During receptor-activated cell signaling cascades, it is important that the resynthesis of biochemical substrates is tightly coupled to their consumption by enzymatic reactions. This is particularly necessary in the context of neuronal signaling cascades, where at high rates of receptor stimulation, significant depletion of substrates might otherwise occur. An essential step in Drosophila phototransduction is the hydrolysis of phosphatidylinositol 4,5 bisphosphate (PI(4,5)P₂) by phospholipase Cβ (PLCβ) to generate a second messenger that opens the light-activated channels TRP and TRPL. Although the identity of this messenger remains unknown, recent evidence has implicated diacylglycerol kinase (DGK), encoded by rdgA, as a key enzyme that regulates levels of the second messenger, mediating both amplification and response termination within photoreceptor cells. This study demonstrates that lazaro (laza: from the Spanish novel Lazarillo de Tormes, in which Lazaro is the boy servant of a blind man and helps him to see) encodes a lipid phosphate phosphohydrolase (LPP) that functions during phototransduction. Whereas RdgA converts DAG to phosphatidic acid (PA) by adding phosphate residues to glycerol, Laza carries out the reverse reaction by removing the phosphate residue. In other words, DGK and LPP operate as a kinase/phosphatase pair that controls the level of PA. The activity of Laza laza rescues degeneration in the blind mutant rdgA and works synergistically with it to regulate amplification and response termination during phototransduction. Analysis of retinal phospholipids revealed a reduction in PA levels and an associated reduction in phosphatidylinositol (PI) levels. Together these results demonstrate the contribution of PI depletion to the rdgA phenotype and provide evidence that depletion of PI and its metabolites might be a key signal for TRP channel activation in vivo (Garcia-Murillas, 2006). Similar results have been obtained by (Kwon and Montell, 2006)

Calcium influx through plasma membrane channels regulates a range of functions in both adult and developing neurons, including the transduction of a number of sensory modalities, learning and memory, and the navigation of dendrites during neuronal development. A number of different ion channels mediate calcium influx that subserves these functions; one major family are the TRP proteins. Several classes of TRP channels have been described, and members of all classes mediate key physiological processes in neurons. Despite the importance of TRP proteins in neuronal function, the signaling mechanisms that regulate channel activity in vivo remain controversial and poorly understood. A common theme underpinning the regulation of several classes of TRP channels is the role of PI(4,5)P₂ and/or its metabolites. When heterologously expressed, TRPC, TRPV, and TRPM channels are all reported to be modulated by changes in PI(4,5)P₂ levels during signaling. However, the role of PI(4,5)P₂ depletion in the activation of endogenous TRP channels remains limited by the lack of suitable in vivo models (Garcia-Murillas, 2006).

Drosophila phototransduction is a well-established model system for the analysis of calcium influx triggered by G protein-coupled phosphoinositide hydrolysis (Hardie, 2001). In photoreceptors, the absorption of a photon of light by rhodopsin triggers PLCβ activity. This signaling cascade triggers the opening of at least two classes of plasma membrane channels, Trp and Trpl, resulting in the influx of calcium into the cell. Despite intense investigation, the mechanism by which PLCβ activity results in Trp and Trpl opening remains unclear. The essential role of PI(4,5)P₂ hydrolysis in phototransduction is well-established; null mutants in norpA, which lack eye-enriched PLCβ, show no response to light. The hydrolysis of PI(4,5)P₂ generates the second messengers inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol (DAG). IP₃ is insufficient for activation of Trp and Trpl in vivo, and the IP₃ receptor does not appear to be essential for phototransduction. However, recent studies have suggested that lipid second messengers generated by PLCβ activity might play a role in Drosophila Trp and Trpl activation. Application of both DAG and PI(4,5)P₂ has been reported to modulate heterologously overexpressed Trpl channels (Estacion, 2001), and polyunsaturated fatty acids (Chyb, 1999), metabolites of DAG, can activate Trp and Trpl (Garcia-Murillas, 2006).

