Role of mitochondrial fission-related genes in mitochondrial morphology and energy metabolism in mouse embryonic stem cells.

Mitochondria, the major organelles that produce energy for cell survival and function, dynamically change their morphology via fusion and fission, a process called mitochondrial dynamics. The details of the underlying mechanism of mitochondrial dynamics have not yet been elucidated. Here, we aimed to investigate the function of mitochondrial fission genes in embryonic stem cells (ESCs). To this end, we generated homozygous knockout ESC lines, namely, Fis1-/-, Mff-/-, and Dnm1l-/- ESCs, using the CRISPR-Cas9 system. Interestingly, the Fis1-/-, Mff-/-, and Dnm1l-/- ESCs showed normal morphology, self-renewal, and the ability to differentiate into all three germ layers in vitro. However, transmission electron microscopy showed a significant increase in the cytoplasm to nucleus ratio and mitochondrial elongation in Dnm1l-/- ESCs, which was due to incomplete fission. To assess the change in metabolic energy, we analyzed oxidative phosphorylation (OXPHOS), glycolysis, and the intracellular ATP concentration. The ESC knockout lines showed an increase in OXPHOS, decrease in glycolysis, and an increase in intracellular ATP concentration, which was related to mitochondrial elongation. In particular, the Dnm1l knockout most significantly affected mitochondrial morphology, energy metabolism, and ATP production in ESCs. Furthermore, RNA sequencing and gene ontology analysis showed that the differentially expressed genes in Mff-/- ESCs were distinct from those in Dnm1l-/- or Fis1-/- ESCs. In total, five metabolism-related genes, namely, Aass, Cdo1, Cyp2b23, Nt5e, and Pck2, were expressed in all three knockout ESC lines, and three of them were associated with regulation of ATP generation.

Introduction

Mitochondria are crucial for cell survival and homeostasis, as they participate in cellular energy metabolism, cell cycle [1], cell state maintenance [2], differentiation [3], and determination of stem cell identity, including decisions impacting cell fate [4]. Mitochondria can not only proliferate via self-division (biogenesis), but can also dynamically change their morphology via fusion and fission in response to the cell state and cellular requirement, which is referred to as ‘mitochondrial dynamics’ [5]. Generally, mitochondrial fusion is known to be directly regulated by fusion proteins such as optic atrophy 1 (Opa1), mitofusin 1 (Mfn1), and mitofusin 2 (Mfn2) [6,7]. Both Mfn1 and Mfn2 function in the outer mitochondrial membrane (OMM), whereas Opa1 is localized in the inner mitochondrial membrane (IMM) during mitochondrial fusion. In contrast, fission is directly regulated by the fission proteins such as dynamin 1 like (Dnm1l), mitochondrial fission factor (Mff), and mitochondrial fission 1 (Fis1) [[8], [9], [10]]. These fission proteins bind to the OMM during the mitochondrial fission [9]. In summary, the fusion and fission proteins physically bind to the OMM and IMM and induce morphological changes in the mitochondria; thus fusion and fission proteins regulate mitochondrial dynamics [11].

Mitochondrial morphology varies with cell types and energy demand. In general, mitochondria in undifferentiated stem cells are relatively fragmented and elongate upon fusion when cells differentiate. However, the extent of differentiation does not always reflect the shape of the mitochondria, which is also dependent on the type, niche, or function of the cells [3,4]. For example, neural stem cells (NSCs), which depend on aerobic glycolytic metabolism, possess elongated mitochondria [3]. Neurons, which are derived from differentiated NSCs, also harbor elongated mitochondria, but use oxidative phosphorylation (OXPHOS), which produces energy via oxygen consumption [4]. In contrast, erythropoietic cells, hepatocytes, and neural progenitor cells possess fragmented mitochondria [4,12,13]. Thus, various factors influence mitochondrial dynamics; however; the detailed underlying mechanism has not yet been elucidated.

In particular, embryonic stem cells (ESCs) exhibit unique mitochondrial morphology and energy metabolism. ESCs are pluripotent stem cells that can self-renew indefinitely in vitro and can differentiate into all three germ layers [14,15]. As ESCs are derived from the inner cell mass of the blastocyst, they harbor globular and immature mitochondria [16]. However, multipotent somatic stem cells, such as NSCs, hematopoietic stem cells, and mesenchymal stem cells show elongated mitochondria with mature cristae [3,4,17], indicating that globular mitochondria are the unique features of pluripotent stem cells. The basic mechanisms underlying the stem cell type-dependent differences in mitochondria are not known; however, ESCs contain globular mitochondria, probably because they are derived from the early stages of development, i.e. from the blastocyst [15].

The relationship between mitochondrial morphology and fusion or fission proteins has been investigated using several somatic cell types. For example, the overexpression of mitochondrial fusion proteins, such as Mfn1/Mfn2 [18] and Opa1 [19], induced mitochondrial elongation in mouse embryonic fibroblasts (MEFs). Similarly, the overexpression of mitochondrial fission proteins, such as Dnm1l [20], Fis1 [21], and Mff [21], induced mitochondrial fragmentation in MEFs and HeLa cells. In contrast, a loss-of-function study showed that Mfn1 or Mfn2 deficiency resulted in mitochondrial fragmentation in MEFs [18]. However, overexpression or loss-of-function studies on fusion/fission proteins using early embryos or ESCs containing immature forms of mitochondria are limited.

