The actions of gamma-aminobutyric acid, glycine and their antagonists upon horizontal cells of the Xenopus retina

Characterization of glycinergic synapses in vertebrate retinas

Glycine is one of the essential neurotransmitters modulating visual signals in retina. Glycine activates Cl(-) permeable receptors that conduct either inhibitory or excitatory actions, depending on the Cl(-) electrical-chemical gradient (E (Cl)) positive or negative to the resting potential in the cells. Interestingly, both glycine-induced inhibitory and excitatory responses are present in adult retinas, and the effects are confined in the inner and outer retinal neurons. Glycine inhibits glutamate synapses in the inner plexiform layer (IPL), resulting in shaping light responses in ganglion cells. In contrast, glycine excites horizontal cells and On-bipolar dendrites in the outer plexiform layer (OPL). The function of glycinergic synapse in the outer retina represents the effect of network feedback from a group of centrifugal neurons, glycinergic interplexiform cells. Moreover, immunocytochemical studies identify glycine receptor subunits (alpha1, alpha2, alpha3 and beta) in retinas, forming picrotoxin-sensitive alpha-homomeric and picrotoxin-insensitive alpha/beta-heteromeric receptors. Glycine receptors are modulated by intracellular Ca(2+) and protein kinas C and A pathways. Extracellular Zn(2+) regulates glycine receptors in a concentration-dependent manner, nanomolar Zn(2+) enhancing glycine responses, and micromolar Zn(2+) suppressing glycine responses in retinal neurons. These studies describe the function and mechanism of glycinergic synapses in retinas.

Amino acids and their transporters in the retina

This review provides an overview of the distributions, properties and roles of amino acid transport systems in normal and pathological retinal tissues and discusses the roles of specific identified transporters in the mammalian retina. The retina is used in this context as a vehicle for describing neuronal and glial properties, which are in some, but not all cases comparable to those found elsewhere an the brain. Where significant departures are noted, these are discussed in the context of functional specialisations of the retina and its relationship to adjacent supporting tissues such as the retinal pigment epithelium. Specific examples are given where immunocytochemical labelling for amino acid transporters may yield inaccurate results, possibly because of activity-dependent conformation changes of epitopes in these proteins which render the epitopes more or less accessible to antibodies.

Long-term light history modulates the light response kinetics of luminosity (L)-type horizontal cells in the roach retina

We have examined the effects of prolonged periods of darkness on the responses of luminosity-type horizontal cells (L-HCs) in the freshwater cyprinid, Rutilus rutilus. Two groups of retinae were compared, those recorded after 10 min dark adaptation (SA) and those recorded after 3 h dark adaptation (LA). The results suggest that long-term light history does not modify the general responsiveness of the L-HCs in this species. However, there are apparent changes in the receptive field of the cells and modifications to the kinetics of the light-evoked response. The kinetics changes involve both a delay in the onset of light response and a selective effect on the hyperpolarizing light-ON response. Thus the mean time constant (tau) for the SA cells was 32.4+/-2.39 ms (n=62), whilst that for the LA cells was 53.4+/-3.03 ms (n=61). These effects occur in the absence of changes in the relative spectral sensitivity or threshold sensitivity of the HCs. The results suggest that in some vertebrate retinae, prolonged darkness (light-history) may regulate long-term plasticity in the kinetics of the cone-HC pathway.

Photoreceptor classes and transmission at the photoreceptor synapse in the retina of the clawed frog, Xenopus laevis

The photoreceptor population in Xenopus consists of a green-sensitive rod (lambda(max) = 523 nm), a blue-sensitive rod (lambda(max) = 445 nm) and three classes of cone. The largest cone is red-sensitive (lambda(max) = 611 nm). The intermediate cone is presumed to be blue-sensitive based on physiological criteria, whereas the miniature cone may be UV-sensitive. Horizontal cells (HC) are of two sorts: axon-bearing and axonless. The axon-bearing HC is of the luminosity type and probably contacts all types of photoreceptor. The axonless HC is of the chromaticity type and contacts only intermediate (blue) cones and at least one type of rod. During development dendrites of HCs and bipolar neurons penetrate photoreceptor bases. A progressive maturation of HC and bipolar synapses with rods and cones occurs between tadpoles stages 37/8 and 46. Neighboring rods and cones are joined by gap junctions. During this same period, the outer segments are laid down and photopigments synthesized. A linear relation was found between the quantum capturing ability of the rod and its absolute threshold. Mature rods of the Xenopus retina release glutamate in a calcium-dependent manner. Glutamate release was found to be a linear function of calcium influx through L-type calcium channels. Both types of HC possess ionotropic glutamate receptors of the AMPA subtype.

Mesopic state: cellular mechanisms involved in pre- and post-synaptic mixing of rod and cone signals

At least twice daily our retinas move between a light adapted, cone-dominated (photopic) state and a dark-adapted, color-blind and highly light-sensitive rod-dominated (scotopic) state. In between is a rather ill-defined transitional state called the mesopic state in which retinal circuits express both rod and cone signals. The mesopic state is characterized by its dynamic and fluid nature: the rod and cone signals flowing through retinal networks are continually changing. Consequently, in the mesopic state the retinal output to the brain contained in the firing patterns of the ganglion cells consists of information derived from both rod and cone signals. Morphology, physiology, and psychophysics all contributed to an understanding that the two systems are not independent but interact extensively via both pooling and mutual inhibition. This review lays down a rationale for such rod-cone interactions in the vertebrate retinas. It suggests that the important functional role of rod-cone interactions is that they shorten the duration of the mesopic state. As a result, the retina is maintained in either in the (rod-dominated) high sensitivity photon counting mode or in the second mode, which emphasizes temporal transients and spatial resolution (the cone-dominated photopic state). Experimental evidence for pre- and postsynaptic mixing of rod and cone signals in the retina of the clawed frog, Xenopus, is shown together with the preeminent neuromodulatory role of both light and dopamine in controlling interactions between rod and cone signals. Dopamine is shown to be both necessary and sufficient to mediate light adaptation in the amphibian retina.

