Biomedical Research (Tokyo) 36 (2) 121-133, 2015

Cellular localization and tissue distribution of endogenous DFCP1 protein Tomohisa NANAO,1* Masato KOIKE,2* Junji YAMAGUCHI,1 Mitsuho SASAKI,3 and Yasuo UCHIYAMA1

Departments of 1Cellular and Molecular Neuropathology and 2Cell Biology and Neurosciences, Juntendo University Graduate School of Medicine, Bunkyo-Ku, Tokyo 113-8421, Japan; and 3Department of Animal Models for Human Diseases, National Institute of Biomedical Innovation, Ibaraki-City, Osaka, 567-0085, Japan (Received 26 December 2014; and accepted 10 January 2015)

ABSTRACT Autophagy is essential for the maintenance of cellular metabolism. Once autophagy is induced in cells, the isolation membrane forms a so-called phagophore. The endoplasmic reticulum (ER) is one of several candidates for the membrane source for phagophores. Recently, LC3-positive isolation membranes were found to emerge from a DFCP1 (double FYVE domain-containing protein)positive, ER-associated compartment called the omegasome. Although the GFP-tagged DFCP1 protein has been examined in cultured cells, little is known about the precise cellular and tissue distribution of this endogenous protein. To determine the expression of the endogenous DFCP1 protein, we produced antibodies specific to mouse DFCP1 protein. The antibody recognized both human and mouse DFCP1 proteins, both of which have molecular masses of approximately 87 kDa. In HeLa cells under normal conditions, immunoreactivity for DFCP1 was found dotted or tubular along Tom20-positive filamentous mitochondria and was only partially co-localized in the ER or Golgi apparatus. Moreover, under starved conditions, distinct DFCP1-positive structures became more dotted and scattered in the cytoplasm, while one part of the LC3-positive autophagosomes were immunopositive for DFCP1. These results indicate that an antibody raised against DFCP1 could be a useful tool in explaining the mechanism of phagophore formation from omegasome compartments.

Macroautophagy, which is simply referred to as autophagy, is an intracellular bulk degradation system that is conserved among eukaryotes and plays a pivotal role in the maintenance of cellular metabolism (9, 20, 21). In addition to basic or constitutive autophagy, there is also induced autophagy that occurs in response to various stresses under pathological conditions such as starvation, inflammation and ischemia (12, 13, 22). Once autophagy is induced in cells, the isolation membrane (IM) forms a cupshaped double membrane sac called a phagophore, Address correspondence to: Masato Koike, Department of Cell Biology and Neurosciences, Juntendo University Graduate School of Medicine, Bunkyo-Ku, Tokyo 1138421, Japan Tel: +81-3-5802-1539, Fax: +81-3-5800-0245 E-mail: [email protected]

which further forms autophagosomes that are labeled with LC3. Autophagosomes receive lysosomal hydrolytic enzymes either via transporting vesicles from the trans-Golgi network (TGN) or by fusing with lysosomes to become autolysosomes (10, 12, 19).   In mammalian cells, the membrane origin of phagophores and/or autophagosomes remains unclear, or rather it has become more controversial. The membrane of the endoplasmic reticulum (ER) is suggested to be responsible for the formation of phagophores (2), which was recently confirmed by electron tomography (7, 31). The membrane origin of phagophores was derived not only from the ER but also from other membrane organelles such as the Golgi apparatus, Atg9-containing compartment, * These authors contributed equally to this work.

122

recycling endosomes, mitochondria, ER-mitochondria contact site, ER exit site, and plasma membranes (3–6, 15, 16). The ER, however, has become a strong candidate for the membrane source when current mechanistic insights into autophagosomal formation are considered (8, 18).  Live-cell imaging has shown phosphatidylinositol 3-phosphate (PI3P)-enriched compartments that are closely associated with the ER serving as sites for the initiation of autophagosome formation; LC3-positive isolation membranes have been discovered at the center of the green fluorescent protein (GFP)tagged, DFCP1 (double FYVE domain-containing protein)-positive, ER-associated compartment, which is called the omegasome because of its shape (1). Electron microscopy has shown an ER-phagophore complex existing between two ER cisternae (7, 28). Thus far, the DFCP1 is the only marker for omegasomes. Although an understanding of the precise behavior and cellular distribution of DFCP1 under both normal and starved conditions is useful for a better understanding of the formation of phagophores and/ or autophagosomes, most studies have been performed using cells expressing GFP-tagged DFCP1 (1, 6, 24, 28–30).  In the present study, in order to examine the precise intracellular and tissue distribution of endogenous DFCP1, we prepared antibodies specific to mouse DFCP1. The antibodies recognized human and mouse DFCP1, as shown by Western blotting and by immunohistochemistry using cultured cells expressing mCherry-tagged DFCP1. Under normal and starved conditions of HeLa cells, immunoreactivity for endogenous DFCP1 basically showed a punctated pattern with positive signals that were associated with Tom20-positive mitochondria. Moreover, DFCP1-positive structures were scattered and partially co-localized in LC3-positive autophagosomes. Our present data indicated that the antibody raised against DFCP1 was a useful tool in explaining the mechanism of phagophore formation. MATERIAL AND METHODS The procedures involving animal care and sample preparation were approved by the Animal Experimental Committee of the Juntendo University Graduate School of Medicine, and were performed in accordance with the NIH guidelines and the regulations and guidelines for the care and use of laboratory animals of the Juntendo University Graduate School of Medicine.

