Rhodopsin kinase activity modulates the amplitude of the visual response in Drosophila

Abstract

A feature shared between Drosophila rhodopsin and nearly all other G protein-coupled receptors is agonist-dependent protein phosphorylation. Despite extensive analyses of Drosophila phototransduction, the identity and function of the rhodopsin kinase (RK) have been elusive. Here, we provide evidence that G protein-coupled receptor kinase 1 (GPRK1), which is most similar to the β-adrenergic receptor kinases, G protein-coupled receptor kinase 2 (GRK2) and GRK3, is the fly RK. We show that GPRK1 is enriched in photoreceptor cells, associates with the major Drosophila rhodopsin, Rh1, and phosphorylates the receptor. As is the case with mammalian GRK2 and GRK3, Drosophila GPRK1 includes a C-terminal pleckstrin homology domain, which binds to phosphoinositides and the Gβγ subunit. To address the role of GPRK1, we generated transgenic flies that expressed higher and lower levels of RK activity. Those flies with depressed levels of RK activity displayed a light response with a much larger amplitude than WT. Conversely, the amplitude of the light response was greatly suppressed in transgenic flies expressing abnormally high levels of RK activity. These data point to an evolutionarily conserved role for GPRK1 in modulating the amplitude of the visual response.

G protein-coupled receptors (GPCRs) are critical for a diversity of processes, which in excitable cells range from vision to analgesia, neuronal differentiation, cardiac function, and synaptic transmission. A key aspect of GPCR signaling is the ability to rapidly adjust receptor activity in response to agonist stimulation. Such agonist-dependent modulation appears to be accomplished in part through phosphorylation of the receptor (1–4). Mammals express >1,000 GPCRs; yet, the human and rodent genomes appear to encode only seven G protein-coupled receptor kinases (GRKs or GPRKs) (3).

GRKs fall into three subgroups, one of which includes the two visual GRKs, GRK1 and GRK7 (1–3). The two nonvisual GRK subgroups consist of GRK2/GRK3, which phosphorylate the β-adrenergic receptor, and GRK4/GRK5/GRK6. GRKs include a central catalytic region and an N-terminal RGS-like domain, but differ primarily in the C-terminal region. A distinguishing feature of the GRK2/GRK3 is a C-terminal pleckstrin homology (PH) domain, which binds both phosphoinositides (PIs) and the Gβγ subunit. In addition, GRK2/GRK3 bind the Gα_q subunit through the N-terminal RGS-like domain, which inhibits the interaction between Gα_q and phospholipase Cβ (5). Interestingly, some functions of GRKs appear to be independent of the catalytic activity (6), which could potentially be mediated through the N- or C-terminal domains.

The biological roles of GRKs have been characterized in a variety of systems. The first GRK to be identified is rhodopsin kinase (RK; also GRK1) (7), and mutations in this enzyme cause defects in the kinetics of deactivation and an increase in the amplitude of the light response (8–10). Defects in the activities of nonvisual GRKs, typically result in reductions in desensitization (2), the phenomenon by which signaling is diminished upon prolonged or repeated exposure to an agonist. Mutations in both visual and nonvisual GRKs can also result in supersensitivity to agonist stimulation (10–13).

Drosophila visual transduction has served as a paradigm to characterize G protein-coupled neuronal signaling, owing to the tractability of fly genetics (14). As in mammals, light-activated rhodopsin is phosphorylated and interacts with a protein, arrestin, which facilitates deactivation of the receptor. However, unlike mammalian phototransduction, light activation in Drosophila is coupled to stimulation of phospholipase C rather than a cGMP-phosphodiesterase.

Two Drosophila genes encoding putative GRKs (GPRK1 and GPRK2) were isolated more than a decade ago (15), but were dismissed as candidate RKs due in part to their much greater similarity to mammalian nonvisual GRKs than to mammalian RKs. Subsequent analysis of GPRK2 demonstrated that it is expressed in the ovaries and required for egg morphogenesis (16). No additional GRK-related genes appear to be encoded in flies. Although the Drosophila RK gene has not been identified, RK activities has been characterized biochemically in larger fly species (17, 18). RKs have been cloned from octopus and squid and shown to bear much greater similarity to the β-adrenergic receptor kinases, GRK2/3 and GPRK1, than to mammalian RKs (19, 20). However, the physiological roles of rhodopsin phosphorylation by these or any other invertebrate RK are not known.

