Hydrogen Sulfide Protects the Retina from Light-induced Degeneration by the Modulation of Ca2+ Influx*

Background: Hydrogen sulfide (H2S) has been recognized as a signaling molecule as well as a cytoprotectant. Results: Ca2+ regulates the 3-mercaptopyruvate sulfurtransferase/cysteine aminotransferase pathway to produce H2S production. H2S, in turn, regulates Ca2+ influx and protects retinal neurons from light-induced degeneration. Conclusion: H2S regulates Ca2+ levels and protects retinal neurons. Significance: It provides a possible role of H2S and its therapeutic application in the retina. Hydrogen sulfide (H2S) has recently been recognized as a signaling molecule as well as a cytoprotectant. Cystathionine β-synthase (CBS) and cystathionine γ-lyase (CSE) are well-known as H2S-producing enzymes. We recently demonstrated that 3-mercaptopyruvate sulfurtransferase (3MST) along with cysteine aminotransferase (CAT) produces H2S in the brain and in vascular endothelium. However, the cellular distribution and regulation of these enzymes are not well understood. Here we show that 3MST and CAT are localized to retinal neurons and that the production of H2S is regulated by Ca2+; H2S, in turn, regulates Ca2+ influx into photoreceptor cells by activating vacuolar type H+-ATPase (V-ATPase). We also show that H2S protects retinal neurons from light-induced degeneration. The excessive levels of light exposure deteriorated photoreceptor cells and increased the number of TUNEL- and 8-hydroxy-2′-deoxyguanosine (8-OHdG)-positive cells. Degeneration was greatly suppressed in the retina of mice administered with NaHS, a donor of H2S. The present study provides a new insight into the regulation of H2S production and the modulation of the retinal transmission by H2S. It also shows a cytoprotective effect of H2S on retinal neurons and provides a basis for the therapeutic target for retinal degeneration.


Hydrogen sulfide (H 2 S) has recently been recognized as a signaling molecule as well as a cytoprotectant. Cystathionine
␤-synthase (CBS) and cystathionine ␥-lyase (CSE) are wellknown as H 2 S-producing enzymes. We recently demonstrated that 3-mercaptopyruvate sulfurtransferase (3MST) along with cysteine aminotransferase (CAT) produces H 2 S in the brain and in vascular endothelium. However, the cellular distribution and regulation of these enzymes are not well understood. Here we show that 3MST and CAT are localized to retinal neurons and that the production of H 2 S is regulated by Ca 2؉ ; H 2

S, in turn, regulates Ca 2؉ influx into photoreceptor cells by activating vacuolar type H ؉ -ATPase (V-ATPase). We also show that H 2 S protects retinal neurons from light-induced degeneration. The excessive levels of light exposure deteriorated photoreceptor cells and increased the number of TUNEL-and 8-hydroxy-2deoxyguanosine (8-OHdG)-positive cells. Degeneration was greatly suppressed in the retina of mice administered with NaHS, a donor of H 2 S. The present study provides a new insight into the regulation of H 2 S production and the modulation of the retinal transmission by H 2 S. It also shows a cytoprotective effect of H 2 S on retinal neurons and provides a basis for the therapeutic target for retinal degeneration.
Hydrogen sulfide (H 2 S) is synthesized by three enzymes: cystathionine ␤-synthase (CBS), 2 cystathionine ␥-lyase (CSE), and 3-mercaptopyruvate sulfurtransferase (3MST) along with cysteine aminotransferase (CAT) (1)(2)(3)(4)(5)(6). 3MST produces H 2 S from 3-mercaptopyruvate (3MP), which is produced from cysteine and ␣-ketoglutarate (␣-KG) by CAT that is identical with aspartate aminotransferase (4,7,8). In the central nervous system, CBS is mainly localized to astrocytes, and 3MST to neurons (4,9,10). H 2 S facilitates the induction of hippocampal long-term potentiation by enhancing the activity of NMDA receptors and induces Ca 2ϩ waves in astrocytes (2,11,12). It also relaxes smooth muscle by activating K ATP channels, regulates insulin release and induces angiogenesis (3,(13)(14)(15)(16).
