A RT I C L E
The mitochondrial protease HtrA2 is regulated by Parkinson’s disease-associated kinase PINK1 Hélène Plun-Favreau1,5,8, Kristina Klupsch1,8, Nicoleta Moisoi4, Sonia Gandhi5, Svend Kjaer2, David Frith3, Kirsten Harvey6, Emma Deas5, Robert J. Harvey6, Neil McDonald2, Nicholas W. Wood5, L. Miguel Martins4,7 and Julian Downward1,7 In mice, targeted deletion of the serine protease HtrA2 (also known as Omi) causes mitochondrial dysfunction leading to a neurodegenerative disorder with parkinsonian features. In humans, point mutations in HtrA2 are a susceptibility factor for Parkinson’s disease (PARK13 locus). Mutations in PINK1, a putative mitochondrial protein kinase, are associated with the PARK6 autosomal recessive locus for susceptibility to early-onset Parkinson’s disease. Here we determine that HtrA2 interacts with PINK1 and that both are components of the same stress-sensing pathway. HtrA2 is phosphorylated on activation of the p38 pathway, occurring in a PINK1-dependent manner at a residue adjacent to a position found mutated in patients with Parkinson’s disease. HtrA2 phosphorylation is decreased in brains of patients with Parkinson’s disease carrying mutations in PINK1. We suggest that PINK1dependent phosphorylation of HtrA2 might modulate its proteolytic activity, thereby contributing to an increased resistance of cells to mitochondrial stress. The serine protease HtrA2 was first identified as a mammalian homologue of the Escherichia coli endopeptidases HtrA (DegP) and DegS1,2. Like its bacterial orthologues, HtrA2 has a PDZ domain in addition to its serine protease domain. In mammalian cells, HtrA2 is localized to the mitochondria. A mature form of HtrA2 can be generated by proteolysis, revealing an amino-terminal sequence motif related to that of the Drosophila death-promoting proteins Reaper, Grim and Hid. During apoptosis, mature HtrA2 is released from the mitochondria into the cytosol, where it binds the inhibitor of apoptosis proteins (IAPs) by means of its N-terminal sequence. Binding of mature HtrA2 to IAPs relieves their inhibitory action towards caspases, contributing to the induction of apoptosis3-6. In addition, this binding results in an increase in the proteolytic activity of HtrA2 (ref. 7), further contributing to apoptotic cell death8. HtrA2 has therefore been proposed to be a pro-apoptotic protein analogous to Reaper.
However, this view of HtrA2 function has been thrown into doubt by the phenotype of mice with a homozygous deletion of the HtrA2 gene or a point mutation causing defective protease activity; these both die prematurely and have a parkinsonian neurodegenerative phenotype9,10. It is possible that HtrA2 acts to protect mitochondria from certain stresses in a manner similar to the homologous stress-adaptive proteases DegP and DegS in bacteria11,12. We therefore studied whether HtrA2 is important in the aetiology of Parkinson’s disease in humans and have identified and characterized two mutations in HtrA2 linked to the disorder that deregulate its serine protease activity13, leading to designation of the HTRA2 gene as the Parkinson’s disease-13 (PARK13) locus. Loss of function of HtrA2 thus leads to neuronal cell death, possibly as a result of mitochondrial dysfunction, both in mice and in humans. Here we set out to identify proteins interacting with HtrA2 and identified PTEN (phosphatase and tensin homologue)-induced putative kinase 1 (PINK1) as a binding partner. We subsequently determined that HtrA2 is phosphorylated on activation of p38, and that PINK1 is important for this phosphorylation event to occur. We also show that phosphoHtrA2 levels are decreased in the brains of patients with Parkinson’s disease with mutations in PINK1. Phosphorylation of HtrA2 is likely to modulate the proteolytic activity of HtrA2 and contribute to increased resistance of cells to mitochondrial stress. RESULTS PINK1 interacts with HtrA2 A carboxy-terminally tandem affinity purification (TAP)-tagged version of HtrA2 was stably expressed in HEK-293 cells, enabling the recovery and purification of tagged HtrA2 along with interacting proteins. With the use of 35 antibodies raised against a range of 29 mitochondrial proteins as well as an antibody against XIAP (X-linked inhibitor of apoptosis) as a positive control, the purified fraction containing HtrA2 was analysed (Table 1). We co-purified HAX-1, a protein previously shown
