 2 /s) or light of different wavelengths and variable intensity for periods ranging 5-300 min (see below). To study blockade of cone elongation, eyecups were prepared as above and placed in darkness or in light of different wavelengths and variable intensity for periods ranging from 15-180 min. Eyecups were incubated in 35-mm Falcon culture dishes in groups of 4 or 5, each dish receiving a defined treatment. They were supported in surgical grade stainless-steel springs attached to the bottom of the culture dish as described previously (Pierce & Besharse, 1987) to insure that they faced the light source throughout the incubation. Light was delivered from an Oriel tungsten-halogen lamp (Oriel, Stratford, CT) via a mirror to a sealed lucite culture chamber. The illuminated field was sufficiently large (15-cm diameter) to permit simultaneous exposure of up to seven culture dishes. To measure effects of different intensities or wavelengths of light simultaneously, calibrated 2 x 2-in. neutral density (Oriel, Stratford, CT) filters were attached to the lid of each dish. The bottom and sides of dishes were masked with black tape to prevent light scatter. Light intensity was calibrated to account for absorption by the lid of the culture dish and medium (4 ml/dish). Intensities were measured (as W/cm 2 /s) incident at the bottom of each culture dish using a calibrated IL700 radiometer with an irradiance detector probe (International Light, Newburyport, MA). To study the effects of different wavelengths of light, we used band-pass filters (half-band width 10 nm; either from Balzers, Liechtenstein or Oriel) placed in a filter holder on the illuminator. Cone length, defined as the distance from the external limiting membrane to the proximal edge of the oil droplet (see Besharse et al., 1982), was measured in light-microscopic preparations. At the end of an experiment, eyecups were fixed under infrared illumination using an FJW Finder Scope or under the ambient illumination of the experiment. Fixation was for 90 min on ice in a mixture of 1.65% glutaraldehyde and 1% OsO 4 in 0.075 M cacodylate buffer (pH 7.4) followed by embedment in Polybed 812/Araldite mixture (Polysciences, Inc., Warrington, PA) and sectioning at 1 /tm. Sections were stained with Azure II or toluidine blue (Sigma Chemical Co., St. Louis, MO). Each experimental treatment included 4 or 5 eyecups from different animals. For each eyecup, lengths of 20 individual cones in the central retina within 1.0 mm of the optic nerve were measured. These measurements were used for frequencydistribution analysis of each treatment. In addition, average cone length in each eyecup was determined and these values were used in subsequent analysis of variance as described in Pierce and Besharse (1985, 1988). In this paper, cone length data are presented as a percent of the maximum movement (±S.E.M.) induced by light referenced to the dark control. Intensity response data were analyzed using a curve-fitting routine called Allfit based on a four-parameter form of the NakaRushton equation (DeLean et al., 1978). This microcomputerbased system was developed for analysis of dose-response data in pharmacology. It permitted simultaneous best-fit analysis of two or more intensity-response curves with independent estimation of the maximum, minimum, half-maximum, and slope of each curve. Standard error estimates for half maxima derived
J.C. Besharse and P.