In addition to its role as a signaling molecule, H 2 S has a cytoprotective effect. H 2 S protects neurons from oxidative stress by enhancing the activity of ␥-glutamylcysteine synthetase (␥-GCS) as well as the transport of cysteine and cystine, leading to reinstating the levels of GSH decreased by oxidative insults (17,18). H 2 S also protects cardiac muscle from ischemia-reperfusion injury (19). H 2 S facilitates the nuclear localization of a transcription factor, nuclear factor erythroid 2-related factor 2 (Nrf2), which increases the expression of antioxidants such as thioredoxin and heme-oxygenase1 (20). H 2 S produced by 3MST along with CAT can directly react with several cytotoxic oxidant species in mitochondria to protect cells (18,21).
Calcium ion (Ca 2ϩ ) acts as a second messenger involved in a broad spectrum of intracellular signaling pathways. A number of enzymes are regulated by Ca 2ϩ . NOS and heme oxygenase-2 are regulated by Ca 2ϩ /calmodulin (22,23). There are several enzymes that are directly regulated by Ca 2ϩ without being mediated by Ca 2ϩ -binding proteins. For examples, three intramitochondrial citrate cycle dehydrogenases; pyruvate dehydrogenase, NAD-isocitrate dehydrogenase, and oxoglutarate dehydrogenase were activated by Ca 2ϩ (24). Recently, we demonstrated H 2 S production by 3MST depends on thioredoxin and dihydrolipoic acid (DHLA) (25). However, the regulation of the activity of 3MST is not well understood.
The present study shows that the H 2 S production by 3MST and CAT is regulated by intracellular Ca 2ϩ . H 2 S, in turn, suppresses Ca 2ϩ channels by activating vacuolar type H ϩ -ATPase (V-ATPase). Under physiological conditions, H 2 S may maintain intracellular Ca 2ϩ in low levels. The regulation of Ca 2ϩ by H 2 S may be failed by the excessive levels of light, and the photoreceptor cell degeneration occurs. Even under such conditions the administration of sodium hydrosulfide (NaHS), a donor of H 2 S, suppresses photoreceptor degeneration. These observations suggest that H 2 S protects photoreceptor cells from the insult caused by excessive levels of light.
Determination of H 2 S Producing Activity-All animal procedures were approved by the National Institute of Neuroscience Animal Care and Use Committee. Homogenates of the mouse retina were prepared in the ice-cold isolation buffer containing 50 mM potassium phosphate (pH7.4), 1 mM DTT, the protease inhibitor mixture complete EDTA-free (Roche Diagnostics, Mannheim, Germany) and 10 mM EGTA, and sonicated for 10 s using a sonicator (Branson Model 450; Branson Ultrasonics, Danbury, CT). Protein concentrations were determined by Bio-Rad D c Protein Assay (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer's instructions.
The enzyme reaction was performed as described (4) with modifications. The substrate was added to 0.1 ml of homogenates or cell lysates in a 15 ml centrifuge tube and incubated at 37°C for 30 min. After adding 0.2 ml of 1 M sodium citrate buffer, pH 6.0, the mixtures were incubated at 37°C for 10 min with shaking on a rotary shaker to facilitate a release of H 2 S gas from the aqueous phase. Two ml of approximate 14.5 ml of head-space gas was applied to a gas chromatograph (GC-2014; Shimadzu, Kyoto, Japan) equipped with a flame photometric detector and a data processor C-R8A Chomatopac (Shimadzu). The concentrations of H 2 S were calculated using a standard curve of 0 to 5 nM of Na 2 S, as a source of H 2 S.
