The FASEB Journal • Research Communication

Correlative light and electron microscopic analyses of mitochondrial distrihution in hlastomeres of early fish embryos Yongming Yuan,* Mingyou Ii,*'^ Ni Hong,* and Yunhan Hong*'^'^ *Department of Biological Sciences, National University of Singapore, Singapore, Singapore; and ^College of Eisheries and Life Science, Shanghai Ocean University, Shanghai, China ABSTRACT Early embryos of vertebrates undergo remarkable dynamic molecular events, such as embryonic gradient, cellular polarity, and asymmetry necessary for cell fate decisions. Correlative light and electron microscopy (CLEM) is a powerful tool to investigate rare or dynamic molecular events and has been developed for relatively small cells in culture and tissues but is not yet available for large cells of early development stage embryos. Here we report the capability of CLEM in blastomeres of medaka fish by using the mitochondria detection system. A short N-terminal signal peptide of the mitochondrial protein Tom20 was linked to green fluorescent protein (GFP), resulting in a fusion protein termed Tom20:GFP. The subcellular location of Tom20:GFP in medaka blastomeres reveals the lack of mitochondrial distribution in pseudopodia as well as inconspicuous redistribution during divisions. Blastomeres, after sample preparation procedures including high-pressure freezing and freeze substitution, are able to preserve fluorescence, antigenicity, and fine structures, which allows for precise correlation between the Tom20:GFP fluorescence and mitochondria on merged light and electron micrographs. Furthermore, nanogold immunostaining for Tom20:GFP and endogenous Tom20 revealed their specific localization on the mitochondrial outer membrane. Our results extend the CLEM approach to early development stage embryos of a vertebrate.—^Yuan, Y., Li, M., Hong, N., Hong, Y. Correlative light and electron microscopic analyses of mitochondrial distribution in blastomeres of early fish embryos. FASEB J. 28, 577-585 (2014). www.fasebj.org

Abbreviations: CLEM, correlative light and electron microscopy; DAPI, 4',6'-diamidino-2-phenylidol; DIC, differential interference contrast; DMEM, Dulbecco's modified Eagle's meditnii; EM, electron microscopy; ES, embi'yonic stem; FS, freeze substitution; GEP, green fluorescent protein; HPE, high-pressure freezing; IM, inner membrane; mt, mitochondrion; LM, light microscopy; NA, numerical aperture; OM, outer membrane; PCC, Pearson correlation coefficient; ROI, region of interest; SP, signal peptide; TEM, transmission electron micro.scopy; TMD, transmembrane domain; TIM, translocases of tbe inner initochondrial membrane; TOM, translocases of the otiter mitoehondrial membrane 0892-6638/14/0028-0577 © FASEB

Key Word.'i: medaka • Tom20 • localization • fluorescent pTotein ' out membrane IN DIVERSE ANIMALS, CELLS of

early developing embryos undergo a series of molecular and cellular events, culminating in cell proliferation and fate decisions. Such events may include embryonic gradient, celhtlar polarity, assembly/disassembly of macromolecular complexes, redistribution of organelles, atid intercellular asymmetry. Elucidation of these events is fundamental to understanding the molecular and cellular mechanisms that cotitrol normal and abnormal processes of early development. Mitochondria are some of the most important organelles for energy metabolism and are involved in controlling cell growth and survival. A typical mitochondrion (mt) comprises the outer membrane (OM), inner membrane (IM), intermembrane space, and cristae containing the matrix of the mitochotidrial genome and oxidative enzymes. Most mitochondrial proteitis are encoded by a nuclear genome and synthesized in the cytosol as precursors. They are sorted and imported into different subcompartments by translocases of the outer mitochondrial membrane (TOM) complex and translocases of the inner mitochondrial membrane (TIM) complex or sorting and assembly machinery (1). These precttrsors make use of two different signals for targeting into subcompartments. One is the signal peptide (SP) of 20-50 amino acid residues at the N terminus, which potetitially forms positively charged amphiphilic helices and directs the precursors across both the OM and IM (2). The other is the signal patch, which is formed from several scattered residues on the folded protein (3). The TOM complex ser\'es as the getieral entry gate for mitochondrial import (4). One of its major components is Tom20, a receptor to recognize precursor proteins. It has a transmembrane ' Correspondence: Department of Biological Seienees, National Universit)' of Singapore, 14 Science Drive 4, Singapore 117543, Singapore. E-mail: [email protected] doi: 10.1096/fj. 13-233635 This article inelttdes supplemental data. Please visit http:// www.fasebj.org to obtain this information. 577