In the current article, we provide evidence that GPRK1 is the Drosophila RK. Consistent with this conclusion, we found that GPRK1 was enriched in photoreceptor cells, interacted with rhodopsin, and phosphorylated the major rhodopsin. In addition, GPRK1 bound to PIs as well as to the Gβγ subunit. Of primary importance here, we found that transgenic flies that displayed a decrease in RK activity in vivo exhibited a much larger photoresponse than WT. Conversely, the amplitude of the photoresponse was diminished in transgenic flies displaying higher levels of RK activity. We conclude that the physiological function of the Drosophila RK is to modulate the amplitude of the light response. Given the known roles, biochemical characteristics, and domain organizations of mammalian GRKs, our results point to striking similarities between GPRK1 and mammalian β-adrenergic receptor kinases.

Materials and Methods

Fly Stocks and Germ-Line Transformation. The following fly strains were reared at 25°C under a 12-h light/12-h dark cycle without or with heat shock (30-min heat shocks at 37°C twice per day for 9 days) as indicated: cn bw (wt), cn bw P[w⁺, hs-gprk1] (ogprk1), cn bw;P[w⁺, hs-gprk1^K220R](ogprk1^K220R), sine oculis (so), rdgA^P38 (rdgA), w;;ninaE^P332 (ninaE), y w;;rdgC³⁰⁶ (rdgC), and ry⁵⁰⁶ P[ry⁺,Gprk2⁶⁹³⁶] (gprk2). The DNA constructs used for the germ-line transformation and other assays in this study are described in Supporting Materials and Methods, which is published as supporting information on the PNAS web site.

Northern Blots, Western Blots, and Coimmunoprecipitation (Co-IP). gprk1 and rp49 DNAs were labeled with [α-³²P]dATP and used to probe Northern blots containing 5 μg poly(A)⁺ mRNAs. For Western blots, 20 heads or three bodies (per 100 μl) were homogenized in TBS [20 mM Tris/150 mM NaCl/1× protease inhibitor mixture (Sigma), pH 7.5] containing 1% Triton X-100 (TBST) and centrifuged for 5 min at 18,000 × g. The supernatants were fractionated by SDS/PAGE, transferred to a poly-(vinylidene difluoride) membrane, and probed with α-GRK2 antibodies. A description of the antibodies is included in Supporting Materials and Methods. Signals were detected by using an ECL kit (PerkinElmer) or ¹²⁵I-labeled protein A (NEN). Separate blots containing the same extracts were probed with the α-transient receptor potential (TRP) or α-myc antibodies. For quantification (see Fig. 2), the membranes were exposed to a BAS-III imaging plate with a PhosphorImager (BAS-1500, Fuji Film). Co-IPs were performed as described (21) with modifications (see Supporting Materials and Methods).

Fig. 2.

Expression levels of GPRK1 in transgenic lines. (A) A Western blot of head extracts from WT (wt: cn bw) as well as ogprk1 and ogprk1^K220R transgenic flies was probed with α-GRK2 antibodies. Extracts were prepared from flies with (+) or without (-) heat shock treatment. Parallel blots were probed with α-myc and α-TRP antibodies. (B) Quantification of GPRK1 was determined with a PhosphorImager. The amount of GPRK1 from WT (-) heat shock is defined as 100%. SEMs are indicated (n = 3).

In Situ Hybridizations. In situ hybridizations on sections of adult fly heads were performed as described (21).

Phosphorylation of Rh1. Five fly heads added to 100 μl of buffer A (TBS plus 1 mM DTT/3 mM MgCl₂/1 mM EGTA) were illuminated for 20 sec with blue light and homogenized under a red photographic safety light, and an additional 100 μl of buffer A containing 0.1 mM [γ-³²P]ATP (ICN) was added. The homogenates were incubated for 10 min at 25°C in the dark and centrifuged at 20,000 × g for 10 min. The pellets were washed once with 500 μl of buffer A, dissolved in 30 μl of SDS sample buffer, and fractionated by SDS/PAGE. The dried gels were analyzed by autoradiography or with a PhosphorImager. Western blots containing the same samples were probed with α-Rh1 antibodies.

