Nature Neuroscience
7, 1070 - 1078 (2004)
Published online: 12 September 2004; | doi:10.1038/nn1313
Rolling blackout, a newly identified PIP2-DAG pathway lipase required for Drosophila phototransductionFu-De Huang1, 3, Heinrich J G Matthies1, 3, Sean D Speese2, Mark A Smith2
& Kendal Broadie11 Department of Biological Sciences, Vanderbilt Kennedy Center, Vanderbilt Brain Institute, Vanderbilt University, Nashville, Tennessee 37235-1634, USA. 2 Department of Biology, University of Utah, Salt Lake City, Utah 84112, USA. 3 These authors contributed equally to this work.
Correspondence should be addressed to Kendal Broadie kendal.broadie@vanderbilt.eduThe rolling blackout (rbo) gene encodes an integral plasma membrane lipase required for Drosophila phototransduction. Photoreceptors are enriched for the RBO protein, and temperature-sensitive rbo mutants show reversible elimination of phototransduction within minutes, demonstrating an acute requirement for the protein. The block is activity dependent, indicating that the action of RBO is use dependent. Conditional rbo mutants show activity-dependent depletion of diacylglycerol and concomitant accumulation of phosphatidylinositol phosphate and phosphatidylinositol 4,5-bisphosphate within minutes of induction, suggesting rapid downregulation of phospholipase C (PLC) activity. The RBO requirement identifies an essential regulatory step in G-protein-coupled, PLC-dependent inositol lipid signaling mediating activation of TRP and TRPL channels during phototransduction.Drosophila phototransduction is mediated by phospholipase C (PLC )-dependent opening of transient receptor potential (TRP) and TRP-like (TRPL) channels1,
2,
3. PLC cleaves phosphatidylinositol 4,5 bisphosphate (PIP2) into inositol triphosphate (IP3) and diacylglycerol (DAG), which is further cleaved by DAG lipase to produce polyunsaturated fatty acids (PUFAs). PIP2, IP3, DAG and PUFAs have all been implicated in the regulation of TRP-TRPL channel activation4,
5,
6,
7,
8. PIP2 reportedly inhibits TRPL channels6, suggesting that PLC-dependent PIP2 cleavage may alleviate inhibition. IP3 may have a role in TRP-TRPL activation9, although recent evidence argues against this mechanism8,
10. DAG may directly or indirectly activate TRP-TRPL channels6,
11, and also stimulates an eye-specific protein kinase C (PKC), which negatively regulates TRP channels12,
13,
14. Finally, DAG hydrolysis generates PUFAs that may independently activate TRP and TRPL channels15. These putative interactive effects of the PIP2 signaling system, including the coincident removal of PIP2 and production of DAG and its metabolites, indicate that the activation of TRP and TRPL channels may result from the coordinated effects of lipid signaling4,
16.
A particularly fruitful genetic screening strategy in Drosophila focuses on the identification of temperature-sensitive conditional mutations blocking neuronal signaling. Such mutants have repeatedly revealed key proteins involved in the acute mediation of signal transduction17,
18,
19. Here, we characterize a temperature-sensitive mutant in the rolling blackout (rbo) gene encoding an integral plasma membrane lipase. Conditional rbo mutants exhibit acute, total blockade of phototransduction within a few minutes at restrictive temperature. The block is activity dependent, and recovers rapidly in the absence of light. Conditional rbo mutants show activity-dependent depletion of DAG and concomitant accumulation of PIP and PIP2. These results reveal a previously unknown regulatory step in the G-protein-coupled PLC cascade upstream of TRP and TRPL channels.
Results
Mapping and identification of the rbo gene
Nearly 20 years ago, an ethylmethane sulfonate (EMS)-induced, temperature-sensitive paralytic mutant named stambhaA (stmA1) was mapped to cytological region 42A8-19 (ref. 20). We found that stmA1 shows acute, temperature-sensitive blindness (see below), and set out to clone and characterize this gene. However, three genomic deficiencies of the 42A8-19 cytological region complemented stmA1, suggesting that the original mapping was incorrect. Therefore, deficiencies of surrounding cytological regions were used for remapping, and the mutant was mapped to 44D2-8 with three overlapping deficiencies (Fig. 1a); a remapping of stmA to 44D1-2 has also been reported21. EMS-induced mutant alleles causing unconditional embryonic lethality failed to complement either the temperature-sensitive mutant or the deficiencies covering 44D2-8 (see Supplementary Fig. 1 online). To transposon-tag the gene, two P-elements in 44D were independently mobilized (see Methods), and two new P-element insertion lines were isolated based on failure to complement both the embryonic lethality and the temperature sensitivity of existing EMS mutants (Fig. 1b and Supplementary Fig. 1). Inverse PCR cloning revealed that both P-element insertions are in the recently cloned gene cmp44E, encoding conserved membrane protein at 44E (Fig. 1b). This gene was reported to encode a predicted transmembrane protein essential for cell viability22.