domain (TMD) within the N-terminal SP of 33 amino acid residues. Tom 20 is N terminally anchored to the OM and exposes the C-terminal receptor domain to the cytosol (4). The N-terminal SP acts as a "signal-anchored" domain and has been demonstrated to be sufficient for intracellular sordng and anchoring to the OM (5, 6). The mitochondrial protein importation provides an excellent model to study macromolecular trafficking by correlative light and electron microscopy (CLEM). CLEM is a powerful approach for investigating rare and/or dynamic processes in living cells. With the help of a tagged fiuorescent probe such as green fluorescent protein (GFP), targets of interest were first identified under fluorescence light microscopy (LM) and subsequendy analyzed by electron microscopy (EM) for fine ultrastructures at high resolution (7-10). Because crosslinkingfixativessuch as formaldehyde normally quench most of the fluorescent signal, CLEM was performed as a preliminary method with 2 sets of samples before and after fixation. First, fresh material was imaged with LM to record the fluorescent signal, and the same sample was subsequently fixed with chemical or plunge freezing for EM analysis. Because the introduced fixadon procedure was interrupted, a major issue is how to precisely trace back the LM recoded area under EM. Recently, progress has been made in methods to perform CLEM with the development of cryofixation via high-pressure freezing (HPF) followed by freeze substitution (FS). The HPF-FS procedure is able to preserve the sample in a nearly native state and to reduce structural artifacts below the detection limit (11-13). More important, this procedure preserves fluorescence and fine structure and thus allows for a correlative analysis of the same section with botb LM and EM. CLEM studies based on the sample prepared with the HPF-FS procedure have been successful in cultured cells (14, 15) and advanced zebrafish embryos at 48-72 h after fertilization (16). However, the applicability of CLEM to early development stage embi^os of vertebrates is tinclear, because the suitability of the HPF cryofixation procedure for markedly larger and thus fragile cells, namely blastomeres of fish embryos until the blasttila stage, is still not known.

that the Tom20 SP is sufficient to direct its fusion partner GFP to the fine-structured mitochondrial OM, to which the endogenous Tom20 is colocalized by nanogold immunostaining.

MATERIALS AND METHODS Fish

Work with fish followed the Guidelines on the Gare aud Use of Animals for Scientific Purposes of the National Advisory Committee for Laboratory Animal Research in Singapore and was approved by this committee (permit 27/09). Medaka strain i was maintained under an artificial photoperiod of a 14:10-h light-dark cycle at 26°G as described previously (20, 25, 26). Plasmids

Extraction of total RNA and cDNA synthesis were done as described previously (27). Plasmid pTom20gfp was constructed by fusing in-frame a cDNA fragment encoding the 34 N-terminal amino acid residues between Kpnl and BamHl to the gfp-coding sequence iu pcDNA3.1-GEP. The cDNA was amplified from a blástula cDNA library of medaka strain i^ by using primers aaaggtaccATGGGGAGGAGGAGGAG and gcgggatccGAAGTTGGGGTGAGTGGG. where lowercase letters indicate the introduced restriction sites Kpnl aud BamH I for clouing, and underlined letters indicate tlie initiation codon. The pritriers were designed oti the basis of a predicted medaka TotTi20 cDNA sequence (ENSORLT00000001496). Gorrect clouing was confirmed by sequencing. The plasmid pGVpr, which expresses the fusion protein of puromycin acetyltrausferase and red fluorescent protein, was descrihed previously (28). Plasmid DNA for cell transfection and RNA synthesis was prepared with a Midiprep kit (Qiagen, Valencia, GA, USA). mRNA synthesis and nMcroinjecdon RJMA synthesis was performed as described previously (25). Iu brief, pTom20gfp and pH2Bcheri7 were linearized with Apal and A'oil, respectively. The linearized plasmid was used as a template for capped mRNA synthesis with the T7 mMessage mMachiue kit (Ambiou, Austin, TX, USA). RNA injection at the l-ce!l stage (at —50 pg/medaka embryo) was performed as described previously (25).

The laboratory fish medaka {Oryzias latipes) is a favorable vertebrate model for stem cell research (17Cell culture, transfection, and staining 22), and the transparent embryo is an ideal organism for developmental studies with LM and EM (23, 24). The medaka embryonic stem (ES) cell line MESl was maintained on gelatin-coated tissue culture plastic ware iu ESM4 However, conventional methods of EM do not provide medium (18, 29) and trausfected as described previously optimal preservation of all embryonic structures and (30). The basis of ESM4 medium is pure Dulbecco's modified are usually incompatible with immunolabeling and Eagle's medium (DMEM) that is supplemented with antibivisualization of expressed fluorescently tagged proteins. otics, fetal boxdue serum, and growth factors (31). MESl cells The aim of this work was to prepare proper samples at 70% confluence were transfected in 6-well plates by using DNAfectiu reagent (Applied Biological Materials, Richmond, of early development stage blastomeres of medaka for BG, Ganada). In brief, 2 |xg of plasmid DNA (pTom20gfp the CLEM procedures with mitochondrial Tom20 loalone or with pGVpr) and 8 |xl of DNAfectin reagent were calization as a model. We show that the blastomeres can mixed iu 200 |xl of pure DMEM. After incubation at room be used for tbe HPF cryofixation procedures without temperature for 20 miu, the trausfection mixture was added losing fluorescence, antigenicity, and fine structures. By drop^vise to cells in a well of the 6-well plate containing 2 ml use of correlative fluorescence and transmission elecof DMEM. After incubation for 6 h at 28°G, the cells were grown in ESM4 for 12 — 72 h before observation. Eor mitotron microscopy (TEM) on the same sample, we reveal 578

