Simultaneous Single-Cell Measurements Demonstrate a Positive Correlation between RNA Copy Number for Mitochondrial Division and Fusion Genes and Mitochondrial Fragmentation

Abstract

Mitochondria are some of the most highly dynamic organelles in eukaryotic cells. The spatial distribution and morphology of mitochondria undergo constant changes via division and fusion to adapt to cellular conditions and to maintain various biological processes. However, how mitochondrial division and fusion are coordinated within a cell is currently unclear. Here, we performed microscopic observation using single-molecule fluorescence in situ hybridization (smFISH) combined with immunostaining to simultaneously measure RNA molecules transcribed from mitochondrial division or fusion genes and the number of mitochondria in an individual cell. The RNA copy number in a single cell varied widely; with up to a 6–10 fold difference among cells. Notably, the expression levels of both types of genes seemed to be positively correlated with the number of fragmented mitochondria. Moreover, multiplexed smFISH imaging confirmed that the expression levels of eight species of mitochondrial division and fusion genes were significantly correlated in individual cells. These findings suggest that the increased expression of mitochondrial division/fusion genes is likely to trigger the homogenization of mitochondrial component molecules such as DNA, RNA, proteins, and lipids among mitochondria through structural remodeling via frequent division and fusion.

Mitochondria, which originated from an ancestral free-living bacterium (Gray 1992, Gillham et al. 1994), supply the chemical energy needed for many types of biological processes in eukaryotic cells via the production of adenosine triphosphate (ATP). The intracellular distribution of mitochondria is a highly dynamic process in which mitochondria constantly divide and fuse to maintain cellular functions (Fig. 1a). Several factors needed for mitochondrial division and fusion have been identified in human (Fig. 1b) (Osteryoung and Nunnari 2003, Kuroiwa et al. 2006, Friedman and Nunnari 2014, Roy et al. 2015). The dynamin-related GTPase protein DRP1 (also known as DNM1 in yeast), which is a key component of the mitochondrial division machinery, localizes to the outer mitochondrial membrane at the division site and assembles into a ring-like structure (Bleazard et al. 1999, Nishida et al. 2003, Mears et al. 2011). The outer mitochondrial membrane protein FIS1 also functions in mitochondrial division by serving as a receptor that tethers DRP1 on the mitochondrial division site of the outer mitochondrial membrane (Mozdy et al. 2000). Additional factors for mitochondrial division have also been identified in animals, such as MFF and MIEF1/2 (Gandre-Babbe and van der Bliek 2008, Otera et al. 2010, Palmer et al. 2011, Losón et al. 2013, Liu and Chan 2015). A series of studies have shown that mitochondrial division is a highly organized process involving many types of protein molecules. In many eukaryotes including higher organisms, mitochondria undergo fusion as well as division. The mitochondrial fusion machinery consists of large GTPases such as MFN1, MFN2, and OPA1 (Shepard and Yaffe 1999, Wong et al. 2000, Olichon et al. 2003, Griparic et al. 2004). MFN1 and MFN2 contain two transmembrane regions and localize to the outer mitochondrial membrane, where their GTPase domain is exposed to the cytosolic side. By contrast, OPA1, a dynamin-related GTPase, is present in the inner membrane space.

Fig. 1. Mitochondrial dynamics via division and fusion. (a) Schematic representation of mitochondrial dynamics via division and fusion. (b) Responsible genes for mitochondrial division and fusion in human. DRP1, FIS1, MFF, MIEF1, and MIEF2 are involved in mitochondrial division, whereas MFN1, MFN2, and OPA1 are involved in mitochondrial fusion. (c) Schematic representation of the imaging technique involving smFISH combined with immunostaining. Target RNA molecules were visualized with fluorescence-labeled probes and mitochondria were immunostained.

The loss of function of mitochondrial division genes causes excess fusion of mitochondria, leading to the presence of long, interconnected mitochondria (Hoppins 2014, Roy et al. 2015). By contrast, the loss of function of mitochondrial fusion genes leads to excess divisions, with extensive fragmentation of the mitochondria (Hoppins 2014, Roy et al. 2015). However, it is unclear how the mitochondrial division and fusion processes are coordinately regulated in the cell. Whereas conventional transcriptomic approaches such as DNA microarray analysis, RNA-seq, and ribosomal profiling can reveal the average number and trends of gene expression profiles in cells, emerging microscopy technology can be used to detect various types of molecules at a single-molecule resolution in a single cell, opening new avenues for biological research. To quantify RNA expression levels in an individual cell, smFISH has been developed to identify the copy number and spatial distribution of an RNA molecule of interest (Raj et al. 2008). The fundamental principle of smFISH is based on the hybridization of fluorophore-labeled oligonucleotide probes to native RNA molecules. After hybridization, the RNA molecules can be observed via highly sensitive imaging using single-molecule microscopy. To date, many types of smFISH imaging methods have been developed, leading to the quantitative measurement of hundreds to thousands of RNA molecules in various organisms and tissues (Lubeck et al. 2014, Chen et al. 2015, Coskun and Cai 2016, Shah et al. 2016). Thus, smFISH imaging has helped uncover gene expression profiles in specific tissues and cells that function in many biological processes.

