The SOCS box protein STOPS is required for phototransduction through its effects on phospholipase C

A mutation causing defective termination of phototransduction

To identify new genes that function in Drosophila phototransduction, we screened for mutations on the 3^rd chromosome that disrupted the electroretinogram (ERG) (Wang et al., 2005). ERGs are extracellular recordings that measure the summed retinal responses to light (reviewed in Wang and Montell, 2007). Upon initiation of a light stimulus, the wild-type ERG displays a rapid corneal negative deflection, which quickly returns to the baseline after cessation of the light stimulation (Figure 1A). One mutant line, referred to as stops¹, displayed slower termination of the light response relative to wild-type (Figure 1B), but exhibited normal ERG amplitudes (Figure 1B; wild-type, 8.96 ±1.79 mV, n=12; 8.79 ±1.20 mV, n=12). The stops¹ mutation complemented all known 3^rd chromosome mutations that disrupted the ERG, including arr2, rdgC, ninaE, trp, calx, ninaB and pinta (data not shown), and therefore appeared to represent a new locus required for the normal light response.

Most genes required for phototransduction are expressed and function in the retina. However, some genes, such as ninaB, ninaD and santa maria are required outside the retina for a normal visual response (Gu et al., 2004; Wang et al., 2007; Yang and O’Tousa, 2007). To test whether stops functions autonomously in the retina, we used a mitotic recombination approach that leads to the generation of homozygous stops¹ mutant eyes in animals that are heterozygous in all other tissues (Stowers and Schwarz, 1999). These mosaic flies displayed a slow termination phenotype similar to the original stops¹ mutant (Figure 1C), indicating that stops is required in the compound eye. The stops¹ phenotype was due to a defect in the photoreceptor cells as the slow termination of the light response was evident in intracellular recordings which measure the light response of single photoreceptor cells (Figures 1D and E).

Identification of the stops gene

To identify the gene responsible for the stops¹ phenotype, we used meiotic recombination to map the mutation to cytogenetic region 100B (see Experimental Procedures). No deficiencies were available to more precisely localize the site of the stops¹ mutation. Therefore, to identify a candidate stops gene, we considered whether expression of any gene near the 100B region was eye-enriched, since most genes that function in phototransduction are expressed primarily in the compound eye (Xu et al., 2004). We have previously performed a genome-wide screen for eye-enriched genes by comparing the expression of genes in wild-type heads and in the heads of an eyeless mutant (sine oculis; so) (Xu et al., 2004). In the chromosomal region encompassing stops, only one gene was enriched in the compound eye (CG31006; 100B4-100B5; wild-type head/sine oculis head ratio was ~22.6).

To confirm and extend the data indicating that CG31006 was eye-enriched, we performed in situ hybridizations and Northern blots. Based on in situ hybridizations, CG31006 RNA was spatially restricted to the retina (Figure 1F). The Northern blots supported these results, as the CG31006 mRNA was present in wild-type but not in sine oculis heads or in wild-type bodies (Figure 1G). In addition, the CG31006 signal was absent from the glass mutant (Figure 1G), which eliminated photoreceptor cells (Moses et al., 1989). These results indicated that CG31006 was expressed primarily in photoreceptor cells and was therefore a good candidate for encoding STOPS.

To test whether mutation of CG31006 was responsible for the stops¹ phenotype, we sequenced the CG31006 genomic region from stops¹ flies and compared the sequence with DNA isolated from the original isogenized stock used to conduct the mutagenesis. In the stops¹ mutant, there was a nonsense mutation in place of residue 92 (Figure 2A TGG to TAG), which truncated the 445 amino acid protein encoded by wild-type CG31006 (Supplemental Figure S1A).

To provide additional evidence that CG31006 was the stops gene, we took advantage of a fly line in which a P-element (KG09937) was inserted into the CG31006 gene. The P-element resulted in lethality, which was possibly due to the disruption of the gycβ 100B gene encoding a guanylyl cyclase, since CG31006 was located in an intron of gycβ100B (Figure 2A). We generated mosaic KG09937 animals containing mutant eyes and found that KG09937 exhibited a slow termination phenotype similar to stops¹ (Figures 2B–D, H). KG09937 failed to complement the stops¹ ERG phenotype (Figures 2E and H), indicating the same gene was disrupted in both mutants.

To confirm that CG31006 was the stops gene, we tested for rescue of the stops¹ phenotype using a transgene containing the wild-type CG31006 genomic region (gstops). Normal termination was restored in the gstops;stops¹ flies (Figure 2F and H). To test whether CG31006 was required in photoreceptor cells, we expressed CG31006 in photoreceptor cells using the major rhodopsin promoter (ninaE; neither inactivation nor after potential E). The ninaE-stops;stops¹ flies showed a normal visual response, demonstrating that CG31006 was the stops gene and was required in photoreceptor cells (Figure 2G and H).

