|
|
|
|
|
|
|
||
Original Article |
m.kamermans{at}ioi.knaw.nl
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 s and can be attributed to processes intrinsic to the cones. It does not require dopamine, is not the result of changes in the kinetics of the cone light response and is not due to changes in horizontal cells themselves. During a flash train, cones adapt to the mean light intensity, resulting in a slight (4 mV) depolarization of the cones. The time constant of this depolarization is 3 s. We will show that at this depolarized membrane potential, a light-induced change of the cone membrane potential induces a larger change in the calcium current than in the unadapted condition. Furthermore, we will show that negative feedback from horizontal cells to cones can modulate the calcium current more efficiently at this depolarized cone membrane potential. The change in horizontal cell response properties during the train of flashes can be fully attributed to these changes in the synaptic efficiency. Since feedback has major consequences for the dynamic, spatial, and spectral processing, the described mechanism might be very important to optimize the retina for ambient light conditions.
Key Words: retina • synapse • goldfish • calcium current
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), play a prominent role in the generation of spectral opponent HCs (Stell and Lightfoot 1975; Kamermans et al. 1991), and form the neural basis of the color constancy (Kamermans et al. 1998).
The strength of the feedback signal from HCs to cones seems to vary strongly with the adaptation state of the retina. For instance, in the light adapted retina, two types of spectrally opponent HCs can be found, whereas in the fully dark adapted retina, spectrally opponent HCs are completely absent (Weiler and Wagner 1984). Changes in response properties of ganglion cells can also be found. In a fully light-adapted condition, these cells have antagonistic surrounds. These surrounds are absent in the dark-adapted condition (Raynauld et al. 1979). Given this important role of negative feedback in the signal processing in the retina, variations of the strength of negative feedback will have strong impact on the whole visual system.
The Mechanism of Negative Feedback in the Goldfish Retina
For goldfish, the negative feedback pathway from HC to cones has been described in detail (Verweij et al. 1996b; for review see Kamermans and Spekreijse 1999). Horizontal cells feed back to cones by modulating the cone Ca-current directly in a 
-aminobutyric acid (GABA)–independent way. Hyperpolarization of the HCs results in a shift of the Ca-current activation function to more negative potentials, yielding an increase in Ca influx in the cone synaptic terminal and thus an increase in glutamate release (Verweij et al. 1996a). The neurotransmitter involved is presently unknown. Although the properties of the negative feedback pathway are beginning to be resolved, the way this pathway is modulated during light/dark adaptation is not clear at all.
In this study, we describe a mechanism by which the efficiency of the feedback signal is changed during adaptation. To change the adaptation state of the retina, a train of bright white flashes was used. It will be shown that during this stimulus protocol, the efficiency of the feedback signal increases strongly due to a slight depolarization of the cones. This cone depolarization is due to intrinsic photoreceptor adaptation. This study illustrates that local photoreceptor adaptation yields a strong change in the network properties of the retina.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 920 nm) illumination.
Patch Clamp Measurements and Intracellular Recordings
Optical stimulator.
A 450-W Xenon lamp supplied two beams of light. These were projected through Uniblitz VS14 shutters (Vincent Associates), neutral density filters (NG Schott), lenses, and apertures. For the patch clamp measurements, the spots, 20 µm in diameter, were projected through a 40x water immersion objective (N.A. = 0.55) of the microscope and the spots, 3,000 µm in diameter, were projected through the condenser (N.A. = 1.25) of the microscope. The full-field white light stimuli used for intracellular recordings were projected onto the retina through a 2x objective lens (N.A. = 0.08) of the microscope. For all experiments, only white light stimuli were used. The light intensities are expressed in log units relative to 4 x 103 cd/m2. Since we used white light stimuli in our experiments, we compared the size of the cone light responses to white light with the responses of M cones to 550-nm light flashes and calculated the amount of effective quanta in our white light to be 1.0 x 105 s–1 µm–1 (550 nm). In goldfish, Malchow and Yazulla 1988 calculated that 50% of the pigment was bleached when the retina was stimulated with 3.16 x 109 photons s–1 µm–1, left on for 4 min. The amount of photons they used exceed our stimulus condition by six log units. This is an indication that in our experimental conditions, bleaching hardly plays a role.
Electrodes and recording equipment.