In this study, we investigated the expression levels of the responsible genes for mitochondrial division and fusion in single cells while simultaneously visualizing mitochondrial morphology by combining smFISH and immunostaining techniques. Our results shed light on the mechanism coordinating mitochondrial division and fusion at the single-cell level.

Materials and methods

Cell culture and fixation

U2-OS cells were cultured on 35 mm φ glass bottom dishes (Matsunami) in FluoroBrite DMEM (ThermoFisher Scientific) supplemented with 10% fetal bovine serum and 1×GlutaMAX-I (GIBCO) and incubated at 37°C with 5% CO₂ for 48 to 72 h. The cultured cells were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 15 min and washed with PBS and methanol. The fixed cells were immediately transferred to a freezer (−20°C) until use.

Probe design

Primary probes contained two oligonucleotide regions: (i) a 20-nt (or 35-nt for SCRN2) priming site on the 5′ side that could specifically hybridize to the target sequence of each mRNA molecule and (ii) a 40-nt site on the 3′ side consisting of two repetitive priming sites and to allow for an additional hybridization with Cy5-labeled secondary probes. Twenty-nt priming sites on the 5′ side were designed using the probe design tool Stellaris probe designer (https://www.biosearchtech.com/stellaris-designer). For each gene, 6–9 primary probes were designed, depending on the gene sequence length (Table 1). Secondary probes were synthesized and labeled with Cy5 at the 5′ ends at Fasmac (Table 2). “/5Cy5/” indicates a 5′ Cy5 modification.

Table 1. Oligonucleotide sequences of primary probes.