NORPA protein level is down-regulated in the stops¹ mutant

The stops gene encodes a 445 amino acid protein including a 40 amino acid SOCS box (Suppressor Of Cytokine Signaling) (Figure 3A; Supplemental Figure S1; residues 294–333). Although only ~20% of the residues in SOCS boxes are widely conserved, a common feature of these domains is that they bind to elongin B/C, which is part of the E3 ubiquitin ligase complex (Kile et al., 2002). SOCS box containing proteins typically include one or more additional motifs. STOPS is most similar to ASB (ankyrin repeat and SOCS box) proteins that include multiple ankyrin repeats (Supplemental Figure S1; nearly 30% identical to residues 191–334). However, STOPS does not appear to include a bona fide ankyrin repeat.

Most SOCS box proteins promote the proteasomal degradation of signaling proteins. The mechanism seems to involve ubiquitination of target proteins by linking the substrates with components of the E3 ubiquitin ligase complex. Therefore, disruption of the stops locus might lead to a defect in response termination as a consequence of overexpression of one or more signaling proteins. Termination of the light response is achieved at multiple levels to ensure that each component in the signaling cascade is properly inactivated. For example, activated rhodopsin is quenched by arrestin binding (Dolph et al., 1993), termination of the Gα_q activity is proposed to be facilitated by the GAP (GTPase activation protein) activity of NORPA (Cook et al., 2000) and closing of the TRP/TRPL channels is regulated by calmodulin (CaM) (Scott et al., 1997) and protein kinase C (INAC) (Popescu et al., 2006). In principle a defect in termination could result if the levels of rhodopsin, Gα_q or the TRP channels were increased relative to the signaling proteins that terminate their activities. Therefore, we checked the relative amounts of these and other proteins by performing Western blots. We found that stops¹ expressed normal levels of Arr2, CalX, CaM, Gα_q, Gβ_e, INAC (PKC), INAD, NINAC (myosin III), Rh1 and TRP (Figure 3B).

In contrast to the expectation that the concentration of one or multiple signaling proteins would be increased, we found that the quantity of NORPA was decreased significantly in the stops¹ mutant as well as in KG09937 mosaic flies and in KG09937/stops¹ trans-heterozygous animals (Figure 3C; ~9% wild-type levels; see below). The level of NORPA reverted to wild-type in gstops;stops¹ and ninaE-stops;stops¹ flies (Figure 3C). These data indicate that STOPS is required specifically for normal expression of NORPA.

STOPS is required for expression of NORPA in photoreceptor cells but not in extra-retinal tissues

Although NORPA is expressed predominately in photoreceptor cells, it is also expressed and functions in other head regions including the brain (Zhu et al., 1993; Kim et al., 1995). To address whether STOPS controls the expression of NORPA in non-retinal tissues, we assessed the relative amounts of NORPA in retinas and in heads without retinas. To conduct the comparison, the extracts were prepared in a constant volume consisting of the same actual numbers of retinas, and heads without retinas, rather than including more retinas to adjust for the lower amounts of total proteins in the retinas only samples. The relative quantity of NORPA was greatly reduced (10%) in the stops¹ retina; however, in the head tissue missing the retina, the concentration of NORPA was similar in wild-type and stops¹ (Figure 3D and Supplemental Figure S2). Furthermore, upon elimination of the photoreceptor cells by expression of the proapoptotic gene hid (Grether et al., 1995), under the control of an eye-specific promoter (GMR) (Hay et al., 1994), the concentration of NORPA was unchanged in stops⁺ and stops¹ flies (Supplemental Figure S3). These results indicate that STOPS is required for NORPA expression specifically in the retina.

The SOCS box is indispensable for STOPS function

The observation that loss of stops caused a decrease rather an increase in NORPA indicates that the SOCS box might not function to promote elongin B and elongin C-dependent proteolysis. To test this proposal, we analyzed lethal mutations in elongin-B and elongin-C in mosaic animals. We found that in elongin-B or elongin-C mutant eyes, NORPA was produced at normal levels (Figure 4A) and the ERG responses were indistinguishable from wild-type (Figures 4B-E). Moreover, the stops¹ phenotype was not suppressed in an elongin-B background (Figure 4F).

The observation that the stops¹ phenotype was independent of elongin B/C raised the question as to whether the SOCS box in STOPS was crucial for function, since elongin B/C proteins bind to SOCS boxes in other proteins. Therefore, we generated a Myc-tagged form of STOPS lacking the entire SOCS box and expressed it using the ninaE promoter (ninaE-stops^dsocs;stops¹; Figure 4G, middle panel). As a control, we generated stops¹ flies expressing full length STOPS fused to a Myc tag (ninaE-stops; Figure 4G, middle panel). Introduction of ninaE-stops^dsocs in stops¹ flies failed to restore normal amounts of NORPA or wild-type termination (Figures 4G and I). Both wild-type NORPA protein concentrations and ERG responses were restored in ninaE-stop;stops¹ flies (Figures 4G and J).

