YY2 in Mouse Preimplantation Embryos and in Embryonic Stem Cells

Yin Yang 2 encodes a mammalian-specific transcription factor (YY2) that shares high homology in the zinc finger region with both YY1 and REX1/ZFP42, encoded by the Yin Yang 1 and Reduced Expression Protein 1/Zinc Finger Protein 42 gene, respectively. In contrast to the well-established roles of the latter two in gene regulation, X chromosome inactivation and binding to specific transposable elements (TEs), much less is known about YY2, and its presence during mouse preimplantation development has not been described. As it has been reported that mouse embryonic stem cells (mESC) cannot be propagated in the absence of Yy2, the mechanistic understanding of how Yy2 contributes to mESC maintenance remains only very partially characterized. We describe Yy2 expression studies using RT-PCR and staining with a high-affinity polyclonal serum in mouse embryos and mESC. Although YY2 is expressed during preimplantation development, its presence appears dispensable for developmental progress in vitro until formation of the blastocyst. Attenuation of Yy2 levels failed to alter either Zscan4 levels in two-cell embryos or IAP and MERVL levels at later preimplantation stages. In contrast to previous claims that constitutively expressed shRNA against Yy2 in mESC prohibited the propagation of mESC in culture, we obtained colonies generated from mESC with attenuated Yy2 levels. Concomitant with a decreased number of undifferentiated colonies, Yy2-depleted mESC expressed higher levels of Zscan4 but no differences in the expression of TEs or other pluripotency markers including Sox2, Oct4, Nanog and Esrrb were observed. These results confirm the contribution of Yy2 to the maintenance of mouse embryonic stem cells and show the preimplantation expression of YY2. These functions are discussed in relation to mammalian-specific functions of YY1 and REX1.


Introduction
Yin Yang 2 (Yy2) is one of a group of related genes encoding DNA-binding transcription factors that also includes Reduced Expression Protein 1/Zinc Finger Protein 42 (Rex1/Zfp42) [1] and the prototype Yin Yang 1 (Yy1) [2]. YY1 is a well-studied, ubiquitously expressed Repeat DNA and transposable elements (TEs) including endogenous retroviruses (ERVs) occupy >50% of mammalian genomes [28,29]. While they contribute to biological processes, their expression is generally considered deleterious to normal physiology and genome integrity [30,31]. Specific TEs including LINE, Intracisternal A-particle (IAP), and MERVL are expressed in a stage-specific pattern in both human and mouse preimplantation embryos [31,32]. Altering normal levels of either LINE-1 [33] or MERVL-containing transcripts [34] during preimplantation stages interferes with normal developmental progress, probably as a result of altered chromatin accessibility [35]. The expression of MERVL in mESC is associated with expanded potency, attributed to so-called two-cell-like cells [32,36]. While sequences related to the YY1-binding motif element are enriched at LINE, IAP, and MERVL elements, and YY2 binding has been reported [24,28], the extent to which YY1 or its paralog YY2 contribute to this regulation during preimplantation development and in mESC has not been extensively studied.
Despite the presence of both YY1 and REX1 in preimplantation embryos [1,13], neither the presence of YY2 at preimplantation stages nor a functional role for Yy2 at stages before blastocyst outgrowth/implantation has been previously addressed. This functional role may reside in the regulation of either Zscan4 [18], as described for REX1 [13], or transposable elements. YY2 binding to the latter has been reported in mESC [24] and TS cells [17], especially to MERVL and IAP. However, as opposed to the regulation of these elements by REX1 in early embryonic stages [17], similar regulation by YY2 has not been previously addressed. It was previously shown that propagation of mESC requires Yy2 [25]. However, in the absence of viable Yy2-deprived cells, gene expression analysis in mESCs with attenuated Yy2 levels was very limited and did not define the role of Yy2 in ES self-renewal and pluripotency. Therefore, while YY2 binding sites in mESC were identified, results on the potential deregulation of pluripotency genes were mostly derived from YY2 overexpression studies. Moreover, deregulation of transposable elements in mESC as a function of Yy2 levels was not addressed at all [25].
In the present manuscript, we first set out to show the presence of YY2 in preimplantation embryos from zygote onwards. Once its presence was established, we addressed both whether attenuation of Yy2 levels might alter preimplantation development in vitro and the possibility that a lack of YY2 may deregulate the expression of co-regulated Zscan4 and MERVL, as well as IAP elements. Furthermore, in an independent effort to confirm and analyze Yy2 function in mESC, we studied mESC with attenuated Yy2 levels. For the first time, we addressed both colony formation capacities and pluripotency-associated gene expression in these cells. Finally, we were then able to test whether YY2 binding to MERVL and IAP elements in ES cells is translated into deregulation of such elements in the absence of YY2.