 | | Figure 1. Mapping, identification and genomic rescue of rbo, and hydrophobicity and membrane association of RBO protein. | | |  | (a) Temperature-sensitive and null rbo mutant alleles were mapped to cytological region 44D2−8 on chromosome 2 on the basis of on complementation failure with the indicated three genomic deficiencies. A small intragenic deficiency of the cmp44E gene, rev499, fails to complement rbo mutants, indicating that rbo is allelic to cmp44E. (b) The rbo gene and protein structures. Top, rbo genomic organization, with the insertion site of the eGFP coding sequence indicated. Two P-elements, rboP1 and rboP2, are inserted in the first exon of the smaller rbo gene transcript. Bottom, the two predicted alternatively spliced isoforms of RBO. Shaded bars indicate predicted transmembrane (TM) domains. The premature stop codon mutation (Q396amber) in rbo3 and missense mutation (G487D) in rbots1 are indicated. (c) Hydrophobicity diagram of RBO protein generated by TMpred software; predicted transmembrane domains (minimum 17 amino acids, maximum 33 amino acids with positive hydrophobicity value) are indicated with arrows. (d) Subcellular location of RBO-eGFP fusion protein in extracts from adult brain. Numbers at right, kDa. The protein ( 117 kDa) is strongly associated with the membrane fraction (T, total lysate; S, supernatant; P, membrane pellet). Right, the detergent NP-40 can release RBO from the membrane, but high pH or high salt concentration cannot. (e) V5-tagged RBO expressed in S2 cells was labeled by cell-surface biotinylation. Left, RBO-V5-His probed with V5 antibody; right, biotin-labeled RBO-V5-His with streptavidin-conjugated alkaline phosphatase.
 Full Figure and legend (44K) |
| |
Complementation tests between stmA and cmp44E mutants confirmed that both affect the same gene (Supplementary Fig. 1). Subsequent sequencing of temperature-sensitive and EMS-induced lethal alleles revealed one missense (G487D) and one nonsense (Q396amber) mutation, respectively (Fig. 1b). A single copy of wild-type genomic DNA fully rescued both the temperature-sensitive defects and the embryonic lethality of all homozygous and transheterozygous mutant alleles (Supplementary Fig. 1). Based on the distinctive phenotype of the conditional mutant, the single descriptive name rolling blackout (rbo) was adopted for this gene, along with the mutation designations rbots1 (temperature-sensitive allele), rboP1 and rboP2 (P-element insertion lines), and rbo2 (genomic deficiency, rev499), rbo3 and rbo4 (EMS lethal mutations) (Supplementary Fig. 1).
RBO pioneers a family of lipolytic enzymes
Homologs of RBO exist in species ranging from yeast to man (Supplementary Fig. 2). Sequence analysis of Drosophila RBO and its homologs in other species indicates conservation around three motifs characteristic of carboxyesterases and lipases (Supplementary Fig. 3). These three motifs contain serine, aspartic acid and histidine residues, which form the enzymatic catalytic triad. The characteristic lipase motif of the serine active site, Gly-X-Ser-X-Gly23,
24, is found in RBO (Supplementary Fig. 3). In addition, RBO shows homology to three known lipases: (i) the recently cloned human membrane-associated sn-1 DAG lipase (42% homology)25 (ii) rat triacylglycerol lipase (35%) and (iii) fungal mono- and diacylglycerol lipases (25−30%). Taken together, these features indicate that RBO is a lipase.
RBO is an integral membrane protein
Hydrophobicity analysis of RBO predicts a 2- or 3-pass transmembrane protein (Fig. 1c)22. In the 3-pass model, the N terminus containing the histidine active site is outside the plasma membrane, whereas the C terminus harboring the aspartic site is inside, rendering this model invalid. In contrast, in the 2-pass model, both N and C termini are inside, placing the histidine and aspartic active sites on the same (internal) side of the membrane. The serine active site is located in the center of a transmembrane domain. Thus, only in the 2-pass model can the three active sites assemble to form the triad catalytic center, indicating that RBO is likely to be a 2-pass transmembrane protein.
To test the prediction that RBO is a transmembrane protein, we assayed the subcellular location of RBO in fractionation experiments (Fig. 1d). The heads of flies expressing RBO-eGFP were homogenized, and soluble and membrane fractions generated. RBO-eGFP was found to be associated only with the membrane pellet fraction (Fig. 1d, left). The membrane fractions were next treated with agents that extract peripheral membrane proteins (high salt concentration and basic pH) or integral membrane proteins (Nonidet P-40 (NP-40) detergent). RBO protein was not appreciably disassociated from the membrane with high salt (1 M) or basic pH (pH = 10), but could be released by detergent (Fig. 1d, right). These results are consistent with the prediction that RBO is a transmembrane protein.
To further test this prediction, Drosophila S2 cells were transfected with a V5 epitope−tagged RBO, after which all cell surface proteins were biotinylated with a cell membrane−impermeable version of biotin. V5-tagged RBO expressed in S2 cells was biotinylated (Fig. 1e), showing that the protein must span the plasma membrane. Taken together, these results indicate that RBO is a transmembrane protein.