To confirm the importance of the SOCS box, we mutated three consecutive amino acids in the SOCS box (LRH to PGG, residues 297–299), two of which are highly conserved in SOCS boxes (Figure 3A). We expressed the Myc-tagged derivative of STOPS in flies under the control of the ninaE promoter (ninaE-stops^m3;stops¹) and found that Myc-STOPS^m3 and wild-type Myc-STOPS were produced at similar levels (Figure 4G). However, expression of STOPS^m3 did not rescue the stops¹ phenotype (Figures 4G and K). Taken together, these data indicated that the SOCS box in STOPS was indispensable, even though STOPS did not appear to function through an elongin B/C dependent mechanism.

STOPS controls NORPA protein concentration post-transcriptionally

The STOPS protein could be required for NORPA expression by controlling the amount of norpA mRNA. To address this question, we performed Northern blots and found that the quantities of norpA mRNAs were indistinguishable in wild-type and stops¹ flies (Figure 5A). This result suggested that transcription of norpA was not controlled by STOPS.

To provide additional evidence that STOPS regulates NORPA protein concentration post-transcriptionally, we generated transgenic flies that expressed the norpA cDNA under the control of the ninaE promoter. Therefore, production of the norpA mRNA was dependent on the ninaE promoter instead of its own promoter, and was independent of mRNA splicing since a cDNA was used to express norpA. Thus, a requirement for STOPS for transcription or mRNA splicing of norpA should be unnecessary in norpA^P24;ninaE-norpA flies. However, the quantity of NORPA protein in ninaE-norpA;stops¹ flies was as low as in stops¹ flies (Figure 5B). This reduced NORPA concentration was not due to a problem in expression of the ninaE-norpA construct since NORPA was detected at levels similar to wild-type in a norpA^P24;stops⁺ background (Figure 5B; norpA^P24;ninaE-norpA). These data indicate that STOPS regulates NORPA expression post-transcriptionally.

To determine whether STOPS controls the spatial localization of the NORPA protein, we performed immunostaining. However, the concentration of NORPA was too low to be detected in sections of the stops¹ compound eye. Therefore, we stained whole isolated ommatidia with anti-NORPA antibodies. Drosophila photoreceptor cells contain a specialized microvillar region, the rhabdomere, which contains the major rhodopsin (Rh1), NORPA and other proteins that participate in phototransduction. We found that the residual NORPA in stops¹ co-localized with the major rhodopsin Rh1, indicating that it was in the rhabdomeres (Figure 5C). Thus, STOPS did not affect rhabdomere localization of NORPA, but specifically regulated the quantity of the NORPA protein.

Reduced NORPA concentration was responsible for the slow termination

To test whether the reduction in the amount of the NORPA protein alone was responsible for the stops¹ phenotype, we examined whether decreased expression of wild-type NORPA mimicked the slow termination phenotype in stops¹ flies. We generated transgenic flies that expressed varying levels of wild-type NORPA under the control of the heat-shock promoter (hs-norpA) and placed the transgene in a null norpA background (norpA^P24). The heat-shock promoter results in expression of norpA mRNAs in both heads and bodies (Supplemental Figure S4A). However, the NORPA protein expressed under the control of the heat-shock promoter was detected predominately in the retina (Supplemental Figure S4B).

We found that the speed of termination of the photoresponse was proportional to NORPA protein expression. After a 15 minute heat shock, norpA^P24;;hs-norpA flies expressed about ~3.8% of wild-type NORPA, and displayed a severe slow termination phenotype (Figures 6A, B and F). As the duration of the heat-shock increased, the expression of NORPA rose gradually and the termination of the photoresponse became faster. Exposing the norpA^P24;;hs-norpA flies to 30, 60 and 120 minute heat-shock treatments resulted in the production of ~7%, ~15% and ~20% of wild-type NORPA protein respectively, and expression of 20% the normal concentration of NORPA restored a wild-type ERG (Figure 6). Thus, there was a strong correlation between the termination time and amount of NORPA. Moreover, stops¹ flies, which expressed NORPA at ~9% of wild-type levels, displayed a similar termination time as norpA^P24;;hs-norpA flies, which expressed similar quantities of NORPA (30 minute heat shock; Figures 6A, B, D and G). These results indicated that the reduction of NORPA protein in stops¹ was the basis underlying the slow termination phenotype.

