Mitochondrial Protein Synthesis Is Essential for Terminal Differentiation of CD45– TER119–Erythroid and Lymphoid Progenitors

Summary p32/C1qbp regulates mitochondrial protein synthesis and is essential for oxidative phosphorylation in mitochondria. Although dysfunction of p32/C1qbp impairs fetal development and immune responses, its role in hematopoietic differentiation remains unclear. Here, we found that mitochondrial dysfunction affected terminal differentiation of newly identified erythroid/B-lymphoid progenitors among CD45– Ter119– CD31– triple-negative cells (TNCs) in bone marrow. Hematopoietic cell-specific genetic deletion of p32/C1qbp (p32cKO) in mice caused anemia and B-lymphopenia without reduction of hematopoietic stem/progenitor cells. In addition, p32cKO mice were susceptible to hematopoietic stress with delayed recovery from anemia. p32/C1qbp-deficient CD51– TNCs exhibited impaired mitochondrial oxidation that consequently led to inactivation of mTORC1 signaling, which is essential for erythropoiesis. These findings uncover the importance of mitochondria, especially at the stage of TNCs during erythropoiesis, suggesting that dysregulation of mitochondrial protein synthesis is a cause of anemia and B-lymphopenia with an unknown pathology.


INTRODUCTION
Mitochondria are cellular organelles involved in multiple cellular functions such as oxidative phosphorylation (OXPHOS), energy metabolism, production of reactive oxygen species, iron homeostasis, signal transduction, and apoptosis (Nunnari and Suomalainen, 2012;Spinelli and Haigis, 2018;Tait and Green, 2012). Mitochondria have their own unique transcriptional system in which mitochondrial DNA (mtDNA) encoding rRNA, tRNA, and proteins comprising the respiratory chain is transcribed in response to cellular dynamics, including mitochondrial replication, which is regulated by mitochondrial transcription factor A (TFAM) (Kang et al., 2007;Uchiumi and Kang, 2017). p32, also known as complement component 1, q subcomponent-binding protein (p32/C1qbp), is a multifunctional chaperone protein associated with TFAM mainly localized in the mitochondrial matrix. p32/ C1qbp interacts with mitochondrial mRNA and is required for mitochondrial ribosome (mitoribosome) formation to synthesize proteins within mitochondria (Leucci et al., 2016;Muta et al., 1997;Petersen-Mahrt et al., 1999;. Dysfunction of p32/C1qbp impairs fetal development and immune responses, and its genetic mutations are related to human diseases such as cardiomyopathy and progressive external ophthalmoplegia (Feichtinger et al., 2017;Gotoh et al., 2018).
Hematopoiesis is a multistep process originating from HSCs at the top of the hematopoietic hierarchy, which is finely regulated by cell-intrinsic transcription factors, extrinsic cytokines, and metabolic controls (Asada et al., 2017;Ito and Suda, 2014;Kunisaki et al., 2013;Lu et al., 2019;Wei and Frenette, 2018). HSCs are located in bone marrow where they differentiate into all blood lineages. In bone marrow, non-hematopoietic stromal cell populations exist as constituents of hematopoietic microenvironments supporting HSC maintenance and differentiation (Mendez-Ferrer et al., 2010;Pinho and Frenette, 2019;Pinho et al., 2013;Sacchetti et al., 2007). Although the non-hematopoietic cell fraction of bone marrow was classically isolated as CD45 -Ter119 -CD31cells (hereafter referred to as triple-negative cells; TNCs), erythroid and lymphoid progenitors, which rapidly expand to replenish progenies in cases of hemolytic crises, have been newly identified among CD51 -TNCs (Boulais et al., 2018). Cell differentiation is a dynamic process during which numerous transcriptional and metabolic changes occur to assign the progenies with specific functions and characteristics. With recent advances in genomic technologies, nuclear transcription factors essential for lineage commitments have been identified. However, the roles of mitochondrial protein synthesis over the course of differentiation are yet to be clarified.
To address this issue, we generated hematopoietic-specific p32/C1qbp-deficient mice, in which mitochondria were found to be structurally and functionally impaired, and investigated the relationships between dysregulation of mitochondrial protein synthesis and hematopoietic differentiation in a steady-state and under hematopoietic stress.