Neuromodulation of ligand- and voltage-gated channels in the amphibian retina

To understand information processing in the retina, it is important to identify and characterize the types of synaptic receptors and intrinsic ion channels in retinal neurons. In order to achieve a high degree of adaptability, retinal synapses have evolved multiple neuromodulatory mechanisms. Light or modulatory agents can alter the efficacies of both electrical and chemical synaptic transmission in the retina. Recent studies indicate that interaction of voltage-gated channels with those activated by neurotransmitters plays a significant role in shaping the light-evoked postsynaptic responses of retinal neurons. The fact that both types of channels are subject to modulation by multiple second messenger-mediated intracellular processes is a clear indicator of the importance of neuromodulation in retinal function. The whole-cell patch clamp technique provides a means to study mechanisms of regulation of ion channels by controlling intracellular as well as the extracellular environment. This review describes the experimental evidence, mostly obtained in our laboratory, which indicates the important role of Ca-dependent neuromodulatory processes in the regulation of signal transmission in the vertical pathway of the amphibian retina.

Modification of the Xenopus electroretinogram by actions of glycine in the proximal retina

The electroretinogram (ERG) was recorded from the Xenopus retina, to examine the effects of glycine and strychnine on these responses and to determine the origins of these changes. Glycine at concentrations between 0.1 and 10 mM reduced the b- and d-waves of the ERG in a dose-dependent manner, while strychnine increased their amplitude. 2-Amino-4-phosphonobutyric acid (APB) reduced the b-wave and blocked the effect of glycine, but not strychnine, on the d-wave. When the d-wave had first been blocked by kynurenic acid (KYN) or reduced by (+/-)cis-2,3-piperidine dicarboxylic acid (PDA) the b-wave was enhanced by glycine, but not by strychnine. N-methyl-DL-aspartate (NMDLA), which alters responses in the proximal retina only, blocked the effects of glycine and strychnine on the ERG. This suggests that the glycinergic effects on the ERG are at least partly mediated by processes in the proximal retina. The results further support the suggestion that inhibitory neurotransmitters in the proximal retina may modulate both the b- and d-waves of the Xenopus ERG.

The role of GABA in modulating the Xenopus electroretinogram

We have recorded the electroretinogram (ERG) from the superfused eyecup of the Xenopus retina in order to assess the effects of the inhibitory neurotransmitter gamma-aminobutyric acid (GABA), and its agonists and antagonists, on individual ERG components. We found that GABA (0.5-10 mM) reduced the amplitudes of both the b- and d-waves of the Xenopus ERG. The GABA uptake blocker nipecotic acid (1 mM) had similar effects on b- and d-waves. GABA at 5 mM and 10 mM also caused an increase in the a-wave. The GABA antagonist picrotoxin (0.1-2 mM) and the GABA/a antagonist bicuculline (0.2 mM) both increased the amplitude of the b- and d-waves of the ERG. The GABA/b agonist baclofen (0.3 mM) reduced the amplitude of the ERG b-wave, enhanced the amplitude of the a-wave, and slightly reduced the amplitude and increased the peak time of the d-wave. The GABA/b antagonists phaclofen and saclofen had no reliable effects on the Xenopus ERG. Glutamate analogs known to affect specific types of retinal neurons were applied to modify the retinal circuitry and then the effects of GABA and its antagonists were examined under these modified conditions. 2-amino-4-phosphonobutyric acid (APB) increased the d-wave, and blocked the b-wave and the effect of GABA on the ERG, but not the antagonist-induced increase in the d-wave. KYN blocked the antagonist-induced increase in the b-wave, while GABA increases the amplitude of the b-wave if the d-wave has been removed by prior superfusion with kynurenic acid (KYN). N-methyl-DL-aspartate (NMDLA), which acts only in the proximal retina, reduced the amplitude of the ERG and blocked the effect of GABA and the antagonist-induced increase in ERG b- and d-waves amplitude. These results suggest that GABAergic mechanisms related to both A and B receptor types can influence the amplitude and light sensitivity of all the components of the Xenopus ERG. Since GABA is found in greatest abundance in the proximal retina, and B type of receptors are present almost exclusively there, the data suggests that most of the effects of GABA agonists and antagonists observed are dependent on proximal retinal mechanisms, and that there are separate mechanisms in the proximal retina related to the b- and the d-waves.

Gain of rod to horizontal cell synaptic transfer: relation to glutamate release and a dihydropyridine-sensitive calcium current

We related rod to horizontal cell synaptic transfer to glutamate release by rods. Simultaneous intracellular records were obtained from dark-adapted rod-horizontal cell pairs. Steady-state synaptic gain (defined as the ratio of horizontal cell voltage to rod voltage evoked by the same light stimulus) was 3.35 +/- 0.60 for dim flashes and 1.50 +/- 0.03 for bright flashes. Under conditions of maintained illumination, there was a measurable increment of horizontal cell hyperpolarization for each light-induced increment of rod hyperpolarization over the full range of rod voltages. In separate experiments we studied glutamate release from an intact, light-responsive photoreceptor layer, from which inner retinal layers were removed. Steady light reduced glutamate release as a monotonic function of intensity; spectral sensitivity measures indicated that we monitored glutamate release from rods. The dependence of glutamate release on rod voltage was well fit by the activation function for a high-voltage-activated, dihydropyridine-sensitive L-type calcium current, suggesting a linear dependence of glutamate release on [Ca]i in the synaptic terminal. A simple model incorporating this assumption accounts for the steady-state gain of the rod to horizontal cell synapse.

Gain of rod to horizontal cell synaptic transfer: relation to glutamate release and a dihydropyridine-sensitive calcium current

We related rod to horizontal cell synaptic transfer to glutamate release by rods. Simultaneous intracellular records were obtained from dark-adapted rod-horizontal cell pairs. Steady-state synaptic gain (defined as the ratio of horizontal cell voltage to rod voltage evoked by the same light stimulus) was 3.35 +/- 0.60 for dim flashes and 1.50 +/- 0.03 for bright flashes. Under conditions of maintained illumination, there was a measurable increment of horizontal cell hyperpolarization for each light-induced increment of rod hyperpolarization over the full range of rod voltages. In separate experiments we studied glutamate release from an intact, light-responsive photoreceptor layer, from which inner retinal layers were removed. Steady light reduced glutamate release as a monotonic function of intensity; spectral sensitivity measures indicated that we monitored glutamate release from rods. The dependence of glutamate release on rod voltage was well fit by the activation function for a high-voltage-activated, dihydropyridine-sensitive L-type calcium current, suggesting a linear dependence of glutamate release on [Ca]i in the synaptic terminal. A simple model incorporating this assumption accounts for the steady-state gain of the rod to horizontal cell synapse.