In most eukaryotic cells, DAG is phosphorylated by DGK to generate PA. In Drosophila, rdgA mutants that lack eye-enriched DGK show severe retinal degeneration that can be rescued by trp mutants. The retinal degeneration phenotype of rdgA has recently been shown to be modulated by light, Gq, and PLCβ (Georgiev, 2005), all of which are key elements required for PI(4,5)P₂ hydrolysis during phototransduction. Electrophysiologically, rdgA mutant photoreceptors show constitutive Trp channel activity as well as defects in amplification and response termination. Thus, the analysis of the rdgA mutant strongly suggests that in vivo the tight regulation of PI(4,5)P₂-derived lipid metabolites is essential for normal Trp channel activation. While the analysis of rdgA demonstrates the key role of DGK in this process, the biochemical and molecular basis for its requirement is unclear. In particular, it has thus far not been possible to identify the specific lipid messenger(s) responsible for the abnormal Trp channel activation in rdgA. One possibility is that reduced DGK activity results in an elevation of DAG and/or its metabolites in the face of ongoing PI(4,5)P₂ hydrolysis. Alternatively, it is possible that a reduction in levels of PA, the product of DGK, is a key determinant of the rdgA phenotype. In addition to being a potential second messenger in its own right, PA can also be recycled to resynthesize PI(4,5)P₂. PA can bind to a number of proteins and in some cases regulate their activity. These include two molecules involved in PI turnover: type I phosphatidylinositol 4 phosphate 5 kinase (type I PIPkin) in mammalian cells and Opi1p, an ER resident transcription factor in yeast. Opi1p, when bound to PA, translocates to the nucleus and stimulates the transcription of genes regulating phosphatidylinositol (PI) biosynthesis. Thus, it is likely that PA levels would be tightly regulated and linked to PI turnover in vivo. It also raises the possibility that a depletion of PI and consequently PI(4,5)P₂ might underlie the rdgA phenotype (Garcia-Murillas, 2006).

As part of a study to understand the biochemical basis of the rdgA phenotype, additional enzymes encoded in the Drosophila genome were examined that might regulate lipid turnover during PI(4,5)P₂ hydrolysis. An immediate possibility is that lipid phosphate phosphohydrolases (LPPs) might act in synergy with DGK in controlling the balance of DAG and PA. LPPs were originally designated as type II PA phosphatases (PAP), since their activity was distinct from the type I PAP activity involved in glycerolipid biosynthesis. However, to date it is unclear whether LPPs can metabolise PA, an intracellular bioactive lipid generated during PI(4,5)P₂ hydrolysis. This hypothesis was tested and the endogenous LPP that functions during Drosophila phototransduction was identified. The biochemical consequence is described of the concerted action of DGK and LPP during phototransduction and its implications for the mechanism of Trp channel activation (Garcia-Murillas, 2006).

During receptor-activated cell signaling cascades, it is important that the resynthesis of biochemical substrates is tightly coupled to their consumption by enzymatic reactions. This is particularly necessary in the context of neuronal signaling cascades, where at high rates of receptor stimulation, significant depletion of substrates might otherwise occur. For example, in bright daylight, Drosophila photoreceptors are able to successfully detect ~10⁶ photons s⁻¹ without significant loss of sensitivity. In the context of a G protein-coupled phosphoinositide signaling cascade, the hydrolysis of the substrate PI(4,5)P₂, a minor membrane phospholipid, needs to be coupled to its resynthesis by the sequential phosphorylation of a cellular pool of PI. Thus, an adequate rate of PI(4,5)P₂ resynthesis requires among other things the maintenance of this pool of PI. PI is synthesized in the ER by the condensation of inositol and CDP-DAG by PI synthase. Previous studies have suggested the importance of a constant supply of inositol generated from IP₃ to maintain ongoing G protein-coupled PLC signaling in neurons. However, the role, if any, for the recycling of DAG to supply CDP-DAG required for PI synthesis has thus far not been addressed. In addition to being an allosteric activator of PKC, PA can be recycled into PI resynthesis by the sequential action of CDP-DAG synthase and PI synthase. In the present study, it was found found that (1) in rdgA³ photoreceptors, levels of total PA are reduced and that (2) this reduction is enhanced by overexpressing LPP. Together with previous findings that DAG levels are not elevated in the rdgA, these observations demonstrate the critical role of PA levels generated by DGK in the rdgA phenotype. However, the findings that elevation of PA levels via the patch pipette does not suppress constitutive Trp channel activity and that elevation of PA levels using the cds¹ mutant did not rescue rdgA³ strongly suggest that PA does not directly influence Trp channel activity or contribute to the degeneration phenotype (Garcia-Murillas, 2006).