p32/C1qbp Is Essential for Development of Erythrocytes and B-Lymphocytes
We previously reported that p32/C1qbp-deficient mice are embryonic lethal owing to loss of mitochondrial translation . To investigate the functions of p32/C1qbp in hematopoietic cells, we generated a hematopoietic-specific p32/C1qbp conditional knockout (p32cKO) mouse strain by crossing p32/ C1qbp flox/flox mice with Vav1-Cre transgenic mice. Protein expression analyses of bone marrow cells isolated from p32cKO mice (p32 flox/flox Vav1-Cre + ) and control littermates (p32 flox/flox Vav1-Cre -) confirmed that p32/C1qbp protein was deleted efficiently in the bone marrow cells of p32cKO mice ( Figure S1A). Coincident with a previous report showing that mitochondrial p32/C1qbp is required for maturation of mitochondrial rRNA and synthesis of mitochondria-encoded proteins (Leucci et al., 2016;, we found that 16S rRNA levels in the bone marrow cells was reduced significantly by the loss of p32/C1qbp (Figures S1B and S1C). Furthermore, we examined the expression levels of respiratory chain proteins by immunoblotting. The protein levels of complex I and IV, which include mtDNA-encoded subunits, were decreased significantly in p32/C1qbp-deficient (p32 À/À ) bone marrow cells ( Figure S1D). These results indicated that synthesis of proteins associated with the mitochondrial respiratory chain was dependent on p32/C1qbp in hematopoietic cells, which prompted us to investigate hematopoiesis in p32cKO mice.
We thus examined relationships between p32/C1qbp and hematopoiesis. p32cKO mice at the age of 8-12 weeks old displayed anemia and a decline of white blood cell (WBC) counts in the peripheral blood ( Figure 1A). Among the WBCs, the number of both B-and T-lymphocytes was prominently affected more than those of myeloid cells (Figures 1B,1C and S2A). To clarify what stages of hematopoietic precursors were functionally impaired, we performed bone marrow reconstitution analyses, in which total bone marrow cells from WT or p32cKO mice were transplanted into lethally irradiated WT recipients. Although anemia and reduction of white blood cells and B-lymphocytes were observed by 4 weeks after bone marrow transplantation with p32cKO BM cells, which sustained until at least 12 weeks, the numbers of T-lymphocytes and myeloid cells were gradually reduced in mice transplanted from p32cKO mice Figure 1. p32/C1qbp is Essential for Development of Erythrocytes and B-Lymphocytes (A) WBCs, RBCs, hemoglobin concentration (Hb), hematocrit (Ht), and the platelet (Plt) count in peripheral blood from 8-12-week-old WT (open circle, n = 8) and p32cKO (closed squares, n = 8) mice. (B and C) Numbers of lymphoid cells (Gr-1 -CD11b -), myeloid cells (Gr-1 + CD11b + ), B-lymphocytes (CD19 + CD3 -Gr-1 -CD11b -), and T-lymphocytes (CD19 -CD3 + Gr-1 -CD11b -) in the peripheral blood. (D) Kaplan-Meier plots of age-matched WT and p32cKO mice (n = 9) treated with 5-FU (250 mg/kg). (E) Appearance of bone marrow pellets from WT and p32cKO mice after 5-FU injection. (F) WBCs, RBCs, Hb, and Plts in peripheral blood from WT (open circle, n = 4-6) and p32cKO (closed squares, n = 4-6) mice after 5-FU injection (250 mg/kg). In (A-C) and (F), data are shown as means G SD. *p < 0.05 versus WT mice. Data are representative at least three (A-F) independent experiments. See also Figures S1 and S2.

OPEN ACCESS
iScience 23, 101654, November 20, 2020 3 iScience Article ( Figures S2B and S2C). These findings suggest that differentiation of erythroid and lymphoid committed progenitors was effected more prominently rather than hematopoietic stem/progenitor cells.

