The fifth and the most well-conserved member of the TLR (Toll-like receptor) adaptor, SARM (sterile α- and HEAT/armadillo-motif-containing protein), has been reported to be an important mediator of apoptosis. However, the exact cellular localization of SARM with respect to its role is unclear. In the present study we show that SARM specifically co-localizes with mitochondria. Endogenous SARM is mainly found in the mitochondria. We demonstrate that the N-terminal 27 amino acids (S27) of SARM, which is hydrophobic and polybasic, acts as a mitochondria-targeting signal sequence, associating SARM to the mitochon-dria. The S27 peptide has an inherent ability to bind to lipids and mitochondria. This sequence effectively translocates the soluble EGFP (enhanced green fluorescence protein) reporter into the mitochondria. Positioning S27 downstream of the EGFP abrogates its mitochondria-targeting ability. Transmission electron microscopy confirms the ability of S27 to import EGFP into the mitochondria. Importantly, by mutagenesis study, we delineated the specificity of the mitochondria-targeting ability to the arginine residue at the 14th position. The R14A SARM mutant also showed reduced apoptotic potential when compared with the wild-type. Taken together, S27, which is a bona fide signal sequence that targets SARM to the mitochondria, explains the pro-apoptotic activity of SARM.

INTRODUCTION

TLRs (Toll-like receptors) are the prime sensors of pathogens. TLRs induce signalling cascades and activate transcription factors such as NF-κB (nuclear factor κB), AP-1 (activator protein 1) and IRF3 (interferon regulatory factor 3) [1]. TLRs act via five major adaptors: MyD88 (myeloid differentiation factor 88), MAL (MyD88 adaptor-like protein), TRIF [TIR (Toll/interleukin-1 receptor) domain-containing adaptor protein inducing interferon β], TRAM (TRIF-related adaptor molecule) and SARM (sterile α- and HEAT/armadillo-motif-containing protein). MAL and TRAM are bridging adaptors of MyD88 and TRIF respectively [25]. MyD88 and TRIF are respectively early and late activators of TLR signalling [68].

SARM (GenBank® accession number NP_055892) is an ancient TLR adaptor having homologues in Drosophila, zebrafish, nematode worm (Caenorhabditis elegans), amphioxus and horseshoe crab [912]. Importantly, phylogenetic analysis indicates that SARM is closely related to bacterial TIR domain-containing protein, highlighting the evolutionary conservation of this protein [13]. SARM is a multi-domain protein possessing two repeats each of ARM (HEAT/armadillo motif) and SAM (sterile α-motif), and a TIR domain, all of which are involved in protein–protein interactions [9] making SARM a unique TLR adaptor.

The C. elegans SARM homologue, TIR-1, protects the nematode worm from fungal and bacterial infections, independent of the TLR pathway [14,15]. SARM is regarded as a negative regulator of both MyD88- and TRIF-mediated TLR signalling [11,12,16,17]; however, macrophages from SARM−/− mice do not show up-regulation of TLR signalling [18,19]. Thus it appears that SARM plays a functional role that is different from the classical TLR signalling pathway. Interestingly, SARM was reported to recruit JNK3 (c-Jun N-terminal kinase 3) to the mitochondria and was observed to be involved in stress-induced neuronal toxicity [18]. In addition, C. elegans TIR-1 was reported to be a key adaptor in mediating anoxic death [20]. Furthermore, SARM has been highlighted as a key player in mediating apoptosis in T-cells (P. Panneerselvam, S. Ye, G.C.L. Wong, V. Selvarajan, W.J. Chng, S.B. Ng, B. Ho, J. Chen and J.L. Ding, unpublished work). Coincidentally, the SARM level is decreased in NK (natural killer)/T-cell lymphoma patient tissue when compared with the healthy cells, thus supporting its apoptotic phenotype ([22] and P. Panneerselvam, S. Ye, G.C.L. Wong, V. Selvarajan, W.J. Chng, S.B. Ng, B. Ho, J. Chen and J.L. Ding, unpublished work). It is therefore pertinent to delineate the exact subcellular localization of SARM with regards to its pro-apoptotic function. Kim et al. [18] highlighted that SARM lacking the mitochondria-targeting sequence is not localized inside the mitochondrial matrix. Furthermore, Peng et al. [17] showed that SARM is localized in the cytoplasm with a dot-like pattern in the nucleus, and claimed that it is anchored to the intracellular membrane. Hence the exact subcellular localization of SARM is still unclear, although its precise identification will offer insights into its functional role.

In the present study we show that SARM co-localizes with mitochondria. SARM is specifically present in the mitochondrial fraction of cells. We highlight an important signature sequence in SARM that is responsible for targeting it into the mitochondria. As a key cellular organelle for the regulation of apoptosis, the mitochondrion harbours various pro-apoptotic proteins such as AIF (apoptosis-inducing factor), Smac/DIABLO [direct IAP (inhibitor of apoptosis)-binding protein with low pI] and cytochrome c [23]. We have mapped the precise residues at the N-terminus of SARM that confer the mitochondrial association. Initially the sequence analysis indicated that the N-terminus of SARM possesses a hydrophobic polybasic domain and has an ability to fold into an α-helix. We then showed the ability of S27 synthetic peptide (the N-terminal 27 amino acids of SARM) to associate with lipids. Using fluorescence microscopy, we demonstrated that fusion of S27 upstream of the reporter, EGFP (enhanced green fluorescent protein), was sufficient to direct the EGFP into the mitochondria. The ability of S27 to act as a mitochondria-targeting leader sequence was further supported by TEM (transmission electron microscopy). Furthermore, using site-directed mutagenesis we showed that Arg14 in S27 is the mitochondria-targeting determinant. Importantly, R14A SARM mutant had reduced apoptogenic activity when compared with wild-type SARM. Taken together, our findings highlight that pro-apoptotic SARM localizes to the mitochondria and it contains a bona fide mitochondria-targeting signal sequence in the first 27 amino acids of SARM, and that Arg14 is the critical potent residue which confers this ability.

EXPERIMENTAL

Cell line and reagents

HEK-293T [human embryonic kidney cells expressing the large T-antigen of SV40 (simian virus 40)] and NIH 3T3 cell lines were maintained in DMEM (Dulbecco's modified Eagle's medium) containing 10% (v/v) fetal bovine serum, 100 units/ml penicillin and 100 μg/ml streptomycin. The mammalian expression constructs: SARM full-length, N ARM1, N ARM2, N124, ARM1 ARM2 and S27 were cloned into pEGFPN3 (Supplementary Table S1 at http://www.BiochemJ.org/bj/442/bj4420263add.htm). The S27 peptide corresponding to the first 27 amino acids at the N-terminus of SARM (MVLTLLLSAYKLCRFFAMSG PRPGAER), TMR (tetramethylrhodamine)-labelled S27 and negative control negS1 (GFELEGMAEISCLPNGQWSNFPPKCIRECAMVSS) [24] were synthesized and purified to >95% purity by Genemed Synthesis. Lipofectamine™ LTX, Mitotracker and far red-labelled transferrin were from Invitrogen.