Vol.28

Eebruary2014

The EASEB journal • www.fasebj.org

YUAN ET AL

chondria staining, live cells were incubated in DMEM containing 100 nM MitoTracker (Molecular Probes, Eugene, OR, USA) at 28°C for 20 min, washed in DMEM with changes, and maintained in ESM4 for observation. Preparation of ultrathin sections Medaka blastotneres contain large and fragile cells; thus, dissection of embryos and the following sample loading should be undertaken with special care to avoid cell disruption. Medaka embryos at the midblastula stage were immersed in DMEM, and the whole blastomere was dissected out from the chorion with fine forceps. Meanwhile, the yolk in the blastomere was drained during dissection. Blastomeres were loaded in aluminum carriers (0.3 mm depth) and rapidly frozen (320 ms, 2100 bar) with an HPF machine (HPF 01; M. Wohlwend, Sennwald, Switzerland). Cryofixed samples were transferred into a 1.5-ml cryotube containing 0.1% (w/v) uranyl acetate in pure acetone for FS. All samples were processed by FS and embedding in a temperature-controlled unit (AFS2; Leica Microsystetns, Wetzlar, Germany). After incubation at —90°C for 48 h, the samples were washed in prechilled acetone with 3 changes during gradual raising of the temperature to -50°C (4.5°C/h). Lowicryl HM20 (Electron Microscopy Sciences, Hatfield, PA, USA) at increasing concentrations (25, 50, and 75%) was added for gradient infiltration, with each step lasting 4 h. After the temperature was raised gradually to -30°C (1.6°C/h), 100% Lowicryl HM20 was added to the samples with 3 changes, each lasting 10 h. The .samples were polymerized with UV light for 48 h at -30°C and for 24 h at 20°C. Polymerized blocks were trimmed with a glass knife and sectioned with a diamond knife (Diatome, Hatfield, PA, USA) on an ultramicrotome (Leica EM FCS; Leica Microsystems). Ultrathin sections (220 nm thick) were picked up on 50-mesh nickel grids with support film for observation or further processing. Immunostaining Immunostaining was done as described previously (25), with modifications to the ultrathin sections on grids. In brief, sections were blocked by incubation in PBS containing 4% hoNiine serum albumin and 0.05% Tween 20 for 20 min. The blocked sections were incubated with primary anti-GFP antibody (aGFP, abl218; Abeam hic, Cambridge, MA, USA) or anti-Tom20 antibody (aTom20, ab56783; Abeam Inc.) at a 1:50 dilution for 3 h and washed thoroughly with PBS, followed by incubation for 3 h with secondary antibodies conjugated with 1.4-nm nanogold (catalog no. 2001; Nanoprobes, Yaphank, NY, USA) or with Alexa 546 (7402; Nanoprobes). In certain experiments, 4',6'-diamidino-2-phenylidol (DAPI; 0.5 |Jig/ml) was added for nuclear staining. After washing in distilled water with 3 changes, the sections were analyzed by fluorescence microscopy. The sections of interest were subjected to the gold-enlargement procedure to an appropriate size for EM analysis, according to the supplier's instructions (2113; Nanoprobes). Colocalization analysis of Toni20:GFP Cultured cells cotransferred with pTom20gfp plus pCVpr and blástula cells from medaka embryos injected with H2B:Cherry mRNA (25 pg/embryo) were imaged to determine the subcellular distribution of Tom20:GFP. To analyze the colocalization of mitochondria and Tom20:GFP, cells transferred with pTom20gfp were stained with MitoTracker. Meanwhile, resin sections of blastomeres after HPF fixation were immtinostained with the primary antibody aGFP or aTom 20 and CORRELATIVE MICROSCOPY IN BLASTOMERES

the secondary antibody conjugated with Alexa 546. Stained cells and resin sections were imaged with UltraView VoX (PerkinElmer, Waltham, MA, USA) confocal microscopy using an Olympus water-immersion objective lens (Olympus, Tokyo,Japan). Numerical apertures (NAs) were 1.15 and 1.20 for the above X40 or X60 lens, respectively. Volocity 6.2.1 software (PerkinElmer) was set at 4 channels for sequentially recording modes with laser lines at 405, 488, and 561 nm and differential interference contrast (DIC) optics. The colocalization index was calculated with the same software. Time-lapse imaging of blastomeres Embryos at the 1-cell stage were injected with tom20:gfp mRNA (50 pg/embryo) alone or together with H2B:cherry mRNA. At the blástula stage, blastomeres were dissociated in DMEM and stained with Hoechst 53342 (10 |JLg/ml; Invitrogen, Carlsbad, CA, USA) and MitoTracker for 10 min at room temperature for those embryos injected with tom20:gfp mRNA alone. After 3 rinses with DMEM, the cells were transferred onto a 35-mm glass-bottom culture dish (MatTek, Ashland, MA, USA) and analyzed on the UltraView VoX confocal microscope system in a 28°C temperature-controlled environment using a X40 water-immersion objective lens (NA 1.15). Acquired images were analyzed and exported to a movie with the same soft\vare. CLEM

