SIRT2 is required for efficient reprogramming of mouse embryonic fibroblasts toward pluripotency

The role of sirtuins (SIRTs) in cancer biology has been the focus of recent research. The similarities between underlying pathways involved in the induction of pluripotent stem cells and transformation of cancer cells revealed the role of SIRTs in cellular reprogramming. Seven SIRTs have been identified in mammals and downregulation of SIRT2 was found to facilitate the generation of primed pluripotent stem cells, such as human induced pluripotent stem cells. Herein, we evaluated the role of SIRT2 in naive pluripotent stem cell generation using murine cells. We found that absolute depletion of SIRT2 in mouse embryonic fibroblasts resulted in a notable reduction in reprogramming efficiency. SIRT2 depletion not only upregulated elements of the INK4/ARF locus, which in turn had an antiproliferative effect, but also significantly altered the expression of proteins related to the PI3K/Akt and Hippo pathways, which are important signaling pathways for stemness. Thus, this study demonstrated that SIRT2 is required for cellular reprogramming to naive states of pluripotency in contrast to primed pluripotency states.


Introduction
Sirtuins (SIRTs) are highly conserved NAD + -dependent deacetylases 1 . In mammals, there are seven different SIRTs (SIRT1-SIRT7) with discrete subcellular localizations and distinct functions 2 . SIRT1, SIRT6, and SIRT7 are mainly located in the nucleus, SIRT2 is mainly in the cytoplasm, and SIRT3, SIRT4, and SIRT5 are localized to the mitochondria 3 . Because SIRTs play a key role in maintaining genomic integrity by coordinating cellular responses to various stresses, their aberrant regulation causes tumorigenesis 4 .
According to previous studies, overlapping mechanisms control induced pluripotent stem cell (iPSC) production and tumorigenesis 5,6 . A study comparing the transcriptomes of iPSCs and oncogenic foci (a tumor cell mass created in vitro) from common parental fibroblasts revealed many similarities 7 . Thus, pluripotency and tumorigenicity appear to be closely associated; consequently, SIRTs may be related to cellular reprogramming.
Several reports have described a correlation between SIRTs and iPSC reprogramming efficiency. SIRT1 not only enhances iPSC generation through p53 deacetylation, but also is required for proficient post-reprogramming telomere elongation 8,9 . Because SIRT1 is the closest mammalian homolog of yeast Sir2, it has been the most extensively studied SIRT in mammals. Other SIRTs (SIRT2-SIRT7) have received less attention in this regard; a previous study revealed that SIRT6 improves iPSC reprogramming efficiency in aged human dermal fibroblasts by regulating miR-766 transcription 10 . Another study showed that pluripotency genes are upregulated by silencing of SIRT3 in bovine fibroblasts; however, the exact role of SIRT3 in iPSC reprogramming remains unclear 11 . SIRT2 is primarily found in the cytoplasm where it transiently localizes to the nucleus during the G2/M phase. As a class III histone deacetylase, it deacetylates histone H4 at lysine 16 upon migration to the nucleus 12 . Thus, SIRT2 has been mainly studied for its role in regulating mitosis 13,14 . Because cancer is a consequence of uncontrolled cell division and proliferation, many researchers have focused on the role of SIRT2 in tumorigenesis, as SIRT2 is involved in cell cycle progression, cellular necrosis, and cytoskeleton reorganization 13,15 . Whether SIRT2 is a tumor suppressor [16][17][18][19] or oncogene 20-23 remains controversial. Recently, it was reported that suppression of SIRT2 by miR-200c alters the acetylation levels of glycolyic enzymes, which in turn facilitates cellular reprogramming during human induced pluripotency 24 . Human iPSCs and mouse iPSCs have different characteristics, including in their metabolic strategies, as they exist in primed and naive states, respectively 25 . However, the role of SIRT2 in murine cell reprogramming toward pluripotency has not been examined.
