HMG Domain Containing SSRP1 Is Required for DNA Demethylation and Genomic Imprinting in Arabidopsis

全文

(1)HMG. domain. containing. SSRP1. is. required. for. DNA. demethylation and genomic imprinting in Arabidopsis. Yoko Ikedaa, Yuki Kinoshitaa, Daichi Susakib,Yuriko Ikedaa, Megumi Iwanoc, Seiji Takayamac, Tetsuya Higashiyamab, Tetsuji Kakutanid, Tetsu Kinoshitaa,1. a. Plant Reproductive Genetics, GCOE Research Group, Graduate School of. Biological Sciences, Nara Institute of Science and Technology, 8916-5 Takayama-cho, Ikoma, Nara 630-0192, Japan. b. Division of Biological Science, Graduate School of Science, Nagoya University,. Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan. c. Intercellular communications, Graduate School of Biological Science, Nara. Institute of Science and Technology, 8916-5 Takayama-cho, Ikoma, Nara 630-0192, Japan. d. Department of Integrated Genetics, National Institute of Genetics, 1111 Yata,. Mishima, Shizuoka 411-8540, Japan. 1. To whom correspondence should be addressed. E-mail: [email protected]. Running title: Role of SSRP1 in DNA demethylation Keywords: DNA demethylation, genomic imprinting, SSRP1, gene activation, Arabidopsis endosperm. 1.

(2) Summary In Arabidopsis, DEMETER (DME) DNA demethylase contributes to reprogramming of the epigenetic state of the genome in the central cell. However, other aspects of the active DNA demethylation processes remain elusive. Here we show that Arabidopsis SSRP1, known as an HMG domain-containing component of FACT histone chaperone, is required for DNA demethylation and for activation and repression of many parentally imprinted genes in the central cell. Although loss of DNA methylation releases silencing of the imprinted FWA-GFP, double ssrp1-3;met1-3 mutants surprisingly showed limited activation of maternal FWA-GFP in the central cell, and only became fully active after several nuclear divisions in the endosperm. This behavior was in contrast to the dme-1;met1 double mutant in which hypomethylation of FWA-GFP by met1 suppressed the DNA demethylation defect of dme-1. We propose that active DNA demethylation by DME requires SSRP1 function through a distinctly different process from direct DNA methylation control.. Introduction The control of DNA methylation and histone modifications has been intensively investigated with respect to parent-of-origin specific gene expression (genomic. imprinting). in. Arabidopsis. (Berger. and. Chaudhury,. 2009).. Establishment of the patterns of asymmetry in DNA methylation is thought to be 2.

(3) achieved through a series of passive and active DNA demethylation steps in the central cell, the progenitor cell of endosperm before fertilization. During maturation of female gametophytic cells, passive demethylation is mediated via the transcriptional repression of METHYLTRANSFERASE1 (MET1) by the Retinoblastoma pathway, which may lead to hemimethylated state of DNA methylation (Jullien et al., 2008). Active demethylation is achieved by the enzyme DEMETER (DME), a 5-methylcytosine glycosylase of the BER pathway (Andreuzza et al., 2010; Choi et al., 2002; Gehring et al., 2006). DME demethylates many of the repetitive transposable elements in the female genome (Gehring et al., 2009a; Hsieh et al., 2009b), and induces activation of imprinted protein-coding genes such as MEDEA (MEA), FERTILIZATION INDEPENDENT SEED 2 (FIS2) and FWA in the central cell (Choi et al., 2002; Jullien et al., 2006b; Kinoshita et al., 2004). By contrast to the demethylation of maternally transmitted alleles in the endosperm, the DNA methylation patterns in the corresponding paternal alleles are preserved by a maintenance DNA methyltransferase during formation of the male gametophyte and in the endosperm after fertilization (Huh et al., 2008; Jullien et al., 2006b; Kinoshita et al., 2004). In addition to control by DNA methylation, repression of the paternally derived MEA allele is auto-regulated by the Polycomb Repressive Complex 2 (PRC2) in the endosperm (Baroux et al., 2006; Gehring et al., 2006; Jullien et al., 2006a). Thus, epigenetic reprogramming of imprinted genes in the female central cell before fertilization is a prerequisite step for establishment of 3.