GST Pull-Down Assays. DNAs encoding Drosophila eye-enriched Gβ and Gγ were used to prepare the ³⁵S-methionine-labeled probe by coupled in vitro transcription/translation (Promega). To perform the pull-down assays, ≈2 μg of each GST–GPRK1 fusion protein was coupled to 30 μl of glutathione beads for 30 min and incubated with ³⁵S-labeled Gβ for 2 h in TBST buffer. The beads were pelleted and washed four times with TBST buffer. The bound proteins were eluted in 2× SDS sample buffer and resolved by SDS/PAGE. The dried gels were exposed to x-ray film.

Arrestin Pelleting Assay. Arr2 binding assays were performed as described (22) with minor modifications. Six fly heads were added to TBS buffer containing 5 mM DTT. The heads were exposed to 20 sec of blue light or 20 sec of blue light followed by 20 sec of orange light, homogenized in the dark, and centrifuged at 20,000 × g for 5 min. The pellet and supernatant fractions were separated in the dark and subjected to SDS/PAGE and Western analysis with antibodies against Arr2 and Rh1. Quantification of the Arr2 was determined by using ¹²⁵I-labeled protein A and a PhosphorImager.

Phospholipid Binding Assays. GPRK1 and Drosophila eye-enriched Gα_q were labeled with ³⁵S-methionine by coupled transcription/translation. The probe (5 μl) was mixed with either PI(4)P, PI(5)P, or PI(4,5)P₂ beads or control beads (40 μl of slurry; Echelon Biosciences, Salt Lake City) in 400 μl of TBST buffer plus 1× protease inhibitor mixture (Sigma). After a 2-h incubation at 4°C, the beads were pelleted by centrifugation and washed three times with TBST. The washed beads were pelleted, and 40 μl of 2× SDS sample buffer was added. The eluted proteins were resolved by SDS/PAGE, and the dried gels were exposed to BioMax MR film (Kodak) and analyzed with a PhosphorImager. PIP-Strips (Echelon Biosciences) binding assays were performed as described (23) (also see Supporting Materials and Methods).

Electroretinogram (ERG) Recordings. ERG recordings were performed as described (24) with white-eyed (cn bw) flies.

Results

Drosophila GPRK1 is an excellent RK candidate as it is much more similar to the squid and octopus RKs than to Drosophila GPRK2. Overall, GPRK1 is 61–62% identical to these invertebrate RKs, whereas GPRK2 shares 39% identity (Figs. 8 and 9A, which are published as supporting information on the PNAS web site). As is the case with the squid and octopus RKs (19, 20), GPRK1 bears greater resemblance to the β-adrenergic receptor kinases (GRK2/3) than to mammalian RKs (Figs. 8 and 9A). Furthermore, GPRK1, but not GPRK2, contains a C-terminal PH domain (Fig. 9B), which is also found in the β-adrenergic receptor kinases (GRK2 and GRK3) and the squid and octopus RKs.

To determine whether gprk1 is expressed in the adult eye, we performed Northern blots. Samples were prepared from WT heads and bodies and from the heads of a mutant referred to as sine oculis, which does not develop eyes (Fig. 1A). The gprk1 RNA was enriched in the eyes because the RNA levels were higher in WT heads than in sine oculis heads or WT bodies (Fig. 1 A). However, as is the case with the gene encoding the rhodopsin phosphatase, rdgC (21, 24, 25), gprk1 is not specifically expressed in the eyes. To directly examine the expression pattern of gprk1 RNA, we performed in situ hybridizations to adult heads. Consistent with the Northern blot data, gprk1 was enriched in the retina (Fig. 1C).

Fig. 1.

