Protein aggregates and damaged organelles are tagged with ubiquitin chains to trigger
selective autophagy. To initiate mitophagy, PINK1 phosphorylates ubiquitin to activate
Parkin, which builds ubiquitin chains on mitochondrial outer membrane proteins where
they act to recruit autophagy receptors. Using genome editing to knock out five autophagy
receptors, we find that two previously linked to xenophagy, NDP52 and Optineurin,
are the primary receptors for PINK1/Parkin-mediated mitophagy. The ubiquitin kinase
PINK1 recruits NDP52 and Optineurin, but not p62, to mitochondria to directly activate
mitophagy independent of Parkin. Once recruited to mitochondria, NDP52 and Optineurin
recruit ULK1, DFCP1 and WIPI1 to focal spots proximal to mitochondria revealing a
function for these autophagy receptors upstream of LC3. This supports a new model
that PINK1 generated phospho-ubiquitin serves as the autophagy signal on mitochondria
and that Parkin amplifies it. This work also suggests direct and broader roles for
ubiquitin phosphorylation in other autophagy pathways.
Selective autophagy clears intracellular pathogens and mediates cellular quality control
by engulfing cargo into autophagosomes and delivering it to lysosomes for degradation.
Autophagy receptors bind ubiquitinated cargo and LC3-coated phagophores to mediate
autophagy
1,2
. Damaged mitochondria are removed by autophagy following activation of the kinase
PINK1 and the E3 ubiquitin ligase Parkin
3,4
. Upon loss of mitochondrial membrane potential or accumulation of misfolded proteins,
PINK1 is stabilized on the outer mitochondrial membrane
3
, where it phosphorylates ubiquitin at Ser65 to activate Parkin ubiquitin ligase activity
5–7
. Although the autophagy receptors p62 and Optineurin (OPTN) have been shown to bind
ubiquitin chains on damaged mitochondria, their roles, and the roles of the other
autophagy receptors in mediating mitophagy is unclear
8–11
.
Autophagy receptors in mitophagy
To clarify autophagy receptor function during mitophagy, genome editing was used to
knock out five autophagy receptors in HeLa cells (pentaKO), which do not express endogenous
Parkin. DNA sequencing (Supplementary Table 1) and immunoblotting of TAX1BP1, NDP52,
NBR1, p62 and OPTN (Fig. 1a, lane 6) confirmed their knockout. We analyzed mitophagy
in pentaKOs by measuring the degradation of cytochrome C oxidase subunit II (CoxII),
a mtDNA encoded inner membrane protein, following mitochondrial damage with oligomycin
and antimycin A (OA). After OA treatment, CoxII was degraded in WT cells expressing
Parkin, but not in pentaKOs or ATG5 KO HeLa cells, indicating a block in mitophagy
(Fig. 1b, c, Supplementary Table 1 and Extended Data Fig. 1a). As a second indicator
of mitophagy, mitochondrial DNA (mtDNA) nucleoids were quantified by immunofluorescence
(Extended Data Fig. 1b). After 24 h OA treatment, WT cells were nearly devoid of mtDNA,
whereas pentaKOs and ATG5 KOs retained mtDNA (Fig. 1d, e). Parkin translocated to
mitochondria (Extended Data Fig. 1c) and Mfn1 and Tom20 were degraded via the proteasome
comparably in WT and pentaKOs (Fig. 1b, Extended Data Fig. 1d). mtDNA nucleoids clump
following OA treatment in ATG5 KO cells but not in pentaKOs, consistent with a reported
role of p62
10,11
.
The five endogenous receptors in WT cells (Extended Data Fig. 1c) and each receptor
re-expressed in pentaKOs (Extended Data Fig. 1e, f) translocated to mitochondria after
OA treatment. However, in pentaKOs only GFP-NDP52, GFP-OPTN and to a lesser extent,
GFP-TAX1BP1, rescued mitophagy (Fig. 1f, g). Another recently reported autophagy receptor,
Tollip
12
, neither recruited to mitochondria nor rescued mitophagy following OA treatment (Extended
Data Fig. 1g–i).
We generated single OPTN, NDP52 KO and NDP52/OPTN double KO (N/O DKO) and NDP52/OPTN/TAX1BP1
triple KO (N/O/Tx TKO) cell lines (Supplementary Table 1, Fig. 1a) and found no compensatory
change in the expression of the remaining receptors. NDP52 or OPTN KO alone caused
no defect in mitophagy, whereas NDP52/OPTN DKO and to a greater extent, NDP52/OPTN/TAX1BP1
TKO inhibited mitophagy (Fig. 2a–d, Extended Data Fig. 2a, b). The robust mitophagy
observed in OPTN KOs contrasts with a report indicating loss of mitophagy using RNAi-mediated
knockdown of OPTN in HeLa cells
9
. Although NDP52 and OPTN redundantly mediate mitophagy, they function non-redundantly
in xenophagy
13
. Their expression levels in human tissues indicate that OPTN or NDP52 may function
more prominently in different tissues (Extended Data Fig. 2c).
Mutations in autophagy receptors can lead to diseases such as primary open angle glaucoma
(POAG, OPTN; E50K)
14
, ALS (OPTN; E478G and Q398X)
15
and Crohn's disease (NDP52; V248A)
16
. Defects in xenophagy occur when OPTN is mutated to block its phosphorylation by
TANK-binding kinase 1 (TBK1; S177A) or ubiquitin binding (D474N)
13,17
. In pentaKOs, the UBAN-domain disrupting mutants OPTN-Q398X, OPTN-D474N and OPTN-E478G
(Extended Data Fig. 2d) failed to translocate to mitochondria (Extended Data Fig.
2e, f) or rescue mitophagy (Fig. 2e, Extended Data 2g). OPTN-S177A weakly rescued
mitophagy and minimally translocated to mitochondria, whereas OPTN-E50K robustly translocated
and substantially rescued mitophagy (Fig. 2e, Extended Data Fig. 2e–g). NDP52-V248A
fully recruited to mitochondria and rescued mitophagy, but a mutant lacking the ZF
ubiquitin-binding domains (NDP52-ΔZF)
18
did not (Extended Data Fig. 2h–k, Fig. 2f). Thus, ubiquitin binding by OPTN and NDP52
is necessary for mitophagy and some disease-causing mutations prevent mitophagy.