T. Nanao et al.

Animals. Male C57BL/6J 8-week-old mice and female Japanese white rabbits were obtained from Charles River Laboratories (Yokohama, Japan) and Oriental Yeast Co. Ltd. (Tokyo, Japan), respectively, and subsequently were housed under specific pathogen-free conditions at the Juntendo University. Generation of an antibody against the mouse DFCP1. The antibody was prepared as previously described (17, 27). Briefly, the two DFCP1 peptides corresponding to its amino acid and carboxyl terminal (#1:Acetyl-EELHRQERLRNHERIRLKAGHVPYCCONH2 and #2: NH2-CDARNVQLDVTEAQADD EGGTLIARK-COOH, respectively) were conjugated with KLH and emulsified individually with Titer Max Gold adjuvant (CytRx Corp., Los Angeles, CA), and the emulsification was injected subcutaneously into female rabbits at 2–4 week intervals. Each polyclonal antibody was purified from antisera using an affinity column conjugated with the antigen. Antibodies. Rabbit polyclonal anti-DFCP1 antibodies for amino acid and carboxyl terminals (hereafter referred to as DFCP #1 and 2, respectively) were generated in the present study, as mentioned above. The commercially available antibodies used for immunohisto-/cytochemistry and Western blotting in the present study were as follows: mouse monoclonal antibody to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Ambion, Austin, TX), β-actin (Sigma, St. Louis, MO), GM130 (BD Biosciences, San Jose, CA), glucose-regulated protein 78/BiP (BD Biosciences), Tom20 (Santa Cruz Biotechnology, Santa Cruz, CA), KDEL (Enzo life sciences, Farmingdale, NY), synapsin1 (Synaptic Systems, Göttingen, Germany), LC3 (Cosmo Bio, Tokyo, Japan), goat polyclonal antibody to MAP2 (Frontier Institute, Hokkaido, Japan), and mitochondrial creatine kinase (CK-Mi) (Frontier Institute), and rabbit polyclonal antibody to LC3A/B (Cell signaling, Danvers, MA). Cell culture and plasmid transfection. HeLa cells were cultured in DMEM (Nacalai Tesque, Tokyo, Japan) supplemented with 100 U/mL penicillin, 100 μg/ mL streptomycin (Nacalai Tesque), and 10% fetal bovine serum (10% FBS DMEM). To induce cell starvation, cells were incubated for 90 min in DMEM (high glucose) with sodium pyruvate and without amino acids (Wako Pure Chemicals, Osaka, Japan) for an efficient induction of autophagy. Cells were transfected using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA), according to the manufacturer’s instructions.

Localization of DFCP1

Primary neuronal cultures. Primary cortical neurons were prepared from the mouse cerebrum at embryonic day (E) 16, based on a method reported previously (14). Briefly, dissociated cortical neurons were plated on poly-L-ornithine (Sigma)-coated glass coverslips. Two hours later, the medium was changed from normal [5% FBS, 2 mM L-glutamine (Nacalai Tesque) in DMEM (Nacalai Tesque)] to maintaining [2% B27 (Invitrogen), 2 mM L-glutamine (Nacalai Tesque) in Neurobasal medium (Invitrogen)]. Plasmid construction and cell transfection. An open reading frame of the mouse DFCP1 gene was subcloned into the BglII/SalI site of the pmCherry-C1 vector (Clontech, Mountain View, CA). The sequences of the primers that were used for the cloning were BglII-mDFCP1-S (5’-gatcagatctatgagtgcccagac ttccctagcag-3’) and SalI-stop-mDFCP1-A (5’-ggaagt cgacttaaaggtcaccgggctttttattg-3’). siRNA transfection. Cells were transfected with either non-targeting unlabeled siRNA (MISSION siRNA Universal Negative Control #1, SIC-001, Sigma) or oligonucleotides targeting DFCP1 [mixture of Hs ZFYVE 6580 and 0204 (Sigma)] using Lipofectamine RNAiMAX™ (Invitrogen) according to the manufacturer’s instructions, and the transfected cells were then cultured for 48 h. RT-PCR of DFCP1 and GAPDH. Total RNAs were extracted from wild-type and DFCP1-knockdown HeLa cells using an RNeasy Mini Kit (QIAGEN, Hilden, Germany). Total RNAs were then reversetranscribed using RevaTra Ace (Toyobo, Osaka, Japan). The primers used for GAPDH were 5’-acctca actacatggtttacatgttc-3’ and 5’-tgtcatcatatttggcaggtt tttct-3’. The primers used for DFCP1 were 5’-aaagtc tcatactctcaaccacact-3’ and 5’-gggttgtaagtcctcgttccc-3’. Immunoblot analysis. HeLa cells were harvested and lysed with a lysis buffer [1% Triton X-100 and protease inhibitor cocktail (Nacalai Tesuque) in PBS]. The lysates were then centrifuged at 20,400 × g for 10 min, and the resultant supernatant was collected. Various tissues were obtained from the mice after the administration of deep anesthesia with pentobarbital (25 mg/kg, intraperitoneal). The tissues were frozen in liquid nitrogen and pulverized using a Cryo-Tec frozen tissue crusher (Microtec Co., Funabashi, Japan). Each pulverized tissue was homogenized with 5–10 volumes of the lysis buffer using a Polytron PT3100 (Kinematica, Littau, Switzerland), and the homogenates were incubated on ice for 30 min and