Site-directed mutagenesis

The mutants Lys11Ala, Arg14Ala, Arg22Ala Arg27Ala and Arg14Ala SARM were made by site-directed mutagenesis according to the manufacturer's guidelines (QuikChange® XL site-directed mutagenesis kit, Stratagene). The oligonucleotides were designed using the web-based QuikChange® Primer design program (Supplementary Table S1; http://www.agilent.com/genomics/qcpd). The site-directed mutagenesis was confirmed by sequencing the mutated plasmids.

Transfection

HEK-293T cells were seeded on 12-well plates (Nunc) at a density of 2.5×105 cells/well in 1 ml of DMEM and grown overnight before transfection. Similarly, NIH 3T3 cells were seeded at a density of 8×104 cells/well in 1 ml of DMEM. Each transfection mixture, containing 800 ng of plasmid in 200 μl of incomplete medium and 2 μl of Lipofectamine™ LTX (Invitrogen), was incubated at room temperature (25°C) for 30 min. The transfection mixture was then added to the cells.

Fluorescence imaging

HEK-293T and NIH 3T3 cells were transfected for 12 h with the indicated truncated constructs of SARM. Following transfection, cells were incubated with 10 nM MitoTracker or 10 μg/ml labelled transferrin for 20 or 35 min respectively. The cells were washed and fixed in 4% formaldehyde for 15 min at room temperature, then washed three times with PBST (PBS containing 0.1% Tween 20). For immunostaining, the transfected cells were fixed and permeabilized. The cells were then stained with anti-(cytochrome c) antibody (1:200 dilution), followed by Alexa Fluor® 546-conjugated donkey anti-mouse antibody (1:1000 dilution). The cells were mounted on glass slides with mounting medium containing DAPI (4′,6-diamidino-2-phenylindole). The slides were viewed using a LSM META 510 confocal laser-scanning microscope with a ×100 oil-objective lens (Carl Zeiss).

Tyrosine/tryptophan fluorescence scanning spectroscopy

Either 250 μM S27 or 250 μM negS1 control peptide was mixed with increasing concentrations of a suspension of phospholipids, POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine)/POPG (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol) (2:1), in a 96-well plate. Samples were excited at either 274 nm (λex for peptide containing tyrosine) or 280 nm (λex for peptide containing tryptophan), and the fluorescence emission spectrum was either scanned from 300 to 500 nm or read at tyrosine λmax (303 nm) respectively, using a BioTek fluorescence reader.

Dot blotting

Equal amounts of POPE (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine) and POPG in 5 μl aliquots were spotted on to a PVDF membrane, air-dried and incubated with 50 μM TMR-labelled S27 peptide at room temperature in the dark for 2 h. After washing three times with 20 mM phosphate buffer (pH 7.0), the fluorescence was measured at λex 495 nm and λem 525 nm using a BioTek reader.

Mitochondria-binding assay

Mitochondria were isolated from HEK-293T cells by differential centrifugation (3000 g for 15 min at 4°C) using a mitochondria isolation kit and following the manufacturer's protocol (Pierce Biotechnology). Mitochondria isolated from 106 cells were resuspended in 100 μl of ice-cold PEB buffer [PBS (pH 7.2), 2 mM EDTA and 0.5% BSA) and the indicated amounts of TMR-labelled peptide were added. After 30 min of incubation in the dark at 4°C, the mitochondria were washed three times with ice-cold PEB buffer and then analysed using a LSRII flow cytometer. Flow Jo software was used for data analysis.

TEM

HEK-293T cells were transfected with S27–EGFP and at 24 h post-transfection, the cells were fixed with 4% paraformaldehyde and 1% glutaraldehyde. The cells were washed twice in distilled water and dehydrated in an ethanolic series. The cells were then gradually infiltrated with LR white resin and polymerized at 50°C for 2 days. Following ultramicrotomy, the sections were transferred on to a nickel grid and stained with rabbit anti-EGFP (1:100 dilution; Abcam) and mouse anti-(PDH E1α) (pyruvate dehydrogenase E1α) antibodies (1:20 dilution; Santa Cruz Biotechnology), followed by incubation with gold-conjugated anti-mouse and anti-rabbit secondary antibodies (Ted Pella). The grids were stained with osmium tetroxide, uranyl acetate and lead citrate. The samples were then observed using a JEM1010 transmission electron microscope (Jeol).

Western blotting

Mitochondria and cytosol fractions obtained from the transfected cells by differential centrifugation (3000 g for 15 min at 4°C) were separated by SDS/PAGE and transferred on to the PVDF membranes for immunodetection with rabbit anti-SARM (Santa Cruz Biotechnology) or anti-EGFP (Clontech) antibodies. The same membrane was stripped and reprobed for loading controls using rabbit anti-VDAC (voltage-dependent anion channel) antibody as a mitochondrial marker (Cell Signaling Technology) and mouse anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) antibody (Santa Cruz Biotechnology) as a cytoplasmic marker.

Apoptosis assay

HEK-293T cells were transfected with either wild-type SARM or the R14A SARM mutant. The transfected cells were collected at various time points post-transfection and stained with the dead cell stain 7-AAD. Cell death was quantified by flow cytometry. Flow Jo software was used for data analysis.

RESULTS

SARM co-localizes with the mitochondria

To precisely delineate the exact intracellular localization of SARM, we costained the HEK-293T cells expressing SARM fused to EGFP with MitoTracker or labelled transferrin (endosomal marker). The fluorescence images clearly indicate that SARM is specifically localized to the mitochondria and not to the endosomes (Figure 1A). To delineate the exact domain of SARM that is important for mitochondrial association, we made various truncated constructs of SARM fused to EGFP. In our analysis, we observed that only the constructs with the N-terminal 124 amino acids of SARM, the N ARM1, N ARM2 and N124 constructs (Figure 1B), showed association with the mitochondria. Consistently, the ARM1 ARM2 construct, lacking the N-terminal 124 amino acids, did not localize to the mitochondria (Figure 1B). The mitochondrial localization of SARM and N124 was further confirmed biochemically by isolating the mitochondrial fraction from untransfected and SARM- and N124-transfected HEK-293T cells. Endogenous and transiently expressed SARM were enriched in the mitochondrial fraction. The N124 protein was also specifically detected in the mitochondrial fraction of the transfected cells. Immunoblot analysis with mitochondria- and cytosol-specific markers revealed that the subcellular fractions were pure (Figure 1C). The mitochondrial association of SARM and its various truncates was further validated in another cell line, NIH 3T3 cells (Figure 1D). Additionally SARM and N124 specifically co-localized with the mitochondrial protein cytochrome c in NIH 3T3 cells (Figure 1E). Collectively, it is evident that SARM is specifically localized in the mitochondria. It is the N-terminal domain and not the ARM domain of SARM that is important for mitochondrial association. Hence it was imperative to delineate the exact sequence at the N-terminal domain of SARM that is important for mitochondrial association.