RBO is enriched in neuropil and photoreceptors
To study the expression of RBO, we generated transgenic lines expressing RBO-eGFP under genomic regulation (Fig. 1b). RBO-eGFP fully rescues both temperature-sensitive blindness and embryonic lethality in all rbo mutant allelic combinations (Supplementary Fig. 1), showing that the protein is properly expressed and functional. RBO-eGFP was expressed in the rbo null mutant background to allow us to assay protein expression throughout development from embryo to adult.
RBO is concentrated in the nervous system beginning during mid-embryogenesis and continuing in the mature fly (Fig. 2). During embryogenesis, RBO is first detected at stage 9 in the central neuropil of the developing ventral nerve cord (VNC) and brain (Fig. 2a). Throughout the larval stages, the protein continues to remain enriched in the neuropil. In double-labeling experiments (Fig. 2b,c; nuclei red), RBO is undetectable in the cortical cell bodies, which is particularly obvious in the brain lobes (Fig. 2c). The RBO enrichment in neuronal processes is clearly observed in peripheral nerves penetrating the cortical somal layers to exit the VNC (Fig. 2b arrow) and in the neuronal axons outside the VNC (Fig. 2c).
| | |
In adult flies, RBO remains enriched in the neuropil of thoracic ganglion and brain, but absent from the cortical cell bodies (Fig. 2d). As far as can be observed, the protein appears to be highly expressed in all neuropil regions and thus seems to be pan-neuronal. RBO is also enriched in the adult retina, in photoreceptor cells (Fig. 2e). Outside the CNS, strong RBO expression occurs in many peripheral sensory neurons throughout development and in the adult (data not show). These observations suggest that RBO may function in multiple forms of sensory transduction, particularly phototransduction.
Required role of RBO in phototransduction is activity dependent
Drosophila phototransduction was assayed by electroretinogram (ERG) recording26,
27. During a light stimulus, the ERG is composed of an initial rapid depolarization (amplitude 1, Fig. 3a) and a sustained depolarization (amplitude 2, Fig. 3a). At permissive temperature (25 °C), rbo temperature-sensitive mutants show normal light-induced ERG responses (Fig. 3). In contrast, at nonpermissive temperature (37 °C), the ERG was completely eliminated, either in constant light (<1 min) or during repetitive light flashes (Fig. 3). This complete blockade of phototransduction is due only to loss of rbo function, as the addition of a single copy of the wild-type genomic rbo gene fully restored the ERG (Fig. 3a). Notably, during the progressive loss of phototransduction, the ERG manifests a transient depolarization (Fig. 3a, at 2.5 min), similar to flies mutant in the trp (transient receptor potential) gene (as discussed later). Quantification of ERG amplitudes show that rbo mutants show normal phototransduction at 25 °C, lack a detectable ERG at 37 °C, but that ERG amplitudes are fully restored by a single copy of wild-type rbo (Fig. 3b). These results demonstrate that RBO is acutely required for phototransduction.
| | |
By altering frequency and duration of light stimuli, the phototransduction block in rbo mutants was found to be activity dependent (Fig. 4). At 37 °C, continuous light blocked phototransduction within 40 s, but repetitive 1 s light flashes at low frequency induced robust ERGs even after 30 min at 37 °C. Repetitive 1 s light flashes separated by darkness of 20 s (rbo mutants at high frequency; rbo-h), caused blockade of phototransduction in <5 min (Fig. 4a). In contrast, repetitive 1 s light flashes separated by darkness of 3 min (rbo mutants at low frequency; rbo-l) allowed phototransduction to persist (Fig. 4a). Quantification of initial ERG depolarization (amplitude 1) showed that rbo mutants maintained normal amplitudes during low-frequency light stimulation, indistinguishable from wild-type, whereas high-frequency stimulation caused a block in <5 min (Fig. 4b, top). The sustained ERG depolarization (amplitude 2) was similarly unchanged compared to control during low-frequency stimulation, whereas high-frequency stimulation blocked the ERG (Fig. 4b, bottom). Note that longer-term stimulation at low frequency did cause a significant reduction in ERG amplitudes (30% amplitude 1, 70% amplitude 2; Fig. 4b). These results demonstrate an acute activity-dependent requirement for RBO in phototransduction.
| | Figure 4. Activity-dependent blockade of phototransduction in rbo mutants. | | |  | Phototransduction block in rbo temperature-sensitive mutants is activity dependent and recovers in the dark at restrictive temperature. (a) Representative traces show ERG recordings with repetitive 1 s light stimulation at 20 s intervals (high (h) frequency) in wild-type (wt-h) and rbo mutants (rbots1/rbo2; rbo-h), at 3 min intervals (low (l) frequency) in mutants (rbo-l), and a mix of high and low frequency in the mutants (rbo-m). Lower traces show the simultaneously recorded temperature, in each case shifted from 25 °C to 37 °C as indicated. At the high frequency of light stimulation, phototransduction is abolished in rbo mutants within 5 min, whereas at the low frequency of light stimulation robust ERG responses persist. (b) Quantification of ERG amplitudes. Top and bottom graphs show mean values of amplitudes 1 and 2, respectively (shown in Fig. 3a). n = 5 in each case. #P < 0.05, *P < 0.01. (c) Recovery of ERG amplitudes after continuous light-induced total blockade. Effects of dark interval and temperature decrement on recovery of initial peak amplitude (1, left), sustained peak amplitude (2, middle) and amplitude at 10 s of illumination (3, right) at room temperature (RT) or 37 °C, as indicated. n = 5 for each. Error bars, s.d.