Here, we aimed to investigate the function of mitochondrial fission genes in ESCs. We generated ESC lines in which the mitochondrial fission protein-encoding genes, such as Dnm1l, Fis1, and Mff, were mutated using the clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9 system [[22], [23], [24]]. Interestingly, ESCs with knockouts in mitochondrial fission genes still maintained the ability to self-renew and differentiate into all three germ layers in vitro. Next, we investigated the effects of these knockouts on mitochondrial morphology, energy metabolism, and global gene expression profiles. We hypothesized that although fission gene knockout does not affect pluripotency, the loss-of-function of mitochondrial fission genes in ESCs may induce changes in mitochondrial dynamics and energy metabolism. For this purpose, mitochondrial morphology, oxygen consumption rate (OCR), extracellular acidification rate (ECAR), and changes in the expression of relevant genes were analyzed using Dnm1l, Fis1, and Mff ESC lines.

Materials and methods

Designing the CRISPR-Cas9 construct

The CRISPR-Cas9 single guide RNA (sgRNA) design tool developed by Dr. Zhang's laboratory (http://crispr.mit.edu/) was used to design single guide (sg)RNAs for Cas9 nuclease targeting Dnm1l, Fis1, and Mff in the mouse genome. The target sequences in the exons of several genes were selected following analysis of each gene. The 20-nt target sequences (5′-AAGTGTCAGGTTGACAACGT-3′ for Dnm1l, 5′-ACCTGGCCGTGGGCAACTAC-3′ for Fis1, 5′-TGGGACTTGCATTATCACAC-3′ for Mff), including the protospacer adjacent motif (PAM; 5′-NGG-3′), were selected for their predicted high score and reduced off-target effects. Three sequences, including the antisense sequence and restriction enzyme site, were synthesized as oligomers (Bionics, Seoul, Republic of Korea) for cloning in the Cas9/sgRNA expression vector [px330 Cas9/sgRNA dual expression system from Hyungbeom Kim's laboratory in Yonsei University (Republic of Korea)]. Cloning was performed as described previously [23].

Transfection

Mouse ESCs (1 × 105) were transfected with the sgRNA cloned in the px330 Cas9 expression vector via electroporation with the Neon transfection system (Thermofisher, Waltham, MA, USA) per the manufacturer's instructions at 1200 V and 10 ms pulse length.

Cell culture

Control ESC and Knockout ESC lines were maintained on a dish layered with MEF in the Mouse Embryonic Stem Cell (mESC) medium, consisting of Dulbecco's modified Eagle's medium (D-MEM) low glucose (Hyclone, 11885–084, GE Healthcare, Melbourne, VIC, Australia) supplemented with 15% heat-inactivated fetal bovine serum (FBS; Hyclone), 1x penicillin/streptomycin/glutamine (P/S/G; Gibco, 10378–016, Grand Island, NY, USA), 0.1 mM nonessential amino acids (NEAA; Gibco, 11140–050), 1 mM β-mercaptoethanol (Gibco, 21985–023), and 103 U/mL leukemia inhibitory factor (ESGRO, Merck Millipore). For three germ layers differentiation, all of the ESC lines were suspended on a petri dish in the DMEM-low medium, consisting of D-MEM low glucose (Hyclone) supplemented with 15% FBS (Hyclone), 1x P/S/G (Gibco), 0.1 mM NEAA, 1 mM β-mercaptoethanol (Gibco) for 7 days until the embryoid bodies were formed. After then, the embryoid bodies were attached to the porcine gelatin (Sigma, G9136) coated dish in DMEM-low medium and culture for 7 days with a medium change every day.

Immunocytochemistry

Cells were fixed for 20 min at 4 °C with 4% paraformaldehyde. After PBS washing, cells were treated with PBS containing 0.03% Triton X-100 (Sigma) for 10 min and then blocked for 30 min with PBS containing 3% bovine serum albumin (Bovogen) at room temperature. After then, cells were probed with primary antibodies against Pou5f1 (Oct 4; monoclonal, 1 μg/ml, Abcam sc-9081), Nanog (monoclonal, 1 μg/ml, Abcam ab80892), tubulin beta III isoform (tuj1; monoclonal, 1 μg/ml, Millipore MAB1637), smooth muscle actin (SMA; monoclonal, 1 μg/ml, Abcam ab7817), and Sox17 (polyclonal, 1 μg/ml, R&D systems AF1924). Finally, cells were labeled with secondary antibodies conjugated to Alexa Fluor 568 (Molecular Probes, Eugene, OR, USA), following the specifications of the manufacturer.

Teratoma formation

Control, Fis1, Mff, and Dnm1l ESCs (1 × 106) were injected into the testis capsule Balb/c Nude (5 weeks, male) mice, which were purchased from Orient Bio (Gyeonggi-do, Korea) to generate teratoma. Teratomas were harvested from mice on 5 weeks post-injection and fixed in 4% paraformaldehyde (Sigma), paraffin-embedded, and sectioned. To analyze the differentiation potential into all three germ layers, the sectioned slides were histologically stained by hematoxylin/eosin (Endoderm), masson's trichrome (Mesoderm), and anti-tuj1 antibody (Ectoderm).