TBK1 and OPTN cooperate in mitophagy
TBK1 phosphorylation of OPTN at S177 increases its association with LC3 during xenophagy
13
, and the OPTN E50K mutation increases TBK1/OPTN binding
19
. TBK1 auto-phosphorylation at Ser172 is indicative of TBK1 activation
20
and occurs in a Parkin-dependent manner following 3 h OA treatment, but only in cells
expressing OPTN (Extended Data Fig. 3b, lanes 4 and 10). Prolonged OA treatment induces
moderate TBK1 phosphorylation in the absence of Parkin but still requires PINK1 (Extended
Data Fig. 3c). To investigate TBK1 function during mitophagy, we generated TBK1 KO,
TBK1/NDP52 (T/N) DKO and TBK1/OPTN (T/O) DKO HeLa cells (Extended Data Fig. 3d, Supplementary
Table 1). Parkin translocated to mitochondria in all lines, however, only TBK1/NDP52
DKOs displayed defective mitophagy (Extended Data Fig. 3e, Fig. 2g–j). Mitophagy in
TBK1/NDP52 DKOs was rescued by WT-TBK1 or phospho-mimetic OPTN (OPTN-S177D), but not
by kinase-dead TBK1 (TBK1-K38M) (Extended Data Fig. 3g–i). Thus, in the absence of
NDP52, TBK1 is critical for effective mitophagy via OPTN.
Ubiquitin phosphorylation in mitophagy
Since many autophagy receptors recruit to mitochondria following Parkin activation,
why do only some function in mitophagy? Parkin-mediated mitophagy is driven by PINK1's
phosphorylation of Ser65 of both ubiquitin
5–7,21,22
and the UBL domain of Parkin
23
. Since Ser65 phospho-ubiquitin is structurally unique, it may differentially interact
with ubiquitin binding proteins
22
. To determine whether OPTN is directly recruited to phospho-ubiquitin on mitochondria,
we conditionally expressed PINK1 on undamaged mitochondria
10
in HeLa cells lacking Parkin (Fig. 3a, Extended Data Fig. 4a). When PINK1Δ110-YFP-2xFKBP
is cytosolic, mCherry-OPTN, mCherry-NDP52 and mCherry-p62 are also cytosolic (Extended
Data Fig. 4b, c). When PINK1Δ110-YFP-2xFKBP is localized to FRB-Fis1 expressing mitochondria
with rapalog, where ubiquitin on surface proteins
24
(Extended Data Fig. 4a) can be phosphorylated
5,21,25,26
, OPTN and NDP52 are recruited (Fig. 3a, b), but p62 remains cytosolic (Fig. 3b, Extended
Data Fig. 4c). OPTN/NDP52 recruitment requires PINK1 kinase activity (Fig. 3a, b)
and receptor-ubiquitin binding, as OPTN-D474N and NDP52-ΔZF fail to recruit following
rapalog treatment (Fig. 3b, Extended Data Fig. 4d, e). Therefore, PINK1 ubiquitin
kinase activity recruits OPTN/NDP52 via ubiquitin binding domains to mitochondria
in the absence of Parkin.
To determine whether the observed autophagy receptor recruitment to mitochondria in
the absence of Parkin can induce mitophagy, we developed a sensitive FACS based mitophagy
assay. We expressed mitochondrial-targeted mKeima (mt-mKeima, see Online Methods)
in WT and pentaKOs also expressing mitochondrial FRB-Fis1 and PINK1Δ110-YFP-2xFKBP.
mt-mKeima engulfment into lysosomes results in a spectral shift due to low pH. Only
1% (range 0.89–1.15) of WT or pentaKO cells display mitophagy when PINK1 is cytosolic.
However, when PINK1 is recruited to mitochondria with rapalog, mitophagy increases
~7-fold in WT cells and ~8-fold with overexpressed OPTN (Table 1, Extended Data Fig.
5a). PentaKOs showed no increase in mitophagy after targeting PINK1 to mitochondria
(Table 1, Extended Data Fig. 5b). When rescued with FLAG/HA-OPTN or FLAG/HA-NDP52,
pentaKOs displayed an increase in mitophagy of more than 5-fold and 4-fold, respectively
(Table 1, Extended Data Fig. 5b–e). Rescue with FLAG/HA-p62 or ubiquitin-binding mutants
(OPTN-Q398X, OPTN-D474N and NDP52ΔZF) failed to increase mitophagy above baseline,
but other mutants (OPTN-E50K, OPTN-S177A and NDP52-V248A) rescued mitophagy (Table
1, Extended Data Fig. 5c–f). OPTN-E50K and S177A restored mitophagy as well as or
better than WT OPTN (Table 1), differing from their response in the presence of Parkin
(Fig. 2e) likely due to the lack of robust TBK1 activation in the absence of Parkin
(Extended Data Fig 3b). Here, enhanced OPTN-E50K binding to TBK1
19
may become advantageous by allowing OPTN phosphorylation by TBK1 in the absence of
Parkin thus improving mitophagy. In the absence of TBK1 activation, WT OPTN is likely
not phosphorylated at S177 and thus is functionally similar to S177A OPTN. Importantly,
ubiquitin kinase activity of PINK1 is required, as kinase-dead (KD) PINK1 did not
induce mitophagy (Table 1, Fig. 3c, Extended Data 5g). Parkin expression dramatically
increased mitophagy in FLAG/HA-OPTN expressing pentaKOs (Table 1, Extended Data Fig.
5h), supporting the model that PINK1-phosphorylated ubiquitin recruits receptors for
mitophagy and Parkin ubiquitination of mitochondrial substrates amplifies this ubiquitin
signal.
Comparing mitophagy induced by OA treatment in WT relative to PINK1KO cells confirmed
that endogenous PINK1 mediates mitophagy in the absence of Parkin (Extended Data Fig.
6a, b). Re-expressing PINK1 in PINK1 KO cells rescued OA-induced mitophagy (Extended
Data Fig. 6c, d). Furthermore, mCherry-OPTN is recruited to mitochondria in the absence
of Parkin in a PINK1-dependent manner following prolonged exposure to OA (Extended
Data Fig. 6e, f).