CLEM was performed on sections prepared according to the HPF procedure. For LM observation, sections were observed under a Zeiss Axiovert 200 upright microscope with a X40 objective lens (NA 0.75; Carl Zeiss, Oberkochen, Germany) and photographed using a Zeiss AxioCam MRc digital camera (25). Micrographs of different optics were merged by using AxioVision 4.6 software (Carl Zeiss). A nickel grid with ultrathin sections was immersed in a drop of distilled water in the circled area of a Teflon printed glass slide (63424-06; Electron Microscopy Sciences). The "sandwich" was sealed with a coverslip to restrict the movement of the grid, whereas the Teflon layer served as spacer between the slide and the coverslip. Sections were first analyzed by fluorescence microscopy. Details of locations of the regions of interest (ROIs) were recorded on graph paper. After LM imaging, the coverslip was removed carefully, and sections were further processed with the gold-enlargement procedure. The sections were then air-dried and examined with a 120-kV transmission electron microscope (Tecnai T12; FEl, Hillsboro, OR, USA). According to the recorded location under LM, the same ROIs were traced back under EM and imaged at a series of magnifications. Electron micrographs were recorded with the Gatan 4K X 4K charge-coupled device camera (Gatan Corp., Pleasanton, CA, USA). Image correlation and analysis Light micrographs and electron micrographs of identical fields in the resin sections were manually adjusted to identical orientations and magnifications and merged by using AxioVision 4.6. RESULTS Experimental design

We made use of the fish medaka to establish the CLEM procedure for studying highly dynamic events and 579

Figure 1. Macromolecule labeling system for correlative microscopy. A) Amino acid sequence alignment between medaka Tom20 (ENSORLT00000001496) and httman homolog (NP_055580.1). B) Consütict pTom20gfp (toja left), which expresses Tom20:GFP, a fusion between the 34 N-termina] amino acid residues (underlined in panel A) of Tom20 and GFP. Tom20 is a component of the TOM complex on the OM of mitochondria. Tom20:GFP contains die TMD (asteiisks in panel A) and r'etains tlie ability to localize into mitochondria (bottom left) via integrating into tlie OM. Other components of the TOM complex are receptors Tom22 and Tom70, the chatinel protein Tom40, and small Tom proteins Tom5, Tom6, and Tom7.

m -i.pPPVFûatTKLFilSORi«AûsjE£)ci

B

Species % identity Medaka 100 Human 92

pTom20:GFP

OM

Mitochondrion

processes in early development stage embryos of vertebrates. To this end, we focused on mitochondria as the detecdon target. Cells of early embryos of medaka possess many mitochondria (Supplemental Eig. SI), which are distributed widely in the cytoplasm for easy observation by EM. We chose the medaka Tom20 as the targedng molecule for mitochondrial localization. The medaka Tom20 shares the same length of 145 amino acid residues and a 92% sequence identity value with its human homolog and contains a highly conserved TMD at the N terminus (Fig. 1.4). The N-terminal peptide of 35 amitio acid residues from the rat Tom20 is able to target the GFP fused to mitochondria in COS-7 cells (6). We adopted this approach and fused the N-termi-

TOM complex

nal peptide of 34 amino acid re.sidttes of the medaka Tom20 in-frame to GFP, resulting iti a fusion protein, Tom20:GFP, which should be capable of localization on the mitochondrial OM, thus making it visible by GFP fluorescence (Fig. IB). Tom20:GFP localization by LM

Tom20:GEP overexpression by zygotic mRNA microinjection did not interfere with embryonic development (Supplemental Fig. S2). Live blastomeres were dissociated at the midblasttila stage for fluorescence intensity analysis and time-lapse imaging. They were 30 — 50 |xm in diameter (Fig. 2A); H2B:Cherry was vflthin the

Figure 2. Intracellular Tom20: GFP distribution. Embryos at tlie 1-cell stage were injected with mRNAs for Tom20:GFP and H2B:Cherry. Blastomeres were dissociated at the blasatia stage for fluorescent microscopic observation. A) Die image of dissociated blastomeres. B—D) Stacks of fluorescence images, showing stibcelltilar distribution of H2B: Cherry (B) and Tom20:GFP (C) and merged images (D). E) Higher magnification of the left ceU in panel D, showing uneveti distribution of the signal. Fotir areas (I-IV) of the GFP signal are seen along the broken arrow. Areas lacking the fluorescent signal are circled, nu, nucleus. F) Fluorescence intensity in the 4 areas along the arrow shown in panel £ Scale bars =

