Photoreceptor IFT complexes containing chaperones, guanylyl cyclase 1 and rhodopsin

Abstract

Intraflagellar transport (IFT) provides a mechanism for the transport of cilium-specific proteins, but the mechanisms for linkage of cargo and IFT proteins have not been identified. Using the sensory outer segments (OS) of photoreceptors, which are derived from sensory cilia, we have identified IFT-cargo complexes containing IFT proteins, kinesin 2 family proteins, two photoreceptor-specific membrane proteins, guanylyl cyclase 1 (GC1, Gucy2e) and rhodopsin (RHO), and the chaperones, mammalian relative of DNAJ, DnajB6 (MRJ), and HSC70 (Hspa8). Analysis of these complexes leads to a model in which MRJ through its binding to IFT88 and GC1 plays a critical role in formation or stabilization of the IFT-cargo complexes. Consistent with the function of MRJ in the activation of HSC70 ATPase activity, Mg-ATP enhances the co-IP of GC1, RHO, and MRJ with IFT proteins. Furthermore, RNAi knockdown of MRJ in IMCD3 cells expressing GC1-green fluorescent protein (GFP) reduces cilium membrane targeting of GC1-GFP without apparent effect on cilium elongation.

Fig. 1. Yeast two-hybrid and GST pull down analysis of MRJ binding to IFT88

A. MRJ protein contains a J domain (aa 1-74) at the N-terminus and a glycine/phenylalanine-rich domain (G/F; aa 75-123) in the middle; the clone recovered in the screen encoded aa 44-242 of bovine MRJ. Co-transformation using this MRJ clone along with the original bait (IFT88(TPR 1-3) resulted in blue colonies comparable to the positive control (T-antigen + p53). Single transformations with IFT88(TPR1-3) or MRJ exhibited no growth and were comparable to the negative control (T-antigen/lamin C). B. IFT88(TPR1-3), IFT88(4-10), or IFT88(TPR1-10) together with MRJ were expressed in yeast cells as fusion proteins with either the GAL4 DNA-binding domain or the GAL4 DNA-activation domain. Interactions between each pair were tested quantitatively by measuring the α-galactosidase activity; single transformations serve as controls. The SV40 large T-antigen/p53 pair is a positive control and SV40 large T-antigen/lamin C pair is a negative control. Results are expressed as means ± the standard error from three separate experiments. C. IFT88(TPR1-10), IFT57, IFT52 and IFT20 together with MRJ were expressed in yeast cells as fusion proteins with either the GAL4 DNA-binding domain or the GAL4 DNA-activation domain. Data are presented as in B. D. Compared to IFT88, MRJ interacts weakly with a kinesin light chain 1 (KLC1) domain containing 5 TPR repeats. The diagram illustrates the domain structure of KLC1 and the region (aa 176-417) used in the assay. Data are presented as in B. E. GST-bMRJ pulls down purified recombinant IFT88. Sepharose beads containing 1 μM GST-bMRJ were used in pull-downs containing 0.25- 2 μM IFT88 (lanes 3-6); controls were no IFT88 (lane 1), 0.2 μM of IFT88 as a loading control (lane 2), and 2 μM IFT88 with GST beads alone (lane 8); lane 7 is blank. IFT88 was detected in the Western blot using goat anti-IFT88C (4). F. GST-bMRJ pull-down of endogenous IFT88, KIF3A, HSC70, and IFT57 (lane 2) from an extract of retinal proteins (RE). Controls include RE alone (lane 1), GST-beads plus RE (lanes 3), GST-bMRJ alone (lane 4), and GST-beads alone (lane 5). The HSC70 antibody (SPA-820) recognizes both HSP70 family members and HSC70.