p32/C1qbp Deficiency Prohibits Terminal Differentiation of Erythrocytes and B-Lymphocytes
We next examined the bone marrow cells in p32cKO mice and found no significant changes in the numbers of HSCs and multipotent, common myeloid, or common lymphoid progenitors ( Figures S2E-S2G). To determine which stage of differentiation was affected, we analyzed lineage-committed progenitors in the bone marrow. Multipotent progenitor cells 2 generates pre-megakaryocyte-erythrocytes (pre-MegEs) (Pietras et al., 2015). Pre-MegEs are bipotent cells that are upstream of more committed erythroid-restricted progenitor (pre-CFU-E) and colony-forming unit erythroid (CFU-E) cells. The numbers and proportions of pre-CFU-E and CFU-E cells were increased significantly in the bone marrow of p32cKO mice (Figures 2A  and 2B, and S3A). We sorted pre-CFU-E and CFU-E cells from p32cKO bone marrow and examined their capacities for differentiation into erythrocytes in vitro. As a result, p32/C1qbp deficiency impaired the differentiation of pre-CFU-E and CFU-E cells into erythroid cells in vitro ( Figures 2C and 2D), suggesting that p32/C1qbp is important for erythroid differentiation after the CFU-E cell stage. Development of B-lymphocytes in bone marrow progresses following the order of pre-pro-B, pro-B, pre-B, immature B, and mature B cells (Nagasawa, 2006). The numbers of B-lymphocytes at the stage later than pre-B cells were reduced significantly in the bone marrow of p32cKO mice compared to those of controls (Figures 2E and S3B). To assess the differentiation capacities of pro-B cells, we sorted pro-B cells and evaluated their ability to form colonies of pre-B cells in culture. As observed in the erythroid lineage, the loss of p32/C1qbp impaired a colony-forming activity of B-lymphoid cells ( Figure 2F). Taken together, these results indicate that terminal differentiation of erythrocytes and B-lymphocytes is disrupted by p32/C1qbp deficiency.

p32/C1qbp Regulates Mitochondrial OXPHOS in CD51 -TNCs
To explore the mechanisms by which p32/C1qbp was involved in the differentiation process of CD51 -TNCs into erythrocytes and B-lymphocytes, we analyzed structure and functions of mitochondria in CD51 -TNCs. Electron microscopy revealed the abnormal morphologies of mitochondria as impaired cristae organization and the appearance of abnormal components in p32 À/À CD51 -TNCs (Figures 4A and S5A), although there was no difference in the mitochondrial mass or membrane potential measured by flow cytometry with MitoTracker Green FM and tetramethylrhodamine methyl ester, respectively ( Figures S5B and S5C). Using an XF-24 extracellular flux analyzer, we next measured the oxygen consumption rate (OCR) as an indicator of mitochondrial OXPHOS and the extracellular acidification rate (ECAR) as an index of lactate production and glycolysis in CD51 -TNCs. As a result, p32 À/À CD51 -TNCs exhibited a lower OCR and higher ECAR compared with controls ( Figures 4B and 4C). These results suggest that p32/C1qbp deletion impairs mitochondrial OXPHOS that promotes a metabolic shift between two major metabolic systems, OXPHOS and glycolysis, to generate ATP in CD51 -TNCs, as observed previously in dendritic cells (Gotoh et al., 2018). In addition, an inhibitor of mitochondrial translation, CAM, exerted inhibitory effects on erythroid and Blymphoid differentiation of CD51 -TNCs, and rotenone and antimycin-A, inhibitors of the mitochondrial electron transport chain, and oligomycin, an inhibitor of mitochondrial ATP synthase, exerted the same effects on CD51 -TNC differentiation ( Figures 4E and 4F) (Sasaki et al., 2017). To directly measure metabolites associated with mitochondrial OXPHOS, we performed a mass spectrometric analysis of sorted WT and p32 À/À CD51 -TNCs. Using a statistical cutoff (p < 0.05), we identified several metabolites that showed differential abundance in p32 À/À CD51 -TNCs ( Figures 4G and S5D). Consistent with a previous study (Gotoh et al., 2018), intermediate metabolites of the tricarboxylic acid cycle, including citrate and isocitrate, were decreased in p32 À/À CD51 -TNCs ( Figure 4H). We also found that pyruvate dehydrogenase (PDH) activity was decreased in p32 À/À CD51 -TNCs ( Figure S5E), suggesting that p32/C1qbp regulates mitochondrial OXPHOS via PDH activity in CD51 -TNCs. Consistent with these results, CPI-613 (6,8-bis octanoic acid), which is a selective inhibitor of PDH and a-ketoglutarate dehydrogenase (Zachar et al., 2011), also exerted inhibitory effects on erythroid and B-lymphoid differentiation of CD51 -TNCs ( Figure 4I).
Mitochondria are critical for heme and iron metabolism because inhibition of mitochondrial translation and OXPHOS are associated with sideroblastic anemia (Ducamp and Fleming, 2019;Fleming, 2011). Heme synthesis begins in mitochondria in which decarboxylative condensation of glycine and succinyl-coenzyme A (CoA) produces 5-aminolevulinic acid (5-ALA) that is a critical product for the porphyrin synthetic pathway. We found no reduction in the amounts of 5-ALA in p32 À/À CD51 -TNCs, although the level of glycine was decreased ( Figures S5D and S5F). We also evaluated expression levels of genes associated with erythroid differentiation by qPCR and p32 À/À CD51 -TNCs exhibited no reduction in these gene expression (Figure S5G). These results suggest that p32/C1qbp plays important roles in terminal differentiation of CD51triple-negative erythroid and B-lymphoid progenitors by regulating mitochondrial OXPHOS rather than heme synthesis or gene transcription.
p32cKO Mice Are Susceptible to Hemolysis Due to Erythroid Differentiation Failure We further investigated whether recovery of erythrocytes could be dependent of CD44 + CD51 -TNCs using a phenylhydrazine (PHZ)-induced hemolytic anemia model. When control and p32cKO mice were treated with a single dose of PHZ (80 mg/kg), all p32cKO mice died due to severe anemia within 6 days after injection, whereas control mice showed a 100% survival rate ( Figures 7A and 7B). In the bone marrow, the proportions and numbers of Ly6D -CD44 + CD51 -TNCs, which were enriched with pre-proerythroblasts (Boulais et al., 2018), were increased significantly in p32cKO mice after PHZ administration ( Figures 7C and 7D). Sorted p32 À/À Ly6D -CD44 + CD51 -TNCs failed to give rise to mature erythroid colonies in vitro ( Figure 7E). We also examined the numbers of pre-CFU-Es and CFU-Es in the bone marrow after PHZ injection ( Figures  S7A and S7B). The numbers of pre-CFU-Es increased and those of CFU-Es decreased in p32cKO mice compared to controls, suggesting erythroid differentiation was blocked at the stage between these two populations. These results imply that expansion of Ly6D -CD44 + CD51 -TNCs contributes to the recovery from hemolytic stresses in cooperation with CFU-Es.
Our data indicate that the p32/C1qbp-mTORC1 axis is essential for terminal differentiation of CD51 -TNCs into erythrocytes ( Figure 6). Next, we treated C57/BL6 mice with Torin 1. As observed in p32cKO mice, Torin 1-treated mice exhibited severe anemia and no mice survived after a single injection of PHZ (80 mg/kg) ( Figures 7F and 7G). In addition, sorted Ly6D -CD44 + CD51 -TNCs after treatment with Torin 1 failed to give rise to mature erythroid colonies in vitro ( Figure 7H). Taken together, these data demonstrate that iScience Article p32/C1qbp is important for activation of mTORC1 signaling necessary for terminal erythrocyte differentiation from CD44 + CD51 -TNCs in acute hemolytic crises.