Full Figure and legend (44K) |
| |
When maintained in darkness for 3 min after complete suppression of phototransduction, rbo mutants showed strong recovery (80%) in ERG amplitude 1 but relatively weak recovery (25%) in ERG amplitude 2, suggesting differential RBO dependence of the initial and sustained phototransduction events (Fig. 4a,b, rbo-m). To study in detail the recovery of phototransduction in rbo mutants, we used continuous light at 37 °C to block phototransduction, then subjected flies to variable dark intervals before the next test light (Fig. 4c). Three ERG amplitudes were quantified: amplitudes 1 (initial peak), 2 (sustained peak) and 3 (amplitude of sustained depolarization at 10 s). Recovery of both amplitudes 1 and 2 was dark interval dependent, with amplitude 1 recovering faster and more completely, whereas amplitude 3 did not recover appreciable even after 3 min (Fig. 4c, right). When rbo mutants were returned to permissive temperature (25 °C) to restore RBO function, all three ERG amplitudes showed complete (>90%), dark interval−dependent recovery within 3 min (data not shown). These results demonstrate that progressively sustained ERG depolarization is more dependent on RBO function.
RBO acts upstream of TRP and TRPL channels
The loss of phototransduction current could be due either to a direct defect in TRP and TRPL channels or to an upstream defect. In darkness, anoxia produces a rapid depolarization of photoreceptor cells that is largely dependent on the functional integrity of TRP and TRPL channels28. Therefore, to test the integrity of TRP and TRPL channels, anoxia was produced in rbo mutants after phototransduction blockade at 37 °C. Brief anoxia produced a similar rapid depolarization in rbo mutants (15.9  1.0 mV) and wild type (16.8 2.1 mV; Fig. 5). The amplitude of anoxia-induced depolarization in both rbo mutants and wild-type controls was twice that of trpl302;trp343double mutants (7.4 1.5 mV, P < 0.01; Fig. 5). In addition to attenuated anoxia-response amplitudes, the response kinetics of trpl;trp mutants was much slower than in either rbo mutants or wild type (Fig. 5)28. These data suggest that normal TRP and TRPL channels persist in rbo mutants under conditions where phototransduction is abolished. Thus, the phototransduction blockade in rbo mutants maps to a defect upstream of TRP and TRPL channels.
| | |
Activation of TRP and TRPL channels is completely dependent on PLC (norpA gene). Reduced PLC activity is reflected by slowed ERG response kinetics, particularly the slope of the initial ERG depolarization29. ERG records from rbo mutants and controls were therefore quantified by measuring slope from lights-on to half amplitude 1 (Fig. 3a). Controls showed a significantly (P < 0.05) increased slope as a function of temperature, from 1.43 0.15 (25 °C) to 2.2 0.18 mV/mS (37 °C) (Fig. 6), consistent with increased PLC activity at elevated temperatures30. In contrast, the ERG slope in rbo mutants was significantly (P < 0.01) decreased during a 25 °C to 37 °C temperature shift, from 1.58 0.18 to 0.45 0.04 mV/mS (Fig. 6). This observation suggests that PLC activity is inhibited by loss of RBO function.
| | |
To determine whether PLC function is impaired in rbo mutants, an in vitro assay of PLC activity was used29,
31. Head extracts were assayed under conditions where PLC is dissociated from membrane by high salt concentration in the presence of detergent (sodium deoxycholate). At both 25 °C and 37 °C, PLC activity in wild-type and rbo mutant extracts was indistinguishable (Supplementary Fig. 4). These results demonstrate that the rbo mutation does not impair PLC activity per se. However, a reduction of PLC activity in vivo could result from a defect in activation or regulation of the enzyme in the absence of RBO function.
Misregulation of PIP-PIP2 and DAG levels in rbo mutants
The above studies show that RBO is essential to an activity-limited phototransduction mechanism upstream of TRP and TRPL channels. The nature of the phototransduction blockade is most consistent with disruption of PLC activity or of the PLC-dependent lipid signaling cascade leading to TRP and TRPL channel activation. To test this hypothesis, PIP2 pathway lipids (PIP, PIP2, DAG and phosphatidic acid; Supplementary Fig. 4) were assayed in the heads of rbo mutant and wild-type flies at 25 °C and 37 °C, in conditions of constant light (active state) or constant darkness (resting state).