Given that PINK1 ubiquitin kinase activity can recruit OPTN and NDP52, we investigated
autophagy receptor binding to phospho-mimetic (S65D) HA-ubiquitin in HeLa cells. Endogenous
OPTN and NDP52 preferentially co-immunoprecipitate (co-IP) with HA-ubiquitinS65D (Extended
Data Fig. 7a). Conversely, p62 was present at equal levels in all co-IPs (Extended
Data Fig. 7a). Ubiquitin-modified and unmodified forms of OPTN and NDP52 were present
in co-IPs, and HA-ubiquitinS65D induced or preserved this modification (Extended Data
Fig. 7a). Co-IP samples treated with the deubiquitinase USP2 removed the ubiquitin-modified
bands on OPTN and NDP52, yet OPTN and NDP52 retained HA-ubiquitinS65D binding (Extended
Data Fig. 7b). Binding of endogenous receptors in HeLa cell cytosol to in vitro phosphorylated
strep-tagged ubiquitin (Extended Data Fig. 7c) showed that OPTN, but not p62, bound
better to phospho-ubiquitin (Extended Data Fig. 7d, e). However, recombinant GST-OPTN
did not bind better to in vitro phosphorylated K63 linked ubiquitin chains
27
indicating that OPTN may need additional factors or modification in vivo to preferentially
bind Ser65 phosphorylated ubiquitin.
OPTN/NDP52 recruit upstream machinery
Autophagy receptors are thought to primarily function by bridging LC3 and ubiquitinated
cargo
1,2
. In mCherry-Parkin WT cells, GFP-LC3B accumulated in distinct puncta adjacent to
mitochondria after OA treatment (Extended Data Fig. 8a). Although OA also induced
GFP-LC3B puncta in pentaKOs, they were fewer and not near mitochondria (Extended Data
Fig. 8a). Conversely, GFP-LC3B in ATG5 KOs was near mitochondria, but not in puncta
(Extended Data Fig. 8a). LC3B lipidation is retained in pentaKOs, but lost in ATG5
KOs (Extended Data Fig. 8b). This indicates that ATG5 is activated downstream of PINK1,
but independently of autophagy receptors, and that LC3 lipidation and mitochondrial
localization are independent steps of mitophagy.
OPTN and NDP52 interact with LC3B and LC3C, respectively, for Salmonella clearance
13,28
. Beyond that, little is known about the specificity of LC3 family members toward
autophagy receptors
29
or their involvement in mitophagy. We examined the recruitment of all LC3/GABARAP
family members to mitochondria in WT, pentaKO and NDP52/OPTN DKO cells. The OA-induced
mitochondrial localization of GFP-LC3s in WT cells was absent in pentaKOs, while only
GFP-LC3B recruitment was inhibited in NDP52/OPTN DKOs (Fig. 4a, Extended Data Fig.
8c). GFP-LC3C recruitment was inhibited in NDP52/OPTN/TAX1BP1 TKOs (Extended Data
Fig. 8d, e), indicating that TAX1BP1 can recruit LC3C during mitophagy. GABARAPs did
not recruit to mitochondria, indicating they likely play no substantial role in mitophagy
(Extended Data Fig. 9a).
We also examined the involvement of WIPI1 and DFCP1, two proteins that mediate phagophore
biogenesis upstream of LC3
30
, in mitophagy. In WT cells, OA induced foci of both GFP-WIPI1 and GFP-DFCP1, mostly
localized on or near mitochondria (Fig. 4b, c, Extended Data Fig. 9b, c). In NDP52/OPTN
DKOs, GFP-WIPI1 and GFP-DFCP1 foci were reduced and were almost undetectable in pentaKOs
(Fig. 4b, c, Extended Data Fig. 9b, c). Despite this, phosphorylation of Beclin1
31
was normal in both pentaKOs and NDP52/OPTN DKOs (Extended Data Fig. 9d), indicating
that failure to recruit WIPI1/DFPC1 was not due to defective Vps34 complex. GFP-DFCP1
recruitment in pentaKOs was rescued by expression of FLAG/HA-OPTN or FLAG/HA-NDP52,
but not by FLAG/HA-p62 (Extended Data Fig. 10a).
Though autophagy receptors are thought to function late in autophagy with LC3
32
, the deficit in WIPI1 and DFCP1 recruitment to mitochondria indicates a defect upstream
in autophagosome biogenesis. ULK1 phosphorylation by AMPK at S317 and dephosphorylation
at S757
33
, required for activation, occurs comparably in WT, NDP52/OPTN DKO and pentaKO cells
(Fig. 4d). Despite this, ULK1 localization to mitochondria
34
following OA is diminished by half in the NDP52/OPTN DKOs and more than 80% in pentaKOs
(Fig. 4e, f). FLAG/HA-OPTN or FLAG/HA-NDP52, but not FLAG/HA-p62, rescued GFP-ULK1
localization in pentaKOs (Extended Data Fig. 10b). Overall, these data indicate that
NDP52 and OPTN recruit ULK1 to initiate mitophagy.
We next assessed if ubiquitin phosphorylation, independent of Parkin, is also sufficient
to recruit ULK1 to mitochondria. Rescue of pentaKOs expressing FRB-Fis1 and PINK1Δ110-YFP-FKBP
with myc-OPTN or myc-NDP52 resulted in mitochondrial ULK1 puncta following rapalog
treatment (Fig. 4g, h). Myc-OPTN-E50K also rescued ULK1 recruitment to mitochondria,
but ALS-associated mutant myc-OPTN-Q398X did not (Fig. 4i, Extended Data Fig. 10d).
ULK1 recruitment was restored by myc-OPTN-F178A (Fig. 4i, Extended Data Fig. 10d),
a mutation that disrupts OPTN association with LC3
12
, indicating that ULK1 recruitment is not through LC3 interaction and occurs upstream
of LC3. Taken together, our data show that PINK1 ubiquitin-kinase activity is sufficient
to recruit the autophagy receptors and upstream autophagy machinery to mitochondria
to induce mitophagy.