Probe #	Target gene	5′-priming site (20nt) for target RNA	3′-priming site (40nt) for Cy5-probe	Complementary Cy5 probe
1	DNM1	cactacgacgatttgaggca	cgtaaacggccacaagttcacgtaaacggccacaagttca	Cy5-probe 1
2		tttctagcactgagctcttt	cgtaaacggccacaagttcacgtaaacggccacaagttca
3		tgacaattccagtacctctg	cgtaaacggccacaagttcacgtaaacggccacaagttca
4		gttgcagaatgagaggtctc	cgtaaacggccacaagttcacgtaaacggccacaagttca
5		gttgttttccgtttatcttc	cgtaaacggccacaagttcacgtaaacggccacaagttca
6		gaaatttaccccattcttct	cgtaaacggccacaagttcacgtaaacggccacaagttca
7		gcttactcccttattatttc	cgtaaacggccacaagttcacgtaaacggccacaagttca
8		gctcaatatccttaggttga	cgtaaacggccacaagttcacgtaaacggccacaagttca
9		ccgaagaatgagctctctga	cgtaaacggccacaagttcacgtaaacggccacaagttca
10	MFF	cagaatctgatctgtgccac	cgacttcttcaagtccgccacgacttcttcaagtccgcca	Cy5-probe 2
11		cagaaatcggtgcctggaac	cgacttcttcaagtccgccacgacttcttcaagtccgcca
12		gatggtgtcacagtgtactc	cgacttcttcaagtccgccacgacttcttcaagtccgcca
13		ggtaacatatggggaggaca	cgacttcttcaagtccgccacgacttcttcaagtccgcca
14		gcagaggaaagattagctcc	cgacttcttcaagtccgccacgacttcttcaagtccgcca
15		catccaagatctgctggtat	cgacttcttcaagtccgccacgacttcttcaagtccgcca
16	FIS1	tactcaaactgcgtgctctt	agcgcaccatcttcttcaagagcgcaccatcttcttcaag	Cy5-probe 3
17		cacgatgcctttacggatgt	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
18		aggtagaagacgtaatcccg	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
19		tattccttgagccggtagtt	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
20		cggacgtactttaaggcctt	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
21		ttcatggccttgtcaatgag	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
22	MIEF1	atcgtacatccgcttaactg	ctgaagggcatcgacttcaactgaagggcatcgacttcaa	Cy5-probe 4
23		gagtcgtgacttcttcaagt	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
24		cggttccggtagtaagtaag	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
25		catatgtccacagcagcttg	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
26		ctcaagtacatgtcccgaag	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
27		acctgcaggtcatcgtagag	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
28		gagttggatgtggtcagctg	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
29		tactctggattctcacgacg	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
30		tacacagcggtcccagtaac	ctgaagggcatcgacttcaactgaagggcatcgacttcaa
31	MIEF2	tgtcaatgaaccgcttcacg	cacaacgtctatatcaccgccacaacgtctatatcaccgc	Cy5-probe 5
32		ctcaggatcagtttctgcag	cacaacgtctatatcaccgccacaacgtctatatcaccgc
33		aacttgctccgaaagtaggc	cacaacgtctatatcaccgccacaacgtctatatcaccgc
34		agaactcaagctgcgtcctg	cacaacgtctatatcaccgccacaacgtctatatcaccgc
35		cagttgacagaagcagccag	cacaacgtctatatcaccgccacaacgtctatatcaccgc
36		acaagcacagccacagtgag	cacaacgtctatatcaccgccacaacgtctatatcaccgc
37		aagctgctgaagaggttcac	cacaacgtctatatcaccgccacaacgtctatatcaccgc
38		tagccaatgtcgtcaatctc	cacaacgtctatatcaccgccacaacgtctatatcaccgc
39		tcctgtaggccactgtatag	cacaacgtctatatcaccgccacaacgtctatatcaccgc
40	MFN1	ttcctgtatgttgcttcaac	atcaaggccaacttcaagatatcaaggccaacttcaagat	Cy5-probe 6
41		ctcggatgctattcgatcaa	atcaaggccaacttcaagatatcaaggccaacttcaagat
42		gaagcagttggttgtgtgac	atcaaggccaacttcaagatatcaaggccaacttcaagat
43		cggtcataaggtaggcttta	atcaaggccaacttcaagatatcaaggccaacttcaagat
44		acagtcttcacacttttctt	atcaaggccaacttcaagatatcaaggccaacttcaagat
45		ctctgtggtgacatctgtac	atcaaggccaacttcaagatatcaaggccaacttcaagat
46		tgttcatcagtgttgattcc	atcaaggccaacttcaagatatcaaggccaacttcaagat
47		tgaagatgttgggcttggag	atcaaggccaacttcaagatatcaaggccaacttcaagat
48		aagcatcccaacggttattc	atcaaggccaacttcaagatatcaaggccaacttcaagat
49	MFN2	ttggcagtgacaaagtgctt	gacaaccactacctgagctagacaaccactacctgagcta	Cy5-probe 7
50		acgtgtcttcaaggaaggtg	gacaaccactacctgagctagacaaccactacctgagcta
51		aacctgttcttctgtggtaa	gacaaccactacctgagctagacaaccactacctgagcta
52		ggataggtaacctttgacgt	gacaaccactacctgagctagacaaccactacctgagcta
53		aaaagccactttcatgtgcc	gacaaccactacctgagctagacaaccactacctgagcta
54		cagagcatggcattgatcac	gacaaccactacctgagctagacaaccactacctgagcta
55		aatcccagagggcagaactt	gacaaccactacctgagctagacaaccactacctgagcta
56		caggaagcaattggtggtgt	gacaaccactacctgagctagacaaccactacctgagcta
57		acatcacactcactaggctg	gacaaccactacctgagctagacaaccactacctgagcta
58	OPA1	tgcttcgtgaaaccagatgt	gtcctgctggagttcgtgacgtcctgctggagttcgtgac	Cy5-probe 8
59		aaaaattcctgcgaggctgg	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
60		aaggtccggtatcatatctt	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
61		ctcatcaatttcccacacaa	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
62		atcttgtcaaagtctggtgc	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
63		actttcagatccacgatctg	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
64		ctgagtgtgcagaagttctt	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
65		gatgaatgcctttgtcatct	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
66		tgggcaatcatttccaacac	gtcctgctggagttcgtgacgtcctgctggagttcgtgac
67	GAPDH	ccaaatccgttgactccgaccttcaccttccccat	cgtaaacggccacaagttcacgtaaacggccacaagttca	Cy5-probe 1
68		taaaagcagccctggtgaccaggcgcccaatacga	cgtaaacggccacaagttcacgtaaacggccacaagttca
69		gtcattgatggcaacaatatccactttaccagagt	cgtaaacggccacaagttcacgtaaacggccacaagttca
70		aacatgtaaaccatgtagttgaggtcaatgaaggg	cgtaaacggccacaagttcacgtaaacggccacaagttca
71		tgccatggaatttgccatgggtggaatcatattgg	cgtaaacggccacaagttcacgtaaacggccacaagttca
72		cacccatgacgaacatgggggcatcagcagagggg	cgacttcttcaagtccgccacgacttcttcaagtccgcca	Cy5-probe 2
73		gatcttgaggctgttgtcatacttctcatggttca	cgacttcttcaagtccgccacgacttcttcaagtccgcca
74		gctaagcagttggtggtgcaggaggcattgctgat	cgacttcttcaagtccgccacgacttcttcaagtccgcca
75		taccaaagttgtcatggatgaccttggccaggggt	cgacttcttcaagtccgccacgacttcttcaagtccgcca
76		gatggcatggactgtggtcatgagtccttccacga	cgacttcttcaagtccgccacgacttcttcaagtccgcca
77		tgacacgttggcagtggggacacggaaggccatgc	agcgcaccatcttcttcaagagcgcaccatcttcttcaag	Cy5-probe 3
78		gcaggtttttctagacggcaggtcaggtccaccac	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
79		cctgcttcaccaccttcttgatgtcatcatatttg	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
80		gtagcccaggatgcccttgagggggccctccgacg	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
81		ctgttgaagtcagaggagaccacctggtgctcagt	agcgcaccatcttcttcaagagcgcaccatcttcttcaag
Probe #	Target gene	Sequence*		Complementary Cy5 probe
82	N.A.	NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN		N.A.