Oocyte and Embryo Collection and In Vitro Culture
Oocyte and embryo collection were performed according to standard procedures [37]. Four-to six-week-old females were superovulated by an intraperitoneal (i.p.) injection of pregnant mare serum gonadotropin (PMSG, 5 IU) followed by human chorionic gonadotropin (hCG; 7.5 IU) 48

Immunofluorescence and Confocal Microscopy
Mouse embryos (CD-1, Charles River, Lyon, France) for immunostaining were obtained from natural matings using standard methods. Immunostaining and confocal microscopy were routinely performed as previously described [13,38,39], with minor modifications. Briefly, mouse embryos from natural matings at the indicated stage were washed in PBS for 5 min, fixed at room temperature (RT) for 20 min in 2.5% paraformaldehyde in PBS (pH 7.4), and permeabilized with 0.2% Triton-X100 in PBS for 20 min at RT. Embryos were then blocked/permeabilized in PBS/0.2% Triton-X100/10% FBS for 1 h at RT. All the following incubations and washes were performed in PBS/0.2%Tx100/2% FBS (PTF). Fixed and permeabilized embryos were then incubated overnight (o.n.) at 4 • C with primary antibodies. Anti-YY2 antibody was used at (1:4800) dilution. Embryos were washed three times for 15 min and incubated with anti-rabbit biotin (1:150) secondary antibody for 1h at RT. Embryos were washed three times for 15 min and incubated with streptavidin-488 (1:400) for 1h at RT, followed by three washing steps in the dark. Nuclei were counterstained with DAPI (10 µg/mL) in PBS for 20 min. Embryos were mounted in glycerol/DABCO 2.5%/Tris pH8.6 (DTG). Negative controls were performed using the same procedure without the addition of primary antibody. Confocal sections were obtained using an Olympus confocal microscope Fluoview FV1000 with 40X objective (Olympus Europe, Hamburg, Germany). Images were pseudocolored as follows: YY2 in green and DAPI in blue (cell nuclei).

Immunological Reagents and Western Blot
A rabbit antiserum raised in house against a GST-YY2∆ protein (amino acids 1 to 173) and IgG-purified according to standard methods has been described in detail elsewhere [24]. To prove specificity for YY2, the serum has previously been tested in Western blot and immunoprecipitation assays, comparing HA-tagged YY2 with HA-tagged YY1 in 293T cells ( Figure S3 in [24]). The experiments demonstrated that this antibody is specific for YY2, showing also absence of binding to YY1, absence of IF staining in non-transfected 293T cells, and absence of cross-reaction with YY1 in transfected 293T cells. Monoclonal α-HA (clone HA-7) was obtained as an unpurified ascites fraction (Sigma H9658). Preparation of cell lysates, separation by SDS-PAGE, and Western blot were carried out as described [17]; equivalent loading in each lane is demonstrated by reprobing the membrane with an αTUBULIN antibody.

Embryo Microinjection with siRNAs
Zygotes were collected from superovulated B6D2F1/J females mated to B6D2F1/J males (Charles River, Lyon, France) at 20 h post-hCG (defined as Day 1), maintained in M2 media, and microinjected into the cytoplasm with siRNA duplexes (100 µM). Genespecific Stealth TM siRNA duplexes were purchased from Invitrogen and prepared as described [13]. Sequences are listed in Table S2; Stealth™ RNAi Negative Control Med GC (Life Technologies Corporation, Ref. 12935300, Carlsbad, CA, USA) was used as a control.
Texas Red detection was used to monitor siRNA microinjections. Degenerated embryos were removed both immediately after injection and 18 h after. The number of embryos and their developmental stage (e.g., number of cells) were evaluated daily. Images were captured using a Leica DFC360Fx camera adapted to a Leica M165FC stereomicroscope (Leica Microsystems AG, Heerbrugg, Switzerland) using Leica Application Suite 3.2.0 (Leica Microsystems Ltd., Heerbrugg, Switzerland). For subsequent gene expression analysis, embryos were harvested at the indicated morphology in TRIzol ® (Life Technologies Corporation, Carlsbad, CA, USA), flash-frozen, and stored at −80 • C.