Witkovsky
from the Allfit program are provided in the captions to Figs. 2, 4, and 8 as an index of the reliability of estimates. Results Light-evoked contraction is sensitive to green light If light absorption by cones controlled their contractile response, cone contraction would be maximally sensitive to red light because Xenopus cones absorb maximally at 611 nm (Witkovsky et al., 1981c). We therefore tested the effects of 30-min treatments of red (650 nm) and green light (533 nm). The choice of wavelengths was dictated by spectrophotometric analysis (Witkovsky et al., 1981a) indicating that the Xenopus retina is dominated by populations of red absorbing cones (611 nm), green absorbing principal rods (523 nm), and a small population of blue absorbing rods (445 nm). The choice of a 30-min light exposure was based on our observation using bright white light that cones contract maximally within 30 min (Fig. 1). Contrary to the hypothesis, cone contraction is more sensitive to green than red light (Fig. 2). From the curves, we estimate a 30-fold (1.5 log unit) greater sensitivity to green light. The half-maximal intensity in green light (2.28 x 109 quanta/ cm 2 /s) corresponds to an estimated absorption of 384 quanta/ rod/s, based on the dimensions of Xenopus rods and their estimated collecting area (Engbretson & Witkovsky, 1978). With alternative assumptions (i.e. different maxima, minima, or slope) about the properties of the curves in Fig. 2, the estimate of a 30-fold greater sensitivity to green light is unaltered. Our confidence in a sensitivity difference of at least this magnitude is further supported by similar differences in response to green and red light in two separate smaller scale experiments. The 1.5 log unit difference in sensitivity to red vs. green light is not statistically different from the 1.55 log unit difference
120
c o £ 3 O X LU
100 80 60
e> o O
40
c u
20
a>
a. -20 -5
5
15
25
35
45
55
65
Time (min) Fig. 1. The time course of light-evoked cone contraction. Eyecups were prepared in room light near the time of light offset and then maintained in darkness for 3 h to fully elongate cones. Eyecups were then exposed to white light (1 x 10~3 W/cm 2 /s) for periods ranging 0-60 min. Cone length is expressed as a percent of the maximum excursion referenced to the dark control. The values plotted are means (open circles) based on 5 eyecups per time point. Vertical bars extend one standard error above and below the mean.
Light-evoked contractions in Xenopus retina
I
ursloi
120 100
u
80
K UJ 0)
•I
1 \
40
c a> o
20
t iX \ \
K \\ \ \i \
1 N
-
60
c o o
V >
K
01
0.
0
245
-
•
-co
i
i
1
5
5
1
|
8 7 Log Quanta
•
•
9
10 -2
cm
i
\l H
l
11
\ i
12
1
-1
sec
Fig. 2. Cone contraction in response to green (533 nm; closed circles) and red (650 nm; open circles) light stimuli of increasing quantal flux compared to a dark control (triangle; -oo). Data are plotted as the mean plus or minus the standard error as in Fig. 1. For some the standard error is plotted only in a single direction. Means are based on 5 eyecups per point except for green stimuli at 11.65 and 10.1 log quanta/cm2/s where JVwas 10 and 3 eyecups, respectively. The template curves are a best fit of the data using Allfit with the assumption that the data for both red and green stimuli share a common maximum, minimum, and slope. The curve fitting the green data is displaced 1.5 log units to the left on the X axis of that fitting the red data. The estimates of half maxima in log quanta/cm2/s with estimated standard error (DeLean et al., 1978) are 650 nm, 10.99 ± 0.74; and 533 nm, 9.49 ±0.49.
predicted from the spectral-sensitivity function for the 523-nm principal rod (see Fig. 4; Witkovsky et al., 19816). Thus, one interpretation of our data is that cone contraction is controlled indirectly through light absorption by the green absorbing, prin-
cipal rod. Nonetheless, we cannot rule out involvement of other photoreceptors, particularly blue absorbing rods. As a test of the hypothesis that green absorbing rods alone control lightevoked cone contraction, we tested the effects of light at 480 and 580 nm which are absorbed nearly equally by the 523-nm rod (see Witkovsky et al., 19816). If the blue absorbing (445 nm) rod were involved in controlling the response, greater blue sensitivity would be expected. To test this possibility, we conducted an experiment in which matched left-right pairs of eyecups were treated simultaneously at each wavelength over the range of expected sensitivity (Fig. 3). Our data show that blue and yellow test lights evoked similar contraction when matched for quantal flux. Thus, as shown in Fig. 4, our sensitivity observations fall close to those expected if the green absorbing principal rod alone controlled light-evoked cone contraction. Characterization of the light response Cone contraction is a slow process in which myoid shortening occurs at a rate of between 1 and 2 ^m/min (Fig. 1). The data in Figs. 2 and 3 were obtained by continuous light exposure throughout the period of cone contraction. To evaluate further the nature of the green light stimulus relevant for cone contraction, we varied the duration of the stimulus at saturating intensity (Fig. 5) and the duration of the exposure at subsaturating intensity (Fig. 6). In the first experiment, the total extent of cone contraction was measured at the end of a 30-min period from the start of the stimulus (533-nm green light at 2.3 x 1012 quanta/cm 2 /s). Bright light stimuli of 180 ms, 1.9 s, and 18 s had no effect on cone position measured at 30 min, whereas a 3-min stimulus caused cones to contract nearly 65% of the full
0
.