Mitochondrial DNA (mtDNA) quantitative PCR analysis

For quantitative PCR, standard curves were created for each target gene primer set using known quantities of total mitochondrial DNA (mtDNA), and nucleus DNA (nDNA) from other cells. The PCR reactions were performed in triplicate using a TOPrealTM qPCR 2X PreMIX (Enzynomics, Daejeon, Republic of Korea) and Roche LightCycler 5480 following instruction. Target genes were amplified 45 cycles at 95 °C, 60 °C, and 72 °C for 10 s each. We corrected the differences in PCR efficiency between the target and reference loci using the efficiency correction in the Relative Quantification Software (Roche LC 480). The primers for real-time PCR are showed in previous report [25].

Electron microscopy

For transmission electron microscope (TEM) observations, the samples were fixed in 4% paraformaldehyde (Sigma) and 2.5% glutaraldehyde (Sigma) in 0.1 M phosphate (Sigma) buffer overnight. After washing in 0.1 M phosphate buffer, the samples were post-fixed for 1 h in 1% osmium tetroxide (Sigma) prepared in the same buffer. The samples were dehydrated with a graded series of ethyl alcohol concentrations, embedded in Epon 812, and polymerized at 60 °C for 3 days. Ultrathin sections (60–70 nm) were obtained using an ultramicrotome (Leica Ultracut UCT). Ultrathin sections collected on grids (200 mesh) were examined in TEM (JEM 1010) operating at 60 kV, and images were recorded by a charge-coupled device camera (SC1000; Gatan).

Mitochondrial length analysis

The images from electron microscopy were analyzed and measured by the Image J 1.43 (NIH) software for calculating the maximum (Max)/minimum (Min) ratio of mitochondrial length. At least over fifty mitochondria were measured and analyzed per sample to obtain data.

Oxygen consumption rate analysis

For measuring the oxygen consumption rate (OCR), we used Seahorse extracellular flux (XF96) analyzer. Total 2 × 104 cells were attached in XF96 Cell Culture Microplate pre-coated with Matrigel in mES medium before 24 h from the assay. After a medium change to XF base media supplemented d-glucose (1 g/L, Sigma, G8769), Sodium pyruvate (1 mM, Gibco, 11360–070) and l-glutamine (4 mM, Gibco, 25030–081), the assay was performed by using XF96 Extracellular Flux Analyzers (Seahorse Bioscience, North Billerica, MA, USA). Five measurements were obtained under basal conditions and after the addition of several chemicals such as oligomycin (1 μM), FCCP (2 μM) and rotenone (1 μM)/antimycin A (1 μM) (Agilent Technologies, 103015–100, Santa Clara, CA, USA). The process after treatment was performed following the manufacturer's instructions.

Extracellular acidification analysis

To measure the extracellular acidification rate (ECAR), we used a Seahorse extracellular flux (XF96) analyzer. In all, a total of 2 × 104 cells were cultured in XF96 Cell Culture Microplate pre-coated with Matrigel in mES medium before 24 h from the assay. After a medium change to XF base media supplemented l-glutamine (4 mM, Gibco, 25030–081), the assay was performed using XF96 Extracellular Flux Analyzers (Seahorse Bioscience, North Billerica, MA, USA). Four measurements were obtained under basal conditions and after the addition of several chemicals such as d-glucose (10 mM), oligomycin (1 μM), and 2-Deoxy-d-glucose (50 mM) (Agilent Technologies, 103020–100). The process after treatment was performed according to the manufacturer's instructions.

Direct measurement of adenosine triphosphate level

To measure the adenosine triphosphate (ATP) levels, we used a luminescent ATP detection assay kit (Abcam, ab113849) with GloMax 96 Microplate Luminometer (Promega, E6521, Madison, WI, USA). A total of 1 × 104 cells were cultured in 96 well cell culture plates (SPL life science, 30296, Gyeonggi-do, Korea) pre-coated with Matrigel in mES before 24 h from the assay. The process after treatment was performed according to the manufacturer's instructions.

RNA sequencing

Total RNA samples were converted into cDNA libraries using the TruSeq Stranded mRNA Sample Prep Kit (Illumina). Starting with 100 ng of total RNA, polyadenylated RNA (mainly mRNA) was selected and purified using oligo dT-conjugated magnetic beads. This mRNA was physically fragmented and converted into single-stranded cDNA using reverse transcriptase and random hexamer primers, with the addition of Actinomycin D to suppress DNA-dependent synthesis of the second strand. Double-stranded cDNA was created by removing the RNA template and synthesizing the second strand in the presence of dUTP (deoxy-ribo-uridine triphosphate) in place of dTTP (deoxythymidine triphosphate). A single A base was added to the 3¢ end to facilitate ligation of sequencing adapters, which contain a single T base overhang. Adapter-ligated cDNA was amplified by a polymerase chain reaction to increase the amount of sequence-ready library. During this amplification, the polymerase stalls when it encounters a U base, rendering the second strand a poor template. Accordingly, the amplified material used the first strand as a template, thereby preserving the strand information. Final cDNA libraries were analyzed for size distribution and using an Agilent Bioanalyzer (DNA 1000 kit; Agilent), quantitated by qPCR (Kapa Library Quant Kit; Kapa Biosystems, Wilmington, MA, USA), then normalized to 2 nM in preparation for sequencing.