Gene Expression in Preimplantation Embryos
Embryos were harvested at 1-cell, 2-cell, 4-cell, 8-cell, morula, blastocyst, and late blastocyst stage, at 20, 43, 55, 66, 80, 92 h post-hCG, and at embryonic day 4.5, respectively. Oocytes were collected at 14 h post-hCG from unmated females. An identical number of experimental and control oocytes or embryos from the same experiment (generally 8 to 10 embryos) were pooled and homogenized in 100 µL TRIzol ® . Total RNA from embryos was purified by a two-round extraction protocol. RNA was firstly isolated from TRIzol ® by chloroform extraction, ethanol precipitation using glycogen (Roche, Basel, Switzerland) as a carrier, pelleted by centrifugation, and re-suspended in DEPC-treated water. RNA was further DNAse-treated with RQ1 RNase-free DNase (Promega), phenol extracted, re-precipitated with ethanol, and re-suspended in DEPC-treated water. Reverse transcription of purified RNA was performed using the ThermoScript ® RT-PCR System (Life Technologies Corporation, Carlsbad, CA, USA) with random hexamer primers. cDNA from one to two oocyte or embryo equivalents were used as a template for each PCR reaction, and amplified using the following PCR program: 35 cycles of denaturation at 94 • C for 30 s, 30 s of annealing and 1 min extension at 72 • C; and a final 5 min extension. PCR products were visualized using ethidium bromide and photographed on a Gel Doc transilluminator (BIO-RAD, BIO-RAD Laboratories, Segrate, Milan, Italy). Primers are listed in Table S2.

Cell Culture, shRNA Knockdown, and RT-qPCR Analysis
Procedures for tissue culture of mESC line E14T [40], colony formation assays after shRNA-mediated knockdown, and RT-qPCR analysis were carried out as described previously [17]. mESC were cultured in in the presence of 10% fetal calf serum, leukemia inhibitory factor (LIF), and 2i as described [41], on gelatin-coated tissue culture dishes. Vectors used for the episomal expression of short hairpin RNAs (shRNAs) [17] contained sequences for episomal maintenance, PGKHygro, a H1 promoter [42], and short hairpin sequences, which were designed using specialized software [43]. Two different sequences were inserted: 5 CAATACCACTCTCCTGTTATT as sh1Yy2; 5 GTAGTAGAGATCATGATAA as sh2Yy2; a GFP sequence was used as a control: 5 AAGCGCGATCACATGGTCCTG.
For the subsequent evaluation of colony formation assays (colony numbers, protein extraction, and RNA extraction), colonies grown in the presence of selective agent for 7 days were carefully washed twice with ice-cold PBS (avoiding colony cell loss). For colony numbers and self-renewal analyses, cell colonies were fixed at 4 • C in 80% ethanol. Posterior analysis included staining for alkaline phosphatase (AP) (Sigma-Aldrich, Ref. 85L3R, St. Louis, MO, USA) and counting of colonies (Nikon Diaphot Inverted Tissue Culture Microscope, (Nikon Corporation, Tokio, Japan)). For each condition, ten fields of predetermined size were counted; colony numbers of three independent experiments are represented as mean ± standard deviation. Pearson's chi-squared test was subsequently used to determine whether the difference between the frequencies observed after treatment with each shRNA were statistically significant. Alternatively, cells were directly harvested in TRIzol ® or in protein lysis buffer for further gene expression (RT-qPCR) and protein level (Western blot) analyses.
Attenuation of Yy2 levels with a third hairpin (5 TTCAGTCCTGAATTTGGAAGC named sh3Yy2) was studied after transient transfection. The construct contained a CAGGdriven EGFP and an H1 promoter for expression of hairpin RNA. Two days after transfection, cells were sorted to separate GFP+ from GFP-cells, which were subsequently lysed in TRIzol ® for posterior gene expression analysis.
Gene expression was analyzed by quantitative PCR; all primers are listed in Table  S2. Expression levels were normalized to Gapdh as a reference gene and depicted as fold expression (2 −∆∆Ct ) [44] relative to mESC-expressing shControl RNAs. The origin and specificity of TE primers has been amply discussed [17]. Statistical analyses were performed using GraphPad software (San Diego, California USA, software version 9.0.2. for MacOS).