580

Vol.28

February 2014

The FASEB lournal • www.fasebj.org

YUAN ET AL

nucleus (Fig. 2B), whereas Tom20:GFP was localized to the cytoplasm. (Fig. 2C). In the cytoplasm, Tom20:GFP distributed evenly except for 1 or 2 spherical areas of 10-15 |xm in diameter, which were essentially free of Tom20:GFP distribution (Fig. 2C, D). This finding is more evident at higher magnification, in which 4 areas were clearly seen according to Tom20:GFP distribution (Fig. 2E,F). The wide distribution of Tom20:GFP in the bulk cytoplasm is reminiscent of the mitochondrial distribution in blastomeres as revealed by EM (Supplemental Fig. SI). Time-lapse imaging revealed that Tom20:GFP-labeled mitochondria concentrated in the perinuclear cytoplasm in all blastomeres (Fig. 3). Notably, mitochondria were absent in psetidopodia and lacked any detectable redistributioti during nuclear cycle phases and cytokinesis (Supplemental Movies SI and S2). The MitoTracker signal was shown in individual channel in Supplemental Movie S2 to highlight the distribution of mitochondrial. After cotransfection of pTom20gfp and pCVpr into MESl, the expressed Tom20:GFP was found to concentrate in certain areas in the cytoplasm, whereas the fusion protein PR was distributed evenly in the cells (Fig. 4A). When pTom20gfp-transfected cells were stained with MitoTracker, a standard fluorescent dye to locate mitochondria, the Tom20:GFP signal was found to colocalize with MitoTracker staining with a Pearson correlation coefficient (PCC) of 0.95 (Fig. 4B). These results suggest that Tom20:GFP can efficiently localize to mitochondria in medaka FS cell culture. We extended our observation to developing me-

daka embryos. Blastomeres were dissociated from the tom20gfp RNA-injected embiyos until the midblastula stage and subjected to the HPF-FS sample preparation procedure mentioned above. The ultrathin sections were imtiuitiostained with anti-GFP atitibody (aGFP) and anti-Tom20 antibody (aTom20), respectively, and analyzed by fluorescence microscopy. The Tom20:GFP signal displayed a colocalized match to GFP immunostaining (PCC 0.93) in the cytoplasm surrounding the nucleus (Fig. 4C). Meanwhile, The Tom20:GFP signal and Tom20 immunostaining also exhibited a perfect match in spatial distribution (PCC 0.94; Fig. 4D). Taken together, the results show that Tom20:GFP retains its fluorescence and antigenicity for immunostaining after the multiple procedures necessary for CLEM. CLEM We furthered our experiments with CLEM. To this end, sections with particularfieldsof interest were identified by LM observation and photographed (Fig. 5A, B). Such samples were then examitied by TEM atid photographed at various magnifications. After adjustment to identical orientations and magnifications, light micrographs and electron micrographs of sections were merged, leading to a precise overlay between LM and EM. For example, the DAPI-stained nuclear signal from LM correlatively matched the electron-dense nuclei from EM (Fig. 5C). A precise correlation between LM and EM was also obtained for Tom20:GFPfluorescenceand mitochondria (Fig. 5D, D', E, E). This is most evident at a higher magnification, at which the fluorescent signal was clearly seen on easily

Figure 3. Live imaging of mitochondria in early embryos. Representative micrographs of Supplemental Movie SI, showing the behavior of Tom20:GFPpositive mitochondtia in dividing blastomeres. Medaka embryos at the 1-cell stage were injected with tom20:gfp mRNA. Embiyonic cells at the blástula stage were dissected and staitied for nuclei with Hoechst 33342 (bitte) before analysis by confocal microscopy. Arrowheads indicate psettdopodia; asterisks indicate anapliase cells; hashtags itidicate lnetaphase cells. Scale bars = 20 (j.m.

CORRELATIVE MICROSCOPY IN BLASTOMERES

581

Tom20:GFP

Figure 4. Mitochonddal locaüzaüon of the Tom20:CEP. A) Live ES cells at 24 b post-transfecüon with pTom20gfp (green) and pCVpr (red), ntt, nucletis. ii) Live ES cells after MitoTraeker staining at 24 h post-transfecdon with pTom20gfp, showing colocalization of the Tom20:GEP signal and mitocbondrial staining. Q Section of blastomeres showing colocalization of Tom20:CEP fluorescence (green) and GFP immttnostaining (red) in mitochondda surrounding the nucleus (blue by DAPI staining). D) Seetion of blastomeres showing eolocaüzaüon of Tom20:GFP fluorescence (green) and Tom20 immunostaining (red) in mitochondda surrounding the nucleus. Scale bars = 5 ixm {A, B); 20 jxm (C, D).

u

Tom20:GFP

Mitoehondrial coloealization of Tom20:GFP and Tom20 The experiments described so far demonstrate specific localization of the Tom20:GEP fusion protein on mitochondria by fluorescence microscopy and CLEM. We asked whether Tom20:GEP could mimic endogenous Tom20 in localizadon on the mitochondrial OM. To this end, ultrathin secdons were immunostained with aGEP for tom20g£p RNA-injected samples and with ctTom20 for noninjected control embryos, the signals were detected by using nanogold-conjugated secondary antibodies, and gold pardcles were enhanced from 1.4 to 10-15 nm in diameter for unambiguous resolution by TEM. The gold Vol.28

February 2014

merge

LU

identifiable mitochondria but was essendally absent on the cell membrane and nuclear membrane (Eig. 5F, F). Therefore, the Tom20:GEP fusion protein is able to localize specifically into mitochondria, and reladvely large cells from early developing embryos can preserve their ultrastructures after muldple procedures of sample p r e p aradon necessary for CLEM.