Conclusions
Through genetic knockout of five autophagy receptors we have defined their relative
roles in mitophagy and identified their unanticipated upstream involvement in autophagy
machinery recruitment. p62 and NBR1 are dispensable for Parkin-mediated mitophagy;
OPTN and NDP52 are the primary, yet redundant, receptors. We also uncovered a new
and more fundamental role for PINK1 in mitophagy: to directly induce mitophagy through
phospho-ubiquitin-mediated recruitment of autophagy receptors. We posit that PINK1
generates the novel and essential signature (phospho-ubiquitin) on mitochondria to
induce OPTN and NDP52 recruitment and mitophagy; Parkin acts to increase this signal
by generating more ubiquitin chains on mitochondria, which are subsequently phosphorylated
by PINK1. Our findings clarify the role of Parkin as an amplifier of the PINK1-generated
mitophagy signal, phospho-ubiquitin, which can engage the autophagy receptors to recruit
ULK1, DFCP1, WIPI1 and LC3 (see model in Extended Data Fig. 10e).
Online Content
Methods, along with any additional Extended Data display items, are available in the
online version of the paper; references unique to these sections appear only in the
online paper.
METHODS
Cell Culture, Antibodies and Reagents
HEK293T, HeLa and PINK1 KO
35
cells were cultured in Dulbecco's modified eagle medium (Life Technologies) supplemented
with 10% (v/v) Fetal Bovine Serum (Gemini Bio Products), 10 mM HEPES (Life Technologies),
1 mM Sodium Pyruvate (Life Technologies), nonessential amino acids (Life Technologies)
and GlutaMAX (Life Technologies). HeLa cells were acquired from the ATCC and authenticated
by the Johns Hopkins GRCF Fragment Analysis Facility using STR profiling. All cells
were tested for mycoplasma contamination bimonthly using the PlasmoTest kit (InvivoGen).
Transfection reagents used were: Effectene (Qiagen), Lipofectamine LTX (Life Technologies),
Avalanche-OMNI (EZ Bio-systems), X-tremeGENE HP (Roche) and X-tremeGENE 9 (Roche).
Rabbit monoclonal and polyclonal antibodies used: Beclin, pULK1-S317, pULK1-S757,
TBK1, pTBK1-S172, NDP52, TAX1BP1, ATG5, Actin, and HA (Cell Signaling Technologies);
GAPDH and LC3B (Sigma); ULK1 and Tom20 (Santa Cruz Biotechnology); Optineurin (OPTN)
(Proteintech); GFP (Life Technologies); pSer65 ubiquitin (Millipore) and Mfn1 was
generated previously
36
. Mouse monoclonal antibodies used: NBR1 and p62 (Abnova), Cytochrome C oxidase subunit
II (CoxII, Abcam), Parkin (Santa Cruz Biotechnology), DNA (Progen Biotechnik), ubiquitin
(Cell Signaling). Chicken anti-GFP (Life Technologies) was also used. For catalog
numbers see Supplementary Table 1. Human tissue panel blots were purchased (NOVUS
Biologicals).
Generation of knockout lines using TALEN and CRISPR/Cas9 gene editing
To generate knockout cell lines, TALENs and CRISPR gRNAs were chosen that targeted
an exon common to all splicing variants of the gene of interest (listed in Supplementary
Table 1). Transcription activator-like effector nuclease (TALEN) was used to generate
the OPTN KO HeLa cell line. The TALEN constructs were generated by sequential ligation
of coding repeats into pcDNA3.1/Zeo-Talen(+63), as previously described
37–39
. The CRISPR/Cas9 system generated by the Church lab
40
, was used to knockout ATG5, NDP52, TAX1BP1, NBR1, p62 and TBK1. Oligonucleotides
(Operon) containing CRISPR target sequences were annealed and ligated into AlfII-linearized
gRNA vector (Addgene)
40
. For CRISPR/Cas9 gene editing, HeLa cells were transfected with gRNA constructs,
hCas9 (Addgene) and pEGFP-C1 (Clontech), or for TALEN gene editing HeLa cells were
transfected with OPTN TALEN constructs and pEGFP-C1. Two days after transfection,
GFP-positive cells were sorted by fluorescence activated cell sorting and plated in
96-well plates. Single colonies were expanded into 24-well plates before screening
for depletion of the targeted gene product by immunoblotting. As a secondary screen
of some knockout lines, genomic DNA was isolated from cells and the genomic regions
of interest were amplified using PCR followed restriction enzyme digestion analysis
(primers listed in Supplementary Table 1). Sequencing of targeted genomic regions
of knockout lines was also conducted to confirm the presence of frameshifting indels
in the genes of interest (Supplementary Table 1). To generate multiple gene knockout
cell lines, parental cell lines were transfected sequentially with one or multiple
gRNA constructs to generate desired knockout lines. Parental cell lines are outlined
in Supplementary Table 1.
Cloning and generation of stable cell lines
pMXs-puro-GFP-WIPI1 and pMXs-puro-GFP-DFCP1 were a kind gift from Dr. N. Mizushima
(University of Toyko, Japan) and pMXs-IP-GFP-ULK1 was purchased from Addgene (#38193).
To generate pBMN-mEGFP-C1, mEGFP-C1 (Addgene #36412) was PCR amplified (together with
the multiple cloning site) and cloned into pBMN-Z at BamHI/SalI sites using the Gibson
Cloning kit (New England BioLabs) according to manufacturer's instructions. The BamHI
and SalI sites used to insert mEGFP-C1 were not regenerated. The following GFP-tagged
plasmids were generated by PCR amplification of open reading frames followed by ligation
into pBMN-mEGFP-C1: OPTN, NDP52, p62, TAX1BP1, NBR1, LC3A, LC3B, LC3C, GABARAP, GABARAPL1,
GABARAPL2. The Gateway Cloning (Invitrogen) system was used to generate GFP-, mCherry-,
myc- and FLAG/HA-constructs. Briefly, TBK1, TBK1-K38M, NDP52, OPTN, p62, DFCP1, WIPI1
and ULK1 were cloned into pDONR2333. Mutations in cDNA sequences were introduced using
PCR site directed mutagenesis in the pDONR2333 vector, (sequences of mutagenesis primers
used are available upon request) then recombined into pHAGE-N-FLAG/HA, pHAGE-N-GFP,
pHAGE-N-mCherry and/or pDEST-N-myc using LR Clonase (Invitrogen) as per the manufacturer's
protocol. All constructs generated in this study were verified by sequencing.