Expression of YY2 during Preimplantation Development
A publicly available database of expressed-sequence tags (ESTs) (Unigene database) was searched for ESTs encoding Yy2. While ESTs for diverse adult body parts including Cells 2021, 10, 1123 6 of 16 bone marrow, bone, lung, mammary gland, skin, spleen, testis, and thymus were listed in the database, no pancreas, eye, or intestine-derived ESTs have been reported. We also found that among 31 cDNA Yy2 clones described, no EST was detected in oocytes or embryos at the zygotic stage, but several originated from libraries of embryos at cleavage, morula, and blastocyst stages ( Figure S1). This in silico analysis indicates that Yy2 is expressed during embryonic preimplantation development.
To experimentally confirm the expression patterns of Yy2 during early development, we performed RT-PCR analysis of pooled oocytes and embryos at preimplantation stages. Yy2 mRNA was detected in oocytes and early embryos at all developmental stages ( Figure 1A). Surprisingly, in contrast with the in silico prediction, Yy2 was expressed in oocytes and zygotes, although its levels of expression were lower than in blastocysts ( Figure 1A, Figure S1B). From the two-cell stage onwards, Yy2 mRNA levels gradually increased, reaching peak levels in the late blastocyst.
To study the temporal and spatial expression pattern of the YY2 protein, we performed indirect immunofluorescence analyses using an affinity-purified polyclonal serum, raised against the aminoterminus of YY2 [24]. The specificity of this serum for YY2 has been extensively confirmed previously; see Materials and Methods for a full description of relevant data. In accordance with the presence of Yy2 transcripts ( Figure 1A), immunostaining revealed YY2 expression at all preimplantation stages tested, showing a more prominent staining in two-cell embryos, which may correspond to either increased protein levels or nuclear clustering ( Figure 1B). Starting at the one-cell stage until the blastocyst stage, anti-YY2 immunoreactivity was detected in all cells, although dynamic subcellular localization patterns were evident. YY2 staining was detected in the nucleus at all embryonic stages analyzed, although some potential cytoplasmic localization might be observed in embryonic cells and metaphase II (MII) oocytes above the antibody unspecific background ( Figure 1B,C). In embryonic day 3.5 (E3.5) blastocysts ( Figure 1B), YY2 showed a mixed perinuclear-cytoplasmic localization pattern in the ICM compartment, while nuclear staining was observed in cells within the trophectoderm. YY2 staining became progressively more nuclear in trophectoderm cells at the late blastocyst stage. YY2 showed dynamic levels of expression during preimplantation, with apparent fluctuations in expression level and intracellular localization.