SARM co-localizes with the mitochondria

Figure 1
SARM co-localizes with the mitochondria

HEK-293T cells were transfected with full-length SARM–EGFP (FL) fusion construct. (A) At 12 h post-transfection, cells were co-stained with MitoTracker (mitochondria) or transferrin (endosomes). (B) At 12 h post-transfection, HEK-293T cells transfected with N ARM1–, N ARM2–, N124– and ARM1 ARM2–EGFP fusion constructs were co-stained with MitoTracker (mitochondria). The cells were fixed and then viewed by confocal microscopy. (C) Mitochondria (Mito) and cytosolic (Cyto) fractions were isolated from untransfected, SARM–EGFP (FL)-transfected and N124–EGFP-transfected HEK-293T cells. The mitochondria and cytosolic fractions were isolated by differential centrifugation and lysates were immunodetected with anti-SARM and anti-EGFP antibodies respectively. The same membrane was stripped and reprobed with anti-VDAC and anti-GAPDH antibodies that served as mitochondrial and cytosolic markers respectively. Molecular mass in kDa is given on the left-hand side of each panel. (D) At 12 h post-transfection, NIH 3T3 cells transfected with full-length SARM (FL)–, N ARM1–, N ARM2–, N124– and ARM1 ARM2–EGFP fusion constructs were costained with mitotracker. The cells were fixed and observed by confocal microscopy. (E) NIH 3T3 cells transfected with SARM (FL)– and N124–EGFP were fixed, permeabilized and stained with anti-(cytochrome c) antibody (mitochondrial marker). The cells were then viewed by confocal microscopy. Scale bars, 5 μm.

Figure 1
SARM co-localizes with the mitochondria

HEK-293T cells were transfected with full-length SARM–EGFP (FL) fusion construct. (A) At 12 h post-transfection, cells were co-stained with MitoTracker (mitochondria) or transferrin (endosomes). (B) At 12 h post-transfection, HEK-293T cells transfected with N ARM1–, N ARM2–, N124– and ARM1 ARM2–EGFP fusion constructs were co-stained with MitoTracker (mitochondria). The cells were fixed and then viewed by confocal microscopy. (C) Mitochondria (Mito) and cytosolic (Cyto) fractions were isolated from untransfected, SARM–EGFP (FL)-transfected and N124–EGFP-transfected HEK-293T cells. The mitochondria and cytosolic fractions were isolated by differential centrifugation and lysates were immunodetected with anti-SARM and anti-EGFP antibodies respectively. The same membrane was stripped and reprobed with anti-VDAC and anti-GAPDH antibodies that served as mitochondrial and cytosolic markers respectively. Molecular mass in kDa is given on the left-hand side of each panel. (D) At 12 h post-transfection, NIH 3T3 cells transfected with full-length SARM (FL)–, N ARM1–, N ARM2–, N124– and ARM1 ARM2–EGFP fusion constructs were costained with mitotracker. The cells were fixed and observed by confocal microscopy. (E) NIH 3T3 cells transfected with SARM (FL)– and N124–EGFP were fixed, permeabilized and stained with anti-(cytochrome c) antibody (mitochondrial marker). The cells were then viewed by confocal microscopy. Scale bars, 5 μm.

The N-terminus of SARM harbours structural features for mitochondrial targeting

We then investigated whether the N-terminus of SARM could act as a mitochondria-targeting signal sequence. Interestingly, the N-terminus of SARM does not bear any sequence similarity with other reported mitochondria-targeting signal sequence, which could explain why Kim et al. [18] reported a lack of a mitochondria signal sequence in SARM. We then examined whether the N-terminal sequence of SARM bears structural features similar to that of mitochondrial signal sequence. First, using Kyte–Doolittle hydropathy plots [25], we found that the N-terminus of SARM harbours a significant hydrophobic patch (Figure 2A). Secondly, we observed the presence of a conserved polybasic domain [17] (Figure 2B). Hence the first 27 amino acids of SARM (S27) possess hydrophobic (leucine and alanine), positively charged (lysine and arginine) and hydroxylated (serine and threonine) amino acids (Figure 2C), which appear to bear a resemblance to other mitochondria-targeting leader sequences [26]. Additionally PSIPRED [27] indicated that S27 (amino acids 2–16 of SARM) has the ability to fold into an α-helix (Figure 2D). However, the α-helix is not amphiphilic as indicated by a helical wheel analysis (http://rzlab.ucr.edu/scripts/wheel/wheel.cgi) (Figure 2E). Thus, except for the lack of a typical α-helix motif with amphiphilicity, it appears evident that the first 27 amino acids of SARM resemble a mitochondria-targeting signal sequence. However, amphiphilicity has been considered to not be essential in some mitochondria-targeting sequences [28,29]. Guided by these observations and reasoning, we hypothesized that the first 27 amino acids of SARM might be important for its association with the mitochondria. Coincidentally, we noted that the first 27 amino acids of SARM are well conserved across various SARM homologues (Figure 2F), suggesting a potential functional significance.

The N-terminus of SARM harbours structural features for mitochondrial targeting

Figure 2
The N-terminus of SARM harbours structural features for mitochondrial targeting

(A) The hydropathy plot was drawn for the first 30 amino acids of SARM based on the hydrophobicity score obtained from ProtScale online server (http://web.expasy.org/protscale) using the Kyte–Doolittle amino acid scale value. (B) Isoelectric points of sequential 30-amino-acid segments of various SARM homologues highlight the conserved polybasic domain at the N-terminus. The isoelectric point was computed using an online ExPASy server (http://web.expasy.org/compute_pi/). (C) N-terminal 27 amino acids of SARM highlighting the hydrophobic (underlined), basic (*) and hydroxy (°) side chain residues. (D) The secondary structure of SARM is predicted using PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/). The Figure depicts the secondary structure of the first 80 amino acids of SARM, indicating that the N-terminus of SARM (amino acids 2–16) is able to form an α-helix. AA, amino acid; Conf, confidence; Pred, prediction. (E) Helical wheel projection of the N-terminus of SARM was plotted using software (http://cti.itc.virginia.edu/~cmg/Demo/wheel/wheelApp.html). The non-polar, polar uncharged and basic residues are indicated as orange, green and blue respectively. (F) The N-terminal 27 amino acids of various SARM homologues were aligned using ClustalX program. Symbols: *, 100% conservation of the same single amino acid;:, conservation of the same single amino acid between groups of strongly similar properties;., conservation of the same single amino acid between groups of weakly similar properties.

Figure 2
The N-terminus of SARM harbours structural features for mitochondrial targeting

(A) The hydropathy plot was drawn for the first 30 amino acids of SARM based on the hydrophobicity score obtained from ProtScale online server (http://web.expasy.org/protscale) using the Kyte–Doolittle amino acid scale value. (B) Isoelectric points of sequential 30-amino-acid segments of various SARM homologues highlight the conserved polybasic domain at the N-terminus. The isoelectric point was computed using an online ExPASy server (http://web.expasy.org/compute_pi/). (C) N-terminal 27 amino acids of SARM highlighting the hydrophobic (underlined), basic (*) and hydroxy (°) side chain residues. (D) The secondary structure of SARM is predicted using PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/). The Figure depicts the secondary structure of the first 80 amino acids of SARM, indicating that the N-terminus of SARM (amino acids 2–16) is able to form an α-helix. AA, amino acid; Conf, confidence; Pred, prediction. (E) Helical wheel projection of the N-terminus of SARM was plotted using software (http://cti.itc.virginia.edu/~cmg/Demo/wheel/wheelApp.html). The non-polar, polar uncharged and basic residues are indicated as orange, green and blue respectively. (F) The N-terminal 27 amino acids of various SARM homologues were aligned using ClustalX program. Symbols: *, 100% conservation of the same single amino acid;:, conservation of the same single amino acid between groups of strongly similar properties;., conservation of the same single amino acid between groups of weakly similar properties.