In this study, we found that complete depletion of SIRT2 prevents the generation of pluripotent stem cells from mouse embryonic fibroblasts (MEFs). We also demonstrated the production of functionally competent naive iPSCs with self-renewal capacity that differentiated into three germinal layers both in vitro and in vivo with blastocyst chimera formation, even from SIRT2-knockout (KO) MEFs; however, reprogramming efficiency was significantly low.

Semi-quantitative and quantitative reverse transcription PCR
Total RNA was extracted from cells or tumor tissues using TRIzol reagent (#15596-026, Thermo Fisher Scientific, MA, USA). Extracted RNA was then used as a template for reverse transcription into complementary DNA (cDNA) with an M-MLV cDNA synthesis kit (#28025013, Thermo Fisher Scientific, MA, USA). Semiquantitative RT-PCR was performed using the primers shown in Table S1. cDNA was also amplified by quantitative PCR using a Rotor-Gene Q (#9001550, Qiagen, Hilden, Germany) with a Rotor-Gene SYBR Green PCR kit (#204174, Qiagen, Hilden, Germany); primers used are listed in Table S2. The results were evaluated by Rotor-Gene Q series software. The threshold cycle was determined and the relative gene expression ratio was calculated as follows, fold-change = 2 ΔΔCt . Experiments were performed in triplicate.

Differentiation of iPSCs
Both SIRT2-WT-iPSCs and SIRT2-KO-iPSCs were randomly differentiated into three germ layers through the formation of embryoid bodies using a suspension culture method on bacterial culture dishes, as described previously 26 .

Teratoma formation and histopathology
Six-week-old male Balb/c nu/nu mice (Orient Bio, Gyeong-gi, South Korea) were used in this assay. Animal experiments were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Kyungpook National University (KNU 2014 − 0167). SIRT2-WT/KO-iPSCs (5 × 10 6 ) were counted and mixed with Matrigel (#8482, BD Biosciences, NJ, USA). The mixture was then injected subcutaneously into the dorsal part of nude mice. After 3 weeks, mice were killed and tumor tissues from the injected sites were isolated for histopathological and molecular analyses. Half of the tumor tissues were frozen for molecular studies and the rest were fixed in 10% neutral-buffered formalin, processed using routine methods, and embedded in paraffin wax. The samples were cut into 4-μm thick sections and then deparaffinized in toluene and rehydrated in a graded alcohol series. The sections were then stained with hematoxylin and eosin.

Chimera formation analysis
iPSCs were electroporated with pCXLE-EGFP (#27082, Addgene, Cambridge, UK) to detect their contribution in chimeric embryos. GFP + iPSCs were aggregated with denuded post-compaction eight-cell-stage embryos to obtain an aggregate chimera. Eight-cell embryos flushed from 2.5 days post coitum (dpc) B6D2F1 female mice (Charles River, Cambridge, UK) were cultured in microdrops of embryo culture medium. After the cells were trypsinized, clumps of iPSCs were selected and transferred into microdrops containing zona-free eight-cell embryos. Morula-stage embryos aggregated with iPSCs were cultured overnight at 37°C in 5% CO 2 . The aggregated blastocysts were then transferred into the uterine horns of 2.5 dpc pseudopregnant recipients.
Cell cycle analysis SIRT2-WT/KO-MEFs were incubated at 100% confluency for 72 h without addition of fresh medium for cell cycle synchronization following a protocol described in a previous study 27 . Synchronized cells were then subcultured to restart the cell cycle and then harvested for propidium iodide staining after 24 h of culture. Cells (1 × 10 6 ) were fixed in cold 70% ethanol for 30 min, washed with phosphate-buffered saline, treated with 100 μg/mL of RNase A (R1253, Thermo Fisher Scientific, MA, USA), and then stained with 400 μL propidium iodide solution (P4864, Sigma Aldrich, MO, USA) at room temperature overnight. Analysis was performed by flow cytometry using a FACSAria III. Experiments were performed in triplicate.