582

MtTrackei

particles of aGEP staining were found in association with mitochondria (Fig. 6A). At higher magnifications, the particles were clearly seen on the mitochondrial OM (Eig. 6B) and were essentially absent in the inner membrane, cristae, and matrix of mitochondria (Eig. 6C—F). Similarly, the nanogold particles of aTom20 staining were found specifically on mitochondrial OM (Eig. 6G-J). Taken together, these resttlts indicate that Tom20:GEP resembles Tom20 in the ability for specific localization onto the mitochondrial OM, and the procedures established are a powerful tool for live cell imaging, CLEM, and precise localization of target proteins by immunostaining in large cells of early development stage embryos.

DISCUSSION In this study, we prepared samples of large cells of early fish embryos that allowed for visualization of endogenous and introduced proteins, which is compatible with xisualization of GEP-tagged proteins by LM and resolu-

The FASEB Journal • www.fasebj.org

YUAN ET AL.

Figure 5. Sttbcellular localizadon of Tom20:GEP by correlative LM and TEM. Embryos at the 1-cell stage were injected with the Tom20:GFP mRNA. Blastomeres were di.ssociated at the blástula stage and fixed by HPE. Ultrathin sections after uuclear staiuiug with DAPI (blue) were transferred to a gdd for analysis by LM and EM. A, B) Low-magnification micrographs oftheTom20:GEP signal (A) and nuclear staitiing {B). Q Merge between tlie fluorescence micrograph (DAPI) and the dausmission electron micrograph, showing the location of nuclear staining. D, D, E, E) Electron micrographs {D, £) and merged fluorescence and electron micrographs (£)', E) showing magnified views of the boxed areas d and e in panel G F, F) Higher magnification of the boxed areas in panels E and E, highlighting the localization of Tom20:GEP in mitochondria (mt) surrounding the nucleus (nu). cm, cell membrane; nm; nuclear membrane. Scale bars = 10 iJim {A-Q; 2 ixm

{D,D',E,E}; 1 |xm {F,F). tion of ultrastructures by EM. To adopt the CLEM procedures for the analysis of molecular events in early medaka embryos, we chose mitochondria as a model system. These organelles are very abundant in early

C

medaka embryos and can be easily visualized in living cells after vital staining with MitoTracker. Many macromolecules are known to localize in different compartments and thus provide excellent sys-

OM D cr

/ ma OM

ma

CORRELATIVE MICROSCOPY IN BLASTOMERES

Figure 6. Electron micrographs showng mitochonddal localization of Tom20:GEP. Ultrathin sections of blastomeres were incubated wth primary antibodies against GEP or Tom20, followed by inctibation with secondary antibodies conjttgated with nanogold ( 1.4 nm). The nanogold was enlarged to 10—15 nm with the gold enhancement reagent. A—F) Immunostaining of GEP iu embryos injected with tom20:gfp mRNA, showing numerous mitochoudda (mt) aud Tom20:GEP localization on the mitochoudrial OM. Panels show 6 individual examples, nu, nucleus. G^/) Immunostaining of Tom20 in control embryos, showing localization of the Tom20 signal in particular regions on the mitochondria OM. Panels show 4 individual examples, cr, cristae; ma, matdx. Scale bars = 2 ixm (A); 200 nm (B); 100 nm

583

tems for tracking d)'namic events of molecular trafficking. In this study, we have chosen medaka Tom20 to construct the Tom20:GFP fusion protein as a tagged reporter for real-time visualization of mitochondrial localization in living cells. Three lines of evidence estabhsh that this fusion protein is a suitable marker for live imaging of localization on the mitochondrial OM. First, the Tom20:GFP fluorescence and MitoTracker staining match in medaka ES cell culture and embryonic cells. Second, immunostaining of Tom20:GFP also overlaps with immunostaining of endogenous Tom2(). Finally, immunoelectron microscopy reveals the precise localization of the introduced Tom20:GFP and endogenous Tom20 on the mitochondrial outer membrane. These results conform to the observation that the N-terminal SP of Tom20 acts as a signal-anchored domain and is sufficient for intracellular sorting and anchoring to the OM (5, 6). The procedure introduced here offers advantages for studying cellular and molecular events of early development stage embryos. A GFP-labeled molecule can be visualized in live embryos and then fast-frozen within seconds and processed for visualization of the same molecule in a sectioned sample by LM and EM. Therefore, large and fragile blastomeres offish are amenable to the CLEM procedures, and the mt-targeting Tom20: GFP offers an excellent system for studying dynamic processes of macromolecular events at a high resolution. The CLEM procedures established in this study will be applicable to detection of both endogenous and GFP-tagged proteins in many structures of early developing embiyos of vertebrates. One interesting finding in our results is the lack of mitochondrial distribution in pseudopodia as well as inconspicuous extensive redistribution during divisions. It remains unclear whether this is a common feature in blastomeres offish species, but mitochondria are rich in pseudopodia in guinea pig trophoblast cells (32) and human gliomas (33), and an apparent redistribution of mitochondria occurs during nuclear and cytoplasmic divisions of animals (34) and plants (35).