Expression pattern of gprk1. (A) Northern blot analysis. mRNAs from WT heads (WH), sine oculis heads (SH), and WT bodies (WB) were probed with ³²P-labeled gprk1 and rp49 DNAs. RNA markers (kb) are indicated. (B) Western blot analysis. α-GRK2 antibodies were used to probe a blot containing extracts from WT bodies (WB), WT heads (WH), sine oculis heads (SH), and heads from rdgA flies. Parallel blots were probed with α-TRP and α-tubulin antibodies. Size markers (kDa) are indicated. (C) Spatial distribution of gprk1 RNA in adult fly heads. In situ hybridizations to frontal sections were performed with antisense and sense gprk1 probes. A phase-contrast image of the same section hybridized with the gprk1 sense probe is shown. B, brain; M, medulla; L, lamina; R, retina.

We examined whether the GPRK1 protein was eye-enriched by performing Western blots using extracts prepared form WT and sine oculis heads. We used antibodies raised against mammalian GRK2 as this protein was significantly more homologous to GPRK1 than to any other Drosophila protein, including GPRK2 (Figs. 8 and 9A). A single protein of the predicted molecular weight was detected in WT heads (Fig. 1B) and at much higher levels in transgenic flies overexpressing GPRK1 (see below). GPRK1 was expressed at lower levels in sine oculis than in WT heads and below the level of detection in WT bodies (Fig. 1B). Because the antibodies were ineffective in immunostaining of tissue sections (data not shown), we used a biochemical approach to address whether GPRK1 was expressed in photoreceptor cells. We found that in a mutant, rdgA, which underwent severe photoreceptor cell degeneration (26), the concentration of GPRK1 was reduced to approximately the same levels as in sine oculis heads (Fig. 1B). These data indicate that most of the GPRK1 present in the eyes was in photoreceptor cells.

Currently, the Berkeley Drosophila Genome Project has not cloned or sequenced the gprk1 genomic region, and there are no known P element insertions near this locus. Thus, it is possible that gprk1 is present in a heterochromatic region, complicating attempts to generate loss-of-function mutations through imprecise excision of P elements or homologous recombination. Therefore, to address the function of this locus, we generated transgenic flies that expressed a mutant form of GPRK1, which contained the corresponding lysine (residue 220; Figs. 8 and 9B) to arginine mutation in the ATP binding motif shown to ablate the kinase activity in mammalian GRK2 (27). This derivative of GPRK1 was fused to a myc tag (ogprk1^K220R) to facilitate biochemical analyses (see below). Both the WT and mutant forms of gprk1 (ogprk1 and ogprk1^K220R, respectively) were expressed by linking the coding regions to the hsp70 promoter. After heat shock treatments, the levels of GPRK1 were significantly higher in ogprk1 and ogprk1^K220R than in nontransgenic flies, as determined by probing Western blots with anti-GRK2 antibodies (Fig. 2). The ogprk1^K220R protein was also detected with anti-myc antibodies (Fig. 2 A). The heat shock treatments did not appear to affect the morphology of the photoreceptor cells in either the WT or transgenic flies (Fig. 10, which is published as supporting information on the PNAS web site).

To test whether phosphorylation of the major rhodopsin (Rh1) was affected in ogprk1 and ogprk1^K220R, we examined the levels of the phosphorylated Rh1 in light-treated fly head extracts. The phosphorylated band detected in this assay was identified as Rh1 as it migrated at the known position of the Rh1 monomer and was absent in Rh1 mutant flies (ninaE^P332; Fig. 3A). Rh1 was hyperphosphorylated in ogprk1 flies, whereas phosphorylation of Rh1 was significantly reduced in ogprk1^K220R (Fig. 3).

Fig. 3.

Phosphorylation of Rh1 in transgenic flies. (A) Heads from the indicated stocks were exposed to blue light, homogenized and incubated with [γ-³²P]ATP. (Upper) After centrifugation, the pellets were fractionated by SDS/PAGE and exposed to film. (Lower) A Western blot containing the same amounts of each sample was probed with α-Rh1 antibodies. (B) Relative phosphorylation levels of Rh1 were measured with a PhosphorImager. Phosphorylated Rh1 in WT was defined as 100%. The error bars represent SEMs (n = 4).