*N indicates any of nucleotide (a or g or c or t)

Table 2. Oligonucleotide sequences of secondary probes.

Probe name	Oligonucleotide sequence
Cy5-probe 1	/5Cy5/TTTTtgaacttgtggccgtttacg
Cy5-probe 2	/5Cy5/TTTTtggcggacttgaagaagtcg
Cy5-probe 3	/5Cy5/TTTTcttgaagaagatggtgcgct
Cy5-probe 4	/5Cy5/TTTTttgaagtcgatgcccttcag
Cy5-probe 5	/5Cy5/TTTTgcggtgatatagacgttgtg
Cy5-probe 6	/5Cy5/TTTTatcttgaagttggccttgat
Cy5-probe 7	/5Cy5/TTTTtagctcaggtagtggttgtc
Cy5-probe 8	/5Cy5/TTTTgtcacgaactccagcaggac

smFISH imaging and immunostaining

Fixed cells were briefly washed once with ice-cold PBS, permeabilized for 5 min with 0.25% v/v Triton X-100 in PBS at room temperature, and washed three times with ice-cold PBS. The cells were incubated for 5 min in wash buffer comprising 2×saline-sodium citrate buffer (SSC) (Sigma) and 20% formamide (Nacalai tesque) and hybridized with 200 µL of 10 nM of the primary probes in hybridization buffer comprising 2×SSC, 20% formamide, 0.2 mg mL⁻¹ Bovine serum albumin, 100 mg mL⁻¹ dextran sulfate (Sigma), 2 mM vanadyl ribonucleoside complex (Sigma), and 1 mg mL⁻¹ Escherichia coli tRNA (Sigma) in a humid chamber inside a 37°C-hybridization oven for 12 h. Following hybridization, the samples were washed twice with wash buffer, 30 min per wash, followed by two washes with 2×SSC at room temperature. Each sample was postfixed in 4% paraformaldehyde in 2×SSC at room temperature for 30 min. The sample was then washed three times with 2×SSC and washed once with wash buffer. Secondary hybridization was performed as follows. Two hundred microliters of 2 nM of the appropriate fluorescently labeled readout probe in hybridization buffer were gently added to the glass bottom region of a dish and spread uniformly. The samples were incubated in a humid chamber inside a 37°C-hybridization oven for 12 h. For sequential smFISH imaging, secondary probes were hybridized for 30 min in 24°C. The samples were washed twice with wash buffer (30 min per wash), twice with 2×SSC, and once with PBS at room temperature. Subsequently, the samples were incubated in blocking solution (Blocking One, Nacalai tesque) for 30 min at 4°C, washed once with blocking solution, and immunostained with anti-TOM20 antibody (1 : 50 dilution, Santa Cruz Biotechnology, F-10) in blocking solution for 1 h at 4°C. The samples were washed twice with blocking solution and immunostained with Alexa Fluor 555 donkey anti-mouse IgG (1 : 1000 dilution, ThermoFisher Scientific) in blocking solution, followed by two washes with PBS at room temperature. Imaging was performed in PBS. As a negative control in smFISH imaging, the cells were hybridized with 75-nt random oligo DNAs for primary hybridization.

Single-molecule microscopy

Fluorescence images were obtained using an Olympus IX83 inverted microscope. The illumination was provided by a 514 nm Argon ion laser (Coherent, Innova 70C) and 640 nm diode lasers (CNI Lasers, MRL-III-640). The laser lights were homogenized using a rotating glass diffuser and introduced into the microscope for wide-field illumination via an objective lens (Olympus, PLAPON 60×, numerical aperture=1.42, oil). The samples were observed through an excitation filter (Semrock, FF02-628/40-25 for 640 nm laser), a dichroic mirror (Semrock, Di02-R514, and FF660-Di02 for 514 and 640 nm laser, respectively), and emission filter (Semrock, FF01-542/27-25 for 514 and Chroma, ET700/75 m for 640 nm laser). Images were acquired with an electron-multiplying charge-coupled device camera (Andor, iXon 897). The effective pixel size was 208 nm.

