Comprehensive Identification of Krüppel-Like Factor Family Members Contributing to the Self-Renewal of Mouse Embryonic Stem Cells and Cellular Reprogramming

Pluripotency is maintained in mouse embryonic stem (ES) cells and is induced from somatic cells by the activation of appropriate transcriptional regulatory networks. Krüppel-like factor gene family members, such as Klf2, Klf4 and Klf5, have important roles in maintaining the undifferentiated state of mouse ES cells as well as in cellular reprogramming, yet it is not known whether other Klf family members exert self-renewal and reprogramming functions when overexpressed. In this study, we examined whether overexpression of any representative Klf family member, such as Klf1–Klf10, would be sufficient for the self-renewal of mouse ES cells. We found that only Klf2, Klf4, and Klf5 produced leukemia inhibitory factor (LIF)-independent self-renewal, although most KLF proteins, if not all, have the ability to occupy the regulatory regions of Nanog, a critical Klf target gene. We also examined whether overexpression of any of Klf1-Klf10 would be sufficient to convert epiblast stem cells into a naïve pluripotent state and found that Klf5 had such reprogramming ability, in addition to Klf2 and Klf4. We also delineated the functional domains of the Klf2 protein for LIF-independent self-renewal and reprogramming. Interestingly, we found that both the N-terminal transcriptional activation and C-terminal zinc finger domains were indispensable for this activity. Taken together, our comprehensive analysis provides new insight into the contribution of Klf family members to mouse ES self-renewal and cellular reprogramming.


Introduction
Mouse embryonic stem (ES) cells are derived from the inner cell mass of the blastocyst and can be maintained indefinitely in a self-renewing state in culture [1,2]. The ability to direct the differentiation of ES cells toward a specific cell fate is a highly pursued goal in regenerative medicine [3]. However, the utilization of ES cells for therapeutic purposes will require a better understanding of the molecular mechanisms underlying the regulation of pluripotency [4,5]. Previous studies revealed that the pluripotency of ES cells is maintained by multiple soluble factors, such as LIF [6,7], and by nuclear factors [8][9][10][11][12][13][14][15], including putative core transcription factors such as Oct3/4, Sox2, Klf family members and Nanog. A reduction in extracellular-signal-regulated kinase (ERK) activity strongly promotes pluripotency. Activation of the fibroblast growth factor (Fgf)4-Fgf receptor(R)-ERK pathway destabilizes the pluripotent state and promotes a primed state [16][17][18]. Furthermore, the simultaneous inhibition of ERK and glycogen synthase kinase-3 (Gsk3) beta (with the inhibitor 2i) dramatically stabilizes the selfrenewal process of mouse ES cells [19]. However, the precise molecular mechanisms of selfrenewal remain elusive.
Induced pluripotent stem cells (iPSCs) can be derived from lineage-restricted cells, such as fibroblasts, by the forced expression of defined transcription factors [20,21]. Although previous studies have indicated that Krüppel-like transcription factors (Klfs) are essential for the reprogramming of somatic cells into a pluripotent state, the molecular mechanisms underlying these processes remain unknown [10,11,22]. The processes involved in cellular reprogramming from somatic cells and the epiblast to generate iPSCs might be similar.
Epiblast stem cells (EpiSCs) are pluripotent stem cells derived from the epiblast of embryos at the egg cylinder stage and retain the ability to differentiate into all three embryonic germ layers [23,24]. The properties and gene expression patterns of EpiSCs are very similar to those of human ES cells derived from embryos at the blastocyst stage. ES cells and EpiSCs are said to be in a naïve or in a primed state of pluripotency, respectively [25]. EpiSCs can be converted into iPSCs by reprogramming factors such as Nanog, Esrrb, and Klf family members including Klf2 and Klf4 [26,27]. The processes of self-renewal and cellular reprogramming share common transcription factors, indicating that both processes might be governed by similar molecular mechanisms.
Previous analysis indicated that expression of Klf2, Klf4 and Klf5 is associated with an undifferentiated state in mouse ES cells, and loss-of-function gene knockout (KO) studies indicated that a triple KO of Klf2, Klf4 and Klf5 resulted in defective self-renewal, and the introduction of Klf2, Klf4, or Klf5, but not Klf10 rescued the defective self-renewal phenotype [11]. Our previous analysis of Klf5 KO ES cells indicated that a lack of Klf5 resulted in the spontaneous differentiation of mouse ES cells and that the phenotype was rescued by Klf4 expression. Overexpression of Klf2, Klf4, or Klf5 is sufficient to maintain the undifferentiated state of mouse ES cells in the absence of LIF [10,28,29], yet it is still unknown whether other Klf family members have such a function, and the functional domains required for self-renewal and cellular reprogramming are not clear.
Here we report a comprehensive analysis of which of the Klf family members could achieve self-renewal in mouse ES cells and found that only Klf2, Klf4, and Klf5 possessed the ability for LIF-independent self-renewal, although most of the Klf proteins (Klf1-Klf10) could occupy the regulatory regions of Nanog. We also found that Klf5, in addition to Klf2 and Klf4, had the ability to reprogram EpiSCs into a naïve pluripotent state. We also delineated the functional domains of Klf that could enable LIF-independent self-renewal and reprogramming. Remarkably, we found that both the N-terminal transcriptional modulation and C-terminal zinc finger DNA-binding domains were required for LIF-independent self-renewal and cellular reprogramming. This indicated that the ability to reprogram EpiSCs into naïve pluripotent stem cells is correlated with the ability to maintain the pluripotent state. Taken together, our results provide a comprehensive view on the mechanisms involved in self-renewal and reprogramming by Klf family members.

