Retinal degeneration-3 protein promotes photoreceptor survival by suppressing activation of guanylyl cyclase rather than accelerating GMP recycling

Abstract

Retinal degeneration-3 protein (RD3) deficiency causes photoreceptor dysfunction and rapid degeneration in the rd3 mouse strain and in human Leber's congenital amaurosis, a congenital retinal dystrophy that results in early vision loss. However, the mechanisms responsible for photoreceptor death remain unclear. Here, we tested two hypothesized biochemical events that may underlie photoreceptor death: (i) the failure to prevent aberrant activation of retinal guanylyl cyclase (RetGC) by calcium-sensor proteins (GCAPs) versus (ii) the reduction of GMP phosphorylation rate, preventing its recycling to GDP/GTP. We found that GMP converts to GDP/GTP in the photoreceptor fraction of the retina ∼24-fold faster in WT mice and ∼400-fold faster in rd3 mice than GTP conversion to cGMP by RetGC. Adding purified RD3 to the retinal extracts inhibited RetGC 4-fold but did not affect GMP phosphorylation in wildtype or rd3 retinas. RD3-deficient photoreceptors rapidly degenerated in rd3 mice that were reared in constant darkness to prevent light-activated GTP consumption via RetGC and phosphodiesterase 6. In contrast, rd3 degeneration was alleviated by deletion of GCAPs. After 2.5 months, only ∼40% of photoreceptors remained in rd3/rd3 retinas. Deletion of GCAP1 or GCAP2 alone preserved 68% and 57% of photoreceptors, respectively, whereas deletion of GCAP1 and GCAP2 together preserved 86%. Taken together, our in vitro and in vivo results support the hypothesis that RD3 prevents photoreceptor death primarily by suppressing activation of RetGC by both GCAP1 and GCAP2 but do not support the hypothesis that RD3 plays a significant role in GMP recycling.

Keywords: GCAP; GMP; RD3; RetGC; calcium-binding proteins; congenital blindness; cyclic GMP; guanylate cyclase (guanylyl cyclase); photoreceptor; retinal degeneration; signal transduction; vision.

Figure 1

Two hypothetical roles of RD3 in preventing photoreceptor degeneration.A, Hypothesis 1: RD3 is required to suppress aberrant activation of RetGC cyclase by GCAPs (17, 18, 19). In normal photoreceptors, the membrane guanylyl cyclase (RetGC) is delivered with the help of RD3 to the outer segment, where it produces cGMP for phototransduction. While in the inner segment, the RetGC is suppressed by RD3 blocking the cyclase activation by GCAPs. In RD3-deficient photoreceptors, RetGC content in the outer segment is strongly reduced, but the cyclase remaining in the inner segment becomes unprotected against activation by GCAPs, and this triggers the apoptotic process. B, Hypothesis 2: RD3 accelerates GMP phosphorylation by guanylate kinase (20). In the light, RetGC rapidly converts GTP to cGMP, and then PDE6 converts cGMP to 5’GMP. RD3 stimulates guanylate kinase activity to convert GMP to GDP and then back to GTP. The RD3-deficient photoreceptors fail to phosphorylate GMP; therefore, the RetGC/PDE6 pathway depletes the GDP/GTP pool and thus causes photoreceptor degeneration (20). GCAP, guanylyl cyclase activating protein; RD3, retinal degeneration 3 protein; RetGC, retinal membrane guanylyl cyclase; PDE6, cGMP phosphodiesterase 6.

Figure 2

Deletion of calcium-sensor activators of guanylyl cyclase rescues rd3/rd3 photoreceptors.A, representative retinal morphology at 2.5 months of age in wildtype, rd3/rd3, GCAP1^-/-rd3/rd3, GCAP2^-/-rd3/rd3, and GCAP1,2^−/−rd3/rd3; hereafter, the retinal histological layers are referred to as RPE (retinal pigment epithelium), ROS (rod outer segments), ONL (outer nuclear layer), OPL (outer plexiform layer), INL (inner nuclear layer), IPL (inner plexiform layer), and GCL (ganglion cell layer). The photoreceptor layer thickness between the retinal pigment epithelium and outer plexiform layer is highlighted in yellow. B, photoreceptor nuclei count in the outer nuclear layer at 2.5 months in wildtype (◇), rd3/rd3 (○), GCAP1^-/-rd3/rd3 (▲), GCAP2^-/-rd3/rd3 (□), and GCAP1,2^−/−rd3/rd3 (♦); each data point is from a different mouse. The differences were tested using ANOVA (F = 158, p < 0.0001); the p-values shown in the graph are from the Tukey’s HSD post hoc comparison (hereafter, confidence level 99%, alpha = 0.01). GCAP, guanylyl cyclase activating protein.

