XB-ART-56349J Cell Biol January 1, 2019; 218 (11): 3753-3772.
Show Gene links Show Anatomy links
An intrinsic compartmentalization code for peripheral membrane proteins in photoreceptor neurons.
In neurons, peripheral membrane proteins are enriched in subcellular compartments, where they play key roles, including transducing and transmitting information. However, little is known about the mechanisms underlying their compartmentalization. To explore the roles of hydrophobic and electrostatic interactions, we engineered probes consisting of lipidation motifs attached to fluorescent proteins by variously charged linkers and expressed them in Xenopus rod photoreceptors. Quantitative live cell imaging showed dramatic differences in distributions and dynamics of the probes, including presynapse and ciliary OS enrichment, depending on lipid moiety and protein surface charge. Opposing extant models of ciliary enrichment, most probes were weakly membrane bound and diffused through the connecting cilium without lipid binding chaperone protein interactions. A diffusion-binding-transport model showed that ciliary enrichment of a rhodopsin kinase probe occurs via recycling as it perpetually leaks out of the ciliary OS. The model accounts for weak membrane binding of peripheral membrane proteins and a leaky connecting cilium diffusion barrier.
PubMed ID: 31594805
PMC ID: PMC6829649
Article link: J Cell Biol
Species referenced: Xenopus laevis
Genes referenced: akap13 cps1 dbt dcc grk1 mtor rho
Article Images: [+] show captions
|Figure 1. Expression and quantification of fluorescent probes in rods. (A) Amino acid sequences of PMP probes. (B) Transmission image of retinal slice preparation. (C) Confocal image in XY and XZ of a live rod expressing EGFP. Colors are a heat map: red is highest fluorescence, blue is lowest. Axial intensity values (lower panel) were taken along black lines. (B and C) Scale bars, 10 µm. (D) Structure, name, binding energy, and estimated membrane affinity of the lipid groups. The myristoylation motif results in heterogeneous acylation in retina, with four acyl varieties. In mammalians, the proportions are 55% C14:2, 20% C12:0, 16% C14:1, and 8% C14:0, as determined from mass spec of Tα (Kokame et al., 1992; Neubert et al., 1992; Johnson et al., 1994; Neubert and Hurley, 1998; Lobanova et al., 2007). Throughout, Myr refers to the myristoylation motif and acyl refers to lipids. Two Kds are given for prenyls, with that for carboxymethylated appearing in parentheses. Binding energies and membrane affinity estimates were based on in vitro partitioning experiments (Peitzsch and McLaughlin, 1993; Silvius and l’Heureux, 1994).|
|Figure 2. Lipidation and surface charge, alone, drive compartmentalization of PMPs. (A) Schematic of amphibian rod. S, Synaptic spherule; Ax, axon; N, nucleus; M, myoid; E, ellipsoid; ND, nascent discs; CC, connecting cilium; DM, mature disc membranes; and PM, plasma membrane. Cyan: Apical membrane. NDs are open to the extracellular milieu and contiguous with the CC and PM. DMs are enclosed within and separate from the OS PM and each other. (B) Confocal images of rod cells expressing PMP probes. Scale bars, 10 µm. (C) Box-whisker plots of average fluorescence in synapse divided by average fluorescence of OS. In all box-whisker plots, circles represent individual cells, red line is median, box limits are 25th and 75th percentiles, whiskers are interquartile range, excluding outliers (+), and green asterisks are means. Presence of differences among all probes was tested by one-way ANOVA, α = 0.05. Black asterisks indicate significant differences from NL0, determined by Tukey-corrected post hoc two-tailed t test. n for each construct: GG+8, 14; Far+8, 13; Far+4, 21; NL0, 29; Myr+8, 27; Far-8, 16; ctTγ, 17; GG-8, 22; Far0, 42; GG0, 20; Myr-8, 15; ntTα, 18; Myr0, 28; and GRK1ct18, 30. ns, not significant.|