In Vitro Development of Yy2 Depleted Embryos
To attenuate Yy2 expression during preimplantation development, we designed siR-NAs (Stealth TM RNAs; see Materials and Methods). The efficiency of siRNAs was tested by co-transfection with HA-YY2 in 293T cells. Expression of HA-YY2 was abolished by co-transfection of S5 Stealth RNAs and, to a much lesser degree, by S6 Stealth RNAs, but not by control siRNAs ( Figure S2A). Efficiency of these Yy2 S5 Stealth RNAs (siYy2 hereafter) and siRNA against Rex1 (siRex1) in mouse embryos was assessed after injection of one-cell embryos. Yy2 and Rex1 mRNA levels were monitored in injected embryos at the morula stage. Expression of Rex1 was severely suppressed in the siRex1-injected embryos ( Figure 2A). The converse was true in siYy2-injected embryos, which displayed unaltered levels of Rex1 and attenuated levels of Yy2. No changes were detected in the expression of the H2afz gene, used as a control.  RT refers to controls without reverse transcription; a PCR reaction without input (-) is shown to the right. The sizes of molecular weight markers run alongside are indicated on the left. (B) Indirect immunofluorescent detection of YY2 in confocal sections of representative mouse embryos depicted as DIC images. DAPI staining indicates chromatin (blue). The antibody used has been shown to specifically detect YY2 in immunofluorescence ( Figure 1A and Figure S4 in [24]). Scale bar, 10 µm. (C) Absence of unspecific background in indirect immunofluorescence. Confocal sections of preimplantation mouse embryos incubated without primary antibody (control), processed in parallel to the indirect immunofluorescent detection of YY2 shown in Figure 1B. Scale bar, 10 µm. but detectable expression of Yy2 in two-cell embryos seemed slightly attenuated by the siRNAs (Figure 2B), we could not detect differences in the levels of Zscan4 expression. Additionally, under these conditions, neither MERVL expression at the morula stage (Figure 2C) nor the blastocyst-specific expression of IAP elements was altered ( Figure 2D). IAP expression was also not affected in Yy2+Rex1 doubly depleted embryos at the blastocyst stage. Altogether, we found no evidence for altered gene expression as a result of Yy2 depletion in preimplantation embryos. were performed in parallel and embryonic development was monitored daily. The ratio of the number of embryos at the specified stage over the total number of embryos, and the average percentage, is indicated. The day of zygotic microinjection was set as Day 1. Cumulative data of three independent experiments are presented, except for data at Day 5, which corresponds to 1 and 2 experiments for siControl and siYy2, respectively. A technical issue for the control condition on Day 5 precluded normal developmental progression; therefore, Day 5 data of this condition are limited to only one experiment (see Table S1). (B-D) Effect of attenuation of Yy2 levels on gene Injections of control siRNAs (siControl) and siRNAs against Yy2 (siYy2) were performed in parallel and embryonic development was monitored daily. The ratio of the number of embryos at the specified stage over the total number of embryos, and the average percentage, is indicated. The day of zygotic microinjection was set as Day 1. Cumulative data of three independent experiments are presented, except for data at Day 5, which corresponds to 1 and 2 experiments for siControl and siYy2, respectively. A technical issue for the control condition on Day 5 precluded normal developmental progression; therefore, Day 5 data of this condition are limited to only one experiment (see Table S1). (B-D) Effect of attenuation of Yy2 levels on gene expression. One-cell embryos were microinjected with control siRNAs (siCon), Yy2-specific siRNAs (siYy2), or a combination of Rex1 and Yy2 siRNAs (siY + R). Embryos were cultured in vitro to the 2-cell (B), morula (C), or blastocyst (D) stages. RNA was extracted, reverse-transcribed, and gene expression of equivalent amounts of embryos was analyzed by semiquantitative RT-PCR with primers specific for the indicated genes; RNA extracted from mESC was processed alongside. H2afz was used as a control. (-) refers to PCR reactions without template added. The sizes of Mw markers run alongside are indicated on the left.
In order to examine whether transient depletion of Yy2 affected early embryonic development, one-cell embryos were collected and injected with control siRNAs (non-Cells 2021, 10, 1123 9 of 16 sense oligonucleotides) or with siYy2 and cultured in vitro. Similar numbers of embryos were injected with either Yy2 or control siRNAs in each experiment, and successful injection was visualized using co-injected Texas Red ( Figure S2B). Two-cell embryos were selected for further culture and developmental progress was examined daily. After injection of 104 and 114 embryos in total (siControl and siYy2, respectively) on three independent experiments, we were unable to observe differences in developmental progress between embryos injected with control siRNAs or siRNAs directed against Yy2 (Table 1,  Table S1). Control and siYy2-injected embryos developed into two-cell embryos (100 and 99.1%, respectively; day 2), 5-8-cell embryos (88 and 82.2%, respectively; day 3), morulae (90 and 93%, respectively; day 4), and blastocysts (80.8 and 76.3%, respectively; day 5) with equal efficiency (Table 1, Figure S2B). No statistically significant differences were found between the two groups at the stages analyzed. Injections of control siRNAs (siControl) and siRNAs against Yy2 (siYy2) were performed in parallel and embryonic development was monitored daily. The ratio of the number of embryos at the specified stage over the total number of embryos, and the average percentage, is indicated. The day of zygotic microinjection was set as Day 1. Cumulative data of three independent experiments are presented, except for data at Day 5, which corresponds to one and two experiments for siControl and siYy2, respectively. A technical issue for the control condition on Day 5 precluded normal developmental progression; therefore, Day 5 data of this condition are limited to only one experiment (see Table S1).
Given the absence of a phenotype in Yy2-depleted embryos, we addressed the effect of simultaneous depletion of both Yy2 and Rex1. Similar injection experiments were performed, resulting in no significant developmental defects or delays (Supplementary  Table S1). Similar to control embryos, the doubly depleted embryos developed with high efficiency into 5-8-cell embryos (95%), morulae (92%), and blastocyst stages (82%) on consecutive days.

Yy2 Loss-of-Function Mouse Embryos Express Normal Levels of ERV Elements
We assayed the gene expression of the embryo-specific gene Zscan4 and selected transposable elements by RT-PCR at different developmental stages, in embryos with attenuated Yy2 levels, which were generated using siRNAs as described above. While weak but detectable expression of Yy2 in two-cell embryos seemed slightly attenuated by the siRNAs (Figure 2B), we could not detect differences in the levels of Zscan4 expression. Additionally, under these conditions, neither MERVL expression at the morula stage ( Figure 2C) nor the blastocyst-specific expression of IAP elements was altered ( Figure 2D). IAP expression was also not affected in Yy2+Rex1 doubly depleted embryos at the blastocyst stage. Altogether, we found no evidence for altered gene expression as a result of Yy2 depletion in preimplantation embryos.

