Heteromeric p97/p97R155C complexes induce dominant negative changes in wild-type and autophagy 9-deficient Dictyostelium strains

Abstract

Heterozygous mutations in the human VCP (p97) gene cause autosomal-dominant IBMPFD (inclusion body myopathy with early onset Paget's disease of bone and frontotemporal dementia), ALS14 (amyotrophic lateral sclerosis with or without frontotemporal dementia) and HSP (hereditary spastic paraplegia). Most prevalent is the R155C point mutation. We studied the function of p97 in the social amoeba Dictyostelium discoideum and have generated strains that ectopically express wild-type (p97) or mutant p97 (p97(R155C)) fused to RFP in AX2 wild-type and autophagy 9 knock-out (ATG9(KO)) cells. Native gel electrophoresis showed that both p97 and p97(R155C) assemble into hexamers. Co-immunoprecipitation studies revealed that endogenous p97 and p97(R155C)-RFP form heteromers. The mutant strains displayed changes in cell growth, phototaxis, development, proteasomal activity, ubiquitinylated proteins, and ATG8(LC3) indicating mis-regulation of multiple essential cellular processes. Additionally, immunofluorescence analysis revealed an increase of protein aggregates in ATG9(KO)/p97(R155C)-RFP and ATG9(KO) cells. They were positive for ubiquitin in both strains, however, solely immunoreactive for p97 in the ATG9(KO) mutant. A major finding is that the expression of p97(R155C)-RFP in the ATG9(KO) strain partially or fully rescued the pleiotropic phenotype. We also observed dose-dependent effects of p97 on several cellular processes. Based on findings in the single versus the double mutants we propose a novel mode of p97 interaction with the core autophagy protein ATG9 which is based on mutual inhibition.

Figure 1. Domain structure of p97 and immuno-verification of mutant Dictyostelium strains.

(A) Domain structure of p97. The 793 amino acid protein is composed of two N-terminal CDC48-like domains, followed by two AAA ATPase domains, D1 and D2, which are separated by the L2 linker region and a C-terminal region of approximately 160 amino acids. (B) The R155C mutation which causes IBMPFD in affected individuals is situated in the second CDC48-like domain. The region surrounding the arginine 155 is highly conserved from yeast to man. (C) Ectopic expression of p97-RFP and p97^R155C-RFP in AX2 wild-type cells and in the ATG9^KO mutant. Top: Verification of expression of endogenous and RFP-fused p97 using pAb p97_8_6841. Middle: Verification of expression of RFP-fused p97 using a polyclonal RFP antibody. Bottom: loading control, actin.

Figure 2. Purified recombinant p97 and p97^R155C form hexamers.

Purified recombinant p97 and p97^R155C were subjected to either blue-native (left image) or denaturing (right image) gel electrophoresis followed by Coomassie brilliant blue staining. Under native conditions, wild-type and mutant p97 migrate at a position corresponding to approximately 600 kDa.

Figure 3. p97 and p97^R155C form heteromers in vivo.

(A) Coomassie stained SDS-PAGE gel of a p97 co-immunoprecipitation experiment. Input: soluble cell lysate of either AX2/p97^R155C-RFP (left column) or AX2 (right column) cells; Control (ctrl): beads incubated with bovine serum albumin instead of the polyclonal RFP antibody; CoIP: co-precipitated proteins by the RFP antibody. (B) Immunoblot verification of the presence of p97 and p97^R155C-RFP in the immunoprecipitate and input. (C) Immunoblot verification of the presence of p97^R155C-RFP.

Figure 4. Cell growth in shaking culture and on Klebsiella aerogenes are altered in mutant strains expressing p97-RFP or p97^R155C-RFP.

(A) Strains expressing p97-RFP or p97^R155C-RFP display specific growth defects in shaking culture. Please note the logarithmic scale of the y-axis. (B) Growth on K. aerogenes lawns. Mutation specific and dose dependent effects are seen in both wild-type and ATG9^KO strains. Growth of AX2 on day 5 was set to 100%.

Figure 5. Subcellular localization of p97 and its co-localization with ubiquitin in wild-type and mutant strains.