Excur;>ion
120
0)
5
100
_
\ \ \
0) (A
E
80
\
3
I -2.0 _
60
O a> o
o
E
_i
40
V
Per
\
-1.0
c
o c
'
«
ivity
^+—
c
c o
CO
m
00
in
m
o
20
Blockade of cone elongation by green light Cones normally elongate at light offset (dusk) and this effect is blocked by continued light exposure. We compared the efficacy of red and green light in blocking cone elongation in order to test the hypothesis that under these light-adapted conditions, maintenance of the contracted state would be mediated by the cones own red absorbing photopigment. We first determined that in darkness cones begin to elongate after a refractory period of about 1 h and are completely elongated within 3 h (Fig. 7). Thus, we chose a treatment period of 3 h during which eyecups were held in red or green light of varying intensity. As shown in Fig. 8, maintenance of contracted cones in the contracted
120 r c o '55
Would a subsaturating green light stimulus cause greater cone contraction if applied over a time interval greater than 30 min? Such a result would be predicted if light intensity was primarily affecting the rate of contraction. To answer this question, we evaluated cone position as a function of time in green light at intensities below (1 x 109 quanta/cmVs) and above (1 x 1010 quanta/cmVs) the half-saturating intensity for green light in Fig. 2. Retinas were evaluated at 30-min intervals up to 300 min at each intensity. At the lower intensity, there appeared to be an initial contraction of about 40% followed by a rebound. At the brighter intensity, cones contracted to just over 60% of the full excursion and remained at roughly that length for the duration of the experiment. All light-exposed eyecups differed significantly from the dark control but there was no significant difference among the light-treated tissues. Although the total number of quanta delivered was the same at 300 min in the lower intensity and 30 min at the higher intensity, the former were not significantly different from dark while the latter achieved a partially contracted state that was sustained for up to 300 min. These data are consistent with the view that light causes contraction to an average cone length that is intensitydependent and sustained.
120 r
Q.
30
90
150
210
270
330
Time (min) Fig. 6. Light-evoked cone contraction as a function of time in green (533 nm) light stimuli of varying intensity. Closed square: darkness; open square: 4.5 x 100" quanta/cm 2 /s; open circles: 1 x 1010 quanta/cm 2 /s; and closed circles: 1 x 109 quanta/cm 2 /s. Stimulus intensity and times are matched so that the total number of quanta delivered for the closed circles at 300 min is equal to that delivered for the open circles at 30 min. Points are means based on 4 or 5 eyecups and are plotted as in Figs. 1-3. The points representing darkness (dark square) and bright green light (open square) were run in parallel separately along with each of the two curves at lower intensities so as to scale the responses.
100
cent Cone Excuirsion
-20 -30
1
80 60 40 20 0 -20 -30
0
30
60
90
120 150
180
Time (min)
excursion (Fig. 5). The 30-min stimulus caused a complete response. Thus, a bright green light stimulus must be integrated over a substantial period (between 18 and 180 s) to cause a significant contractile event. After this, however, substantial contraction will occur in continued darkness.
Fig. 7. Time course of cone elongation in eyecups incubated in darkness. Eyecups were prepared in room light near the time of normal light offset and were then fixed at times ranging from 0-180 min in darkness. Data points (solid circles) are means based on 4 or 5 eyecups per point. Data are plotted as in Figs. 1-3.