Signal Transduction, 2Structural Biology and 3Protein Analysis Laboratories, Cancer Research UK London Research Institute, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK. 4Cell Death Regulation Laboratory, MRC Toxicology Unit, Lancaster Road, Leicester LE1 9HN, UK. 5Department of Molecular Neuroscience, Institute of Neurology, and National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK. 6Department of Pharmacology, The School of Pharmacy, 29/39 Brunswick Square, London WC1N 1AX, UK. 7 Correspondence should be addressed to L.M.M. (
[email protected]) or J.D. (
[email protected]) 8 These authors contributed equally to this work. 1
Received 8 June 2007; accepted 7 September 2007; published online 30 September 2007; DOI: 10.1038/ncb1644
nature cell biology volume 9 | number 11 | NOVEMBER 2007 © 2007 Nature Publishing Group
1243
SH-SY5Y
Mr(K) 75
Buffer
HAX-1
30
WT
35
HtrA2 KO
IP: HtrA2
WT
Mr(K)
b TAP
a
HtrA2 KO
HtrA2–TAP
A RT I C L E
WB: PINK1
75 XIAP
PINK1
PINK 1-581
Mr(K) 35
PINK 1–507
50
PINK 150–507
35
PINK1
Mock
Mr(K)
PINK1
e
Mock
PINK 150–581
50 35
HtrA2
581
PINK 1–507
1
d
507
Kinase domain (Pfam)
PINK 150–507
156
Mock
c
PINK 150–581
50
HtrA2 + PINK 1–507
75
WB: HtrA2
*
HtrA2 + PINK 150–507
35
HtrA2 + PINK 150–581
50
IP: Myc (PINK1) WB: Flag (HtrA2)
IP: Myc (PINK1) WB: Myc (PINK1)
IP: HtrA2 WB: Myc (PINK1)
IP: HtrA2 WB: Myc (PINK1)
50 *
IP: HtrA2 WB: HtrA2
35 Cytosol
Mitochondria
Figure 1 HtrA2 binds PINK1. (a) Purification of HtrA2-interacting proteins. HtrA2–TAP or TAP were purified from HEK-293 stable cell lines and eluates were probed with the indicated antibodies. (b) Immunoprecipitation of endogenous PINK1 together with endogenous HtrA2 from MEFs or SH-SY5Y cells. KO, knockout; WB, western blot; WT, wild type. The asterisk indicates an IgG background band. (c) Schematic representation of three deletion proteins of PINK1 (Pfam is a database of alignments of protein domains)
(d) Immunoprecipitation (IP) of HtrA2 together with the kinase domain of PINK1. HEK-293 cells transiently expressing full-length HtrA2–Flag and truncated PINK1–Myc were used. (e) Immunoprecipitation of HtrA2 together with PINK1 in the mitochondria-enriched fraction. PINK1–Myc was expressed in HEK-293 cells, and cytosol and mitochondria-enriched fractions were prepared. Uncropped images of blots in d and e are shown in Supplementary Information, Fig. S5.
to interact with HtrA2 (ref 14), as well as XIAP, demonstrating the validity of this approach. One other protein was found to purify together with HtrA2, the putative mitochondrial kinase PINK1 (Fig. 1a). Mutations in the PINK1 gene are associated with the recessive Parkinsonism locus PARK6 (refs 15–20). Given that dysfunction in both HtrA2 and PINK1 has been associated with parkinsonian neurodegeneration in mammals, we proposed that they might both be components of the same stress-sensing pathway. To confirm the interaction between endogenous PINK1 and HtrA2, we immunoprecipitated endogenous HtrA2, probed with anti-PINK1 antibody and detected endogenous PINK1 in complexes from wild-type mouse embryonic fibroblasts (MEFs) and SH-SY5Y human neuroblastoma cells, but not from cells lacking HtrA2 (HtrA2 knockout MEFs) or in lysis buffer alone (Fig. 1b). To determine the regions of PINK1 involved in interaction with HtrA2, various deletion mutants of PINK1 were created (Fig. 1c). PINK1 mutant proteins were all at least partly localized together to the mitochondria when overexpressed in U2OS cells, including PINK 150-507, which encoded only the kinase domain (see Supplementary
Information, Fig. S1). Simultaneous purification of Myc-tagged PINK1 deletion mutants and Flag-tagged HtrA2 showed that the kinase domain of PINK1 was sufficient for the interaction with HtrA2 (Fig. 1d). Immunoprecipitation experiments from subcellular fractions showed that the interaction between HtrA2 and PINK1 was detected predominantly in mitochondrial fractions (Fig. 1e). Although the interaction between PINK1 and HtrA2 is therefore likely to occur in the mitochondria, the possibility cannot be excluded that the initial contact occurs in the cytosol before mitochondrial import.
