Drosophila visual transduction has served as a paradigm to characterize G protein-coupled neuronal signaling. As in mammals, light-activated rhodopsin is phosphorylated and interacts with a rhodopsin regulatory protein, arrestin, which facilitates deactivation of the receptor. However, unlike mammalian phototransduction, light activation in Drosophila is coupled to stimulation of phospholipase C rather than a cGMP-phosphodiesterase. The visual arrestin, undergoes light-dependent trafficking in mammalian and Drosophila photoreceptor cells, though the mechanisms underlying these movements are poorly understood. In Drosophila, the movement of the visual arrestin, Arr2, functions in long-term adaptation and is dependent on interaction with phosphoinositides (PIs). However, the basis for the requirement for PIs for light-dependent shuttling has been unclear. This study demonstrates that the dynamic trafficking of Arr2 into the phototransducing compartment, the rhabdomere, requires the eye-enriched myosin III, NINAC. Defects in ninaC result in a long-term adaptation phenotype similar to that which occurs in arr2 mutants. The interaction between Arr2 and NINAC is PI dependent and NINAC binds directly to PIs. These data demonstrate that the light-dependent translocation of Arr2 into the rhabdomeres requires PI-mediated interactions between Arr2 and the NINAC myosin III (Lee, 2004a).

The activity of signaling cascades can be profoundly affected by modulating the levels of regulatory proteins in specialized cellular compartments. The concentrations of such proteins can be altered through changes in protein synthesis or degradation. However, a faster mechanism for regulating the levels of a signaling protein involves protein trafficking in response to agonist stimulation. Dynamic movements of signaling proteins are of particular importance to neurons, such as photoreceptors, which are highly polarized and respond to their external stimulus, light, with great rapidity (Lee, 2004a).

During the last two decades, it has become clear that several key signaling proteins in mammalian and Drosophila photoreceptor cells undergo light-dependent translocations in and out of the phototransducing compartments, the outer segments, and rhabdomeres, respectively. These include the Drosophila TRPL cation channel (Bähner et al., 2002) and the Gα and Gβ subunits (Sokolov, 2002; Kosloff, 2003), which are concentrated in the phototransducing compartments only in dark-adapted animals. Upon exposure to light stimulation, these proteins shuttle into the cell bodies or inner segments over the course of several minutes. Visual arrestin, which participates in termination of the photoresponse, also undergoes dramatic light-dependent translocation, though the direction of movement is opposite to that of the TRPL and the Gα and Gβ subunits. In response to light stimulation, arrestin migrates into the outer segments/rhabdomeres of vertebrate and Drosophila photoreceptor cells (Broekhuyse, 1985; Philp, 1987; Whelan, 1988; Kiselev, 2000; Lee, 2004a and references therein).

The light-dependent translocations of signaling proteins in photoreceptor cells appear to function in long-term light adaptation (reviewed in Arshavsky 2003; Hardie 2003). Photoreceptor cells adapt to increasing intensities of background illumination by increasing the rate of termination of the photoresponse. The rate of termination of the light response is relatively slow in animals that are initially dark-adapted. Prior exposure to background illumination accelerates the rate of response termination. However, mutations that decrease the rate of translocation of the Drosophila visual arrestin from the cell body to the rhabdomere cause corresponding defects in this mode of adaptation (Lee, 2003; Lee, 2004a and references therein).

A critical question concerns the mechanisms underlying the light-dependent shuttling of signaling proteins, such as visual arrestin. Since mammalian rhodopsin undergoes light-dependent phosphorylation, which promotes binding to visual arrestin (reviewed in Arshavsky, 2002), it has been proposed that visual arrestin becomes concentrated in the outer segments due to binding to phosphorylated rhodopsin. However, it has been shown recently that translocation of visual arrestin into the outer segments occurs normally in rhodopsin kinase-deficient mice (Mendez, 2003; Zhang, 2003). It has been shown that the major Drosophila visual arrestin, Arr2, binds to phosphoinositides (PIs) and that this interaction is necessary for light-induced trafficking. However, the basis for this PI requirement and the identity of proteins that promote the light-dependent translocation of Arr2 or any other signaling protein have not been previously identified (Lee, 2004a).

It has now been shown that the NINAC myosin III, which consists of linked protein kinase and myosin head domains (Montell, 1988), is required for movement of Arr2 into the rhabdomeres. NINAC is expressed as two isoforms, p132 and p174, which are detected exclusively in the cell bodies and rhabdomeres, respectively (Porter, 1992). While p174 is required for response termination (Porter, 1992), no role for p132 in the photoresponse has been described. This study shows that p132 is the primary isoform necessary for light-dependent trafficking of Arr2 into the rhabdomeres. Moreover, flies that do not express p132 display a defect in long-term adaptation, consistent with its role in Arr2 shuttling. This study shows that the interaction between Arr2 and PIs is required for Arr2 to interact with NINAC, NINAC is also a PI binding protein, and that these interactions promoted the light-dependent movement of Arr2 into the rhabdomeres. These data indicate that the light-dependent shuttling of Arr2 into the rhabdomeres requires an association with the NINAC myosin III that is mediated through PIs (Lee, 2004a).