247
Light-evoked contractions in Xenopus retina 120 o 100 "3 3 80 x
UJ
• 60 o
- 4° c
£
20
Q.
0 5
6
7
8
9
10
11
12
13
14
Log Quanta cm" 2 sec" 1 Fig. 8. Blockade of cone elongation by green (533 nm; closed circles) and red (650 nm; open circles) light stimuli of increasing quantal flux compared to a dark control (triangle; —oo). Eyecups were prepared in room light near the time of normal light offset and were incubated for 3 h in darkness or in red (650 nm) or green light of variable quantal flux. Data are plotted as in Figs. 1-3. For some the standard error is plotted only in a single direction. Means are based on 3 or 4 eyecups per point except for data points for green light at quantal fluxes of 10.4, 9.5, 8.5, 8.4, and 7.1 which are means based on pooled data of replicate samples of 4 eyecups each (N = 8). The estimates of half maxima in log quanta/cm 2 /s with estimated standard error (DeLean et al., 1978) are 650 nm, 9.8 ± 0.95; and 533 nm, 6.7 ± 0.75.
state was substantially more sensitive to green than red light. Again the curves in Fig. 8 represent a best fit of the data using the curve-fitting approach of DeLean et al. (1978). The estimated difference in half-maximal effect of 3.1 log units is substantially larger than the 1.55 log units predicted from Witkovsky et al. (19816). In addition, the half-maximal intensity for green light estimated from the curve in Fig. 8 corresponds to the absorbance of only about 1 quantum/rod/s. Although substantial variability both within and among treatment groups places limits on our confidence in these estimates, our data suggest that blockade of cone elongation may be more sensitive to green light than the induction of contraction (see Fig. 2). Discussion Our principal finding is that cone contraction in Xenopus eyecups is much more sensitive to green than to red light. This is of interest because prior microspectrophotometric analysis revealed only red absorbing (612 nm) cones in Xenopus (Witkovsky et al., 1981a). If light regulated motility of cones via the cone's own photopigment and transduction machinery, such responses would be expected to be red sensitive. Thus, our findings indicate that light-regulated cone motility is controlled through an indirect mechanism involving some other photoreceptor. The only other known photoreceptor pigments in Xenopus are those of the principal, green absorbing rod (523 nm) and the blue sensitive rod (445 nm). Although we cannot rule out a contribution of the blue absorbing rods, our data favor the green absorbing rod because 480- and 580-nm stimuli matched for equal absorbance by the 523-nm rod were equally effective in initiating cone contraction.
Cone motility also exhibits green sensitivity when light is used to block elongation of cones at the time of normal light offset (dusk). We conducted these experiments because in the light-evoked contraction experiments cones are initially elongated and, therefore, maximally screened from light by the rods. Thus, a cone-based mechanism could be obscured by the conditions of the experiment. Since the blockade experiments begin with light-adapted eyecups, we were able to determine if light absorption by cones would keep them from elongating under starting conditions in which they were optimally positioned for photon absorption. We found that when placed in darkness, cones remain contracted during a refractory period of about 1 h and then slowly elongate. An additional feature is that blockade of elongation was considerably more sensitive to green light than was initiation of cone contraction. The estimated half-maximal intensity for blockade of contraction by green light (about 1 quantum/rod/s) is near the threshold for the aspartate isolated, slow PHI response in superfused Xenopus eyecups (Witkovsky et al., 19816). The difference in sensitivity of the two responses may be related to the fact that the motile machinery necessary for elongation (microtubule based) is different from that mediating contraction (actin based) and that the two processes may be differentially sensitive to the second-messenger systems regulating motility (Porrello & Burnside, 1984; Gilson et al., 1986; Burnside & Dearry, 1986). In nature, high sensitivity to light and a refractory period prior to initiation of elongation would insure that cones remain in their light-adapted position during transient periods of darkness and during any