Image processing and analysis

For smFISH imaging, Z-stacks of 50 images were taken every 100 nm. Thirty sequential images were extracted from the Z-stack images and processed using a rolling-ball background subtraction algorithm (FIJI) (Schindelin et al. 2012) with a 10 pixel radius, the images were also processed using a mean filter with a 0.5 pixel radius. Subsequently, a maximum-intensity Z-projection of these images was generated. The projection image was processed using a rolling-ball background subtraction algorithm with a 5 pixel radius. Fluorescent spots of RNA molecules were identified and localized in each image as follows: 1) the image was thresholded at the top 0.5% to identify potential RNA spots distinguished from the background that originated from nonspecific hybridization of Cy5-labeled probes; 2) the image was processed using the Bernsen method of auto local threshold algorithm (FIJI) with a 15 pixel radius; and 3) the resulting spots, which were considered to represent RNA molecules, were counted in individual cells. For mitochondrial imaging, Z-stacks of 50 images were also taken every 100 nm, 30 sequential images were extracted and processed using a rolling-ball background subtraction algorithm with a 50 pixel radius, and a maximum-intensity Z-projection image was generated. After image thresholding using the Bernsen method of auto local threshold algorithm with a 15 pixel radius, segmented signal regions that were visualized as more than five pixels were distinguished as individual mitochondria.

Results

To investigate the expression levels of mitochondrial division and fusion genes in individual human cultured cells, we performed RNA imaging via smFISH using a two-step labeling scheme. First, intracellular RNA molecules were hybridized with a set of primary probes containing two oligonucleotide regions: (i) a 20-nt RNA targeting sequence and (ii) a 40-nt sequence containing two repetitive readout sequences (Fig. 1c, see materials and methods). Cy5-labeled secondary probes were then hybridized to the readout sequences of the primary probes, allowing target RNA molecules to be visualized in a single cell (Figs. 1c and 2a). After a series of probe hybridizations, we also performed immunostaining to visualize mitochondria using an antibody against the mitochondrial outer membrane protein TOM20. As a result, we simultaneously observed RNA molecules expressed from a single mitochondrial division/fusion gene and the shapes of the mitochondria in a single cell (Fig. 2).

Fig. 2. Measurement of RNA molecules and visualization of mitochondrial distribution in U2-OS cells. smFISH and immunofluorescence images of a human cultured U2-OS cell. GAPDH (a) or DRP1 RNA molecules (b) are shown as dots. Mitochondria were immunostained using a mitochondrial marker antibody (anti-TOM20 antibody). Oligo DNA probes for GAPDH were hybridized as primary probes as an experimental control in smFISH imaging. Arrowheads indicate transcription sites. Outlines show cell shapes, which were defined by phase contrast imaging. Scale bars: 10 µm. (c) Boxplots of the copy numbers of DRP1, FIS1, MFN1 and MFN2 RNA in single cells. Green and magenta dots indicate cells containing low (green) or high copy numbers of RNA (magenta). Mitochondrial morphology versus RNA abundance in cells. Average size (d) and average number (e) of fragmented mitochondria in cells containing low (green) or high (magenta) copy numbers of RNA for DRP1, FIS1, MFN1, or MFN2 shown in (c). Error bars represent S.E.M. Data are from 12 cells containing low or high copy numbers of RNA.

Using these imaging approaches, we investigated the RNA expression levels of two mitochondrial division genes, DRP1 and FIS1, in 29 and 21 cells, respectively (Fig. 2b, c). Simultaneously, we recorded the mitochondrial morphology in each cell. To detect all fluorescent signals corresponding to target RNA molecules in a single cell, z-stacks of 50 images were taken every 100 nm, and 2D maximum projections were created for analysis. The copy number of DRP1 RNA ranged from 15 to 62 copies per cell (average copy number, 37.9 copies per cell; fluctuation width, 4.1-fold), while the copy number of FIS1 RNA ranged from 16 to 77 copies per cell (average copy number, 38.7 copies per cell; fluctuation width, 4.8-fold). Thus, the expression levels and properties of DRP1 and FIS1 appeared to be similar. We also investigated the expression of two mitochondrial fusion genes, MFN1 and MFN2, in 21 and 23 cells, respectively (Fig. 2c). The MFN1 RNA copy number ranged from 22 to 113 copies per cell (average copy number, 55.7 copies per cell; fluctuation width, 5.1 times), whereas the MFN2 RNA copy number ranged from 17 to 153 copies per cell (average copy number, 62.2 copies per cell; fluctuation width, 9-fold). Therefore, the variability in MFN1 and MFN2 RNA expression was seemingly larger than that of DRP1 and FISH.