To examine whether GPRK1 physically interacted with Rh1 in vivo, we performed co-IP with fly head extracts. Because the anti-GRK2 antibodies were ineffective for IPs, we immunoprecipitated myc-GPRK1^K220R from the ogprk1^K220R flies and probed the Western blots with antibodies to Rh1 and a variety of other signaling proteins. The mutation in myc-GPRK1^K220R could potentially stabilize interactions with substrate(s) for the protein kinase because it eliminates phosphotransferase activity in mammalian GRK2 (27). Rh1 was present in the immune complexes, suggesting that GPRK1 interacted with rhodopsin (Fig. 4A). We did not detect specific signals corresponding to most other phototransduction proteins, even after very long exposures of the Western blots. These include the Gα_q, RDGC (rhodopsin phosphatase), NORPA (phospholipase C), INAC (PKC), Arr2, INAD, TRP, or NINAC (myosin III) (Fig. 4A and data not shown). Because GRK2/3 bound to Gα_q in a dependent manner (5), we tested whether GPRK1 and Gα_q coimmunoprecipitated after adding 30 μM to the extracts. We also performed in vitro pull-down assays as described (5). However, we did not find evidence that GPRK1 bound to Gα_q (data not shown).

Fig. 4.

GPRK1 interacts with Rh1 and Gβγ. (A) Co-IPs between GPRK1^K220R and signaling proteins in vivo. Head extracts were prepared from ogprk1^K220R (R) or WT (wt) flies. IPs were performed with α-myc antibodies or nonimmune serum (NIS), and Western blots were probed with antibodies to the indicated proteins. (B) Gβ interacts with GPRK1 in vitro. Full-length ³⁵S-labeled Gβ and Gγ were incubated with GST, GST–GPRK1, GST–GPRK1^K220R, GST–GPRK1 N (amino acids 1–193), GST–GPRK1 M (amino acids 188–454; catalytic domain), and GST–GPRK1 C (amino acids 450–700; PH domain). The inputs represent 5% of the total. The GST fusions (Lower) and bound [³⁵S]Gβ (Upper) were fractionated by SDS/PAGE and either stained with Coomassie blue or detected by exposure to film.

We found that Gβγ coimmunoprecipitated with GPRK1^K220R as the Gβ signal was detected in ogprk1^K220R but not in WT extracts (Fig. 4A). To address whether there was a direct interaction between GPRK1 and Gβγ, we performed GST pull-down assays. We found that ³⁵S-labeled, full-length Gβγ associated with GST-GPRK1 and GST-GPRK1^K220R but not with GST alone (Fig. 4B). Gβγ bound to the C-terminal portion of GPRK1, which contained the PH domain, but not to the N-terminal or central regions of GPRK1 (Fig. 4B).

Phosphorylation of Rh1 has been suggested to stabilize arrestin/Rh1 complexes (28). Therefore, we examined whether the association and/or dissociation of the major arrestin (Arr2) with Rh1 were affected in ogprk1 and ogprk1^K220R flies. To characterize the Arr2/Rh1 interaction, we used an arrestin pelleting assay (29, 30). As expected, in WT most of the Arr2 bound to Rh1 after exposure to blue light, which stabilizes the active form of Rh1 and promotes the Rh1/Arr2 interaction (Fig. 5). After an identical blue light treatment, slightly less Arr2 bound to Rh1 in ogprk1^K220R than in WT flies (Fig. 5). Surprisingly, the proportion of bound Rh1/Arr2 was even lower in ogprk1 than in either WT or ogprk1^K220R. Because the effects on the level of Rh1 phosphorylation were opposite in ogprk1 and ogprk1^K220R flies (Fig. 3), these data suggested that the GPRK1 might have a phosphorylation-independent role affecting Rh1/Arr2 binding. In contrast to these results, we did not detect any differences between fly strains in the dissociation of the Rh1/Arr2 complexes promoted by exposure to orange light (Fig. 5).

Fig. 5.