Fig. 2. Co-immunoprecipitation of MRJ and HSC70 with IFT proteins

A. Co-IP of IFT88 and MRJ with antibodies directed against MRJ (lane 3), HSC70 (lane 8), and 4 different IFT proteins (lanes 4-7 and lane 9); IP antibodies are shown at top and western blot antibodies are shown on left. 150 μg of retinal extract (RE) was loaded as a positive control (lane 10). Controls are protein G beads plus RE but without primary antibody (lane 1) and antibody plus beads without RE (lane 2). MRJ was detected in each IP, but that for IFT57 (lane 7) was weak. B. Co-IP of IFT88, IFT57, KIF3A, and HSC70 with antibodies against IFT proteins (lanes 1-4), HSC70 (SPA-820, lane 5), and MRJ (lane 7); IP antibodies are shown at the top and Western blot antibodies are shown on left. 150 μg of retinal extract (RE) was loaded as a positive control (lane 6). Controls are as in A and are found in lanes 8 and 9.

Fig. 3. GST-bMRJ pulls down GC1 from bovine photoreceptors

A. Silver stained gel from a GST-bMRJ pull-down assay using the NP-40 soluble fraction from isolated photoreceptor rod outer segments attached to their inner segments (ROS); lanes are molecular weight markers (1) and the GST-bMRJ plus ROS extract (2). Controls are GST-beads plus ROS (3) and GST-bMRJ alone (4). The prominent band at ≈115 kDa specifically pulled down by GST-bMRJ was excised from the gel and identified by MALDI-TOF as including retinal guanylyl cyclase 1 (GC1, Gucy2e). B. The identification of GC1 was confirmed by western blotting GST-bMRJ (lane 1) using a rabbit antibody specific for GC1 and a monoclonal antibody to rhodopsin (Rho); controls were GST-bMRJ alone (lane 2), GST plus ROS (lane 3), and a sample of 100 μg of the ROS soluble fraction (lane 4). GC1 was readily identified, but Rho, the most abundant protein in RIS-ROS preparation, was not identified. C. GST pull down assay using recombinant GST-bMRJ and His-tagged GC1 (lane 1, His-GC1); controls are GST beads plus His-GC1 (lane 2), His-GC1 alone to mark the position of GC1 (lane 3), GST-bMRJ alone (lane 4) and GST alone (lane 5). D. Reciprocal GST pull down assay using recombinant GST-GC1 plus His-mMRJ (lane 1); controls are GST plus His-mMRJ (lane 2), His-mMRJ alone (lane 3), GST-GC1 alone (lane 4) and GST alone (lane 5). Western blots in both C and D are with an anti-His antibody. E. Semi-quantitative analysis of GST-MRJ pull-down of GC1_494-844. GST-MRJ and cleaved GC1_494-844 were expressed and purified (see materials and methods), and GC1_494-844 at concentrations from 0.03 to 0.5 μM (lanes 5-9) was incubated with 1.0 μM of GST-bMRJ for 45 min in 1% Nonidet P-40 homogenization buffer. Protein complexes were captured using 60 μl of glutathione sepharose beads (see Methods), and after washing were resuspended in 1× SDS sample buffer and resolved on SDS-PAGE. GC1 was detected by western blotting with anti-mouse GC1 antibody. Controls included beads plus GC1 (lane 1), GST plus GC1 (lane 2), and GST-bMRJ only (lane 3). In addition, 0.06 μM (lane 4) and 0.75 μM GC1 (lane 10) was loaded as a control for estimating GC1 concentration in lanes 5-9. F. Co-IP of GC1 (GC1), IFT88 and HSC70 with antibodies directed against GC1 (lane 4), MRJ (lane 5), and 4 different IFT proteins (lanes 6-9); IP antibodies shown at top and Western blot antibodies shown on left. 150 μg of retinal extract (RE) was loaded as a positive control (lane 3). Controls are protein G beads plus RE but without primary antibody (lane 1) and antibody plus beads without RE (lane 2).