To generate stably transfected cell lines, retroviruses (for pBMN-mEGFP-C1 constructs,
pBMN-mCherry-Parkin, pBMN-puro-P2A-FRB-Fis1, pCHAC-mt-mKeima-IRES-MCS2) and lentiviruses
(for pHAGE- and pDEST- constructs) were packaged in HEK293T cells. HeLa cells were
transduced with virus for 24 h with 8 μg/ml polybrene (Sigma) then optimized for protein
expression via selection (puromycin or blasticidin) or fluorescence sorting.
Translocation and mitophagy treatments
Cells were either left untreated or treated with 10 μM Oligomycin (Calbiochem), □
μM Antimycin A (Sigma) (referred to as OA) in fresh growth medium for different periods
of time as indicated in the figures. Some experiments were performed with 10 μM Carbonyl
cyanide m-chlorophenyl hydrazine (CCCP) as indicated (Sigma-Aldrich). We chose to
use OA to depolarize mitochondria in most of our experiments, as they are specific
mitochondrial respiratory complex inhibitors and less toxic. Long treatment time points
of both OA and CCCP were also supplemented with the apoptosis inhibitor 20 μM QVD
(ApexBio) to prevent cell death.
Immunoblotting and Phos-Tag gels
HeLa cells seeded into 6-well plates were either untreated or treated with 10 μM Oligomycin
(Calbiochem), □ μM Antimycin A (Sigma) and 20 μM QVD (ApexBio) in fresh growth medium
for different periods of time as indicated in figure legends. Cells were lysed in
1X LDS sample buffer (Life Technologies) supplemented with 100 mM dithiothreitol (DTT,
Sigma) and heated to 99 °C with shaking for 7–10 minutes. 25–50 μg of protein per
sample was separated on 4–12% Bis-Tris gels (Life Technologies) according to manufacturer's
instructions and then transferred to polyvinyl difluoride membranes and immunoblotted
using antibodies as indicated in figure legends. To assess mitophagy, CoxII quantification
was conducted using ImageLab software (BioRad). For uncropped images of all immunoblots,
see Supplementary Information.
To dephosphorylate samples, cells were collected as above and lysed in 1X NEB Buffer
3 (New England BioLabs) supplemented with 1% Triton X-100 and passed through a 26.5
gauge needle. Calf intestinal phosphatase (CIP, New England BioLabs) was added to
half the cell lysate and the other half was used as an untreated control. Both samples
were incubated for 1 h at 37 °C and analyzed by SDS-PAGE and immunoblotting.
To analyze Beclin phosphorylation, lysates were prepared in sample buffer lacking
EDTA and run on 8% Tris-Glycine gels containing 20 μM Phos-Tag (Wako) and 40 μM MnCl2
as described previously
31
. Gels lacking Phos-Tag were run simultaneously as a negative control. Electrophoresis
and western transfer were carried out using standard protocols with the exception
that Phos-Tag gels were incubated in 10 mM EDTA for 10 min to remove excess Mn2+ prior
to transfer.
Immunoprecipitation
WT or PINK1 KO HeLa cells were transiently transfected with HA-tagged ubiquitin WT,
S65A or S65D with or without mCherry-Parkin for 24 h. Cells were harvested, lysed
and the HA-ubiquitin was immunoprecipitated as reported previously
5
, using anti-HA conjugated beads (Pierce). To deubiquitinate the bound proteins, after
binding the HA-ubiquitin, beads were washed three times and incubated in 50 mM Tris-Cl
(pH 7.5), 150 mM NaCl, 5 mM DTT and 1.47μg USP2 (Boston Biochem) at 37 °C for 1 h.
The reaction was stopped and the remaining bound were proteins were washed 5 times
with 1 mL of buffer (50 mM Tris-Cl (pH 7.5), 150 mM NaCl), then eluted by boiling
with 1X LDS sample buffer.
In vitro phosphorylation
Strep-tagged ubiquitin was incubated with either TcPINK1 WT or kinase-dead as previously
reported
5
. This ubiquitin was then incubated with cytosol from WT HeLa cells in 20 mM HEPES-KOH,
pH 7.6, 220 mM mannitol and 70 mM sucrose at 4°C for 1 h. Strep-Tactin beads (Qiagen)
were then added to bind the strep-ubiquitin for an additional 1 h at 4°C. The ubiquitin
and bound proteins were then eluted with 50 mM biotin in 50 mM Tris for 15 min at
room temperature (RT). Samples were then diluted in LDS sample buffer prior to SDS-PAGE
and immunoblot analysis.
Immunofluorescence microscopy
HeLa cells, seeded in 2-well chamber slides (Lab-Tek), were treated as indicated in
the figures legends. Following treatment, cells were rinsed in PBS and fixed for 15
min at RT with 4% paraformaldehyde. Cells were then permeabilized and blocked with
0.1% Triton X-100, 3% goat serum in PBS for 40 minutes at RT. For immunostaining,
cells were incubated with antibodies (as indicated in figure legends) diluted in 3%
goat blocking serum overnight at 4 °C, then rinsed with PBS and incubated with either
anti- rabbit or mouse Alexa Fluor- 488 and 633 conjugated secondary antibodies (Life
Technologies), or anti-chicken Alexa Fluor 488 conjugated antibody (Life Technologies)
for 1 h at RT. Cells were washed 3 times for 5 min each with 1% Triton X-100, PBS.
During the final wash step, cells were incubated with DAPI (10 μg/mL DAPI, Sigma)
in PBS for 5 min. To measure mitophagy by mitochondrial DNA (mtDNA) immunostaining;
images were collected from samples stained with DAPI and immunostained for DNA using
a plan-Apochromat 63×/1.4 oil DIC objective on an LSM 510 microscope (Zeiss). Four
image slices were collected through the Z plane encompassing the top and bottom of
the cells. Image analysis was performed on all images collected in the Z plane using
Volocity software (Perkin Elmer v6.0.1). The percent mtDNA stain remaining was calculated
using the following formula: (cDNAv-nDNAv)/n, where cDNAv= the total cellular DNA
volume determined by staining using anti-DNA antibodies and, nDNAv = the total nuclear
DNA stain volume determined using DAPI, and n= the number of cells. The mtDNA stain
volume in untreated cells was normalized to 100% and the amount of mtDNA stain remaining
after drug treatment was subsequently determined. Final values represent data acquired
from 50–200 cells from three independent experiments.