The patch pipettes were pulled from borosilicate glass (GC150TF-10; Clark) and the intracellular microelectrodes were pulled from aluminosilicate glass (o.d. = 1.0 mm, i.d. = 0.5 mm; Clark) with a P-87 micropipette puller (Sutter Instruments Co.).
Patch pipettes had impedances between 5 and 10 M
when filled with pipette medium and measured in Ringer's solution. The series resistance during the whole cell recording was <12 M. Electrodes were mounted on an MP-85 Huxley/Wall-type micro manipulator (Sutter Instruments Co.) and connected to a 3900A Integrating Patch Clamp (Dagan Corp.).
Microelectrodes had impedances ranging from 100 to 200 M when filled with 4 M KAc. The intracellular voltages were measured with an electrometer (S7000A; World Precision Instruments) and recorded on paper (Graphtec Linearcorder).
Data acquisition, control of the optical stimulator, and control of the patch clamp were done with a 1401 AD/DA convertor (Cambridge Electronic Design Ltd.) and an MS-DOS–based computer system.
Ringer Solutions and Pipette Medium
The Ringer solution contained (mM): 102.0 NaCl, 2.6 KCl, 1.0 MgCl2, 1.0 CaCl2, 28.0 NaHCO3, 5.0 glucose, and was continuously gassed with 2.5% CO2 and 97.5% O2, yielding a pH of 7.8. Some Ringer solutions contained drugs as indicated in the text, and in the figure legends: flupentixol (50 µM; Research Biochemicals, Inc.), niflumic acid (100 µM; Sigma-Aldrich Chemical Co.), CoCl2 (2 mM; Merck). In tetraethylammonium (TEA)-Cl Ringer's solution, 15 mM of the NaCl was replaced by 5 mM CsCl and 10 mM TEACl.
The standard patch pipette medium contained (mM): 12.0 KCl, 61.0 D-gluconic-K, 1.0 MgCl2, 0.1 CaCl2, 1.0 EGTA, 5.0 HEPES, 5.0 ATP-Na2, 1.0 GTP-Na3, 0.2 3':5'-cGMP-Na, 20 Phosphocreatine-Na2, 50 U/ml creatine phosphokinase. To change the Cl-equilibrium potential, KCl was exchanged for equimolar D-gluconic-K. The pH of the pipette medium was adjusted to 7.25 with KOH. Cesium pipette mediums contained equimolar D-gluconic-Cs and 12.0 CsCl instead of the KCl and D-gluconic-K. All chemicals were obtained from Sigma-Aldrich Chemical Co.
Liquid Junction Potential
The liquid junction potential was measured with a patch pipette filled with the pipette medium, and positioned in a bath filled with pipette medium. The reference electrode was filled with 3 M KCl. After the potential was adjusted to zero, the bath solution was replaced with Ringer's solution. The resulting potential change was considered as the junction potential. The liquid junction potential was determined for the various pipette solutions and all data were corrected accordingly.
Dopamine Depletion of the Retina
Dopaminergic interplexiform cells were destroyed with intraocular injections of 5 µl Ringer solution containing 6-hydroxy-dopamine (5 µl) and paragyline (10 µl; both from Sigma-Aldrich Chemical Co.), administered on two successive days (Negishi et al. 1982; Lin and Yazulla 1994). These fish were used 14–33 d after injection. All dopamine-depleted retinas used for intracellular recordings were tested afterwards for the presence of dopamine-containing cells using the tyrosine hydroxylase method (Yazulla and Studholme 1997). Data were taken only from retinas that did not contain any tyrosine hydroxylase–positive cells.
Cell-type Classification
Cones were selected under visual control and spectrally classified (Kraaij et al. 1998). Only L, M, and S cones were found. HCs were classified on their spatial and spectral properties (Norton et al. 1968; Kaneko 1970; Hashimoto et al. 1976; Kaneko and Stuart 1984). Of the three classes of HCs, only monophasic HCs were used in the present study.
Stimulation Protocols
Presynaptic calcium current protocol.