Figure 3

GMP phosphorylation assay. The GMP phosphorylation was assayed as described in the Experimental Procedures. A, The [³H]GMP was produced by converting [³H]cGMP by PDE6, and the resultant preparation (analyzed by TLC on PEI cellulose plates) was used as the substrate in the GMP phosphorylation assay. B, Separation of GTP, GDP, and GMP using TLC. C, Incubation of [³H]GMP with the retinal extract converts it to the phosphorylated forms, GDP + GTP. D, E, GMP phosphorylation yield was directly proportional to the amount of retinal homogenate (D) and was linear with time (E) in the conditions of the assay; mean ± SD, n = 3 technical repeats. PDE6, cGMP phosphodiesterase 6; PEI, polyethylenimine.

Figure 4

GMP phosphorylation rate in RD3-deficient photoreceptors is similar to that in normal photoreceptors.A, The representative morphology of the retina: wildtype (Rd3^+/+) at 1 month of age, rd3/rd3 (Rd3^-/-) at 1 month of age, line 362 (30, 31) (Rd3^+/+) at 7 months, and rd3/rd3 (Rd3^-/-) at 7 months. Note the reduction of the photoreceptor-layer thickness in rd3/rd3 and the complete lack of photoreceptors in the aged line 362 and rd3/rd3. The photoreceptor layer is marked yellow, and the inner retina layers are marked blue. B and C, GMP phosphorylation in whole-retina homogenates from wildtype (Rd3^+/+) (○), rd3/rd3 mice (Rd3^-/-) (◇), in the inner Rd3^+/+ retina from the aged 7 months line 362 mice (●), and the inner Rd3^-/- retina from the aged rd3/rd3 (♦) mice; not corrected for the reduction of photoreceptor count (B) and corrected for the ∼29% loss of photoreceptor count in rd3/rd3 retinas at 1 month of age (C); the p-values in the graph are from the ANOVA (F = 107, p < 0.0001)/Tukey’s HSD post hoc test. The activities in the inner Rd3^+/+retina (line 362) or the inner Rd3^-/- (aged 7 months rd3/rd3 retinas) were subtracted from the respective total activities in the wildtype and the young rd3/rd3 retinas to assess the GMP phosphorylating activity that belonged to the layer of photoreceptors (summarized in Table 1). The sample of the rd3/rd3 retinas marked with an asterisk was subsequently used to test the effect of the addition of purified recombinant RD3 to the assay (see Fig. 5). See the Experimental Procedures section for other details. RD3, retinal degeneration 3 protein.

Figure 5

RD3 strongly inhibits RetGC activity in the retina but does not affect GMP phosphorylation.A, Guanylyl cyclase activity assay. Purified recombinant RD3 was added to the retinal homogenates in the presence of 10 mM MgCl₂ and 2 mM EGTA; right inset: Coomassie-stained 15% SDS-PAAG of the purified RD3; left inset: RetGC activity in the retina in the absence (●) and in the presence (○) of 5 μM RD3. B, GMP phosphorylation activity in Rd3^+/+ wildtype (●,○), rd3/rd3 (♦,◇), and Rd3^+/+ line 362 inner retina (■, □) in the absence (●,♦, ■) or in the presence (○,◇, □) of 5 μM recombinant RD3; there was no detectable guanylate kinase activity in the RD3 preparation. The p-values are from Student’s unpaired/unequal variance t test. RD3, retinal degeneration 3 protein; RetGC, retinal membrane guanylyl cyclase.

Figure 6

Degeneration of the photoreceptor layer is unaffected in 2.5-month-old rd3/rd3 mice reared in the dark.A, Prolonged dark adaptation does not improve ERG responses in rd3/rd3 retinas. Dark-adapted rod-cone full-field ERG response to a bright flash (5.4 ×10⁵ photons/rod), averaged from different mice in each series (solid lines; gray error bars—standard deviation). The a-wave amplitude in rd3/rd3 mice reared in the dark and under 12h dark/12h light cycle was equally suppressed (82 ± 14 μV, and 83 ± 24 μV, n = 5; p = 0.9992) when compared with wildtype (353 ± 62 μV, n = 6; p < 0.0001, ANOVA (F = 79, p < 0.0001)/Tukey’s HSD). B, Representative morphology of the retina, top to bottom: wildtype and rd3/rd3 reared for 2.5 months under 12h dark/12h light cycle, and rd3/rd3 mice reared for 2.5 months in the dark. C. Photoreceptor nuclei count in wildtype (◇) and rd3/rd3 retinas (●,○) from mice reared under the cyclic light conditions (○,◇) or in the dark (●). All data points are from different mice; ANOVA (F = 382; p < 0.0001); the p-values in the graph are from the Tukey’s HSD post hoc test. ERG, electroretinography.