Constructs and Transient Transfection of HEK 293-F Cells-The constructs of 3MST and CAT expression plasmids were described previously (4). Transient transfection of HEK 293-F cells in suspension cultures was performed using a FreeStyle 293 Expression System (Invitrogen Life Technologies Corp., Carlsbad, CA). HEK 293-F cells efficiently express externally applied expression plasmids and grow in higher densities in suspension cultures than regular plating. For transfection, 10 g of expression plasmids were mixed with 15 l of transfection reagent 293 fectin (Invitrogen) and then added to 1 ϫ 10 7 cells in 100 ml Erlenmeyer flasks with 10 ml of FreeStyle 293 expression medium. Cells were incubated with shaking at 125 rpm on a rotary shaker at 37°C in a humid atmosphere with 10% (v/v) CO 2 . Transfection efficiency was Ͼ90%. Cells were harvested at 48 h post-transfection. The HEK 293-F cells were precipitated by centrifugation at 1000 ϫ g for 5 min. After washing with ice-cold PBS, cell pellets were resuspended in ice-cold buffer A and sonicated.
Preparation of Mouse Retinal Slices and Ca 2ϩ Imaging-The mice were kept in a room maintained under a 12 h/12 h light cycle. Mice were dark-adapted for 2 h before experiments. Subsequent manipulations were performed under dim red light. The mice were anesthetized and killed by cervical dislocation. The eyes were enucleated and immediately put in ice-cold bicarbonate-buffered saline (BBS), composed of: 120 mM NaCl, 22.6 mM NaHCO 3 , 3 mM KCl, 0.5 mM KH 2 PO 4 , 2 mM CaCl 2 , 0.5 mM MgSO 4 , and 6 mM glucose. The BBS solution (pH 7.4) was continuously bubbled with 95% (v/v) O 2 and 5% (v/v) CO 2 . After enucleation, the anterior segment of the eye including the lens was removed. The resulting eyecup was detached from the pigmented epithelium, and a section of retina was placed vitreal side down on a piece of filter paper (13 mm in diameter, Type HAWP, 0.45 m pores, Millipore, Billerica, MA). After adher-ing to the filter paper, retinal slices with 200 -250 m thickness were manually cut with a razor blade.
For Ca 2ϩ imaging, retinal slices were loaded with a cell-permeant Ca 2ϩ indicator, Calcium Green-1 AM (Invitrogen), by incubating slices for 50 min in the dark with 1 M Calcium Green-1 AM and 0.01% (v/v) CremophorEL in BBS solution. The slices were positioned in the glass-bottom dish (Matsunami) with high-pure white vaseline (Wako) for viewing the retinal layers. The dish was mounted on an upright fixed stage microscope (Leica DMLFS, Leica Microsystems). The images were acquired using a xenon lamp (Osram, Augsburg, Germany), a water immersion objective (40X; 0.80 NA; Leica Microsystems) and a CCD camera (C4742-95-12ER; Hamamatsu Photonics, Shizuoka, Japan). Excitation and emission were controlled by LEP MAC5000 filter wheel and a shutter controller (Ludl electronic products, Hawthorne, NY). Frame duration ranged from 20 -35 ms and each image was acquired at 5-s intervals. Images were acquired and analyzed using Aquacosmos software Ver. 2.0 (Hamamatsu Photonics).
Light Damage-Male ICR mice (Japan Clea), aged 8 -10 weeks were kept under controlled lighting conditions (12hlight/12h-dark). After dark adaptation for 24 h, NaHS (0.4375 mol/kg) or vehicle (PBS) was administered intraperitoneally to mice 15 min before exposure to light. Mice were dilated with 5% (w/v) phenylephrine eye drops (Kowa, Tokyo, Japan) 5 min before exposure to light. Nonanesthetized mice were exposed to while fluorescent light (1220 lm) for 2 h in cage (375 cm 2 ) with reflective interior. After light exposure, animals remained in darkness until they are analyzed.
