Date Published: September 28, 2018
Publisher: Public Library of Science
Author(s): Olaf Voolstra, Lisa Strauch, Matthias Mayer, Armin Huber, Alfred S. Lewin.
Drosophila retinal degeneration C (RDGC) is the founding member of the PPEF family of protein phosphatases. RDGC mediates dephosphorylation of the visual pigment rhodopsin and the TRP ion channel. From the rdgC locus, three protein isoforms, termed RDGC-S, -M, and -L, with different N-termini are generated. Due to fatty acylation, RDGC-M and -L are attached to the plasma membrane while RDGC-S is soluble. To assign physiological roles to these RDGC isoforms, we constructed flies that express various combinations of RDGC protein isoforms. Expression of the RDGC-L isoform alone did not fully prevent rhodopsin hyperphosphorylation and resulted in impaired photoreceptor physiology and in decelerated TRP dephosphorylation at Ser936. However, expression of RDGC-L alone as well as RDGC-S/M was sufficient to prevent degeneration of photoreceptor cells which is a hallmark of the rdgC null mutant. Membrane-attached RDGC-M displayed higher activity of TRP dephosphorylation than the soluble isoform RDGC-S. Taken together, in vivo activities of RDGC splice variants are controlled by their N-termini.
Drosophila retinal degeneration C (RDGC) is the founding member of the protein phosphatases with EF hands (PPEF) family. In rdgC null mutant flies, rhodopsin becomes light-dependently hyperphosphorylated . This hyperphosphorylation causes light-dependent retinal degeneration and premature entrance into the prolonged depolarizing afterpotential (PDA) [1–3]. A PDA manifests in the persistence of photoreceptor depolarization after cessation of the light stimulus and is caused by an excess of activated rhodopsin (metarhodopsin) molecules over available arrestin molecules. Light-dependent rhodopsin hyperphosphorylation in rdgC mutant flies results in a stable interaction between rhodopsin and arrestin 2 molecules. These rhodopsin/arrestin complexes are internalized from the rhabdomeric membranes into the cell body and trigger apoptosis leading to photoreceptor degeneration [4–6]. The stable interaction between arrestin 2 and rhodopsin probably also underlies the premature entrance into the PDA since it reduces the number of available arrestin molecules. RDGC is activated by Ca2+ through interaction with calmodulin . RDGC contains an IQ-motif that binds to Ca2+/calmodulin and a mutation in this motif, rdgCI12E (but not rdgCI12A), abolished calmodulin binding and resulted in retinal degeneration . Besides dephosphorylation of rhodopsin, we recently showed that RDGC is also involved in the dephosphorylation of the TRP cation channel at Ser936 . Interestingly, Ser936 is the only known TRP phosphorylation site that is phosphorylated in the dark and becomes dephosphorylated in the light [8–10], consistent with Ca2+-dependent activation of RDGC. As revealed by exchange of Ser936-TRP to Ala, preventing phosphorylation, or Asp, mimicking phosphorylation, dephosphorylation of TRP at Ser936 enables photoreceptor cells to process oscillating light stimuli of higher frequencies. Thus, the phosphorylation state of TRP at Ser936 plays a role in the temporal aspect of light adaptation .
Phosphoprotein phosphatases are known to be multifunctional. We recently showed that the Drosophila RDGC phosphatase, besides dephosphorylation of rhodopsin, mediates the dephosphorylation of the TRP ion channel at S936 . Light-dependent Ca2+ influx into the photoreceptor cells activates RDGC. Ca2+/calmodulin binds to an IQ motif and Ca2+ possibly binds directly to the EF hands . These motifs are present in all three RDGC protein isoforms. RDGC activity is elevated after activation of the visual signaling cascade via Ca2+ influx through TRP channels resulting in enhanced dephosphorylation of rhodopsin and TRP (at the S936 site) [1,8]. While RDGC-L is apparently dispensable for RH1 dephosphorylation, lack of RDGC-S and -M results in RH1 hyperphosphorylation although not to an extent that is observed in rdgC null mutant flies. It has to be noted, though, that RDGC-S is by far the most abundant RDGC variant resulting in drastically reduced overall RDGC levels in the rdgCΔSM fly. Expression of RDGC-L (in rdgCΔSM flies) or RDGC-S and -M (in rdgCΔL flies) is sufficient to ensure long-term dephosphorylation of pS936-TRP. However, in both cases, the kinetics of TRP dephosphorylation is delayed. The phosphorylated C-termini of both rhodopsin and TRP are located intracellularly, and are thus potentially accessible by both soluble RDGC-S and membrane-bound RDGC-M/L. Thus, the RDGC-S variant, although soluble, has still the potential to enter the cytosolic portion of the rhabdomeric microvilli. We observed accelerated kinetics of pS936-TRP dephosphorylation in the rdgCΔS fly lacking RDGC-S but expressing elevated amounts of RDGC-M. We conclude that, in vivo, RDGC-M has a higher activity compared to RDGC-S. This difference could stem from a generally different catalytic activity of RDGC-M and -S or result from the membrane attachment of RDGC-M that potentially brings this RDGC variant into closer proximity to its membrane-bound substrates. Furthermore, possible differences in catalytic activity could result from the different N-terminal sequences of RDGC isoforms per se. Binding of Ca2+/calmodulin to the IQ motif has been demonstrated to abolish the interaction between the N-terminus and the catalytic domain of RDGC . This interaction between the N-terminus and the catalytic domain has been proposed to control the catalytic activity. Thus, different N-termini might influence the catalytic activity of RDGC. Taken together, our data suggest that usage of alternative N-termini controls the solubility and subcellular targeting of the three Drosophila RDGC phosphatase isoforms and ultimately their in vivo activity.