PIP, PIP2 and phosphatidic acid were assayed by immunostaining of high-performance thin-layer chromatography (HPTLC) plates (Fig. 7). At 25 °C, under either light or dark conditions, there was no detectable difference in the levels of PIP, PIP2, or phosphatidic acid between rbo mutants and wild-type (Fig. 7). At 37 °C, both PIP and PIP2 were elevated relative to wild type. Analysis of PIP levels showed a highly significant elevation in rbo mutants in both light (1.97 0.13 rbo versus 0.8 0.05 pmol in wild type) and dark (1.6 0.07 rbo versus 1.0 0.09 in wild type; Fig. 7a). As expected from PIP2 consumption in light, the level of PIP2 in wild-type was strongly reduced by shifting from dark (2.34 0.09) to light (1.3 0.13 pmol), whereas no change at all occurred in the rbo mutant (1.7 0.07 dark versus 1.85 0.16, light; Fig. 7b). This defect strongly suggests a block in light-activated PLC activity in rbo mutants. As a consequence, PIP2 levels are significantly elevated in rbo mutants in the light, under conditions that block phototransduction (Fig. 7b). Phosphatidic acid abundance appeared comparable between wild-type flies and rbo mutants under all conditions (Fig. 7c). This observation is consistent with the fact that phosphatidic acid abundance is independent of PLC activity32.
| | |
The level of DAG was assayed using the DAG kinase detection method33 (Fig. 8). The most notable finding was a reduction of DAG levels in the absence of RBO function and in the light (Fig. 8). In wild type, DAG levels increased at 37 °C, consistent with the effect of elevated temperature, which increases PLC activity30. In contrast, in rbo mutants at 37 °C, the level of DAG was half that of wild-type flies (200 10.9 in wild-type versus 97 4.1 pmol in rbo flies; Fig. 8b). In the dark, wild-type flies showed lower DAG levels. In contrast, in rbo mutants the level of DAG in the dark and light was the same, consistent with a failure to activate PLC in the mutants (Fig. 8b). DAG levels of the mutant in the light were indistinguishable from those of the wild type in the dark, suggesting a complete failure of light-dependent PLC activity in the mutant. In the dark, DAG levels were comparable in rbo mutants and controls (78.7 3.6 pmol in wild-type versus 69.3 4.0 pmol in rbo flies; Fig. 8b). There was a 50% increase in DAG levels in the mutant compared to wild type in both light and dark conditions at 25 °C (Fig. 8b). The significance of this DAG elevation under conditions in which phototransduction is not detectably altered is unclear. Taken together, these data show that the RBO-dependent mechanism modifies PIP2-DAG signaling, consistent with disruption of PLC activity.
| | |
Discussion
Chromosomal deficiency mapping, transposon mutageneses and cloning, mutant genomic DNA sequencing, and genomic DNA rescue of mutant phenotypes have combined to demonstrate that the rolling blackout (rbo) gene is allelic to both stampha A (stmA)20,
21 and conserved membrane protein at 44E (cmp44E)22. Hydrophobicity analysis, membrane fractionation and disassociation studies, and surface biotinylation assays show that RBO is an integral plasma membrane protein. RBO is the first of a new protein family with a serine-aspartate-histidine catalytic triad and homology to known DAG lipases. RBO is enriched in the nervous system, apparently in all neurons, where it is subcellularly restricted to neuropil and to sensory neurons, including photoreceptors. Null rbo mutants are embryonically lethal, showing that RBO has an essential function. Temperature-sensitive rbo mutants show a complete blockade of phototransduction, indicating that RBO has a function acutely required in G-protein-coupled PLC signaling activation of TRP and TRPL channels. The accumulation of PIP-PIP2 and depletion of DAG after conditional blockade of RBO function suggest that RBO is required for PLC activity.
Activity-dependent requirement for RBO in phototransduction
Drosophila phototransduction involves PLC-dependent opening of TRP and TRPL channels4,
5,
8,
16,
34. Conditional removal of RBO function blocks this signaling pathway completely in less than 1 min under constant stimulation. The widespread expression of RBO, and its essential role in the embryo, argue that the protein is not required in a photoreceptor-specific mechanism (such as rhodopsin function). Likewise, the persistence of normal anoxia-induced depolarization under conditions where phototransduction is blocked indicates that RBO is not required for the functional integrity of TRP and TRPL channels28. Thus, RBO function maps between light-activated rhodopsin and the phototransduction channels, to the PLC cascade. The conditional nature of the phototransduction blockade, and the persistence of normal PLC activity isolated from head extracts of rbo mutant flies, show that RBO is not required for PLC expression or function per se. Therefore, RBO seems to be required for PLC activation in vivo.
Two lines of evidence are indicative of an acute blockade of PLC activity in rbo mutants. First, rbo mutants show rapid depletion of DAG and concomitant accumulation of PIP-PIP2. In the absence of RBO function, the level of PIP2 does not change with light stimulation, suggesting a complete absence of light-activated PLC cleavage of PIP2. Likewise, the PLC product (DAG) is not elevated by light stimulation in the absence of RBO function. These results suggest that loss of RBO function may equate to loss of PLC function. A caveat concerning these total head lipid analyses is that rbo mutations block all light-activated neuronal signaling, and this could potentially cause indirect alterations in lipid metabolism. The second line of evidence is that the loss of RBO function causes a marked reduction in the slope of the initial photoreceptor depolarization, which correlates with reductions in PLC activity29. Taken together, these results suggest that RBO may be required for regulation of PLC in photoreceptors.