Fig. 4. Co-immunoprecipitation of GC1, rhodopsin, and MRJ with IFT88 is enhanced by addition of Mg/ATP

The IP antibody under all conditions is IFT88C; Western blot antibodies are indicated on the left. IFT88 IPs using equal 150 μg aliquots of retinal extract (RE) were conducted with no addition (lane 3), Mg/ATP (lane 4), Mg/AMP-PNP (lane 5), Mg/ATPgS (lane 6), Mg/ATP plus EDTA (lane 7), ATP without Mg⁺² (lane 8), and Mg/ATP plus NaN₃ (lane 9). 150 μg of retinal extract (RE) was loaded as a positive control (lane 10). Controls are protein A beads plus RE but without primary antibody (Prot-G, lane 2) and antibody plus beads without RE (IgG, lane 1). Bars on right of the RHO blot indicate position of RHO oligomers that typically form in SDS PAGE when RHO is abundant. Note that only monomeric RHO co-IPs with IFT88.

Fig. 5

A. Dose dependent effect of ATP ATP concentration (mM) is indicated at the top and Western blot antibodies are shown on the right. 150 μg of retinal extract (RE) was loaded as a positive control (lane 9). Controls are protein G beads plus RE but without primary antibody (Prot-G, lane 1) and antibody plus beads without RE (IgG, lane 2). Note that two different HSC70 antibodies were used; HSC70(A1) was the SPA-820 antibody (Stressgen Bioreagents) that recognizes both the constitutive heat shock cognate protein 70 (HSC70) as well as heat shock protein 70 (HSP70). In this experiment the membrane was stripped and re-probed HSC70(A2) using ab19136 (Abcam, Cambridge, MA); the latter is a rat monoclonal which is specific for HSC70. Under the loading conditions used, neither GC1 nor RHO were detectible at 0 mM ATP, but increased in the 0.5 to 4 mM concentration range. Similar amounts of HSC70 and IFT88 were present in all IPs. B. Effect of MgCl₂ and NaCl concentration. Effect of 150 and 250 mM NaCl at 1 and 4 mM MgCl₂ (labeled at top). At each NaCl and MgCl₂ concentration 150 μg of retinal extract (RE) was compared to the IP with 0 and 4 mM ATP. 150 μg of retinal extract (RE) was loaded as a positive control (lanes 5, 8, 11, and 14). Controls are protein G beads plus RE but without primary antibody (lane 2) and antibody plus beads without RE (lane 1). Note that the effects of MgCl₂ and NaCl in these experiments was negligible compared to the effect of adding ATP.

Fig. 6. Localization of MRJ and HSC70 in photoreceptors

A-F. Immunocytochemistry images of sections from fresh-frozen bovine retina. A-C. Low power images of the whole retina and D-F are higher magnification images of the photoreceptor layer; scale bars are in C and F, which are merged images. Sections were stained with an antibody against MRJ (A and D, green) and the K26 monoclonal antibody (B and E, red) to indicate the position of connecting cilia (arrow in E) of the photoreceptors. Labels in B and E are: OS, outer segment layer; IS, inner segment layer; ONL, outer nuclear layer; INL, inner nuclear layer; G, ganglion cell layer. G-L. Two examples (G-I and J-L) of isolated mouse photoreceptors stained with an antibody to MRJ (green, G and J) and acetylated alpha-tubulin (red, H and K); G-I correspond to one cell and J-L correspond to another cell. The magnification bars in G-I correspond to 7.8 μm, in J-L to 9.8 μm. I and L are the merged images for each cell. In J-L the arrow indicates the position of the ciliary rootlet within the inner segment and the arrowhead indicates the position of the distal axoneme within the OS. Note that MRJ is present in the OS but is not specifically localized on the axoneme. M-Q. A whole mounted mouse OS with attached inner segment (arrow) stained for HSC70 (green, O) and acetylated alpha-tubulin (red, N). M. Bright field image (phase-like). N. Same as M stained for acetylated a-tubulin (red). O. Same image stained for HSC70 (green). P. Merge image of N and O. Q. Merged image of M, N and O. Magnification bars corespnd to 7.3 μm. Arrow points to the inner segment and the arrowhead points to the axoneme within the OS. The asterisk indicates the position of the basal body/connecting cilium. Note that HSC70 is abundant in the inner segment and basal body region and is diffusely distributed in the OS.