(A) Visualization of subcellular localization of p97 in AX2 wild-type cells with polyclonal antibodies p97_8_6841 directed against amino acids 23–73 and p97_9_6574 directed against amino acids 254–310. (B) Subcellular localization of p97 (left panel) using the p97_8_6841 polyclonal antibody and ubiquitin (middle panel) using the P4D1 monoclonal antibody (NEB, Germany) in AX2 wild-type cells and mutant strains. Merged images and DAPI staining to visualize nuclei (right panel). Upper row: AX2 wild-type cells; middle row: ATG9^KO mutant; bottom row: ATG9^KO/p97^R155C-RFP double mutant. Please note that ubiquitin positive protein aggregates frequently co-localize with p97 in the ATG9^KO mutant (arrows) but not in the ATG9^KO/p97^R155C-RFP double mutant (double arrowheads). The ATG9^KO mutant also contains protein aggregates that are either positive for p97 (arrowhead) or ubiquitin (double arrowhead). Cells were fixed with cold methanol and stained with the indicated antibodies. Scale bars are 10 µm and 2 µm in inset.

Figure 6. Ubiquitinylation, proteasomal activity and levels of SU5 and ATG8.

Strains expressing p97-RFP or p97^R155C-RFP display specific changes in ubiquitinylation (A), proteasomal activity (B), and levels of SU5 and ATG8 (C). Values of AX2 have been set to 1; normalization was to actin (see Fig. S1). For detection of ubiquitin, SU5, and ATG8 in Western blots the monoclonal antibodies P4D1 (NEB, Germany) and proteasomal subunit 5 (SU5) as well as the ATG8_6080 polyclonal antibody were used, respectively. The proteasomal activity assay was performed by adding the Ultra-Glo™ Luciferase and the signal peptide specific for chymotrypsin-like activity coupled to luciferin (Promega, Germany) to cell lysates. Proteasomal activity was normalized to proteasome content, and the chymotrypsin-like activity of AX2 was set to 1.

Figure 7. Expression of p97^R155C-RFP impairs phototaxis in AX2 cells and rescues the phototaxis defect of ATG9^KO cells.

The ability of AX2 wild-type and of mutant slugs to migrate towards a light source (wavy line) was tested. While AX2 slugs (top left image) and AX2 slugs expressing p97-RFP (bottom left image) nicely migrated towards the light source, phototactic ability was severely impaired in the AX2/p97^R115C-RFP strain (middle left image). In the ATG9^KO strain phototaxis is completely lost (top right image). While expression of p97^R155C-RFP in the ATG9^KO strain partially rescued the phototactic ability (middle right image), no rescue of phototactic ability was observed upon expression of p97-RFP (bottom right image). The phototaxis assay was performed as described .

Figure 8. Expression of p97^R155C-RFP rescues fruiting body formation in ATG9^KO cells.

Neither expression of p97-RFP nor p97^R115C-RFP changed fruiting body formation in AX2 wild-type cells (left column). In the ATG9^KO strain fruiting body formation is completely lost (top right image). While expression of p97^R155C-RFP in the ATG9^KO strain fully rescued the fruiting body formation (middle right image), no obvious rescue of fruiting body formation was observed upon expression of p97-RFP (bottom right image). The assay was performed as described .

Figure 9. Interaction and mutual inhibition of p97 and ATG9.

(A) The analysis of phototactic ability illustrates mutual inhibition of p97 and ATG9 in conjunction with inhibition of this process in AX2 cells. (B) Expression of p97^R155C in AX2 wild-type cells frees ATG9 to exert its inhibitory activity. (C) Expression of wild-type p97 in wild-type AX2 cells results in slightly reduced phototaxis which can be explained by a dose dependent effect. (D) In the absence of ATG9, its inhibitory activity versus p97 is lost leaving p97 free to strongly inhibit phototaxis. (E) Partial rescue of phototaxis in the ATG9^KO/p97^R155C-RFP double mutant. (F) Since phototaxis is already lost in ATG9^KO cells, the expression of p97-RFP cannot lead to a further impairment. Line width correlates with the strength of the inhibitory effect. Dotted line, residual activity. Please refer to the Discussion section for further details on this model of interaction and mutual inhibition.