Next, we analyzed the relationship between the RNA expression levels of mitochondrial division/fusion genes and mitochondrial morphology in individual cells. We focused on cells containing high and low copy numbers of the RNA (magenta and green dots in Fig. 2c, respectively). Mitochondrial morphologies in these cells showed interesting trends. The cells containing high copy numbers of the RNA molecules had smaller mitochondria than cells containing low copy numbers of the RNA (Fig. 2d). Thus, the RNA expression levels of both division and fusion genes appear to be negatively correlated with the average size of fragmented mitochondria. Comparisons of the copy numbers of RNA versus the total number of fragmented mitochondria demonstrated this relationship more clearly. Cells containing high copy numbers of RNA (for both mitochondrial division and fusion genes) contained more fragmented mitochondria than cells containing low copy numbers of the RNA (Fig. 2e). These results pose an interesting question about mitochondrial dynamics. Considering their functions, the expression of the division genes would be expected to positively correlate with the number of mitochondria, as we observed. However, the observed positive correlation between mitochondria number and RNA for fusion genes was unexpected. These results suggest that high expression levels of genes for both mitochondrial division and fusion are required to produce and retain large numbers of mitochondria.

To help confirm this notion, we simultaneously measured the RNA expression levels of the eight mitochondrial division and fusion genes in individual cells by performing sequential rounds of probe hybridization, imaging, and photobleaching (Fig. 3). The sequential smFISH was carried out as follows. Each RNA species was labeled with a mixture of primary probes comprising an RNA-targeting region and a specific readout region for each RNA species. Each primary probe that hybridized with a target mRNA was further associated with one of eight types of fluorescence readout probes. After each round of hybridization, the cells were imaged by single-molecule microscopy. Fluorescent spots corresponding to individual RNA molecules were visualized and detected, after which the spots were efficiently extinguished via photobleaching (Fig. 3a). This hybridization round was repeated eight times for DRP1, FIS1, MFF, MFN1, MFN2, MIEF1, MIEF2, and OPA1 in order (Fig. 3b, c). After sequential smFISH imaging, the cells were immunostained with anti-TOM 20 antibody to visualize the mitochondrial structure. During these series of experimental procedures, the sample dish was not removed from the microscope. We carried out imaging for twenty-one cells (Fig. 4a, b). RNA copy numbers for DRP1, FIS1, MFN1 and MFN2 showed similar trends as in conventional smFISH imaging (Figs. 2c and 4a), indicating that the results from sequential smFISH imaging are reliable. Although the correlation between the RNA expression level of OPA1 and mitochondrial number was weak and not significant, the RNA expression levels of the other seven mitochondrial division/fusion genes were significantly correlated with mitochondrial numbers (Fig. 4c).

Fig. 3. Sequential smFISH imaging of RNA species from eight mitochondrial division and fusion genes. (a) Image acquisition process for sequential smFISH. Primary probe mixtures for RNA species from eight genes, including DRP1, FIS1, MFF, MIEF1, MIEF2, MFN1, MFN2, and OPA1, were hybridized to the cells, followed by post-fixation. Cy-5-labeled secondary probes were then hybridized to the cells for detection. Secondary probe hybridization, imaging, and bleaching were sequentially carried out for each RNA species. After eight rounds of sequential smFISH imaging, the cells were immunostained to visualize mitochondrial morphology. Scale bar=20 µm. (b) The localization of all single RNA molecules detected by sequential smFISH. Scale bar=20 µm. (c) False-colored fluorescence images of the eight RNA species for mitochondrial division and fusion genes after eight rounds of sequential hybridizations. Circles indicate identified RNA molecules that were distinguished from background signals (see Materials and methods). As a negative control in sequential smFISH imaging, random oligo DNAs were hybridized as primary probes, and no significant signals were detected by eight rounds of sequential hybridizations. Scale bar=2 µm.

Fig. 4. The number of RNA copy number for eight RNA mitochondrial division and fusion genes and mitochondria in twenty one analyzed by sequential smFISH imaging. (a) The copy number of RNAs. (b) The number of mitochondria. (c) Scatter plots of copy number of each RNA species per cell versus the number of mitochondria per cell. Pearson correlation coefficients (r) is also shown in each scatter plots. p values for DRP1, MIEF1, MIEF2, MFN1, and MFN2 are less than 0.01, and p values for MFF and FIS1 are less than 0.05.

Discussion

Gene expression is one of the most fundamental biological processes in living cells. Identifying the species and number of RNA molecules in a cell can shed light on the biological activity and conditions of the cell. However, well-established transcriptomic techniques such as DNA microarray analysis, RNA-seq, and ribosome profiling have only limited sensitivity and cannot be used to count the number of RNA molecules in a single cell. By contrast, single-molecule RNA imaging can be used to precisely determine RNA expression levels in individual cells (Symmons and Raj 2016). We do not exclude a possibility that RNA expression levels of other genes are also increased with mitochondrial dynamics and the expression patterns of the genes might be a specific trend for a cell type of the U2-OS. However, in this study, we showed the cell-to-cell variability in the expression levels of mitochondrial division and fusion genes at single-cell resolution. The expression levels of the genes varied approximately 4- to 9-fold depending on the cell, suggesting that the activity level of mitochondrial dynamics depends on cellular conditions in each individual cell. As expression patterns of other genes have not been investigated in this research, it is hypothesized that some of the genes are also expressed for the regulation of mitochondrial dynamics. Further single cell genome-wide RNA profiling will identify the uncharacterized genes relating with mitochondrial dynamics.