conditions of dim illumination. Another conclusion suggested by our experiments is that light stimuli leading to cone contraction must be integrated over a period of minutes to cause a major contractile response. Furthermore, a given ambient light intensity appears to control the extent of cone contraction. We compared the temporal effects of subsaturating green light stimuli above and below the halfmaximal intensity necessary to cause complete contraction within 30 min. These stimuli were applied over periods of 30 to 300 min. Thus, a stimulus near the threshold for a significant contractile response at 30 min had no additional effect when extended for a 10-fold longer period. Likewise, a 10-fold brighter stimulus that caused a 65% contraction at 30 min did not cause significant additional contraction when extended to 300 min. These results suggest that light of intermediate intensities can cause contraction to intermediate positions that are sustained over extended periods of time. We believe that the indirect mechanism for light-regulated cone motility implied by our spectral-sensitivity experiments is likely to involve dopamine as a light signal. A substantial body of data in Xenopus (reviewed in Besharse et al., 1988) and green sunfish (reviewed in Dearry & Burnside, 1989a; Witkovsky & Dearry, 1991) suggests that dopamine is a principal effector of light-adaptive retinomotor activity. These data indicate that dopamine and D2-selective agonists cause light-adaptive motility and that D2-selective antagonists block the effects of dopamine in light. In Xenopus, dopamine antagonists block both lightevoked cone contraction and, in the light, induce dark-adaptive elongation, suggesting a tonic role for dopamine in maintenance of the light-adapted state (Pierce & Besharse, 1985). These findings form the basis for a model for control of cone motility of Xenopus (Fig. 9). Although many details of the model remain to be worked out, such a model may have broader applicability in regulation of other events in the outer retina. For example,
J.C. Besharse and P. Witkovsky
248
"ON" Bipolar
GABAergic Amacrine
f~\
Y
Dopaminergic Interplexiform
Fig. 9. A diagrammatic model illustrating the retinal circuitry thought to be involved in the regulation of retinomotor cone contraction in Xenopus. Light absorbed by rod photoreceptors (this study) initiates signaling through an indirect pathway that involves "ON"-bipolar neurons as evidenced by the sensitivity of light-evoked contraction to L-APB (Besharse, 1991). Light causes release of dopamine (Boatright et al., 1989) which directly affects photoreceptors (Muresan & Besharse, 1991) through a paracrine pathway (Pierce & Besharse, 1985). Cone movement is tonically influenced by GABAergic neurons that act through the regulation of dopamine release (Pierce & Besharse, 1988).
dopamine appears to play an important role in control of lightadaptation of horizontal cell responses (Witkovsky et al., 1988, 1989; Witkovsky & Shi, 1990) and in the control of light-evoked disc shedding (Pierce & Besharse, 1986; Besharse et al., 1988). The fundamental features of the model are that light absorbed by the principal rods controls cone motility (see above) through an indirect mechanism involving dopamine release from inner retinal neurons. Dopamine is the major catecholamine in Xenopus and appears to be released from interplexiform-like cells with limited distal ramifications (Schiitte & Witkovsky, 1991). The D2-like dopamine receptors are located on cone cell bodies as evidenced by dopamine-induced contractility of isolated cones (Dearry & Burnside, 1986), and by localization of binding to those sites with a D2-antagonist carrying a fluorescent probe (Muresan & Besharse, 1991). Because cones do not appear to receive contacts from dopaminergic cells, our model suggests that dopamine acts as a paracrine effector by diffusing from the inner retina to binding sites on the cones. This part of the model is supported by release studies of endog-