|Figure 3. Distribution of prenylated probes dependds on linker charge. (A–C) Upper panels, representative confocal images. Scale bars, 10 µm. Lower panels, average compartment fluorescence normalized to total cell fluorescence. (D) OSEI, defined as the ratio of average OS fluorescence to average myoid fluorescence, FOS/FM. (E) SynEI, defined as ratio of average synaptic fluorescence to average myoid fluorescence, Fsyn/FM. (D and E) Lower panels are significance tables. *, P = 0.05; NS, not significant. Box fills as in Fig. 2 C key. Box-whisker plots as described in Fig. 2. n for each construct: NL+8, 14; NL0, 13; NL-8, 17; GG+8, 14; GG0, 20; GG-8, 22; Far+8, 13; Far+4, 21; Far0, 42; and Far-8, 16.|
|Figure 4. Positive prenylated probes are enriched in the cell body PM, perinuclear membranes, and the OS–IS junction. (A) Representative confocal images. Scale bars, 10 µm. Arrowheads: OS–IS junction. (B) 3D renderings of the Far+8 rod shown in A, at different angles (see angle indicators). Green: 50th percentile, red: 90th percentile intensities. White arrowheads: CPs (Videos 1, 2, and 3). Scale bar, 5 µm. (C) Confocal images of broken OS–IS showing fluorescence-containing CPs separated from OS. Scale bars, 5 µm. (D) OS–IS junction enrichment width estimate. Upper panel: Confocal image of Far+8 rod in A with OS–IS junction expanded. Scale bar, 5 µm. Middle panel: Fluorescence intensities as a function of axial distance; zero is the OS–IS junction. Arrowheads show half maximum fluorescence. Bottom panel, Junction widths for Far+8 and GG+8 cells were not different as determined by two-tailed T test assuming equal variance. Box-whisker plot as described in Fig. 2. n for each construct: Far+8, 9; and GG+8, 12. (E) Fluorescence as a function of distance within the OS. Left panels: Representative profiles from positive prenylated probes. Right panels: OS distributions of other PMPs. (F) FRAP of Far+8 in the synapse. Red box: Spherule region of interest (ROI); yellow box: OS–IS junction ROI. Scale bar, 5 µm. (G) Time course of synapse fluorescence recovery and OS–IS junction fluorescence loss.|
|Figure 5. Prenylated PMP transport to OS does not require association with LBC protein, PrBPδ. (A–C) GFP trap of prenylated probes fail to pull down PrBPδ. Western blots probed with anti PrBPδ antibody. (D and E) Upper panels: Representative confocal images of GRK1 and Tγ probes. Scale bars, 10 µm. Lower panels: Normalized compartment distributions. (F) OSEIs showing higher OS enrichment of GRK1. Significance table: P ≤ 0.05; NS, not significant. (G and H) GFP trap of GRK1 and Tγ probes show that GRK1ct18 associates with PrBPδ while Tγct15 does not. Box-whisker plots as described in Fig. 2. n for each construct: GRK1ct18, 30; Tγct19, 17; Far0, 42; and NL0, 29. (A–C, G, and H) MW, molecular weight markers; I, input; FT, flow through; and E, eluate.|
|Figure 6. Myristoylated probes are enriched in the OS regardless of linker charge. (A) Upper panels: Representative confocal images of rods expressing Myr probes. Lower panels: Compartment distribution profiles. (B) Confocal image and compartment distribution profile of Tαnt16. (A and B) Scale bars, 10 µm. (C) OSEIs of myristoylated probes. Box-whisker plots as described in Fig. 2. Significance table shows that all probes had significantly higher OSEIs than NL0. (D) SynEIs show that Myr+8 is the only presynapse enriched probe. Significance tables: P ≤ 0.05; NS, not significant. n for each construct: Myr+8, 27; Myr0, 28; Myr-8, 15; and Tαnt16, 18.|