1244
HtrA2 is phosphorylated on activation of MEKK3 A mutation screen of the HTRA2 gene performed in German patients with Parkinson’s disease recently led to the identification of novel heterozygous G399S and A141S mutations that were associated with the disease13. These mutations localize to regions previously shown to be important in the regulation of the proteolytic activity of HtrA2 (ref. 7), namely the PDZ domain and the N-terminal portion of the mature form of HtrA2, respectively. An analysis with the Scansite algorithm21 nature cell biology volume 9 | number 11 | NOVEMBER 2007
© 2007 Nature Publishing Group
A RT I C L E a
Potential proline-dependent Ser/Thr kinase phosphorylation sites IAP-binding domain * A V P S P P P A S P R S
Mr(K) 35 35
50
MEKK3:ER – +
Akt:ER – +
Raf:ER – +
Protease domain
c
– –
4OH-Tx
Mr(K) P-S142 HtrA2 35
HtrA2
35
P-Akt
35
PDZ domain 1
+
–
5 +
–
10 – +
+
BIRB796 (µM)
d
4OH-Tx
siRNA MKK3–6
b
TM
Control
MT
* H K V I L G S P A H R A
P-S142 HtrA2
Mr(K)
HtrA2
35
P-S142 HtrA2
35
HtrA2
P-p38 p38
50
35 P-ERK1/2 P-p38
35 P-JNK
e
P-p38
Mr(K) 35
50 35
AA
WT
–
–
AA
WT
HtrA2
p38
p38γ P
35
33
35
HtrA2
50
35
Coomassie 1
2
3
4
5
6
Figure 2 HtrA2 is phosphorylated on MEKK3 activation. (a) Schematic representation of human HtrA2. MT, mitochondrial targeting sequence; TM, transmembrane domain. The IAP-binding domain at the N terminus of mature HtrA2 is indicated. Residues mutated in Parkinson’s disease (underlined) and potential Ser/Thr kinase phosphorylation sites (asterisks) are shown. (b) HtrA2 is phosphorylated on Ser 142 on activation of MEKK3. ∆MEKK3:ER (MEKK3:ER), myrAkt:ER (Akt:ER) and ∆Raf-DD:ER (Raf: ER) stable cell lines transiently expressing full-length HtrA2 were activated with 4OH-Tx. Lysates were analysed by western blotting with the indicated antibodies. (c) p38 inhibition by BIRB796 inhibits MEKK3-induced HtrA2 phosphorylation on Ser 142. ∆MEKK3:ER cells transiently expressing
full-length HtrA2 were incubated with BIRB796 and analysed by western blotting. (d) Transient downregulation of MKK3 and MKK6 with the use of siRNA decreases endogenous Ser 142 HtrA2 and p38 phosphorylation. Uncropped images of blots in c and d are shown in Supplementary Information, Fig. S5. (e) Ser 142 and/or Ser 400 are required for HtrA2 phosphorylation by p38γ. In vitro kinase assay with recombinant wild-type or S142A,S400A (AA) HtrA2. In parallel, gels were stained with colloidal Coomassie blue. Images are from the same exposures. Results are representative of three independent experiments. Quantification indicated a 2.0-fold increased signal in lane 5 relative to lane 3 and a 2.5-fold increased signal in lane 6 relative to lane 4.