enous dopamine from Xenopus eyecups (Boatright et al., 1989) showing that extracellular levels of dopamine increase fourfold in light under the same conditions used in studies of cone movement. As measured by Witkovsky etal. (1991), dopamine would achieve concentrations in these experiments sufficient to activate D2-receptors. Other aspects of the model take into account the retinal mechanisms that control dopamine release. The idea that the light-signaling pathway includes sign-inverting "ON"-bipolar neurons is supported by the finding that the "ON"-bipolar agonist, L-2-amino-4-phosphonobutyrate (Slaughter & Miller, 1981) blocks light-evoked cone contraction (Besharse, 1991) and dopamine-mediated horizontal cell coupling (Dong & McReynolds, 1991). One possibility is that this pathway is essential to convey the light signal to inner retinal neurons. It is also clear that GABA plays a role. GABA and its agonists cause darkadaptive cone elongation that is blocked by dopamine, and the GABA antagonist, picrotoxin, causes cone contraction that is blocked by the dopamine antagonist, spiperone (Pierce & Besharse, 1988). The details of this analysis are fully consistent with the idea that GABAergic effects on cone motility reflect its well-established role in the tonic regulation of dopamine release (reviewed in Besharse et al., 1988). However, interpretation is clouded by the fact that GABA is found in both amacrine and horizontal cells and that a role for the latter through feedback synapses cannot be ruled out (Pierce & Besharse, 1988). Indirect control through a rod pathway that regulates dopamine release may be a widespread feature of retinomotor motility. It has recently been shown in a cichlid fish that movement of both red and green absorbing cones is highly sensitive to a green stimulus maximally absorbed by rods (Kirsch et al., 1989). Furthermore, the spectral sensitivity for melanin pigment migration in a ranid frog matches that of the green absorbing rod (Liebman et al., 1969), and motility of long wavelength cones is markedly green sensitive as well (Grigonis & Fite, 1983). A possible interpretation is that an indirect pathway such as that described above (see Fig. 9) coordinately controls multiple motility events in the outer retina. This interpretation is consistent with the finding in green sunfish that dopamine is involved in the regulation of motility of rods, cones, and melanin pigment (Dearry & Burnside, 1986, 19896), and that dopamine plays a role in the regulation of pigment movement in a ranid frog (Dearry et al., 1990). Nonetheless, it should be emphasized that, in green sunfish, motility of both rods and cones can be induced directly in isolated cells (Dearry & Burnside, 1986), and that light has been reported to initiate cone movement directly in intact catfish (Douglas & Wagner, 1984). The latter findings leave open the question of the relative importance of direct and indirect mechanisms among teleost species. In our view, the indirect, polysynaptic pathway outlined above offers at least three advantages when compared with a direct pathway. First, it provides a mechanism for control of cone motility by the highly sensitive rod system. Thus, low light stimuli, at either dawn or dusk, would tend to place cone photoreceptors in optimal anatomical position for photon capture. Second, it provides a mechanism for humoral regulation of events that occur on a slow time scale relative to that of visual signaling through the cone pathway. Third, it provides a mechanism for the regulation of the cytoskeletal machinery necessary for cone contraction through alterations in second-messenger systems (i.e. cAMP) that are not directly coupled to the cone's own phototransduction system (Besharse et al., 1982; Porrello & Burn-
Light-evoked contractions in Xenopus retina side, 1984; Gilson et al., 1986). Finally, it should be emphasized that cone contraction is an instructive model system that may provide general insight into the regulation of other rhythmic events in the retina.
249 DOUGLAS, R. & WAGNER, H.-J. (1984). Action spectrum of photomechanical cone contraction in the catfish retina. Investigative Ophthalmology and Visual Science 25, 534-538. ENGBRETSON, G.A. & WITKOVSKY, P. (1978). Rod sensitivity and visual pigment concentration in Xenopus. Journal of General Physiology 72, 801-819. GILSON, C.A., ACKLAND, N. & BURNSIDE, B. (1986). Regulation of re-
Acknowledgments We would like to thank Susan Stone, Greg Cahill, and Mary Pierce for helpful discussions and Gwen Spratt and Jim Geiser for technical assistance. This work was supported by National Institutes of Health Research Grants EY02414 (J.C.B.), EYO357O (P.W.), and Core Grant EY01842 and by a Senior Investigator Award (P.W.) from Research to Prevent Blindness, Inc.