How might reduced levels of photoreceptor PA contribute to the rdgA phenotype? In this study, two observations provide an insight into this question. (1) The cds¹ mutant that elevates photoreceptor PA levels while reducing PI and PIP levels enhances rather than suppressed degeneration in rdgA³. (2) The reductions in PA level are associated with a change in PI levels in rdgA³ retinae, and PI levels undergo a dramatic reduction when PA levels were further reduced by LPP overxpression. Taken together, it is likely that these findings reflect the requirement for a constant supply of PA generated by rdgA so that CDP-DAG can be generated by CDP-DAG synthase to be used as a substrate for PI synthesis. The finding that transcript levels of PI synthase in photoreceptors are directly correlated with PA levels suggests one mechanism by which PA might regulate PI resynthesis linked to the level of ongoing PI(4,5)P₂ hydrolysis. In this analysis, colocalization of the enzyme within the cell body in a pattern highly reminiscent of the ER within these cells was found. This membrane compartment is the site at which the resynthesis of PI(4,5)P₂ is initiated by the enzyme CDP-DAG synthase using PA as the substrate. In photoreceptors, DGK (encoded by rdgA) is also localized on the submicrovillar cisternae thought to be a specialization of the ER. Thus, three key enzymes that together contribute to the regulation of cellular PA and generation of PI appear to be localized to the ER, consistent with a role for PA in stimulating resynthesis of PI (Garcia-Murillas, 2006).

Since the sequential phosphorylation of PI is a major route of PI(4,5)P₂ generation, changes in PI levels could impact on photoreceptor PI(4,5)P₂ resynthesis. To test this possibility, the effect of laza activity was analyzed in rdgB, a mutant with reduced rates of PI(4,5)P₂ resynthesis at the rhabdomeral plasma membrane (Hardie, 2001). The effect was analyzed of altering PA levels by both overexpressing LPP as well as using laza mutants to remove the endogenous LPP that functions during phototransduction and it was found that the retinal degeneration phenotype of rdgB can be modulated by levels of LPP activity in the retina. Together with biochemical data on the effects of altered PA (the substrate of Laza) levels in photoreceptors, these results suggest a role for PA levels in regulating the supply of PI to be transported to the plasma membrane for PI(4,5)P₂ resynthesis during phototransduction. How might PA levels modulate the rdgB phenotype? It has previously been shown that PA does not bind to and alter phosphatidylinositol-transfer protein (PITP) function in vitro. One obvious mechanism is that the level of PI (a key substrate that is transported by PITP) is altered by changes in laza activity. An alternative mechanism is suggested by the finding that in vitro PA binds to and stimulates type I phosphatidylinositolphosphate kinase (PIPkin) activity. The finding that reduced type 1 PIPkinase activity (sktlΔ20/sktlΔ1-1) enhances the rate of degeneration in rdgA³ supports this mechanism. Thus, there are at least three mechanisms by which PA derived from PI(4,5)P₂ can stimulate PI(4,5)P₂ resynthesis: (1) as a unique substrate for CDP-DAG synthesis, (2) as a transcriptional regulator of PI synthase, (3) as an allosteric regulator of type I PIPkin. In the present study, evidence is provided in support of these three mechanisms in photoreceptors. It is therefore proposed that the generation of PA by DGK during PLCβ-mediated signaling provides a key signal to regulate the resynthesis of PI(4,5)P₂ (Garcia-Murillas, 2006).