Since the large modulation of the light-sensitive conductance masks the much smaller presynaptic calcium current of the cone, it is not possible to measure light-induced changes in presynaptic calcium currents of cones directly. To overcome this problem, previously recorded light responses of a cone to spot stimuli (20 µm diameter) of 500-ms durations were "played back" as command voltages in a condition where the light conductance of the cone is blocked with an intense (–0.15 log) white spot (20 µm diameter). In this way, "light-induced" changes in presynaptic calcium current can be measured without the interference of the light-sensitive conductance. Fig. 1 explains the voltage clamp protocol used to isolate light-induced changes in calcium current. The insert displays the original hyperpolarizing response of a cone to a 500-ms flash that was used to construct the voltage clamp protocol. Trace 1 consists of this hyperpolarizing response repeated nine times. Trace 2 is similar to trace 1, but now with a depolarization of 4 mV superimposed, applied in discrete 1-mV steps just before the second, third, fourth, and fifth hyperpolarizing responses. Because the 1-mV voltage steps were taken in between the "light flashes," the whole light response is shifted to a 1-mV more-depolarized level. During the last four hyperpolarizing steps, this final 4-mV depolarized level was maintained. For trace 3, a similar protocol as for trace 2 was generated, but the depolarizing steps were twice as large, yielding a total depolarization of 8 mV. The final amount of depolarization is indicated by the numbers to the right of the voltage traces. After the "flashes," a 100-ms step to –7 mV was applied, followed by a 1,500-ms step to –77 mV. Finally, the cone membrane potential was stepped back to the initial potential. This last part of the protocol was used to study the tail currents. During the step to –77 mV, a 50-ms step to –87 mV was applied. This step was used for leak subtraction to isolate the presynaptic currents of the cone. The part of the protocol used for determining the tail and leak currents are not shown in the figures.
|
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Definitions
HCs feed back to cones by shifting the calcium current activation function of the cones using an unknown neurotransmitter. Because we cannot measure the amount of this neurotransmitter, we will define the feedback signal as the size of the shift of the calcium current activation function of the cone. The resulting change in the calcium influx will be called the efficiency of the feedback signal.
HC Kinetics during the Flash Train
Fig. 2 A shows HC responses to a train of 10 white light flashes of 500-ms durations with an inter-stimulus interval of 200 ms for three different intensities (–2.0, –1.0, and 0.0 log). For the –2.0 log intensity, the responses are small and sustained. For the –1.0 log intensity, the responses are larger and show a pronounced secondary depolarization; i.e., the difference between peak and plateau value. This secondary depolarization has been shown to be the consequence of negative feedback from HCs to cones (Piccolino et al. 1981; Wu 1994; Kamermans and Spekreijse 1999; Kraaij et al. 1998).
|
Changes in HC Kinetics Do Not Depend on Dopamine
It is known that dopamine modulates the feedback signal from HCs to cones (Kirsch et al. 1990). To test whether dopamine can account for the observed changes of HC light responses during the flash train, the experiment of Fig. 2 was repeated in dopamine-depleted animals. Fig. 3 A shows the first and last light responses of a HC from a retina, without dopaminergic interplexiform cells (IPCs). Like Fig. 2, the gray bar and the dashed lines indicate the amount of secondary depolarization of the first and last flash, respectively. The retina was isolated 14 d after intraocular injection of 6-hydroxy-dopamine, which is known to kill the IPCs. In these dopamine-depleted retinas, the changes in HC kinetics during the flash train remain present. This result was found in eight retinas that did not stain for tyrosine hydroxylase. Furthermore, blocking the D1 and D2 receptors in control retinas with the antagonist flupentixol (Fig. 3 B) did not influence the changes in response shape during the flash train.
|
Presynaptic Changes during the Flash Train
One possibility is that changes in the cone light response during the flash train can account for the effects observed in the HC response. Fig. 4 A shows the voltage responses of a cone under whole-cell configuration to the same flash train as used for the HCs. Cones hyperpolarize in response to repetitive stimulation, but, unlike HCs, the kinetics of the responses show a decrease in transientness (Fig. 4 B). This result was obtained for both full-field as well as small-spot stimulation (20 µm in diameter; not shown). However, the mean membrane potential of the cone depolarizes with the number of flashes. In 12 cells tested, the mean depolarization was 3.3 ± 1.4 mV (31% of the amplitude of the light response of each cone). The time constant of this depolarization was 3.3 ± 0.5 s, which is of the same order as the time constant of the change of the secondary depolarization of the HCs.
|
|
); i.e., the inward current activated by hyperpolarization (Ih), the calcium current (ICa), the calcium-dependent chloride current (IClCa), and the delayed rectifier (IKdr). Blocking the IKdr with extracellular TEA and intracellular Cs2+ did not change the increase in current during the voltage protocol (not shown). This result excludes the possibility that IKdr is modulated during this protocol. The increase in the modulation depth of a light flash due to depolarization of the cone is not likely to be due to modulation of Ih because depolarization of the cone would then decrease the amount of current. The two remaining currents, ICa and IClCa, are located in the terminal of the cone (Taylor and Morgans 1998) and could therefore be involved in synaptic transmission.