Binding and release of Arr2 from Rh1 in gprk1 transgenic flies. (A) Relative proportion of Arr2 proteins in the supernatant and pellet fractions after exposure to light. Blue light (480 nm) converts Rh1 to metaRh1, resulting in Arr2 binding. Orange light (580 nm) converts metaRh1 back to Rh1, resulting in dissociation of Arr2. Arr2 bound to Rh1 in the pellet (P). Unbound Arr2 remains in the supernatant (S). Isolated heads were exposed to either blue light (B) or blue light followed by orange light (BO). Western blots were probed with α-Arr2 and α-Rh1 antibodies. (B) Percent of Arr2 bound to Rh1-containing membranes after blue light (B) or blue light followed by orange light (BO): wt, B = 89.2 ± 9.2%, BO = 36.6 ± 3.5%; oGprk1,B = 62.5 ± 4.3%, BO = 39.1 ± 8.7%; oGPRK1^K220R,B = 75.9 ± 1.4%, BO = 35.0%± 6.4%. Error bars are the SEM (n = 3).

Because GPRK1 has a C-terminal PH domain (Fig. 9B), which could potentially bind to PIs, we probed a membrane containing various PIs (PIP-Strip) with the full-length GPRK1 fused to GST (GST–GPRK1). GST–GPRK1 bound to several PIs, whereas GST did not (Fig. 6A). We mapped the PI-interacting region in GPRK1 to the PH domain, as the C-terminal region (GST–GPRK1 C) but not the N-terminal (GST–GPRK1 N) or middle (GST–GPRK1 M) domains bound to PIs (Fig. 6A). To confirm the PI/GPRK1 interaction, we performed pull-down assays with PIP beads. We chose PI(4)P and PI(5)P beads because these PIs bound strongly to GST–GPRK1 and GST–GPRK1 C in the filter binding assay (Fig. 6A). We also used PI(4,5)P₂ because it is one of the most abundant PIs in vivo. We found that ³⁵S-labeled-GPRK1 bound to all three PIP beads but not to control beads (Fig. 6B). Consistent with PIP-Strip assay, GPRK1 interacted more strongly with PI(4)P and PI(5)P than with PI(4,5) beads (Fig. 6 B and C). Gα_q did not interact with the PIP beads (Fig. 6B).

Fig. 6.

Interaction of GPRK1 with PIs. (A) Phospholipid filter binding assay. Purified GST or the indicated fusion proteins were incubated with PIP-Strips, and bound proteins were detected with α-GST antibodies. (B) PIP bead binding assay. Full-length GPRK1 or Gα_q (labeled with ³⁵S-methionine) was incubated with PI(4)P, PI(5)P, PI(4,5)P₂ beads or agarose beads alone (con). The bound probes were eluted, resolved by SDS/PAGE, and exposed to film. (C) Quantification of the bound GPRK1 (from B). The amount of GPRK1 bound to PI(4,5)P₂ beads was set at 100%. SEMs are indicated (n = 3).

At least three possible roles for GPRK1 in Drosophila phototransduction can be envisioned. First, GPRK1 phosphorylates Rh1 and thereby promotes termination of signaling through the receptor. Second, phosphorylation of Rh1 by GPRK1 may control the amplitude of the visual response. Third, GPRK1 may have a phosphorylation-independent role, which could potentially be mediated through binding to Gβγ or PIs. Either of the first two proposals would predict that ogprk1 and ogprk1^K220R would display opposite phenotypes, whereas the third potential role would predict that ogprk1 and ogprk1^K220R should display similar phenotypes.

To distinguish between the potential roles for GPRK1, we performed ERGs, which measure the summed responses of all retinal cells to light. WT flies displayed a corneal negative response to orange light, which returned to baseline after cessation of the light (Fig. 7A). We found that ogprk1^K220R flies, which expressed lower RK activity than WT, showed a pronounced increase in the amplitude of the ERG (Fig. 7 C and E). Conversely, the ERG amplitude was much smaller than WT in ogprk1 flies, which displayed a higher level of RK activity (Fig. 7 B and E). Consistent with these results, disruption of the rhodopsin phosphatase, rdgC, resulted in a similar decrease in the ERG amplitude as an increase in RK activity (Fig. 7 D and E). Although the amplitudes of the ERGs were affected in the gprk1 transgenic flies, there was no apparent effect on the termination of the photoresponse (Fig. 7 A–C). Termination was also normal in gprk1 transgenic flies challenged with consecutive pulses of blue light, which results in the production of stable metarhodopsin (Fig. 11, which is published as supporting information on the PNAS web site). In contrast to the ERG phenotype resulting from a decrease in the activity of GPRK1, the ERG amplitude was normal in gprk2 mutant flies (data not shown), which are female-sterile and display reduced viability (16). These data indicate that the primary role of GPRK1 kinase activity in photoreceptor cells is to modulate the amplitude of the visual response.