For analysis with a light microscopy, the eyes were enucleated under dim red illumination, fixed with 4% (w/v) paraformaldehyde in PBS at 4°C overnight and immersed in 5, 15, and 30% (w/v) sucrose in PBS in order at 4°C. The eyes were then submerged into an O.C.T. compound and frozen. The retinal sections with 12-m thickness were cut by a cryostat at Ϫ20°C and stored at Ϫ80°C until staining. The sections were stained with Mayer's hematoxylin and eosin for histological analysis.
To detect the retinal cell death, TUNEL staining was performed. The sections were washed three times in PBS and incubated in PBS containing 2 g/ml proteinase K (Merck) at 37°C for 10 min. The sections were washed three times in PBS and incubated in 0.3% (w/v) hydrogen peroxide with methanol at room temperature for 5 min to block endogenous peroxidase activity. The sections were washed and incubated with terminal deoxyribonucleotidyl transferase (TdT; Invitrogen) with biotin-labeled dCTP (Invitrogen) at 37°C for 1 h. The biotin-labeled dCTP was detected colorimetrically using peroxidaseconjugated streptavidin (Nichirei) and its chromogenic substrate 3, 3Ј-diaminobenzidine (Dojindo, Kumamoto, Japan). Sections were imaged by an epifluorescence microscopy (Axiophot, Carl Zeiss, Germany) using a Plan-NEOFLUAR 40ϫ objective (Carl Zeiss). Three light-microscope images were photographed from each eye sample, and the number of TUNEL-positive cells in the outer nuclear layer was averaged for five eye samples.
Statistical Analysis-All statistical analyses of the data were performed using Microsoft Excel 2004 for Mac (Microsoft, Redmond, WA) with the add-in software Statcel2 (OMS, Saitama, Japan). Differences between two groups were analyzed with Student's t test. Differences between three or more groups were analyzed with one-way analysis of variance (ANOVA). Post hoc multiple comparisons were made using the Bonferroni/Dunn test.

RESULTS
3MST and CAT Are Localized to the Retinal Neurons-Because H 2 S-producing enzymes 3MST and CAT are localized to neurons in the brain (4), we examined the localization of both enzymes in the retina by immunohistochemistry (Fig. 1A). Both 3MST and CATs were localized to the inner plexiform layer, the outer plexiform layer, the inner nuclear layer, the outer nuclear layer and the outer segments of photoreceptors (Fig. 1,  B--I). Especially 3MST co-localized with calbindin, a specific marker for horizontal cells (Fig. 1, J-Q). Neither CBS nor CSE was found in the retina (Fig. 1, R-U).
The Production of H 2 S by 3MST and CAT Is Regulated by Ca 2ϩ -Because both 3MST and CAT are localized in the retina, it is possible that the retina can produce H 2 S. The production of H 2 S by lysates of the retina was examined. Lysates produced H 2 S in the presence of cysteine, ␣-KG and pyridoxal 5Ј-phosphate (PLP) (Fig. 2A). In the absence of ␣-KG, which is required for CAT but not for CBS nor CSE, lysates produced little H 2 S (Fig. 2A). These observations suggest that H 2 S is produced by 3MST along with CAT in the retina.
Because Ca 2ϩ regulates the activity of many enzymes, we investigated the effect of Ca 2ϩ on the activity of 3MST and CAT by measuring the amount of H 2 S produced by retinal lysates. H 2 S production was the maximum in the absence of Ca 2ϩ and suppressed by Ca 2ϩ in a dose-dependent manner (Fig. 2B). The production of H 2 S was completely suppressed at 2.9 M Ca 2ϩ . Since the intracellular concentrations of Ca 2ϩ change from ϳ600 nM in darkness to less than 10 nM during illumination in mouse retinal photoreceptor cells (28), H 2 S is likely to be produced by 3MST and CAT when photoreceptor cells are exposed to light.
Because of its role in Ca 2ϩ -dependent regulation, the involvement of calmodulin in the Ca 2ϩ regulation of H 2 S production was examined. Neither calmodulin nor a calmodulin specific inhibitor, W-7, showed any effect on the production of H 2 S in cell lysates (Fig. 2C). These observations suggest that calmodulin is not involved in the regulation of H 2 S-producing activity of 3MST and CAT.