Both high- and low voltage-activated calcium currents contribute to the light-evoked responses of luminosity horizontal cells in the Xenopus retina

We examined the contribution of two intrinsic voltage-dependent calcium channels to the light-evoked responses of a non-spiking retinal neuron, the horizontal cell (HC). HC's isolated from the Xenopus retina were studied by the whole cell version of the patch clamp. In a mixture of agents which suppressed Na- and K-dependent currents, we identified a transient, low voltage-activated Ca current suppressed by Ba2+ and blocked by Ni2+ (T-type) and a sustained, high voltage-activated, dihydropyridine-sensitive Ca current that was enhanced by Ba2+ (L-type). We made simultaneous intracellular recordings from rods and HC's in the intact, dark-adapted Xenopus retina. Under certain stimulus conditions, transient oscillations appeared in HC responses but were absent in rod light-evoked waveforms. One type of transient was seen at relatively hyperpolarized potentials (< -45 mV), was enhanced by Sr2+ and inhibited by Ni2+. It thus appears to depend on a T-type Ca-current. A second type of oscillation was seen to be superimposed on a prolonged depolarizing wave following light off in the HC and as spike-like depolarizations in rods. These oscillations were enhanced by Ba2+ and Sr2+, but blocked by the dihydropyridine, nifedipine, indicating their dependence on an L-type calcium conductance. All calcium-dependent oscillations were suppressed by 0.05-0.5 mM Co2+. Suppression of glutamate neurotransmission with CNQX or kynurenate, or glycine neurotransmission with strychnine, enhanced the HC oscillations.

Functional role of GABA in cat retina: II. Effects of GABAA antagonists

Putative GABAergic mechanisms were studied in the cat retina by exogenous application of the GABAA antagonists picrotoxin (PTX), native bicuculline (BCC), and bicuculline methyl bromide (BCC MeBr). When recording intracellular responses from horizontal cells (HCs) and amacrine cells as well as electroretinograms (ERGs), drugs were added to the perfusate used to maintain the isolated eyecup; when recording extracellular spikes from ganglion cells of anesthetized cats, drugs were introduced by iontophoretic injection. Both PTX and BCC MeBr had relatively little influence upon the response properties of HCs. In contrast, native BCC tended to decrease the amplitude of and to slow the photic response to light onset and both to quicken and to increase the amplitude of response to light offset; in the presence of native BCC, HC responses were dominated by a prominent spike-like "Off-overshoot." The influence of GABAA agonists upon HC responses was not blocked by GABAA antagonists. ERG b-wave amplitude was reduced both by PTX and by native BCC, but was not influenced by BCC MeBr. Latency (time to half-peak) was increased by low doses of native BCC, and to a lesser extent PTX but not BCC MeBr. Rod-amacrine On-transient responses were increased in amplitude by PTX. Extracellular recordings from On- and Off- X and Y ganglion cell types became considerably more transient with application of either PTX, native BCC, or BCC MeBr; this tendency was greater in Off-type ganglion cells. Collectively, these results strengthen conclusions from the previous paper suggesting that GABA serves to slow onset and offset kinetics of retinal neurons, making them more sustained and less phasic. They also suggest that in mammalian retina heterogeneous types of GABAA receptors exist, segregated into different zones: a distal zone, sensitive only to native BCC, a central zone sensitive to both native BCC and PTX, and a proximal zone sensitive to native BCC, BCC methyl halides (BCC MeH), and PTX. Only the proximal zone obeys conventional GABAA pharmacology.

Functional role of GABA in cat retina: I. Effects of GABAA agonists

Putative GABAergic mechanisms were studied in perfused cat retina by means of intracellular recording and application of GABA and the GABAA agonists delta-amino valeric acid (dAVA), muscimol, and THIP. In contrast to results reported previously for cold-blooded vertebrates, introduction of 20 mM GABA into the superfusate had no influence upon the response properties of cat retinal horizontal cells (HCs). In common with results reported in cold-blooded vertebrates, introduction of the GABAA agonists dAVA (2-12 mM) and THIP or muscimol (0.2-1 mM) had four consistent reversible influences upon the response properties of cat retinal HCs: (1) they reduced photic-response amplitude, (2) slowed response onset, (3) slowed response offset, and (4) depolarized the dark membrane potential. Both rod and cone signal components were affected. GABAA agonists had similar influences upon both the time course and amplitude of responses recorded from amacrine and ganglion cells. In all cell types examined, the influence upon response kinetics was made particularly apparent with rapidly flickering stimuli. Flicker responses were reduced in amplitude much more than sustained responses. These results suggest that, in addition to other influences, GABAergic action serves to modify the time course of photic responses in both the inner and outer plexiform layer of mammalian retina making responses slower and less phasic.

Feedback from luminosity horizontal cells mediates depolarizing responses of chromaticity horizontal cells in the Xenopus retina

It has been proposed that the depolarizing responses of chromaticity horizontal cells (C-HCs) to red light depend on a feedback signal from luminosity horizontal cells (L-HCs) to short-wavelength-sensitive cones in the retinas of lower vertebrates. In this regard we studied the C-HCs of the Xenopus retina. C-HCs and L-HCs were identified by physiological criteria and then injected with neurobiotin. The retina then was incubated with peanut agglutinin, which stains red-but not blue-sensitive cones. Electron microscopic examination revealed that L-HCs contact all cone classes, whereas C-HCs contact only blue-sensitive cones. Simultaneous recordings from C-HC/L-HC pairs established that when the L-HC was saturated by a steady bright red light, C-HCs alone responded to a superimposed blue stimulus. In response to red test flashes, the C-HC response was delayed by approximately 30 msec with respect to the L-HC response. Isolated HCs of both subtypes were examined by whole-cell patch clamp. Both responded to kainate with sustained inward currents and to quisqualate or alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) with desensitizing currents from a negative holding potential; i.e., both have AMPA-type glutamate receptors. gamma-Aminobutyric acid or glycine opened a chloride channel in the L-HC, whereas the C-HC was unresponsive to either inhibitory amino acid. Since glycine has been shown to abolish selectively the depolarizing response of the C-HC, this finding and other pharmacological data strongly implicate the L-HC in the underlying circuit. Moreover, because the C-HC does not respond to gamma-aminobutyric acid, the neurotransmitter of the L-HC, by elimination, a feedback synapse from L-HC to blue cone is the most plausible mechanism for the creation of depolarizing responses in C-HCs.

