Paternally Expressed Peg3 Controls Maternally Expressed Zim1 as a Trans Factor

The expression of two adjacent imprinted genes, Peg3 and Zim1, is inversely correlated: down-regulation of Peg3 coinciding with up-regulation of Zim1. The current study characterized this inverse correlation using a mutant allele targeting Peg3. According to the results, the mutation on the paternal allele of Peg3 resulted in a dramatic increase in the transcription levels of the maternal allele of Zim1, suggesting the involvement of unknown trans factors in this trans-allelic event. Subsequent ChIP experiments revealed that the protein encoded by Peg3 itself binds to the zinc finger exon of Zim1, which is modified with the repression mark H3K9me3. Interestingly, the levels of H3K9me3 on Zim1 are also reduced in the mutant cells lacking the protein PEG3, suggesting potential roles for PEG3 in establishing H3K9me3 on Zim1. Reintroducing PEG3 into the mutant cell restored down-regulation of Zim1, confirming the predicted repressor role for Peg3 on Zim1. Overall, these results demonstrated that paternally expressed Peg3 controls maternally expressed Zim1 as a trans factor. The current study also provides the first case for the trans-allelic interaction of two oppositely imprinted genes through their gene products.


Introduction
In mammalian imprinted domains, two genes with opposite imprinting are quite often localized right next to each other, and such examples include maternally and paternally expressed H19/ Igf2 and Gtl2/Dlk1. This genomic layout is related to the fact that two genes are usually co-regulated through shared cis elements, such as Imprinting Control Regions (ICRs) [1]. As such, one gene is very closely associated with the other gene in terms of their transcription levels and allele-specific expression patterns. This has been well demonstrated through a series of mouse mutagenesis experiments. For instance, mutating an endoderm-specific enhancer located in the 39-side of H19 caused down-regulation for both H19 and Igf2 [2], yet repositioning this enhancer from the downstream region of H19 to an intergenic region between the two genes resulted in the reactivation of the silenced maternal allele of Igf2, causing biallelic expression of Igf2 [3]. This has been a basis for identifying an enhancer-blocking function of the ICR that is located immediate upstream of H19 [4]. It is most likely that similar regulatory mechanisms operate in the other imprinted domains with this genomic layout, the juxtaposition of two adjacent genes with opposite imprinting.
A similar genomic layout is also found in the Peg3 domain, which contains 7 imprinted genes: paternally expressed Peg3, Usp29, APeg3, Zfp264 and maternally expressed Zim1, Zim2, Zim3 [5]. Among these genes, paternally expressed Peg3 and maternally expressed Zim1 are localized right next to each other, suggesting potential co-regulation of these two genes through shared cis elements. As expected, this domain is regulated through an ICR, termed the Peg3-DMR (Differentially Methylated Region), the 4-kb genomic region surrounding the 1 st exons of Peg3 and Usp29 [6]. Deleting part of this ICR, the 2.5-kb genomic region harboring multiple YY1 binding sites, caused global changes in the expression levels and imprinting status of several genes within this domain [7]. In particular, the expression levels of Peg3 and Zim1 were affected in a dosage-dependent manner: 4-fold down-regulation of Peg3 coinciding with 4-fold upregulation of Zim1. Interestingly, the observed up-regulation of Zim1 was still derived from the maternal allele although the mutation causing down-regulation of Peg3 was on the paternal allele [7]. This trans-allelic outcome by a mutation has not been observed before, and thus suggests the presence of different regulatory mechanisms involving possible trans factors rather than the known mechanisms involving shared cis-regulatory elements as seen from the H19/Igf2 pair.
According to recent studies, Peg3 encodes a DNA-binding protein with transcriptional repression function [8]. Given the observed tight correlation between Peg3 and Zim1, it is possible that Peg3 may control directly the transcription of Zim1 as a trans factor. In this case, the absence or reduced protein levels of PEG3 might be responsible for the observed up-regulation of Zim1. To further test this possibility, the Peg3/Zim1 pair was analyzed using a new mutant model targeting Peg3. In this new model, the mutation truncates the transcription of Peg3, thus removing the PEG3 protein [9]. The results revealed that the removal of PEG3 through the mutation on the paternal allele caused up-regulation of Zim1 without disrupting its maternal-specific expression. PEG3 also binds to the Zim1 locus as a trans factor, yet this binding by PEG3 is closely associated with the histone modification mark H3K9me3, suggesting a potential repression mechanism for PEG3. More detailed results have been described in the following sections.