Here we reported that the large and fragile blastomeres of fish are amenable to HFP fixation and subsequent correlative microscopy procedures. The mt-targeting Tom20;GFP offers an excellent system for studying the dynamic processes of macromolecular events at a high resolution. Fltiorescence-labeled blastomeres are cryofixed via HPF and FS. Ultrathin sections of embedded samples are transferred to grids and immunostained with nanogold-conjugated antibodies against the fltiorescence tag or endogenous proteins. After fluorescence imaging, the same sections are subjected to a gold enlargement step for TEM. With this procedure, we can monitor the localization at nanometer resolution of the mitochondrial protein Tom20 and GFPtagged Tom20 on the outer membrane in blastomeres

584

Vol.28

February 2014

REFERENCES L

2.

3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

CONCLUSIONS

of the fish medaka as a vertebrate model.

The atuhors thank J. R. Deng for fish breeding and C. M. Foong for laboratory management. This work was supported by the National Research Foundation Singapore (NRF-CRP72Ö10-03). Author contributions: Y.Y. and Y.H. designed the experiment and prepared the manuscript; Y.Y., M.L., and N.H. performed the research work. The authors declare no conflicts of interest.

IFTI

13.

14.

15.

16.

17.

Pfanner, N., Wiedemann, N., Meisinger, C , and Lithgow, T. (2004) Assembling the mitochondrial onter membrane. Nat. Struct. Mol. Biol. U , 1044-1048 Hoogenraad, N.J., Ward, L. A., and Ryan, M. T. (2002) Import and assembly of proteins into mitoehondria of mammalian eells. Biochim. Biophys. Acta 1592, 97-105 Pfanner, N., and Geissler, A. (2001) Versatility of the mitoehondrial protein import machinery. Nat. Rev. Mot. Cell Biol. 2, 3.S9-349 Endo, T., and Yaniano, K. (2010) Transport of proteins across or into the mitoehondrial onter membrane. Biochim. Biophys. Acta 1803, 706-714 Waizenegger, T., Stan, T., Neupert, W., and Rapaport, D. (2003) Signal-anchor domains of proteins of the outer membrane of mitochondria: structural and functional charaeteristies. / Biol. Chem,. 278, 42064-42071 Kanaji, S., Iwahashi, J., Kida, V., Sakagnehi, M., and Mihara, K. (2000) C^liaraeterizaiion of the signal that directs Tom20 to the mitoehondrial outer m e m b r a n e . / Cell Biol. 151, 277-288 Giepmans, B. N. (2008) Bridging fluorescence microscopy and electron microscopy. Histochem. Cell Biol. 130, 211-217 Shu, X., Lev-Ram, V., Deerinck, T. J., Qi, Y., Ramko, E. B., Davidson, M. W.,jin,Y., Ellisman, M. H., andTsien, R. Y. (2011) A genetically encoded tag for correlated light and electron microscopy of intact cells, tissues, and organisms. PLoS Biol. 9, el001041 Van Rijnsoever, C , Oorschot, V., and Klumperman, J. (2008) Correlatix'e lighl-electron microscopy (CLEM) combining livecell imaging and immunolabeling of ultrathin ciyosections. Nat. Methods 5, 973-980 Muller-Reichert, T., Srayko, M., Hyman, A., O'Toole, E. T., and McDonald, K. (2007) Correlative light and electron microscopy of early Caenorhabditis elegans embryos in mitosis. Methods Cell Biol. 79, 101-119 McDonald, K., and Morphevv, M. K. (1993) Improved preservation of ukrastructure in clifficult-to-fix organisms by high pressure freezing and freeze substitution: I. Drosophila melanogaster and Stron^'locentrotus purpuratus embryos. Microsc. Res. Tech. 24, 465-473 ' Vanhecke, D., Graber, W., and Studer, D. (2008) Close-to-native iiltrastructural preservation by high pressure freezing. Methods Cell BioL SS, ÏOI-1&4 He. W., Ladin.sky, M. S., Huey-Tubman, K. E., Jensen, G. J., Mclntosh, J. R., and Bjorkman, P. J. (2008) FcRn-mediated antibody transport across epithelial cells revealed by electron tomography. Nature 455, 542-546 Brown, E., Mantell, J., Carter, D., Tilly, G., and Verkade, P. (2009) Studying intracellular transport using high-pressure freezing and correlative light electron microscopy. Semin. Cell Dev. Bi'oL 20, 910-919 Kukulski, W., Schorb, M., Welsch, S., Picco, A., Kaksonen, M., and Briggs, J. A. (2011) Correlated fluorescence and 3D eleetron microscopy with high sensitivitj' and spatial precision./ Cell Biot. 192, 111-119 Nixon, S. J., Webb, R. I., Floetenmeyer, M., Schieber, N., Lo, H. P., and Parton, R. G. (2009) A single method for ciyofixation and correlative light, electron mieroscopy and tomography of zebrafish embryos. Traffic 10, 131-136 Hong, Y., Liu, T., Zhao.' H., Xu, H., Wang, W., Liu, R., Chen, T., Deng,J., and Gui,J. (2004) Establishment of a normal medakafish spermatogonial cell Une capable of sperm production in vitro. Proc. Nati Acad. Sei. U. S. A. 101, 8011-8016

The FASEB Journal • www.fasebj.org

YUAN ET AL

18.