RBO is required in a strictly activity-dependent manner. Only under conditions of high signaling demand (constant light or rapid light flashes) does loss of RBO function result in complete blockade of phototransduction signaling. Removing the demands of light-activated signaling restores phototransduction in the absence of RBO function. Activity-dependent phototransduction blockade in other mutants has been linked to light-dependent exhaustion of limiting signaling factors35,
36,
37. For example, mutation of the gene encoding CDP-diacylglycerol synthase (cds) causes activity-dependent phototransduction blockade because the protein is required for regeneration of PIP2 from phosphatidic acid36. Similarly, the activity-dependent phototransduction blockade in rbo mutants must be due to exhaustion of a critical RBO-dependent factor. The rapid decay of sustained photoreceptor depolarization in rbo mutants indicates that this factor is consumed during sustained light activation. The dark interval−dependent recovery of phototransduction at restrictive temperature may indicate a secondary pathway for regenerating this factor, which is rate limiting in the absence of RBO function owing to rapid, light-activated consumption. A caveat to this conclusion is that the supposed 'secondary pathway' may simply reflect the temperature-sensitive mutant not producing the complete null condition. We conclude that RBO is required to sustain phototransduction but is not directly involved in the activation of TRP and TRPL channels.
Activation of TRP and TRPL channels and RBO function
The mechanism by which TRP and TRPL channels are activated during Drosophila phototransduction remains hotly debated, but there is a clear consensus that channel activation depends absolutely on PLC. Interest has focused on DAG, as a direct activator, or DAG hydrolysis products (PUFAs) as secondary activators of TRP and TRPL channels4,
8,
15,
34. The fact that a characterized protein with over 40% similarity to RBO is a DAG lipase25 immediately suggests that RBO-mediated hydrolysis of DAG might produce the putative PUFA trigger signal. However, the amino acid identity between RBO and the sn-1 DAG lipase is low (12%), and RBO clearly belongs to a separate lipase family that does not as yet include known DAG lipases. Moreover, conditional removal of RBO results in rapid depletion of DAG, which is the opposite result to that predicted for a DAG lipase mutant. Finally, given the activity-dependent nature of the rbo phototransduction blockade, as discussed above, it does not seem probable that RBO could have any direct role in TRP and TRPL channel activation. Thus, the data do not support the hypothesis that RBO acts as a DAG lipase producing PUFAs that directly gate the phototransduction channels.
Acting as a DAG lipase, RBO might acutely regulate phototransduction through a mechanism that does not involve direct gating of the TRP and TRPL channels. Similar to rbo mutants, mouse DAG kinase  mutants38 and Drosophila rdgA (encodes eye-specific DAG kinase) mutants39 fail to accumulate DAG, as would be predicted by the enzymatic function of their products. Indeed, the mouse mutants show DAG depletion, similar to rbo mutants. Moreover, PLC activity in rdgA mutants is reduced39,
40, and downregulation of PLC activity has similarly been suggested in the mouse mutants38. These observations suggest positive feedback loops, by which products downstream of DAG feed back to upregulate PLC activity. In the absence of this feedback, in DAG kinase mutants or, by analogy, DAG lipase mutants, PLC activity would be downregulated, resulting in accumulation of PIP2 and loss of DAG. Moreover, disruption of such a feedback loop would be secondary to depletion of the signaling intermediates, and therefore it would take time for PLC downregulation to be manifested. All of these conclusions fit nicely with the observed phenotypes of rbo mutants. Thus, RBO could be a DAG lipase, as suggested by closest homology, but rather than acting to produce excitatory PUFAs that directly gate TRP and TRPL channels, it may act instead as a positive regulator of PLC activity.
In summary, RBO as an integral plasma membrane lipase that is essential for PLC-mediated phototransduction. In the absence of RBO function, the PLC-PIP2 signaling pathway arrests acutely in an activity-dependent manner. Multiple aspects of the rbo mutant phenotype are consistent with the hypothesis that RBO function is required for PLC function. We therefore propose that RBO acts as a positive regulator of PLC during G-protein-coupled inositol phospholipid signaling.
Methods
Genetics.
Drosophila stocks were cultured on standard medium on a 12 h:12 h light/dark cycle at 25 °C. The wild-type strain w1118 was used as control. Genomic deficiencies used include (i) 42A region: Df(2R)nap1, Df(2R)cn88b, and Df(2R)nap9, and (ii) 44D region: Df(2R)H3D3 (31), Df(2R)Np3 (258) and Df(2R)H3C1 (198). P-element stocks used for local hop mutagenesis were P{lacW}l(2)44Db[k08904] and P{lacW}ptc[k0507]. The temperature-sensitive stmA1 mutant and lethal stmA18.2 mutant were provided by S. Chandrashekaran and the cmp44E mutants (EMS mutant cmp44E1 and intragenic deficiency (cmp44Erev499) by T. Jongens. The local P-element mutagenesis was an F3 screen based on a failure to complement existing EMS mutants. Complementation tests were done for both temperature-sensitive blindness and embryonic lethality.