Results
Removal of the PEG3 protein results in the up-regulation of Zim1 According to the results from the previous study, deletion of part of the Peg3-DMR derived a concurrent 4-fold down and upregulation of Peg3 and Zim1, respectively [7]. However, it is currently unknown the causal relationship between the observed down and up-regulation of the two genes since the mutation also caused other changes within the Peg3 domain. To further clarify the observed effects on the Peg3/Zim1 pair, a new mutant model targeting Peg3 was used for the current study (Fig. 1A). This model was originally constructed with a combinatory scheme of knock-in/knock-out, thus will be referred to as a KO model hereafter for the simplicity. In this model, the mutant allele carries an expression cassette containing two ORFs (Open Reading Frames) within its 5 th intron of Peg3: the promoterless bgalactosidase gene and the neomycin resistance gene driven by the human b-actin promoter [9]. Because of the two poly(A) signals within the cassette, the mutant allele truncates the transcription of Peg3, thus removing the PEG3 protein [9]. Subsequent global gene expression analyses revealed that a large number of genes were affected by the removal of PEG3 protein in both the embryos and placentas of 14.5 d.p.c. (days post coitum) [9]. As expected, Zim1 was also found to be up-regulated in this survey: 2-fold up-regulation in both tissues, which is consistent with the results from the mutant allele deleting part of the Peg3-DMR [7].
To further follow-up this initial result, we have derived a set of mouse embryonic fibroblast (MEF) cells from the mutant animals. Two litters of 14.5-dpc embryos were prepared through timed mating of the female and male heterozygotes for the mutant allele with male and female wild-type littermates, respectively. The first litter inheriting the mutant allele maternally does not have any mutational effects on Peg3 since Peg3 is expressed mainly from the paternal allele. In contrast, the second litter inheriting the mutant allele paternally has an effect on Peg3, removing the PEG3 protein completely [9]. After genotyping and gender determination, each MEF line from a given embryo was individually cultured, and subsequently used for isolating total RNA for RT-PCR analyses. According to the results from the female set (Fig. 1B), the expression levels of Zim1 were 2.5-fold greater in the paternal heterozyote cell (+/2p) than the wild type cell (+/+). Since this mutant allele disrupts the transcription and translation of Peg3 only, the observed up-regulation of Zim1 is most likely an outcome of the removal of PEG3. We repeated this series of expression analyses using another set of MEFs, which also showed a consistent up-regulation of Zim1 ( Figure S1). The allele-specific expression of the observed up-regulation of Zim1 was also tested using the F1 hybrid animals obtained from the inter-specific crossing of C57BL/6J and PWD/PhJ (Fig. 1C). The results from the total RNA of the neonatal brain indicated that the up-regulated expression of Zim1 was still derived from the maternal allele. Overall, this series of expression analyses derived a consistent outcome as seen from the other mutant allele, the upregulation of Zim1 on the maternal allele coinciding with the mutation on the paternal allele of Peg3. Since both mutant alleles share one common feature, the reduced protein levels of PEG3, the observed up-regulation of Zim1 is most likely caused by the changed protein levels of PEG3 in both mutant models.