Given the critical role of PA during phototransduction, LPP was identified as an enzyme that might work in conjunction with DGK to regulate PA levels. LPP was originally designated as a type II PA phosphatase (PAP) (Jamal, 1991; Kai, 1996) and shows remarkable substrate promiscuity in vitro (reviewed in Brindley, 1998). When overexpressed in cell culture models, LPPs are able to dephosphorylate extracellular lipids presented to intact cells (Jasinska, 1999 and Roberts, 1998) and localize mainly to the plasma membrane (Alderton, 2001; Burnett, 2003; Jasinska, 1999), with residues involved in catalysis facing externally (Zhang, 2000). These findings have led to the idea that LPPs are ectoenzymes catalyzing the hydrolysis of extracellular bioactive lipid, although, to date, the in vivo substrate of any LPP has not been identified. In particular, it is unclear whether LPPs can metabolise PA generated by hydrolysis of PI(4,5)P₂. During this study, the enzymatic activity was analyzed of a number of LPP gene products during Drosophila phototransduction. Overexpressing four different Drosophila LPPs individually can exacerbate the phenotype of hypomorphic mutants in DGK. Significantly, the effects of LPP overexpression are evident only in the sensitized background of DGK loss of function, providing strong evidence that LPP can function as antagonists of DGK. The finding that the reduced levels of PA in rdgA³ retinae are further reduced on LPP overexpression provides biochemical support for the genetic data. Further, it was also demonstrate that laza²² retinae show light-induced accumulation of PA. This finding supports the idea that laza regulates PA levels during phototransduction. Together, these findings provide compelling evidence that in vivo DGK and LPP operate as a kinase/phosphatase pair that controls the level of PA. The finding that overexpression of LPP in flies lacking both DGK and PLCβ in their photoreceptors (norpA^P24, rdgA³) did not result in retinal degeneration strongly supports the conclusion that in vivo LPP can dephosphorylate PA generated from PI(4,5)P₂ hydrolysis by the sequential enzymatic activity of PLCβ and DGK. This study identifies an in vivo substrate of LPP and also demonstrates a role for LPP in regulating the DAG/PA balance during G protein-coupled PI(4,5)P₂ signaling (Garcia-Murillas, 2006).

During this study, it was found that when overexpressed LPPs encoded by a number of different gene products, including CG11426, laza, wun, and wun2, are all able to antagonize DGK function, reminiscent of the ability of all three isoforms of vertebrate LPPs (LPP1, LPP2, and LPP3) to dephosphorylate PA (Roberts, 1998). In contrast, it was found, using loss-of-function analysis, that the only LPP that is able to functionally antagonise DGK and hence suppress degeneration in rdgA is the eye-enriched LPP laza. It has recently been shown that when overexpressed, LPPs can form oligomers (Burnett, 2004). It is possible that high levels of expression result in the assembly of oligomers that lack substrate specificity (Garcia-Murillas, 2006).