To determine which enzyme, 3MST or CAT, is regulated by Ca 2ϩ , H 2 S production from lysates of HEK 293-F cells overexpressing 3MST and CAT was examined. To evaluate Ca 2ϩ dependence of 3MST, the amount of H 2 S was measured using 3MP as a substrate in the absence of CAT. The production of H 2 S with 3MP as a substrate was not changed by Ca 2ϩ (Fig. 2D). This observation suggests that the activity of 3MST is not regulated by Ca 2ϩ . Because 3MP is unstable and difficult to measure, the amount of H 2 S produced from cysteine and ␣-KG by combined 3MST and CAT was measured to evaluate the Ca 2ϩ dependence of CAT. The production of H 2 S was decreased by Ca 2ϩ in a concentration-dependent manner (Figs. 2E and 2F). These observations suggest that the activity of CAT is regulated by Ca 2ϩ .
H 2 S Suppresses Ca 2ϩ Influx by Activating V-ATPase-High concentrations of K ϩ evokes Ca 2ϩ influx in photoreceptor cells (29). Because H 2 S regulates Ca 2ϩ channels in astrocytes (11), the effect of H 2 S on Ca 2ϩ influx evoked by high K ϩ was examined. Na 2 S, a donor of H 2 S, suppressed Ca 2ϩ influx in the outer nuclear layer (ONL) (Fig. 3, A-C). Na 2 S alone did not induce Ca 2ϩ influx. A similar observation was made in the outer plexiform layer (OPL) (Fig. 3D). In contrast, the photoreceptor outer segment only weakly responded to high-K ϩ , and the responses were not affected by Na 2 S (Fig. 3E). These observations suggest that H 2 S may regulate voltage-dependent Ca 2ϩ influx in cells of ONL and OPL.
The center-surround organization of receptive field is one of the most important characteristics of the retinal neurons. The negative feedback from horizontal cells to photoreceptor cells plays an important role in the center-surround organization (30). Feedback from horizontal cells to photoreceptor cells is mediated by the suppression of L-type Ca 2ϩ channels on photoreceptor cells by protons released from V-ATPase on horizontal cells (31)(32)(33)(34). V-ATPase has two cysteine residues at the ATP-binding site of catalytic subunit and reversible formation of disulfide bond results in inactivation of the V-ATPase (35). Because H 2 S has reducing ability, it can activate the V-ATPase by reducing the disulfide bond in the catalytic site. We there-fore examined the type of calcium channels mediating Ca 2ϩ influx and the involvement of V-ATPase in H 2 S-induced suppression of Ca 2ϩ influx. Inhibitors specific to L-type voltage gated calcium channels, nifedipine and diltiazem, greatly suppressed Ca 2ϩ influx (Fig. 4A), indicating that L-type Ca 2ϩ channels are activated by high K ϩ in agreement with the previous study (29,36,37). Because protons released from V-ATPase cause acidification, whether or not acidification suppresses Ca 2ϩ influx was examined. The acidification from pH 7.4 to pH 7.2 suppressed the Ca 2ϩ influx (Fig. 4B). This observation suggests that V-ATPase mediates H 2 S-induced suppression of Ca 2ϩ influx.
To confirm the involvement of V-ATPase, the suppressing effect of H 2 S on Ca 2ϩ influx was investigated in the presence of inhibitors of V-ATPase, bafilomycin A1 and N-ethylmaleimide (NEM) (34). 100 nM bafilomycin A1, a specific inhibitor of a V-ATPase, abolished the suppressing effect of H 2 S (Fig. 4C). Five hundred M NEM, which masks SH-groups at the active site of V-ATPase (35), also abolished the suppressing effect of H 2 S on Ca 2ϩ influx (Fig. 4D). These observations suggest that H 2 S activates V-ATPase on horizontal cells to release protons that suppress L-type Ca 2ϩ channels on photoreceptor cells.