123

centrifuged at 20,400 × g for 10 min, after which the resultant supernatants were collected. After measuring the protein concentration using a BCA Protein Assay Reagent (Thermo Fisher scientific, Rockford, IL), equal amounts of total protein (10 μg/lane) were subjected to 12.5% sodium dodecylsulfate (SDS)polyacrylamide gel electrophoresis (PAGE), and were then transferred electrophoretically onto a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore Co., Bedford, MA), as described previously (11, 25). The sheets were soaked in PBS containing 0.1% Tween 20 (TPBS) and 5% bovine serum albumin (Sigma) to block non-specific binding, and were then incubated overnight with antibodies to DFCP1 #1 (1 : 1,000), LC3A/B (1 : 1,000), β-actin (1 : 100,000) and GAPDH (1 : 100,000). They were then washed 3 times for 10 min in TPBS, and further incubated for 1 h at room temperature (RT) with horseradish peroxidase-labeled pig anti-rabbit or goat anti-mouse IgG (1 : 1,000) (DAKO, Glostrup, Denmark). After three 10 min washes in TPBS, the membranes were treated with Immobilon Western Chemiluminescent HRP Substrate (Millipore, Billerica, MA) for 2 min and then observed using an LAS4000 mini system (Fuji Photo Film, Tokyo, Japan). Immunocytochemistry. Immunostaining was used to examine tissue and cellular distribution, as reported previously (14, 27). Cells grown on coverslips were fixed in 4% paraformaldehyde in 0.1 M PB for 10 min. After fixation, cells were permeabilized by 0.5% Tween 20 in 0.1 M PB for 10 min. They were then blocked with a TNB blocking buffer containing 0.1 M tris-HCl pH 7.5, 0.15 M NaCl and 0.5% blocking reagent (PerkinElmer, Waltham, MA). For a single staining of DFCP1, samples were incubated with anti-DFCP1 #2 (1 : 200) overnight at 4°C and with Cy3-conjugated donkey anti-rabbit IgG (Jackson Laboratory, Bar Harbor, ME) (1 : 300) for 1 h at RT. For double staining, samples were incubated with a mixture of anti-DFCP1 #2 (1 : 200) and antiGM130 (1 : 200), BiP (1 : 200), and either Tom20 (1 : 200) or LC3 (1 : 100) overnight at 4°C and with a mixture of Alexa 488-conjugated donkey antimouse IgG (Life Technologies, Rockville, MD) (1 : 400) and Cy3-conjugated donkey anti-rabbit IgG (1 : 200) for 1 h at RT. Cultured cortical neurons were incubated overnight at 4°C with mixtures of the following primary antibodies: mouse anti-synapsin 1 (1 : 200), goat anti-MAP2 (1 : 200), rabbit antiDFCP1 #2 (1 : 200), mouse anti-GM130 (1 : 200), and either KDEL (1 : 200) or guinea pig anti-CK-Mi (1 : 200). The coverslips were further incubated for

124

1 h at room temperature with a mixture of the appropriate secondary antibodies: Alexa405-, Alexa488-, Cy5-, and Cy3-conjugated donkey anti-goat, mouse, rabbit or guinea pig IgG (1 : 300) (Life technologies and Jackson Laboratory), respectively.  Samples were then viewed with a confocal laser scanning microscope (FV1000; Olympus, Tokyo, Japan) after counterstaining with 4’-6’-diamindino-2phenylindole (DAPI). RESULTS Generation of a specific antibody against mouse DFCP1 To examine both the tissue distribution and cellular localization of endogenous DFCP1, specific antibodies against synthetic peptides that had 25 amino acid sequences of 47–71 and 656–680 of mouse DFCP1 near both the amino-terminus (DFCP1 #1) and the carboxyl terminus (DFCP1 #2) were prepared. Primary amino acid sequences of DFCP1 are highly conserved between mice and humans (about 95% sequence identity) (Fig. 1A), and the amino acid sequence of DFCP1 #1 is the same in both species (Fig. 1B). In mice, the amino acid sequence of DFCP1 #2 differs from that of humans by three amino acid residues (Fig. 1B). We found that the antibody against DFCP1 #1 was only available for Western blotting, whereas that against DFCP1 #2 worked only for immunocytochemistry. To verify the specificity of the former and the latter antibodies, non-treated (WT) HeLa cells and those treated with siRNA to knockdown DFCP1 expression (knockdown (KD) cells) were used for Western blotting and immunocytochemistry, respectively. Decreased levels of DFCP1 mRNA and protein expressions were confirmed in the KD cells by RT-PCR and Western blotting, respectively (Fig. 1C and 1D). With Western blotting, protein bands immunopositive to anti-DFCP1 #1 that appeared at a molecular mass of approximately 87 kDa, which was identical to the expected molecular weight of DFCP1, were detected in both WT and siRNA-negative control cells, but not in KD cells. A protein band was also detected at a molecular mass below 50 kDa, which did not change its intensity in lysates from the KD cells, and indicated that this protein band was considered to be non-specific (Fig. 1D).  For immunocytochemistry, siRNA was transfected with GFP to show the transfected cells. In the KD cells with intense GFP signals, no immunopositive signals were detected in the cytoplasm, whereas in the negative control cells, intense signals for DFCP1

T. Nanao et al.

were distinct around the nuclei, suggesting that the antibody to DFCP1 #2 specifically detected endogenous DFCP1 by immunocytochemistry (Fig. 2A). To further examine the antibody specificity, mCherryDFCP-expressing cells under normal conditions and at 90 min after starvation were stained with anti-DFCP #2 and observed by confocal laser scanning microscopy. Western blotting revealed that the amount of the membrane bound LC3 (LC3-II) was increased in wild-type HeLa cells at 90 min after starvation, while that of the cytosolic LC3 (LC3-I) was decreased in them, suggesting that the starvation in HeLa cells induced autophagy. On the other hand, the protein amount of endogenous DFCP1 did not differ after starvation (Fig. 2B). Under non-treated conditions, intense granular immunosignals for DFCP1 were observed in the perinuclear region and weak meshlike structures were positive in the peripheral region, whereas dotted immunopositive signals were scattered in the cytoplasm after starvation (Fig. 2C). The perinuclear granular signals disappeared after starvation. Under both conditions, mCherry-positive signals indicative of mCherry-DFCP1 were largely coincident with positive signals immunostained with anti-DFCP1 #2 (Fig. 2C).  These results confirmed that the DFCP #1 and #2 antibodies were specific for DFCP1 and could be used in subsequent experiments to study its cellular and tissue distributions. Distinct cellular distribution patterns of endogenous DFCP1 in wild-type and mCherry-DFCP1-introduced HeLa cells As demonstrated above, the anti-DFCP #2 antibody specifically recognized endogenous DFCP1 in the cytoplasm of HeLa cells. However, the distribution pattern of endogenous DFCP1 in the non-transfected HeLa cells (Fig. 2A) was basically tubular or dotted around the nuclei and was somewhat different from that of mCherry-DFCP1 (Fig. 2C). However, when the distribution pattern of endogenous DFCP1 was examined in the mCherry-DFCP-expressing HeLa cells using anti-DFCP1 #2, it was altered and became similar to that of mCherry-DFCP1, which was similar to that of GFP-DFCP1, as shown previously (1, 8, 28) (Fig. 2C). This tendency was clearly shown by comparing the immunoreactivity of DFCP1 in the HeLa cells with or without the expression of mCherry-DFCP1 cultured under normal and starved conditions that were located in the same visual fields (Fig. 2D). Thus we could examine cells with or without the expression of mCherry-DFCP1 in the same visual field as indicated with asterisks (*) or pound