To analyze LC3/mitochondria protein colocalization; cells were treated, fixed and
immunostained as above. Between 5–8 slices were imaged through the Z plane using either
a plan-Apochromat 63× or 100×/1.4 oil DIC objective on a CW STED confocal microscope
(Leica). Volocity software (Perkin Elmer, v6.0.1) was used to measure intensity of
the GFP signal representing LC3 in the volume occupied by mitochondria (as defined
by Tom20 positive region) and the cytosol (as defined by Tom20 negative region). “Normalized
mitochondrial LC3” was calculated using the following formula: Normalized mitochondrial
LC3 = (mi/mv)/(ci/cv), where mi = mitochondrial GFP intensity, mv = mitochondrial
volume, ci = cytosolic GFP intensity and cv= cytosolic volume. The resulting Normalized
mitochondrial LC3 is equal to 1 if the intensity of GFP is equal per volume in the
cytosolic and mitochondrial volumes (no translocation) and is above one if the mitochondrial
intensity is higher per volume (translocation). Final values for Normalized mitochondrial
LC3 represents data acquired from 50–105 cells from three independent experiments.
For GFP-DFCP1, GFP-WIPI1 and GFP-ULK1 puncta analysis; cells were treated, prepared
and imaged on the CW STED as above with the addition of immunofluorescence using either
rabbit or chicken GFP antibodies to enhance the signal in the green channel. For GFP-DFCP1,
puncta were quantified using Volocity software (Perkin Elmer v6.0.1) and for GFP-WIPI1
and GFP-ULK1 puncta were quantified manually. Colocalization of autophagy receptors
with GFP-DFCP1 or GFP-ULK1 was assessed with line scans using LAS AF software (Leica,
v.2.6.0.7266).
Heterodimerization
The C-terminal Fis1 tail of human Fis1 (amino acids 92–152) was cloned into pC4-RhE
vector (ARIAD) at SpeI/BamHI sites to make FRB-Fis1 construct, the insert of which
was then PCR amplified and cloned into pBMN-Z vector together with Puro-P2A sequence
at HindIII, XhoI and NotI sites by In-Fusion kit from Clontech to make pBMN-puro-P2A-FRB-Fis1.
For receptor translocation assays, WT HeLa cells stably expressing FRB-Fis1 were generated
using retroviral transduction as described above. Previously generated PINK1Δ110-YFP-2xFKBP
41
WT and KD and individually each mCherry-tagged autophagy receptor were transfected
into FRB-Fis1 stable HeLa cells for 24 h. Cells were then treated with 0.5 μM rapalog
(Clontech) for 8 h as previously described
41
. Cells were then fixed and stained as described above. Cells were manually counted
for translocation of mCherry-tagged autophagy receptors to mitochondria. Final values
represent data collected from 100–150 cells for three independent experiments.
For ULK1 and DFCP1 rescue analysis, pentaKO HeLa cells stably expressing FRB-Fis1
were transiently transfected with PINK-1Δ110-YFP-2xFKBP, mCherry-ULK1 and one of the
autophagy receptors: myc-OPTN, myc-NDP52, myc-OPTN-F178A, myc-OPTN-E50K or myc-OPTN-Q398X
for 18–24 h. Cells were treated with 0.5 μM Rapalog or 100% ethanol (vehicle) for
7 h and imaged live on a Zeiss 780 in a humidified 37°C/5% CO2 chamber. To visualize
mitochondria in vehicle -treated controls, cells were pre-incubated for 10 minutes
in 75 nM Mitotracker Deep Red (Invitrogen) prior to imaging. Fields of PINK1-YFP positive
cells were imaged blindly to the mCherry-containing channel. The images were then
blinded and counted manually for translocation of mCherry-tagged ULK1 to mitochondria.
Final values represent >75 cells counted over at least two independent experiments.
Mito-Keima mitophagy assay
mt-mKeima
42
(a gift from A. Miyawaki, Brain Science Institute, RIKEN, Japan) was cloned into pCHAC-MCS-1-IRES-MCS2
vector (Allele Biotechnology). PINK1(Δ110)-YFP-2xFKBP WT and KD were PCR amplified
from the original pC4 and cloned into pRetroQ-AcGFP-C1 at NheI/XhoI sites by Gibson
assembly kit. HA-tag was removed and stop codon is introduced. WT and 5KO pentaKO
HeLa cells stably expressing FRB-Fis1, mt-mKeima and either WT or KD PINK1Δ110-YFP-2xFKBP
were generated using retroviral transduction as described above. FLAG/HA-receptors
were stably-expressed in these cells by lentivirus transduction as described above
then treated with 0.5 μM rapalog for 24 h. Cells were then resuspended in sorting
buffer (145 mM NaCl, 5 mM KCl, 1.8 mM CaCl2, 0.8 mM MgCl2, 10 mM HEPES, 10 mM glucose,
0.1% BSA) containing 10 μg/mL DAPI. Analysis was performed using Summit software (v6.2.6.16198)
on a Beckman Coulter MoFlo Astrios cell sorter. Measurements of lysosomal mt-mKeima
were made using dual-excitation ratiometric pH measurements at 488 (pH 7) and 561
(pH 4) nm lasers with 620/29 nm and 614/20 nm emission filters, respectively. For
each sample, 50,000 events were collected and subsequently gated for YFP/mt-mKeima
double-positive cells that were DAPI-negative. Data were analyzed using FlowJo (v10,
Tree Star).
Statistical calculations
All statistical data were calculated and graphed using GraphPad Prism 6. To assess
statistical significance, data from three or more independent experiments were analyzed
using one-way ANOVA and Tukey's post-test with a confidence interval of 95%. All error
bars are expressed as mean ± standard deviation (s.d.). In Fig.4h, i outliers were
removed using ROUT in GraphPad Prism 6 with a Q=1%, 1–2 values from each condition
were removed.