It was reported that GABA mediates the feedback signal from horizontal cells to photoreceptor cells (38). We examined whether or not GABA mediates the suppressing effect of H 2 S on Ca 2ϩ influx using inhibitors specific for GABA receptors, picrotoxin and bicuculline. Neither picrotoxin nor bicuculline changed the suppressing effect of H 2 S on Ca 2ϩ influx (Fig. 4, E  and F). These observations suggest that GABAergic input is not involved in the suppressing effect of H 2 S on Ca 2ϩ influx.
H 2 S Protects Retinal Photoreceptor Cells from Light-induced Degeneration-The retina is susceptible to oxidative stress because of its high consumption of oxygen and daily exposure to light. Excessive light exposure leads to photoreceptor degeneration in the retina (39). Photoreceptor cell death is an irreversible injury that is caused by various factors, such as reactive oxygen species (ROS) and elevated intracellular concentrations of Ca 2ϩ (40). Because H 2 S protects neurons from oxidative stress (17,18,21), it is possible that H 2 S protects photoreceptor cells from light-induced retinal degeneration. To examine this possibility, we investigated the effect of H 2 S on retinal degeneration caused by exposing the retina to excessive levels of light. The cytoprotective effect of H 2 S in vivo was investigated with NaHS, and NaHS as well as Na 2 S suppressed Ca 2ϩ influx in the retina (Fig. 5, A-C) (18). For these reasons we administered NaHS to mice and examined its effect on photoreceptor degeneration induced by light exposure. The photoreceptor outer segments were deteriorated (Fig. 5, E and H). In contrast, the light-induced damage was significantly suppressed in mice administered NaHS (Fig. 5, F and I). Many cells in the outer nuclear layer, in which rod inner segments are located, became TUNEL-positive following exposure to light (Fig. 5, K and N). The administration of NaHS decreased the number of TUNELpositive cells by ϳ80% relative to a control without any effect on the density of cells (Fig. 5, L and O-Q).
To confirm the protective effects of H 2 S on the retinal neurons from light-induced degeneration, we examined the levels of 8-hydroxy-2Ј-deoxyguanosine (8-OHdG), which is a product of DNA damaged by ROS. We used immunohistochemistry with an antibody against 8-OHdG. Light exposure produced a lot of 8-OHdG positive cells in the outer nuclear layer (Fig. 5T). In contrast, the number of cells positive to 8-OHdG was decreased in NaHS-treated mice (Fig. 5V). These observations suggest that H 2 S protects photoreceptor cells from light-induced retinal degeneration and oxidative stress.

DISCUSSION
The Localization of 3MST and CAT in the Retina-Both 3MST and CAT were localized to the retinal neurons, while CBS and CSE were not found, and H 2 S was not produced from cysteine in the absence of ␣-KG (Figs. 1, B-U and 2A) (41,42), indicating that the 3MST along with CAT can produce H 2 S in the retina. Although both CBS and CSE were found in salamander retina (42), the 3MST/CAT pathway is a major pathway to produce H 2 S in mammalian retina.
Ca 2ϩ Regulation on 3MST/CAT Pathway-The range of intracellular Ca 2ϩ is shifted to the lower concentrations in the retinal neurons compared with the other types of cells in which the intracellular concentrations of Ca 2ϩ are between 100 nM and 1-2 M (43). In brightness, the voltage-gated Ca 2ϩ channels are closed, and the intracellular concentrations of Ca 2ϩ in photoreceptor cells are reduced to less than 10 nM. When photoreceptor cells are hyperpolarized in brightness, a release of glutamate from photoreceptor terminals is decreased, resulting in horizontal cells being in a quiescent state. The intracellular concentrations of Ca 2ϩ in horizontal cells are maintained ϳ50 -75 nM in the resting state (44). In darkness, Ca 2ϩ ions enter the photoreceptor cells, and the intracellular concentrations of Ca 2ϩ reach ϳ600 nM (28). The present study shows that the H 2 S-producing activity of 3MST/CAT pathway is increased at low concentrations of Ca 2ϩ that is achieved in brightness (Fig. 2B).