Molecular techniques.
To identify the rbo gene, we sequenced the flanking genomic DNA (inverse PCR) of P-element insertions in rbop1 and rbop2. To find EMS point mutations, both genomic rbo DNA and cDNA were sequenced. To generate the RBO-GFP fusion construct, eGFP coding sequence was fused with rbo genomic DNA (RF II in pBluescript SK+ vector)22 by overlap extension PCR and ligated into pCaSpeR4. This plasmid containing genomic rbo-egfp hybrid DNA was injected into w1118 embryos with p 2−3 helper plasmid. To express RBO in S2 cells, the cDNA of the longer RBO splice form was cloned by RT-PCR and ligated into pMT/V5-His expression vector.
Membrane association analyses.
rbo2/rbo2; rbo-egfp/rbo-egfp transgenic fly heads were homogenized with lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM KCl, 1 mM EGTA, 1 mM EDTA, 5 mM DTT and Complete protease inhibitors (Roche). The lysate was centrifuged at 1,000g for 10 min. The supernatant (total lysate) was centrifuged at 100,000g for 60 min to separate pellet from supernatant. The total lysate, pellet and supernatant were western blotted and probed with a monoclonal GFP antibody (Covance). To examine whether RBO was an integral or peripheral membrane protein, the membrane fraction was treated with PBS, 0.1 M NaCO3 (pH 10), 1% NP-40 or 1.0 M NaCl for 30 min, then centrifuged at 30,000g for 30 min to generate supernatant and pellet.
To test whether RBO is an integral membrane protein, S2 cells were transfected with V5 epitope−tagged RBO, and cell surface proteins biotinylated at 4 °C with a membrane-impermeable version of biotin (1 mM) and washed with buffers (TBS and 100 mM glycine). Membranes were treated with RIPA buffer containing Complete protease inhibitors and centrifuged (100,000g for 1 h). The supernatant was used for immunoprecipitation with V5 antibody. The immune complex was washed with RIPA buffer and probed by western blotting with either the V5 antibody or streptavidin conjugated to alkaline phosphatase.
Imaging and immunocytochemistry.
RBO-eGFP imaging was performed on rbo-null mutant flies (rbo2) rescued with rbo-eGFP hybrid DNA using a Bio-Rad Radiance 2000 laser confocal scanning system. In the embryo, all analyses were done on living animals. The larval CNS was treated with PBS−Triton X-100 (0.3%) for 30 min and then with RNase A (0.025  g/ml) for 20 min, both at room temperature. Propidium iodide (1.25 g/ml) for nuclear staining was applied for 30 min at room temperature. Adult heads were fixed in 4% paraformaldehyde for 2 h at room temperature.
Electroretinogram recording.
A standard ERG recording configuration was used41. Briefly, flies were anesthetized with CO2 and then embedded in dental wax. A reference glass microelectrode filled with standard saline was inserted into the thorax. A recording glass microelectrode filled with 3 M KCl was inserted into the center of the fly's eye. Flies were dark adapted for 5 min. Light was generated with a 1460X GE lamp and controlled with a VMM-T1 shutter (Uniblitz). Anoxia was induced by acute application of N2.
Lipid biochemistry.
Assay of PIP, PIP2 and phosphatidic acid mass. PIP, PIP2 and phosphatidic acid mass were determined by HPTLC immunohistochemistry42. Flies were dark adapted for 30 min, and then placed at 25 °C or 37 °C in the light or dark for 5 min. They were next frozen in liquid nitrogen, sieved heads pulverized, and homogenized in lysis buffer (1:2:0.8 (v/v) chloroform/methanol/1 M NaCl containing 2 mM EDTA and EGTA) (10 min at room temperature), chloroform and 1 M NaCl, 2 mM EDTA and EGTA were added, and the lower phase collected43. The interface was re-extracted with lysis buffer containing 0.2 M phosphoric acid. The pooled lower phases were dried and separated on HPTLC plates (Whatman LHPKD) with a standard curve using 46:17:15:14:8 chloroform/acetone/methanol/acetic acid/water. The plate was treated with PBS containing 1% polyvinylpyrrolidone and 0.5% bovine serum albumin (PBS-PV-BSA), incubated with PIP and PIP2 antibody (1:1,000, Assay Designs, Inc) followed by alkaline phosphatase−conjugated secondary antibody and developed.
Assay of DAG mass. Frozen adult fly heads were isolated and extracted as described above. This mixture was incubated, chloroform and 1 M NaCl containing 2 mM EDTA and EGTA were added, and then the mixture was centrifuged and re-extracted and the lower phases collected. The dried, pooled lower phases were resuspended in chloroform and DAG levels were assayed based on an Escherichia coli DAG kinase assay33. The washed products and standard curve were separated by HPTLC using 65:15:5 chloroform/methanol/acetic acid. 32P incorporation was assayed with a Typhoon 9400 (Amersham).