activated elongation in lysed cell models of teleost retinal cones by cAMP and calcium. Journal of Cell Biology 102, 1047-1059. GRIGONIS, A.M. & FITE, K.V. (1983). Photomechanical responses of visual receptors in the retina of the bullfrog (Rana catesbeiana). Brain, Behavior, and Evolution 22, 212-222. KIRSCH, M., WAGNER, H.-J. & DOUGLAS, R.H. (1989). Rods trigger
light-adaptive retinomotor movements in all spectral cone types of a teleost fish. Vision Research 29, 389-396. LIEBMAN, P.A., CARROLL, S. & LATIES, A. (1969). Spectral sensitivity
References BESHARSE, J.C. (1992). The "ON"-bipolar agonist, L-2-amino-4phosphonobutyrate, blocks light-evoked cone contraction in Xenopus eye cups. Neurochemical Research 17, 75-80. BESHARSE, J.C. & DUNIS, D.A. (1983). Rod photoreceptor disc shedding in eye cups: relationship to bicarbonate and amino acids. Experimental Eye Research 36, 567-580. BESHARSE, J.C. & WITKOVSKY, P. (1988). Light-evoked contraction of red cones in Xenopus eye cups is highly sensitive to green light. Investigative Ophthalmology and Visual Science (Abstract Suppl.) 29, 107. BESHARSE, J . C , TERRILL, R.O. & DUNIS, D.A. (1980). Light-evoked disc
shedding by rod photoreceptors in vitro: relationship to medium bicarbonate concentration. Investigative Ophthalmology and Visual Science 19, 1512-1517. BESHARSE, J . C , DUNIS, D.A. & BURNSIDE, B. (1982). Effects of cyclic
adenosine 3',5'-monophosphate on photoreceptor disc shedding and retinomotor movement. Journal of General Physiology 79, 775-790. BESHARSE, J . C , IUVONE, P.M. & PIERCE, M.E. (1988). Regulation of
rhythmic photoreceptor metabolism: a role for post-receptoral neurons. Progress in Retinal Research 7, 21-61. BOATRIOHT, J.H., HOEL, M.J. & IUVONE, P.M. (1989). Stimulation of
endogenous dopamine release and metabolism in amphibian retina by light- and K + -evoked depolarization. Brain Research 482, 164-168. BURNSIDE, B. & DEARRY, A. (1986). Cell motility in the retina. In The Retina: A Model for Cell Biological Studies, Part I, ed. ADLER, D. & FARBER, D., pp. 151-206. New York: Academic Press. DEARRY, A. & BURNSIDE, B. (1986). Dopaminergic regulation of cone retinomotor movement in isolated teleost retinas: I. Induction of cone contraction is mediated by D2 receptors. Journal of Neurochemistry 46, 1006-1021. DEARRY, A. & BURNSIDE, B. (1989a). Regulation of cell motility in teleost retinal photoreceptors and pigment epithelium by dopaminergic D2 receptors. In Extracellular and Intracellular Messengers, ed.