The present study shows that Ca 2ϩ regulates the CAT activity but that calmodulin is not involved in the regulation (Fig. 2, C, E, and F). The regulation by Ca 2ϩ without the involvement of calmodulin has been demonstrated for three mitochondrial dehydrogenases; pyruvate dehydrogenase, NAD-isocitrate dehydrogenase and oxoglutarate dehydrogenase (24). These enzymes do not have any Ca 2ϩ binding domain, but they are stimulated by Ca 2ϩ . The other example is serine racemase, which is a PLP-dependent enzyme similar to CAT, is activated by Ca 2ϩ (45). In this case Ca 2ϩ binds to the site formed by two carboxylate-containing residues (glutamate and aspartate) (45,46). No such Ca 2ϩ binding site has been identified in CAT.
H 2 S Regulates Ca 2ϩ Influx-The center-surround organization is caused by the inhibitory feedback from horizontal cells (30). The proposed mechanisms for the feedback are the changes in pH in the synaptic cleft and the involvement of GABAergic neurons (31)(32)(33)(34)38). The present study shows that H 2 S activates V-ATPase to suppress high K ϩ evoked Ca 2ϩ influx mediated by L-type Ca 2ϩ channels (Fig. 4, A-D). The influence of pH changes caused by Na 2 S and NaHS on the suppression of Ca 2ϩ influx is negligible for the following reasons. 1) Because bicarbonate buffered saline (BBS) contains 22.6 mM NaHCO 3 continuously bubbled with 95% (v/v) O 2 and 5% (v/v) CO 2 , concentrations less than 10 M Na 2 S or NaHS do not have any effect on pH. 2) Na salts shift pH, if any, to alkaline, while suppression of Ca 2ϩ influx occurs by acidification. Because photoreceptor cells and horizontal cells have 3MST and CAT to produce H 2 S, it is possible that H 2 S activates V-ATPase in horizontal cells to release protons to the synaptic cleft that suppress Ca 2ϩ channels in photoreceptor cells.
GABA receptor inhibitors did not affect the suppression of Ca 2ϩ influx induced by H 2 S (Fig. 4, E and F). GABAergic input from horizontal cells to cone photoreceptor cells is present but its contribution is weak and limited, and picrotoxin could not completely suppress the cone receptive field surround (32,47). These observations suggest that GABA may not mediate H 2 Sinduced suppression of Ca 2ϩ influx.
Ca 2ϩ Homeostasis and Photoreceptor Damage-H 2 S suppresses the elevation of Ca 2ϩ in the photoreceptor cells by activating V-ATPase in horizontal cells and maintains Ca 2ϩ homeostasis (Figs. 3 and 4). Intracellular concentrations of Ca 2ϩ are increased during photoreceptor apoptosis, but the L-type Ca 2ϩ channel blocker diltiazem prevents light-induced photoreceptor cell death (48). The lack of the V-ATPase a3 subunit causes retinal degradation in mice (49). These observations suggest that the failure of Ca 2ϩ homeostasis, which is regulated by Ca 2ϩ channels and V-ATPase, by excessive levels of light may cause retinal cell degeneration. The regulation of Ca 2ϩ and cytoprotective effect of endogenous H 2 S may fail when photoreceptor cells are exposed to excessive levels of light. Even under such conditions the administration of H 2 S protects cells from lightinduced degeneration.

CONCLUSION
H 2 S is produced by 3MST along with CAT in retinal neurons and may modulate the negative feedback from horizontal cells to photoreceptor cells by regulating Ca 2ϩ influx through the activation of V-ATPase. Because H 2 S protects retinal neurons from light-induced degeneration, the enhancement of 3MST/ CAT pathway or the administration of H 2 S may have clinical benefit for diseases with retinal cell degeneration.