Proteomics
Proteomic analyses were conducted using SIRT2-WT/ KO-MEFs. In brief, 44 paired spots that showed differential expression were selected from 2D gel electrophoresis (2-DE) gels. Spots of interest were digested by trypsin and subjected to liquid chromatography-tandem mass spectrometry analysis. Peptides were searched based on the Mascot algorithm (Matrix Science, MA, USA) and filtered with a significance threshold of P < 0.05. Details are presented in the supporting information. Among the proteins identified with > 95% confidence and having two or more matched peptides, the rank 1 protein of each spot, except for unnamed proteins, was included in the final data analysis. Categorization of proteins and pathway analysis were performed using the DAVID program.

Statistical analysis
All values are presented as the mean ± standard error of the mean. Statistical analyses were performed by Student's t test or one-way analysis of variance followed by Bonferroni or Tukey's honest significant difference post hoc tests for multiple comparisons. Statistical significance values were defined as *P < 0.05 or **P < 0.01.

SIRT2 expression depends on cell potency
SIRT2 was differentially expressed in several cell types depending on their potency. We examined various murine cells, including pluripotent stem cells such as embryonic stem cells and iPSCs, adipose-derived stem cells as multipotent stem cells, and unipotent somatic cells including MEFs and dermal fibroblasts. Expression of SIRT2 showed an inverse correlation with the pluripotency marker OCT4 and naive state marker REX1. Unipotent somatic cells exhibited the most robust SIRT2 expression, whereas pluripotent cells showed weak expression (Fig. 1a). A similar tendency was observed during the spontaneous differentiation of embryonic stem cells (Fig. 1b). SIRT2 expression increased as differentiation progressed; however, a slight decrease was observed by day 9 along with further differentiation. This was not a lineage-specific phenomenon, as the same expression pattern was observed during retinoic acid-induced embryonic stem cell differentiation (data not shown). We also examined whether the expression of SIRT2 continuously decreased during pluripotency reprogramming using a feeder-free reprogramming strategy. Unexpectedly, SIRT2 increased before the mid-stage of reprogramming, after which a sudden decrease was observed (Fig. 1c).
SIRT2 deficiency inhibits cellular reprogramming to naive pluripotent stem cells SIRT2-KO-MEFs showed delayed progression in reprogramming. Mesenchymal-epithelial transition was not observed in KO cells by day 8 of reprogramming and most colonies formed by KO cells on day 15 were non-ESlike, exhibiting a flat shape and blunt border (Fig. 2a). Alkaline phosphatase (AP) staining revealed only a few positive colonies of KO cells; however, there was no distinct difference between wild-type (WT) cells and heterotype (HT) cells (Fig. 2b). Thus, further experiments without HT cells were performed to compare WT and KO cells. Based on FACS analysis, the CD44 − /SSEA-1 + and CD44 − /OCT4 + fractions indicative of cells in early or late stages of reprogramming, respectively, were also significantly diminished in KO cells compared with in WT cells (Fig. 2c, d).
SIRT2-KO-iPSCs are in a state of naive pluripotency SIRT2-KO cells showed decreased ES-like iPSC colony formation compared with their WT counterparts. Thus, it was unknown whether SIRT2-KO-iPSCs were functional pluripotent stem cells. To address this, both SIRT2-WT-iPSCs and SIRT2-KO-iPSCs were simultaneously characterized and compared. iPSC lines were established through picking and propagation of independent ES-like colonies. Finally, five pairs of SIRT2-WT-iPSC and SIRT2-KO-iPSC lines were selected, as these clones were highly proliferative and could be passaged for long periods of time without differentiation. When these undifferentiated cells were subjected to immunofluorescence, both iPSC lines expressed the pluripotent stem cell marker OCT4 in the nucleus, irrespective of SIRT2 expression in the cytoplasm (Fig. 3a, b). Furthermore, these cell lines differentiated into tissues from all three germinal layers upon differentiation through embryoid body formation in vitro (Fig. 3c, 3d).