(4) asymmetric gene expression in the endosperm (Hsieh et al., 2011; Huh et al., 2008). To elucidate additional molecular mechanisms of genomic imprinting and DNA demethylation, we identified and characterized mutants that are defective in activation of an imprinted FWA-GFP reporter. Two of the isolated mutants encode an Arabidopsis homologue to SSRP1 (STRUCTURE SPECIFIC RECOGNITION PROTEIN1), an HMG (high mobility group) domain containing non-histone chromosomal protein, which was originally identified as a protein that shows binding affinity for specific DNA structures (Bruhn et al., 1992; Shirakata et al., 1991). SSRP1 was subsequently shown to form a heterodimer with SPT16 in a FACT (facilitates chromatin transcription/transaction) histone chaperone complex in humans, Drosophila and Arabidopsis (Duroux et al., 2004; Lolas et al., 2009; Orphanides et al., 1999; Shimojima et al., 2003). This heterodimer contributes to the remodeling of chromatin by displacing histones H2A and H2B, and influences initiation of transcription, transcription elongation, DNA replication, DNA repair and centromere function (Belotserkovskaya et al., 2003; Formosa, 2008; Heo et al., 2008; Lejeune et al., 2007; Tan et al., 2006). In addition to its role in the FACT histone chaperone, SSRP1 may have other, independent roles as it was purified as a co-activator of p63 transcription factor without SPT16 (Zeng et al., 2002), and transcriptome analysis of knockdown mutants in human cell cultures showed SSRP1 specific targets (Li et al., 2007). Here, we demonstrate another role for SSRP1 in DNA demethylation and control 4.

(5) of imprinted gene expression in the Arabidopsis central cell before fertilization.. Results and Discussion Isolation of DNA demethylation mutants We designed a genetic screen using a pFWA::ΔFWA-GFP reporter construct that has the promoter and 5’ SINE-related sequence, and whose DNA methylation status determines imprinted gene expression (Kinoshita et al., 2007). We screened approximately 1,200 M1 plants and found mutations that displayed a 1:1 segregation ratio of GFP positive to negative fluorescence both before and after fertilization (Figures 1A-1D) (see also Supplemental Experimental Procedures). By using a map-based cloning strategy, we found alleles of SSRP1, encoding a protein known to be a component of a FACT histone chaperone (Figure S1A-S1D). One of these alleles, ssrp1-3, had a premature termination in the conserved structure specific recognition domain in the N-terminus that resulted in a null mutation of ssrp1. A second mutant, ssrp1-4, had truncation of the conserved HMG domain in the C-terminus. Plants heterozygous for this mutation produced viable seeds, while the homozygous were dwarf and sterile. This phenotype is similar to that reported for the ssrp1-2 mutant (Lolas et al., 2009).. Phenotypic analyses of ssrp1 mutants related to genomic imprinting. We asked the question whether SSRP1 affects only FWA or controls other 5.

(6) imprinted genes such as MEA or FIS2. In the female gametophytes of plants with mutations of MEA or FIS2, central cell proliferation (autonomous endosperm) is not suppressed, and silique elongation occurs without fertilization. When mea or fis2 heterozygous mutants are fertilized with wild type pollen, endosperm development proceeds to the syncytial nuclear division stage in a 1:1 ratio, and cellularization of the endosperm fails, resulting in the arrest of embryo growth at the late heart stage (Guitton et al., 2004; Ingouff et al., 2005; Rodrigues et al., 2010). If SSRP1 controls imprinting of MEA or FIS2, then we would expect to observe overlap between the phenotypes of ssrp1 mutants and mutants of other imprinted genes. For example, while dme does not show an autonomous endosperm phenotype, it shows endosperm over-proliferation after fertilization (Choi et al., 2002; Guitton et al., 2004). We observed that heterozygous ssrp1-3 mutant similarly showed nuclear division in the central cell (Figures 1E-1G; Figure S1E) and silique elongation also occurred without fertilization (Figure 1H; Figure S1F). The dividing nucleus also acquired an endosperm character as evidenced by expression of the endosperm marker KS22 (Figures 1I and 1J). In addition to the autonomous endosperm phenotype of maternal ssrp1-3, we observed reduced paternal transmission rates for the ssrp1 mutations (Table S1). This suggests that SSRP1 may also have a role in the vegetative or sperm cells in the male gametophyte, although the maturation of pollen grains in heterozygous mutant plants appeared normal (Figures S2A and S2B) (Lolas et al., 2009). A similar decreased paternal transmission rate is 6.

(7) observed in dme mutants (Schoft et al., 2011). Next, we investigated seed phenotypes in self-pollinated ssrp1-3 plants, and observed two types of abnormal seed. One type aborted at a relatively early stage of seed development, and contained an arrested embryo at the octant stage of embryo development (Figures S2C-S2F). The other showed a brown coloration and aborted at a later stage of seed development, generally 7 to 12 days after pollination (DAP) (Figure S2D). Since the fis mutants show this phenotype at a much higher frequency, SSRP1 is interpreted as having a role in the control of other imprinted genes, and/or additional roles in the embryo and the endosperm. Further analysis confirmed that although paternal transmission of ssrp1-3 is low, the seeds of crosses between WT females and male ssrp1-3 heterozygotes appear normal (Table S2). In the reciprocal cross between female ssrp1-3 heterozygotes and WT males, the rate of early seed abortion is reduced compared to self-pollinated ssrp1-3 heterozygotes but late aborting seeds were still found. The characteristics of this maternal effect on late abortion, namely white and plump seeds, are due to endosperm over-proliferation and abortion with accumulation of brown pigment (Figures 2A-2F). These characteristics overlapped with those reported for mutations of the imprinted MEA and FIS2 genes, although the rate of seed abortion was low compared to those mutants. To investigate the role of SSRP1 in early seed development, we used the weak ssrp1-2 allele (Lolas et al., 2009) as a pollen donor to create the heteroallelic combination of ssrp1-3 and ssrp1-2. Early in seed development, the 7.