It is known that large divalent cations like cobalt (Co2+) block Ca-dependent synaptic transmission. To determine the contribution of ICa to the presynaptic currents, the effect of Co2+ on the experiments of Fig. 5 was studied. To check the effectiveness by which we could block ICa, we first determined the effect of Co2+ on the current–voltage relations of the cones. Fig. 6 (top) shows the leak-subtracted current responses to a voltage ramp protocol from –70 to 20 mV (drawn at the bottom of the figure) before (1), during (2), and after (3) cobalt. Co2+ shifts the calcium current activation function to more positive potentials (n = 4). This is a general feature of large divalent cations (Kostyuk et al. 1982; Hille 1992; Piccolino et al. 1996). This shift will block the synaptic transmission from cones to HCs because ICa is shifted out of the operating range of the cone. Fig. 7 shows that the presynaptic currents are almost completely blocked with Co2+. Fig. 7 (top, Control) shows the increase in light-induced changes in the presynaptic currents due to the voltage trace that depolarizes to 8 mV. This trace is similar to the 8-mV trace, presented in Fig. 5. Fig. 7 (middle) is the response of the same cell, but now in the presence of 2 mM Co2+. Almost no modulation of the presynaptic currents is present (n = 4). After washing out the Co2+ (Fig. 7, bottom), the presynaptic currents reappear.
|
|
|
). Fig. 9 B shows that application of 100 µM niflumic acid did not block the light-induced increase in synaptic currents (n = 4), whereas tail currents present in control (Fig. 9 A, 1) disappeared in the presence of niflumic acid (2). Furthermore, changing the calculated equilibrium potential for chloride from –47 to –28 mV did not change the depolarization-induced increase of the presynaptic currents (not shown). These experiments show that IClCa is not involved in the increase in presynaptic currents. Therefore, we can conclude that the increase in outward current, observed when the cone is allowed to depolarize is actually a decrease in the standing inward ICa.
|
Changes in HC Properties during the Flash Train
One of the striking changes in the HC response properties is the change in the secondary depolarization. This secondary depolarization is due to negative feedback from HCs to cones. If the changes in HC kinetics are due to changes in HC properties, one should expect that the output of the HCs would also change. Because we cannot measure the amount of neurotransmitter released by the HCs, we looked at the effect of feedback on the cones; i.e., the change in calcium influx in the cone terminal. This effect of the feedback signal on the cone calcium current can be measured most effectively by clamping the cone at one potential, saturating the cone with a small spot and stimulating the retina with a full field flash stimulus (for details of the feedback protocol, see MATERIALS AND METHODS). The result of this feedback protocol is that the cone recorded from responds only to the signal it receives from the HCs. In this way, the feedforward signal can be separated from the feedback signal, which cannot be done measuring HCs themselves. Fig. 10 shows a voltage-clamped cone stimulated with the feedback protocol. It is obvious from this figure that the feedback-induced inward current shows no pronounced changes during the flash train (n = 4). So, contrary to the increase in secondary depolarization, the feedback signal as measured in the cones has not increased during the flash train.
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the next sections, we will discuss this mechanism in detail and present results of a quantitative computer simulation showing that cone polarization alone is enough to generate the observed results. Finally, we will discuss the relevance of this mechanism in relation to the literature concerning adaptational changes in retinal processing.
Changes Seen during the Adapting Flash Train
Two things happen with the HC responses during the adapting flash train used in this study that need to be accounted for: (a) the onset response becomes faster and (b) the secondary depolarization becomes larger.
The only significant change found in the cones during the flash train is that the mean cone membrane potential slightly depolarized during the flash train. Both the transient character of the first light response of the cone and the depolarization of the mean membrane potential are presumably due to the Ca-dependent feedback on the phototransduction pathway (Haynes and Yau 1985a; Ripps and Pepperberg 1987). The question now is, can the changes in HC kinetics during the flash train be accounted for by only a slight depolarization of the cone membrane potential?