Moreover, SIRT2-KO-iPSCs and SIRT2-WT-iPSCs formed tumors upon injection into immunodeficient mice (Fig. 4a). The size and weight of tumor tissues derived from SIRT2-WT-iPSCs and SIRT2-KO-iPSCs were similar (Fig. 4b). Based on histopathology, the tumors were diagnosed as teratomas, which contain derivatives of all three germ layers. Pigment cells or epidermis were identified as ectoderm-derived components, skeletal muscle, and cartilage as mesoderm-derived components, and respiratory epithelia or acinar cells as endoderm-derived components (Fig. 4c). Thus, both iPSC lines are functional pluripotent stem cells, as they could differentiate into all three germ layers in vivo and in vitro.

SIRT2 depletion impairs cell cycle progression
Although SIRT2-KO-MEFs were hypothesized to be reprogrammed better than their WT counterparts as was observed for human cells, the results did not support this hypothesis. However, the resultant KO-iPSCs functioned similarly to wild-type iPSCs. Therefore, we focused on the differences between SIRT2-WT-MEFs and SIRT2-KO-MEFs to identify the reason for the decreased reprogramming efficiency in KO cells. During cell culture, SIRT2-KO-MEFs required more time to reach confluency. Based on manual counts (four times at 24-h intervals) and WST-1 assays, KO cells showed reduced proliferation compared to normal cells (Fig. 5a, b). Because SIRT2 is a well-known cell cycle regulator, we predicted that the difference was because of the different cell cycle distribution between SIRT2-KO-MEFs and WT MEFs. As expected, SIRT2-KO-MEFs displayed an abnormal pattern, as 2 N and 4 N peaks partially overlapped without separation by an S-phase fraction (Fig. 5c). This indicates that depletion of SIRT2 resulted in aberrant cell cycle progression, in turn causing a delay in cell growth and reprogramming. Because genes involved in both cell cycle regulation and iPSC reprogramming required further investigation, we focused on the INK4/ ARF locus based on a previous study 28 . The main transcriptional products of INK4/ARF and their downstream gene expression were analyzed by quantitative PCR. p16 Ink4a , p15 Ink4b , Rb, p19 Arf , and p21, but not p53, were upregulated in SIRT2-KO-MEFs compared with in WT cells. Particularly, p16 Ink4a was markedly differentially expressed between the cell types, showing an~80-fold increase in mRNA expression levels in KO cells compared with in WT cells (Fig. 5d). The difference in p16 Ink4a expression between the two cell types was also confirmed at the protein level. INK4/ARF products are rarely expressed in growing primary cells; indeed, WT MEFs exhibited virtually undetectable expression. However, INK4/ARF protein expression was observed in KO cells (Fig. 5e). Moreover, all members of the SIRT family were analyzed because the deletion of SIRT2 may affect the expression of other SIRTs, some of which are known to regulate reprogramming efficiency. Lower expression levels of SIRT1 were observed in KO-MEFs, whereas SIRT5 expression was higher in KO-MEFs, showing an inverse correlation with SIRT2 expression. SIRT6 expression did not significantly differ between the two cell types (Fig. 5f). Although neither SIRT1 and SIRT5 expression levels was affected when the SIRT2 was artificially downregulated by transfection of siRNA for SIRT2 ( Figure S4E), the inverse correlation between SIRT2 and SIRT5 was consistently observed when their natural expression levels were compared in between the wild-type cells and SIRT2-KO cells, including iPSCs ( Figure S4F).