(8) female ssrp1-3 showed a 1 : 1 segregation ratio with respect to activation of FWA-GFP (see also in Figure 4). In this cross, therefore, half of the seeds should be heteroallelic i.e. ssrp1-2/ssrp1-3, and the other half heterozygous for ssrp1-2. The observed phenotypic segregation ratio of early aborting to green (normal-looking) seeds was consistent with this expectation (Table S2). The early aborting seeds (Figures 2G-2I), expected to be ssrp1-2/ssrp1-3 heteroalleles, showed arrested embryo development up to the dermatogen stage (Figures 2J-2M), suggesting that SSRP1 is also required for early embryogenesis. To clarify the role of SSRP1 and its dependency on SPT16, its binding partner in a FACT complex, we examined spatial expression patterns of the pSSRP1::SSRP1-mRFP reporter, and also measured SSRP1 and SPT16 transcripts in female gametophytic cells isolated by a micromanipulation system. We found that the fusion protein was localized in the nuclei of the root apex and the lateral primordial cells and in female gametophytic cells; the latter localization was supported by detection of SSRP1 transcripts in the egg and central cells (Figures S2G-S2J). Although SPT16 transcripts can also be detected in these cells, surprisingly, hemizygous T-DNA insertions of spt16-1 or spt16-2 and homozygous insertion of spt16-2 did not cause any defect in activation of FWA-GFP, while heterozygous ssrp1-4 (a weak allele) clearly reduce FWA-GFP expression (Figures 2N-2T). Analyses here and elsewhere (Lolas et al., 2009) have shown that SSRP1 and SPT16 are commonly involved 8.

(9) in many aspects of plant development, which likely represent their roles in a FACT histone chaperone. By contrast, our contrary observation that the FWA-GFP reporter is normally activated even in the homozygous spt16-2 mutant ovule might reflect an SPT16 independent role for SSRP1 (Li et al., 2007; Zeng et al., 2002). However, further analyses, possibly using an spt16 null allele if this can be created are required to confirm an SPT16-dependent or -independent role for SSRP1. In any case, the SSRP1 genotype of the female gametophytic central cell determines the imprinted pattern of FWA expression and the maternal effect on endosperm phenotype.. SSRP1 controls expression of the parental imprinted genes We next investigated expression of imprinted genes using reporter constructs and qRT-PCR. Activation of FWA, and to a lesser extent of FIS2, were impaired in mutant ovules before and after fertilization by a wild type male, although the non-imprinted FIE and DME genes displayed normal expression (Figures 3A and 3B; Figures S3A-S3F). This is consistent with the observation of an overlap in phenotypes between ssrp1-3 and mutants of other imprinted genes. In our qRT-PCR experiment, the level of the MEA transcript in ovules carrying the ssrp1-3 mutant was lower before but not after fertilization (Figures 3A and 3B). This contrasting behavior in gene expression can be explained by the fact that PRC2 represses MEA itself (Baroux et al., 2006; Gehring et al., 2006; Jullien et al., 2006a), hence, absence of PRC2 activity due to the lack of MEA and FIS2 9.

(10) gene expression before fertilization could release repression of MEA in ssrp1-3 after fertilization. Recently genome-wide analyses identified parental imprinted genes and DNA demethylation in the endosperm (Gehring et al., 2009a; Hsieh et al., 2009a; Hsieh et al., 2011; Wolff et al., 2011). We extended our qRT-PCR analysis to some of these genes (Figure 3C). DNA demethylation usually causes increased transcription, so the failure to demethylate should lead to decreased transcription, as we have observed with maternally expressed genes in an ssrp1-3 mutant. However our data also show that these mutants fail to repress paternally expressed genes so their transcription increases. This unexpected result suggests that the epigenetic mechanisms that control gene expression are more complex than anticipated, and are not solely based on the level of DNA methylation (Makarevich et al., 2008). In addition, a recent report suggested that DNA methylation may exclude PRC2 and its repressive effects from some target genes in Arabidopsis (Weinhofer et al., 2010). This may offer a clue to explain our observations, since the maternal alleles of HDG3 and VIM5 are up-regulated in both PRC2 and DNA demethylase mutants (Hsieh et al., 2011). DME has been reported to be predominantly expressed in the central cell (Choi et al., 2002); however, a recent report has indicated that DME is also expressed in proliferating vegetative cells (Kim et al., 2008), raising the possibility that DME might be expressed in the sporophytic cells of the ovule. Therefore, we analyzed transcripts of genes directly in the central cells of WT, 10.