DISCUSSION
Collectively, the present data show that p32/C1qbp, which has been shown to regulate mitochondrial protein synthesis , is essential for terminal differentiation of CD45 -TER119 -CD31erythroid/ B-lymphoid progenitors. CD45 -TER119 -CD31 -TNCs were classically isolated as non-hematopoietic stromal cells in bone marrow, but a recent study found that this population contains common erythroid/ Erythropoiesis is a multistep process as a series of erythroid-committed progenitors, erythroid burst-forming unit cells, CFU-E cells, proerythroblasts, and erythroblasts, during which numerous transcriptional and metabolic changes occur (Hattangadi et al., 2011). Among them, CFU-E cells have the potential to proliferate rapidly in response to acute anemia, and the importance of mTOR signaling at the proerythroblast stage has been reported (Hattangadi et al., 2011). Our findings show that inhibition of mTOR signaling by p32/C1qbp deficiency or a selective ATP-competitive inhibitor of mTOR, Torin 1, blocked CD51 -TNCs including pre-proerythroblasts from differentiating into erythrocytes. Expansion of CD44 + CD51 -TNCs contributes to recovery from hemolytic crises in cooperation with the bona fide erythroid precursor, CFU-Es. Furthermore, p32/C1qbp-mediated pathways in erythroid differentiation are shared by CD44 + CD51 -TNCs and CFU-E cells.
mTORC1 is a protein complex composed of mTOR, a key regulator of protein synthesis that also supports mitochondrial functions. mTORC1-mediated protein translation is closely associated with mitochondrial biogenesis in erythropoiesis (Liu et al., 2017) (Knight et al., 2014) (Malik et al., 2019. During erythropoiesis, mTORC1 signaling is upregulated and accompanied by increases of the mitochondrial mass, mtDNA, MMP, and protein synthesis. Our electron microscopic analysis showed that CD51 -TNCs possessed abundant mitochondria and ribosomes unlike other hematopoietic cells ( Figure 4A). These findings may explain our observations that mitochondrial dysfunctions caused by p32/C1qbp deficiency prominently influenced the differentiation of CD51 -TNCs, and that Torin 1 was less effective for B-lymphocyte differentiation.
Furthermore, our data suggest that p32/C1qbp regulates mTORC1 signaling by upregulating ATF4 and Sestrin2 that negatively regulate mTORC1. p32/C1qbp deficiency impairs the mitochondrial structure and OXPHOS in CD51 -TNCs, which may be involved in the molecular pathway inactivating mTORC1 whose activity is influenced by intra-cellular and extra-cellular factors such as iron intake, nutrients, hypoxia, and DNA damage (Knight et al., 2014). Previous studies have shown that inhibition of mTORC1 signaling is a major event that causes defective erythropoiesis and vulnerability to hemolytic crisis. Understanding the molecular mechanisms by which mitochondrial protein synthesis is associated with the mTORC1 activation pathway may provide therapeutic cues for anemia and B-lymphopenia.
In the aspect of clinical relevance to this study, several studies report that mutations and single nucleotide polymorphisms (SNPs) of genes related to mitochondrial DNA, OXPHOS and translation were discovered in human patients exhibiting anemia (Ducamp and Fleming, 2019) and that 78.2% of the patients with mitochondrial disorders had anemia (Finsterer and Frank, 2015), implying a close association between mitochondrial dysfunctions and erythroid differentiation. Although, thus far, mutations or SNPs of p32/ C1qbp have found to associate with progressive external ophthalmoplegia or exacerbations of myopathy and influenza infection in human, respectively (Feichtinger et al., 2017) (Chatzopoulou et al., 2018), further studies might reveal genetic anomalies of p32/C1qbp in mitochondrial disorders exhibiting anemia and B-lymphopenia with currently unidentified pathologies. In this study, we have shown that mitochondrial p32/C1qbp is essential for terminal differentiation of CD51 -TNCs into erythrocytes and B-lymphocytes. In the aspect of mechanisms, we focused on the CD51 -TNCs. Genetic deletion of p32/C1qbp in hematopoietic cells causes not only reductions of lymphocytes but also decreases of myeloid cells, which was not significant after the bone marrow transplantation, implying that p32/C1qbp has roles on HSCs. Further investigation under severe hematopoietic stresses as competitive or serial transplantation is required to conclude the functions of p32/C1qbp in HSCs.