Recent studies using synchronized primitive unicellular organisms showed that many of RNAs and proteins expressed from mitochondrial division genes exist only during the mitochondrial division phase (Nishida et al. 2003, Yoshida et al. 2017). On the other hand, we did not detect any cells lacking RNA transcribed from mitochondrial division/fusion genes in human cultured U2-OS cells in this study. Given that each cell contained (at least) approximately 10 copies of RNA from these genes, it is hypothesized that there are always at least a few mitochondria within a cell undergoing division or fusion in animals. The positive correlation between the number of fragmented mitochondria and the RNA copy numbers from mitochondrial fusion genes suggests that in addition to the division process, the fusion process play important roles in the production of the many mitochondria in each cell. Previous studies demonstrated that mitochondrial fusion is important for mixing of mitochondrial component molecules (such as DNA, RNA, protein, and lipid molecules) and maintaining the electrical state of the mitochondria (Hayashi et al. 1994, Hoppins 2014). Therefore, perhaps such frequent structural remodeling via mitochondrial division and fusion induces the mixing of mitochondrial component molecules among individual mitochondria in a single cell, increasing the homogeneity of mitochondria within each cell (Fig. 5). To explore these hypotheses, further investigations using single cell omics analyses in other cell types and other organisms are needed, including transcriptomic, proteomic, and metabolomic analyses, which should provide insights into the molecular mechanisms of mitochondrial dynamics in living organisms.

Fig. 5. Schematic representation of regulation of mitochondrial quality control in a single cell. (a) The high copy number of mitochondria induce the heterogeneity of mitochondrial quality. (b) Frequent structural remodeling via mitochondrial division and fusion induces the mixing of mitochondrial component molecules among individual mitochondria.

Acknowledgment

This work was supported by Human Frontier Science Program Career Development Award (CDA00049/2018-C to Y.Y.); Japan Society for the Promotion of Science KAKENHI (JP18K06325 to Y. Y.); JGC-S Scholarship Foundation (to Y.Y.); the Sumitomo Foundation (to Y.Y.); PRESTO, Japan Science and Technology Agency (JPMJPR15F7 to Y.T.); Grants-in-Aid for Young Scientists (A) (24687022 to Y.T.), Challenging Exploratory Research (26650055 to Y.T.) and Scientific Research on Innovative Areas (23115005 to Y.T.), Japan Society for the Promotion of Science; the Takeda Science Foundation (to Y.T.); and the Mochida Memorial Foundation for Medical and Pharmaceutical Research (Y.T.).