Proteome profile depends on SIRT2 expression of MEFs
The greatest difference in proliferation rates was observed between wild-type MEFs and SIRT2-KO-MEFs, and thus cell cycle and related regulatory genes were analyzed. In addition, the proteomes of WT and KO cells were compared with identify other underlying causes of decreased reprogramming efficiency of SIRT2-knockout MEFs. In 2-DE gel analysis, 215 paired protein spots and 237 non-paired protein spots were detected. Among the paired spots, 25 were morphometrically increased by more than twofold in KO cells compared with in WT cells. In addition, 19 spots decreased by > 50% in KO cells compared with in WT cells. Differentially expressed spots were subjected to liquid chromatography-tandem mass spectrometry. The cutoff criteria at a P value < 0.05 and fold-ratio difference > 2 was strictly adjusted and the rank 1 protein for each spot satisfying the criteria is summarized in Table 1. Subsequently, differentially expressed proteins were analyzed for biological process, cellular component, and molecular function. Among the 242 affected biological processes, seven were significant at P < 0.05 with Bonferroni correction; all were related to conformational changes ( Figure S5A). Among the 65 affected cellular components, 18 were significant at P < 0.05 with Bonferroni correction; the top 10 components, which contained cytoplasm or exocytosis-related compartments, are shown in Figure S5B. Among the 43 affected molecular functions, nine were significant at P < 0.05 with Bonferroni correction; these were all binding functions, which were further classified into various biological activities ( Figure S5C). Next, pathways affected by SIRT2 depletion were analyzed using the KEGG database. The top 10 significant pathways (among 19 pathways) at P < 0.05 are presented in Figure S5D. Consistent with these findings, the cell cycle-related pathway was identified. Moreover, the pathways most affected by SIRT2 depletion were the PI3K-Akt signaling pathway followed by the Hippo signaling pathway, both of which are known to be closely related to iPSC reprogramming.

Discussion
Because iPSCs were first generated using Yamanaka factors with an efficiency of~0.1%, numerous studies have been conducted to improve reprogramming efficiency. Typically, introduction of genes that are highly expressed in ESCs, such as Nanog and Sall4, provokes more efficient pluripotency reprogramming 29,30 . Similarly, artificial inhibition of genes that show low expression in ESCs, such as Mbd3 and p53, typically augments reprogramming 31,32 . Because SIRT2 displayed low expression levels in ESCs, its inhibition was expected to increase reprogramming efficiency, as reported for human cell reprogramming toward pluripotency 24 .
However, we demonstrated that complete depletion of SIRT2 paradoxically decreased reprogramming efficiency in murine cells. This result may be attributed to the high transient expression of SIRT2 during the mid-stage of reprogramming, rather than a gradual decrease from day 0 to the end of iPSC reprogramming (Fig. 1c).  The epithelial-mesenchymal transition factor SNAIL paradoxically enhances reprogramming efficiency 33 ; similarly, stage-specific elevation of SIRT2 appears to be required for successful reprogramming. After SIRT2 depletion, the kinetics of reprogramming and resultant iPSC colonies were decreased compared with those of WT or HT cells. Based on a previous study 34 that used the same inducible lentiviral system for reprogramming as this study, AP was found as the most initially activated pluripotency marker and subsequent activation of SSEA-1 represented a transgene-dependent partially reprogrammed stage. When the MEFs were fully reprogrammed and independent from the transgenes, OCT4 was expressed 34 . We found that initial activation of AP was most severely inhibited in SIRT2-KO-MEFs during the initial stage of reprogramming, although sequential activation of SSEA-1 and OCT4 were also depressed under SIRT2-KO conditions (Fig. 2). Because SIRT2 expression gradually increased from the initial stage to the mid-stage of reprogramming, SIRT2 may play a positive role in AP activation during this period and before activation of SSEA-1.
As the induction and maintenance of pluripotency both require activation of similar gene sets, we examined whether SIRT2-KO-iPSCs are fully functional iPSCs. Although depletion of SIRT2 prevents the reprogramming process, established ES-like iPSC lines from SIRT2-KO cells showed characteristic features of pluripotent stem cells, such as a large N/C ratio, rapid proliferation, and OCT4 expression, when cultured on feeder cells with ES media. This result is also consistent with that of a previous study, which demonstrated that SIRT2 knockdown did not affect OCT4 and Nanog expression in mouse ESCs 35 . When cultured without leukemia inhibitory factor, SIRT2-KO-iPSCs spontaneously formed embryoid bodies in non-adherent conditions and were further differentiated into the ectodermal, mesodermal, and endodermal linages. The functionality of these cells was also demonstrated in vivo by teratoma and chimera formation. Although significantly reduced in number, SIRT2-KO-iPSCs retained pluripotent competency, indicating that there is functional redundancy of SIRT2. There is evidence that not only SIRT2 but also SIRT1 possess H4K16Ac deacetylase activity; this suggests a synergistic relationship or functional redundancy between SIRT1 and SIRT2 36 . Because SIRT1 also plays a positive role during reprogramming 9 , it may be a functional substitute for SIRT2 in SIRT2-KO cells.