(11) dme-1 and ssrp1-3 using a micromanipulation system (Figure 3D, see also Figure S2J). We found that the levels of FWA transcripts were decreased in the central cells of both dme-1 and ssrp1-3 mutants compared to WT; FIS2 transcripts showed a similar but weaker effect. By contrast, control studies of DME transcripts showed a decrease in the dme-1 central cells, but no difference from WT in ssrp1-3 cells. The relatively low level of MET1 transcripts in the central cell was also not altered by ssrp1-3 (Figure 3D). Thus, by measuring transcripts of DME and MET1 in the single cell type, we excluded the possibility that SSRP1 indirectly affects the level of transcripts of FWA through either DME DNA demethylase or MET1 methyltransferase.. ssrp1 affects DNA demethylation of SINE-related cis-element for FWA We then asked if a maternally-inherited ssrp1-3 mutation affected embryonic or endosperm DNA demethylation of the SINE-related repetitive cis-element that controls FWA imprinting (Kinoshita et al., 2007)(see Supplemental Experimental Procedures)(Figure 3E). Seeds carrying a maternally-inherited ssrp1-3 allele were first identified by their larger size (see Figure 2D), and their genotypes were then confirmed (Figure S3G). In the wild type embryos, the bisulfite sequence analysis of the SINE-related cis-element revealed a high level of CG methylation with moderate levels of CHG and CHH methylation. By contrast, in the wild type endosperm, decreased levels of DNA methylation were found at CG, CHG and CHH sites in the maternally derived FWA allele, which is consistent 11.

(12) with our previous report (Kinoshita et al., 2004). In an embryo inheriting a maternal ssrp1-3 allele, the levels of DNA methylation increased at CG sites in the maternally derived SINE-related cis-element, but was lower at CHG and CHH sites in both the maternal and paternal alleles. Importantly, an endosperm inheriting the maternal ssrp1-3 allele had a considerably higher methylation level of the maternal SINE-related cis-element at all CG, CHG and CHH sites as compared to the wild type. The levels of methylation at CG, CHG, and CHH sites on the paternal allele were not significantly changed in the ssrp1-3 endosperm (Figure 3E). These results suggest that SSRP1 is necessary for DNA demethylation and activation of the FWA maternal allele in the central cell, and therefore in the endosperm. We also examined DNA methylation status by McrBC restriction enzyme digestion followed by PCR and confirmed that ssrp1-3 affected DNA demethylation of the repetitive sequences in the MEA 3’ - ISR region and in the Helitron transposon of MEA 5’ region (Figure S3H).. Genetic dissection of the role of SSRP1 DNA methylation at imprinted genes is antagonistically controlled by the MET1 maintenance DNA methyltransferase and DME DNA demethylase (Jullien et al., 2008; Xiao et al., 2003). We used plants heterozygous for met1-3, dme-1 or ssrp1-3, and also double heterozygotes that were combinations of these mutants, to elucidate the relationship of SSRP1 to DNA demethylation. We crossed female mutant plants bearing pFWA::ΔFWA-GFP with male wild type 12.

(13) Col-0 plants in all experiments. Because the maternal plants are heterozygotes, each mutant allele only shows its defect in the haploid generation of the female gametophyte before fertilization, and its epigenetic state is inherited in the primary endosperm after fertilization. The activation of the FWA reporter gene in heterozygous dme-1 was observed in only 50 % of ovules, whether before or after fertilization with wild type pollen, indicating that maternal transmission of mutant dme-1 conformed to a 1:1 ratio, and confirming that dme-1 mutant cannot activate FWA (Kinoshita et al., 2004). By contrast, in the dme-1;met1-3 double heterozygote, expression of the FWA reporter gene was rescued, such that about 75% of ovules activated the reporter, both before (73.5%; chi square test, χ = 1.11, P>0.25) and after fertilization (75.2%; χ = 0.021, P>0.8) (Figure 4A; see also an additional control in Figure S4), indicating that hypomethylation caused by the met1 mutant allele suppresses the dme-1 mutant phenotype. Thus, even without the active DNA demethylase, maternal FWA alleles became fully active in the central cells when the maintenance DNA methyltransferase was not functional in the gametophytic generation (see also Xiao et al., 2003). The ssrp1-3 heterozygote showed the expected 50% activation both before and after fertilization regardless of the presence of the paternally derived WT SSRP1 allele (Figure 4A). Since ssrp1-3 affects DNA methylation similar to dme-1, we asked if met1-3 suppresses ssrp1-3 defects in the central cell and the endosperm. Interestingly, our analyses of the pFWA::ΔFWA-GFP activation in an ssrp1-3;met1-3 double heterozygote revealed differences compared to those 13.