Extended Data
Extended Data Figure 1
Analysis of knockout cell lines and characterization of autophagy receptor translocation
to damaged mitochondria
a, ATG5 KO cell line confirmed by immunoblotting. b, Representative images of mitochondrial
DNA nucleoids in HeLa cells immunostained with an α-DNA antibody (green) confirming
colocalization with the mitochondrial marker Tom20 (red) (n=3). c, Mitochondrial fractions
from mCherry-Parkin (mCh-Parkin) expressing pentaKO and WT cells were assessed by
immunoblotting. d, mCh-Parkin expressing WT, pentaKO and ATG5 KOs were treated with
OA or OA and MG132. Cell lysates were assessed by immunoblotting. e, Expression levels
of GFP-tagged OPTN, NDP52, p62, NBR1 and TAX1BP1 re-expressed in pentaKOs by immunoblotting.
f, Representative images of mCh-Parkin expressing pentaKOs from e immunostained for
Tom20 (n=3). g, Expression of GFP-Tollip in mCh-Parkin pentaKOs. h, pentaKOs mCh-Parkin
and with or without GFP-Tollip expression were immunoblotted. i, Representative images
of mCh-Parkin pentaKOs expressing GFP-Tollip immunostained for Tom20 (n=3). Scale
bars, 10 μm.
Extended Data Figure 2
OPTN, NDP52 and TAX1BP1 triple knockout analysis and disease-associated mutations
a, KO cell lines with or without mCherry-Parkin (mCh-Parkin) expression were immunoblotted
and b, CoxII levels were quantified. c, A panel of human tissue lysates was immunoblotted.
d, Expression of WT or mutant GFP-OPTN in mCh-Parkin pentaKOs. e, Quantification of
cells in f. >100 cells per condition. f, Representative images of mCh-Parkin pentaKOs
expressing GFP-OPTN mutants immunostained for Tom20 (n=3). g, pentaKOs expressing
mCh-Parkin were rescued with WT or mutant GFP-OPTN, analyzed by immunoblotting. See
Fig. 2e for quantification of CoxII. h, Expression of WT or mutant GFP-NDP52 in mCh-Parkin
pentaKOs. i, Quantification of cells in j. >100 cells per condition. j, Representative
images of mCh-Parkin pentaKOs expressing WT or mutant GFP-NDP52 were immunostained
for Tom20 (n=3). k, pentaKOs expressing mCh-Parkin rescued with WT or mutant GFP-NDP52
were analyzed by immunoblotting. See Fig. 2f for quantification of CoxII. Quantification
in b and i are displayed as mean ± s.d. from 3 independent experiments using one-way
ANOVA tests (***P<0.001, ns, not significant) and in e as mean from 2 independent
experiments. Scale bars, 10 μm.
Extended Data Figure 3
TBK1 in activates OPTN in PINK1/Parkin mitophagy
a, Representative images of untreated mCherry-Parkin (mCh-Parkin) cells and merged
images of treated cells as indicated immunostained for DNA. See Fig. 2a for anti-DNA/DAPI
images of treated samples (n=3). b, Cell lysates from WT, N/O (NDP52/OPTN) DKO, OPTN
KO and NDP52 KO cells with or without mCh-Parkin expression were immunoblotted for
TBK1 activation. c, Cell lysates from WT and PINK1 KO cells without Parkin expression
were immunoblotted for TBK1 activation (S172 phosphorylation). d, Confirmation of
T/N (TBK1/NDP52) DKO, T/O (TBK1/OPTN) DKO and TBK1 KO by immunoblotting. e, KO cell
lines from d were immunostained for Tom20 (n=3). f, Representative images of untreated
mCh-Parkin WT and KO cells, and merged images of treated cells as indicated were immunostained
for DNA. See Figure 2g for anti-DNA/DAPI images treated samples (n=3). g, T/N DKO
cells rescued with GFP-TBK1 WT or K38M, or GFP-OPTN S177D and were assessed by immunoblotting.
h, Cells in g were assessed by immunoblotting. i, Quantification of CoxII levels in
h displayed as mean ± s.d. from 3 independent experiments and use one-way ANOVA tests
(**P<0.005, ns, not significant). OA, Oligomycin and Antimycin A. Scale bars, 10 μm.
Extended Data Figure 4
Parkin-independent recruitment of receptors to mitochondria through PINK1 activity
a, Isolated mitochondria from WT and pentaKOs with or without FRB-Fis1 and with WT
or kinase-dead (KD) PINK1Δ110-YFP-2xFKBP were immunoblotted. b–e, Representative images
of pentaKOs expressing FRB-Fis1, WT (PINK1-WT) or kinase-dead (PINK1-KD) PINK1Δ110-YFP-2xFKBP
and either (b) mCherry-OPTN or mCherry-NDP52, (c) mCherry-p62, (d) mCherry-OPTN-D474N
or (e) mCherry-NDP52-ΔZF. Cells were (b) untreated or (c–e) treated with rapalog then
immunostained for Tom20. All images are representative of three independent experiments.
See Figure 3b for quantification. Scale bars, 10 μm.
Extended Data Figure 5
PINK1 directly stimulates mitophagy in the absence of mitochondrial damage
a, b, Cells were treated with rapalog and analyzed by FACS for lysosomal positive
mt-mKeima. Representative data for WT HeLa (a) and pentaKO (b) without or with FLAG/HA-OPTN.
c, d, Cell lysates from pentaKOs expressing FRB-Fis1, PINK1Δ110-YFP-2xFKBP, mt-mKeima
and (c) WT FLAG/HA-OPTN or mutants, (d) FLAG/HA-p62, WT FLAG/HA-NDP52 or NDP52 mutants
as indicated were assessed for receptor expression by immunoblotting. e, f, Cells
from c and d were rapalog treated analyzed by FACS for lysosomal positive mt-mKeima.
Representative data of two experiments is presented. g, Cell lysates from pentaKOs
expressing FRB-Fis1, with or without FLAG/HA-OPTN and WT or kinase-dead (KD) PINK1Δ110-YFP-2xFKBP
were assessed for OPTN by immunoblotting. h, FLAG/HA-OPTN pentaKOs expressing FRB-Fis1,
PINK1Δ110-YFP-2xFKBP, mt-mKeima transfected and either vector or untagged Parkin were
analyzed by FACS. Representative data of two experiments is presented.
