Discussion

Atg8 constitutively associates with autophagosomes after retrieval of other Atgs from the isolation membrane, and plays a critical role in the maturation of autophagosomes [205]. Residues on the surface of Atg8, especially those at the hydrophobic patch, mediate protein-protein interactions that are essential for autophagosome formation in yeast [192, 195, 198]. The existence of a conserved hydrophobic patch on the surface of LC3 has led to the hypothesis that LC3, like its yeast counterpart Atg8, could employ an evolutionarily conserved mechanism to govern cargo sorting and maturation of autophagosomes in mammalian systems [140, 141, 196]. Because selective autophagy has been suggested to play a pivotal role in the clearance of aberrant protein aggregates, elucidation of the molecular mechanism mediating the interaction between LC3 and its cargo receptor p62 could have profound implications in understanding the pathogenesis of various human diseases, including HD and other neurodegenerative disorders.

In this study, we investigate the LC3-p62 binding mechanism by overexpressing individual mutant LC3 cDNA constructs. Although the levels of exogenous GFP-LC3 protein were not always equivalent among transfectants, a phenomenon commonly seen in transient transfection of mammalian cells, the dominant-negative effects of these exogenous LC3 mutants on LC3-p62 interaction can still be vividly observed

70

even with the modest level of LC3-K51A, suggesting that these exogenous LC3 mutant proteins are sufficiently effective in substituting endogenous LC3 in autophagy. Given that Phe⁵², Leu⁵³, and the first N-terminal α-helix of LC3 have been found to be essential for p62 binding [206], the present evidence strongly supports a model in which the ubiquitin core protruding from the surface of the LC3 protein is indispensable for the electrostatic and hydrophobic interactions between LC3 and p62 [141]. Our data, in accordance with X-ray crystallographic and NMR analyses of Atg8-Atg19 and LC3-p62 complexes [207], further support the notion that Atg8 homologues from yeast to higher eukaryotes could employ an evolutionarily conserved mechanism to bind and deliver specific cargo proteins into autophagosomes.

Before our studies, there was no documentation discussing the residue Lys30 of LC3. Unlike Lys51 and Leu53 which locate in the central core of the hydrophobic pockets, Lys30 resides at the head of beta1-sheet, slightly away from the central ubiquitin core, (and composing a protrusion on the surface of LC3). It is likely that Lys30 serves as a recognition site for p62 through electrostatic attraction, and then promotes later association of p62 with LC3. This assumption is in accordance with our previous results that substitution of Arg28 in Atg8 (which corresponds to Lys 30 in LC3) to Asp significantly impaired its binding with Atg19, while mutation to Ala

71

had no such blocking effect [192]. Interestingly, although K30D mutant only half attenuates the recruitment of p62, it almost completely, like K51A and L53A, suppresses the clearance of N’Htt109Q. The p62 protein utilizes a C-terminal UBA domain to bind with Lys 63-linked polyUb chains of its substrates (herein, Htt), and this associating process involves a conformational switch of UBA domain [208], indicating a high flexibility of this structure. Since the LC3 recognition sequence (LRS) locates adjacent to UBA domain of p62, it is plausible that replacing a positive charge with a negative one at Lys30 site on LC3 might cause a conformational change of its binding partner p62 which leads to reduce the interaction between p62 and Htt.

p62 has also been found in neuronal inclusion bodies of various neurodegenerative disorders. In cultured cells, RNAi-mediated down-regulation of LC3s or overexpression of p62 defective in the LC3-recognition sequence (LRS) or UBA can lead to a significant increase in cytosolic ubiquitin-positive inclusion bodies and cell death [145, 209, 210]. For the first time, we focused on a specific protein, the aggregate-prone mutant of Htt (N’Htt109Q), which is cytotoxic and induces pathological phenotypes of Huntintung’s disease in mice models [211, 212]. Our results revealed that when LC3 mutants failed to trap p62 into autophagosomes, not only the degradation of p62 via autophagy was impaired, more accumulations of Htt aggregates in cells were observed, strongly prompting the pivotal function of p62 as a