Sequenced read processing

The validity of sequenced reads was accessed with FastQC (v0.11.5) and the reads were aligned with STAR (v2.5.2b) to UCSC mouse mm10 genome assembly. StringTie (v1.3.3b) was used to process STAR results in transcript assemblies. Results from StringTie were used to calculate average expression of each gene in FPKM and the acquired values were used for drawing scatter plots in R (v3.6.1). For heatmaps, the FPKM value of individual samples and heatmap 2 function of R's gplots (v3.0.1.1) package were used. In all cases, DEGs were identified with the aforementioned cutoff value using calculations with R. With the resulting DEGs from scatter plot and heatmap, gene ontology and KEGG pathway analysis were conducted through DAVID (v6.8).

Statistical analysis

All experiments were performed in triplicate and data represented as means ± standard deviation. The significance of differences was assessed by an unpaired two-tailed student's t-test, one-way analysis of variance (ANOVA) with Tukey's Honestly Significant Difference post hoc or Brown-Forsythe and Welch ANOVA with Tamhane's T2 post hoc for multiple comparisons. p-values <0.05 were considered statistically significant.

Animal use ethical statement

All methods used in this study were carried out in accordance with national animal care and use guidelines laws, and all experimental protocols were approved by the Institutional Animal Care and Use Committee of Konkuk University.

Results

Generation of mouse ESCs with mitochondrial fission gene knockout

To engineer biallelic mutations of the mitochondrial fission genes Dnm1l, Fis1, and Mff in ESCs, we introduced the px330 backbone-based sgRNA-Cas9 dual expression vectors, which harbor specific sgRNAs for each gene (Fig. 1A). Using our optimized protocol for the selection of the knockout ESCs (Fig. 1B), we successfully obtained homozygous and heterozygous knockout ESC lines for each gene (Fig. 1C–E). We targeted the exon of each gene for introducing frameshift in the open reading frames (Fig. 1C–E). First, we obtained Dnm1l homozygous (Dnm1l) and heterozygous (Dnm1l) knockout ESCs (Fig. 1C). Both alleles of Dnm1l in Dnm1l ESCs harbored a +1 or −29 nucleotide (nt) frameshift in exon 5, while one allele harbored a −7 nt frameshift in Dnm1l+/- ESCs (Fig. 1C). Second, we also obtained Fis+/- and Fis ESCs (Fig. 1D). Both Fis1 alleles in Fis ESCs contained a −8 nt frameshift in exon 3, while one allele harbored a −7 nt frameshift in Fis+− ESCs (Fig. 1D). Finally, we obtained Mff and Mff ESCs (Fig. 1E). Both alleles of Mff in Mff ESCs contained −14 or −28 nt frameshift in exon 3, while one allele harbored −2 nt frameshift in Mff ESCs. Collectively, we confirmed that homozygous and heterozygous knockout ESC lines of three mitochondrial fission genes had been established successfully. However, owing to negligible difference in many aspects between the wild-type and heterozygous knockout ESCs, we used the homozygous knockout ESC lines for further analyses.

Fig. 1

Generation of mitochondrial fission gene knockouts in ESCs. (A) Cas9/sgRNA dual expression vector construct used in this study. (B) Schematic illustration showing the generation of mitochondrial fission gene knockout ESC lines. (C–E) Information regarding the mutated sequence of the targeted region of each knockout ESC line (red; protospacer adjacent motif (PAM) sequence, blue; guide RNA target sequence, orange; inserted sequences, gray; deleted sequences, underline; targeted sequence). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Mitochondrial fission genes were not essential for the maintenance of ESC pluripotency

We predicted that ESCs with homozygous knockout in mitochondrial fission genes may not be pluripotent, which can be inferred from the cell state-dependent differences in mitochondrial morphology [2], such as the differentiated state [3], stem cell state [4], and pluripotent state [26]. Hence, we investigated the molecular characteristics of these homozygous knockout ESC lines. First, we compared the expression of the major pluripotency markers, OCT4 and NANOG, using immunocytochemistry in each ESC line (Fig. 2A). All the ESC lines stained positively for OCT4 and NANOG. Next, we compared the proliferation rate of the three knockout ESCs (Fis1, Mff, and Dnm1l ESCs) (Supplementary Fig. 1). Interestingly, only Dnm1l ESCs showed significantly decreased proliferation rate compared with control ESCs.

Fig. 2

Pluripotency of mitochondrial fission gene knockout ESCs. (A) Immunocytochemical analysis of pluripotency markers, OCT4 and NANOG, in control, Dnm1l−/−, Fis1−/−, and Mff−/− ESCs. Scale bars: 200 μm. (B)In vitro differentiation via embryoid body formation in control, Dnm1l−/−, Fis1−/−, and Mff−/− ESCs. All ESC lines were able to differentiate into the mesoderm (smooth muscle actin, SMA), endoderm (Sox17), and ectoderm (neuron-specific class III beta-tubulin, Tuj1) lineages. Scale bars: 50 μm. (C) Heat map showing the expression patterns of pluripotency-related genes in control, Dnm1l−/−, Fis1−/−, and Mff−/− ESCs.