Early development of GABA-like immunoreactive cells in the retina of turtle embryos

Gamma aminobutyric acid (GABA) is one of the earliest neuroactive substances appearing in the developing central nervous system. The distribution and the time course of the appearance of GABA-like immunoreactivity in the retina of the turtle Emys orbicularis were investigated from embryonic stage 13 to hatching. The first GABA-like immunoreactive cells were observed at stage 14. These cells were located in both the scleral third of the neuroblastic layer and the inner layers of the retina. They were identified as presumptive immature horizontal cells and amacrine cells, respectively. The observation of numerous labelled fibers in the nerve fiber layer suggests that some of the GABA-like immunoreactive cells in the layers were ganglion cells. The development of GABA-like immunoreactive cells followed a gradient of maturation from central to peripheral retina. At hatching, the central retina appeared nearly morphologically mature. In conclusion, GABA is present before the morphofunctional maturation of the retina and this precocious existence supports the idea of its involvement in a neurotrophic role preceding the establishment of synaptic connections and neurotransmitter function.

Regulation of endogenous dopamine release in amphibian retina by gamma-aminobutyric acid and glycine

Endogenous dopamine release in the retina of the African clawed frog (Xenopus laevis) increases in light and decreases in darkness. The roles of the inhibitory amino acid transmitters gamma-aminobutyric acid (GABA) and glycine in regulating this light/dark difference in dopamine release were explored in the present study. Exogenous GABA, the GABA-A receptor agonist muscimol, the GABA-B receptor agonist baclofen, and the GABA-C receptor agonist cis-aminocrotonic acid (CACA) suppressed light-evoked dopamine overflow from eyecups. The effects of GABA-A and -B receptor agonists were selectively reversed by their respective receptor-specific antagonists, whereas the effect of CACA was reversed by the competitive GABA-A receptor antagonist bicuculline. The benzodiazepine diazepam enhanced the effect of muscimol on light-evoked dopamine release. Both GABA-A and -B receptor antagonists stimulated dopamine release in light or darkness. Bicuculline was more potent in light than in darkness. These data suggest that retinal dopaminergic neurons are inhibited by GABA-A and -B receptor activation in both light and darkness but that GABA-mediated inhibitory tone may be greater in darkness than in light. Exogenous glycine inhibited light-stimulated dopamine release in a concentration-dependent and strychnine-sensitive manner. However, strychnine alone did not increase dopamine release in light or darkness, nor did it augment bicuculline-stimulated release in darkness. Additionally, both strychnine and 7-chlorokynurenate, an antagonist of the strychnine-insensitive glycine-binding site of the N-methyl-D-aspartate subtype of glutamate receptor, suppressed light-evoked dopamine release. Thus, the role of endogenous glycine in the regulation of dopamine release remains unclear.

White noise analysis of a chromatic type horizontal cell in the Xenopus retina

The dynamics of color-coded signal transmission in the light-adapted Xenopus retina were studied by a combination of white noise (Wiener) analysis and simultaneous recordings from two types of horizontal cells: chromatic-type horizontal cells (C-HCs) are hyperpolarized by blue light and depolarized by red light, whereas luminosity-type horizontal cells (L-HCs) are hyperpolarized by all wave-lengths. The retina was stimulated by two superimposed fields of red and blue light modulated by two independent white noise signals, and the resulting intracellular responses were decomposed into red and blue components (first-order kernels). The first-order kernels predict the intracellular responses with a small degree of error (3.5-9.5% in terms of mean square error) under conditions where modulated responses exceeded 30 mV in amplitude peak-to-peak, thus demonstrating that both red and blue modulation responses are linear. Moreover, there is little or no interaction between the red- and blue-evoked responses; i.e., nearly identical first-order kernels were obtained for one color whether the other color was modulated or not. In C-HCs (but not L-HCs), there were consistent differences in the dynamics of the red and blue responses. In the C-HC, the cutoff frequency of the red response was higher than for the blue (approximately 12 vs 5 Hz), and the red kernel was more bandpass than the blue. In the L-HC, kernel waveform and cutoff frequencies were similar for both colors (approximately 12 Hz or greater), and the time-to-peak of the L-HC kernel was always shorter than either the red or blue C-HC kernel. These results have implications for the mechanisms underlying color coding in the distal retina, and they further suggest that nonlinear phenomena, such as voltage-dependent conductances in HCs, do not contribute to the generation of modulation responses under the experimental conditions used here.

Effects of submicromolar concentrations of dopamine on photoreceptor to horizontal cell communication

Dopamine has been postulated to act as an intraretinal messenger for light adaptation by biasing retinal circuits to favor cone over rod inputs to second- and third-order neurons. As an experimental test, we studied the effects of dopamine and related ligands on rod to horizontal cell synaptic transfer. Rods and horizontal cells (HC) were recorded from simultaneously. Red and green light-emitting diodes were modulated sinusoidally in counterphase at 1 Hz and their relative intensities adjusted to elicit a rod null. Dark-adapted HC's also showed a null, but exposure to 0.5-1.0 microM dopamine, which corresponds to the endogenous levels, elicited a large imbalance in the HC response while the rod null was maintained. Similar effects were achieved with either a D1 dopamine agonist, SKF 38393 (20 microM) or a D2 dopamine agonist, quinpirole HCl (1 microM). Correspondingly, the frequency range over which the HC responded to red flickering lights increased very substantially. Exposure to a D2, but not a D1 dopamine agonist increased rod flicker, but not as dramatically as in the HC. These data indicate that the synaptic gains of rod and cone transmission to a second order retinal neuron are modifiable by endogenous levels of dopamine. Secondly, the bandpass of rod flicker is altered by dopamine, acting through a D2 dopamine receptor.

Synaptic feedback, depolarization, and color opponency in cone photoreceptors

For some 20 years, synaptic feedback from horizontal cells to cones has often been invoked, more or less convincingly, in discussions of retinal action and vision. However, feedback in cones has proved to be rather complex and difficult to study experimentally. The mechanisms and consequences of feedback are therefore still only partly understood. This review attempts to assess the knowns and unknowns. The limitations of the evidence for feedback are reviewed to support the position that unequivocal evidence still largely rests on intracellular recording from cones. Of the three distinct types of depolarization observed in cones, the graded depolarization is taken as the fundamental manifestation of feedback. The evidence for the hypothesis that GABA is the neurotransmitter for feedback appears reasonably strong but several complications will have to be resolved to make the hypothesis more secure. There is evidence that feedback contributes to aspects of light adaptation and spatiotemporal processing of visual information. The contributions seem modest in magnitude. The role of feedback in shaping the color-opponent responses of retinal neurons is evaluated with particular emphasis on pharmacological studies, spatial and temporal aspects of the response of chromatic horizontal cells, and the enigmatic nature of depolarizations in blue- and green-sensitive cones. On this and other evidence, it is suggested that feedback may impress some detectable wavelength dependency in some cones but the dominant mechanisms for color opponency probably reside beyond the photoreceptors.