Depletion of Yy2 Expression Alters Self-Renewal in Mouse ES Cells
To attenuate Yy2 expression in mESC, we designed two hairpin constructs, termed Sh1Yy2 and Sh2Yy2. YY2 protein levels in mESC were severely reduced in cultures over-expressing either hairpin RNA, as opposed to cells expressing the control shRNA ( Figure 3A). We subsequently investigated Yy2 function in mESC by using stable episomal expression of these shRNAs. Cells were electroporated and selected for expression of Hygromicine resistance co-carried on the shRNA plasmids, seeded at clonal densities, and colony formation was analyzed as a measure of self-renewal. Cells expressing shRNAs targeting Yy2 formed fewer colonies compared to cells electroporated with control shRNAs (Figure 3B), accounting for 81% and 37% of the total number of colonies in Sh1Yy2 and Sh2Yy2 conditions, respectively, relative to the total number of colonies in shControl ( Figure 3C). ES colonies were classified into three different categories based on morphology and alkaline phosphatase (AP) staining: undifferentiated-colonies with smooth edges and rounded shape, limited presence of differentiated cells and strong AP positive staining; intermediate colonies-those with irregular edges, considerable number of differentiated cells, and faint AP staining; and differentiated colonies-with no AP positive staining. Colony morphology was affected by Yy2 depletion, as the number of undifferentiated colonies diminished 1.5-and 8.8-fold in colonies expressing sh1Yy2 and sh2Yy2, respectively ( Figure 3D). In contrast, the percentage of colonies with either an intermediate or differentiated morphology increased in both conditions (3-and 7.3-fold for sh1Yy2 and sh2Yy2, respectively ( Figure 3D)). The observed difference in colony morphology was significant between control and both sh1Yy2 (X 2 = 106.64, df = 2, p-value < 0.05) and sh2Yy2 (X 2 = 604.9, df = 2, p-value < 0.001). Altogether, these data show that reduced expression of Yy2 was coincident with reduced self-renewal in mESC.

Gene Expression in Yy2-Depleted ES Cells
To further understand the molecular changes associated with the depletion of YY2 in mESC, we extracted RNA from Yy2-depleted colonies (and controls) and performed RT-qPCR gene expression analyses. We first compared the expression levels of pluripotency factors in the colonies analyzed, to mESCs obtained from regular culture. Results ( Figure S3) indicate that the control colonies (shControl) express levels of core pluripotency factors comparable to mESC. We therefore assume that the cells making up these colonies can be considered mostly undifferentiated as expected, although some differentiation is apparent in the colonies after 7 days of culture ( Figure 3B). We then analyzed whether changes in the levels of Yy2 ( Figure 4A) could affect the expression of other members of the transcription factor family. In Yy2-depleted cells, Rex1 levels appeared slightly but significantly increased (p-value ≤ 0.05), while Yy1 levels were not changed, although a trend of downregulation and some variability was observed between independent experiments ( Figure 4A). We next measured the expression of several pluripotency markers and no significant differences of expression were detected for Oct4/Pou5f1, Sox2, Nanog, and Esrrb mRNAs as a result of attenuation of Yy2 levels ( Figure 4A, Figure S3A). Using primers that amplify regions conserved among several copies of transposable elements, we did not detect significant differences in the expression levels of either MERVL or IAP ( Figure 4A). We finally investigated the expression of Zscan4 in Yy2-depleted cells. We detected a significant induction of Zscan4 levels in Yy2-depleted cells as opposed to control cells (3.6-fold change, p-value ≤ 0.01, Figure 4A).
To validate these results, we transfected mESCs with an shRNA against Yy2 containing an EGFP cassette (sh3Yy2) and compared the levels of gene expression of GFP-expressing (GFP+) versus GFP-non-expressing (GFP-) cells by RT-qPCR. As expected, GFP + sorted cells had lower levels of Yy2 compared to GFP-cells ( Figure 4B). Similarly to the results obtained in the previous approach, Yy2 family members were somewhat affected in Yy2-depleted cells GFP + cells. While no significant changes were observed for Rex1, Yy1 mRNA levels were significantly downregulated. Consistent with the results shown in Figure 4A, no significant changes in gene expression levels were detected in the pluripotency markers analyzed (Oct4/Pou5f1, Sox2, Nanog, and Esrrb), although the level of Sox2 trended downwards. As before, the Yy2-depleted GFP + cells showed a 2.4-fold increase in Zscan4 levels compared to GFP-cells ( Figure 4B). From the combined experiments, we conclude that YY2 contributes to the regulation of Zscan4 levels in mESC.