S27 peptide directly binds to lipids and mitochondria

SVMPROT (http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi) predicted SARM to be a lipid-binding protein. Hence we reasoned that the direct lipid-binding ability of S27 might aid in the mitochondrial import of SARM. We therefore assessed the ability of the leading 27 amino acids of SARM to associate with lipids. To achieve this, we chemically synthesized the S27 peptide. The S27 peptide possesses one tyrosine and two phenylalanine residues (Figure 3A). The presence of tyrosine conferred the fluorescence property on the S27 peptide, which made it possible to measure its interaction with lipids. Interestingly, in the presence of POPC/POPG (at a ratio of 2:1), the inherent fluorescence emission of the S27 peptide gradually increased, indicating potential interaction between S27 and lipid (Figures 3B and 3C). The control peptide, negS1, which possesses a tryptophan residue (measured at λex 280 nm) did not show any significant change in the inherent fluorescence emission in the presence of POPC/POPG (Supplementary Figure S1 at http://www.BiochemJ.org/bj/442/bj4420263add.htm). Furthermore, the TMR-labelled S27 peptide showed preferential binding to POPG compared with POPE (Figure 3D). It is noteworthy that POPG is predominantly found in the bacterial cell wall, whereas POPE is present in the mammalian cell membrane [30]. As the mitochondria are evolutionary endosymbionts of bacterial origin, POPG is present in the mitochondrial membrane lipids [31,32]. Therefore the stronger association of S27 to POPG corroborates the preferential anchorage of S27 on the mitochondrial membrane lipid. To substantiate our finding that S27 peptide associates with the mitochondrial membrane, we further isolated mitochondria from HEK-293T cells by differential centrifugation. When TMR-labelled S27 was incubated with the isolated mitochondria, we observed a dose-dependent increase in the binding of the S27 peptide to the cell-free mitochondria (Figures 3E and 3F). Taken together, the results of the present study showed that the N-terminal 27-amino-acid leader sequence of SARM could potentially interact with lipids and, in particular, with the mitochondrial membrane lipids.

S27 peptide associates with lipids and mitochondria

Figure 3
S27 peptide associates with lipids and mitochondria

(A) Amino acid sequence of S27 highlighting the residues which contribute to the intrinsic fluorescence measurement of the peptide. (B) The tyrosine fluorescence emission spectrum of the S27 peptide with increasing concentrations of POPC/POPG (0, 50, 100, 250 and 500 μM). (C) The tyrosine fluorescence emission of S27 peptide at 303 nm was measured with increasing concentrations of POPC/POPG (0, 50, 100 and 250 μM). Results are means±S.D. of triplicate measurements. (D) Solid-phase interaction of S27 with POPE and POPG. Equal amounts of POPE and POPG were dot-blotted on to PVDF membranes and incubated with TMR-labelled S27. The fluorescence emission was measured at λex 495 nm and λem 525 nm. Mitochondria were isolated from HEK-293T cells by differential centrifugation following the manufacturer's protocol and incubated in the dark with indicated amounts of the labelled peptide. The labelled mitochondria were then washed and analysed by flow cytometry. (E) Representative flow cytometry data depicting the overlay of mitochondrial fluorescence of unlabelled (grey) and peptide-incubated (black) mitochondria. (F) Histogram indicating the proportion of mitochondria labelled with TMR-labelled S27 with increasing amounts of peptide. Results are means±S.D. of triplicate measurements.

Figure 3
S27 peptide associates with lipids and mitochondria

(A) Amino acid sequence of S27 highlighting the residues which contribute to the intrinsic fluorescence measurement of the peptide. (B) The tyrosine fluorescence emission spectrum of the S27 peptide with increasing concentrations of POPC/POPG (0, 50, 100, 250 and 500 μM). (C) The tyrosine fluorescence emission of S27 peptide at 303 nm was measured with increasing concentrations of POPC/POPG (0, 50, 100 and 250 μM). Results are means±S.D. of triplicate measurements. (D) Solid-phase interaction of S27 with POPE and POPG. Equal amounts of POPE and POPG were dot-blotted on to PVDF membranes and incubated with TMR-labelled S27. The fluorescence emission was measured at λex 495 nm and λem 525 nm. Mitochondria were isolated from HEK-293T cells by differential centrifugation following the manufacturer's protocol and incubated in the dark with indicated amounts of the labelled peptide. The labelled mitochondria were then washed and analysed by flow cytometry. (E) Representative flow cytometry data depicting the overlay of mitochondrial fluorescence of unlabelled (grey) and peptide-incubated (black) mitochondria. (F) Histogram indicating the proportion of mitochondria labelled with TMR-labelled S27 with increasing amounts of peptide. Results are means±S.D. of triplicate measurements.

S27 is sufficient to target EGFP to the mitochondria

The ability of S27 to act as a mitochondria-targeting signal sequence was then directly visualized in a live cell. We prepared two different fluorescent fusion constructs of S27, one with C-terminal EGFP fusion (S27–EGFP) and the other with N-terminal EGFP fusion (EGFP–S27). The transfected HEK-293T cells were co-stained with MitoTracker and endosomal marker. Confocal microscopy revealed that S27–EGFP was co-localized with the mitochondria, but not with the endosomes (Figure 4A). Thus S27 specifically directs the EGFP to the mitochondria. In contrast, EGFP–S27 did not associate with the mitochondria; instead it was distributed throughout the cell (Figure 4B). Consistently S27–EGFP was specifically detected in the mitochondrial fraction of the transfected cells (Figure 4C). S27–EGFP also co-localized with mitochondria and cytochrome c in NIH 3T3 cells (Figures 4D and 4E). Thus it appears that S27 is a specific mitochondria-targeting signal sequence when positioned as a leader sequence.

S27 is sufficient to target EGFP to the mitochondria

Figure 4
S27 is sufficient to target EGFP to the mitochondria

HEK-293T cells were transfected either with S27–EGFP or EGFP–S27 fusion constructs. (A) At 12 h post-transfection, S27–EGFP-transfected cells were co-stained with either MitoTracker (mitochondria) or labelled transferrin (endosomes). The cells were fixed and then viewed by confocal microscopy. (B) Similarly, EGFP–S27-transfected cells were labelled with MitoTracker and viewed by confocal microscopy. (C) Mitochondria (Mito) and cytosolic fractions (Cyto) were isolated from S27–EGFP-transfected HEK-293T cells by differential centrifugation and lysates were Western blotted with anti-EGFP antibody. The same membrane was stripped and reprobed with anti-VDAC and anti-GAPDH antibodies, which served as mitochondrial and cytosolic markers respectively. (D) At 12 h post-transfection, NIH 3T3 cells transfected with S27–EGFP fusion constructs were co-stained with MitoTracker (mitochondria). The cells were fixed and observed by confocal microscopy. (F) NIH 3T3 cells transfected with S27–EGFP were fixed, permeabilized and stained with anti-(cytochrome c) antibody (mitochondrial marker). The cells were then viewed by confocal microscopy. Scale bars, 5 μm.