Next, we assessed whether these knockout ESC lines differentiated into all the three germ layers in vitro (Fig. 2B). In vitro differentiation through embryoid body formation revealed that all the ESC lines had successfully differentiated into the mesoderm, endoderm, and ectoderm, as indicated by positive staining for SMA, SRY-box 17 (Sox17), and neuron-specific class III β-tubulin (Tuj1), respectively (Fig. 2B). Next, differentiation potential was confirmed by in vivo differentiation through teratoma formation (Supplementary Fig. 2). All three knockout ESCs (Fis1, Mff, and Dnm1l ESCs) could differentiate into endoderm (gut epithelium), mesoderm (muscle), and ectoderm (neuron), which were indistinguishable from control ESCs. Furthermore, RNA sequencing analysis confirmed that the expression of the pluripotency-related genes, including Sox 2, Esrrb, Klf4, Dppa5a, and Dnmt3l, did not differ significantly from those of the control ESCs (Fig. 2C). Hence, we concluded that knockout of the mitochondrial fission genes in ESCs did not affect the maintenance of pluripotency and differentiation potential.

Knockout of mitochondrial fission gene affected mitochondrial morphology in ESCs

Although all the knockout ESC lines were pluripotent, we speculated that knocking out the mitochondrial fission genes might induce morphological changes in the mitochondria, as previous reports have shown that the inhibition of mitochondrial fission was accompanied by mitochondrial elongation in somatic cells [27]. Hence, we performed transmission electron microscopy to investigate the morphologies of intracellular organelles, including the mitochondria, in each ESC line (Fig. 3A–F). Although the cytoplasm to nucleus ratios of the control, Fis1, and Mff ESCs were similar, the Dnm1l ESCs showed an increase in the cytoplasm to nucleus ratio (Fig. 3A and B). Next, we compared the mitochondrial morphology of fission gene knockout and control ESCs. Generally, mitochondria in pluripotent stem cells are globular-shaped with only a few cristae [28]. Hence, we investigated the extent of mitochondrial elongation in the ESC knockout lines (Fig. 3C). The Dnm1l ESCs in particular showed many elongated mitochondria that had failed to complete fission (white arrows in Fig. 3C). In addition, to measure the number of mitochondria, we analyzed the mitochondrial DNA. The amount of mitochondrial DNA was decreased in Dnm1l ESCs compared with that of control ESCs (Supplementary Fig. 3). The extent of elongation or globularity was determined by measuring the ratio of the mitochondrial maximal (max) to minimal (min) length (max/min ratio) (Fig. 3D). Interestingly, the max/min ratio of mitochondria had increased only in the Dnm1l ESCs, whereas the Fis1 and Mff ESCs showed max/min ratio similar to that of the control ESCs (Fig. 3E). In agreement with these results, the globularity of the mitochondria increased only in the Dnm1l ESCs (Fig. 3F). The mean value of the globularity in Dnm1l ESCs was 2.30, which was significantly higher than those of the control (1.46), Fis1 (1.62), and Mff ESCs (1.82). Collectively, we concluded that Dnm1l is the crucial gene that controls mitochondrial morphology and cytoplasm to nucleus ratio, and these aspects were not affected by the other fission genes, Fis1 and Mff.

Fig. 3

Ultrastructural analysis of ESCs harboring mitochondrial fission gene knockouts. (A) Representative image of each knockout ESC line observed using transmission electron microscopy. Nuclei (N), cytoplasm (C) Scale bars: 2 μm. (B) Cytoplasm to nuclear ratio of each knockout ESC line. Brown-Forsythe ANOVA test: p = 0.11.(C) Enlarged TEM images of each knockout ESC line focused on the mitochondria. Most mitochondria in Dnm1l ESCs were elongated owing to fission failure (arrows). Scale bars: 0.5 μm. (D) Criteria used for analyzing mitochondrial maximal (Max) or minimal (Min) length of all ESC lines. (E) Calculated length of mitochondrial max/min values using the ImageJ software among all ESC lines. (F) Globularity (calculated (c)-Max/c-Min axes) of mitochondria. One-way ANOVA test: ***p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Knockout of mitochondrial fission genes in ESCs affected cellular energy metabolism

Mitochondria play a pivotal role in providing the energy required for cell survival via OXPHOS, a process for producing energy via oxygen consumption [29]. Previous reports regarding mitochondria and OXPHOS have shown that mitochondrial morphology is involved in the metabolism of various cell types [30]. Most studies on the upregulation or downregulation of mitochondria-related genes have revealed that the changes in the expression level of mitochondrial fission genes induce morphological changes in the mitochondria. However, studies on whether mitochondria-related genes affect cellular metabolism are limited. Hence, we aimed to compare the metabolic status in ESC lines with knockouts in mitochondrial fission genes.