(14) seen in dme-1;met1-3 plants. Before fertilization, only half of the ovules (46.5%; χ = 2.45, P>0.1) from the ssrp1-3;met1-3 double heterozygotes had a strong GFP signal (Figures 4A and 4C). A small population of the ovules (15.9%) had an unusually weak GFP signal, while the remainder showed no GFP signal (37.7%) (Figures 4A, 4D and 4E). After fertilization, the percentage of ovules that had intense GFP expression increased (Figures 4A and 4F-4K), and the proportion of ovules showing either strong or weak GFP expression at 3 DAP (74.2 %; χ = 0.33, P>0.5) ultimately achieved the expected 3:1 segregation ratio that would occur if met1-3 suppressed ssrp1-3, as observed in the dme-1;met1-3 double heterozygote. Although GFP expression in the ovules from dme-1;met1-3 heterozygotes suggests that the SINE-related cis-element in the FWA promoter is fully hypomethylated in met1-3 ovules, met1-3 does not fully suppress ssrp1-3 before or just after fertilization. Full suppression of the ssrp1-3 allele by met1-3 was observed only after 6-7 nuclear divisions of the primary endosperm. This phenomenon was also seen in autonomous endosperm that lacked a paternal genome contribution (up to 3 nuclear divisions; Figures 4B and 4L-4N). If. maintenance. DNA. methylation. is. non-functional. during. female. gametophyte development, then the three rounds of haploid nuclear division that occur during gametophytic development will dilute the amount of CG methylation that megaspore inherited. Consequently, heterozygotes for met1-3 show ectopic expression of FWA and delayed flowering associated with a loss of DNA 14.

(15) demethylation at the 5’ SINE-related sequence in the vegetative life cycle (Saze et al., 2003; Soppe et al., 2000). However, it was unclear whether transcriptional activation of FWA was caused directly by DNA demethylation or if additional mechanisms are required. Our results demonstrated that while DNA hypomethylation caused by met1-3 in the central cell suppressed the defect in the DME DNA demethylase, met1-3 did not induce full activation of the FWA reporter gene when the SSRP1 gene was mutated. By contrast, after 3 or 6-7 further nuclear divisions in the autonomous endosperm or fertilized endosperm, respectively, the FWA reporter was converted to an almost fully active state. A clue to understanding these observations comes from the fact that DNA replication in a hypomethylated genome induces an alteration from the silent chromatin state to the active chromatin state (Groth et al., 2007). Thus, the ssrp1 mutant’s defects in activation of FWA expression in a hypomethylated genome might be suppressed by this kind of chromatin state transition. The order of the predicted steps for DNA demethylation, i.e., the chromatin based mechanism relating the removal of methylcytosine to transcription activation, are unknown. However, we observed that DNA methylation of FWA remained at a high level and that gene activation was impaired in ssrp1-3, suggesting that DME alone does not induce complete DNA demethylation and full activation of imprinted genes in the central cell. Alternatively, transcriptional activation by SSRP1 might be required for DME demethylation. The latter hypothesis could not be applied to the loss of transcriptional repression of the 15.

(16) maternal alleles for imprinted VIM5 and HDG3 genes by SSRP1 in the central cell. Similarly, it seems unlikely to explain the DNA demethylation of the 5’ promoter region of MEA, the 3’ MEA-ISR and the Helitron transposon (Figure S3H), all of which are located outside the MEA transcription unit mediated by RNA polymerase II. Alteration of the chromatin state by SSRP1 may be relevant to DME-mediated genome-wide DNA demethylation in the central cell. The results presented here, in addition to the role of SSRP1 demonstrated in other organisms, suggest that SSRP1 contributes, at least in part, to a chromatin state transition upon DNA demethylation in the central cell before fertilization. Our analyses of double heterozygous mutants showed differential effects of dme-1 and ssrp1-3 on the release of FWA-GFP silencing in a hypomethylated genome induced by met1-3. This suggests a possible link, which was not evident before, between the mechanisms that affect chromatin configuration and DNA demethylation during epigenetic reprogramming.. Experimental Procedures Procedures for basic experiments and mutant screening are described in the Supplemental Experimental Procedures.. RNA extraction from ovules and qRT-PCR analysis For gene expression analyses using ovules, 30 hand-dissected ovules with GFP fluorescence (inherited wild type allele) or without GFP fluorescence (inherited 16.