To further explore this phase-specific effect mediated by SIRT2 depletion during reprogramming, we compared WT MEFs and SIRT2-KO-MEFs. Although many previous studies of SIRT2 have used immortalized cells for research 17,37,38 , immortalized cells may contain malfunctioning tumor suppressor genes. Although immortalized MEFs generated by the classic 3T3 protocol are known to possess functional p53 39 , these cells are not identical to primary MEFs, as they were shown to be resistant to miR-290 senescence cues, irrespective of p53 status 40 . In addition, the miR-290 family is known to be expressed abundantly in mouse embryonic stem cells and aid in reprogramming of somatic cells toward pluripotency 41 . Therefore, immortalized MEFs may not reproduce normal cell characteristics, particularly in the field of iPSC research. Thus, primary MEFs between passage numbers 1 and 5 were used in the present study.
SIRT2 is known for its cell cycle regulatory function. Previous studies showed that SIRT2 is upregulated during mitosis and that its overexpression prolongs mitosis 42 and shortens the G1 phase 43 . Another study using primary MEFs derived from SIRT2-knockout mice reported longer G1 and shorter S phases, with rarely observed effects on mitosis duration 12 . Kim et al. 17 recently reported that SIRT2 controls mitotic exit by regulating the anaphasepromoting complex and cyclosome activity. We also confirmed that SIRT2-knockout MEFs displayed aberrant cell cycle progression and thus considered the correlation between cell cycle and iPSC generation. The INK4/ARF tumor suppressor locus is known not only for its role in cell cycle control 44 but also as a barrier for reprogramming 28 . p16 Ink4a and p15 Ink4b bind to and inhibit the cyclin D-dependent kinases Cdk4 and Cdk6 to subsequently relieve the cell cycle inhibitory effects of Rb, whereas p19 Arf binds to and inhibits Mdm2 to stabilize the tumor suppressor p53 28 . Thus, genes in the INK4/ ARF tumor suppressor locus are not activated under normal physiological conditions in young animals, but show increased expression during aging to mitigate the threats of tumorigenesis by activating both Rb and p53.
(see figure on previous page) Fig. 4 In vivo developmental potential of iPSCs. a WT (left) and KO (right) iPSCs formed tumors in immunodeficient mice. Representative images are from five mice per group. b Tumor tissues from both groups showed no significant differences in size and weight. c Tumor tissues from both groups were diagnosed as teratomas. A WT-iPSC-derived teratoma is shown in the upper row and a KO-iPSC-derived teratoma is shown in the bottom row. From left, ectoderm, mesoderm, and endoderm tissues are indicated by arrows. Scale bars = 200 μm. *P < 0.05, **P < 0.01 compared with wild-type (n = 5). d Chimera formation by aggregation with SIRT2-WT-iPSCs (upper panel) or SIRT2-KO-iPSCs (middle and lower panel). Embryos (13.5 dpc) showed that GFP + cells (SIRT2-WT-iPSCs and SIRT2-KO-iPSCs) contribute to all three germ layers of ectoderm, mesoderm, and endoderm as well as gonads If Rb and p53 become inactivated, cells gain abnormally enhanced proliferative ability 45 . Meanwhile, the INK4/ ARF locus is actively silenced in stem cells, such as embryonic and adult stem cells, to maintain their characteristic proliferative ability 45 . During reprogramming, exogenous transcription factors such as Klf4 and c-Myc can activate INK4/ARF-associated pathways, as they are well-known oncogenes 46,47 . Thus, the INK4/ARF locus should be silenced to facilitate reprogramming, although factors used for reprogramming ultimately activate that  Because p16 Ink4a is rarely expressed in normal primary cells, as shown by our results (Fig. 5e), and the absence of INK4/ARF products in proteomics may be because 2-DE spots from one of the three WT MEF protein samples were set as the standard and only paired spots were used for mass spectrometry. Among the 41 differentially expressed proteins, isocitrate dehydrogenase 2 mitochondrial precursor (IDH2), heat shock protein 90-alpha, endoplasmin precursor, 14-3-3 protein beta, 14-3-3 protein theta, and 14-3-3 protein zeta expression patterns strongly support other data in this study. IDH2 was more abundant in WT MEFs than in KO and is responsible for the production of NADPH, whereas NADPH is essential for the proliferation of both normal and tumor cells 52 .