To determine the metabolic state of these cells, we first compared the OCR, which represents the OXPHOS activity in cells (Fig. 4A). Basal respiration represents the oxygen consumption that is used for cellular ATP production under normal conditions. The basal respiration levels from all samples decreased and reached the basal level after treatment with oligomycin (an ATP synthase inhibitor) (Fig. 4A and B). Maximal respiration was observed after treatment with FCCP, which collapses the proton gradient and disrupts the mitochondrial membrane potential (Fig. 4A and C). In particular, the highest level of maximal respiration was observed in Dnm1l ESCs, which showed distinct mitochondrial morphology compared to the other knockout ESC lines and control ESCs (Fig. 4A and C). Next, we analyzed the spare respiratory capacity (SRC), which represents extra oxygen consumption of mitochondria. Control ESCs showed a decrease in SRC (Fig. 4A and D), indicating that the mitochondria of control ESCs have higher membrane potential than those in the knockout ESC lines. The mitochondrial membrane potential is generated by proton pumps during the process of energy storage in OXPHOS. A previous report also showed reduction in SRC in T cells, which have higher membrane potential [31]. In contrast, high SRC was observed in all knockout ESC lines, especially in Dnm1l ESCs (Fig. 4D). This indicated that knockout of the mitochondrial fission gene increased SRC. SRC might be strongly related to mitochondrial dynamics. Collectively, these data indicated that the knockout of mitochondrial fission genes induced changes in metabolism, as well as in mitochondrial morphology.

Fig. 4

Metabolic analysis of ESCs with mitochondrial fission gene knockouts. (A) Measurement of oxygen consumption rate (OCR) in all ESC lines. Measurement of (B) Basal respiration, (C) maximal respiration, (D) spare respiratory capacity. (E) Measurement of extracellular acidification rate (ECAR) in all ESC lines. Measurement of (F) glycolysis, (G) glycolytic capacity, (H) glycolytic reserve in all ESC lines. (I) Measurement of energy phenotype in all ESC lines. The empty square shows the basal cell state. The filled square shows the cell state response to oxygen stress. (J) ATP production in all ESC lines. All ESC lines; control, Dnm1l, Fis1, and Mff ESCs. Welch's ANOVA test: *p < 0.05, **p < 0.01.

Next, we compared the extracellular acidification rate (ECAR), which represents the glycolytic capacity (Fig. 4E). We hypothesized that the mitochondria of the knockout ESCs were elongated and showed an eventual decrease in glycolytic function, as they were morphologically favorable for OXPHOS. ECAR increased rapidly after treatment with oligomycin, which inhibits the ATP synthase during OXPHOS; this was because of the dependence on glycolysis for cellular energy production. Interestingly, ECAR varied with the knockout cell line (Fig. 4E and F). Control ESCs showed higher glycolytic function, including glycolysis, glycolytic capacity, and glycolytic reserve, than Dnm1l−/− and Mff−/− ESCs (Fig. 4F–H). However, the changes in the glycolytic categories of Fis1−/− ESCs did not differ significantly from those of control ESCs (Fig. 4F–H). Taken together, these results demonstrated that among the mitochondrial fission genes examined, Dnm1l caused mitochondrial elongation, increased OXPHOS, and reduced glycolytic metabolic changes. In addition, XF Cell Energy Phenotype analysis showed that the knockout ESC lines (Dnm1l−/−, Fis1−/−, and Mff−/−) moved to the energetic phase (metabolic state focused on energy production), whereas the control ESCs moved to the glycolytic phase (metabolic state focused on producing cellular compounds) (Fig. 4I), indicating the shift from glycolytic to OXPHOS property due to the knockout of mitochondrial fusion genes in ESCs. Next, we determined the total cellular ATP concentration using the luminescent ATP detection assay (Fig. 4J). As expected from the higher SRC levels in the knockout lines than in the control ESCs, knockout ESC lines produced more intracellular ATP, with the highest level in Dnm1l−/− ESCs (Fig. 4J). Collectively, these results indicated that knockout of mitochondrial fission genes increased OXPHOS, decreased glycolysis, and increased intracellular ATP concentration, which are related to mitochondrial elongation, due to reduced mitochondrial fission.

Molecular signature of ESCs with knockout in mitochondrial fission genes

Next, we performed RNA-seq analyses to compare the global gene expression patterns of the mitochondrial fission gene knockout ESC lines and control ESCs. In total, 768 genes (vs. Dnm1l−/− ESCs), 639 genes (vs. Fis1−/− ESCs), and 1029 genes (vs. Mff−/− ESCs) were differentially expressed when compared with control ESCs (fragment per kilobase of transcript per million mapped reads (FPKM) > 2, fold change (FC) > 2) (Fig. 5A and B). Among the 768 differentially expressed genes (DEGs), 390 genes were overexpressed and 378 were downregulated in Dnm1l−/− ESCs. Among the 639 DEGs in Fis1−/− ESCs, 280 and 359 genes were overexpressed and downregulated, respectively. Among the 1029 DEGs in Mff−/− ESCs, 444 and 585 genes were overexpressed and downregulated, respectively (Fig. 5A and B, Supplementary Figs. 4–5).

