*Phosphoinositides and the BBSome*

The BBSome, a heterooctameric protein complex involved in ciliary trafficking, whose defects lead to the blinding ciliopathy, Bardet–Biedl syndrome, binds to membranes, and shows a preference for acidic lipids, including phosphoinositides [67,68], and isolated BBS5 which contains pleckstrin homology domains, binds to phosphoinositides, especially PI(3)P [68]. More recently, it was shown that a core BBSome complex containing BBS 1, 4, 5, 8, 9 and 18 and a smaller sub-complex lacking BBS1 and BBS5 bind phosphoinositides with similar specificities. A caveat for these studies is that the commercial "PIP strips" used have local surface densities of phosphoinositides that far exceed anything found under physiological conditions, so further investigation of phosphoinositide-binding of the BBSome and its sub-complexes is warranted.

### **10. Autophagy and other Stress Responses Involving Redirection of Membrane Tra**ffi**c and Phosphoinositides**

It has been reported that light exposure induces elements of the autophagy pathway in rods and that autophagy plays an important role in photoreceptor homeostasis [69–73]. This pathway has been suggested to be a neuroprotective one that forestalls apoptosis under conditions of stress [69,74]. This process may be part of a more general neuroprotective response involving re-direction of membrane traffic and phosphoinositides. As discussed below, Synaptojanin-1 has been implicated in autophagy in zebrafish cones [75].

### **11. Evidence for E** ff**ects of Light on Phosphoinositide Metabolism**

A number of early reports in the 1980s suggested that light had measurable impacts on PI metabolism in photoreceptors or the retina generally. Based on measurement of 32Pi incorporated into PtdIns by metabolic labeling, it was reported that exposure of isolated frog retina to light decreased levels of PIP2 by 14% after 5 s and 37% after 15 s, while levels of PI(4)P, PtdIns and other acidic lipids remained essentially constant [76]. Subsequent publications from the same group reported increased levels of 3H inositol and 32P into phosphoinositides upon illumination [77]. They also reported PI(4,5)P2-specific phospholipase C (PLC) activity in frog photoreceptors, and PLC immunoreactivity in bovine rod outer segments [78,79], as well as PI-kinase and PIP-kinase activities in frog ROS [19]. Another group found that exposure of rat retinas to light led to decreased staining of rod outer segments with anti-PI(4,5)P2 antibodies [80,81]. The caveats of those experiments are that it is known that physical properties of outer segmen<sup>t</sup> membranes and their protein composition are altered by bright light exposure, and that it is di fficult to establish the specificity of such antibody staining. In addition to light, reports on regulation of photoreceptor PLC by Ca2+ [82] or by subunits of the phototransduction G protein, transducin [83], were published, also suggesting a possible influence of light exposure, which is known to control levels of Ca2+ and active and inactive forms of transducin subunits.

In contrast, another group [84] reported that bovine rod outer segments have very little PIP kinase activity as compared to the rest of the retina (consistent with the previously reported low level of PI(4,5)P2 in outer segmen<sup>t</sup> membranes), and that light adaptation had no measurable e ffect on phosphoinositide metabolism as compared to in vivo dark adaptation. Yet another group, using metabolic labeling with [3H]inositol, reported that light led to decreases in PIP2 levels without generation of InsP3, suggesting light-dependent activity of a phosphatase rather than of phospholipase C [85]. In vitro studies demonstrated that the presence of PI(4,5)P2 in membranes could a ffect the activity of components of the phototransduction cascade [86,87], including the cGMP-gated cation channel and the cGMP phosphodiesterase-transducin complex. The physiological relevance of e ffects of PI(4,5)P2 on phototransduction or of e ffects of light on PI(4,5)P2 metabolism remains untested, and it seems to be possible to explain the entire time-course of rod light responses without invoking any participation by phosphoinositide metabolism [88]. Antibodies specific for isoforms of phospholipase C or <sup>G</sup>αq/<sup>11</sup> PLC-coupled subunits revealed the presence of PLCβ4 and <sup>G</sup>αq/<sup>11</sup> in rod outer segments, and <sup>G</sup>αq and other PKC isoforms elsewhere in the retina [89]. The functional roles of these proteins in outer segments are not known. There have been suggestions that slower e ffects of prolonged light exposure, such as arrestin translocation from the inner to outer segments of rod cells, may be mediated by a phospholipase C cascade [90].