In addition, NADPH is needed to establish cancer-like glycolytic phenotypes, which is further required for establishing and maintaining a DNA methylation pattern in iPSCs resembling that of ESCs 53 . Recently, it was reported that primed human pluripotent stem cells require SIRT2 downregulation to maintain pluripotency via altering metabolic pathways toward aerobic glycolysis 24 . In contrast, naive state mouse ESCs are bivalent in their energy production, as they can switch from glycolysis to oxidative phosphorylation on demand 54 . Thus, different metabolic states between human and mouse cells may provoke different responses upon SIRT2 depletion. HSP90 has two cytosolic isoforms, HSP90α and HSP90β, and is an important stress protein 55 ; both isoforms were upregulated in WT cells and are known to be crucial for maintaining pluripotency in mouse ESCs by regulating OCT4 and Nanog 56 . In contrast, 14-3-3 protein isoforms were upregulated in SIRT2-knockout cells; these proteins bind to phosphorylated TRIM32 and induce an increase in soluble free TRIM32 57 , which is known to repress iPSC reprogramming by modulating OCT4 stability 58 .
Based on gene ontology and KEGG pathway analyses, differentially expressed proteins between WT MEFs and SIRT2-KO-MEFs are important for conformational changes and binding processes. Furthermore, many of these proteins were found to be associated with the PI3K-Akt signaling pathway, which is essential for iPSC survival 59 , and Hippo signaling pathway, which acts a barrier to iPSC reprogramming 60 .
In addition, SIRT1 and SIRT5 showed altered expression in SIRT2-depleted MEFs. SIRT1, previously shown to enhance reprogramming 9 , exhibited lower expression levels in SIRT2-KO-MEFs; this may be an additional reason for the poor reprogramming efficiency of these cells. However, it is unlikely that the expression level of SIRT2 directly affects to that of SIRT1 because SIRT1 expression does not changed in case of SIRT2 knockdown by siRNA transfection. Meanwhile, the expression pattern For proteomic analysis, we performed quantification based on biological triplicate experiments. Among 44 spots detected by LC-MS/MS, unnamed rank 1 proteins of three spots were excluded from further analysis. See also Figure S2. b Peptides with ion scores above the 95% confidence level were counted. c KO/WT ratio represents the ratio of proteins detected in SIRT2-knockout MEFs to those of wild-type MEFs. Increased expression, fold-change ≥ 2; decreased expression, fold-change ≤ 0.5.
of SIRT5 contrasted that of SIRT2, but the implication of this phenomenon is unknown as the exact association of SIRT5 with stem cells or tumorigenesis remains unclear.
The present study provides important information regarding SIRT2-mediated alteration of gene sets required for successful reprogramming towards pluripotency. SIRT2 itself is dispensable for reprogramming towards naive pluripotency and subsequent differentiation of murine cells. However, its depletion exacerbates reprogramming-induced stress through senescence marker activation and significantly reduces reprogramming efficiency. Although the role of SIRT2 in pluripotent stem cells during phase change from naive to primed states requires further analysis, SIRT2 may assist in the efficient reprogramming of murine somatic cells to naive pluripotency (Fig. 6).