72

specific cargo receptor involved in autophagy-mediated clearance of Htt, which greatly depends on the interaction between p62 and LC3. These findings are also correspondent with a previous finding that depletion of p62 levels by siRNA treatment or overexpressing p62 mutant lacking UBA domain which directly binds with Htt both results in severe cell death induced by mutant Htt [154]. Interestingly, some recent studies suggest that the deposition of inclusion bodies is rather a kind of protective mechanism to prevent the deleterious effects from toxic diffuse form of Htt [213]. We also observed the same inhibitory effects of LC3 mutants on the clearance of 1% Triton-soluble Htt detected by Western blot under denaturing condition, indicating that autophagy is also responsible for the degradation of monomeric or oligomeric soluble Htt proteins, mediated by interactions between LC3 and p62. We propose that rescue effect of autophagy on cell death induced by N’Htt-109Q could be reversed by expression of LC3 mutants.

Here, we displayed that mutant LC3s with less binding ability to p62 disrupt the transportation of aggregate-prone N’Htt109Q into autophagy. Developing strategies to upregulate autophgy or to maintain the level of functional p62 and LC3 would be beneficial for prevention and treatment of HD, and even other neurodegenerative diseases.

73

Chapter 3

Presenilin-1 regulates the expression of sequestosome-1/p62 and governs p62-dependent degradation of Tau independent of

γ-secretase activity

74

3.1 Abstract

Mutations in the gene encoding presenilin-1 (PS1) account for most cases of early-onset familial Alzheimer’s disease (FAD), which is characterized by

extracellular amyloid plaques and the accumulation of intracellular Tau. In addition to being the catalytic subunit of γ-secretase to generate amyloid-β, PS1 has been shown

to regulate diverse cellular functions independent of its proteolytic activity. We have found that cells deficient in PS1 exhibit reduced levels of p62 protein, a cargo receptor shuttling Tau for degradation in proteasomes or autophagosomes. The RNAi-mediated downregulation of PS1 also led to a significant decrease in both the protein and mRNA transcript of p62, concomitant with attenuated p62 promoter

activity. This PS1-dependent transcriptional regulation of p62 was mediated through an Akt/AP-1 pathway that does not require the proteolytic activity of PS1/γ-secretase.

In accordance with the role of p62 in mediating Tau degradation, the clearance of Tau was significantly impaired in PS1-deficient cells. Interestingly, such accumulation of Tau can be rescued by ectopic expression of either p62 or wild-type PS1 but not mutant PS1 containing FAD-linked mutations. Our data suggest that these FAD mutations of PS1 may cause a partial loss-of-function in PS1 to exacerbate tauopathies by suppressing p62 expression. Our study suggests a new role for PS1 in modulating the expression of p62 to control the degradation of aggregation-prone Tau

75

protein.

3.2 Introduction

Alzheimer’s disease (AD), the most prevalent type of dementia in the elderly,

affects more than 35 million people worldwide and imposes tremendous social costs in many countries. Extracellular plaques of amyloid-β (Aβ) and intracellular

neurofibrillary tangles (NFTs) in the brains of AD patients are two major pathological hallmarks of the disease and are widely regarded as the main causes of neurodegeneration in AD [214]. Genetic studies on cases of familial AD (FAD) have found that the genes encoding presenilin-1 (PS1) and its homologue, PS2, harbor 90%

of the mutations that cause AD [215]. PSs consist of the catalytic subunit of γ-secretase complexes, which cleave amyloid precursor protein (APP) to generate Aβ.

The majority of FAD-linked PS mutations show a change in the proteolytic activity of γ-secretase, leading to increased relative production of aggregation-prone Aβ42

peptides that accelerate the formation of plaques and contribute to neurodegeneration [215].