Localization of DFCP1

125

Fig. 1 Characterization of anti-DFCP1 antibody #1. (A) Amino acid sequence of mouse DFCP1. The peptide sequences near amino- and carboxyl-terminal portions that were used as antigens are indicated in red and blue, respectively. (B) Comparison of the peptide sequences of human (Hs) DFCP1 with that of mouse (Ms) used as antigens. Non-conserved amino acids are underlined. (C) Suppression of DFCP1 expression by siRNA. mRNAs obtained from wild-type (WT), knock-down (KD) control, and DFCP1 KD HeLa cells were reverse transcribed, and the DFCP1 gene was amplified by PCR. The upper and lower panels show DFCP1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), respectively. (D) Western blot analysis of endogenous DFCP1 in WT, KD control, and DFCP1 KD HeLa cells. The upper panel shows protein bands immunopositive for DFCP1, and the lower panel shows actin. Expression of the DFCP1 (87-kDa bands) was significantly suppressed in the KD cells. The asterisk indicates a nonspecific band, which is not suppressed in the KD cells.

signs (#), respectively (Fig. 2D). As shown in Figure 2D, the fluorescent intensity of endogenous DFCP1 obtained by anti-DFCP1 #2 was much lower in the HeLa cells without mCherry-DFCP1 than in those with mCherry-DFCP1. We therefore adjusted the fluorescent intensity in both cells to obtain the appropriate localization patterns of labeled or non-labeled DFCP1 in the cells (see asterisks for endogenous DFCP1 in non-expressing cells and pound signs for mCherrry-DFCP1 in expressing cells). Under non-

starved conditions (upper panels in Fig. 2D), granular immunosignals for DFCP1 were concentrated in the perinuclear region of mCherry-DFCP1-expressing cells and weaker signals were observed in their peripheral region (see Fig. 2C), whereas in non-expressing cells tubular or dotted immunosignals for DFCP1 were evenly distributed in the cytoplasm. Under starved conditions (lower panels in Fig. 2D), dotted immunosignals for DFCP1 were no longer concentrated in the perinucler region but scattered in

126

T. Nanao et al.

Fig. 2 Characterization of anti-DFCP1 antibody #2. (A) Immunofluorescent staining for the DFCP1 in control or DFCP1 siRNA transfected HeLa cells using anti-DFCP1 antibody #2. Immunofluorescence of DFCP1 (red) in the cytoplasm disappears in the knockdown cell. GFP expression vectors were co-transfected as a transfection marker (green). Nuclei were stained with DAPI (blue). (B) Western blot analysis of endogenous DFCP1 and LC3 in wild-type HeLa cells with or without starvation. Western blot of β-actin represent internal controls. (C) Immunofluorescent staining using anti-DFCP1 antibody #2 for the DFCP1 in mCherry-DFCP1-expressing HeLa cells under normal (control; upper panels) and starved (lower panels) conditions. Squares with dashed and solid lines correspond to the peripheral and perinuclear legions of the control cells, respectively, and those with a solid line in the starved cells are enlarged and shown as enlarged images. The enlarged images in the peripheral legion are enhanced to demonstrate clearer immunosignals. Immunopositive signals for DFCP1 (green) and mCherry fluorescence (red) overlapped significantly. Under normal conditions they are detected as perinuclear granular signals and as weak peripheral mesh-like structures (upper and middle panels of the enlarged images, respectively), whereas under starved conditions their dotted signals are scattered in the cytoplasm (lower panels of the enlarged images). Nuclei were stained with DAPI (blue). (D) Immunofluorescent staining using the anti-DFCP1 antibody #2 in mCherryDFCP1-expressing and non-expressing HeLa cells with and without starvation (upper and lower panels, respectively) to compare the distribution patterns of the DFCP1-immunopositive signals in these cells. Boxed areas are enlarged and shown in the next panels on the right (enlarged). Distribution pattern of DFCP1 in mCherry-DFCP1-expressing and non-expressing cells are examined in normal and high intensity images, respectively. Nuclei of the mCherry-DFCP1-expressing and non-expressing HeLa cells examined are indicated with asterisks (*) and pound signs (#). Arrowheads indicate granular immunosignals for DFCP1 in the perinuclear region under non-starved condition and in the cytoplasm under starved conditions. bars: 10 μm and 5 μm (enlarged images in C and D)