PLC activity of NORPA is dispensable for termination of the light response

NORPA has been suggested to function as a GAP to promote hydrolysis of the GTP bound to the Gα_q (Cook et al., 2000), in addition to its established role as a phospholipase C, which is necessary for activation of the TRP and TRPL channels. Therefore, the slow termination of the light response caused by reduced NORPA expression could result from either of two distinct mechanisms. Decreased Ca²⁺ influx due to low PLC activity might underlie the phenotype, since termination of the photoresponse is dependent on light-dependent Ca²⁺ influx. Alternatively, lower GAP activity might be responsible for the stops phenotype due to slow turn-off of the Gα_q. To differentiate between these two possibilities we used genetic approaches.

If the basis of the termination defect in stops¹ was diminished Ca²⁺ influx during light stimulation, then the termination phenotype might be suppressed by a genetic manipulation that increased intracellular Ca²⁺ during light stimulation. We have shown previously that elimination of the Na⁺/Ca²⁺ exchanger, CalX, results in an increased Ca²⁺ concentration during light stimulation (Wang et al., 2005). Therefore, to test whether an increase in intracellular Ca²⁺ can suppress the stops¹ ERG phenotype, we examined the light response in calx^B, stops¹ flies. Although the calx^B photoresponse terminates rapidly, calx^B, stops¹ flies display the slow termination characteristic of stops¹ flies (Figure 7). These results suggest that diminished Ca²⁺ due to low PLC activity is not the major factor affecting termination of the light response in stops¹ flies.

To address whether the diminished NORPA GAP activity underlies the stops¹ phenotype, we tested whether we could restore normal termination in stops¹ flies by expressing a NORPA derivative that retained only GAP activity. Currently, mutations that disrupt the GAP activity of PLCβs have not been described. However, specific residues have been identified that are critical for the phospholipase C activity. We reasoned that if the stops¹ phenotype was due to lower GAP activity, then we might suppress the slow termination by expressing normal amounts of a NORPA derivative with a mutation in a critical residue in the PLC catalytic domain, but which does not affect GAP activity. To test this possibility, we introduced a point mutation in the Y domain of the PLC catalytic site of NORPA (serine 559 to alanine; Figure 8A), which has been shown in other PLCs to reduce PLC activity >10,000-fold (Essen et al., 1996; Ellis et al., 1998). The norpA^S559A was expressed in norpA^P24 flies under the control of the ninaE promoter (ninaE-norpA^S559A), resulting in a normal rhabdomeral distribution of NORPA^S559A (Figure 5C), and expression at a level comparable to NORPA in wild-type flies (Figure 8B).

Expression of NORPA^S559A did not restore a light response in norpA^P24 null mutant flies (Figure 8F) consistent with previous studies that mutation of serine 559 has a profound effect on PLC activity. Although the norpA^P24;ninaE-norpA^S559A flies were devoid of an ERG response, they displayed normal light-induced GAP activity (Figure 8C).

Expression of NORPA^S559A under the control of the ninaE promoter was low, if the ninaE-norpA^S559A transgene was placed in a stops¹ background (Figure 5B). Therefore, to test whether the slow termination due to a reduction in wild-type NORPA protein could be suppressed by NORPA^S559A, we used an alternative approach. We expressed NORPA^S559A in a norpA^P24;;hs-norpA genetic background (norpA^P24;ninaE-norpA^S559A;hs-norpA) and induced low amounts of wild-type NORPA protein using a 30 minute heat shock. After expressing low amounts of wild-type NORPA alone in a norpA^P24 background (30 minute heat shock of norpA^P24;;hs-norpA), the termination of the photoresponse was slow relative to wild-type (Figures 8D, G and I). Of significance here, production of normal levels of NORPA^S559A restored wild-type response termination kinetics in flies expressing low quantities of wild-type NORPA (30 minute heat shock of norpA^P24;ninaE-norpA^S559A;hs-norpA) (Figure 8G and H). The suppression of the termination phenotype could be due to the increase in GAP activity. Alternatively, expression of NORPA^S559A might increase the concentration of wild-type NORPA and raise the PLC activity. However, we found that the NORPA PLC activities were not raised in norpA^P24;ninaE-norpA^S559A;hs-norpA relative to norpA^P24;;hs-norpA flies, indicating that the level of wild-type NORPA was not increased by expression of NORPA^S559A (Supplemental Figure S5). These results indicate that the slow termination phenotype resulting from a low amount of NORPA protein is due to the reduced GAP activity of NORPA.

Low levels of PLC cause slow termination of phototransduction

Many PLCβ isoforms, such as the NORPA protein, which functions in fly phototransduction, are expressed in distinct subsets of cells. In the current work, we demonstrate that NORPA expression in photoreceptor cells depends on the STOPS protein. Mutations in stops caused a decrease in NORPA protein concentration, resulting in slow termination of the photoresponse.