Fig. 7.

Altered ERG amplitudes in gprk1 transgenic flies. ERGs were performed on white-eyed WT (wt) (A), ogprk1 (B), ogprk1^K220R (C), and rdgC (D) flies. ERGs were obtained in the presence and absence of heat shock, except for rdgC. Heat shock caused lethality of rdgC flies. The flies were stimulated for 5 sec with orange (70 mW/cm²) light. (E) Average amplitudes - or + heat shock. Error bars indicate the standard errors (n = 9–17).

Discussion

Before the current article, the function of a RK had not been addressed in any invertebrate organism to our knowledge. We conclude that GPRK1 is the Drosophila RK as it is enriched in photoreceptor cells and interacts with and phosphorylates Rh1. As is the case with the octopus and squid RKs (19, 20), GPRK1 bears much greater similarities to the β-adrenergic receptor kinases, GRK2 and GRK3, than to mammalian RKs. In addition, we found that the C-terminal PH domain in GPRK1 bound to PIs and the Gβγ subunit, consistent with other β-adrenergic receptor kinases (2, 3). Surprisingly, the proportion of arrestin bound to Rh1 was higher in WT than in transgenic flies displaying either increased or decreased levels of RK activity. Thus, the question arises as to whether GPRK1 displays a protein kinase-independent function.

We found that flies containing abnormally high and low levels of RK activity displayed reciprocal phenotypes. While a decrease in GPRK1 activity caused a large increase in the amplitude of the ERG, flies expressing elevated levels of RK activity exhibited a significantly smaller ERG than WT. These data indicate that the effect of GPRK1 on response amplitude depended on the kinase activity. The decreased levels of arrestin bound to Rh1 in the transgenic flies might have been a consequence of competitive interactions between arrestin and GPRK1 for binding to the receptor. In further support for a role for Rh1 phosphorylation in attenuating the amplitude of the light response, the size of the ERG was reduced in rdgC flies deficient for the rhodopsin phosphatase.

The function of GPRK1 bears some similarities to those ascribed to mammalian GRKs. Reminiscent of the findings that decreased activity of GPRK1 causes an increase in the amplitude of the ERG, elimination of the mouse nonvisual GRKs (GRK2/3/5/6) results in enhanced sensitivity to agonist stimulation (11–13). Mutation of the rod RK, GRK1, also results in supersensitivity to light (10), although GRK1-/- mice also display a defect in termination of the photoresponse.

A surprising finding was that an ≈2-fold increase or decrease in GPRK1 activity in vivo had major effects on the amplitude of the ERG. However, this result is consistent with the observation that the severity of the phenotypes was similar in GRK6 heterozygous and homozygous mutant mice (11). Furthermore, a 2-fold decrease in the number of phosphorylated residues can have a profound effect on the concentration of arrestin required for inactivation of bovine rhodopsin (31).

The much closer similarity between invertebrate RKs and GRK2/3 raises the possibility that other mammalian proteins related to those that regulate Drosophila rhodopsin may also modulate the activities of β-adrenergic receptors. Consistent with this proposal, the visual arrestins required for Drosophila phototransduction are more closely related to β-arrestins than to the mammalian visual arrestins (32). Currently, the phosphatases that dephosphorylate mammalian β-adrenergic receptors are not known. It is intriguing to speculate that one or both of the mouse RDGC homologs, which are dispensable for the visual response (33), are actually β-adrenergic receptor phosphatases. Finally, although Drosophila GPRK1 and RDGC are enriched in photoreceptor cells, they are also expressed in the heads of eyeless flies. Whether these proteins also participate in other G protein-coupled receptor-dependent sensory modalities, such as taste and smell, remains to be determined.