Discussion
The data on YY2/Yy2 expression in time and space are by no means complet had previously shown that YY2 protein is expressed in mouse ES and TS cells [24 now report the presence of Yy2 mRNA and protein starting in oocytes and in preim tation embryos at all stages. We show that YY2 is present during all stages of preim tation development, and dynamically localized between the nucleus and cytoplas opposed to a steady and gradually increasing expression of Yy2 throughout preimp tion development, we detected apparently higher levels of YY2 specifically at the tw stage, showing clear nuclear expression above the cytoplasmic staining. This pattern be indicative of a potential clustering of YY2 protein (and associated functional rele at these particular stages) or directly result from higher expression levels. The latte is compatible with a non-linear relationship between mRNA and protein levels desc in mESC [25]. While knockdown of Yy2 mRNA was clearly achieved, development blastocyst under these conditions proceeded in the absence of temporal or morphol alterations. While depletion of Yy2 has been shown to interfere with blastocyst outgr [25], we find no evidence for an earlier role in vivo according to morphological and v ity criteria. However, more subtle defects cannot be ruled out and defects may have further masked by the continued presence of protein from maternal deposits (see the ence of Yy2 mRNA and YY2 protein in the oocyte, Figure 1). Furthermore, implan and posterior development of either the embryos or extra-embryonic tissues is lik Gene expression in mESC colonies transfected with either sh1Yy2-or shCon-expressing plasmids was measured by qPCR. Expression levels are depicted as fold expression in depleted cells relative to mESC-expressing shControl. Data presented for sh1Yy2 are the mean of 3 biological replicates, each performed on different days and using a different preparation of cells (except for Yy1 and Esrrb data that correspond to two replicates and one experiment, respectively), and three technical replicates for each experiment. The statistical significance (one sample t-test) of differences in gene expression is indicated, ** p-value ≤ 0.01, * p-value ≤ 0.05; all other differences were not significant. (B) Expression levels of Yy2 gene family members, pluripotency markers, and Zscan4, measured in cells transfected with a GFP-expressing sh3Yy2 construct. Cells were sorted based on GFP expression, and expression is depicted as fold expression of GFP+ relative to GFP-cells. Data shown are based on technical replicates. ** p-value ≤ 0.01, * p-value ≤ 0.05 (one sample t-test).