Materials and Methods Plasmids
The coding regions of mouse Klf1-Klf10 were amplified from cDNAs derived from mouse ES cells or tissues, sequenced, and then introduced into the multicloning site of a pPB-FLA-G-HA-CAG-ireshygroR vector (Sanger Institute, Cambridge, UK).

Culture of Mouse ES Cells and Overexpression of Klf
The OCRG9 ES cell line expressing green fluorescent protein (GiFP) (Oct3/4-CFP-irespuroR:: Rex1-GFP) was generated as described previously [30] and a gift from Dr. Hitoshi Niwa (Kumamoto University). All ES cells were maintained in mouse ES medium consisting of Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1 mM nonessential amino acids, 100 μM 2-mercaptoethanol, 1 mM L-glutamine and 1,000 U/ml LIF on gelatin coated dishes as described [10]. In experiments, 1 × 10 7 ES cells were electroporated with 15 μg of total DNA (piggyBac transposase and Klf plasmid; pPB-CAG-FLAG-HA-Klf-ire-shygroR) and cultured in the presence of 150 μg/ml hygromycin B with or without LIF for 7 days; 1 μg/ml puromycin was included in the medium used for the LIF-independent selfrenewal assay to select Oct3/4-positive puromycin-resistant colonies.

Generation of EpiSCs with Stable Overexpression of Klfs
Mouse ΔPE-Oct3/4 EpiSCs were kindly provided by Dr. Austin Smith (Cambridge University) [27]. Transient transfection of mouse EpiSCs was performed on fibronectin-coated dishes seeded with 1 × 10 6 cells using 12 μl Lipofectamine 2000 reagent (Invitrogen) with 4 μg total DNA (piggyBac transposase and Klf plasmid) per well of a six-well culture plate. The EpiSCs were cultured in N2B27 medium supplemented with human activin A (10 ng/ml; R&D) and basic fibroblast growth factor (bFgf) (5 ng/ml; Wako Pure Chemical Industries) as described previously [27] and then selected with 250 μg/ml hygromycin B and 1 μg/ml puromycin for 7 days starting from 24 h post-transfection.
For the chimera experiment, 129 EpiSCs were generated as described previously and obtained from Dr. Paul Tesar (Case Western Reserve University) [24] and transfected with a piggyBac transposase and floxed Klf5 plasmid, and cultured in the presence of 200 μg/ml G418.