PEG3 binds to the zinc finger exon of Zim1
Since Peg3 encodes a DNA-binding protein with repression activity, we tested the possibility that the protein PEG3 may control directly the transcription of Zim1 as a trans factor. As an initial step, the binding of PEG3 to the Zim1 locus was investigated using Chromatin ImmunoPrecipitation (ChIP) technique (Fig. 2). For this series of ChIP experiments, 4 genomic regions were selected to scan the entire locus of mouse Zim1: Region 1 (promoter), 2 (intron), 3 (zinc finger exon) and 4 (39-UTR) ( Fig. 2A). The primer set amplifying the promoter of Pgm2l1 (phosphoglucomutase 2-like 1) was also included as a positive control since this region has been shown to be a target locus of the PEG3 protein [8]. Three different sets of cross-linked chromatins were immunoprecipitated with anti-PEG3 polyclonal antibodies. First, the immunoprecipitated DNA from the two MEF cells representing wild-type (WT) and mutant (KO) cells were individually analyzed with PCR ( Fig. 2B). As expected, the enrichment at the Pgm2l1 locus was detected only in the WT cells but not in the KO cells lacking PEG3, confirming the binding of PEG3 to this locus and also the specificity of our ChIP experimental system. The same set of immunoprecipitated DNA was further tested using the 4 primer sets of Zim1. Although the two sets, Region 2 and 4, detected some levels of the enrichment, the detections were not specific to the WT cells, indicating nonspecific binding of the anti-PEG3 antibody to other unknown proteins. On the other hand, the promoter region of Zim1 (Region 1) did not show any level of the enrichment, indicating no binding of PEG3 to this region. In contrast, the zinc finger exon (Region 3) showed much higher levels of the enrichment in the WT cells than in the KO cells, indicating that the observed enrichment likely represents the genuine binding of PEG3 to this region. This has been further confirmed through a set of independent ChIP experiments with a custom-made antibody against PEG3 ( Figure  S2).
Another set of ChIP experiments was also performed using the second set of chromatin prepared from the brains of the WT and KO (+/2p) neonates (Fig. 3B). According to the results, the enrichment at the promoter of Pgm2l1 was also detected higher levels in the WT than in the KO sample, confirming again the specificity of the anti-PEG3 antibody and the ChIP system. A similar pattern of WT-specific enrichment was observed in both Region 1 and 3, indicating the potential binding of PEG3 to these regions. However, the enrichment at Region 3 is much more obvious than Region 1. Also, the observed enrichment at Region 3 is consistent with the results from the MEF cells, thus confirming the genuine binding of PEG3 to this region in the neonatal brain. Finally, the third set of ChIP experiments were performed using the chromatin prepared from the F1 hybrid between C57BL/6J and PWD/PhJ (Fig. 3C). Restriction enzyme digestion scheme was employed to differentiate the two alleles of the immunoprecipitated DNA. According to the results, the immunoprecipitated DNA at Region 1 and 3 both were derived equally from the two parental alleles, indicating that PEG3 likely binds to both alleles of these two regions. Taken together, this series of ChIP experiments concluded that PEG3 binds to the zinc finger exon of Zim1 on both alleles, further supporting the possibility that PEG3 may control directly the transcription of Zim1 as a trans factor.

Reduced levels of H3K9me3 in the mutant cells lacking PEG3
The mouse locus of Zim1 was carefully examined using the genome browser (genome.ucsc.edu) to obtain hints regarding the potential functional outcomes of the observed PEG3 binding. Histone modification levels of H3K4me3 and H3K9me3 derived from ES cells and whole brain tissues are presented as Fig. 4A. This examination revealed that the zinc finger exon of Zim1 is marked with the histone modification H3K9me3 (Fig. 4A). Although this histone mark is relatively rare in gene-rich euchromatic regions, it is well known that this modification is quite often associated with the genomic regions encoding the Kruppel-type zinc finger gene family [10]. Zim1 is a member of this gene family [11], thus the detection of this histone mark at the Zim1 locus is consistent with the pattern observed from previous studies. This modification at Zim1 is most obvious in ES cells but some levels of this modification are also detected in other cell types, including adult whole brain. To further test a potential connection between PEG3 binding and H3K9me3, the modification levels of this repression signal were compared between the WT and KO cells. According to the results (Fig. 4B), the modification levels of H3K9me3 at Region 3 were 2 fold lower in the KO cells than those in the WT cells although the modification levels at other genomic regions, such as the ICR of H19/Igf2, were similar between the two MEF cells (Fig. 4BC). This indicates that the observed reduction of H3K9me3 in the KO cells is specific to the zinc finger exon of Zim1. Since H3K9me3 is known to be a repression mark for transcription, the reduced levels of H3K9me3 is also consistent with the observed up-regulation of Zim1 in the KO cells lacking PEG3 (Fig. 1B). Overall, the genomic region of Zim1 bound by PEG3 is closely associated with H3K9me3, further suggesting that PEG3 might repress Zim1 through H3K9me3.

Restoring the protein levels of PEG3 down-regulates Zim1
To further test PEG3's repressor role in the transcription of Zim1, we performed the following set of in vitro transfection experiments (Fig. 5).  Green Florescent Protein vector (GFP) (Fig. 5). As shown in Fig. 5B, the transient expression of FLP indeed removed the inserted cassette based on the detection of a genomic fragment without the inserted cassette. This removal of the cassette subsequently restored the expression of Peg3 based on RT-PCR. In these cells with the restored PEG3, the transcriptional levels of Zim1 was 2.5 and 1.5-fold reduced as compared to those observed from the two control cells (Fig. 5C). It is also prudent to mention that the transcriptional levels of Zim1 were further reduced in the set transfected with FLP than in the set with GFP. These results demonstrated that the restored expression of Peg3 is responsible for the down-regulation of Zim1, confirming again the inverse correlation between Peg3 and Zim1. Down-regulation of Zim1 was further tested with the over-expression of Peg3 ( Figure S3A). The results confirmed the down-regulation of Zim1 ( Figure S3B). Overall, this series of transfection experiments confirmed that PEG3 functions as a repressor for the transcription of Zim1.