A correlation between low NORPA expression and slow response termination has been proposed previously since multiple norpA alleles as well as inaD² express reduced concentrations of NORPA and display termination defects (Pearn et al., 1996; Shieh et al., 1997; Tsunoda et al., 1997). However, these mutants have additional alterations, precluding the conclusion that low levels of NORPA prevent normal termination of the photoresponse. For example, hypomophic norpA mutants such as norpA^P57 and norpA^P16 express mutant rather than wild-type NORPA (Pearn et al., 1996). Moreover, the interaction of NORPA with the INAD scaffold protein is disrupted in inaD² and norpA^C1094S, suggesting that the INAD-NORPA interaction contributes to the termination phenotype in these mutant flies (Shieh et al., 1997; Tsunoda et al., 1997).

We provide two lines of evidence that a normal concentration of NORPA is essential for stopping signaling after cessation of the light stimulus. First, the stops mutant expresses decreased quantities of wild-type NORPA but not other proteins in photoreceptor cells, resulting in slow termination of the light response. Second, we manipulated the amount of wild-type NORPA using the heat shock promoter. Our results showed a strong relationship between the termination rate and the concentration of NORPA protein.

Low GAP activity underlies the termination defect due to reduced NORPA levels

A question concerns the mechanism through which reduced amounts of NORPA cause a termination defect. In principle, the slow termination could result from a combination of low NORPA levels, and a decrease in intrinsic enzyme activity, due to a requirement for STOPS as a cofactor. However, this does not appear to be the case as the termination defect is indistinguishable between stops⁺ flies that express the same low levels of NORPA (expressed under the control of the hsp70 promoter in a norpA mutant background) as is produced in stops¹ flies. STOPS does not co-immunoprecipitate with the TRP channels, suggesting that it does not cause defects in termination through direct interactions with the TRP channel (Supplemental Figure S6). Moreover, the ERG response is the same in trp^P343 and trp^P343, stops¹ flies (Supplemental Figure S7), suggesting that STOPS does not alter the activity of the TRPL channel. Furthermore, STOPS does not co-immunoprecipitate with the scaffold protein, INAD (Supplemental Figure S6), or alter the concentration of any known member of the signalplex other than NORPA, indicating that the termination phenotype in stops¹ does not result from destabilization of the signalplex.

In addition to serving as a phospholipase C, mammalian PLCβs and NORPA have been proposed to be GAPs (Berstein et al., 1992; Biddlecome et al., 1996; Cook et al., 2000). Therefore, according to one model, low NORPA levels cause a termination defect by decreasing GAP activity, thereby leading to sustained activity of the Gα_q following light stimulation (Cook et al., 2000). However, previous work did not exclude that the slow termination resulting from a reduced NORPA concentration was due to a requirement for PLC activity. Thus, an alternative possibility is that the slow termination is a consequence of decreased light-dependent Ca²⁺ influx, since the NORPA phospholipase C activity leads to opening of the TRP and TRPL channels. A rise in Ca²⁺ is important for termination, as the rate is decreased in flies overexpressing the Na⁺/Ca²⁺ exchanger, CalX, or if the photoreceptor cells are illuminated in a Ca²⁺-free bath (Henderson et al., 2000; Wang et al., 2005).

Our results indicate that the slow termination due to a low concentration of NORPA is caused by reduced GAP activity rather than diminished intracellular Ca²⁺. In support of this conclusion, the delayed termination was not suppressed by a mutation in calx, which eliminates Ca²⁺ extrusion. Rather, we found that slow termination of the photoresponse was fully reversed by expression of a PLC-derivative, NORPA^S559A, which contains a mutation in a residue critical for activation of the photoresponse. Since NORPA^S559A fully retained GAP activity, these data indicate that the GAP activity of NORPA is essential for arresting signaling following cessation of the light stimulus. The observation that introduction of normal levels of NORPA^S559A, suppresses the termination defect resulting from expression of a low concentration of wild-type NORPA, indicates that more than one NORPA molecule can interact successively with the same activated Gα_q. This proposal is consistent with the evidence that NORPA forms a homodimer (Kimple et al., 2001). In addition, it appears that a single activated Gα_q-GTP is capable of interacting sequentially with multiple independent NORPA molecules (Hardie et al., 2002).

Following activation, rapid termination of the light response is essential for high temporal resolution to ensure appropriate responses to subsequent stimuli. Since G-protein deactivation appears to be the rate-limiting step for termination in mammalian phototransduction cascade (Sagoo and Lagnado, 1997), timely deactivation of G-protein may be equally important for regulation of Drosophila phototransduction. Given that NORPA is the direct target for the G-protein, stimulation of the GTPase activity of the Gα_q represents a highly efficient and rapid mode of negative feedback regulation.