Figure 4
S27 is sufficient to target EGFP to the mitochondria

HEK-293T cells were transfected either with S27–EGFP or EGFP–S27 fusion constructs. (A) At 12 h post-transfection, S27–EGFP-transfected cells were co-stained with either MitoTracker (mitochondria) or labelled transferrin (endosomes). The cells were fixed and then viewed by confocal microscopy. (B) Similarly, EGFP–S27-transfected cells were labelled with MitoTracker and viewed by confocal microscopy. (C) Mitochondria (Mito) and cytosolic fractions (Cyto) were isolated from S27–EGFP-transfected HEK-293T cells by differential centrifugation and lysates were Western blotted with anti-EGFP antibody. The same membrane was stripped and reprobed with anti-VDAC and anti-GAPDH antibodies, which served as mitochondrial and cytosolic markers respectively. (D) At 12 h post-transfection, NIH 3T3 cells transfected with S27–EGFP fusion constructs were co-stained with MitoTracker (mitochondria). The cells were fixed and observed by confocal microscopy. (F) NIH 3T3 cells transfected with S27–EGFP were fixed, permeabilized and stained with anti-(cytochrome c) antibody (mitochondrial marker). The cells were then viewed by confocal microscopy. Scale bars, 5 μm.

TEM confirms S27 as a unique mitochondria-targeting signal sequence

We further performed TEM to confirm the specificity of S27 as a mitochondria-targeting signal sequence. HEK-293T cells transfected with S27–EGFP was used for the TEM studies. Immunogold labelling against the EGFP tag showed that EGFP was specifically localized inside the mitochondria. The untransfected negative control did not show any gold particles in the mitochondria (Figures 5A and 5B). To further confirm the specific localization of S27–EGFP in the mitochondria, we performed double labelling of S27–EGFP with mitochondrial matrix protein PDH E1α. The TEM images showed the co-localization of S27–EGFP with the matrix protein PDH E1α (Figure 5C). Hence it appears that S27 specifically targets and translocates the EGFP into the mitochondrial matrix.

TEM confirms that S27 is a unique mitochondria-targeting signal sequence

Figure 5
TEM confirms that S27 is a unique mitochondria-targeting signal sequence

HEK-293T cells transfected with S27–EGFP fusion construct were embedded in LR white resin. TEM of untransfected (A) or S27–EGFP-expressing (B) HEK-293T cells. Ultrathin sections were labelled with rabbit anti-EGFP antibody followed by labelling with 10 nm gold-conjugated anti-rabbit secondary antibody. Scale bar, 0.2 μm. (C) Ultrathin gold sections of embedded S27–EGFP-transfected HEK-293T cells were stained with mouse anti-(PDH E1α) and rabbit anti-EGFP antibodies. The sections were then stained with 10 nm gold-conjugated anti-mouse and 20 nm gold-conjugated anti-rabbit secondary antibodies. Scale bar, 100 nm. Following immunogold labelling, the sections were stained with osmium chloride, lead citrate and uranyl acetate. The grids containing the sections were then observed by TEM. Regions of the cell that are shown at higher magnification are marked with boxes. N, nucleus.

Figure 5
TEM confirms that S27 is a unique mitochondria-targeting signal sequence

HEK-293T cells transfected with S27–EGFP fusion construct were embedded in LR white resin. TEM of untransfected (A) or S27–EGFP-expressing (B) HEK-293T cells. Ultrathin sections were labelled with rabbit anti-EGFP antibody followed by labelling with 10 nm gold-conjugated anti-rabbit secondary antibody. Scale bar, 0.2 μm. (C) Ultrathin gold sections of embedded S27–EGFP-transfected HEK-293T cells were stained with mouse anti-(PDH E1α) and rabbit anti-EGFP antibodies. The sections were then stained with 10 nm gold-conjugated anti-mouse and 20 nm gold-conjugated anti-rabbit secondary antibodies. Scale bar, 100 nm. Following immunogold labelling, the sections were stained with osmium chloride, lead citrate and uranyl acetate. The grids containing the sections were then observed by TEM. Regions of the cell that are shown at higher magnification are marked with boxes. N, nucleus.

Arg14 is critical for mitochondrial targeting

We then delineated the key residues in S27 that might be determinants for the mitochondrial targeting. Guided by the hypothesis that basic residues might be important for mitochondrial import, we mutated Lys11, Arg14, Arg22 and Arg27 to alanine (Figure 6A). The HEK-293T cells transfected with these mutant constructs (Lys11Ala, Arg14Ala, Arg22Ala and Arg27Ala) were stained with MitoTracker to observe potential co-localization. All of the mutants except Arg14Ala retained their ability to localize to the mitochondria (Figure 6B). Interestingly mutation of only Arg14 abolished the mitochondrial association of S27, suggesting the importance of this single residue for translocating the protein into the mitochondria (Figure 6B). The critical role of Arg14 in determining the mitochondrial localization was further confirmed by transfecting each of the four mutant constructs into NIH 3T3 cells. Consistently, the Arg14Ala mutant failed to localize to the mitochondria (Figure 6C).

Arg14 is critical for mitochondrial localization

Figure 6
Arg14 is critical for mitochondrial localization

(A) Amino acid sequence of S27 highlighting the basic residues that are mutated. (B) HEK-293T cells grown on glass cover slips were transfected with the S27–EGFP mutant plasmids Arg14Ala, Lys11Arg, Arg22Ala and Arg27Ala. At 12 h post-transfection, cells were co-stained with MitoTracker (mitochondria). The cells are fixed and then viewed by confocal microscopy. (C) At 12 h post-transfection, NIH 3T3 cells transfected with the S27–EGFP mutant constructs Arg14Ala, Lys11Arg, Arg22Ala and Arg27Ala were co-stained with MitoTracker (mitochondria). The cells were fixed and observed by confocal microscopy. Except for the Arg14Ala mutant, all other mutants (Lys11Ala, Arg22Ala and Arg27Ala) associate with mitochondria in both HEK-293T and NIH 3T3 cells. Scale bar, 5 μm.

Figure 6
Arg14 is critical for mitochondrial localization

(A) Amino acid sequence of S27 highlighting the basic residues that are mutated. (B) HEK-293T cells grown on glass cover slips were transfected with the S27–EGFP mutant plasmids Arg14Ala, Lys11Arg, Arg22Ala and Arg27Ala. At 12 h post-transfection, cells were co-stained with MitoTracker (mitochondria). The cells are fixed and then viewed by confocal microscopy. (C) At 12 h post-transfection, NIH 3T3 cells transfected with the S27–EGFP mutant constructs Arg14Ala, Lys11Arg, Arg22Ala and Arg27Ala were co-stained with MitoTracker (mitochondria). The cells were fixed and observed by confocal microscopy. Except for the Arg14Ala mutant, all other mutants (Lys11Ala, Arg22Ala and Arg27Ala) associate with mitochondria in both HEK-293T and NIH 3T3 cells. Scale bar, 5 μm.