Discussion
In the current study, we tested the possibility that paternally expressed Peg3 may control the transcription of maternally expressed Zim1 as a trans factor. According to the results, the reduced protein levels of PEG3 is indeed responsible for the upregulation of Zim1. The PEG3 protein also binds to the zinc finger exon of Zim1 that is marked with the repression mark H3K9me3. Furthermore, the KO cells lacking PEG3 have the reduced levels of H3K9me3 at the zinc finger exon of Zim1, suggesting that PEG3 might control Zim1 through H3K9me3. In vitro transfection experiments further demonstrated that reintroducing the PEG3 protein into the KO cells restores the down-regulation of  Zim1. Overall, these results confirm that paternally expressed Peg3 controls maternally expressed Zim1 as a trans factor.
The imprinted gene pair of Peg3/Zim1 is unique based on the following reasons. As seen in the other pairs of oppositely imprinted genes, the transcriptional level of Peg3 is also inversely correlated with those of Zim1 ( Fig. 1B and Fig. 5). However, this inverse correlation does not involve any change in their allelespecific expression pattern (Fig. 1C), which is quite different from the other pairs of oppositely imprinted genes [1]. This suggests the involvement of some unknown trans factors in the inverse correlation between Peg3 and Zim1. The results form the current study further indicated that the protein encoded by Peg3 itself is likely this unknown trans factor ( Fig. 2 and Fig. 3). According to previous studies, Peg3 encodes a DNA-binding protein with repression activity [8]. Also, there is a very tight inverse correlation between Peg3 and Zim1: 4-fold down regulation of Peg3 coinciding with 4-fold up-regulation of Zim1 [7]. This line of evidence supports the idea that Peg3 controls the transcription of Zim1 as a trans factor. Nevertheless, the results from the current study also confirm another important aspect of Zim1 that PEG3 is unlikely involved in regulating the allele-specific expression of Zim1 since the removal of PEG3 does not have any effect on the maternal expression of Zim1. The observed PEG3-mediated regulation of the transcriptional levels of Zim1 appears to be separate from some unknown mechanisms controlling the maternal-specific expression of Zim1. In summary, the transcription of Zim1 is regulated through two separate mechanisms: one controlling the allele-specific expression and the other controlling the expression level through PEG3 (Fig. 6).
The binding of PEG3 to the zinc finger exon of Zim1 is consistent with several known facts about the evolutionary origin of Peg3 as well as the repression mark H3K9me3. First, Peg3 is localized in the middle of a Cys2His2-type zinc finger gene cluster [12], yet Peg3 itself encodes a DNA-binding protein with C2H2 zinc finger motifs [8]. This suggests that Peg3 may have originated from this type of zinc finger genes (ZNFs). The C2H2-type ZNFs are known to interact the H3K9 methylase SETDB1 via the corepressor protein KAP-1 [13]. Thus, the binding of PEG3 to the genomic region with the H3K9me3 modification makes sense given the evolutionary origin of Peg3, and further implies that PEG3 might still recruit SETDB1, possibly through the interaction with KAP-1 (Fig. 6). Second, it is well known that ZNFs, such as Zim1, are usually regulated through H3K9me3 [10]. This is further supported by the fact that the zinc finger-coding region of Zim1 is indeed modified by this repression mark in ES and other somatic cells (Fig. 4A). Yet, the KO cells lacking PEG3 have the reduced levels of H3K9me3 (Fig. 4B). This further suggests that PEG3 likely controls the transcription of Zim1 through H3K9me3. According to the results from the mutant mouse model targeting Peg3, many placenta-specific gene families are also up-regulated in the brains of KO mice, yet all of these gene families are known to be regulated through similar repression mechanisms involving H3K9me3 [9]. Therefore, it is most likely that PEG3's regulation on Zim1 may be also through H3K9me3.

Ethics Statement
All the experiments related to mice were performed in accordance with National Institutes of Health guidelines for care and use of animals, and also approved by the Louisiana State University Institutional Animal Care and Use Committee (IACUC), protocol #13-061.