Localization of DFCP1

the cytoplasm of mCherry-DFCP1-expressing cells, while the immunosignals for DFCP1 in the cytoplasm of non-expressing cells became intense, and dotted and tubular in shape.  These results were strong arguments for the presence of a cellular distribution pattern of endogenous DFCP1 in wild-type cells that was distinct from those of DFCP1 tagged with either GFP or mCherry. Moreover, these results also indicated that in cells with the expression of DFCP1 tagged with either GFP or mCherry, the distribution pattern of endogenous DFCP1 was also altered and became similar to that of exogenously expressed DFCP1. Comparison of the cellular distribution of endogenous DFCP1 in HeLa cells with or without starvation By double staining with the anti-DFCP1 #2 antibody and antibodies for several organelle markers, we precisely checked the cellular distribution of endogenous DFCP1 in HeLa cells under normal and starved conditions. We first examined the correlation of endogenous DFCP1 with GM130 or BiP, which are markers for either the Golgi apparatus or the ER, respectively (Fig. 3A and B). Under normal conditions, immunosignals for DFCP1 were only partially co-localized for the GM130-immunopositive Golgi apparatus (Fig. 3A, upper panel), whereas these immunosignals were distinct from each other under starved conditions (Fig. 3A, lower panel). With or without starvation, the immunosignals for DFCP1 and BiP were partially co-localized (Fig. 3B). By double immunostaining for DFCP1 and Tom20, which is a marker for mitochondria, intense DFCP signals were found dotted or tubular along Tom20positive filamentous mitochondria under normal conditions (Fig. 3C, upper panel). Under starved conditions, immunoreactivity for Tom20 was changed to a coarse and round appearance, while that for DFCP1 became more punctated than that detected under normal conditions, and was well co-localized with immunosignals for Tom20 (Fig. 3C, lower panel). These results suggested that immunoreactivity for endogenous DFCP1 was preferentially localized in mitochondria-related membrane compartments either with or without starvation. Localization of endogenous DFCP1 in omegasomes According to the previous live-cell imaging and immunocytochemistry using the GFP antibody in GFP-DFCP1 expressing cells, LC3-positive isolation membranes are known to emerge from the center of GFP-positive omegasomes (1). To examine whether endogenous DFCP1 in the omegasomes was also de-

127

tected by the anti-DFCP1 #2 antibody, we performed double immunostaining of DFCP1 and LC3, using HeLa cells with or without starvation. Immunosignals for LC3 were rarely punctated in the cytoplasm of HeLa cells under normal conditions (Fig. 4A, upper panel). In HeLa cells under starved conditions, however, LC3-positive dots were increased in number and size, and some appeared as ring-like structures, some of which were closely associated with the DFCP1-positive puncta (arrowheads in Fig. 4B). Notably, one part of LC3-positive ring-like structures was co-localized with DFCP1-positive puncta (arrows in Fig. 4B), indicating that LC3-positive ring-like membrane structures were deeply associated with DFCP1-positive puncta. These results suggested that DFCP1 was detected as omegasome compartments. Tissue distribution of DFCP1 As stated above, the antibody recognized endogenous DFCP1 in the HeLa cells. To examine the levels of DFCP1 expression in various mouse tissues, extracts from each tissue were subjected to Western blotting using the antibody. As shown in Fig. 5, DFCP1 was ubiquitous in mouse tissues, and the main bands observed in each tissue had a molecular mass at approximately 87 kDa (Fig. 5A). Densities of main bands for DFCP1 in each tissue were measured and the relative intensity of each band is shown in Fig. 5B. The expression levels were higher in the pancreas, spleen and central nervous system (CNS), and lowest in the muscular tissues, indicating that the expression of DFCP1 depends on the tissue type. In addition, protein bands other than the main 87 kDa band were discerned in the CNS, kidney and testis, appearing below the main band. Localization of DFCP1 in mouse primary cerebral cortical neurons Because the expression of DFCP1 was high in CNS tissues, we prepared primary cerebral cortical neurons, which were cultured for 10 days in vitro (10 DIV) and then subjected to immunocytochemistry for DFCP1. A positive signal for DFCP1 was basically a punctated appearance in somatodendrites, as well as in the axons and their terminals (Fig. 6). Immunoreactivity for DFCP1 in the MAP2-positive neuronal perikarya were closely associated or partially co-localized with GM130 or KDEL markers for ER and Golgi apparatus (arrows in Fig. 6A and 6B). By double immunostaining of DFCP1 and CK-Mi, which is a marker for mitochondria (26), DFCP positive signals were also associated or partially co-local-

128

T. Nanao et al.

Fig. 3 Double immunofluorescent staining of endogenous DFCP1 and various organelle markers under normal (control: upper panels) and starved (lower panels) conditions in HeLa cells. Boxed areas in the images under low magnification are enlarged and shown in the next panels on the right. (A) GM130 (Golgi marker, green) is partially co-localized with DFCP1 (red) under normal conditions (arrowheads), but they distinctly differ from each other under starved conditions. (B) BiP (ER marker, green) is closely associated and partially co-localized with DFCP1 (red) under both normal and starved conditions (arrowheads). (C) Immunopositive signals for DFCP1 (red) are detected along Tom20-positive mitochondria (green) and they are largely co-localized under both normal and starved conditions. Nuclei were stained with DAPI (blue). bars: 10 μm (low magnification) and 5 μm (enlarged images).