Arg14Ala SARM mutant has reduced apoptotic activity

Having shown that Arg14 is critical for mitochondrial localization, we checked whether mutation of Arg14 affects the pro-apoptotic potential of SARM. HEK-293T cells transfected with wild-type or Arg14Ala mutant SARM were stained with the dead cell stain 7-AAD. Cell death was quantified by flow cytometry. Interestingly, we observed that the Arg14Ala SARM mutant lost 50% of its apoptotic potential compared with the wild-type SARM (Figure 7). These data suggest that mitochondrial localization of SARM is important for the pro-apoptotic phenotype of SARM.

Arg14Ala SARM mutant has reduced apoptotic activity

Figure 7
Arg14Ala SARM mutant has reduced apoptotic activity

HEK-293T cells in 12-well plates were transfected with either wild-type SARM or Arg14Ala SARM mutant constructs. At 12 and 18 h post-transfection, the cells were stained with dead cell stain 7-AAD, and cell death was quantified by flow cytometry. Results are means±S.D. of triplicate measurements. *P<0.005.

Figure 7
Arg14Ala SARM mutant has reduced apoptotic activity

HEK-293T cells in 12-well plates were transfected with either wild-type SARM or Arg14Ala SARM mutant constructs. At 12 and 18 h post-transfection, the cells were stained with dead cell stain 7-AAD, and cell death was quantified by flow cytometry. Results are means±S.D. of triplicate measurements. *P<0.005.

DISCUSSION

Earlier studies had reported that SARM associates to the outer membrane of the mitochondria, and, owing to the lack of observation of a typical mitochondria-targeting signal, the ARM domain of SARM was purportedly important for this association [12,18]. Moreover, transient expression of SARM GFP (green fluorescent protein) appeared as dots in the nucleus and elsewhere in the cell [17]. Therefore the precise localization of SARM in relation to its cellular function has hitherto been unclear. In the present study we demonstrated the mitochondria-targeting ability of the most conserved TLR adaptor, SARM, by immunofluorescence and Western blotting (Figure 1). Furthermore, we delineated the mitochondria-targeting ability of SARM to its N-terminal signal sequence and showed that the N-terminal sequences are well conserved across various SARM homologues (Figure 2). Importantly, the presence of a mitochondria-targeting signal sequence in SARM reflects its biological significance and helps to explain how SARM might mediate apoptosis in neurons and T-cells, and in C. elegans ([18,20] and P. Panneerselvam, S. Ye, G.C.L. Wong, V. Selvarajan, W.J. Chng, S.B. Ng, N.S. Tan, B. Ho, J. Chen and J.L. Ding, unpublished work).

There is no sequence similarity between S27 and other known mitochondria-targeting signal sequences. However, S27 possesses structural features of hydrophobicity, positively charged residues and α-helicity, which are important for targeting the mitochondria (Figure 2 and Supplementary Figure S2 at http://www.BiochemJ.org/bj/442/bj4420263add.htm). Nevertheless, it is important to note that the α-helix in S27 lacks amphiphilicity, which is purportedly critical for binding to the mitochondria import machinery (Supplementary Figure S2) [33]. However, our results suggest that the hydrophobic and the conserved polybasic residues serve as a bipartite signal and are sufficient to target the mitochondria. We further show evidence that the S27 peptide has an inherent ability to interact with lipids (Figure 3), which might facilitate its import into the mitochondria. In addition, TEM indicated that S27 is able to translocate EGFP into the mitochondria. Further confirmation was obtained by double immunogold labelling of S27–EGFP with mitochondrial matrix-targeting protein PDH E1α (Figure 5). Interestingly, the mitochondria-targeting ability of S27 was abolished when S27 was cloned downstream of the EGFP reporter, indicating that S27 is only functional as a mitochondrial target when it is at the leading end of the protein (Figure 4). The mitochondria-targeting ability of SARM and S27 was further supported by Western blot analysis of mitochondrial fraction isolated from the transfected cells (Figures 1 and 4).

Since the positively charged amino acids in a signal sequence are earlier documented to be important for the transport across the membrane via the TOM (translocase of the mitochondrial outer membrane)/TIM (translocase of the mitochondrial inner membrane) import machinery [3437], we mutated all four positively charged amino acids to uncharged alanine. Surprisingly, only the Arg14Ala mutant effectively abolished the mitochondria-targeting ability of S27 (Figure 6). The ability of Arg14 in determining the mitochondrial localization was confirmed using another cell line, NIH 3T3 (Figure 6). Coincidentally, Arg23 in the mitochondria signal sequence of ornithine transcarbamylase was also reported to be very critical for import [38]. In contrast, the mitochondria-targeting ability of cytochrome P450 2E1 and aldehyde dehydrogenase was abrogated only when at least two or all of the arginine residues at the leader sequences were mutated simultaneously [34,39]. Thus arginine appears to be the common critical determinant for the interaction with either the mitochondria import machinery or directly with the negatively charged head groups of the mitochondrial phospholipid bilayer.

Importantly, mutating Arg14 to alanine caused SARM to lose 50% of its apoptotic potential (Figure 7). This clearly shows that mitochondrial localization of SARM is important for its pro-apoptotic activity. However, the pro-apoptotic potential of the Arg14Ala mutant was not completely abolished, suggesting the contribution of other residues and/or existence of other factors that are important for the full pro-apoptotic potential of SARM.

Currently drugs have been aimed at targeting antioxidants or membrane-disrupting agents to the mitochondria as a therapy for various oxidative-stress-related diseases and tumours [4044]. The knowledge gained in the present study can be used to design therapeutics specifically targeting the mitochondria. Taken together, the results of the present study have shown the mitochondria-targeting ability of SARM, a unique TLR adaptor, and we have further isolated the crucial sequence and identified the single residue Arg14, which determines the mitochondria-targeting specificity. Concomitant to the mutation of Arg14 was a substantial reduction in the apoptotic potential of SARM. Thus the mitochondria-targeting characteristic of SARM reflects its biological importance in mediating apoptosis in neurons and T-cells, and in C. elegans. In the future, it will be interesting to examine whether this signal sequence follows the classical mitochondrial import mechanism and also to identify the molecular target on the mitochondria with which the SARM docks/interacts.

Abbreviations

     
  • ARM

    HEAT/armadillo motif

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • EGFP

    enhanced green fluorescent protein

  •  
  • GAPDH

    glyceraldehyde-3-phosphate dehydrogenase

  •  
  • HEK-293T

    human embryonic kidney cells expressing the large T-antigen of SV40 (simian virus 40)

  •  
  • MyD88

    myeloid differentiation factor 88

  •  
  • MAL

    MyD88 adaptor-like protein

  •  
  • PDH

    E1α, pyruvate dehydrogenase E1α subunit

  •  
  • POPC

    1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

  •  
  • POPE

    1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine

  •  
  • POPG

    1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol

  •  
  • SARM

    sterile α- and HEAT/armadillo-motif-containing protein

  •  
  • TEM

    transmission electron microscopy

  •  
  • TIR

    Toll/interleukin-1 receptor domain, TLR, Toll-like receptor

  •  
  • TMR

    tetramethylrhodamine

  •  
  • TRIF

    TIR domain-containing adaptor protein inducing interferon β

  •  
  • TRAM

    TRIF-related adaptor molecule

  •  
  • VDAC

    voltage-dependent anion channel

AUTHOR CONTRIBUTION

Porkodi Panneerselvam and Jeak Ling Ding conceived the study and designed the experiments. Porkodi Panneerselvam and Laishram Pradeepkumar Singh performed the experiments and analysed the data. Bow Ho and Jianzhu Chen provided some suggestions. Porkodi Panneerselvam and Jeak Ling Ding wrote the paper.