(17) mutant allele) were obtained from ssrp1-3+/- mutant plants using a MVX10 microscope (Olympus). To obtain immature seeds at 3 DAP, ssrp1-3+/- plants were pollinated with wild type pollen, and then 15 GFP-positive and 15 GFP-negative seeds were harvested under a fluorescence microscope. Total RNAs were isolated using the PicoPure RNA isolation kit (ARCTURUS) according to the manufacturer’s instructions. Each sample of total RNA was treated with DNase (Promega) and reverse-transcribed in a 20 µl reaction mixture using the Primescript Reverse Transcriptase (TaKaRa) with an oligo dT-21 primer. Ten-fold diluted cDNA was used as a qRT-PCR template. qRT-PCR analysis was performed using SYBR Premix Ex Taq (TaKaRa) and Thermal Cycler Dice (TaKaRa). Gene expression was quantified by absolute quantification. The primer sequences used for PCR are listed in supplemental experimental procedures.. Isolation of central cells and cDNA synthesis Unfertilized ovules from ssrp1-3+/- and dme-1+/- mutant plants were soaked in an enzyme solution [1 % cellulase (Worthington), 0.3 % macerozyme R-10 (Yakult), 0.05 % pectolyase (Kyowa Kasei), 0.45 M mannitol, pH 7.0]. After a few minutes, the cells in the embryo sac were released from the micropyle of the ovule. The released central cells were identified by their relatively large cell volume and amyloplast-rich cytoplasm. Five central cells with or without GFP fluorescence, were collected from ssrp1-3+/- and dme-1+/- mutant ovules using 17.

(18) a Pico-Pipet (Altair) with glass-capillary and a micromanipulator (Eppendorf) under an Axiovert microscope (Carl Zeiss). Total RNAs were isolated using the PicoPure RNA isolation kit (Arcturus) with on-column DNase digestion (Qiagen). Each sample was reverse-transcribed and amplified by in vitro transcription using. SuperScript. III. Reverse. Transcriptase. (Invitrogen). and. MessageBOOSTER cDNA Synthesis Kit for qPCR (Epicentre). Five-fold diluted cDNA was used as the qRT-PCR template. Gene expression was quantified by absolute quantification. The transcript level was normalized against ACT11 expression. The primer sequences used for PCR are listed in supplemental experimental procedures.. References Andreuzza, S., Li, J., Guitton, A.E., Faure, J.E., Casanova, S., Park, J.S., Choi, Y., Chen, Z., and Berger, F. (2010). DNA LIGASE I exerts a maternal effect on seed development in Arabidopsis thaliana. Development 137, 73-81. Baroux, C., Gagliardini, V., Page, D.R., and Grossniklaus, U. (2006). Dynamic regulatory interactions of Polycomb group genes: MEDEA autoregulation is required for imprinted gene expression in Arabidopsis. Genes Dev 20, 1081-1086. Belotserkovskaya, R., Oh, S., Bondarenko, V.A., Orphanides, G., Studitsky, V.M., and Reinberg, D. (2003). FACT facilitates transcription-dependent nucleosome alteration. Science 301, 1090-1093. 18.

(19) Berger, F., and Chaudhury, A. (2009). Parental memories shape seeds. Trends Plant Sci 14, 550-556. Bruhn, S.L., Pil, P.M., Essigmann, J.M., Housman, D.E., and Lippard, S.J. (1992). Isolation and characterization of human cDNA clones encoding a high mobility group box protein that recognizes structural distortions to DNA caused by binding of the anticancer agent cisplatin. Proc Natl Acad Sci U S A 89, 2307-2311. Choi, Y., Gehring, M., Johnson, L., Hannon, M., Harada, J.J., Goldberg, R.B., Jacobsen, S.E., and Fischer, R.L. (2002). DEMETER, a DNA glycosylase domain protein, is required for endosperm gene imprinting and seed viability in arabidopsis. Cell 110, 33-42. Duroux, M., Houben, A., Ruzicka, K., Friml, J., and Grasser, K.D. (2004). The chromatin remodelling complex FACT associates with actively transcribed regions of the Arabidopsis genome. Plant J 40, 660-671. Formosa, T. (2008). FACT and the reorganized nucleosome. Mol Biosyst 4, 1085-1093. Gehring, M., Bubb, K.L., and Henikoff, S. (2009a). Extensive Demethylation of Repetitive Elements During Seed Development Underlies Gene Imprinting. Science 324, 1447-1451. Gehring, M., Huh, J.H., Hsieh, T.F., Penterman, J., Choi, Y., Harada, J.J., Goldberg, R.B., and Fischer, R.L. (2006). DEMETER DNA glycosylase establishes MEDEA polycomb gene self-imprinting by allele-specific 19.

(20) demethylation. Cell 124, 495-506. Gehring, M., Reik, W., and Henikoff, S. (2009b). DNA demethylation by DNA repair. Trends Genet 25, 82-90. Groth, A., Rocha, W., Verreault, A., and Almouzni, G. (2007). Chromatin challenges during DNA replication and repair. Cell 128, 721-733. Guitton, A.E., Page, D.R., Chambrier, P., Lionnet, C., Faure, J.E., Grossniklaus, U., and Berger, F. (2004). Identification of new members of Fertilisation Independent Seed Polycomb Group pathway involved in the control of seed development in Arabidopsis thaliana. Development 131, 2971-2981. Heo, K., Kim, H., Choi, S.H., Choi, J., Kim, K., Gu, J., Lieber, M.R., Yang, A.S., and An, W. (2008). FACT-mediated exchange of histone variant H2AX regulated by phosphorylation of H2AX and ADP-ribosylation of Spt16. Mol Cell 30, 86-97. Hsieh, T.-F., Ibarra, C.A., Silva, P., Zemach, A., Eshed-Williams, L., Fischer, R.L., and Zilberman, D. (2009a). Genome-Wide Demethylation of Arabidopsis Endosperm. Science 324, 1451-1454. Hsieh, T.F., Ibarra, C.A., Silva, P., Zemach, A., Eshed-Williams, L., Fischer, R.L., and Zilberman, D. (2009b). Genome-wide demethylation of Arabidopsis endosperm. Science 324, 1451-1454. Hsieh, T.F., Shin, J., Uzawa, R., Silva, P., Cohen, S., Bauer, M.J., Hashimoto, M., Kirkbride, R.C., Harada, J.J., Zilberman, D., et al. (2011). Inaugural 20.