A novel role for a SOCS box protein

In the current work, we demonstrate that the SOCS box protein, STOPS, is required specifically for expression of NORPA in photoreceptor cells. SOCS proteins were initially identified as suppressors of cytokine signaling and contain a common C-terminal 40 amino acid SOCS box (Starr et al., 1997). Genomic analyses led to the identification of many additional SOCS box-containing proteins, which are subdivided into groups based on their diverse N-terminal domains. Examples include an SH2 domain (SOCS), WD40-repeats (WSB), a SPRY-domain (SSBs), a RAB domain (RAR), a domain conserved in the Neuralized family of proteins (Neuralized-like) and ankyrin-repeats (ASBs) (Supplemental Figure S1B) (Hilton et al., 1998; Hilton, 1999; Kile et al., 2002). Among the mammalian proteins, STOPS is most similar to the human CRA-b isoform of ASB15, since STOPS and CRA-b share additional homology (STOPS domain) that extends N-terminal to the SOCS box.

Current understanding about SOCS box containing proteins is that they function as adaptor molecules for the E3 ubiquitin ligase complex to target signaling proteins to the protein degradation pathway (Kile et al., 2002). The variable N-terminal domain interacts with target proteins thereby defining the substrate specificity for the E3 ubiquitin ligase complex. The SOCS box binds to a heterodimer composed of ubiquitin-like elongin B and Skp1-like elongin C, and forms an E3 ubiquitin ligase complex with Cullin-2 and Rbx-1 (Kamura et al., 1998; Zhang et al., 1999; Wilcox et al., 2004). In C. elegans, the ZIF SOCS box protein functions in the exclusion of germ-line proteins from somatic lineages by interacting with the elongin B/C complex and targeting the somatic CCCH finger protein for degradation in a cullin and Rbx-dependent manner (DeRenzo et al., 2003).

In contrast to all other characterized SOCS box proteins, which function in protein degradation, STOPS has an opposite role in promoting expression of NORPA in photoreceptor cells. Consistent with this observation, the elongin B/C pathway was not required for expression of NORPA. The elongin-B and elongin-C mutants expressed normal amounts of NORPA and displayed wild-type light responses. Thus, as expected the elongin B/C complex does not function as a positive regulator of NORPA. Furthermore, a double elongin-B, stops¹ mutant displayed the same decreased NORPA concentration and termination defect as the single stops¹ mutant, indicating that the elongin B/C complex does not function in a pathway opposing STOPS. Nevertheless, mutations in the SOCS box disrupt STOPS function. Taken together, these data indicate that although STOPS is dependent on the SOCS box, it acts independently of the elongin B/C complex in vivo. Thus, STOPS is distinct from previously characterized SOCS box proteins.

Control of NORPA expression exclusively in photoreceptor cells

The expression patterns of many signaling proteins are restricted to ensure that they function in a cell-type specific manner. In wild-type flies, NORPA is expressed predominately in photoreceptor cells, but is still detected in other tissues. We found that the STOPS protein is required for expression of NORPA only in photoreceptor cells since in the stops¹ retina, NORPA is reduced by >90% without affecting NORPA in other tissues. In hs-norpA flies, which expressed NORPA under the control of heat-shock promoter, the NORPA protein was expressed primarily in the head but not in the body, although the mRNAs were present in both the head and body (Supplemental Figure S4C). Moreover, in the head of hs-norpA flies, NORPA protein was enriched in the retina (Supplemental Figure S4D).

Although STOPS is required for relatively high expression of NORPA, it is not sufficient. In heat-shocked norpA^P24;hs-stops;hs-norpA flies, which expressed both the stops and norpA RNAs broadly, NORPA was still expressed principally in the retina (Supplemental Figure S4). In HEK293T cells, coexpression of STOPS with NORPA did not result in an elevation of NORPA levels beyond that obtained in cells expressing NORPA alone (unpublished observations). Moreover, NORPA and STOPS did not co-immunoprecipitate either in fly heads or after expressing the two proteins in HEK293T cells (Supplemental Figure S6 and unpublished observations). Thus, there appear to be one or more additional factors that remain to be identified that function in concert with STOPS to promote NORPA expression. Finally, the current work describes a novel mode for controlling the expression of a PLCβ and raises the possibility that there exist mammalian SOCS box proteins that function independently of the elongin B/C complex to promote the expression of PLCβ and other signaling proteins in distinct subsets of cells.

Fly stocks

The stops¹ flies were isolated by performing EMS (ethylmethylsulfonate) mutagenesis and by screening for 3^rd chromosome mutations affecting the ERG (Wang and Montell, 2005; Wang et al., 2005). The Bloomington Stock Center was the source for the “3^rd chromosome deficiency kit” and the following stocks: 1) KG09937, 2) P{lacW}Idh^L3852, 3) y¹w*;P{neoFRT}42D P{Car20y}44B P{GMR-hid}SS2 l(2)*/CyO;P{GAL4-ey.H}SS5 P{UAS-FLP1.D}JD2, 4) y¹w*;P{GAL4-ey.H}DH1 P{UAS-FLP1.D}JD1;P{neoFRT}82B P{GMR-hid}SS4 l(3)*/TM2, 5) P{EPgy2}elongin-B^EY04022 (elongin-B), 6) P{lacW}elongin-C^SH1299 (elongin-C), 7) so¹ (sine oculis), and 8) gl¹ (glass). EP3575 and EP3089 were from the Szeged Drosophila Stock Center. Dr. W. Pak provided the norpA^P24 stock. The “wild-type” control described in this work is w¹¹¹⁸. The stops¹, norpA^P24 and calx^B mutants were in a w¹¹¹⁸ background. The transgenic flies were red-eyed since the marker used for screening for the transgenic flies was w⁺.