Blastocyst Injection
Two independent iPS cell lines reprogrammed from 129 EpiSCs were transfected with a piggy-Bac transposase, UbC-GFP plasmid (from the Sanger Institute) and pCAG-NLS-Cre (a generous gift from Dr. Andras Nagy, Samuel Lunenfeld Research Institute). GFP-positive and Klf5 cassette-negative colonies were screened by polymerase chain reaction (PCR). The GFPlabeled iPSCs were injected into blastocysts that were transferred into surrogate mice. Embryos (9.5 and 13.5 dpc) were recovered and inspected under a fluorescence microscope.

Immunohistochemistry and Alkaline Phosphatase Assay
Cells were fixed in 4% paraformaldehyde for 10 min at 4°C and incubated overnight at 4°C with anti-Nanog antibodies. After a brief wash with phosphate-buffered saline + 0.1% Tween-20, the cells were detected in parallel by Cy3 fluorescence emissions, respectively. Nuclei were stained with Hoechst 33342 (Molecular Probes). The images were captured using a BIOREVO BZ-9000 fluorescence microscope (Keyence). Alkaline phosphatase assays to evaluate the pluripotent state were performed using leukocyte Alkaline Phosphatase kits (Sigma-Aldrich).

Reverse-Transcription (RT) qPCR Analysis
For RT-qPCR analysis, first-strand cDNA was synthesized from total RNA using a QuantiTect Reverse Transcription kit (Qiagen). qPCR was performed with SYBR Premix Ex Taq II (Takara) and analyzed on a Thermal Cycler Dice Real Time System (TP850; Takara). The amount of target RNA was estimated using an appropriate standard curve and divided by the estimated amount of β-actin.

Statistical Analysis
Statistical analyses were performed using unpaired Student's t tests with Microsoft Office Excel. Data are expressed as the mean with standard error. Differences between means were considered significant at P < 0.05.