Discussion
The data on YY2/Yy2 expression in time and space are by no means complete. We had previously shown that YY2 protein is expressed in mouse ES and TS cells [24]. We now report the presence of Yy2 mRNA and protein starting in oocytes and in preimplantation embryos at all stages. We show that YY2 is present during all stages of preimplantation development, and dynamically localized between the nucleus and cytoplasm. As opposed to a steady and gradually increasing expression of Yy2 throughout preimplantation development, we detected apparently higher levels of YY2 specifically at the two-cell stage, showing clear nuclear expression above the cytoplasmic staining. This pattern may be indicative of a potential clustering of YY2 protein (and associated functional relevance at these particular stages) or directly result from higher expression levels. The latter case is compatible with a non-linear relationship between mRNA and protein levels described in mESC [25]. While knockdown of Yy2 mRNA was clearly achieved, development until blastocyst under these conditions proceeded in the absence of temporal or morphological alterations. While depletion of Yy2 has been shown to interfere with blastocyst outgrowth [25], we find no evidence for an earlier role in vivo according to morphological and viability criteria. However, more subtle defects cannot be ruled out and defects may have been further masked by the continued presence of protein from maternal deposits (see the presence of Yy2 mRNA and YY2 protein in the oocyte, Figure 1). Furthermore, implantation and posterior development of either the embryos or extra-embryonic tissues is likely to depend on YY2, as Yy2 depletion interferes with blastocyst outgrowth in vitro [25]. The continued presence of Yy2 mRNA and protein during preimplantation embryos at all stages overlaps with the YY1 transcription factor [45] and ZFP42/REX1 [13]. As Yy1 and Rex1 contribute to mammalian-specific processes such as regulation of transposable elements and XCI, contributions of Yy2 to these processes may still be envisaged.
YY2 shares homology in the DNA-binding zinc fingers with both YY1 and ZFP42/REX1. Despite a high degree of homology in the DNA-binding domain between YY1 and YY2, target genes for both genes are mostly unique to either of them [24,25], although some overlap exists [23,25]. Certainly, overlap exists to some extent in YY1 and YY2 binding to specific TE [17,46] and YY2 associates with MERVL and IAP elements in mESC [24]. In contrast to previously reported regulation of IAP and MERVL elements by YY1 [46] and REX1 [17], respectively, we were unable to detect altered expression of these elements in mouse embryos or mESC with attenuated levels of Yy2. However, similar to YY1, YY2 may interact with the TRIM28-mediated silencing machinery through LSD1 [18,32]. Therefore, Yy2 may regulate the expression of other TEs, at stages not investigated, or be supplanted by YY1 or REX1 in its absence.
It was previously shown that propagation of mESC requires Yy2 [25]. However, in the absence of viable Yy2-deprived cells, data presented in this report did not define the role of Yy2 in ES self-renewal and pluripotency. Careful titration of selective agent and the use of episomal vectors allowed us to obtain ES colonies with constitutively diminished Yy2 levels. Reduced expression of Yy2 in mESC compromised their colony-forming capacities, resulting in lower numbers of colonies and more differentiated cells (Figure 3). Surprisingly, we failed to detect alterations in the expression levels of a set of pluripotency-associated genes including Sox2, Oct4, Esrrb, and Nanog ( Figure 4A, Figure S3). This result was repeated when analyzing cells sorted for sh3Yy2 expression, with the exception of a weak downregulation of Sox2 levels ( Figure 4B). Tahmasebi et al. have argued that the essential function of YY2 in mESC is most likely mediated by direct transcriptional regulation of ESC genes that they identified in genome-wide YY2 binding studies, which include Oct4, Esrrb, Eras, Tet1, Tet2, and Tdgf1 [25]. This proposal was further based on results of YY2 overexpression studies and attenuation studies in which transient high overexpression of Yy1 was also observed. We show that Yy2 levels can be attenuated in mESC, while still affecting colony formation. Yy2 depletion did not affect Oct4 and Esrrb levels in our experiments. This may be compatible with the nature of the colonies obtained with sh1Yy2, 80% of which were either undifferentiated or consisted of a core of undifferentiated cells. A second explanation is the enhanced expression of these factors upon attenuation of Yy2 levels (as reported previously [25]) in a relatively diminished number of pluripotent cells, resulting in the observed lack of overall change in gene expression levels ( Figure 4). It also remains possible that the effect of Yy2 depletion on the levels of core pluripotency factors is indeed temporal, as suggested by the absence of effects after 16 h ( Figure 4B), or after 7 days ( Figure 4A), in contrast to increased levels observed after 48 h [25], in different studies. Independent of the explanation, our results do not provide further proof for the role of core pluripotency factors in Yy2-mediated mESC self-renewal. Future studies aimed at understanding the contribution of Yy2 to the expression of core pluripotency factors will have to consider off-targeting effects of shYy2, establish appropriate time-courses, and test for genes beyond the restricted set analyzed in our studies.
Although Zscan4 was not identified as a YY2-binding target in mESC [25], we observed that the attenuation of Yy2 levels was concomitant with higher Zscan4 expression. Surprisingly, however, Zscan4 levels in preimplantation embryos seemed indifferent to Yy2 depletion. However, given the weak effect of Yy2 siRNAs in two-cell embryos ( Figure 2B), an effect of YY2 on Zscan4 levels cannot be excluded at this point. Temporal differences in regulation were also observed for REX1-regulated expression of ERV [17]. This observation or phenomenon could be explained by the presence of cis-regulatory regions specific for either stage, the epigenetic modifications that allow YY2 binding in one cell type but not the other, or the requirement for stage-specific cofactors for binding or regulation.
In addition to so-called core pluripotency factors [12], several genes which are essential for long-term maintenance of mESC in culture are expressed only in a subset of cells [47]. Cells shuttle between high-and low-expressing states and the resulting heterogeneity reflects the presence of subpopulations of cells. The non-uniform expression of Zfp42/Rex1 [48] and Stella (Dppa3/Pgc7) [49] is related to a primed or naive state of pluripotency [50,51]. A small subset of cells expressing Zscan4 [36] display a two-cell stage embryo signature marked by the activation of MERVL endogenous retrovirus [32] and the expression of Zscan4 co-regulated genes including Prame, Thoc, and Tcstv [52,53]. Both Yy2 (Figure 4) and Rex1 [13] deficiency alter Zscan4 levels in mESC and two-cell embryos, respectively, raising the possibility of co-regulation of this locus by these two factors. We do not know whether they converge at the level of DNA binding or through LSD1 interaction, which is shared by both factors. Attempts to quantitate Zscan4-positive cells in Yy2or double Yy2 + Rex1-depleted mouse ES colonies may help to clarify this issue. We entertain the possibility that Yy2 exerts an essential function in the regulation of the expression of Zscan4, whose expression in turn is essential for long-term culture of mESC [36].