19. 20.

21.

22.

23. 24.

25.

26.

27.

Hong, Y., Winkler, C , and Schartl, M. (1998) Production of medakafish chimeras from a stable embryonic stem cell line. Proc. Nati Acad. Sei. U. S. A. 95, 3679-3684 Yl, M., Hong, N., and Hong, Y. (2009) Generation of medaka fish haploid embryonic stem cells. Science 326, 430-433 Hong, N., Li, M., Zeng, Z., Yi, M., Deng, J., Gui, j . . Winkler, C , Schartl, M., and Hong, Y. (2010) Accessibility of host cell lineages to medaka stem cells depends on genetic background and irradiation of recipient embryos. Cell Mol Life Sei. 67, 1189-1202 Yl, M., Hong, N., Li, Z., Yan, Y, Wang, D., Zhao, H., and Hong, Y (2010) Medaka fish stem cells and their applications. Sei. Ghina Life Sei. 53, 426-434 Hong, N., He, B. P., Schartl, M., and Hong, Y. (2013) Medaka embryonic stem cells are capable of generating entire organs and embryo-like miniatures. Stem Gells Dexi. 22, 750-757 Wittbrodt, J., Shima, A., and Schard, M. (2002) Medaka—a model organism from the far East. Nat. Rev. 3, 53-64 Herpin, A., Rohr, S., Riedel, D., Kluever, N., Raz, E., and Schartl, M. (2007) Specification of primordial germ cells in medaka (Oryzias latipes). BMG Dev. Biol 7, 3 Li, M., Hong, N., Xu, H., Yr, M., Li, G., Gui, J., and Hong, Y (2009) Medaka vasa is reqtrired for migration but not survival of primordial gei'm cells. Mech. Dev. 126, 366-381 Hong, N., Ghen, S., Ge, R., Song, [., Yi, M., and Hong, Y (2012) Interordinal chimera formation between medaka and zebrafish for analyzing stem cell differentiation. Stem Gelts Dev. 21, 23332341 Liu, L., Hong, N., Xu, H., Li, M., Yan, Y, Purwanti, Y, Yi, M., Li, Z., Wang, L., and Hong, Y (2009) Medaka dead end encodes a

CORRELATIVE MICROSCOPY IN BLASTOMERES

28.

29. 30.

31.

32. 33.

34. 35.

cytoplasmic protein and identifies embryonic and adult germ cells. Gene Expr. Patterns 9, 541-548 Zhao, H., Hong, N., Lu, W., Zeng, H., Song, J., and Hong, Y (2012) Fusion gene vectors allowing for simultaneous dr'ug selection, cell labeling, and reporter assay in vitro and in vivo. Anal Ghem. 84, 987-993 Hong, Y., Schard, Manfred. (1996) Establishment and growth responses of early medakafish (Oryzias Uitipes) embryonic cells in feeder layer-free cultures. Mol Mar. Biol Biotechnol 5, 93-104 Hong, Y, Ghen, S., Gui, J., and Schartl, M. (2004) Retention of the de\'elopmental pluripotency in medaka embryonic stem cells after gene transfer and long-term drug selection for gene targeting in fish. Transgenic Res. IS, 41-50 Yl, M., Hong, N., and Hong, Y. (2010) Derivation and characterization of haploid embryonic stem cell cultures in medaka fish. Nal Protoc. 5, 1418-1430 Garris, D. R. (1984) tJltrastructural aspects of the appositional stage of interstitial blastocyst implantation. Qynecol Obstet. Invest. 17, 10-17 Ar-ismendi-Morillo, G., Hoa, N. T., Ge, L., and Jadus, M. R. (2012) Mitochondr-ial network in glioma's invadopodia displays an activated state both in situ and in vitro: potential firnctional implications. Ultrastruct. Pathol. 36, 409-414 Yamano, K., and Youle, R.J. (2011) Goupling mitochondrial and cell diráion. Nal Gell Biol 13, 1026-1027 Nebenfuhr, A., Frohlick, J. A., and Staehelin, L. A. (2000) Redistribution of Golgi stacks and other or-ganelles during mitosis and cytokinesis in plant cells. Plant Physiol 124, 135-151 Received for publication April 19, 2013. Accepted for publication October 3, 2013.

585

Copyright of FASEB Journal is the property of Federation of American Society for Experimental Biology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.