(21) Article: Regulation of imprinted gene expression in Arabidopsis endosperm. Proc Natl Acad Sci U S A. Huh, J.H., Bauer, M.J., Hsieh, T.-F., and Fischer, R.L. (2008). Cellular Programming of Plant Gene Imprinting. Cell 132, 735-744. Ikeda, Y., and Kinoshita, T. (2009). DNA demethylation: a lesson from the garden. Chromosoma 118, 37-41. Ingouff, M., Haseloff, J., and Berger, F. (2005). Polycomb group genes control developmental timing of endosperm. Plant J 42, 663-674. Jullien, P.E., Katz, A., Oliva, M., Ohad, N., and Berger, F. (2006a). Polycomb group complexes self-regulate imprinting of the Polycomb group gene MEDEA in Arabidopsis. Curr Biol 16, 486-492. Jullien, P.E., Kinoshita, T., Ohad, N., and Berger, F. (2006b). Maintenance of DNA methylation during the Arabidopsis life cycle is essential for parental imprinting. Plant Cell 18, 1360-1372. Jullien, P.E., Mosquna, A., Ingouff, M., Sakata, T., Ohad, N., and Berger, F. (2008). Retinoblastoma and its binding partner MSI1 control imprinting in Arabidopsis. PLoS Biol 6, e194. Kim, M., Ohr, H., Lee, J.W., Hyun, Y., Fischer, R.L., and Choi, Y. (2008). Temporal and spatial downregulation of Arabidopsis MET1 activity results in global DNA hypomethylation and developmental defects. Mol Cells 26, 611-615. Kinoshita, T., Miura, A., Choi, Y., Kinoshita, Y., Cao, X., Jacobsen, S.E., 21.

(22) Fischer, R.L., and Kakutani, T. (2004). One-way control of FWA imprinting in Arabidopsis endosperm by DNA methylation. Science 303, 521-523. Kinoshita, Y., Saze, H., Kinoshita, T., Miura, A., Soppe, W.J., Koornneef, M., and Kakutani, T. (2007). Control of FWA gene silencing in Arabidopsis thaliana by SINE-related direct repeats. Plant J 49, 38-45. Lejeune, E., Bortfeld, M., White, S.A., Pidoux, A.L., Ekwall, K., Allshire, R.C., and Ladurner, A.G. (2007). The chromatin-remodeling factor FACT contributes to centromeric heterochromatin independently of RNAi. Curr Biol 17, 1219-1224. Li, Y., Zeng, S.X., Landais, I., and Lu, H. (2007). Human SSRP1 has Spt16-dependent and -independent roles in gene transcription. J Biol Chem. 282, 6936-6945. Lolas, I.B., Himanen, K., Gronlund, J.T., Lynggaard, C., Houben, A., Melzer, M., Van Lijsebettens, M., and Grasser, K.D. (2009). The transcript elongation factor FACT affects Arabidopsis vegetative and reproductive development and genetically interacts with HUB1/2. Plant J 61, 686-697. Makarevich, G., Villar, C.B., Erilova, A., and Kohler, C. (2008). Mechanism of PHERES1 imprinting in Arabidopsis. J Cell Sci 121, 906-912. Orphanides, G., Wu, W.H., Lane, W.S., Hampsey, M., and Reinberg, D. (1999). The chromatin-specific transcription elongation factor FACT comprises human SPT16 and SSRP1 proteins. Nature 400, 284-288. Rodrigues, J.C., Luo, M., Berger, F., and Koltunow, A.M. (2010). Polycomb 22.