PSs are also involved in diverse cellular functions independent of γ-secretase activity, such as calcium homeostasis and the modulation of intracellular signaling [216-218]. PS1 can bind the p85 subunit of phosphatidylinositol 3-kinase (PI3K) to

76

activate the pro-survival PI3K/Akt pathway [219]. Neurons from PS1-null (PS1^-/-) embryonic mouse brains have reduced levels of PI3K/Akt activity and increased caspase 3-dependent neuronal apoptosis [220]. Interestingly, such defects can be rescued by the reintroduction of wild-type (WT) PS1 but not the FAD mutant PS1, suggesting that the loss-of-function effects of FAD mutations on PS1 could alter the PI3K/Akt pathway to elicit neurodegeneration in AD. Consistent with this notion, the conditional double-knockout (DKO) of PSs in the forebrain of adult mice triggers robust AD-like neurodegeneration, including brain shrinkage, hippocampal atrophy, and NFT-like structures, further substantiating the correlation between the loss of PS function and the pathogenesis of AD [221]. The manifestation of NFT-like structures built up by hyperphosphorylated and ubiquitinated Tau in the brains of PS-DKO mice and FAD patients with PS1 mutations further suggests that PSs could play a role in the modulation of Tau aggregation [221-223].

The proteostasis of Tau can be mediated by both the ubiquitin/proteasome system (UPS) and the autophagosomal/lysosomal pathway [160, 224, 225].

Deregulation of both protein degradation systems has been shown to contribute to tauopathies in AD. Sequestosome-1/p62 colocalizes with certain polyubiquitinated protein aggregates related to neurodegeneration [145, 146] and associates with ubiquitin-dependent Tau clearance [160]. The binding of p62 to

77

K63-polyubiquitinated proteins and the proteasomal Rpt1 protein is mediated by its ubiquitin-associated (UBA) and PB1 domains, respectively [153, 159]. The LC3 interacting region (LIR) encoded within p62 mediates p62-LC3 binding and the subsequent entrapment of p62 into autophagosomes for degradation [140, 141]. These findings strongly suggest a critical role for p62 in governing the homeostasis of Tau.

Our preliminary study revealed a significantly reduced level of p62 protein in PS1-knockdown cells. Given that AD brains exhibit reduced expression of p62 and that genetic inactivation of p62 in a mouse model leads to tauopathy [161, 169], we hypothesize that PS1 modulates p62 expression to control p62-dependent Tau degradation and that FAD-linked mutations of PS1 manifest tauopathy by impairing the p62-dependent clearance of Tau. We report the first direct evidence that PS1 deficiency results in reduced p62 expression through Akt/AP-1-dependent transcriptional regulation. The downregulation of PS1 and expression of mutant PS1 harboring FAD-linked mutations significantly impair p62-dependent Tau degradation.

Our results unveil a novel function of PS1 that may have profound implications in the pathogenesis of AD and the development of advanced therapeutics for AD.

78

3.3 Materials and Methods Reagents

Mouse anti-p62 (2C11) antibody was from Abnova. Rabbit anti-GAPDH (FL-335) antibody, horseradish peroxidase (HRP)-conjugated anti-rabbit IgG, HRP-conjugated anti-rat IgG, and HRP-conjugated anti-mouse IgG were from Santa Cruz Biotechnology, Inc. Rabbit anti-Akt, rabbit anti-phospho-Akt (Ser 473), and mouse anti-Tau (tau46) antibodies were from Cell Signaling Technology.

N-[N-(3,5-Difluorophenacetyl-L-alanyl)]-S-phenylglycine t-Butyl Ester (DAPT); rat anti-presenilin-1 monoclonal, rabbit anti-presenilin-1-loop, and rabbit anti-APP (C-terminal) antibodies; Immobilon Western Chemiluminescent HRP Substrate, and

protease inhibitor cocktail (set III) were from Merck Millipore. Mouse anti-β-actin

antibody (clone AC-15) and tetracycline were purchased from Sigma-Aldrich, Inc.