Extended Data Figure 6
PINK1 directly stimulates mitophagy upon mitochondrial damage
Representative data of mt-mKeima-expressing a, WT, PINK1 KO or c, PINK1 KO rescued
with PINK1-WT cells treated with OA then analyzed by FACS. b, d, Average percent mitophagy
for two replicates of a and c, respectively. e, Representative images of WT HeLa cells
expressing mCherry-OPTN and treated with OA as indicated were immunostained for Tom20
(n=3). f, Quantification of mCherry-OPTN translocation from cells in e. Data displayed
as mean ± s.d. from 3 independent experiments and using one-way ANOVA tests (***P<0.001,
ns, not significant).
Extended Data Figure 7
OPTN and NDP52 preferentially bind phospho-mimetic ubiquitin
a, HeLa cells expressing mCherry-Parkin (Parkin) and HA-ubiquitin (HA-UB) WT, S65D
or S65A were treated with CCCP. HA-UB was co-immunoprecipitated and the bound fraction
was analyzed by immunoblotting. Quantification of the total bound fraction of OPTN,
NDP52 and p62 are shown. b, HA-ubiquitin transfected into HeLa cells with mCherry-Parkin
were treated with CCCP. HA-ubiquitin was immunoprecipitated. The bound fraction was
treated with the deubiquitinase USP2 and washed to remove all unbound protein following
deubiquitination. Quantification of the total bound fraction of OPTN, NDP52 and p62
are shown in the right panel. c,d, Strep-tagged ubiquitin (Strep-UB) was incubated
with either WT or kinase-dead (KD) PINK1 in an in vitro phosphorylation reaction,
immunoblotted with an anti-phosphoS65 ubiquitin antibody (c) and was then incubated
with cytosol harvested from untreated, WT HeLa cells. The ubiquitin was then pulled
down using Strep-Tactin beads and (d) analyzed by immunoblotting. e, Quantification
of bound OPTN and p62 normalized to total ubiquitin. Data displayed in a, b and e
as mean ± s.d. from 3 independent experiments and use one-way ANOVA tests. (***P<0.001,
**P<0.005, *P<0.05). †, non-specific band. a.u., arbitrary units.
Extended Data Figure 8
Analysis of LC3 family members and their translocation to damaged mitochondria in
autophagy receptor KO cell lines
a, Representative images of WT, pentaKO and ATG5 KO HeLa cells expressing mCherry-Parkin
(mCh-Parkin) and GFP-LC3B were immunostained for Tom20 (n=3). b, Cell lysates from
mCh-Parkin expressing WT, pentaKO and ATG5 KO cells were immunoblotted. c, Representative
images of WT, N/O (NDP52/OPTN) DKO and pentaKOs expressing mCh-Parkin and either GFP-tagged
LC3A, LC3B or LC3C were immunostained for Tom20 (n=3, see Figure 4a for quantification).
d, Representative images of WT and N/O/Tx (NDP52/OPTN/TAX1BP1) TKO cells expressing
mCh-Parkin and GFP-LC3C were immunostained for Tom20 (n=3) and e, quantified for GFP-LC3C
translocation to mitochondria. Quantification in e is displayed as mean ± s.d. from
3 independent experiments and use one-way ANOVA tests (***P<0.001). OA, Oligomycin
and Antimycin A. Scale bars, 10 μm.
Extended Data Figure 9
GABARAPs do not translocate to damaged mitochondria and early stages of autophagosome
biogenesis mediated by WIPI1 and DFCP1 are inhibited in autophagy receptor deficient
cell lines
Representative images of WT, N/O (NDP52/OPTN) DKO and pentaKOs expressing mCherry-Parkin
(mCh-Parkin) and either (a) GFP-tagged GABARAP, GABARAPL1 or GABARAPL2, (b) GFP-WIPI1
or (c) GFP-DFCP1 immunostained for Tom20 (n=3 for each condition, see Figure 4b, c
for quantification of b and c). d, mCh-Parkin cell lines as indicated were subjected
to either Phos-Tag SDS-PAGE or standard SDS-PAGE followed by immunoblotting. Arrows
indicate the position of phosphorylated Beclin species. e, Representative images of
untreated WT, N/O (NDP52/OPTN) DKO and pentaKO cell lines expressing mCh-Parkin and
GFP-ULK1 were immunostained for Tom20 and GFP (n=3). OA, Oligomycin and Antimycin
A. Scale bars, 10 μm.
Extended Data Figure 10
OPTN and NDP52 rescue DFCP1 and ULK1 recruitment deficit in pentaKOs
a, Representative images of pentaKOs expressing mCherry-Parkin (mCh-Parkin), GFP-DFCP1
and the indicated FLAG/HA-tagged autophagy receptors immunostained for HA (n=2). Right-hand
panels display co-localization of FLAG/HA-tagged constructs and GFP-DFCP1 by fluorescence
intensity line measurement. b, Representative images of pentaKOs expressing mCherry-Parkin
and GFP-ULK1 were rescued with FLAG/HA-OPTN, FLAG/HA-NDP52, and FLAG/HA-p62, and immunostained
for HA and GFP. Arrows indicate HA-tagged receptor puncta (n=2). Right panels display
colocalization of HA and GFP by fluorescence intensity line measurement. c, d, Representative
images of pentaKOs stably expressing FRB-Fis1 and transiently expressing PINK1Δ110-YFP-2xFKBP
and vector or myc-tagged receptors, were (c) untreated or (d) treated with rapalog
and imaged live (n=3, see Figure 4h, i for quantification of c, d). OA, Oligomycin
and Antimycin A. Scale bars, 10 μm. e, Old and new models of PINK1/Parkin mitophagy.
The old model is dominated by Parkin ubiquitination of mitochondrial proteins. Here
PINK1 plays a small initiator role whose main function is to bring Parkin to the mitochondria.
The new model depicts Parkin-dependent and independent pathways leading to robust
and low-level mitophagy, respectively. Based on our data, PINK1 is central to mitophagy
both before and after Parkin recruitment by phosphorylating UB to recruit both Parkin
and autophagy receptors mitochondria, to induce clearance. In the absence of Parkin
(right panel), this occurs at a low level due to the relatively low basal UB on mitochondria.
When Parkin is present it serves to amplify the PINK1 generated UB-PO4 signal, allowing
for robust and rapid mitophagy induction.
Supplementary Material
1
2
supp_table