Mapping the stops mutation by genetic recombination

We performed deficiency mapping by crossing the stops¹ mutant to the fly stocks that comprised the “3^rd chromosome deficiency kit”. However, stops¹ was not uncovered by any deficiency. Therefore, we used recombination mapping in combination with ninaE (92B), trp (99C) and three P-element lines: P{lacW}Idh^L3852 (66C), EP3575 (94A) and EP3089 (97E). The following recombination rates were observed: trp (1.1%), ninaE (37.5%), P{lacW}Idh^L3852 (45.5%), EP3575 (25%), EP3089 (11.2%). Based on these findings, we localized the stops gene to genomic region 100B.

Generation of transgenic flies

The stops (CG31006) genomic DNA was subcloned from BAC11C03 between the SmaI and SacII sites of the pCaspeR4 vector (Thummel and Pirrotta, 1992). To express the stops cDNA under the control of the ninaE promoter, the cDNA (EST clone GH07253, DGRC) was subcloned between the NotI and XbaI sites of pCNX (Wang and Montell, 2006). To express Myc-tagged wild-type or mutated STOPS in flies, the sequence encoding a Myc tag was fused to the 3′ end of the wild-type stops cDNA. Mutations were introduced using the QuickChange method (Stratagene). The wild-type and mutant stops cDNAs were subsequently subcloned between the NotI and XbaI sites of pCNX.

To express norpA under control of the hsp70 and ninaE promoters, the full length norpA cDNA was excised from pHKX-norpA and subcloned into pCaspeRhs (Thummel and Pirrotta, 1992) and between the NotI and XbaI sites of pCNX. To express the mutated norpA under control of the ninaE promoter, the mutations were first introduced into the norpA cDNA using the QuickChange method, before subcloning into pCNX. The constructs were injected into of w¹¹¹⁸ embryos and transformants were identified by w⁺ pigmentation.

Electroretinogram and intracellular recordings

ERG recordings were performed as previously described (Wes et al., 1999). Briefly, two glass microelectrodes filled with Ringer’s solution were inserted into small drops of electrode cream placed on the surfaces of the compound eye and the thorax. A Newport light projector (model 765) was used for stimulation. The ERGs were amplified with a Warner electrometer IE-210 and recorded with a MacLab/4s A/D converter and the Chart v3.4/s program. Intracellular recordings were performed by placing a low-resistance microelectrode filled with 3 M KCl into a small hole on the compound eye and a reference electrode on the thorax as described previously (Xiong et al., 1994).

Western blots

To perform the Western blots, fly heads were homogenized in SDS-sample buffer, the proteins were fractionated by SDS-PAGE and transferred to Immobilon-P transfer membranes (Millipore) in Tris-glycine buffer. The blots were probed with mouse anti-Tubulin primary antibodies (1:3000 dilution, Developmental Studies Hybridoma Bank), mouse anti-Rh1 antibodies (1:2000 dilution, Developmental Studies Hybridoma Bank), mouse anti-Gβ_e antibodies (1:1000 dilution), rabbit anti-INAC antibodies (1:2000 dilution), rabbit anti-NINAC antibodies (1:2000 dilution), rabbit anti-Arr2 antibodies (1:2000 dilution), rabbit anti-Gα_q antibodies (1:2000 dilution), rabbit anti-INAD antibodies (1:2000 dilution), rabbit anti-TRP antibodies (1:2000 dilution), rabbit anti-CalX antibodies (1:2000 dilution), rabbit anti-NORPA antibodies (1:2000 dilution) and rabbit anti-Myc antibodies (1:1000, Santa Cruz). The blots were subsequently probed with either anti-rabbit, mouse IgG peroxidase conjugate (Sigma), ¹²⁵I-labeled protein A (NEN), ¹²⁵I-labeled anti-mouse IgG (NEN), IRDye 680 goat anti-Rabbit IgG (LI-COR) or IRDye 800 Donkey anti-mouse IgG (LI-COR), and the signals were detected using ECL reagents (Amersham), a phosphoimager or the Odyssey infrared imaging system (LI-COR).