Extracellular dopamine concentration in the retina of the clawed frog, Xenopus laevis

Dopamine reaches targets in the outer retina of the clawed frog (Xenopus laevis) by diffusion from a network of dopaminergic cells and processes located predominantly at the junction of inner nuclear and inner plexiform layers. We obtained values for the steady-state release, uptake, and extracellular concentration of dopamine in the retina by a combination of HPLC (with electrochemical detection), scintillation spectroscopy, and fast-scan cyclic voltammetry. Vitreal concentrations of dopamine varied from 564 +/- 109 nM in light-adapted eyes near the time of subjective dawn to 156 +/- 12 nM in dark-adapted eyes. The data are consistent with a simple model for steady-state dopamine diffusion from an appropriately sited thin-sheet source. This model was used to generate a profile of extracellular dopamine concentration as a function of retinal depth. The model predicted an increase in the dopamine concentration from the vitreous to the layer of dopaminergic cells, remaining constant from that layer to the distal tips of the photoreceptors. This prediction was borne out by comparing fast-scan voltammetric measures of dopamine at the distal tips of the receptors with the vitreal concentrations determined by HPLC using electrochemical detection.

Influence of rod adaptation upon cone responses to light offset in humans: I. Results in normal observers

Dark-adapted rods exert a tonic suppressive influence upon cone-mediated sensitivity to rapid flicker, a phenomenon called suppressive rod-cone interaction (SRCI). However, rod dark adaptation has negligible influence upon cone-mediated thresholds measured with more usual psychophysical procedures. The present study separately examined the influences of rod light and dark adaptation upon cone-mediated sensitivity to transient increases or decreases in illumination using sawtooth flicker with rapid-on (ramp-off) or rapid-off (ramp-on) waveforms. In the parafoveal retina, cones alone were stimulated with flicker by spatially superimposing long- and short-wavelength stimuli presented in counterphase and matched in scotopic illuminance. Several different adaptation procedures were used. For higher (greater than 4 Hz) frequencies, sensitivity of cones to both waveforms is nearly identical under any condition of adaptation; sensitivity decreases as rods progressively dark adapt. A considerably different situation exists for slower frequencies (1-4 Hz). Sensitivity of cones to rapid-off flicker is appreciably greater under light-adapted conditions confirming recent observations by Bowen et al. (1989). But as rods progressively dark adapt, sensitivity of cones to rapid-off waveforms decreases considerably while sensitivity to rapid-on waveforms is much less affected; in the totally dark-adapted eye, sensitivity to both waveforms is identical. These results confirm and extend recent physiological observations in amphibian retina (Frumkes & Wu, 1990) suggesting that SRCI specifically involves responses to transient decreases in illumination.

Physiological and morphological properties of off- and on-center bipolar cells in the Xenopus retina: effects of glycine and GABA

We studied the morphology and center-surround organization of Lucifer Yellow injected OFF- and ON-center bipolar cells in the light-adapted Xenopus retina and the effects of glycine and GABA on their cone-mediated light responses. In both classes of cell, prominent antagonistic surround responses up to 20 mV in amplitude could be evoked without first suppressing the center responses with steady illumination. An additional feature of the light-evoked bipolar cell response was a pronounced (up to -24 mV) delayed hyperpolarizing after potential (DHAP) which followed the depolarizing responses of both classes of bipolar cell. The morphological features of dye-injected bipolar cells conformed to the general idea of segregation of ON and OFF pathways in the inner and outer interplexiform layer, however, the morphology of axonal arborizations was different for both classes. OFF-center cells ramified symmetrically around the primary branchpoint, whereas ON-center cells had a strongly asymmetrical arrangement of their axonal tree. The center and surround responses were differentially sensitive to glycine and GABA. Glycine eliminated the antagonistic surround responses in both OFF and ON cells; the center responses were reduced to some extent but were not eliminated. In contrast, GABA affected the hyperpolarizing responses much more strongly than the depolarizing response components. That is, the amplitude of the center response in the OFF cell and the surround response in the ON cell was reduced 80-90% during exposure to GABA, whereas the surround and center depolarizations of OFF and ON cells, respectively, were reduced only 0-10%. Our findings implicate a role for GABAergic and glycinergic pathways in the center-surround organization of bipolar cells in Xenopus retina. In addition, the results suggest that the pathways mediating center-surround antagonism may be different in OFF-bipolar cells vs. ON-bipolar cells.

The organization of dopaminergic neurons in vertebrate retinas

A survey of the shapes of dopaminergic (DA) neurons in the retinas of representative vertebrates reveals that they are divisible into three groups. In teleosts and Cebus monkey, DA cells are interplexiform (IPC) neurons with an ascending process that ramifies to create an extensive arbor in the outer plexiform layer (OPL). All other vertebrates studied, including several primate species, have either DA amacrine cells or IPCs with an ascending process that either does not branch within the OPL or does so to a very limited degree. DA neurons of non-teleosts exhibit a dense plexus of fine caliber fibers which extends in the distal most sublamina of the inner plexiform layer (IPL). Teleosts lack this plexus. In all vertebrates, DA cells are distributed more or less evenly and at a low density (10-60 cells/mm2) over the retinal surface. Dendritic fields of adjacent DA neurons overlap. Most of the membrane area of the DA cell is contained within the plexus of fine fibers, which we postulate to be the major source of dopamine release. Thus, dopamine release can be modeled as occurring uniformly from a thin sheet located either in the OPL (teleosts) or in the distal IPL (most other vertebrates) or both (Cebus monkey). Assuming that net lateral spread of dopamine is zero, the fall of dopamine concentration with distance at right angles to the sheet (i.e. in the scleral-vitreal axis) will be exponential. The factors that influence the rate of fall-diffusion in extracellular space, uptake, and transport--are not yet quantified for dopamine, hence the dopamine concentration around its target cells cannot yet be assessed. This point is important in relation to the thresholds for activation of D1 and D2 dopamine receptors that are found on a variety of retinal cells.