We thank Associate Professor Thorsten Wohland, Mr Jagadish Sankaran and Mr Nirmalya Bag for phospholipids; Ms Sun Ye for assisting in the lipid-binding assays and cloning; Mdm Josephine Howe for help with the TEM. Porkodi Panneerselvam is a graduate scholar of the Singapore-Massachusetts Institute of Technology Alliance. We thank SMART ID-IRG (Singapore-Massachusetts Institute of Technology Alliance for Research and Technology Interdisciplinary Research Group in Infectious Disease) for the usage of facilities.

FUNDING

This work was financed by SMA-CSB (Singapore-Massachusetts Institute of Technology Alliance Computation and Systems Biology) and A*STAR BMRC (Agency for Science, Technology and Research Biomedical Research Council) [grant number 08/1/21/19/574].

References

References
1
Imler
J. L.
Hoffmann
J. A.
Toll and Toll-like proteins: an ancient family of receptors signaling infection
Rev. Immunogenet.
2000
, vol. 
2
 (pg. 
294
-
304
)
2
Kagan
J. C.
Medzhitov
R.
Phosphoinositide-mediated adaptor recruitment controls Toll-like receptor signaling
Cell
2006
, vol. 
125
 (pg. 
943
-
955
)
3
Fitzgerald
K. A.
Palsson-McDermott
E. M.
Bowie
A. G.
Jefferies
C. A.
Mansell
A. S.
Brady
G.
Brint
E.
Dunne
A.
Gray
P.
Harte
M. T.
, et al. 
Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction
Nature
2001
, vol. 
413
 (pg. 
78
-
83
)
4
Fitzgerald
K. A.
Rowe
D. C.
Barnes
B. J.
Caffrey
D. R.
Visintin
A.
Latz
E.
Monks
B.
Pitha
P. M.
Golenbock
D. T.
LPS–TLR4 signaling to IRF-3/7 and NF-κB involves the toll adapters TRAM and TRIF
J. Exp. Med.
2003
, vol. 
198
 (pg. 
1043
-
1055
)
5
Yamamoto
M.
Sato
S.
Hemmi
H.
Uematsu
S.
Hoshino
K.
Kaisho
T.
Takeuchi
O.
Takeda
K.
Akira
S.
TRAM is specifically involved in the Toll-like receptor 4-mediated MyD88-independent signaling pathway
Nat. Immunol.
2003
, vol. 
4
 (pg. 
1144
-
1150
)
6
Kawai
T.
Adachi
O.
Ogawa
T.
Takeda
K.
Akira
S.
Unresponsiveness of MyD88-deficient mice to endotoxin
Immunity
1999
, vol. 
11
 (pg. 
115
-
122
)
7
Hoebe
K.
Du
X.
Georgel
P.
Janssen
E.
Tabeta
K.
Kim
S. O.
Goode
J.
Lin
P.
Mann
N.
Mudd
S.
, et al. 
Identification of Lps2 as a key transducer of MyD88-independent TIR signalling
Nature
2003
, vol. 
424
 (pg. 
743
-
748
)
8
Tanimura
N.
Saitoh
S.
Matsumoto
F.
Akashi-Takamura
S.
Miyake
K.
Roles for LPS-dependent interaction and relocation of TLR4 and TRAM in TRIF-signaling
Biochem. Biophys. Res. Commun.
2008
, vol. 
368
 (pg. 
94
-
99
)
9
Mink
M.
Fogelgren
B.
Olszewski
K.
Maroy
P.
Csiszar
K.
A novel human gene (SARM) at chromosome 17q11 encodes a protein with a SAM motif and structural similarity to Armadillo/β-catenin that is conserved in mouse, Drosophila, and Caenorhabditis elegans
Genomics
2001
, vol. 
74
 (pg. 
234
-
244
)
10
Meijer
A. H.
Gabby Krens
S. F.
Medina Rodriguez
I. A.
He
S.
Bitter
W.
Ewa Snaar-Jagalska
B.
Spaink
H. P.
Expression analysis of the Toll-like receptor and TIR domain adaptor families of zebrafish
Mol. Immunol.
2004
, vol. 
40
 (pg. 
773
-
783
)
11
Belinda
L. W.
Wei
W. X.
Hanh
B. T.
Lei
L. X.
Bow
H.
Ling
D. J.
SARM: a novel Toll-like receptor adaptor, is functionally conserved from arthropod to human
Mol. Immunol.
2008
, vol. 
45
 (pg. 
1732
-
1742
)
12
Yuan
S.
Wu
K.
Yang
M.
Xu
L.
Huang
L.
Liu
H.
Tao
X.
Huang
S.
Xu
A.
Amphioxus SARM involved in neural development may function as a suppressor of TLR signaling
J. Immunol.
2010
, vol. 
184
 (pg. 
6874
-
6881
)
13
Zhang
Q.
Zmasek
C. M.
Cai
X.
Godzik
A.
TIR domain-containing adaptor SARM is a late addition to the ongoing microbe–host dialog
Dev. Comp. Immunol.
2011
, vol. 
35
 (pg. 
461
-
468
)
14
Couillault
C.
Pujol
N.
Reboul
J.
Sabatier
L.
Guichou
J. F.
Kohara
Y.
Ewbank
J. J.
TLR-independent control of innate immunity in Caenorhabditis elegans by the TIR domain adaptor protein TIR-1, an ortholog of human SARM
Nat. Immunol.
2004
, vol. 
5
 (pg. 
488
-
494
)
15
Liberati
N. T.
Fitzgerald
K. A.
Kim
D. H.
Feinbaum
R.
Golenbock
D. T.
Ausubel
F. M.
Requirement for a conserved Toll/interleukin-1
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
6593
-
6598
)
16
Carty
M.
Goodbody
R.
Schroder
M.
Stack
J.
Moynagh
P. N.
Bowie
A. G.
The human adaptor SARM negatively regulates adaptor protein TRIF-dependent Toll-like receptor signaling
Nat. Immunol.
2006
, vol. 
7
 (pg. 
1074
-
1081
)
17
Peng
J.
Yuan
Q.
Lin
B.
Panneerselvam
P.
Wang
X.
Luan
X. L.
Lim
S. K.
Leung
B. P.
Ho
B.
Ding
J. L.
SARM inhibits both TRIF- and MyD88-mediated AP-1 activation
Eur. J. Immunol.
2010
, vol. 
40
 (pg. 
1738
-
1747
)
18
Kim
Y.
Zhou
P.
Qian
L.
Chuang
J. Z.
Lee
J.
Li
C.
Iadecola
C.
Nathan
C.
Ding
A.
MyD88-5 links mitochondria, microtubules, and JNK3 in neurons and regulates neuronal survival
J. Exp. Med.
2007
, vol. 
204
 (pg. 
2063
-
2074
)
19
Szretter
K. J.
Samuel
M. A.
Gilfillan
S.
Fuchs
A.
Colonna
M.
Diamond
M. S.
The immune adaptor molecule SARM modulates tumor necrosis factor α production and microglia activation in the brainstem and restricts West Nile Virus pathogenesis
J. Virol.
2009
, vol. 
83
 (pg. 
9329
-
9338
)
20
Hayakawa
T.
Kato
K.
Hayakawa
R.
Hisamoto
N.
Matsumoto
K.
Takeda
K.
Ichijo
H.
Regulation of anoxic death in Caenorhabditis elegans by mammalian apoptosis signal-regulating kinase (ASK) family proteins
Genetics
2011
, vol. 
187
 (pg. 
785
-
792
)
21
Reference deleted
22
Ng
S. B.
Selvarajan
V.
Huang
G.
Zhou
J.
Feldman
A. L.
Law
M.
Kwong
Y. L.
Shimizu
N.
Kagami
Y.
Aozasa
K.
, et al. 
Activated oncogenic pathways and therapeutic targets in extranodal nasal-type NK/T cell lymphoma revealed by gene expression profiling
J. Pathol.
2011
, vol. 
223
 (pg. 
496
-
510
)
23
Green
D. R.
Reed
J. C.
Mitochondria and apoptosis
Science
1998
, vol. 
281
 (pg. 
1309
-
1312
)
24
Leptihn
S.
Har
J. Y.
Chen
J.
Ho
B.
Wohland
T.
Ding
J. L.
Single molecule resolution of the antimicrobial action of quantum dot-labeled sushi peptide on live bacteria
BMC Biol.
2009
, vol. 
7
 pg. 
22
 