In vitro PLC activity assay. PLC activity was assayed essentially as in refs. 29,31. Homogenate was pre-equilibrated at either 25 °C or 37 °C for 5 min, and the reaction was allowed to proceed for 5 min and stopped by addition of trichloroacetic acid. The mixture was centrifuged and an aliquot of the supernatant used for liquid scintillation counting.
Note: Supplementary information is available on the Nature Neuroscience website.
Received 1 June 2004; Accepted 19 July 2004; Published online: 12 September 2004.
REFERENCES
- Bloomquist, B.T. et al. Isolation of a putative phospholipase C gene of Drosophila, norpA, and its role in phototransduction. Cell 54, 723−733 (1988). | PubMed | ISI | ChemPort |
- Niemeyer, B.A., Suzuki, E., Scott, K., Jalink, K. & Zuker, C.S. The Drosophila light-activated conductance is composed of the two channels TRP and TRPL. Cell 85, 651−659 (1996). | Article | PubMed | ISI | ChemPort |
- Scott, K., Becker, A., Sun, Y., Hardy, R. & Zuker, C. Gq
 protein function in vivo: genetic dissection of its role in photoreceptor cell physiology. Neuron 15, 919−927 (1995). | Article | PubMed | ISI | ChemPort |
- Hardie, R.C. Phototransduction in Drosophila melanogaster. J. Exp. Biol. 204, 3403−3409 (2001). | PubMed | ISI | ChemPort |
- Montell, C. Visual transduction in Drosophila. Annu. Rev. Cell Dev. Biol. 15, 231−268 (1999). | Article | PubMed | ISI | ChemPort |
- Estacion, M., Sinkins, W.G. & Schilling, W.P. Regulation of Drosophila transient receptor potential-like (TrpL) channels by phospholipase C-dependent mechanisms. J. Physiol. (Lond.) 530, 1−19 (2001). | Article | PubMed | ISI | ChemPort |
- Minke, B. & Cook, B. TRP channel proteins and signal transduction. Physiol. Rev. 82, 429−472 (2002). | PubMed | ISI | ChemPort |
- Hardie, R.C. Regulation of trp channels via lipid second messengers. Annu. Rev. Physiol. 65, 735−759 (2003). | Article | PubMed | ISI | ChemPort |
- Hardie, R.C. & Minke, B. The trp gene is essential for a light-activated Ca2+ channel in Drosophila photoreceptors. Neuron 8, 643−651 (1992). | Article | PubMed | ISI | ChemPort |
- Acharya, J.K., Jalink, K., Hardy, R.W., Hartenstein, V. & Zuker, C.S. InsP3 receptor is essential for growth and differentiation but not for vision in Drosophila. Neuron 18, 881−887 (1997). | Article | PubMed | ISI | ChemPort |
- Raghu, P. et al. Constitutive activity of the light-sensitive channels TRP and TRPL in the Drosophila diacylglycerol kinase mutant, rdgA. Neuron 26, 169−179 (2000). | Article | PubMed | ISI | ChemPort |
- Smith, D.P. et al. Photoreceptor deactivation and retinal degeneration mediated by a photoreceptor-specific protein kinase C. Science 254, 1478−1484 (1991). | PubMed | ISI | ChemPort |
- Huber, A., Sander, P., Bahner, M. & Paulsen, R. The TRP Ca2+ channel assembled in a signaling complex by the PDZ domain protein INAD is phosphorylated through the interaction with protein kinase C (ePKC). FEBS Lett. 425, 317−322 (1998). | Article | PubMed | ISI | ChemPort |
- Liu, M., Parker, L.L., Wadzinski, B.E. & Shieh, B.H. Reversible phosphorylation of the signal transduction complex in Drosophila photoreceptors. J. Biol. Chem. 275, 12194−12199 (2000). | Article | PubMed | ISI | ChemPort |
- Chyb, S., Raghu, P. & Hardie, R.C. Polyunsaturated fatty acids activate the Drosophila light-sensitive channels TRP and TRPL. Nature 397, 255−259 (1999). | Article | PubMed | ISI | ChemPort |
- Hilgemann, D.W., Feng, S. & Nasuhoglu, C. The complex and intriguing lives of PIP2 with ion channels and transporters. Sci. STKE RE19 (2001). | PubMed | ChemPort |
- Ordway, R.W., Pallanck, L. & Ganetzky, B. Neurally expressed Drosophila genes encoding homologs of the NSF and SNAP secretory proteins. Proc. Natl Acad. Sci. USA 91, 5715−5719 (1994). | PubMed | ChemPort |
- Loughney, K., Kreber, R. & Ganetzky, B. Molecular analysis of the para locus, a sodium channel gene in Drosophila. Cell 58, 1143−1154 (1989). | Article | PubMed | ISI | ChemPort |
- van der Bliek, A.M. & Meyerowitz, E.M. Dynamin-like protein encoded by the Drosophila shibire gene associated with vesicular traffic. Nature 351, 411−414 (1991). | Article | PubMed | ChemPort |
- Shyngle, J. & Sharma, R.P. Studies on paralysis and development of second chromosome temperature sensitive paralytic mutants of Drosophila melanogaster. Indian J. Exp. Biol. 23, 235−240 (1985). |
|