Dopaminergic interplexiform cells and centrifugal fibres in the Xenopus retina

Putative dopaminergic neurons in the Xenopus retina were identified using an immunoreaction against tyrosine hydroxylase. A single class of cell was stained whose perikaryon (12-15 microns in diameter) was located at the border of the inner nuclear and inner plexiform layers. About 2% of the stained cell bodies were located in the ganglion cell layer, but the distribution of the processes of displaced cells had the same geometry as for the majority of stained cells. Tyrosine hydroxylase-like immunoreactive perikarya gave rise to one to four stout processes that descended to the most proximal level of the inner plexiform layer, within which they branched repeatedly to generate a diffuse network of fine processes. Secondary branches ascended to the most distal sublayer of the inner plexiform layer where they ramified into fine processes that joined other fibres arising horizontally from the cell body and confined to the distal inner plexiform layer throughout their course. The diameter of the dendritic arbor of stained cells was in the range of 350-600 microns. The dense network of fine fibres within the distal inner plexiform layer was arrayed in rings that surrounded other amacrine cells; using an antiserum against glycine we found that at least some of these were glycinergic neurons. Most tyrosine hydroxylase-positive neurons emitted one or two fine ascending processes that arose from the perikaryon, traversed the inner plexiform layer and arborized within the outer plexiform layer. Additionally, fine varicose fibres arising from the sublayer 1 of the inner plexiform layer and running to the outer retina were observed. Thus, based on light microscopic criteria, dopaminergic neurons in the Xenopus retina appeared to be interplexiform cells. A few tyrosine hydroxylase-immunoreactive fibres were observed in the optic nerve, some of which entered the inner retina where they ramified, thus indicating that they were centrifugal axons. In addition, a small number of stout smooth processes were observed to traverse the entire inner nuclear layer and course laterally at the level of the photoreceptor bases. Whether this second class of ascending process arises from the tyrosine hydroxylase-like immunoreactive efferents remains to be determined. The total number of dopaminergic neurons per retina was 750-800, equivalent to an average density of 30 cells mm-2. The dendritic fields of adjacent cells strongly overlapped, with an estimated coverage factor of 4.8.

Slow light and dark adaptation of horizontal cells in the Xenopus retina: a role for endogenous dopamine

A role for endogenous dopamine in the control of rod and cone contributions to a second-order retinal neuron, the horizontal cell (HC) was studied in the Xenopus retina. Relative rod and cone contributions were estimated from HC responses to scotopically balanced 491- and 650-nm flashes. In eyecups prepared in light then placed in darkness, cone input to the HC slowed and diminished on a time scale of hours. The decline in cone input was balanced by a slow growth of rod input to the HC. Administration of D-amphetamine, a dopamine releasing agent, restored the light-adapted waveform. The kinetics of slow light adaptation were examined by recording HC responses from eyecups that had been dark-adapted previously for 11-14 h. When test flashes fell on a dark field, cone input to the HC grew for 2-4 h, reached a plateau, and later declined. If, however, flashes were superimposed on a weak background field, cone input to the HC continued to increase monotonically at about 10%/h. This increase was abolished by superfusion with a nonspecific dopamine receptor blocker, cis-flupenthixol (50 microM), resulting in the complete suppression of cone-to-horizontal cell synaptic transfer and the enhancement of rod-to-horizontal cell communication. Subcutaneous injection of reserpine, a drug that depletes dopamine stores (2 mg/kg on 1-4 successive days), or intraocular injection of the dopamine neurotoxin, 6-hydroxydopamine (10-30 micrograms) slowed and reduced the amplitude of cone input to the HC, even in completely light-adapted eyes.(ABSTRACT TRUNCATED AT 250 WORDS)

Glycinergic contacts in the outer plexiform layer of the Xenopus laevis retina characterized by antibodies to glycine, GABA and glycine receptors

Electrophysiological experiments have predicted a direct synaptic input from glycinergic interplexiform cells (IPCs) to GABAergic horizontal cells in the Xenopus retina. However, previous ultrastructural studies failed to demonstrate this input. Here, we used three immunocytochemical approaches to investigate this issue. First, double-label postembedding immunocytochemistry with GABA- and glycine-like immunoreactivity (GABA-LI and glycine-LI) was used to study possible interactions of the glycinergic IPC with GABAergic horizontal cells. Processes postsynaptic to glycine-LI IPC terminals in the outer plexiform layer (OPL) fell into two groups, small microtubule-filled processes and larger electron-lucent processes with sparse microtubules and occasional mitochondria. In no case did we find glycine-LI synapses onto GABA-LI cells or processes. Second, pre-embedding immunocytochemistry was used to label GABA-LI cells and processes in the OPL. GABA-LI was sparse in horizontal cell axons and more intense in horizontal cell somas and in small processes. In agreement with our first set of experiments, GABA-LI profiles did not receive input from conventional synapses. Third, we localized glycine-receptor-like immunoreactivity (GlyR-LI) to several types of apparent synapses in the OPL. As expected, it was found at IPC synapses. Unexpectedly, GlyR-LI was also subsynaptic at photoreceptor synapses onto second order neurons, both at ribbon and basal junction type synapses. At least some of the GlyR-LI photoreceptor synapses were from cones. Also, GlyR-LI was apposed to photoreceptors and to unidentified small diameter processes, where no other indication of synaptic input was evident. Because glycine-LI is not found in photoreceptors, we suggest that glycine receptors at photoreceptor synapses are stimulated by glycine that diffuses from other sites, possibly from IPCs. This interpretation is consistent with available physiological studies of glycinergic effects in this retina.

Glycine stimulates calcium-independent release of 3H-GABA from isolated retinas of Xenopus laevis

A perfusion system was used to monitor the release of [3H]-GABA from isolated retinas of Xenopus laevis. Measurable release was stimulated by glycine at concentrations as low as 200 microM. Glycine-stimulated release was blocked by strychnine, and was not reduced in "calcium-free" Ringer's solution (0 Ca2+/20 mM Mg2+). Glutamate also stimulated calcium-independent release, using concentrations as low as 100 microM. In contrast, release stimulated by 25 mM potassium was reduced by 80% in calcium-free medium. In most experiments, agonists were applied in six consecutive 4-min pulses separated by 10-min washes with Ringer's solution. Under these conditions, the release stimulated by 0.5 mM glutamate or 25 mM potassium decreased by at least 50% from the first to the second pulse, and then gradually decreased with successive applications. In contrast, the response to 0.5 mM glycine at first increased and then only gradually decreased with successive pulses. These patterns of response to different agonists were similar in calcium-free medium. Somatostatin (-14 or -28) also stimulated release, and this effect was inhibited by AOAA, an inhibitor of GABA degradation. In the presence of AOAA, somatostatin had little effect, except at high concentrations of somatostatin (5 microM), which increased both basal and glycine-stimulated release. In contrast to somatostatin, glycine-stimulated release was much larger in the presence of AOAA. Autoradiography was used to investigate which cell types released [3H]-GABA under our conditions. Autoradiograms showed that horizontal cells and a population of apparent "off" bipolar cells were well-labeled by [3H]-GABA high-affinity uptake. In addition, light labeling was seen over numerous amacrine cells. After application of glycine, glutamate, or potassium, there was a decrease in label density over horizontal cells.