Localization of DFCP1

129

Fig. 4 Double immunofluorescent staining of endogenous DFCP1 and LC3 in HeLa cells under normal (control: upper panels) and starved (lower panels) conditions. Boxed areas in the images under low magnification are enlarged and shown in the next panels on the right. (A) Without starvation immunoreactivity for LC3 is barely detectable under low magnification. (B) Immunoreactivity for LC3 is punctated and becomes distinct, increasing in number, size, and intensity under starved conditions (lower panels), while it is easily detectable even under a low magnification image. Some of the LC3-positive dots were closely associated with DFCP1 (arrowheads). Furthermore, some of the LC3-positive dots appear as ring-like structures, one part of which is also immunopositive for DFCP1 (arrows). Nuclei were stained with DAPI (blue). Bars: 10 μm (low magnification) and 5 μm (enlarged).

ized with CK-Mi in the perikaryal (arrows Fig. 6C) and dendritic regions (Fig. 6C). Triple immunostaining for DFCP1, CK-Mi, and synapsin 1 showed that immunoreactivity for DFCP1 was also associated or partially co-localized with CK-Mi signals in the presynaptic region where synapsin 1 immunosignals were detected (Fig. 6D). These results suggested that endogenous DFCP1 was associated with the ER, Golgi apparatus and mitochondria in the neuronal perikarya and with the mitochondria in the pre-synaptic terminals. DISCUSSION The main findings of the present study are summarized as follows: 1) Western blotting and immunocytochemistry confirmed the specificity of two anti-DFCP1 antibodies; 2) immunoreactivity for endogenous DFCP1 was basically punctated and tubular, and often positive along the mitochondria with or without

starvation; 3) endogenous DFCP1 was also positive in omegasomes and associated with LC3-positive autophagosomes after starvation; 4) DFCP1 in mice was ubiquitously expressed but in a tissue-dependent manner, and was high in CNS tissue; and, 5) immunoreactivity for endogenous DFCP1 in the primary cortical neurons at 10 DIV was partially co-localized with that for the ER, Golgi apparatus, and mitochondria in the somatodendritic portion, and with that for mitochondria in the pre-synaptic region.  The biogenesis of autophagosomes entails the nucleation and growth of a double-membrane sheet, the phagophore. In mammalian cells, vesicles containing Atg9, the only integral membrane protein, are believed to play an important role in the nucleation of the phagophore (3, 23). The omegasome is a PI3P-enriched, cup-shaped membrane compartment arose from the ER, and could be identified by the presence of DFCP1; the omegasome is consid-

130

T. Nanao et al.

Fig. 5 Distribution of DFCP1 in various mouse tissues. (A) Western blot analysis of various mouse tissue lysates using antiDFCP1 (upper panel) or anti-GAPDH (lower panel) antibodies. Western blot of GAPDH represent internal controls. (B) The intensity of each band in (A) was measured, and the ratio of the amount of the DFCP1 to GAPDH protein was calculated.

ered to serve as a platform for autophagosome biogenesis in mammalian cells (1). Through analyses of the behavior of autophagy-related proteins including Atg9A and DFCP1, the membrane origin of autophagosomes has been proposed, and/or the problem that should be settled concerning the membrane source has been highlighted. Until now, the most problematic aspect of these studies has been an inability to detect endogenous DFCP1.  As for Atg9, we generated an anti-Atg9 antibody for the specific recognition of endogenous Atg9 by Western blotting and immunohisto-/cytochemistry of various mouse tissues and cells (27). After specific antibodies to Atg9 were produced, several antibodies that were confirmed to work well became commercially available. Here, we also generated the antibodies and confirmed their specificity to endogenous DFCP1 either by western blotting or immunocytochemistry. The anti-DFCP1 #2 antibody was considered suitable for immunohistochemistry when the dot-like and tubular localization pattern was con-

firmed in HeLa cells under normal conditions and suppressed in cells with the expression of DFCP1 mRNA that was knocked down by siRNA. The antiDFCP1 #2 antibody did not recognize the SDS-treated antigen in the lysates of HeLa cells. Therefore, it may be reasonable to assume that this antibody recognizes the three-dimensionally arranged amino acid residues as the antigenic sites of DFCP1. Indeed, when considering that the antibody recognized a portion of the ring-shaped, LC3-positive autophagosome, it would also be true that the antibody for DFCP1 recognizes membrane compartments that are considered to be omegasomes. As far as we could ascertain, this is the first study showing the cellular and tissue distribution of endogenous DFCP1 and the evidence of endogenous omegasomes in wildtype cells.  When the specificity of the DFCP1 #2 antibody was examined via immunocytochemistry using mCherryDFCP1-expressing and DFCP1-KD HeLa cells as positive and negative controls, we noticed that the

Localization of DFCP1

131

Fig. 6 Immunofluorescent staining of endogenous DFCP1 and various organelle markers in mouse primary cerebral cortical neurons. The neurons were cultured for 10 days in vitro (10 DIV). (A) Triple staining for DFCP1 (red), GM130 (Golgi marker, green), and MAP2 (blue). Immunosignals for GM130 are partially co-localized with those for DFCP1 (arrows) in MAP2-positive neuronal perikarya. (B) Triple staining for DFCP1 (red), KDEL (ER marker, green), and MAP2 (blue). Immunosignals for KDEL are closely associated or partially co-localized with those for DFCP1 (arrows) in MAP2-positive neuronal perikarya. (C, D) Quadruple staining for DFCP1 (red), CK-Mi (mitochondria marker, green), MAP2 (blue) and synapsin 1 (cyan). Boxed areas in the perikarya and neurites at low magnification are enlarged and shown in the next panels on the right and (D), respectively. In MAP2-positive neuronal perikarya, immunoreactivity for DFCP1 is co-localized or closely associated with CK-Mi-positive mitochondria (arrows in the enlarged images on the right). (D) Immunoreactivity for DFCP1 is also co-localized or closely associated with CK-Mi-positive mitochondria in dendrites and synapsin 1-positive pre-synapses (arrowheads). Bars: 10 μm (low magnification in C) and 5 μm (other images).

distribution pattern of DFCP1 in mCherry-DFCP1expressing cells differed from that of endogenous DFCP1 in wild-type HeLa cells even under normal conditions. In mCherry-DFCP1-expressing cells, the immunoreactive signal for DFCP1 was enriched in

the perinuclear Golgi area and was weakly detected in the periphery corresponding to ER, which is similar to the distribution pattern of GFP-DFCP1, as shown previously (1, 8, 28). However, the distribution pattern of endogenous DFCP1 in wild-type HeLa