of retinal screening pigment migration in the frog. Vision Research 9, 377-384. MURESAN, Z. & BESHARSE, J.C. (1991). Identification of dopamine D2 receptors in the Xenopus retina by fluorescence microscopy. Investigative Ophthalmology and Visual Science (Abstract Suppl.) 32, 1260. PIERCE, M.E. & BESHARSE, J.C. (1985). Circadian regulation of retinomotor movements 1. Interaction of melatonin and dopamine in the control of cone length. Journal of General Physiology 86, 671-689. PIERCE, M.E. & BESHARSE, J.C. (1986). Melatonin and dopamine interactions in the regulation of rhythmic photoreceptor metabolism. In Pineal and Retinal Relationships, ed. O'BRIEN, P.J. & KLEIN, D.C., pp. 219-237. New York: Academic Press. PIERCE, M.E. & BESHARSE, J.C. (1987). Melatonin and rhythmic photoreceptor metabolism: melatonin-induced cone elongation is blocked at high light intensity. Brain Research 405, 400-404. PIERCE, M.E. & BESHARSE, J.C. (1988). Circadian regulation of retinomotor movements: II. The role of GABA in the regulation of cone position. Journal of Comparative Neurology 270, 279-287. PORRELLO, K. & BURNSIDE, B. (1984). Regulation of reactivated contraction in teleost retinal cone models by calcium and cyclic adenosine monophosphate. Journal of Cell Biology 98, 2230-2238. ROHLICH, P., SZEL, A. & PAPERMASTER, D.S. (1989). Immunocytochem-
ical reactivity of Xenopus laevis retinal rods and cones with several monoclonal antibodies to visual pigments. Journal of Comparative Neurology 290, 105-117. SCHUTTE, M. & WITKOVSKY, P. (1991). Dopaminergic interplexiform cells and centrifugal fibres in the Xenopus retina. Journal ofNeurocylology 20, 195-207. SLAUGHTER, M.M. & MILLER, R.F. (1981). 2-Amino-4-phosphonobu-
tyric acid: a new pharmacological tool for retina research. Science 211, 182-184. WITKOVSKY, P. & SHI, X.-P. (1990). Slow light and dark adaptation of horizontal cells in the Xenopus retina: a role for endogenous dopamine. Visual Neuroscience 5, 405-413. WITKOVSKY, P. & DEARRY, A. (1991). Functional roles of dopamine in the vertebrate retina. Progress in Retinal Research (in press).
REDBURN, D. & PASANTES-MORALES, H., pp. 229-256. New York:
WITKOVSKY, P., LEVINE, J.S., ENGBRETSON, G.A., HASSIN, G. & MAC-
Alan R. Liss, Inc. DEARRY, A. & BURNSIDE, B. (19896). Light-induced dopamine release from teleost retinas acts as a light-adaptive signal to the retinal pigment epithelium. Journal of Neurochemistry 53, 870-878.
NICHOL, E.F. (1981o). A microspectrophotometric study of normal and artificial visual pigments in the photoreceptors of Xenopus laevis. Vision Research 21, 867-873.
DEARRY, A., EDELMAN, J.L., MILLER, S. & BURNSIDE, B. (1990). Do-
pamine induces light-adaptive retinomotor movements in bullfrog cones via D2 receptors and in retinal pigment epithelium via Dl receptors. Journal of Neurochemistry 54, 1367-1378. DELEAN, A., MUNSON, P.J. & RODBARD, D. (1978). Simultaneous anal-
ysis of families of sigmoid curves: application to bioassay, radioligand assay, and physiological dose-response curves. American Journal of Physiology 235, E97-E102. DONG, C.-J. & MCREYNOLDS, J.S. (1991). The relationship between light, dopamine release and horizontal cell coupling in the mudpuppy retina. Journal of Physiology 440, 291-309.
WITKOVSKY, P., YANG, E.-Y. & RIPPS, H. (19816). Properties of the blue
sensitive rod in the Xenopus retina. Vision Research 21, 875-883. WITKOVSKY, P., STONE, S. & BESHARSE, J.C. (1988). Dopamine modi-
fies the balance of rod and cone inputs to horizontal cells of the Xenopus retina. Brain Research 449, 332-336. WITKOVSKY, P., STONE, S. & TRANCHINA, D. (1989). Photoreceptor to
horizontal cell synaptic transfer in the Xenopus retina: modulation by dopamine ligands and a circuit model for interactions of rod and cone inputs. Journal of Neurophysiology 62, 864-881. WITKOVSKY, P., RICE, M. & NICHOLSON, C. (1991). High extracellular
levels of dopamine in Xenopus retina detected by high speed cyclic voltametry. Society of Neuroscience Abstracts 624, 1565.