Uptake and K+-stimulated release of [14C]glycine from frog retinal synaptosomal fractions

The uptake of [14C]glycine and the effect of depolarizing potassium concentrations on its release was investigated in the whole frog retina and its synaptosomal fractions. The uptake of [14C]glycine in retina and synaptosomal fractions was found to be saturable as well as energy and Na+-dependent. The Km value for glycine uptake was found to be 46 microM for P2 fraction and 100 microM for P1 fraction, with a Vmax of 3.5 and 3.8 nmol/mg protein/min respectively. The release of [14C]glycine from P1 and P2 synaptosomal fractions was markedly increased by raising potassium concentration in the medium, in a partially Ca2+-dependent manner. Evoked glycine release was 50% reduced when calcium was omitted from the medium. The K+-stimulated release of glycine from P2 fraction was significantly reduced in the presence of TTX. The cellular origin of the P1 and P2 synaptosomal fractions releasing glycine is discussed.

GABA as a potential transmitter in lizard photoreceptors: immunocytochemical and biochemical evidence

The retina of the desert night lizard, Xantusia vigilis, was examined for immunoreactivity to antibodies against gamma-aminobutyric acid and L-glutamate decarboxylase. At the electron microscopic level it was found that a distinct population of the photoreceptor cells was immunoreactive to both antibodies. Computer-assisted reconstruction of serial sections positively identified the immunoreactive receptors as cones. These cones constituted 15% of the photoreceptors in the retinal sections, and they were morphologically distinct. The mean diameter of the labeled cone synaptic pedicles was 5.8 micron whereas that of the unlabeled pedicles was 7.9 micron, a statistically significant difference. L-glutamate decarboxylase was extracted from the lizard brain, positively identified radiometrically, and shown by immunodiffusion to crossreact with the antibody used for localization. The authors suggest that the immunoreactive cones synthesize and accumulate gamma-aminobutyric acid. Whether or not it is used by those cones as a neurotransmitter should be tested directly.

Localization of putative GABAergic neurons in the larval tiger salamander retina by immunocytochemical and autoradiographic methods

Putative GABAergic neurons in the larval tiger salamander retina were localized by a comparative analysis of glutamate decarboxylase immunoreactivity (GAD-IR), GABA-like immunoreactivity (GABA-IR), and high-affinity 3H-GABA uptake at the light microscopical level. Preliminary data showed that all GAD-IR neurons were double labeled for GABA-IR. However, because the weak somatic labeling with GAD-IR, we could not determine if the converse were true. Neurons commonly labeled with GABA-IR and 3H-GABA uptake include horizontal cells, type I (outer) and type II (inner) bipolar cells, type I (inner) and type II (outer) amacrine cells, and cell bodies in the ganglion cell layer (GCL). In addition, interplexiform cells were identified with GABA-IR. The presence of GABA-IR ganglion cells was indicated by GABA-IR fibers in the optic fiber layer and optic nerve as well as by a GABA-IR cell in the GCL that included a labeled axon. The percentage of labeled somas in the inner nuclear layer (INL) compared to all cells in each layer was similar for the two methods: 30% in INL 1 (outer layer of somas), 15% in INL 2 (middle layer), 43-52% in INL 3 (inner layer), and about 21-26% in the GCL. Labeled processes were found in three bands in the inner plexiform layer, with the densest band located in the most proximal part. Postembedding labeling of 1-micron Durcupan resin sections for GABA-IR showed the same general pattern as obtained with 10-microns cryostat sections, with additional staining, however, of type II (inner) bipolar cell Landolt's clubs. Extensive colocalization of labeling was indicated, and we conclude that GABA-IR can serve as a valid and reliable marker for GABA-containing neurons in this retina and suggest that GABA serves as a transmitter for horizontal cells, several types of amacrine cell, a type of interplexiform cell, and perhaps a small percentage of type I and type II bipolar cells and ganglion cells.

Morphology and synaptic connections of HRP-filled, axon-bearing horizontal cells in the Xenopus retina

Axon-bearing horizontal cells of the Xenopus retina were studied by intracellular injection of HRP following physiological characterization. The profile of the cell viewed in whole mount consisted of a round or oval perikaryon about 50 microns in diameter and an axon about 1 mm long which lacked a prominent terminal expansion. The axonal diameter was 0.5-1.0 microns in its proximal third but 2-4 microns in its distal portion. Along its course the axon emitted 25-40 branchlets each 0.2 micron in diameter, up to 10 micron long and terminating in a cluster of two to six synaptic knobs. Electron microscopic examination revealed that both perikaryal dendrites and axon branchlets ended in both rod and cone synaptic bases; cone contacts outnumbered rod contacts by two- to threefold. We were unable to document synapses of presumed interplexiform cells onto identified horizontal cells. Horizontal cell axons are joined in their distal portions by numerous, small (0.2 micron long) gap junctions. Other gap junctions were noted between horizontal cell processes within the synaptic endings of photoreceptors. An hypothesis is advanced whereby the cluster of axon branchlet synaptic knobs permits dynamic interaction of rod and cone synaptic inputs to the horizontal cell.

Somatostatin-like immunoreactivity and glycine high-affinity uptake colocalize to an interplexiform cell of the Xenopus laevis retina

Antibodies directed against somatostatin have been used to label a population of interplexiform cells (IPCs) in the Xenopus laevis retina. These cells have spherical soma which lie in the inner nuclear layer (INL), adjacent to or one cell distal to the inner plexiform layer (IPL). Processes from these cells project throughout the IPL, with a fairly dense accumulation of labeled dendrites in the upper two-fifths of the IPL and a dense, narrow band of labeled dendrites adjacent to the ganglion cell layer. These cells also have finer processes, originating at the cell body, that traverse the INL and ramify in the outer plexiform layer (OPL). Double label experiments show that all of the cells that contain somatostatin-like immunoreactivity (SOM-LI) in the INL are also labeled by high-affinity uptake with 3H-glycine. Immunocytochemistry of retinal whole mounts shows that these cells are evenly distributed across the retina at a density of 542 +/- 65 cells/mm2. On the basis of the colocalization experiments and the morphological homogeneity of these cells, we suggest that they represent a single cell type. Interplexiform cell processes were further characterized by electron microscopy after immunocytochemistry or 3H-glycine autoradiography. In the IPL, IPC processes are seen to be postsynaptic at both ribbon and conventional synapses. This input is found almost entirely in the distal two-fifths of the IPL. Interplexiform cell processes are presynaptic to unlabeled processes in both the distal and proximal IPL. In the OPL, labeled processes are found near or contiguous with photoreceptor bases, and are often presynaptic to small-diameter processes. The postsynaptic processes have been identified as bipolar cell dendrites in six cases. Interplexiform cell processes may also contact horizontal cell processes in the OPL.