In contrast to the Lee (2004a) study, a study by Satoh (2005) found no requirement for NINAC in Arr1 or Arr2 translocation. The Satoh study concludes that the two Drosophila photoreceptor arrestins mediate distinct and essential cell pathways downstream of rhodopsin activation. The majority arrestin, Arr2, quenches rhodopsin signaling, while Arr1 promotes light-induced rhodopsin endocytosis. It is proposed that Arr1 mediates an endocytotic cell-survival activity, scavenging phosphorylated rhodopsin, thereby countering toxic Arr2/Rh1 accumulation; elimination of toxic Arr2/Rh1 in double mutants could thus rescue arr1 mutant photoreceptor degeneration (Satoh, 2005). Light-dependent subcellular translocation of Gqalpha in Drosophila photoreceptors has been shown to be facilitated by NINAC (Cronin, 2004).

Therefore, NINAC represents the first protein required for light-dependent movement of any signaling protein in vertebrates or invertebrates. In the ninaC null mutant (ninaC^P235), there was no detectable increase in rhabdomeral Arr2, even after a 1 hr exposure to light. Consistent with the requirement for NINAC for Arr2 translocation, long-term light adaptation is severely disrupted in the ninaC null mutant, ninaC^P235. This defect is more pronounced than that observed in mutant flies expressing a derivative of Arr2 (Arr2^3K/Q), which displays a large reduction in PI binding (Lee, 2003). The stronger light adaptation phenotype in ninaC^P235 than in arr2^3K/Q is consistent with the findings that the PI/Arr2 interaction and Arr2 translocation is reduced but not eliminated in arr2^3K/Q photoreceptor cells (Lee, 2004a).

The current results also address a long-standing question concerning the role for the cell body-enriched isoform of NINAC p132. NINAC p174 is enriched in the rhabdomeres and is required for termination of the phototransduction (Porter, 1992). However, a role for p132 in the photoresponse had not been described, which was not surprising given that p132 is detected exclusively in the cell bodies and phototransduction takes place in the rhabdomeres. Nevertheless, while both NINAC proteins participate in the light-dependent translocation of Arr2 into the rhabdomeres, it was found that p132 is the primary isoform required. This conclusion is further supported by the long-term adaptation defect in ninaC^Δ132 flies (Lee, 2004a).

Both NINAC isoforms consist of linked protein kinase and myosin domains, either of which might promote Arr2 shuttling. It is suggested that the myosin domain functions in the trafficking of Arr2, since a lysine to arginine mutation in the protein kinase domain, which eliminates enzymatic activity in other protein kinases, has no impact on either Arr2 translocation or long-term adaptation. The relatively rapid movement of Arr2 into the rhabdomeres, which occurs over the course of a few minutes, suggests that the myosin motor activity may promote the movement. Human photoreceptor cells express a NINAC homolog, Myo3A (Dosé, 2000), which moves toward the plus end of actin filaments (Komaba, 2003). As is typical of other microvillar structures, the plus end of the filaments is oriented near the distal tips of the microvilli. By analogy to Myo3A, Drosophila NINAC also is likely to be a plus-ended myosin, which could potentially shuttle cargo from the cell bodies into the rhabdomeres. However, it was not possible to demonstrate that the NINAC motor activity per se was required for movement of Arr2; point mutations in conserved residues required for motor activity of other myosins, such as those in the actin and ATP binding sites, cause instability of NINAC in vivo (Lee, 2004a).

Translocation of Arr2 has been shown to require direct interaction with PIs (Lee, 2003); however, the basis of this requirement has been unclear. In the current work, the association of Arr2 and NINAC was shown to be dependent on PIs and NINAC was shown to bind PIs. In support of this conclusion, it was found that the Arr2/NINAC interaction is disrupted by detergent and further augmented by the addition of exogenous PIs. Furthermore, NINAC binds to PIs in vitro. The interaction between NINAC and Arr2 is dependent on PIs and not on calmodulin. Calmodulin could not serve as a link between NINAC and Arr2; it was shown that Arr2 does not bind calmodulin. Although the NINAC proteins used for the biochemistry were characterized as partially purified, p132 and p174 were the only proteins clearly seen on a Coomassie-stained gel. Therefore, all other proteins were present at substoichiometric concentrations and would therefore not be effective linkers between NINAC and Arr2. In addition, p132 has only one calmodulin binding site. Consequently, the association of Arr2 to p132 immobilized on calmodulin-agarose could not have occurred through a second calmodulin bound to p132 (Lee, 2004a).