(http://scansite.mit.edu) indicates that Ser 142 and Ser 400, residues immediately adjacent to the mutations found in the patients with Parkinson’s disease, are putative phosphorylation sites for prolinedirected serine/threonine kinases (Fig. 2a). Alignment of HtrA2 homologues of different species shows that Ser 142 and Ser 400 are conserved in human, rodent and chicken, and in addition Ser 400 is further conserved in fly and baker’s yeast (see Supplementary Information, Fig. S2a). Ser 142 is a putative phosphorylation site for Cdc2, Cdk5, extracellular signal-regulated kinase (ERK)1 and p38, and Ser 400 for Cdc2 and Cdk5. We tested whether recombinant HtrA2 could be phosphorylated by any of these kinases, finding that p38β, p38γ, p38δ, ERK1, Cdc2 and Cdk5 could do so in vitro (see Supplementary Information,
Fig. S2b). To determine the position of the phosphorylation sites on HtrA2, we subsequently performed mass spectrometric analysis using HtrA2 phosphorylated in vitro by p38γ. This analysis detected a phosphoserine at position 142 (data not shown). Subsequently, we raised an antibody that specifically recognized HtrA2 only when phosphorylated on Ser 142. Because of the behaviour of the Ser 400-containing peptide in mass spectrometric analysis and the failure of our attempts to generate phospho-specific antibody against Ser 400 that would recognize HtrA2 under any circumstances, we were unable to address the phosphorylation state of this site in cell extracts directly. To investigate the nature of the kinase phosphorylating HtrA2 on Ser 142 in intact cells, we used cell lines expressing 4-hydroxytamoxifen (4OH-Tx)inducible versions of MAP kinase/ERK kinase kinase kinase 3 (MEKK3),
nature cell biology volume 9 | number 11 | NOVEMBER 2007 © 2007 Nature Publishing Group
1245
A RT I C L E Table 1 Identification of proteins complexed with HtrA2 Company (clone)
Antigen
Detected in HtrA2–TAP eluate
SCBT (Q-18 and N-19)
ANT
No
SCBT (C-20), BD (clone 48)
Bad
No
Stressg (2-14), Calbio (Ab-1)
Bak
No
Mito (YTH-6A7)
Bax
No
SCBT (N-19), Pharm (4D7)
Bcl-2
No
BD (clone 64)
Bcl-XL
No
Abcam (30A882.1.1)
Bif-1
No
SCBT (N-19, FL-60)
Bik
No
SCBT (N-20)
Bim
No
Calbio (77-92)
Bnip3L
No
MolProbes (21C11)
17-kDa subunit (Complex I)
No
MolProbes (2E3)
Complex II
No
MolProbes (7H10)
Complex V subunit
No
MolProbes (3D5)
Complex V subunit
No
MolProbes (7F9)
Complex V subunit d
No
MolProbes (4C11)
Complex V subunit OSCP
No
MolProbes (5E2)
Complex V inhibitory protein
No
Mito (12F4AD8AF8)
Complex V
No
Pharm (7H8.2C12)
Cytochrome c
No
V. Dawson
DJ-1
No
Stressg (30A5)
Grp-75
No
BD (clone 52)
HAX-1
Yes
SCBT (N-20)
Hrk
No
Stressg (mAb 11-13)
Hsp-60
No
CRUK generated
PINK1
Yes
Abcam (II-14-10)
Prohibitin
No
CRUK generated
Smac
No
Calbiochem (Q-18)
VDAC
No
SCBT (C-17)
VDAC2
No
BD (clone 48)
XIAP
Yes
TAP epitope-tagged HtrA2 was purified from HEK-293 cells stably expressing the fusion protein, then resolved by SDS–PAGE and probed with antibodies raised against mitochondrial proteins. SCBT, Santa Cruz Biotechnology; BD, BD Transduction Laboratories; Stressg, Stressgen; MolProbes, Molecular Probes; Calbio, Calbiochem; Mito, Mitosciences.