|Figure 7. OS mobility shows low affinity binding of prenylated probes to disc membranes and that binding affinity of myristoylated probes scales with charge. (A) Representative time course images of FRAPa or FRAPb experiments. Pre-blast images: Distributions before two-photon photoactivation or photobleaching. Scale bars, 10 µm. (B) Time courses of fluorescence relaxation at the center of the photoconversion sites fitted with a 3D cylindrical diffusion model, magenta lines. Black lines: Data-model difference. (C) Root-mean-square (RMS) errors plotted against D. red circle indicates D with lowest error, and value shown above plots. (D) Bar chart of average DOSs. Error bars are SEM. ANOVA followed by two-tailed, homoscedastic t test with Tukey correction showed that DOSs for all lipidated probes were significantly different from the DOS for PAGFP. DOSs among 0Far and GRK1 were not different. Significance tables: P ≤ 0.05; NS, not significant. n for each construct: 0Far, 6; GRK1, 5; PAGFP, 6; Myr-8, 8; Myr0, 3; and Myr+8, 3.|
|Figure 8. Lipidated probes move through the connecting cilium via impeded diffusion. (A and C) Time course images of PAGFP or Far0-PAGFP flux into the IS. Vertical lines indicate the OS–IS junction. At time zero, PAGFP was activated by two photon scanning over 70–80% of the OS. ROIs were drawn to find cOS (ROI1), cIS (ROI2); and MIS (ROI3; Eq. 1). Scale bars, 10 µm. (B and D) Flux normalized to OS–IS concentration gradient (grad [OS-IS]). Line is linear regression. DCC was estimated according to Eq. 1. (E) Average DCCs; error bars: SEM. Asterisk indicates DCC of PAGFP was significantly higher than those of the lipidated probes at P = 0.05, as determined by two-tailed t test.|
|Figure 9. Contributions of local binding, active transport, and cilium diffusion impediment to OS and synaptic enrichment of PMPs, computed by the DBT model. (A) Schematic of model rod with compartments labeled as in Fig. 2 A. Representative electron micrographs show the geometries of the compartments. Note the many synaptic vesicles filling the spherule (a) and the density and order of OS disc membranes (c). Parameters for each compartment used in all computations are listed: L, length; Ac, area of cross-section; and DGFP, diffusion coefficient of unmodified EGFP. EM images reproduced with permission from a (Schacher et al., 1976), b (Peters et al., 1983), and c (Townes-Anderson et al., 1985). a and b, Scale bars, 1 µm. c, Magnification = 45,000. dia, diameter. (B–E) DBT predictions of EGFP-GRK1ct18 distribution show that active transport and the connecting cilium diffusion impediment, together, are the major contributors to OS enrichment; local OS binding played a minor role. (F) Predictions of OSEI, F(OS)/F(IS) from model traces, and effective DOS, found by fitting model FRAPs as described in Fig 7, plotted versus equivalent binding power. Arrowhead on the DOS line shows that at EBP = 1.5, DOS was approximately twofold lower than that for no binding, as found experimentally for EGFP-GRK1ct18 and EGFP-Far0 versus EGFP (Fig. 7). However, the OSEI was <1. The OSEI line shows that tighter OS binding can lead to fivefold (arrowhead) or better OS enrichment, but at the cost of mobility. BP, binding power. (G) Transport velocity versus predicted OSEI, given OS EBP = 1.5. Arrowhead indicates the velocity that produces the observed approximately fivefold OS enrichment, as found experimentally for EGFP-GRK1ct18. (H–J) Higher synapse affinity resulted in significant enrichment (H) and FRAP recovery t1/2 ∼2 min (I and J), similar to experimental results for Far0 (Fig. 4 G), suggesting the distribution of Far0 within the IS is mediated by equilibrium binding alone.|
|Figure 10. The photoreceptor PMP compartmentalization code. In the absence of interactions with chaperones, PMPs tend to distribute according to a few simple rules. Positive charge near any lipid moiety leads to synaptic enrichment. Myristoylated proteins with any charge and neutral prenylated proteins slightly favor the ciliary OS. All others are distributed like unmodified EGFP (A). To achieve “super” enrichment of the OS or synapse compartments, active mechanisms to enrich or deplete the ciliary OS must come into play (B).|
References [+] :
Baehr, Membrane protein transport in photoreceptors: the function of PDEδ: the Proctor lecture. 2015, Pubmed