Results
Although previous reports indicated that overexpression of Klf2, Klf4, or Klf5 achieves LIFindependent self-renewal of mouse ES cells [10,11,28,29], it is still unknown whether other Klf family members might have similar activity. Therefore, we attempted a comprehensive investigation into whether overexpression of representative Klf family members would be sufficient to maintain the undifferentiated state of mouse ES cells. First, we surveyed the phylogenetic tree of the Klf family, which is composed of four subclasses, and selected Klf1-Klf10 out of 18 members, because they represent the four subfamilies (data not shown) [32]. After a FLAG-HA epitope tag was introduced into the N-terminus of each Klf to monitor the protein expression level (Fig 1A), the expression vector was introduced into OCRG9 (Oct3/4-CFP-irespuroR:: Rex1-GFP) ES cells [30], which possesses a Rex1-GFP pluripotency marker. Western blot analysis confirmed that the epitope-tagged Klf proteins were overexpressed in the ES cells at variable levels (Fig 1B). Evaluation of relative expression levels of epitope-tagged Klf proteins normalized to β-actin showed that levels of overexpressed Klf6, Klf7, Klf8 and Klf10 protein were low, but not significantly different from that of Klf5 (S1 Fig). The ES cells were cultured in the presence of puromycin to select Oct3/4-positive cells, in the presence or absence of LIF (Fig 1C and 1D). All the samples, including those expressing Klf1-Klf10, showed many undifferentiated alkaline phosphatase (AP)-positive colonies in the presence of LIF, yet only cells overexpressing Klf2, Klf4 or Klf5 led to significant numbers of AP-positive colonies in the absence of LIF (Fig 1D and 1E). We performed immunohistochemistry on colonies cultured in the absence of LIF and confirmed high levels of endogenous Nanog protein (Fig 2A). We also confirmed GFP fluorescence driven by the Rex1 promoter (Fig 2B), indicating that the undifferentiated state of pluripotency had been maintained.
Overexpression experiments clearly indicated that Klf2, Klf4 and Klf5, but not any other Klf family member, were able to maintain the undifferentiated state in the absence of LIF. It is not known why only these three members have this ability. To investigate the molecular mechanism underlying the pluripotency maintained by Klf family members, we focused on the Nanog locus, because this gene is important for the self-renewal of mouse ES cells [9,14,33] and is controlled by the Klf2, Klf4, and Klf5 transcription factors [10,11]. Previous reports indicated that Klf2, Klf4 and Klf5 activate Nanog through binding to the promoter and its 3 0 enhancer [11]. When the overexpressed Klf proteins were examined for binding affinities to the promoter and enhancer (Fig 3A), most of them except for Klf9 and Klf10 showed efficient recruitment to the regulatory regions at similar levels to Klf2, Klf4 and Klf5 (Fig 3B). Currently, it is not clear why Klf1, Klf3, Klf6, Klf7, and Klf8 are not sufficient to maintain self-renewal of mouse ES cells, although they have the ability to bind to Nanog regulatory regions.
Mouse EpiSCs can spontaneously reprogram into the naïve state of iPSCs, but at a very low rate [34]. Transient expression of somatic reprogramming factors, such as Klf4, Nanog, Esrrb, and Klf2, cause more efficient and quicker reversion of EpiSCs into iPSCs, implying that this reversion mimics cellular reprogramming from fibroblasts into iPSCs. Previous analysis also indicated that Klf2 and Klf4, but not Klf5, enhance reprogramming of EpiSCs into iPSCs [28]. We comprehensively examined the reprogramming ability of Klfs (Fig 4A). Expression vectors carrying epitope-tagged Klf genes were introduced into EpiSCs and protein levels were confirmed using an anti-HA antibody (Fig 4B). When cells were transferred from EpiSC cultures containing bFGF and activin A into ES medium containing LIF and 2i, only the EpiSCs overexpressing Klf2, Klf4, or Klf5 generated significant numbers of AP-positive colonies (Fig 4C). A previous report indicated that Klf5 expression could not enhance reprogramming of EpiSCs to iPSCs [28]; therefore, we carefully evaluated whether iPSCs generated by Klf5 overexpression were indeed pluripotent. The RT-qPCR analysis showed elevated expression levels of naïve markers, such as Nanog, Klf2, Klf4, Esrrb, Stella, and Rex1, while the expression levels of primed state markers, such as Fgf5 and Brachyury were reduced (Fig 4D). Endogenous Nanog protein was also present in the iPSCs (Fig 4E). The Oct3/4 gene contains distal and proximal enhancers; the distal one drives naïve pluripotent stem cells, while both enhancers are required for generating primed pluripotent stem cells [27]. The iPSCs generated by transient Klf5 expression showed strong GFP fluorescence without the presence of the proximal enhancer (Fig 4F), also indicating that the iPSCs were indeed naïve pluripotent stem cells. We then performed chimera formation, which is the gold standard assay for pluripotency (Fig 5). The iPSCs reprogrammed by overexpression of Klf5 were injected into blastocysts and the resultant blastocysts were returned into the uterus of a surrogate mother. E9.5 fetuses generated from two independent iPSCs exhibited strong contribution of GFP-positive iPSCs ( Fig 5A). Furthermore, E13.5 fetuses exhibited strong contribution of GFP-positive iPSCs throughout the whole embryo including the genital ridge (Fig 5B), indicating that iPSCs have the ability to differentiate into all three germ layers, including the germ cell lineage.
To gain insight into how Klf2, Klf4 and Klf5 might regulate reprogramming of EpiSCs to iPSCs, we performed ChIP assays to assess the occupation of Klf factors at the Nanog locus under EpiSC culture conditions. Most of the Klf proteins examined, except for Klf9, occupied the Nanog promoter and enhancer, indicating that occupation itself was not sufficient for reprogramming (Fig 6). Our analysis indicated that Klf2, Klf4, and Klf5 did not affect Nanog transcription under EpiSC culture conditions (data not shown).
To delineate the functional domains required for LIF-independent self-renewal, we compared the amino acid sequences of Klf2, Klf4 and Klf5, and found that these three Klf proteins share homologous regions in the N-terminal half, which acts in transcription, as well as in the C-terminal zinc finger DNA-binding domain (S2 Fig). Because Klf2 is the smallest of these proteins, deletion mutants for Klf2 were created (Fig 7A). Western blot analysis confirmed that exogenous Klfs were overexpressed in ES cells in the presence of LIF (Fig 7B). When LIF-independent self-renewal activity was evaluated, N-terminal deletion resulted in a reduction in APpositive clones (Fig 7C). Deletion of two zinc fingers also abolished the efficiency of this selfrenewal activity. Thus, our analysis clearly indicates that both the N-terminal transcriptional modulation domain and the C-terminal zinc finger DNA-binding domain are required for the full activity of Klfs (Fig 7). We also delineated the functional domains required for reprogramming of EpiSCs into iPSCs (Fig 8). When various Klf deletion mutants were overexpressed in EpiSCs in the presence of bFGF and activin A (Fig 8A), we found that both the N-terminal transcription modulation domain and the C-terminal zinc finger DNA-binding domain were required for full activity (Fig 8B). This was similar to the domains required for self-renewal.