Akt and Raf (∆MEKK3:ER, myrAkt:ER and ∆Raf-DD:ER). Only activation of MEKK3 induced efficient phosphorylation of both full-length and processed HtrA2 on Ser 142 (Fig. 2b and see Supplementary Information, Fig. S2c). Activation of MEKK3 resulted in strong activation of endogenous p38, weaker activation of ERK or c-Jun N-terminal kinase (JNK) and no activation of Akt (Fig. 2b). To test whether chemical inhibition of p38 affected MEKK3-dependent phosphorylation of HtrA2 on Ser 142, we treated cells expressing ∆MEKK3:ER with increasing concentrations of BIRB796, a broad-spectrum inhibitor of all p38 MAPK isoforms22. This partly decreased the levels of phospho-p38 and phopho-HtrA2, confirming the importance of p38 for the phosphorylation of HtrA2 (Fig. 2c). Downregulation of MAPK kinase (MKK)3 and MKK6 also led to a decrease in the levels of phospho-p38 and phospho-Ser 142 HtrA2 (Fig. 2d). To determine whether p38γ phosphorylation could target either of the two identified serine residues, we performed an in vitro kinase assay using S142A,S400A HtrA2. This mutant protein was considerably less phosphorylated by p38γ in vitro (Fig. 2e). Nevertheless, a slight increase in the incorporation of radioactive phosphate into HtrA2 remained, 1246
possibly explained by phosphorylation on Ser 137, which was observed by mass spectroscopy with wild-type HtrA2 phosphorylated in vitro by p38γ (data not shown). This site was not pursued further, because it is not conserved between human and murine HtrA2 (see Supplementary Information, Fig. S2a). In vitro, p38 is also able to phosphorylate a peptide containing the sequences around Ser 142. As a control, we used a peptide in which Ser 142 was mutated into Ala, further confirming the specificity of the assay. The Ser 142 peptide could be phosphorylated only by p38, not by ERK1, Cdc2 or Cdk5 (see Supplementary Information, Fig. S2d). PINK1 modulates the levels of phosphorylated HtrA2 We next sought to determine whether PINK1 could modulate the phosphorylation of HtrA2 in cells. We first determined that transfection of short interfering RNA (siRNA) against PINK1 resulted in a significant decrease of both PINK1 mRNA and PINK1 protein (see Supplementary Information, Fig. S3a, b). Transient downregulation of PINK1 resulted in a decrease in the levels of endogenous phospho-Ser 142 HtrA2 (Fig. 3a), suggesting that PINK1 is capable of modulating HtrA2 phosphorylation. nature cell biology volume 9 | number 11 | NOVEMBER 2007
© 2007 Nature Publishing Group
Cdc2
Cdk5
Control +
–
+
–
+
4OH-Tx P-S142 HtrA2
35
P-S142 HtrA2
35
HtrA2
35
HtrA2
+
Mr(K) 35
CoCl2 P-S142 HtrA2
35
Mr(K)
PINK1 Y431H
e
C575R
–
–
35
IPD 4
+
+
HtrA2
IPD 3
–
–
35
IPD 2
+
+
Mr(K)
IPD 1
–
–
P-S142 HtrA2
Control 2
Control
d
4OH-Tx
35
Control 1
siRNA
HtrA2
35
c
Control
Mr(K) P-S142 HtrA2
+
S142A HtrA2 – +
PINK1
PINK1
–
WT HtrA2 – +
PINK1
Control
Vector
PINK1
Mr(K) 35
b
siRNA
MKK3–6
a
MKK3–6
A RT I C L E
35
P-S142 HtrA2
35
HtrA2
HtrA2
P-p38
35
p38 35 1
2
3
4
5
75 PINK1 50
6
Figure 3 PINK1 is necessary for HtrA2 phosphorylation. (a) HtrA2 phosphorylation on Ser 142 is decreased by downregulation of PINK1. Mitochondrial fractions from control or PINK1 siRNA (Ambion)-transfected SH-SY5Y cells were analysed. (b) MEKK3-induced phosphorylation of HtrA2 is decreased by downregulation of PINK1. ∆MEKK3:ER cells were transiently transfected with full-length wild-type (WT) or S142A HtrA2 and PINK1-specific siRNA (Ambion), stimulated with 4OH-Tx and analysed by western blotting. (c) Downregulation of Cdc2 and Cdk5 had minimal impact on MEKK3-induced phosphorylation of HtrA2 Ser 142. ∆MEKK3: ER cells were transiently transfected with processed HtrA2 and the indicated
siRNAs, stimulated with 4OH-Tx and analysed by western blotting. (d) CoCl2 induces phospho-Ser 142 HtrA2, which is decreased by downregulation of PINK1 (Dharmacon) and MKK3 and MKK6. HEK-293 cells were transfected with processed HtrA2 and the respective siRNA and subjected to western blotting. For quantification see Supplementary Information, Fig. S3d, e. (e) HtrA2 phosphorylation is decreased in Parkinson’s disease brains carrying mutant PINK1 (C575R and Y431H). Human brain tissue samples from the caudate region were analysed by western blotting. Idiopathic Parkinson’s disease brain (IPD); normal control brain (control). Uncropped images of blots in a, c, d and e are shown in Supplementary Information, Fig. S5.