Depolarization of the Cone Can Account for the Changes in HC Kinetics during the Train of Flashes
Fig. 12 A shows the current-voltage relation of the ICa of a cone. Since there is a linear relation between ICa of the cone and the amount of glutamate release, the ICa can be taken as the amount of glutamate release (Witkovsky et al. 1997). In the dark, cones rest at about –40 mV. At this potential, there is a certain amount of glutamate release (Fig. 12 A, 1). Hyperpolarization of the cone by light results in a reduction in ICa; i.e., in a decrease in glutamate release (Fig. 12 A, 2). Due to this decrease in glutamate release, HCs will hyperpolarize, leading to a change in feedback signal to the cones. By some yet unknown mechanism, the change in feedback shifts the calcium activation function to more negative potentials (dashed curve). This in turn causes an increase in ICa (Fig. 12 A, 3), which leads to an increased glutamate release, leading to the secondary depolarization of the HC responses.
|
4 mV, so the cone rests at –36 mV. Because the cone gradually becomes depolarized during the train of flashes, the range of its light response is shifted to a steeper part of the calcium activation function. Thus, the same hyperpolarizing cone light response will cause a larger decrease in calcium current, and a larger reduction in glutamate release, accounting for the increase in the feedforward HC response. The gradual depolarization of the cone during light adaptation can also explain the increase in the secondary depolarization in HCs. During the flash train (i.e., during light adaptation), the feedback signal from HCs to cones does not change. This means that in the light-adapted state the feedback from the HCs during each flash causes the same leftward shift of the calcium activation function as in the dark-adapted state. However, since the cone potential is in a steeper part of the calcium-activation function, this same leftward shift now causes a larger increase in presynaptic calcium current, a larger increase in glutamate release, and a larger secondary depolarization in HCs. So, now that we have discussed the increase in secondary depolarization of the HC light response due to the small depolarization of the cone, we can discuss how both the increase in feedforward signal and the increase in feedback efficiency can lead to a faster time to peak of the HC light response. The feedforward signal increases less than the increase in feedback efficiency (compare Fig. 5 and Fig. 11). The reason for this is that the 4-mV depolarization is a small depolarization relative to the cone light response, whereas it is a large depolarization relative to the feedback-induced shift in calcium activation function. Since feedback is slower than feedforward, and since the feedback efficiency increases more than the feedforward signal, the result is that the time to peak of the HC responses decreases. Now the question arises why this decrease in time to peak of the HC response and the increase in rollback cannot be seen in the feedback signal received by the cones. Since feedback is slow, the peak of the HC light response might be cut off. Furthermore, the rollback is small, relative to the total response of the HCs, which actually generates the feedback signal. Together, they may explain why the changes in HC kinetics cannot be seen in the feedback signal received by the cones. To test whether the effect of the depolarization on the ICa of the cone was enough to account for the observed changes in the kinetics due to this increase in synaptic efficiency, a quantitative model simulation was performed.
Simulation
The model consists of a cone and a HC. The cone projects via a Ca-dependent glutamatergic pathway to the HC and the HC feeds back to the cone via modulation of the Ca current by an unknown neurotransmitter (Verweij et al. 1996a). Cones are modelled using a simple resistive network consisting of a voltage-sensitive calcium conductance, the light conductance, and a passive potassium conductance. HCs are modelled with a glutamate conductance and a passive potassium conductance. A detailed description of the model is given in the .
With the model, three simulations were performed. (a) The response of the model was studied using a similar flash train as used for the physiological experiments. To mimic the adaptation-induced cone depolarization as found experimentally, the cone membrane potential was depolarized by 1, 2, 3, and 4 mV during the second, third, fourth, and fifth and the rest of the flash train, respectively. The results are shown in Fig. 13 A. It is clear from this figure that the secondary depolarization increases with depolarization of the cone. This is a similar behavior as found in the physiological cone/HC system. Note that only the mean cone membrane potential is changed in the model during the adaptation by the flash train, showing that this small change is enough to induce the changes in kinetics of the HC response.
|
Finally, the feedback signal in the model cones was studied (Fig. 13 C). This was done in two conditions: (a) when the model cone was clamped at –45 mV and (b) when the cone was initially clamped at –60 mV and subsequently depolarized during the first five flashes by a total of 4 mV. Fig. 13 C shows that when the cone is clamped at –45 mV, feedback-induced currents remain nearly equal in size throughout the protocol. This result is similar to the physiological result. However, when the cone is clamped at the light membrane potential (–60 mV) and is depolarized by a total of 4 mV, the feedback-induced currents are almost doubled at the end of the stimulus train, just as found physiologically.