References

• Bleazard, W., McCaffery, J. M., King, E. J., Bale, S., Mozdy, A., Tieu, Q., Nunnari, J. and Shaw, J. M. 1999. The dynamin-related GTPase Dnm1 regulates mitochondrial fission in yeast. Nat. Cell Biol. 1: 298–304.
• Chen, K. H., Boettiger, A. N., Moffitt, J. R., Wang, S. and Zhuang, X. 2015. Spatially resolved, highly multiplexed RNA profiling in single cells. Science 348: aaa6090.
• Coskun, A. F. and Cai, L. 2016. Dense transcript profiling in single cells by image correlation decoding. Nat. Methods 13: 6–11.
• Friedman, J. R. and Nunnari, J. 2014. Mitochondrial form and function. Nature 505: 335–343.
• Gandre-Babbe, S. and van der Bliek, A. M. 2008. The novel tail-anchored membrane protein Mff controls mitochondrial and peroxisomal fission in mammalian cells. Mol. Biol. Cell 19: 2402–2412.
• Gillham, N. W., Boynton, J. E. and Hauser, C. R. 1994. Translational regulation of gene expression in chloroplasts and mitochondria. Annu. Rev. Genet. 28: 71–93.
• Gray, M. W. 1992. The endosymbiont hypothesis revisited. Int. Rev. Cytol. 141: 233–357.
• Griparic, L., van der Wel, N. N., Orozco, I. J., Peters, P. J. and van der Bliek, A. M. 2004. Loss of the intermembrane space protein Mgm1/OPA1 induces swelling and localized constrictions along the lengths of mitochondria. J. Biol. Chem. 279: 18792–18798.
• Hayashi, J., Takemitsu, M., Goto, Y. and Nonaka, I. 1994. Human mitochondria and mitochondrial genome function as a single dynamic cellular unit. J. Cell Biol. 125: 43–50.
• Hoppins, S. 2014. The regulation of mitochondrial dynamics. Curr. Opin. Cell Biol. 29: 46–52.
• Kuroiwa, T., Nishida, K., Yoshida, Y., Fujiwara, T., Mori, T., Kuroiwa, H. and Misumi, O. 2006. Structure, function and evolution of the mitochondrial division apparatus. Biochim. Biophys. Acta 1763: 510–521.
• Liu, R. and Chan, D. C. 2015. The mitochondrial fission receptor Mff selectively recruits oligomerized Drp1. Mol. Biol. Cell 26: 4466–4477.
• Losón, O. C., Song, Z., Chen, H. and Chan, D. C. 2013. Fis1, Mff, MiD49, and MiD51 mediate Drp1 recruitment in mitochondrial fission. Mol. Biol. Cell 24: 659–667.
• Lubeck, E., Coskun, A. F., Zhiyentayev, T., Ahmad, M. and Cai, L. 2014. Single-cell in situ RNA profiling by sequential hybridization. Nat. Methods 11: 360–361.
• Mears, J. A., Lackner, L. L., Fang, S., Ingerman, E., Nunnari, J. and Hinshaw, J. E. 2011. Conformational changes in Dnm1 support a contractile mechanism for mitochondrial fission. Nat. Struct. Mol. Biol. 18: 20–26.
• Mozdy, A. D., McCaffery, J. M. and Shaw, J. M. 2000. Dnm1p GTPase-mediated mitochondrial fission is a multi-step process requiring the novel integral membrane component Fis1p. J. Cell Biol. 151: 367–379.
• Nishida, K., Takahara, M., Miyagishima, S., Kuroiwa, H., Matsuzaki, M. and Kuroiwa, T. 2003. Dynamic recruitment of dynamin for final mitochondrial severance in a primitive red alga. Proc. Natl. Acad. Sci. U.S.A. 100: 2146–2151.
• Olichon, A., Baricault, L., Gas, N., Guillou, E., Valette, A., Belenguer, P. and Lenaers, G. 2003. Loss of OPA1 perturbates the mitochondrial inner membrane structure and integrity, leading to cytochrome c release and apoptosis. J. Biol. Chem. 278: 7743–7746.
• Osteryoung, K. W. and Nunnari, J. 2003. The division of endosymbiotic organelles. Science 302: 1698–1704.
• Otera, H., Wang, C., Cleland, M. M., Setoguchi, K., Yokota, S., Youle, R. J. and Mihara, K. 2010. Mff is an essential factor for mitochondrial recruitment of Drp1 during mitochondrial fission in mammalian cells. J. Cell Biol. 191: 1141–1158.
• Palmer, C. S., Osellame, L. D., Laine, D., Koutsopoulos, O. S., Frazier, A. E. and Ryan, M. T. 2011. MiD49 and MiD51, new components of the mitochondrial fission machinery. EMBO Rep. 12: 565–573.
• Raj, A., van den Bogaard, P., Rifkin, S. A., van Oudenaarden, A. and Tyagi, S. 2008. Imaging individual mRNA molecules using multiple singly labeled probes. Nat. Methods 5: 877–879.
• Roy, M., Reddy, P. H., Iijima, M. and Sesaki, H. 2015. Mitochondrial division and fusion in metabolism. Curr. Opin. Cell Biol. 33: 111–118.
• Schindelin, J., Arganda-Carreras, I., Frise, E., Kayning, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J. Y., White, D. J., Hartenstein, V., Eliceiri, K., Tmancak, P. and Cardona, A. 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9: 676–682.
• Shah, S., Lubeck, E., Zhou, W. and Cai, L. 2016. In situ transcription profiling of single cells reveals spatial organization of cells in the mouse hippocampus. Neuron 92: 342–357.
• Shepard, K. A. and Yaffe, M. P. 1999. The yeast dynamin-like protein, Mgm1p, functions on the mitochondrial outer membrane to mediate mitochondrial inheritance. J. Cell Biol. 144: 711–719.
• Symmons, O. and Raj, A. 2016. What’s luck got to do with it: single cells, multiple fates, and biological nondeterminism. Mol. Cell 62: 788–802.
• Wong, E. D., Wagner, J. A., Gorsich, S. W., McCaffery, J. M., Shaw, J. M. and Nunnari, J. 2000. The dynamin-related GTPase, Mgm1p, is intermembrane space protein required for maintenance of fusion competent mitochondria. J. Cell Biol. 151: 341–352.
• Yoshida, Y., Kuroiwa, H., Shimada, T., Yoshida, M., Ohnuma, M., Fujiwara, T., Imoto, Y., Yagisawa, F., Nishida, K., Hirooka, S., Misumi, O., Mogi, Y., Akakabe, Y., Matsushita, K. and Kuroiwa, T. 2017. Glycosyltransferase MDR1 assembles a dividing ring for mitochondrial proliferation comprising polyglucan nanofilaments. Proc. Natl. Acad. Sci. U.S.A. 114: 13284–13289.