Resource Availability Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Kazuhito Gotoh (gotou.kazuhito.712@m.kyushu-u.ac.jp).

Materials Availability
All mouse lines and reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

Data and Code Availability
The data that support the findings of this study are available from the Lead Contact upon reasonable requests. RNA sequence data have been deposited to Mendeley Data: https://doi.org/10.17632/ w25nchf7cp.1.

METHODS
All methods can be found in the accompanying Transparent Methods supplemental file.

ACKNOWLEDGMENTS
This work was supported by JSPS KAKENHI Grant Numbers JP18K11077 and JP16K19196 to K.G., JP18H02841 to Y.K., JP17H04208 and JP19K22638 to F.A., JP15H04764 and JP24590387 to T.U., JP20H00530 and JP17H01550 to D.K.. This work was supported by a grant from the Takeda Science Foundation (to K.G. and Y. K.). We would like to acknowledge all of our colleagues in Dr. Kang's and Dr. Arai's laboratory for their support throughout this project. We appreciate the technical support from the Research Support Center, Graduate School of Medical Sciences, Kyushu University, and the Medical Institute of Bioregulation, Kyushu University. We thank R. Ugawa for performing the transmission electron microscopic observations. We also thank M. Arico from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.  Muta, T., Kang, D., Kitajima, S., Fujiwara, T., and Hamasaki, N. (1997). p32 protein, a splicing factor 2-associated protein, is localized in mitochondrial matrix and is functionally important in maintaining oxidative phosphorylation. J. Biol. Chem. 272, 24363-24370. Nagao, T., and Mauer, A.M. (1969). Concordance for drug-induced aplastic anemia in identical twins. N. Engl. J. Med. 281, 7-11. Nagasawa, T. (2006). Microenvironmental niches in the bone marrow required for B-cell development. Nat. Rev. Immunol. 6, 107-116.