Derivation of MEF (Mouse Embryonic Fibroblast) cells
Two litters of 14.5-dpc embryos were harvested through timed mating of the male and female mutant mice heterozygous for the KO allele with the female and male wild-type littermates, respectively. The mutant allele of Peg3 used for the current study has been previously reported and characterized in detail [9]. The head portion and the red tissues were removed from the embryos, and the remaining portions were minced with razor blades. These minced tissues were transferred to a 15 mL conical tubes containing 1 mL trypsin (Invitrogen, Cat. No. 25300062). After 5 min incubation at 37uC, the cells were harvested with centrifugation, and later resuspended in 15 mL media (Life technologies, Cat. No. 10566024). Finally, the resuspended cells were plated onto a T-75 flask. MEF from each embryo was first genotyped using the following primer set: Peg3-for (59-ATGAGTCTCGATCCCAGGTATGCC-39) and LoxR  (59-TGAACTGATGGCGAGCTCAGACC-39). The gender of each MEF was also determined using the following primer set: mSry-F (59-GTCCCGTGGTGAGAGGCACAAG-39) and mSry-R (59-GCAGCTCTACTCCAGTCTTGCC-39).

Chromatin ImmunoPrecipitation (ChIP) analyses
Chromatins were prepared from two different types of samples, MEF and neonatal brains, according to the method previously described [8]. In brief, the homogenized samples were first crosslinked with 1% formaldehyde for 20 mins, and then lysed with the buffer containing protease inhibitor cocktail (Millipore, Cat. No. 539131). The released nuclei were fractionated with sonication to a pool of DNA fragments size-ranging from 300 to 1,000 bp in length. The prepared chromatin was immunoprecipitated with the following two antibodies: PEG3 (Abcam, Cat. No. ab99252) and H3K9me3 (Abcam, Cat. No. ab8898). Each immunoprecipitated DNA was dissolved in 100 ml of TE for either PCR or qPCR analyses.

Transfection experiments
MEF cells were transfected with the following two constructs, GFP (pIRES-puro-GFP) and FLP (pIRES-puro-FLP), using the GenJect transfection reagent (Cat. No. SL100489-MEF) according to the manufacturer's protocol. Transfection efficiency was monitored through GFP expression under a fluorescence microscope after 24 hours. The transfected cells were harvested at 72hour post transfection for RNA and DNA isolation. The reverted allele of Peg3 by FLP was detected through PCR with the following primer set: Flpko-F (59-CCCTCAGCA-GAGCTGTTTCCTGCC-39) and Flpko-R (59-AAGC-TACCTGGGAAATGAGTGTGG-39).

RNA isolation and qRT-PCR analyses
Total RNA was isolated from either MEF or the brains of oneday-old neonates using a commercial kit (Trizol, Invitrogen) according to the manufacturer's protocol. The total RNA was then reverse-transcribed using the M-MLV kit (Invitrogen), and the subsequent cDNA was used as a template for quantitative PCR. The qRT-PCR analysis was performed with SYBR Select Master Mix (Applied Biosystems, Life Technologies) using the iCycler iQTM multicolor real-time detection system (Bio-Rad). All qRT-PCR reactions were carried out for 40 cycles under standard PCR conditions with internal controls (28S and b-actin). The results derived from qRT-PCR were further analyzed using the threshold (Ct) value. The DCt value was initially calculated by subtracting Ct value of a testing replicate of a given gene from the average Ct value of the internal control (28S and b-actin). The fold difference for each replicate was then calculated by raising the DDCt value as a power of 2 [14]. The average and standard deviation for each sample were then calculated by compiling the normalized values.
The information regarding individual primer sequences is also available (Data S1). Figure S1 (A) A series of RT-PCR analyses using another set of MEF cells show a consistent up-regulation of Zim1 by the mutation on Peg3. This analysis was performed using two internal controls, 28S and b-actin. (B) The up-regulation of Zim1 in KO MEF cell was further analyzed using qPCR. (JPG) Figure S2 (A) A set of independent ChIP experiments with a custom-made anti-PEG3 antibody using the two sets of chromatins prepared from WT and KO (+/2) MEF cells. The DNA from Inputs, Negative controls (Neg), and Immunoprecipitates with anti-PEG3 antibody (PEG3 IP) was used for PCR amplification. This series of ChIP experiments also included another locus, Pgm2l1, as a positive control. (B) qPCR analyses using these ChIP DNA derived from MEF cells. Regions 2 and 3 showed some levels of the enrichment, but Region 3 showed the highest enrichment levels. However, no significant enrichment was detected in Region 1. (C) Western blotting testing the specificity of a new custom-made antibody using the two sets of total protein prepared from WT and KO (+/2) MEF cells.