Is PINK1 involved in the phosphorylation of HtrA2 induced by MEKK3? PINK1 knockdown resulted in a significant decrease in the levels of phospho-Ser 142 HtrA2 in ∆MEKK3:ER-expressing HEK-293 cells treated with 4OH-Tx (Fig. 3b and see Supplementary Information, Fig. S3c), further confirming the role of PINK1 in an MEKK3–p38 pathway leading to HtrA2 phosphorylation. Downregulation of MKK3 and MKK6 similarly resulted in a decrease in HtrA2 phosphorylation on activation of MEKK3 (Fig. 3c). By contrast, siRNA-mediated knockdown of cdk5 and cdc2 (see Supplementary Information, Fig. S3b) had minimal impact on MEKK3-induced phospho-Ser 142 HtrA2 (Fig. 3c). Another way to induce p38 phosphorylation is to stimulate the cells with CoCl2, a hypoxia-mimetic agent that increases levels of reactive oxygen species23. Strikingly, when phospho-p38 levels were increased in HEK-293 cells on treatment with CoCl2, phospho-Ser 142 HtrA2 levels were also increased. CoCl2-stimulated HtrA2 phosphorylation is decreased by MKK3, MKK6 and PINK1 siRNA (Fig. 3d; for quantification see Supplementary Information, Fig. S3d, e). The data so far reported suggest that PINK1 might be involved in a stress response at least in part by modulating the phosphorylation of HtrA2 by p38 on Ser 142. On the basis of the importance of these two proteins in the pathogenesis of nigral neurodegeneration we investigated the PINK1–HtrA2 pathway in human brain with Parkinson’s disease. We compared the level of phosphorylation of HtrA2 in control brain, brain with
idiopathic Parkinson’s disease, and brain with Parkinson’s disease associated with mutations in the PINK1 gene (Y431H and C575R). Western blot analysis performed with the HtrA2 phospho-specific antibody revealed a low level of phospho-Ser 142 HtrA2 in control brain tissue, with significantly higher levels in brains with idiopathic Parkinson’s disease. Strikingly, the phosphorylation of HtrA2 was virtually abolished in Parkinson’s disease brains with mutations in PINK1 (Fig. 3e). Simultaneous expression of Myc-tagged PINK1 Parkinson’s disease mutants with Flag-tagged HtrA2 followed by purification of the complexes indicated that neither C575R, Y431H nor any of seven other Parkinson’s disease-associated mutant forms of PINK1 had lost their ability to interact with HtrA2 (see Supplementary Information, Fig. S3f). It therefore seems likely that HtrA2 phosphorylation is decreased in brains of patients with Parkinson’s disease carrying mutations in PINK1 and that this requires an activity of PINK1 in addition to its physical interaction with HtrA2. Mutations mimicking phosphorylated HtrA2 enhance its protease activity To determine whether HtrA2 phosphorylation is likely to affect its proteolytic activity, we produced S142D and S400D phospho-mimetic HtrA2 mutants and compared their protease activity with that of the wild-type enzyme by using a previously described assay7. Both phospho-mimetic mutations increased the basal ability of HtrA2 to cleave
nature cell biology volume 9 | number 11 | NOVEMBER 2007 © 2007 Nature Publishing Group
1247
A RT I C L E a
* * *
3
* * *
2
* * *
1
Mr(K) 50
WT
b
S142A
S400D
S400A
S400A
S142D
S400D
WT
S142A
0
S142D
Protease activity (AFU min–1)
4
the peptide with the G399S Parkinson’s disease mutation was decreased (Fig. 4c), indicating that this mutation in HtrA2 might affect its protease activity, at least in part by affecting the phosphorylation of Ser 400.
* * *
56 nM HtrA2 209 nM HtrA2
35 30 Coomassie
c
G399S
A400
S400
A141S
A142
S142
p38γ
P
33
Figure 4 Phosphorylation of HtrA2 increases its proteolytic activity. (a) Protease activity of wild-type HtrA2 and the S142D and S400D phospho-mimetic mutants. Ser to Ala mutations at the phosphorylation sites show no significant effect compared with the wild-type (WT) protein. Results are shown as means ± SD; n = 4 (S400D), n = 6 (WT, S142D, S192A, S400A). Statistically significant values (one-way ANOVA with Bonferroni post-test, compared with wild-type HtrA2) three asterisks, P