Fig. 7. Localization of endogenous MRJ in kidney epithelial cells

A. Cultured LLC-PK1 (pig kidney tubular epithelial) cells at day 4 after plating were stained with an antibodies against MRJ, IFT88, or IFT57 (red, Ab at top) and counterstained with an to antibody against acetylated α-tubulin (green, Tub at top), and Hoechst (blue, Hoechst at top); merged images are on right. Two horizontal panels for MRJ and one each for IFT88 and IFT57 are provided. Note that MRJ, IFT88, and IFT57 all exhibit a punctate pattern of staining along cilia, but cell body staining is much greater for MRJ. The calibration bar corresponds to 10 μm̃. B. Production of stably transfected inner medullary collecting duct (IMCD3) kidney epithelial cells expressing His-mMRJ (His-tagged full-length mouse MRJ). Lysates from un-transfected cells (lane 1), transiently transfected cells (lane 2), non-stably-transfected clone #19 (lane 3), and stably-transfected clone #20 (lane 4). An anti-His antibody used for the Western blot recognized a band of ∼32 kDa in samples from transient transfected cells and stably-transfected clone #20. C. His-mMRJ is found in cilia of stably-transfected IMCD3 cells. Stably-transfected IMCD3 clone #20 (lane 4 in B) was stained with the anti-his antibody (green) and counterstained with Hoechst (blue) to show the positions of nuclei. The anti-His antibody detected His-MRJ in both cilia and in the cell body. The calibration bar corresponds to 10 μm.

Fig. 8. GC1-GFP in cilia of stably transfected IMCD3 cells

A. Western blot with an anti-GC1 antibody showing a low level of GC1-GFP in a stable line and a much higher level in a transiently transfected culture; GC1 is not present in un-transfected cells. GC1-GFP is at a higher molecular weight because of the in frame fusion of GFP. Lanes are equally loaded at 150 μg protein per lane. B. IC showing that antibodies to GC1 (green) co-localize with acetylated-α tubulin (red) in apical sensory cilia. Merged images are on the right. Bars equal 8 μm. C. GC1 and acetylated-α tubulin co-localization similar to B except at a higher magnification of an especially long sensory cilium. Bars equal 4 μm. Single label IC with an anti-GFP antibody and counterstaining with Hoechst (not shown) indicates that most cells make cilia containing GC1-GFP.

Fig. 9. MRJ Knockdown reduces GC1-GFP in sensory cilia

A. Western blot of lysates from three IMCD3 cultures all expressing GC1-GFP; shRNA control (Con Vec), shRNA knockdown vector (MRJ Vec) on right compared to control (Con Vec, left) or No Vec (middle). Blotting antibodies are on left. B. IMCD3 cells were stably co-transfected with the GC1-GFP plasmid and the MRJ shRNA negative control plasmid. GC1-GFP (green) co-localizes with cilia labeled with anti-acetylated-α tubulin (red); merged images are on right. Bars= 8 μm. C. IMCD3 cells as in B except that they were co-transfected with the expermental MRJ shRNA (M1); this was the same transfection illustrated on right in A. Real-time RT-PCR of MRJ mRNA showed that MRJ mRNA was reduced 70% in the experimental group compared to the control (not shown). In the knockdown group very little GC1-GFP is detected in cilia; it is expressed as seen in the positive diffuse staining in B (left and right) and in the western blot in A. Bars= 8 μm. D. Higher power images showing GC1-GFP (green) and acetytylated-α tubulin (red) in controls (Con on left) and shRNA treated cultures (right). Bars = 4 μm.

Fig. 10. Conceptual models for the role of MRJ/HSC70 in binding of GC1 to the IFT particle

The two models are functionally equivalent, but differ in that HSC70 is constitutively present on the IFT particle and also interacts with GC1 in A. In B MRJ recruits HSP70 to the IFT particle. Although we have not studied direct HSC70 or GC1 binding, the fact that HSC70 co-IPs with IFT proteins under all conditions used in this study favors model A. Although both GC1 and RHO are present in IFT cargo complexes, the models include GC1 only because we have demonstrated direct binding between the GC1 and MRJ. The fact that RHO was not present in GST-bMRJ assays that readily identified GC1 suggests that the mode of interaction of RHO with the IFT complex is independent of MRJ.