These simulations show that a very simple model can account for the complex change in the HC response characteristics observed during adaptation to a flash train. Since this behavior can be generated with a model that only includes a Ca current and cone adaptation, it indicates that no other pathways or processes are involved. The relative position of the cone membrane potential to the Ca-current activation function is essential for the size and the kinetics of the HC response. Depolarization of the cone membrane potential due to adaptation of the cone leads to an increase in efficiency of the feedback pathway of the HCs to the cones.
Changes in HC/Cone Network Induced by Adaptation
Many papers have reported changes in the cone/HC network as a consequence of light adaptation. The mechanisms underlying these changes can be separated into two broad categories: "local" and "network" adaptive mechanisms. Local mechanisms are intrinsic to photoreceptors such as bleaching adaptation (Ripps and Pepperberg 1987) and Ca-dependent feedback on the phototransduction cascade, which modulates the cGMP-gated channels in the cone outer segment (Haynes and Yau 1985b). Both local mechanisms tend to depolarize the mean membrane potential of the cone (Burkhardt 1994) when the retina becomes light adapted. The time constant of this process is between 0.1 and 100 s (Baylor and Hodgkin 1974).
Network adaptation refers to mechanisms outside the photoreceptor, such as modulation of the electric coupling between HCs by dopamine (Mangel and Dowling 1985; Tornqvist et al. 1988; Yang et al. 1988; Negishi et al. 1990), the change in sensitivity of the glutamate and GABAc receptors on the HCs (Knapp and Dowling 1987; Dong and Werblin 1994), and the change in the amount of GABA released by the HCs (Yazulla and Kleinschmidt 1982). These dopaminergic processes function on a time scale of minutes to hours (Witkovsky and Shi 1990). GABA also plays an important role in determining the network properties in the outer retina (Yang and Wu 1989; Kamermans and Werblin 1992; Verweij et al. 1998) by modulating both the size and kinetics of the HC responses by means of a mechanism intrinsic to the HCs.
The present paper describes changes in HC kinetics without a change in the intrinsic HC properties, indicating that GABA does not play a role. The described mechanism is also present in dopamine-depleted animals, indicating that dopamine does not play a role in this short-term change of the HC kinetics. Other groups have also reported changes in HC kinetics due to background illumination or repetitive stimulation without the interference of GABA or dopamine. Akopian et al. 1991 reported changes in the responses of turtle HCs during repetitive stimulation with saturating white light flashes. These HCs show an increase of the Off overshoot, while the On response remained unaffected. Since it has been reported that the HC Off overshoot is mediated by feedback (Kamermans et al. 1989), the findings of Akopian et al. 1991 could be accounted for by the mechanism described in this paper. The absence of any effect on the On response in their experiments can be justified by the fact that they used saturating light stimuli. That condition might be similar to the HC responses to 0.0 log intensity light flashes shown in Fig. 2. For that intensity, we also did not find any effect on the On responses of HCs.
Normann and Perlman 1990 showed that background illumination reduces the sensitivity of the HC responses in turtle. High intensities of background light did not change the kinetics of the HC light response, whereas lower intensities lead to a reduction of the time to peak and an increase in secondary depolarization of the HC light response. Their lower intensities are comparable with the intensities in our experiments, eliciting the changes in HC response kinetics. Their results are similar to the results described in this paper despite their use of background light instead of a train of white light flashes. In their paper, no explanation was given why the time to peak of the HC response decreased when they used moderate light intensities. In light of our experiments, their results can be accounted for by the light adaptation–induced depolarization of the cone, which will make the synaptic efficiency between cones and HCs higher, leading to a reduction of the time to peak of the HC response.