Fig. 5

RNA sequencing analysis of ESCs harboring mitochondrial fission gene knockouts. (A) Scatterplot analysis of all ESC lines. (B) Gene ontology: biological process (GO:BP) analysis of enriched genes among the DEGs in all ESC lines. (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of enriched genes among the DEGs in all ESC lines. (D) Correlation matrix analysis showed the distinct gene expression patterns of Mff−/− ESCs compared to Dnm1l and Fis1 ESCs. (E) Venn diagram analysis for DEGs in metabolic pathways in Dnm1l−/−, Fis1−/−, and Mff−/− ESCs. (F) Heat map clusters of 1674 DEGs among all ESC lines. All ESC lines; control, Dnm1l−/−, Fis1−/−, and Mff−/− ESCs.

Next, we attempted to identify the biological functions of the DEGs in each KO ESC line (Fig. 5B). Gene ontology-biological process (GO: BP) analysis showed that the DEGs in Dnm1l−/− ESCs were highly enriched in the terms ‘multicellular organism development’, ‘cell differentiation’, and ‘axon guidance’. Similarly, in Fis1−/− ESCs, DEGs were enriched in the terms ‘multicellular organism development’, ‘cell differentiation’, and ‘regulation of cell shape’. In contrast, the DEGs in Mff−/− ESCs were different from those in Dnm1l−/− or Fis1−/− ESCs. DEGs in Mff−/− ESCs were enriched in the terms ‘regulation of cell adhesion’, ‘angiogenesis’, and ‘negative regulation of transcription from RNA polymerase II promoter’. In addition, we performed the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis to determine the main signaling pathway in each KO ESC line. Surprisingly, the DEGs of all the three KO ESC lines were enriched in ‘metabolic pathways’ in the top 10 categories (Fig. 5C). Correlation matrix analysis showed that the gene expression patterns of Dnm1l−/− and Fis1−/− ESCs were similar; however, they differed from those of Mff−/- ESCs (Fig. 5D). Furthermore, Venn diagram analysis showed that almost half of the genes (49.2%, 31/63) were common between Dnm1l−/- and Fis1−/− ESCs in the metabolic pathway. However, > 50% genes (50.8%, 65/128) in the DEGs of metabolic pathways were expressed distinctly in Mff−/- ESCs (Fig. 5E). In total, five genes, Aass, Cdo1, Cyp2b23, Nt5e, and Pck2, were overlapping in all three knockout ESC lines (Fig. 5E). In particular, three of the five genes, namely, Cdo1, Nt5e, and Pck2, were associated with the ATP regulation process in cancer [[32], [33], [34], [35], [36]], which explains the subsequent change in ATP production rate after knocking out these genes. Hierarchical clustering analysis of DEGs revealed that Mff−/- ESCs were relatively closer to the control ESCs than the Dnm1l−/- and Fis1−/− ESCs (Fig. 5F). Taken together, although all three mitochondrial fission genes affected metabolic pathways in the KEGG pathway analysis, Dnm1l−/- and Fis1−/− ESCs shared a molecular signature, such as development and differentiation in biological process, which was not observed in Mff−/- ESCs.

Discussion

In this study, we assessed the mitochondrial morphology, energy metabolism, and global gene expression profiles of pluripotent ESCs in which the mitochondrial fission genes (Dnm1l, Fis1, and Mff) were knocked out using the CRISPR-Cas9 system. Mitochondrial fission gene knockouts have been carried out in somatic cells but have not yet been attempted in pluripotent stem cells. As observed in the previous reports, the control ESCs showed immature mitochondria, which were globular-shaped with negligible cristae. However, knockout ESCs, which lacked one of the mitochondrial fission-related genes, displayed relatively mature and elongated mitochondria with more cristae. In particular, the effect of the Dnm1l knockout on mitochondria and metabolism was the most striking compared to the knockouts in the other genes; mitochondria were elongated by approximately 57.5% (c-max/c-min). Furthermore, Dnm1l−/− ESCs were defective in the fission process and did not complete mitochondrial fission, indicating that Dnm1l performs a critical role, similar to ‘scissors’ [37]. Mitochondria in Fis1−/− and Mff−/− ESCs also showed 11.0 and 24.7% increase in globularity, although we did not detect any failure of mitochondrial fission in these two KO ESC lines (Fig. 3C). Thus, Dnm1l appears to be the most important gene regulating mitochondrial fission and morphology. This is reasonable, as Fis1 and Mff function as adaptors for recruiting Dnm1l to the OMM, and the scissoring function is mainly associated with Dnm1l [11]. Furthermore, Fis1 is possibly involved more than Mff in recruiting Dnm1l, as Dnm1l ESCs have a molecular signature more similar to that of Fis1 ESCs than that of Mff ESCs (Fig. 5E).

In addition, the mitochondrial fission is linked to mitochondrial morphologies, which affect the cellular metabolic state [38]. In conditions of starvation, Drp1 in human (Dnm1l in mouse) is phosphorylated and this leads to mitochondrial fusion and elongation, which prevents the mitochondrial autophagic degradation and cell death [39]. The AMP-activated protein kinase (AMPK) also senses the metabolic status and ATP level of cells and regulates mitochondrial biogenesis, quality, and morphologies [40]. Generally, mature mitochondria are expected to produce energy more efficiently than immature mitochondria due to their complex cristae structure, which contains more electron transport chains required for energy production [41]. However, unlike our prediction, the basal mitochondrial respiration of the three knockout ESC lines were lower than those of control ESCs despite the presence of more mature mitochondria. Among the three knockout ESC lines, maximal respiration and SRC were most remarkably increased in Dnm1l−/− ESCs, indicating that knockout of mitochondrial fission genes affected mitochondrial function, as well as mitochondrial morphology. This is because Dnm1l−/− ESCs showed the most dramatic change in mitochondrial morphology and energy metabolism. Similarly, ATP production rates were increased in all knockout ESC lines (Fig. 4E), indicating that ESCs with mature mitochondria produce more ATP, which may be because ATP production is more efficient in OXPHOS than in glycolysis.