T. Nanao et al.

132

cells that we observed was enriched in membrane compartments other than either the Golgi apparatus or the ER. The distribution pattern of endogenous DFCP1 in wild-type HeLa cells was dot-like and tubular around the nuclei, and it was associated with Tom20-positive mitochondria under both normal and starved conditions. These results suggest that exogenously expressed DFCP1 behaves differently from endogenous DFCP1, even when autophagy is not induced. Moreover, in cells expressing mCherry/GFPDFCP1, the distribution pattern of endogenous DFCP1 is altered and adjusted to that of mCherry-DFCP1, indicating that the distribution of DFCP1 may be strictly regulated by the expression amount of DFCP1 protein within a cell.  As far as we know this is the first report that endogenous DFCP1 is associated with mitochondria with or without induction of autophagy. Upon the induction of autophagy after starvation or mitophagy, GFP-DFCP1 puncta are known to be co-localized with mitochondria or parkin-positive damaged mitochondria, respectively, suggesting that omegasome formation would be associated with mitochondria as well as with the ER (28–30). Contact sites of the ER-mitochondria are considered to play important roles in autophagosome biogenesis (5, 6); all components of the autophagy-specific class III PI3K (Atg14L, Beclin1, Vps34, and Vps15) and GFPDFCP1 accumulate in the fraction of mitochondriaassociated ER membrane (MAM), which includes ER-mitochondria contact sites after starvation. Considering our present data, endogenous DFCP1 was deeply associated with MAM even under normal conditions. In other words, it is possible that DFCP1 is not only a marker for omegasomes after starvation but also could be a marker for MAM. Of course, further studies using a genetic approach and immunoelectron microscopy for endogenous DFCP1 are necessary to confirm these theories.  In the present study we examined the distribution patterns of DFCP1 in primary cerebral cortical neurons and found that positive signals were basically dot-like and tubular at 10 DIV under both normal and starved conditions. The localization patterns of endogenous DFCP1 in cultured neurons were similar to that in HeLa cells, and were associated with the Golgi apparatus, the ER and the mitochondria. It seems likely that punctate localization would be related to MAM or ERGIC (the ER-Golgi intermediate compartment). In the presynaptic region, endogenous DFCP1 was also detected and associated with the mitochondria. These results indicate that the use of immunoelectron microscopy and the generated

antibody to DFCP1 could promote a further understanding of the molecular mechanisms of autophagosome formation in axons.  Collectively, the present study used specific antibodies to DFCP1 to reveal that the endogenous DFCP1 was preferentially associated with mitochondria rather than with the Golgi apparatus or the ER both with and without starvation, and that it was associated with the endogenous omegasome after starvation in HeLa cells. In primary cortical neurons, endogenous DFCP1 showed better association with the ER and Golgi apparatus but it also associated with mitochondria similar to that in HeLa cells. The antibodies to DFCP1 produced in the present study would be a powerful tool that could help to reveal the real structures of omegasomes without the influence of artificial structures generated by exogenously expressed DFCP1. Acknowledgements This work was supported in part by Grants-in-Aid for Challenging Exploratory Research (Y.U.), Scientific Research on Innovative Areas (Y.U. and M.K.), Grants-in-Aid for Scientific Research (B) (Y.U.), Grants-in-Aid for Scientific Research (C) (M. K.), a Grant-in-Aid for challenging Exploratory Research (Y.U.), a Grant-in-Aid for the “High-Tech Research Center” Project for Private Universities, a matching fund subsidy (Y.U.), and the Program for the Strategic Research Foundation at Private Universities (Y.U.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. REFERENCES   1. Axe EL, Walker SA, Manifava M, Chandra P, Roderick HL, Habermann A, Griffiths G and Ktistakis NT (2008) Autophagosome formation from membrane compartments enriched in phosphatidylinositol 3-phosphate and dynamically connected to the endoplasmic reticulum. J Cell Biol 182, 685–701.   2. Dunn WA, Jr. (1990) Studies on the mechanisms of autophagy: formation of the autophagic vacuole. J Cell Biol 110, 1923–1933.   3. Ge L, Baskaran S, Schekman R and Hurley JH (2014) The protein-vesicle network of autophagy. Curr Opin Cell Biol 29, 18–24.   4. Ge L, Zhang M and Schekman R (2014) Phosphatidylinositol 3-kinase and COPII generate LC3 lipidation vesicles from the ER-Golgi intermediate compartment. eLife 3, e04135.   5. Hailey DW, Rambold AS, Satpute-Krishnan P, Mitra K, Sougrat R, Kim PK and Lippincott-Schwartz J (2010) Mitochondria supply membranes for autophagosome biogenesis during starvation. Cell 141, 656–667.   6. Hamasaki M, Furuta N, Matsuda A, Nezu A, Yamamoto A, Fujita N, Oomori H, Noda T, Haraguchi T, Hiraoka Y, Amano A and Yoshimori T (2013) Autophagosomes form at