It is proposed that p132 binds to PI-containing vesicles, which bind simultaneously to Arr2, facilitating the trafficking of Arr2 into the rhabdomeres. Interestingly, there is an age-dependent accumulation of vesicles in the cell bodies of ninaC^Δ132 (Hicks, 1996), which may result from the absence of p132-dependent vesicular movement into the rhabdomeres of these mutant flies. Although the null allele, ninaC^P235, displays slow termination of the light response and retinal degeneration, which could indirectly affect Arr2 movement, ninaC^Δ132 flies do not undergo retinal degeneration or exhibit defects in activation or termination of the photoresponse (Porter, 1992). While it cannot be excluded that changes in calmodulin distribution associated with elimination of p132 (Porter, 1992) contribute to the defect in Arr2 translocation, several observations support the conclusion that p132 functions directly in light-dependent movement of Arr2. These include the findings that p132 associates with Arr2, the demonstration that this association is dependent on PIs, and the previous report that the Arr2-PI interaction is required for translocation (Lee, 2003; Lee, 2004a).

A remaining question concerns the mechanism underlying the reciprocal translocation of Arr2 from the rhabdomeres to the cell bodies in the dark. While p132 is required for the rapid light-dependent shuttling of Arr2 into the rhabdomeres, it was dispensable for the dark-associated movement of Arr2 out of the rhabdomeres. One possibility is that the transfer of Arr2 from the rhabdomeres to the cell bodies may occur via myosin VI, which is a minus end-directed motor. However, photoreceptor cell-specific expression of a myosin VI antisense transgene, which has been shown to suppress myosin VI activity in vivo, has no impact on Arr2 shuttling back to the cell bodies. Alternatively, the retrograde movement of Arr2 out of the rhabdomeres may occur by diffusion, since this movement occurs on the order of several hours, while the NINAC-dependent trafficking into the rhabdomeres is essentially complete in 10 min (Lee, 2004a).

The current study raises the question as to whether light-stimulated translocation of vertebrate visual arrestin into the outer segment of photoreceptor cells occurs through interaction with a myosin or kinesin and, if so, whether the association is PI dependent. Mammalian visual arrestin binds to inositol phosphates (Palczewski, 1991) and associates with vesicle-like structures in the inner segments (McGinnis, 2002). Kinesin-II is one candidate that could potentially participate in light-stimulated trafficking, because both visual arrestin and opsin accumulate in the inner segments in KIF3A knockout mice (Marszalek, 2000). However, it is not known if the kinesin-II participates in light-induced translocation of visual arrestin or whether the observed effect on arrestin localization is due to retinal degeneration. Unlike kinesin-II-deficient photoreceptor cells, absence of NINAC p132 in Drosophila does not lead to retinal degeneration. It is intriguing to speculate that the vertebrate Myo3A or Myo3B might function in a manner analogous to NINAC. Consistent with this possibility are the observations that the cilium connecting the inner and outer segments contains actin, in addition to tubulin. Moreover, Myo3B is expressed in the retina, and Myo3A is enriched in photoreceptor cells in addition to the cochlea. Recently, one form of nonsyndromic deafness has been attributed to mutations in human MYO3A. Whether these individuals also have a defect in long-term light adaptation is an open question, which remains to be addressed (Lee, 2004a and references).

GENE STRUCTURE

Bases in 5' UTR - 50

Exons - 3

Bases in 3' UTR - 65

PROTEIN STRUCTURE

Amino Acids - 401

Structural Domains

The gene encoding the 49-kilodalton protein that undergoes light-induced phosphorylation in the Drosophila photoreceptor has been isolated and characterized. The encoded protein has 401 amino acid residues and a molecular mass of 44,972 daltons, and it shares approximately 42 percent amino acid sequence identity with arrestin (S-antigen), which has been proposed to quench the light-induced cascade of guanosine 3',5'-monophosphate hydrolysis in vertebrate photoreceptors. Unlike the 49-kilodalton protein, however, arrestin, which appears to bind to phosphorylated rhodopsin, has not itself been reported to undergo phosphorylation. In vitro, Ca2+ was the only agent found that would stimulate the phosphorylation of the 49-kilodalton protein. The phosphorylation of this arrestin-like protein in vivo may therefore be triggered by a Ca2+ signal that is likely to be regulated by light-activated phosphoinositide-specific phospholipase C (Whelan, 1988).

Absorption of a photon of light by rhodopsin triggers mechanisms responsible for excitation as well as regulation of the phototransduction cascade. Arrestins are a family of proteins that appear to be responsible for terminating the active state of G-protein-coupled receptors. One of the major substrates of light-dependent phosphorylation in the visual cascade of Drosophila was purified and partially sequenced. The complete primary structure of the protein was determined by isolating the corresponding gene, which revealed it to be a new isoform of arrestin, Arr2. Arr2 is 401 residues in length, and shares 47% sequence identity with the Drosophila Arr1 protein and 42% with human arrestin. The visual arrestins required for Drosophila phototransduction are more closely related to ß-arrestins than to the mammalian visual arrestins. The two Drosophila arrestin genes are differentially regulated, and Arr2 is a specific substrate for a calcium-dependent protein kinase. This is the first demonstration of in vivo regulation of arrestins in a transduction cascade, and provides a new level of modulation in the function of G-protein-coupled receptors (le Vine, 2000).

date revised: 22 January 2006

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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