Lipofectamine 2000 transfection reagent, SuperScript® III Reverse Transcriptase, Dulbecco's modified Eagle's medium (DMEM), DMEM/F12, fetal bovine serum (FBS), and TRIzol Reagent were from Invitrogen. SYBR Green I Master reagent was from Roche. The QuikChange Site-Directed Mutagenesis kit was from Stratagene.

The BCA protein assay reagent kit was purchased from Pierce. Dual-Glo luciferase assay reagents and Steady-Glo luciferase assay reagents were from Promega.

Pepstatin A, EST, U0126, LY294002, PI-103, SB202190, SB203580, SP600125,

79

MG132, and chloroquine were from EMD Millipore Bioscience. Cellulose acetate membranes (OE66, 200 nm pore size) were from GE Healthcare. All other reagents were at least reagent grade and obtained from standard suppliers.

Cell culture

Human embryonic kidney cells (HEK)293 and mouse embryonic fibroblasts (MEF) generated from PS1^+/+PS2^+/+ (WT), PS1^-/-PS2^+/+, PS1^+/+PS2^-/-, and PS1^-/-PS2 -/-double-knockout (DKO) mice (kindly provided by Dr. Bart De Strooper, Flanders Interuniversity Institute for Biotechnology (VIB4) and K. U. Leuven, Belgium [226, 227]) were maintained in DMEM supplemented with 10% FBS. SH-SY5Y human neuroblastoma cells were maintained in DMEM/F12 medium supplemented with 10%

FBS. Cells were incubated in a humidified incubator at 37 ºC in 5% CO₂.

Virus preparation and infection

Lentivirus encoding gene-targeting or control shRNAs were produced in 293T cells transfected with pLKO.1-shRNA-PS1 or pLKO.1-shRNA-LacZ (transducing vectors), pCMV-dR8.91 (packaging vector), and pMD2.G (a plasmid expressing vesicular stomatitis virus G (VSV-G) glycoprotein) by the calcium phosphate transfection method. Forty-eight to seventy-two hours post-transfection,

80

virus-containing supernatant was collected, concentrated, and stored at -80 ºC. Viruses were added to HEK293 or SH-SY5Y cells in growth media. After 24 h, the media were replaced with fresh growth medium, followed by further treatment or assay.

Knockdown efficiency of the target gene was determined by Western blotting.

Quantitative real-time PCR

Total RNAs of transfected cells were prepared using TRIzol (Invitrogen) and were used to generate first-strand complementary DNA (cDNA) with the SuperScript™ III First-strand cDNA Synthesis Kit (Invitrogen). Equivalent amounts of cDNA were used in quantitative PCR on a LightCycler^® 480 system (Roche) with the following isoform-specific primer pairs: for mouse p62, forward primer

5´-GCTGCCCTATACCCACATCT-3´ and reverse primer

5´-CGCCTTCATCCGAGAA-3´; for human p62, forward primer

5´-ATCGGAGGATCCGAGTGT-3´ and reverse primer

5´-TGGCTGTGAGCTGCTCTT-3´. The levels of mouse S18-rRNA (forward primer

5´-AGGGGAGAGCGGGTAAGAGA-3´ and reverse primer

5´-GGACAGGACTAGGCGGAACA-3´) and human S18-rRNA (forward primer

5´-CAGCCACCCGAGATTGAGCA-3´ and reverse primer

5´-TAGTAGCGACGGGCGGTGTG-3´) were determined as internal controls. For

81

quantification, the transcript level of p62 was individually normalized to that of S18-rRNA. Quantitative data were expressed as the average (± SD) of triplicate measurements from at least three independent experiments.