The action of gamma-aminobutyric acid on the horizontal cells of the skate retina

The effects of gamma-aminobutyric acid (GABA) were studied in the superfused retina of the skate. Intracellular recordings were made from horizontal cells. After application of 500 microM GABA there was a depolarization of the membrane potential, a decrease in the light-evoked amplitude of the response and an increase in the duration of the waveform.

Dopamine modifies the balance of rod and cone inputs to horizontal cells of the Xenopus retina

Dopamine (greater than or equal to 2 microM) increased the cone input and suppressed the rod input to axon-bearing horizontal cells of the Xenopus retina. Dopamine (10 microM) also depolarized the horizontal cell by about 9 mV. The D2-dopamine antagonists spiperone and metoclopramide had the opposite action to dopamine, whereas the D1-dopamine antagonist SCH 23390 was without effect. None of the agents tested modified the light-evoked responses of rods.

GABA release from Xenopus retina does not correlate with horizontal cell membrane potential

The relationship between horizontal cell membrane potential and the release of GABA was explored in the retina of Xenopus laevis. The intracellularly recorded membrane potential of horizontal cells was monitored while the retina was exposed to different concentrations of depolarizing agents. The dose-response curves obtained revealed a rise from 5 to 95% maximum depolarization in 0.5-1.5 log unit concentration change. The molar concentrations that elicited a 20 mV depolarization were 40 mM (potassium), 0.8 mM (glutamate), 0.8 mM (glycine), 5 microM (kainate) and 1.3 microM (quisqualate). Autoradiography revealed that radiolabel was accumulated almost exclusively by horizontal cells when isolated retinas were incubated in medium containing 1 microM [3H]GABA. Thus, retinal release of radioactivity was used as a measure of [3H]GABA release from horizontal cells. Endogenous GABA released from retinas was measured using high performance liquid chromatography and was taken to reflect both amacrine and horizontal cell GABA pools. The release of both [3H]GABA and endogenous GABA was stimulated by glutamate, kainate and potassium, but not by glycine or quisqualate. Similar dose-response curves for GABA release and for depolarization were obtained in the case of potassium and kainate but not for glutamate. Potassium-evoked release either of endogenous GABA or [3H]GABA was both calcium- and sodium-dependent, whereas kainate- or glutamate-evoked GABA release was sodium-dependent but calcium-independent. The results indicate that depolarization per se is not necessarily associated with transmitter release in Xenopus retinal horizontal cells. It is suggested that the action of a given neurotransmitter upon the efflux of GABA from horizontal cells may depend on the degree to which it modifies the sodium conductance of the horizontal cell.

Stimulation of rods can increase cone flicker ERGs in man

Recent psychophysical studies in man and electrophysiological studies in lower vertebrates show that dark adapted, unstimulated rods inhibit cone mediated flicker. This investigation uses comparable psychophysical and ERG procedures in man to demonstrate rod-cone interaction of this type. With either procedure the rod cone interaction cannot be demonstrated with Ganzfeld stimulation. A single small, red, flickering test field, which is a common psychophysical stimulus for testing rod-cone interaction, elicits an immeasurably small cone ERG. But an array of many such targets, flickering synchronously, is an effective psychophysical stimulus and produces an ERG with larger cone than rod components. With such an array, it can be shown that a steady, rod-stimulating background selectively enhances cone ERG components.

Pharmacological modification of the light-induced responses of Müller (glial) cells in the amphibian retina

The light-evoked responses of Müller (glial) cells were monitored by intracellular recording in the isolated, superfused retina of Xenopus laevis. Müller cells had dark resting potentials of -88.5 +/- 6.9 mV and small 1-2 mV light responses of variable waveform in normal Ringer's solution. Exposure to picrotoxin (0.5-1.0 mM) greatly enhanced the light response which then consisted of depolarizing transients (Vmax 5-15 mV) at stimulus onset and offset. GABA (5-10 mM) antagonized the picrotoxin effect and suppressed the light response, whereas 2-amino 4-phosphonobutyrate (0.10-0.15 mM) blocked selectively the 'on' transient. None of these agents appreciably modified the glial cells resting potential level. On the other hand, veratrine (6-9 micrograms/ml) depolarized the Müller cell by 4-13 mV and slowed and greatly reduced the light response. These effects were antagonized by tetrodotoxin (1-4 microM) which itself reduced the light response by 30-50% without altering its shape. On the basis of these findings, we suggest that alterations in the activity of the inner retinal neurons, i.e. amacrine and ganglion cells, are primarily responsible for the drug-induced changes in the membrane potential and light-evoked responses of the Müller cell.

Uptake of 3H-glycine in the outer plexiform layer of the retina of the toad, Bufo marinus

The uptake of 3H-glycine in the retina of the toad, Bufo marinus, was investigated by light and electron microscopical autoradiography. Uptake of 3H-glycine was very prominent in large cell bodies in the inner nuclear layer as well as in discrete clusters in both the outer plexiform layer (OPL) and the inner plexiform layer. This pattern in similar to that described for 3H-glycine-accumulating putative interplexiform cells in goldfish, frog, and Xenopus retinas. Electron microscopical autoradiography of the OPL revealed large, grain-containing varicosities which had electron-lucent cytoplasm and contained both small, agranular and large, dense-core vesicles. The varicosities made extensive en passant and spine synapses in the OPL. Definitive identification of their postsynaptic targets was not achieved. However, autoradiographic analysis with 3H-GABA uptake as well as electrophysiological evidence suggests that axons but not cell bodies or dendrites of 3H-GABA-accumulating horizontal cells (H1 cells) are postsynaptic targets of the varicosities. The presence of dense-core vesicles in the varicosities suggested co-occurrence of glycine and a biogenic amine or neuropeptide. The indirect immunofluorescence technique was used to determine whether any such substances were present in the OPL of the toad retina. However, no specific labeling was found in the OPL for any of 19 substances tested. The extensive synaptic output provided by glycine-accumulating varicosities in the toad OPL may indicate an important role of glycine in the synaptic function of the distal toad retina. We suggest that these varicosities derive from a presumably glycinergic interplexiform cell.