OPEN ACCESS
iScience 23, 101654, November 20, 2020 15 iScience Article Figure S1. Mitochondrial mRNA, and Protein Expression Levels in BM, Related to Figure 1 (A) Deletion of p32 was analyzed using western blotting in BM from WT and p32cKO mice.
-Actin was evaluated as an internal control.
(B) Subcellular distribution of p32 in BM determined by biochemical fractionation. Tom20 and GAPDH were evaluated as mitochondrial and cytoplasmic markers, respectively.

Contact for Reagent and Resource Sharing
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Kazuhito Gotoh (gotou.kazuhito.712@m.kyushu-u.ac.jp).

Animals
C57BL/6 mice were purchased from Japan Clea. Vav1-iCre mice (Stock no: 008610) were obtained from Jackson Laboratory. p32 flox/flox mice have been described previously . Age-and sex-matched p32 flox/flox Vav1-iCre + (p32 cKO) and control littermate p32 flox/flox Vav1-Cre -(WT) mice were used in this study. All mice were maintained on the C57BL/6 background and kept under specific pathogen-free conditions in the animal facility at Kyushu University. The animal protocols were approved by the Committee of Ethics on Animal Experiments, Faculty of Medical Sciences, Kyushu University.

In vivo treatments
WT and p32cKO mice were matched for age and sex. The mice were anesthetized and intraperitoneally injected with 5-FU (150 or 250 mg/kg body weight) or phenylhydrazine (80 mg/kg body weight) at 8-12 weeks of age. Mouse survival was monitored for up to 60 days after injection. Hematological parameters were determined using a K-4500 automatic analyzer (Sysmex).
For bone marrow transplantation, 1×10 6 bone marrow cells from WT and p32cKO mice were transplanted into lethally irradiated recipient.

Flow cytometric analysis and cell sorting
To analyze hematopoietic cells, BM cells were flushed and dissociated by gently passing through a 21 G needle. Ammonium chloride was used for red blood cell lysis. To analyze TNCs, BM plugs were flushed and digested sequentially in HBSS buffer containing collagenase type IV (2 mg/mL, GIBCO) and dispase (1 mg/mL, GIBCO) three times for 10 minutes each at 37°C. The supernatant was collected between digestions and pooled in a tube containing ice-cold FACS buffer (PBS with EDTA 2 mM, BSA 0.1%, and 0.05% NaN3). Cells were stained in PEB buffer (PBS with 0.5% BSA and 2 mM EDTA) for 30 min on ice. Multiparametric flow cytometric analyses were performed on a FACS Verse with BD FACSuite software (BD Biosciences). Data were analyzed by FlowJo software (Tree Star). Cell sorting was performed using an SH800 (Sony) and Aria Cell Sorter (BD).
For intracellular staining, BM cells were fixed with 4% (wt/vol) paraformaldehyde/PBS (Wako Pure Chemical Industries) and permeabilized with 0.2% (wt/vol) Triton X-100/PBS for 15 min at room temperature. After blocking with 1% bovine serum albumin (BSA)/PBS for 30 min, the cells were incubated with primary antibodies in 1% BSA/PBS for 1 hour. Then, the cells were washed with PBS and incubated with an Alexa 488-labeled anti-rabbit secondary antibody for 1 hour.

Immunoblot analysis
For direct immunoblotting, BM cells were lysed in cell lysis buffer (Cell Signaling Technology) and subjected to immunoblotting using specific antibodies.

Quantitative real-time PCR analyses
Total RNA was extracted with an RNeasy Tissue Kit (QIAGEN) and CellAmp Direct Lysis and RT Kit (Takara). Reverse transcription of approximately 650 ng total RNA was performed with random hexamer primers using a PrimeScript RT Reagent Kit (Takara). Expression of mitochondrial genes was detected by qPCR with a thermal cycler (StepOne plus; Applied Biosystems). Ribosomal 18S rRNA was evaluated as an internal control.

Transmission electron microscopy
Sorted CD51 -TNCs were immersed in 0.