We also investigated the DEGs in the knockout and control ESC lines using RNA sequencing analysis. The high-ranked GO terms of Dnm1l−/− and Fis1−/− ESCs were identical, namely, ‘multicellular organism development’, ‘cell differentiation’, and ‘axon guidance’. As Fis1 functions as a receptor for the cytoplasmic Dnm1l [42], the knockouts of these two genes might display similar phenotypes. On the contrary, high-ranked GO terms of Mff−/− ESCs were distinct from those of Dnm1l−/− and Fis1−/− ESCs. This may be because Mff acts with MiD49 and MiD51, in addition to acting with Dnm1l [10]. Thus, the number of DEGs in Mff−/− ESCs might be more than that in Dnm1l−/− or Fis1−/− ESCs. In addition, the Venn diagram also represented the similarity in DEG expression in Dnm1l−/− and Fis1−/− ESCs, because almost 50% DEGs (49.2%, 31/63) of the metabolic pathway were overlapping, while > 50% genes (50.8%, 65/128) were differentially expressed only in Mff−/− ESCs.

Five metabolism-related genes, Aass, Cdo1, Cyp2b23, Nt5e, and Pck2, were differentially expressed in all knockout ESC lines compared to in the control ESCs. Three of these five genes, Cdo1, Nt5e, and Pck2, were reported as regulators of ATP production. Cdo1, encoding cysteine dioxygenase 1, is associated with cysteine sulfinic acid accumulation in cysteine metabolism [32]. It is also linked to the regulation of OXPHOS [32]. Nt5e encodes ecto-5′-nucleotidase, which hydrolyzes extracellular adenosine monophosphate into adenosine and inorganic phosphate [43]. Pck2 encodes a mitochondrial phosphoenolpyruvate carboxykinase, a hub enzyme associated with the tricarboxylic acid cycle, glycolysis, and gluconeogenesis, owing to the conversion of mitochondrial oxaloacetate to phosphoenolpyruvate [36]. Therefore, these three genes were directly related to the regulation of ATP production. A hierarchical clustering and correlation assay also confirmed that the gene expression profile of Mff−/− ESCs clustered differently from the group that include Dnm1l−/− and Fis1−/− ESCs.

We attempted to assess energy metabolism more accurately in knockout and control ESCs using multiple analyses such as OCR, ECAR, and intracellular ATP measurement. However, certain limitations impeded our ability to comprehend the link between the knockout effect of mitochondrial fission genes and metabolic changes in ESCs. First, we measured oxygen consumption and hydrogen ion concentration (pH) using OCR and ECAR analysis, respectively, in the three knockout and control ESC lines. Mammalian cells also use beta-oxidation as a catabolic process that catabolizes fatty acid molecules to produce acetyl-CoA [44]. The beta-oxidation process for energy production is primarily facilitated by peroxisomes instead of mitochondria. In the present study, we focused on the changes in OXPHOS and glycolysis induced by the knockout of the mitochondrial fission genes. Thus, it will be interesting to check whether the knockout of the fission genes affect the beta-oxidation in ESCs for a more accurate interpretation of the effects on energy metabolism. Second, changes in global gene expression patterns were obtained by knocking out a single gene related to mitochondrial fission. However, the changes in gene expression due to fission gene knockout may affect other cellular processes, including various cell signaling pathways and cell survival. Therefore, a detailed investigation regarding the association between mitochondrial dynamics and various cell signaling pathways may be helpful for understanding the roles of mitochondrial fission genes.

This study was performed using pluripotent mouse ESCs, which have rarely been used for studying mitochondrial OXPHOS compared to differentiated somatic cell types. Our data clearly showed that changes in mitochondrial morphology may be induced by knocking out mitochondrial fission genes in pluripotent ESCs, which also affected energy metabolism and global gene expression profiles, especially those related to cell differentiation, organism development, and metabolic pathways. Mff was the least affected gene, as its gene expression profile was similar to that of the control ESCs. Unlike previous studies focusing on somatic cells, genetic changes in the mitochondrial fission machinery in ESCs have not been studied extensively so far. Thus, although ESCs defective in mitochondrial fission maintain the normal features of pluripotency, further studies might reveal how mitochondrial dynamics and the resulting metabolic changes affect cell differentiation, as differentiation into a specific lineage(s) may be affected by mitochondrial function.

Author contributions

Conceived and designed the experiments: BJS, JTD. Performed the experiment: BJS JC. Analyzed the data: BJS, HL, KH. Technical support: YC, JTD. Wrote the paper: BJS JTD.

Declaration of competing interest

The authors declare that they have no conflicts of interest.