PARIS (ZNF746) repression of PGC-1α contributes to neurodegeneration in Parkinson's disease

Abstract

A hallmark of Parkinson's disease (PD) is the preferential loss of substantia nigra dopamine neurons. Here, we identify a new parkin interacting substrate, PARIS (ZNF746), whose levels are regulated by the ubiquitin proteasome system via binding to and ubiquitination by the E3 ubiquitin ligase, parkin. PARIS is a KRAB and zinc finger protein that accumulates in models of parkin inactivation and in human PD brain. PARIS represses the expression of the transcriptional coactivator, PGC-1α and the PGC-1α target gene, NRF-1 by binding to insulin response sequences in the PGC-1α promoter. Conditional knockout of parkin in adult animals leads to progressive loss of dopamine (DA) neurons in a PARIS-dependent manner. Moreover, overexpression of PARIS leads to the selective loss of DA neurons in the substantia nigra, and this is reversed by either parkin or PGC-1α coexpression. The identification of PARIS provides a molecular mechanism for neurodegeneration due to parkin inactivation.

Figure 1. Identification of a novel parkin interacting substrate, PARIS

(A) Schematic representation of PARIS. The conserved Kruppel associated Box (KRAB) and Zinc Finger motifs and their location are indicated. (B) Regional analysis and levels of PARIS protein expression via immunoblot in various brain regions. Data = mean ± S.E.M., n = 3. (C) Immunohistochemical distribution of PARIS in a sagittal and coronal sections of six-week-old male C57BL mouse brain. Right upper panel shows an antigen preabsorption control. Ctx, cerebral cortex; Hip, hippocampus; CPu, caudate putamen; SNpr, SN pars reticulata; SNpc, SN pars compacta; Th, thalamus; Cb, cerebellum; (Cb); PC, Purkinje cells; ML, molecular cell layer; GL, granule cell layer; DG, dentate gyrus. Dashed (−) line outlines SN. High power view of SNpc (rectangles in third row) is shown in the third row middle and right panels and lower middle and right panel. Scale bars, 200 µm unless indicated, n = 3. (D) Confocal microscopy demonstrates that endogenous PARIS and parkin are co-localized mainly in the cytoplasm of rat cortical neurons. Top Panel, Parkin-red; PARIS-green. Inset – high power view of an individual neuron. Bottom panel, Parkin-green; PARIS-red, Nucleus-DAPI-blue, merge-yellow, n = 4. See also Figure S1.

Figure 2. Parkin interacts with PARIS

(A) Immunoprecipitated (IP) FLAG-PARIS interacts with MYC-Parkin, but not MYC-XIAP or the PARIS homologue, V5-ZNF398 (lane 6), n = 3. (B) GST-pull down assay between parkin and PARIS indicates a robust interaction between parkin and PARIS, n = 4. (C) Immunoprecipitated FLAG-PARIS interacts with WT Parkin and Parkin mutants (C431F, G430D, R275W, Q311X) in SH-SY5Y cells. Lower panel, relative binding, data = mean ± S.E.M., n = 3, *p < 0.05, Student’s t-test. (D) Immunoblot (IB) shows that Parkin and PARIS co-immunoprecipitate in human striatum, n = 3. (E) Parkin and PARIS co-immunoprecipitate from WT mouse ventral midbrain, but not parkin KO ventral midbrain, n = 3. See also Figure S2.

Figure 3. Parkin ubiquitinates and regulates the ubiquitin proteasomal degradation of PARIS

(A) WT MYC-Parkin ubiquitinates FLAG-PARIS (lane 3). Parkin mutants (C431F, G430D, Q311X) are unable to efficiently ubiquitinate FLAG-PARIS (lanes 4–7). Ubiquitination (Ub(n)) is indicated on right with brackets, n = 3. (B) Endogenous ubiquitination of PARIS (lane 2) is enhanced with exogenous WT Parkin (lane 3) and it is eliminated with shRNA-Parkin (lane 4). In the presence of shRNA-Parkin, robust ubiquitination of PARIS is observed via co-expression of shRNA-resistant WT parkin (WT^R) but not shRNA-resistant Q311X mutant parkin (Q311X^R), n = 3. (C) 10 µM MG-132 increases PARIS steady state levels compared to DMSO control. Bottom panel, relative PARIS levels normalized to β-actin, n = 3. (D) Increasing ratio (1:1 to 4:1) of MYC-Parkin results in decreased steady-state levels of FLAG-PARIS (lanes 1–4). Bottom panel, relative PARIS and parkin levels normalized to β-actin, n = 3; regression analysis, R²=0.9985, p<0.05). (E) WT Parkin decreases the steady-state levels of PARIS compared to mutant Q311X parkin or GFP transfected control cells in cyclohexamide (CHX)-chase experiments in SH-SY5Y cells transiently expressing FLAG-PARIS. Bottom panel, relative PARIS levels normalized to β-actin, n = 3. (F) MYC-parkin leads to degradation of FLAG-PARIS at a 4 to 1 ratio, respectively. 10 µM MG-132 prevents the degradation of FLAG-PARIS and MYC-Q311X parkin has no effect, n = 3) (G) PARIS accumulates after shRNA-Parkin and co-expression of shRNA resistant parkin (MYC-Parkin^R) leads to robust degradation of PARIS, n = 3. Data = mean ± S.E.M., *p < 0.05, **p < 0.01 and ***p < 0.001, ANOVA with Student-Newman-Keuls post-hoc analysis (E, F, G) or Student’s t-test (C). See also Figure S3.