Construction of reporter plasmids

DNA encoding the SQSTM1/p62 promoter was isolated from human genomic DNA (Clontech). A reverse primer

(5´-CCTCATGACGAGCGGCGAGCTGGCGGAA-3´), appended with a Bsp HI restriction site, and a forward primer

(5´-CCAAGCTTCCCAAATCCTTCCACTTCAGCCCC-3´), affixed with a Hind III restriction site, were used to amplify a 2335-bp fragment of genomic DNA that encoded the 5´UTR (-2140 to +48) of the p62 allele. The PCR product was digested with Bsp HI and Hind III and subcloned into pGL3-Basic (Promega) to generate the full-length WT p62 promoter-driven luciferase reporter construct (pGL3.p62[FL]).

To generate p62 promoter reporters with truncation or point mutations within sequences corresponding to particular transcription-factor binding elements, site-directed mutagenesis using the pGL3.p62[FL] plasmid as a template was performed with the QuikChange Site-Directed Mutagenesis kit according to the supplier’s instructions (Stratagene). The truncated pGL3.p62[0.3K] (-357 to +48)

82

promoter was generated by Kpn I digestion of pGL3.p62[FL] to remove 1700 upstream nucleotides, followed by self-ligation to recircularize the plasmid. All reporter constructs were verified by DNA sequencing.

Luciferase reporter assay

HEK293 or MEF cells were seeded onto 12-well microplates and transiently transfected with 0.5 µg of pGL3.p62[FL] or various mutant reporter constructs and 0.1 µg of pBud-Rrenilla (as a control for transfection efficiency) by Lipofectamine 2000 for 36-48 h. Expression of luciferase reporter genes in transfected cells was determined by using the Dual-Glo luciferase assay reagent kit. Luminescence emitted by the firefly luciferase reporter driven by p62 promoters (WT or mutant) was normalized to that of Renilla luciferase (pBud-Renilla). Relative promoter activity was expressed as the ratio of the normalized luminescence of the indicated p62 promoter reporter construct to that of pGL3-Basic. For kinase-inhibitor treatments,

LY294002 (10 µM), PI-103 (10 or 20 µM), U0126 (10 µM), sp600125 (10 µM), SB202190 (5 µM), SB203580 (10 µM), or Go6976 (0.1 µM) were added to individual

wells 24 h prior to harvesting the transfected cells.

83

Generation of expression vectors

The expression vector encoding the full-length WT Tau (containing exons 2, 3, and 10) N-terminally fused with EGFP (pEGFP-Tau) was kindly provided by Dr.

Pei-Jung Lu (Graduate Institute of Clinical Medicine, Medical College, National Cheng Kung University, Tainan, Taiwan; [228]). The pEGFP-Tau-P301L mutant was

generated from pEGFP-Tau by using the QuikChange Site-Directed Mutagenesis kit.

The DNA fragment encoding EGFP-Tau-P301L was subcloned into pcDNA5 to generate a tetracycline-inducible pc5-EGFP-Tau-P301L construct.

The expression vector encoding the full-length WT human PS1 (pcDNA3.1-PS1) was kindly provided by Dr. Michael Wolfe (Brigham and Women’s Hospital and Harvard Medical School, Boston, MA; [229]). FAD-linked mutations in mutant PS1 alleles, including A79V, M146L, G384A, and D257A, and ∆E9, were individually introduced into pcDNA3.1-PS1 by using the QuikChange Site-Directed Mutagenesis kit. The cDNA fragments encoding WT and mutant PS1 were subsequently subcloned into pAS2.puro or pAS3.zeo (National RNAi Core Facility, Academia Sinica, Taipei, Taiwan) separately for viral transduction.

The cDNA amplimer encoding the ORF of p62 was isolated from cDNAs generated from HEK293 cells and subcloned into pAS2.puro for viral transduction.

All constructs were verified by DNA sequencing.