A common but poorly resolved theme underlying the signaling mechanisms regulating TrpC channel activation is the role of PI(4,5)P₂ or its metabolites in activation. Recently, a number of studies have shown that a range of lipids generated by PI(4,5)P₂ hydrolysis can activate both vertebrate and invertebrate TrpC channels when applied exogenously to overexpressed channels (reviewed in Hardie, 2003). However, to date, the activation of endogenous TrpC channels has been studied in only two model systems: activation of Trp and Trpl during Drosophila phototransduction and the activation of TrpC2 channels in neurons of the mouse vomeronasal organ. In both of these systems, the inactivation of DGK (using the rdgA mutant in Drosophila and pharmacology in the mouse study; Lucas, 2003; Raghu, 2000) results in a sustained activation of TrpC channels in the face of ongoing PI(4,5)P₂ hydrolysis. While these findings are consistent with a role for DGK in regulating response termination, they have thus far not allowed a definitive analysis of the messenger of activation. In the mouse vomeronasal organ, the finding that TrpC2-like channels can be activated by exogenous DAG (Lucas, 2003) implies that the build-up of this messenger is the biochemical basis of abnormal TrpC channel activation when DGK is inhibited. However, a build-up of DAG in the vomeronasal organ during DGK inhibition has yet to be demonstrated. A previous study showed that in Drosophila photoreceptors of the rdgA mutant, there was no elevation of DAG (Inoue, 1989), while demonstrating that light-induced PA levels were reduced. In the present study, it was also demonstrated that in addition to a reduction in retinal PA levels, there is also a reduction in the level of PI, the precursor for PI(4,5)P₂ resynthesis, and an associated change in the levels of PI synthase, a key enzyme for PI resynthesis. These findings raise the possibility that the reduced level of PI might also result in a reduction in the rate of PI(4,5)P₂ resynthesis. The finding that reduction in type I PIP kinase activity enhances degeneration in rdgA³ supports this hypothesis. Previous studies have shown that both rdgB photoreceptors as well as wild-type photoreceptors (in zero extracellular calcium) show prolonged activation of Trp channels following a response to light (Hardie, 2001). In addition, Trpl channels expressed in S2 cells have been reported to be inhibited by PI(4,5)P₂ (Estacion, 2001). Taken together with the findings reported in this paper, it is proposed that an imbalance between the rate of PI(4,5)P₂ hydrolysis by PLCβ and its resynthesis from PA might underlie the excessive activation of endogenous Trp channels in photoreceptors of the rdgA mutant (Garcia-Murillas, 2006).

GENE STRUCTURE

cDNA clone length - 2368

Bases in 5' UTR - 625

Exons - 1

Bases in 3' UTR - 738

PROTEIN STRUCTURE

Amino Acids - 334

Structural Domains

To identify LPP that might function during phototransduction, a bioinformatic analysis was carried out of the completed Drosophila genome. A BLAST search using human LPP2 (Hooks, 1998) identified seven genes that showed homology to LPP. Two of these genes, wunen (wun) and wunen-2 (wun2), have been studied extensively for their role in Drosophila germ cell migration. When expressed in cell culture models, they show a catalytic activity profile similar to that of vertebrate LPP. Five others, namely CG11425, CG11426, CG11437, CG11438, and CG11440, are clustered within an ~8 kb genomic region at 79E on chromosome 3. CG11440 was termed lazaro. Analysis of the protein sequences of Drosophila LPP-like genes shows that they all contain six putative transmembrane domains as well as three domains previously identified as being characteristic of vertebrate LPP (Brindley, 1998). Indeed, 5 of the 7 residues shown to be essential for catalysis in human LPP arte identical in all seven Drosophila genes and the three human orthologs. To identify LPPs that are expressed in the retina, RT-PCR was performed on dissected retinae. This revealed that RNA for five of the seven Drosophila LPP genes, wun, wun2, laza, CG11425, and CG11426, are expressed in adult photoreceptors. No expression could be detected for the other two genes, CG11437 and CG11438. These results show that LPPs are expressed in adult Drosophila photoreceptors and might play a signaling role in these cells. Whether transcripts for any of these LPP genes are enriched in the retina, usually an indication of a role in photoreceptor function, was examined. Transcript levels were compared for all five genes in total RNA from wild-type heads with those from so^D (a mutant that lacks eyes) using semiquantitative RT-PCR. This analysis revealed that only laza transcripts are enriched in the eye (Garcia-Murillas, 2006).

date revised: 1 October 2006

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.