Discussion
Our analysis identified that only Klf2, Klf4, and Klf5 have the abilities to both maintain the undifferentiated state of mouse ES cells and to reprogram EpiSCs into iPSCs when overexpressed. We also delineated functional domains of Klf for LIF-independent self-renewal and reprogramming and found that both the N-terminal transcription modulation domain and the C-terminal zinc finger DNA-binding domain were required for LIF-independent self-renewal and reprogramming. This finding indicates that the ability to reprogram EpiSCs into naïve pluripotent stem cells is correlated with the ability to maintain the pluripotent state.
A previous report by Wang et al. [35] indicated that the reprogramming ability of Oct3/4, Nanog and Sox2 could be boosted by using a chimeric protein including VP16, a potent transcriptional activator, in place of the inherent transcriptional modulation domain, indicating that transcriptional activation is important for the reprogramming ability of those core factors. On the other hand, only a Klf4-VP16 chimeric protein provided a similar level of reprogramming activity, suggesting that the transcriptional modulation domain exerts inherent activity.
EpiSCs can be converted into iPSCs by naïve-state transcription factors such as Nanog, Esrrb, Klf2 and Klf4 [26,27]. Overexpression of Klf2, Klf4 or Klf5 maintain the undifferentiated state of mouse ES cells in the absence of LIF [10,28,29]. It is of note that induced expression of Klf4 and Klf5 are associated with the naïve state of rabbit iPS and ES cells [36], indicating a potential conserved role of Klf factors in the naïve state among species.
Previous "omics" studies using mass spectroscopy indicated that Oct3/4 exists in a large complex containing Sall4, Tcfcp2l1, Dax1, Esrrb, and other epigenetic modifiers in mouse ES cells [37]. Similarly, Nanog binds to Sox2, Mta2, Sall4, and PolII [38]. Interestingly, Klf5 is a protein that forms a complex with Oct3/4 and Nanog [37,38]. Our previous study [10] indicated that Klf5 is in a complex with Metastasis-associated protein (MTA2), a component of Nucleosome Remodeling Deacetylase (NuRD). Our results here will be useful for the further exploration of functional domains that contribute to self-renewal and reprogramming. It will be interesting to identify possible cofactors that are important for self-renewal and reprogramming, leading to elucidation of the molecular mechanisms involved.   Table. Primers used for RT-qPCR analysis. (PDF)