25
Kyte
J.
Doolittle
R. F.
A simple method for displaying the hydropathic character of a protein
J. Mol. Biol.
1982
, vol. 
157
 (pg. 
105
-
132
)
26
Claros
M. G.
Vincens
P.
Computational method to predict mitochondrially imported proteins and their targeting sequences
Eur. J. Biochem.
1996
, vol. 
241
 (pg. 
779
-
786
)
27
McGuffin
L. J.
Bryson
K.
Jones
D. T.
The PSIPRED protein structure prediction server
Bioinformatics
2000
, vol. 
16
 (pg. 
404
-
405
)
28
Horwich
A. L.
Kalousek
F.
Fenton
W. A.
Furtak
K.
Pollock
R. A.
Rosenberg
L. E.
The ornithine transcarbamylase leader peptide directs mitochondrial import through both its midportion structure and net positive charge
J. Cell. Biol.
1987
, vol. 
105
 (pg. 
669
-
677
)
29
Allison
D. S.
Schatz
G.
Artificial mitochondrial presequences
Proc. Natl. Acad. Sci. U.S.A.
1986
, vol. 
83
 (pg. 
9011
-
9015
)
30
Li
P.
Sun
M.
Ho
B.
Ding
J. L.
The specificity of Sushi peptides for endotoxin and anionic phospholipids: potential application of POPG as an adjuvant for anti-LPS strategies
Biochem. Soc. Trans.
2006
, vol. 
34
 (pg. 
270
-
272
)
31
van Meer
G.
Voelker
D. R.
Feigenson
G. W.
Membrane lipids: where they are and how they behave
Nat. Rev. Mol. Cell Biol.
2008
, vol. 
9
 (pg. 
112
-
124
)
32
Dolezal
P.
Likic
V.
Tachezy
J.
Lithgow
T.
Evolution of the molecular machines for protein import into mitochondria
Science.
2006
, vol. 
313
 (pg. 
314
-
318
)
33
von Heijne
G.
Mitochondrial targeting sequences may form amphiphilic helices
EMBO J.
1986
, vol. 
5
 (pg. 
1335
-
1342
)
34
Neve
E. P.
Ingelman-Sundberg
M.
Identification and characterization of a mitochondrial targeting signal in rat cytochrome P450 2E1 (CYP2E1)
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
11317
-
11322
)
35
Komiya
T.
Rospert
S.
Koehler
C.
Looser
R.
Schatz
G.
Mihara
K.
Interaction of mitochondrial targeting signals with acidic receptor domains along the protein import pathway: evidence for the ‘acid chain’ hypothesis
EMBO J.
1998
, vol. 
17
 (pg. 
3886
-
3898
)
36
Bauer
M. F.
Sirrenberg
C.
Neupert
W.
Brunner
M.
Role of Tim23 as voltage sensor and presequence receptor in protein import into mitochondria
Cell
1996
, vol. 
87
 (pg. 
33
-
41
)
37
Bauer
M. F.
Hofmann
S.
Neupert
W.
Brunner
M.
Protein translocation into mitochondria: the role of TIM complexes
Trends Cell Biol.
2000
, vol. 
10
 (pg. 
25
-
31
)
38
Horwich
A. L.
Kalousek
F.
Fenton
W. A.
Pollock
R. A.
Rosenberg
L. E.
Targeting of pre-ornithine transcarbamylase to mitochondria: definition of critical regions and residues in the leader peptide
Cell
1986
, vol. 
44
 (pg. 
451
-
459
)
39
Hammen
P. K.
Weiner
H.
Mitochondrial leader sequences: structural similarities and sequence differences
J. Exp. Zool.
1998
, vol. 
282
 (pg. 
280
-
283
)
40
Rocha
M.
Hernandez-Mijares
A.
Garcia-Malpartida
K.
Banuls
C.
Bellod
L.
Victor
V. M.
Mitochondria-targeted antioxidant peptides
Curr. Pharm. Des.
2010
, vol. 
16
 (pg. 
3124
-
3131
)
41
Sheu
S. S.
Nauduri
D.
Anders
M. W.
Targeting antioxidants to mitochondria: a new therapeutic direction
Biochim. Biophys. Acta
2006
, vol. 
1762
 (pg. 
256
-
265
)
42
Szewczyk
A.
Wojtczak
L.
Mitochondria as a pharmacological target
Pharmacol. Rev.
2002
, vol. 
54
 (pg. 
101
-
127
)
43
Anders
M. W.
Robotham
J. L.
Sheu
S. S.
Mitochondria: new drug targets for oxidative stress-induced diseases
Expert Opin. Drug Metab. Toxicol.
2006
, vol. 
2
 (pg. 
71
-
79
)
44
Ellerby
H. M.
Arap
W.
Ellerby
L. M.
Kain
R.
Andrusiak
R.
Rio
G. D.
Krajewski
S.
Lombardo
C. R.
Rao
R.
Ruoslahti
E.
, et al. 
Anti-cancer activity of targeted pro-apoptotic peptides
Nat. Med.
1999
, vol. 
5
 (pg. 
1032
-
1038
)

Supplementary data