84

SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis Cells were lysed in 1% Triton X-100 lysis buffer (20 mM HEPES, pH 7.4, 1%

Triton X-100, 10% glycerol, 2 mM EDTA, 50 µM β-glycerophosphate, 1 mM Na3VO4, and protease inhibitor cocktail set III) on ice for 30 min. Following removal of cell debris by centrifugation, protein concentrations of clarified lysates were determined by using the BCA protein assay reagent kit. Cell lysates containing equal amounts of proteins were mixed with the 6X sample-loading buffer and boiled at 100

°C for 10 min. To detect PS1 proteins, samples mixed with sample-loading buffer were directly subjected to SDS-PAGE without boiling to avoid aggregation of PS1 proteins. To examine the level of total Tau protein, SH-Tau cells were resuspended in 10 volumes (w/v) of PBS (37 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, and 1.47 mM KH₂PO₄) containing 5 mM EDTA and protease inhibitor cocktail set III, followed by sonication. The homogenate was spun for 10 min at 1500 g to remove nuclei and unbroken cells. Protein concentrations of clarified lysates were determined by using the BCA protein assay reagent kit. Cell lysates containing equal amounts of proteins were mixed with the 6X sample-loading buffer and boiled at 100 °C for 10 min. To detect insoluble Tau protein, clarified lysates were mixed with non-reducing sample-loading buffer and boiled at 100 °C for 10 min.

85

Proteins were resolved on 10% Tris-glycine polyacrylamide gels and then transferred electrophoretically to polyvinylidene difluoride membranes (Pall).

Membranes were blocked by 5% BSA in TBST (blocking buffer) at room temperature for 1 h, followed by incubation with appropriate primary antibodies in blocking buffer (1:5000) at 4 °C overnight. After three washes in TBST, the membranes were incubated with HRP-conjugated secondary antibodies in TBST (1:5000) at room temperature for 1 h. Following extensive washes with TBST, antibody-reactive proteins were visualized by the Immobilon Western Chemiluminescent HRP Substrate.

Images were captured and processed with ChemiGenius2 (Syngene).

Generation of tetracycline-inducible SH-Tau stable cell line

SH-SY5Y neuroblastoma cells were transfected with pcDNA6 (Invitrogen) encoding the Tet repressor by using Lipofectamine 2000 transfection reagent.

Transfected cells were cultured in DMEM supplemented with 10% FBS and 5 µg ml blasticidin. Individual colonies resistant to blasticidin selection were

isolated and tested for the tetracycline-inducible expression of EGFP reporter by transient transfection of the pcDNA5-EGFP plasmid. The clone (SH-pc6) with maximal induction of EGFP expression by tetracycline was chosen for subsequent experiments. SH-pc6 cells were further transfected with the pc5-EGFP-Tau-P301L

86

plasmid and cultured in DMEM supplemented with 10% FBS, 200 µg/ml hygromycin,

and 5 µg/ml blasticidin. Colonies resistant to antibiotic selection were isolated individually and screened for the tetracycline-inducible expression of GFP-tagged Tau^P301L. The stable clone exhibiting maximal tetracycline induction of EGFP-Tau^P301L expression was selected and named SH-Tau for subsequent experiments.

Tau degradation assay

Tetracycline-inducible SH-Tau cells were treated with 0.2 µg/ml tetracycline in growth medium for 3-5 d to induce Tau expression (tet-on). The media were replenished with fresh tetracycline-containing medium every 2-3 d. To terminate Tau expression, induced SH-Tau cells were washed twice with PBS and maintained in fresh culture medium without tetracycline (tet-off). Cells were harvested at various intervals after terminating the tetracycline-induced expression of nascent Tau protein.

In some experiments, 0.2 µM MG132 or 20 µM chloroquine (CQ) were added to the tet-off media, and treated cells were harvested after 24 h. The level of nascent Tau protein in clarified lysates was analyzed by Western blotting, followed by densitometric quantification using Image J software (NIH). The intensity of Tau was normalized with that of actin protein, and the normalized level of Tau at tet-off day 0