Northern blots

Total RNAs were prepared using Trizol (Invitrogen) and the samples were fractionated on 3% formaldehyde, 1.0% agarose gels. The RNAs were transferred to nitrocellulose membranes and allowed to hybridize with ³²P-labeled probes, which were prepared using random primers in combination with the following DNA templates as described by Roche: norpA (nucleotides 1288–2211 from the norpA cDNA), stops PCR product (nucleotides 675–1500 obtained using the stops cDNA), rh1 PCR product (nucleotides 300–900 of ninaE cDNA) or an rp49 PCR product (nucleotides 24–585 of rp49 cDNA). The hybridizations were performed at 65 °C in 7% SDS, 2 mM EDTA, 0.5 M Na₂HPO₄ (pH 7.2) and the membranes were washed at 65 °C in 0.5X SSC, 0.1% SDS.

In situ hybridizations

Frozen 8 μm sections of adult fly heads were prepared and hybridized to DIG labeled sense and anti-sense RNA probes prepared from a full-length stops cDNA clone (pcDNA3-stops) using either SP6 polymerase or T7 polymerase. The hybridizations were at 70 °C, 50% formamide, 5X SSC 0.02% SDS, 1X Blocking Buffer (Roche). The hybridized sections were incubated with anti-DIG alkaline phosphatase conjugated antibodies (Roche) and the signals were detected using NBT and BCIP.

GTPase Assay

Total membrane proteins were prepared from fly heads that were dark-adapted overnight. The GTPase assays were performed essentially as described (Cook et al., 2000) in 100 μl containing 50 mM KH₂PO₄/K₂HPO₄ buffer pH 6.8, 5 mM MgCl₂, 2.5 mM NaF, 0.25 μ;M [γ-P³²]GTP (300–500 cpm/nmol), 0.5 mM NaATP, 2 mM DTT and 10 mg total membrane proteins. After a 5 min incubation at room temperature under bright blue light or in the dark, the reactions were terminated by addition of 0.6 ml 5% charcoal suspension (in 20 mM phosphoric acid). After centrifugation at 12,000 RPM for 10 min, the ³²P_i generated by hydrolysis of GTP in the supernates were assayed using a scintillation counter. Light-dependent GTPase activities were determined by calculating the differences between the samples kept in the dark with those exposed to blue light.

Immunofluorescence staining of isolated ommatidia

Ommatidia from adult flies were dissected as described (Hardie, 1991) except that they were performed in glass bottom culture dishes (MatTek, MA) filled with Schneider’s Drosophila medium. The dissociated ommatidia were attached to poly-D-lysine coated glass surfaces by placing in an incubator for at least 6 hr at room temperature. The stainings were performed in the same culture dishes. The ommatidia were washed once in PBS, fixed with 3.5% formaldehyde in PBS for 20–30 min, washed 2–3 times in PBS and then treated with 1% Triton X-100 in PBS for 5 min. The samples were then washed 3 times in PBS, blocked with PBSG (0.2% BSA, 1% goat serum, 0.01% saponin) and then incubated with the following primary antibodies in PBSG for 1 hr: mouse anti-Rh1 antibodies (1:500 dilution) and rabbit anti-NORPA antibodies (1:100 dilution). After washing 3–5 times in PBS with 0.01% saponin, the samples were incubated with the following secondary antibodies in PBSG for 1 hr: Alexa Fluor 568 anti-mouse IgG and Alexa Fluor 488 anti-rabbit IgG (Invitrogen). After washing in PBS with 0.01% saponin 3–5 times, the samples were mounted by adding Vectashield (Vector Laboratories, CA) to the cover glasses in the culture dishes. Confocal images were acquired using an UltraView LCI System (Perkin-Elmer).

Phospholipase C activity assay

Phospholipase C activity in fly heads was assayed according to procedures similar to those described (Zhu et. al, 1993; Pearn et. al, 1996). Briefly, we homogenized 20 compound eyes in 50 mM Tris-Cl, pH 7.5, 250 mM KCl, 0.05% (w/v) sodium deoxycholate, 0.1 mM dithiothreitol, and 1x protease inhibitor (Sigma). We combined 10 μl head extracts and 90 μl of a buffer containing 50 mM Tris-HCl, pH 7.5, 1.5 mM CaCl₂, 1 mM EDTA, 0.1 mg/ml BSA, 0.2 mM PI (L-α-phosphatidylinositol Na⁺ salt, Sigma), 44,000 dpm [³H]PIP₂ (phosphatidylinositol-4,5-biphospate, [Inositol-2-³H(N)], PerkinElmer), which were incubating for 5 min at room temperature. The reactions were stopped by addition of 100 μl 5% trichloroacetic acid followed by 50 μl 10 mg/ml BSA. After 15 min incubation on ice, the unhydrolyzed [³H]PIP₂ was separated from [³H]IP₃ by centrifugation at 5000 g for 4 min. The [³H]IP₃ in the supernatants were assayed using a scintillation counter.