The Possible Role of Cone Depolarization in Other Adaptive Processes
This study illustrates that the resting membrane potential of the cone can strongly modulate the size and kinetics of the light responses of neurons in the retina. Small changes in this potential can lead to drastic changes in synaptic transmission, and feedback efficiency in particular, influencing possibly all kinds of visual processing. As mentioned in the INTRODUCTION, several papers have reported changes in the amount of feedback during light adaptation. Both the surround of ganglion cells and the spectral opponent responses of HCs disappear when a retina is dark adapted. Light adapting these retinae results in the reappearance of these feedback-mediated responses. Most of these changes during light–dark adaptation are assumed to be mediated by GABA or dopamine. Although the mechanism described in this paper is not GABA or dopamine mediated, it could very well play a role in the above-mentioned GABA- and dopamine-mediated processes. For instance, GABA is known to modulate the membrane potential of cones. GABA slightly hyperpolarize cones in goldfish, tiger salamander, and turtle by activating GABAA receptors (Kaneko and Tachibana 1986; Kamermans and Werblin 1992). By hyperpolarizing the resting membrane potential of the cone, GABA could decrease the feedback efficiency and thereby inhibit feedback-mediated responses. So, in conclusion, the mechanism to modulate the synaptic transmission between HCs and cones described in this paper can be driven by a variety of pathways. The effect will always be a strong modulation of feedback.
This appendix presents a very simple quantitative model to account for the experimental findings presented in this paper with a minimum of components. The main question addressed is whether the characteristics of the Ca channel are sufficient to account for the observed changes in HC dynamics.
The model consists of a cone and a HC. The cone projects via a Ca-dependent glutamatergic pathway to the HC and the HC feeds back to the cone via modulation of the Ca current by an unknown neurotransmitter (Verweij et al. 1996a).
| The Cone The cone membrane potential (Vcone) depends on the light stimulus [I(t)]. The relation between the I(t) and Vcone is given by :
 cone is the time constant of the cone light response in milliseconds, Vrest is the cone resting membrane potential in millivolts, Vresp is the amplitude of the cone light response in millivolts, and I(t) varies between 0 and 1. The timing of I(t) is given in the figures.
Vcone modulates a voltage-dependent Ca channel (ICa). The relation between ICa and Vcone is taken from Verweij et al. 1996a
The Horizontal Cells
The relation between gcat and the HC membrane potential (VHC) is given by
HC is the time constant of the HCs, Ecat is the equilibrium potential of the glutamate-gated channel, gKHCis the relative passive potassium conductance in the HC, and EKHCis the equilibrium potential of the glutamate-gated current in the HC.
HCs feedback to cones by directly modulating the Ca-current activation range. HCs feedback to cones in an almost linear fashion (our unpublished observations). The time constant of the feedback signal is
fb is the time constant of the feedback signal, and Afb and Bfb are constants.
Parameters and Methods
|
|
-aminobutyric acid; HC, horizontal cell; ICa, calcium current; IClCa, calcium-dependent chloride current; TEA, tetraethylammonium To determine the amount of sustained depolarization during the flash train, we determined for each flash response the mean membrane potential in a time window of 225–700 ms after the onset of the flash. Through these mean membrane potentials, an exponential curve was fitted. The amplitude and time constant (
) of this curve were determined and can be interpreted as the amplitude and time constant of the flash train–induced depolarization. The rational for this approach is that during the first light response the cone already starts to adapt, and thus starts to depolarize. Just determining the membrane potentials of the light response at the sustained potentials will in that way underestimate the change in cone membrane potential. Taking the depolarized levels as a measure for the cone membrane potential also yields an incorrect value because sometimes after the first flash an overshoot occurs that is of course absent at the onset of the first flash.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by the Netherlands Organization for Scientific Research (NWO).
Submitted: 25 March 1999
Revised: 2 August 1999
Accepted: 9 August 1999
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  M. VanLeeuwen, I. Fahrenfort, T. Sjoerdsma, R. Numan, and M. Kamermans Lateral Gain Control in the Outer Retina Leads to Potentiation of Center Responses of Retinal Neurons J. Neurosci., May 13, 2009; 29(19): 6358 - 6366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kamermans, D. Kraaij, and H. Spekreijse The dynamic characteristics of the feedback signal from horizontal cells to cones in the goldfish retina J. Physiol., July 15, 2001; 534(2): 489 - 500. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Kraaij, H. Spekreijse, and M. Kamermans The Open- and Closed-Loop Gain-Characteristics of the Cone/Horizontal Cell Synapse in Goldfish Retina J Neurophysiol, September 1, 2000; 84(3): 1256 - 1265. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