Figure 4. PARIS accumulates in AR-PD, sporadic PD and in animal models of parkin inactivation

(A) Immunoblot analysis of PARIS and β-actin in cingulate cortex from age-matched controls and AR-PD patient brains with parkin mutations (B) Quantitation of the immunoblots in panel A normalized to β-actin, n = 4. (C) PARIS levels in cerebellum (CBM), frontal cortex (FC), striatum (STR) and SN of sporadic PD patient brains compared to age-matched controls. (D) Relative PARIS levels normalized to β-actin in panel B, Controls n = 4; PD n = 5. (E) Immunoblot analysis of PARIS in cortex (CTX), STR and ventral midbrain (VM) from WT and parkin exon 7 KO 18–24 month old mice. (F) Relative protein levels of PARIS normalized to β-actin for panel E, WT n = 9; parkin KO n = 10. (G) PARIS mRNA levels in indicated brain regions from WT and parkin exon 7 KO 18–24 month old mice. (H) Top panel, experimental illustration of stereotaxic intranigral virus injection. Bottom panels, immunofluorescent images of TH (red), GFP (green) and merged (yellow) in exon 7 floxed parkin mice (parkin^Flx/Flx) after stereotactic delivery of Lenti-GFP or Lenti-GFPCre into the SNpc. 84.9±1.9% and 78.1±2.6% of TH neurons express GFP and GFPCre, respectively, n = 3 per group. Enlarged images in the right bottom panels were taken from the white rectangle region from the merged images of Lenti-GFPCre and Lenti-GFP, bar = 100 µm. (I) Immunoblot analysis of parkin, PARIS, actin and GFP 4 weeks after intranigral Lenti-GFPCre or Lenti-GFP injection into parkin^Flx/Flx mice. (J) Relative protein levels of PARIS normalized to β-actin for panel I. Data = mean ± S.E.M., *p < 0.05, **p < 0.01 and ***p < 0.001, unpaired two-tailed Student’s t-test (B, J) and ANOVA test with Student-Newman-Keuls post-hoc analysis (D, F, G). See also Table S1 and S2.

Figure 5. PARIS acts as transcriptional repressor of PGC-1 α

(A) Identification and MACAW alignment of the PARIS DNA-binding sequence as determined by CASTing. Darker colors represent a greater degree of overlap of the segment pairs (bottom right - % overlap). *Duplicate sequence tags. (B) Relative luciferase activity of the 1-kilobase human PGC-1α (−992 to +90) compared to Renilla luciferase ± PARIS or ± parkin or ± familial mutant Q311X parkin, n = 3. Location of IRS, CRE motifs and oligos for human ChIP (arrows) in the PGC-1α promoter construct (top of panel). Immunoblot analysis confirms the expression of FLAG-PARIS, MYC-Parkin and MYC-Q311X parkin (right panel). (C) EMSA of GST-PARIS of ³²P-labeled WT (WT-³²P) IRS oligonucleotides (IRS1-WT, IRS2-WT, IRS3-WT) of the human PGC-1α promoter and ³²P-labeled mutant (T→) (MT-³²P) IRS oligonucleotides (IRS1-MT, IRS2-MT, IRS3-MT). Unlabeled WT (WT cold) IRS oligonucleotides disrupt the GST-PARIS-DNA protein complexes with the WT-³²P IRS oligonucleotides, n = 3. Unlabeled mutant probes (MT cold) has no effect on mutant (MT-³²P). Arrow indicates specific PARIS-shifted probe; NS, nonspecific; FS, fragmented PARIS-shifted probe; FP, free probe. (D) PARIS occupies the endogenous PGC-1 α promoter as determined by ChIP assay with anti-PARIS polyclonal antibodies in SH-SY5Y cells, n = 3. (E) PARIS occupies the endogenous mouse PGC-1 α promoter as determined by ChIP in mouse whole brain, n = 3. Mouse specific IRS primers are indicated in Figure S4B. (F) ChIP assay of endogenous PARIS binding to the IRS region of the human PGC-1a promoter in human PD and aged-matched control (CTL) striatum, control n = 3; PD n = 4. (G) Quantitation of ChIP in panel F. Human specific IRS primers for D and F are indicated in panel B. (H) Real-time qRT-PCR of PGC-1α, GFP and β-actin following transient transfection of GFP, GFP-PARIS or GFP-C571A PARIS mutant, n = 4. (I) Immunoblot analysis of PGC-1α, GFP and β-actin following transient transfection of GFP, GFP-PARIS or GFP-C571A PARIS mutant, n = 4. (J) Quantitation of the immunoblots in panel I normalized to β-actin, n = 4. (K) Immunoblot analysis of parkin, PARIS, PGC-1α and β-actin in double knockdown experiments via lentiviral transduction of shRNA-parkin and/or shRNA-PARIS in SH-SY5Y cells, n = 3. (L) Quantitation of the immunoblots in panel K normalized to β-actin, n = 3, sh = shRNA. (M) Relative mRNA levels of PGC-1α normalized to GAPDH, n = 3. Quantitative data = mean ± S.E.M., *p < 0.05, **p < 0.01, ***p < 0.001, unpaired two-tailed Student’s t-test (G), ANOVA with Student-Newman-Keuls post-hoc analysis (B, H, J, L, M). See also Figure S4, Table S2, S3, S5.