Localization of DFCP1 ER-mitochondria contact sites. Nature 495, 389–393.   7. Hayashi-Nishino M, Fujita N, Noda T, Yamaguchi A, Yoshimori T and Yamamoto A (2009) A subdomain of the endoplasmic reticulum forms a cradle for autophagosome formation. Nat Cell Biol 11, 1433–1437.   8. Itakura E and Mizushima N (2010) Characterization of autophagosome formation site by a hierarchical analysis of mammalian Atg proteins. Autophagy 6, 764–776.   9. Kim KH and Lee MS (2014) Autophagy—a key player in cellular and body metabolism. Nat Rev Endocrinol 10, 322– 337. 10. Klionsky DJ and Emr SD (2000) Autophagy as a regulated pathway of cellular degradation. Science 290, 1717–1721. 11. Koike M, Shibata M, Ezaki J, Peters C, Saftig P, Kominami E and Uchiyama Y (2013) Differences in expression patterns of cathepsin C/dipeptidyl peptidase I in normal, pathological and aged mouse central nervous system. Eur J Neurosci 37, 816–830. 12. Koike M, Shibata M, Tadakoshi M, Gotoh K, Komatsu M, Waguri S, Kawahara N, Kuida K, Nagata S, Kominami E, Tanaka K and Uchiyama Y (2008) Inhibition of autophagy prevents hippocampal pyramidal neuron death after hypoxicischemic injury. Am J Pathol 172, 454–469. 13. Koike M, Shibata M, Waguri S, Yoshimura K, Tanida I, Kominami E, Gotow T, Peters C, von Figura K, Mizushima N, Saftig P and Uchiyama Y (2005) Participation of autophagy in storage of lysosomes in neurons from mouse models of neuronal ceroid-lipofuscinoses (Batten disease). Am J Pathol 167, 1713–1728. 14. Koike M, Tanida I, Nanao T, Tada N, Iwata J, Ueno T, Kominami E and Uchiyama Y (2013) Enrichment of GABARAP relative to LC3 in the axonal initial segments of neurons. PLoS One 8, e63568. 15. Lamb CA, Yoshimori T and Tooze SA (2013) The autophagosome: origins unknown, biogenesis complex. Nat Rev Mol Cell Biol 14, 759–774. 16. Longatti A, Lamb CA, Razi M, Yoshimura S, Barr FA and Tooze SA (2012) TBC1D14 regulates autophagosome formation via Rab11- and ULK1-positive recycling endosomes. J Cell Biol 197, 659–675. 17. Lu Z, Dono K, Gotoh K, Shibata M, Koike M, Marubashi S, Miyamoto A, Takeda Y, Nagano H, Umeshita K, Uchiyama Y and Monden M (2005) Participation of autophagy in the degeneration process of rat hepatocytes after transplantation following prolonged cold preservation. Arch Histol Cytol 68, 71–80. 18. Matsunaga K, Morita E, Saitoh T, Akira S, Ktistakis NT, Izumi T, Noda T and Yoshimori T (2010) Autophagy requires endoplasmic reticulum targeting of the PI3-kinase complex via Atg14L. J Cell Biol 190, 511–521.

133 19. Mizushima N (2007) Autophagy: process and function. Genes Dev 21, 2861–2873. 20. Mizushima N and Komatsu M (2011) Autophagy: renovation of cells and tissues. Cell 147, 728–741. 21. Mortimore GE and Poso AR (1987) Intracellular protein catabolism and its control during nutrient deprivation and supply. Annu Rev Nutr 7, 539–564. 22. Murrow L and Debnath J (2013) Autophagy as a stressresponse and quality-control mechanism: implications for cell injury and human disease. Annu Rev Pathol 8, 105–137. 23. Orsi A, Razi M, Dooley HC, Robinson D, Weston AE, Collinson LM and Tooze SA (2012) Dynamic and transient interactions of Atg9 with autophagosomes, but not membrane integration, are required for autophagy. Mol Biol Cell 23, 1860–1873. 24. Ridley SH, Ktistakis N, Davidson K, Anderson KE, Manifava M, Ellson CD, Lipp P, Bootman M, Coadwell J, Nazarian A, Erdjument-Bromage H, Tempst P, Cooper MA, Thuring JW, Lim ZY, Holmes AB, Stephens LR and Hawkins PT (2001) FENS-1 and DFCP1 are FYVE domain-containing proteins with distinct functions in the endosomal and Golgi compartments. J Cell Sci 114, 3991–4000. 25. Suyama M, Koike M, Asaoka D, Mori H, Oguro M, Ueno T, Nagahara A, Watanabe S and Uchiyama Y (2014) Increased immunoreactivity of cathepsins in the rat esophagus under chronic acid reflux esophagitis. J Histochem Cytochem 62, 645–660. 26. Tachikawa M, Fukaya M, Terasaki T, Ohtsuki S and Watanabe M (2004) Distinct cellular expressions of creatine synthetic enzyme GAMT and creatine kinases uCK-Mi and CK-B suggest a novel neuron-glial relationship for brain energy homeostasis. Eur J Neurosci 20, 144–160. 27. Tamura H, Shibata M, Koike M, Sasaki M and Uchiyama Y (2010) Atg9A protein, an autophagy-related membrane protein, is localized in the neurons of mouse brains. J Histochem Cytochem 58, 443–453. 28. Uemura T, Yamamoto M, Kametaka A, Sou YS, Yabashi A, Yamada A, Annoh H, Kametaka S, Komatsu M and Waguri S (2014) A cluster of thin tubular structures mediates transformation of the endoplasmic reticulum to autophagic isolation membrane. Mol Cell Biol 34, 1695–1706. 29. Wong YC and Holzbaur EL (2014) Optineurin is an autophagy receptor for damaged mitochondria in parkin-mediated mitophagy that is disrupted by an ALS-linked mutation. Proc Natl Acad Sci USA 111, E4439–4448. 30. Yang JY and Yang WY (2013) Bit-by-bit autophagic removal of parkin-labelled mitochondria. Nat Commun 4, 2428. 31. Yla-Anttila P, Vihinen H, Jokitalo E and Eskelinen EL (2009) 3D tomography reveals connections between the phagophore and endoplasmic reticulum. Autophagy 5, 1180–1185.

Cellular localization and tissue distribution of endogenous DFCP1 protein.

Autophagy is essential for the maintenance of cellular metabolism. Once autophagy is induced in cells, the isolation membrane forms a so-called phagop...
26MB Sizes 1 Downloads 12 Views