### *Light Regulation of PI-3 Kinase*

Interest has turned toward Type I PI-3 kinase and PI(3,4,5)P3, with reports of e ffects of light on the activity of this enzyme in bovine rod outer segments [91,92], which were reported to be mediated by light-stimulated activation and tyrosine phosphorylation of the insulin receptor [93–95]. Deletion of the p85 α regulatory subunit of Type I PI-3 kinase in cone cells resulted in progressive degeneration of cones, without observable e ffects on rod survival [57]. Likewise, cone-specific inactivation of the gene encoding the p110 α catalytic subunit also resulted in defects in cone survival [96].

In contrast, Type I PI-3 kinase and its product, PI(3,4,5)P3, seem to be much less important for rod function. Rod-specific ablation of the p85 α gene using two di fferent rod-specific Cre transgenes yielded no obvious defects in retinal morphology or rod cell survival [22,97], although modest e ffects on kinetics of light response recovery and arrestin translocation were reported in one case [97]. Quantitative analysis of PI(3,4,5)P3 levels in rods isolated from dark-adapted or light-adapted retinas revealed levels of this phosphoinositide of more than one order magnitude lower than those found for PI(3)P in the light, or at least two orders of magnitude lower than light-stimulated levels of PI(4,5)P2 [22].

### **12. PI(4,5)P2 and Phospholipase C in Intrinsically Photosensitive Ganglion Cells**

In addition to image-forming light detection mediated by rods and cones, the vertebrate retina also contains intrinsically photosensitive retinal ganglion cells. These contain phototransduction cascades reminiscent of that found in invertebrate rhabdomeric photoreceptors [98–104]. Light activates melanopsin, encoded by the *Opn4* gene, a visual pigment which is more closely related to invertebrate opsins than to vertebrate opsin [105]. In M1-type ganglion cells, melanopsin photoisomerization leads to activation of a <sup>G</sup>αq/11/<sup>14</sup> class G-protein, which activates the phosphoinositide-specific phospholipase C isoform, PLCβ4; the phospholipase presumably acts on PI(4,5)P2 as in other cell types, including *Drosophila* photoreceptors, which contain a homologous phospholipase, to produce diacylglycerol and InsP3. PI(4,5)P2 hydrolysis, in turn, leads to the activation of the cation channels TRPC6 and TRPC7 [104]. In M4 ganglion cells, a di fferent phototransduction cascade involving cyclic nucleotides and cyclic nucleotide-regulated HCN channels, whereas in M2 ganglion cells, both of these cascades operate [102].

### **13. Studies of PI Metabolism in the RPE**

A variety of extracellular stimuli acting on tyrosine kinase-associated receptors or G protein-coupled-receptors have been reported to stimulate release of inositol phosphates in cultured RPE cells, presumably derived from PLC action of PI(4,5)P2, on a timescale of tens of minutes; e ffective stimuli included fetal bovine serum, agonists for muscarinic, histamine, and serotonin, peptides, including bradykinin, arginine vasopressin, bombesin and oxytocin, [106–109]. The physiological relevance of these observations was not explored, but an in vivo study using frogs demonstrated dramatic acceleration of inositol phosphate release, especially of InsP3, following stimulation by light [21]. Acutely isolated rat RPE cells were reported to release InsP3 in response to induction of phagocytosis by addition of isolated rod outer segments [110]; this InsP3 release was not observed in cells from Royal College of Surgeon (RCS) rats, which have a defect in OS phagocytosis due to a deficiency in the receptor tyrosine kinase, MERTK [111]. As in many other cell types, insulin has been reported to stimulate activity of Type I PI-3 kinase to produce PI(3,4,5)P3 [95,112,113]. Responses to hypoxia [112,114,115] and elevated glucose [116,117] are also reported to involve this pathway in RPE.

A number of important processes in RPE are known to rely on phosphoinositides, but how they are regulated in these cells is not well understood. Phagocytosis, autophagy, endocytosis and endosome processing, establishment of epithelial cell polarity and extension of microvilli membranes are all known to critically depend on phosphoinositides. For example, both autophagy and phagocytosis involve the recruitment of the ubiquitin-like protein, LC3 [118–120], whose recruitment to membranes depends on PI(3)P. Phagocytosis is also thought to require PI(4,5)P2 and lysosomal fusion may involve other phosphoinositides such as PI(3,5)P2 and PI(5)P [22,121–125].

### **14. Phosphoinositide Kinases and Phosphatase**

In mammals, there are 47 genes encoding 19 PI-kinase and PIP-kinases and 28 PIP phosphatases [126].