Figure 6. Identification of PGC-1α and NRF-1 as pathological in vivo targets of accumulated PARIS in PD brain and conditional parkin KO mice

(A) Real-time qRT-PCR of IRS (PEPCK-like motif)-containing genes and PGC-1α dependent genes in PD SN compared to age-matched CTL-SN normalized to GAPDH, n = 3–4 per group. (B) Immunoblots of PARIS, PGC-1α, parkin and NRF-1 in soluble and insoluble fractions of PD SN compared to age-matched CTL-SN. (C) Quantitation of the immunoblots in panel B normalized to β-actin, PD, n = 4; Control, n = 3. (D) Top panel, experimental illustration of stereotaxic intranigral virus injection. Below are immunoblots of parkin, PARIS, PGC-1α, NRF-1, β-actin and GFP, 4 weeks after stereotactic delivery of Lenti-GFP, Lenti-GFPCre, Lenti-GFPCre + shRNA-dsRed, or Lenti-GFPCre + shRNA-PARIS into exon 7 floxed parkin mice (parkin^Flx/Flx), n = 3 per group. *nonspecific band. (E) Quantitation of the immunoblots in panel D normalized to β-actin, n = 3 per group. (F) The alteration of PGC-1α and NRF1 shown in panel D and E results from transcriptional changes as determined by real-time qRT-PCR, n = 3 per group. (G) Top panel, experimental illustration of stereotaxic intranigral virus injection. Below is TH immunostaining of representative mice midbrain sections from SN of parkin^Flx/Flx mice injected with Lenti-GFP, and Lenti-GFPCre ± Lenti-shRNA-PARIS 10 months post-injection of virus. (H) Stereological assessment of TH and Nissl positive neurons in the SN of parkin^Flx/Flx mice injected with Lenti-GFP, and Lenti-GFPCre ± Lenti-shRNA-PARIS 3 (n = 3 per group) and 10 months (n = 7 per group) after injection of virus. Data = mean ± S.E.M., *p < 0.05, **p < 0.01 and ***p < 0.001, unpaired two-tailed Student’s t-test (panel A and C) and ANOVA with Student-Newman-Keuls post-hoc analysis (panel E, F and H). See also Figure S5, Table S4 and S5.

Figure 7. Introduction of AAV1-Parkin or Lenti-PGC-1α in mice SN protects from AAV1-PARIS-mediated selective dopaminergic neuronal toxicity

(A) Schematic illustration of intranigral viral injection and transduced brain regions. (B) Immunoblot analysis of PARIS, PGC-1α, parkin and NRF-1 four-weeks post intranigral injection of AAV1-PARIS, n = 5 per group. (C) Quantitation of the immunoblots in panel B normalized to β-actin. (D) TH staining of a representative section of mice injected with AAV1-GFP, AAV1-PARIS ± AAV1-parkin or AAV1-PARIS ± Lenti-PGC-1α. Each panel shows the noninjected side (Non) and contralateral injected side (Inj) and white pentagonal box indicates the SNpc. Enlarged images containing SNpc and SNpr are shown on the right panels. AAV1 encoding GFP was used as transduction control in all injection procedures. Broad regions including SNpc and SNpr were successfully transduced (left top panel). In right bottom panel, yellow rectangle indicates the region that PARIS and lenti- PGC-1α co-transduced. Approximately 30% of the SNpc was transduced with lenti-PGC-1α and this is the region, which is protected from PARIS toxicity, n = 6 per group. (E) Stereological TH, Nissl-positive neuronal counting, n = 6 per group. (F) Immunoblot analysis of PARIS, PGC-1α, parkin and NRF-1, n = 3. (G) Quantitation of the immunoblots in panel F normalized to β-actin. (H) Parkin-PARIS-PGC-1α pathway as a model in PD. Endogenous PARIS acts to maintain the balance of PGC-1α levels. In PD, parkin is inactivated by diverse insults such as familial mutations, reactive oxygen species (ROS), nitrosative (NO) and dopamine (DA) stress and PARIS accumulates. Accumulated PARIS continuously inhibits PGC-1α transcription leading to reduction in PGC-1α dependent genes. Ultimately this situation results in neurodegeneration in PD. Data = mean ± S.E.M. *p < 0.05, **p < 0.01, ***p < 0.001; ANOVA with the Student-Newman-Keuls post hoc test. See also Figure S6.