(23) group gene function in sexual and asexual seed development in angiosperms. Sex Plant Reprod 23, 123-133. Saze, H., Mittelsten Scheid, O., and Paszkowski, J. (2003). Maintenance of CpG methylation is essential for epigenetic inheritance during plant gametogenesis. Nat Genet 34, 65-69. Schoft, V.K., Chumak, N., Choi, Y., Hannon, M., Garcia-Aguilar, M., Machlicova, A., Slusarz, L., Mosiolek, M., Park, J.S., Park, G.T., et al. (2011). Function of the DEMETER DNA glycosylase in the Arabidopsis thaliana male gametophyte. Proc Natl Acad Sci U S A 108, 8042-8047. Shimojima, T., Okada, M., Nakayama, T., Ueda, H., Okawa, K., Iwamatsu, A., Handa, H., and Hirose, S. (2003). Drosophila FACT contributes to Hox gene expression through physical and functional interactions with GAGA factor. Genes Dev 17, 1605-1616. Shirakata, M., Huppi, K., Usuda, S., Okazaki, K., Yoshida, K., and Sakano, H. (1991). HMG1-related DNA-binding protein isolated with V-(D)-J recombination signal probes. Mol Cell Biol 11, 4528-4536. Soppe, W.J., Jacobsen, S.E., Alonso-Blanco, C., Jackson, J.P., Kakutani, T., Koornneef, M., and Peeters, A.J. (2000). The late flowering phenotype of fwa mutants is caused by gain-of-function epigenetic alleles of a homeodomain gene. Mol Cell 6, 791-802. Tan, B.C., Chien, C.T., Hirose, S., and Lee, S.C. (2006). Functional cooperation between FACT and MCM helicase facilitates initiation of 23.

(24) chromatin DNA replication. EMBO J 25, 3975-3985. Weinhofer, I., Hehenberger, E., Roszak, P., Hennig, L., and Kohler, C. (2010). H3K27me3 profiling of the endosperm implies exclusion of polycomb group protein targeting by DNA methylation. PLoS Genet 6. Wolff, P., Weinhofer, I., Seguin, J., Roszak, P., Beisel, C., Donoghue, M.T., Spillane, C., Nordborg, M., Rehmsmeier, M., and Kohler, C. (2011). High-Resolution Analysis of Parent-of-Origin Allelic Expression in the Arabidopsis Endosperm. PLoS Genet 7, e1002126. Xiao, W., Gehring, M., Choi, Y., Margossian, L., Pu, H., Harada, J.J., Goldberg, R.B., Pennell, R.I., and Fischer, R.L. (2003). Imprinting of the. MEA Polycomb gene is controlled by antagonism between MET1 methyltransferase and DME glycosylase. Dev Cell 5, 891-901. Zeng, S.X., Dai, M.S., Keller, D.M., and Lu, H. (2002). SSRP1 functions as a co-activator of the transcriptional activator p63. EMBO J 21, 5487-5497. Zhu, J.K. (2009). Active DNA demethylation mediated by DNA glycosylases. Annu Rev Genet 43, 143-166.. Acknowledgements We thank Akiko Terui and Shoko Fukudome for technical assistance with the genetic mapping; Diana Buzas, Mitsuhiro Aida, Yutaka Sato, Susumu Hirose, Matthew Bauer and Robert Fischer for critical comments on the manuscript; Robert Fischer, Hidetoshi Saze, Yoenhee Choi, Klaus Grasser and Frederic 24.

(25) Berger provided seed stocks. Supported by Grant-in-Aid for Scientific Research on Priority Areas (18075010 to T.K.) and Grant-in-Aid for Scientific Research on Innovative Areas (23113003 to T.K.) and Global COE Program (Frontier Biosciences: strategies for survival and adaptation in a changing global environment) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and by the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN 03-01 to T.K.), and the Sasakawa Scientific Research Grant from the Japan Science Society to Y.I.. Figure legends Figure 1. SSRP1 is required for activation of imprinted FWA-GFP (A-D) Fluorescence images of pFWA::ΔFWA-GFP ovules before fertilization (A and B), and of seeds at 2 days after pollination (DAP) seeds (C and D). (E-G) The central cell in WT (E) and ssrp1-3 ovules (F and G) at 6 days after emasculation (DAE). Egg cell (blue), central cell nucleus and primary endosperm nuclei (pink) are pseudocolored. (H) Silique elongation in ssrp1 and wild type at 7 DAE. (I and J) Expression pattern of the KS22 endosperm marker in the mature female gametophyte (I) and in the dividing nuclei of the central cell without fertilization (J). Note that all ovules containing autonomous endosperm expressed the KS22 marker. Background expression of KS22 can also be seen in the integuments at the micropylar end. Bars = 50 µm. See also Figure S1. 25.

(26) Figure 2. Effects on imprinted seed phenotypes in Arabidopsis (A-C) Phenotypes of developing F1 seeds at 4 DAP (A), 7 DAP (B) and 15 DAP (C), in a cross between female ssrp1-3+/- and male wild type. Orange arrowheads indicate a class of seeds with a white and plump phenotype; some of these are aborted at a late stage as indicated by the orange arrowhead in (C). (D-F) A wild type seed (left in D and E) and a seed showing endosperm over-proliferation (right in D and F) in a cross between female ssrp1-3+/- and male wild type. (G-I) Phenotypes of heteroallelic seeds from a cross between female ssrp1-3 +/crossed with male ssrp1-2 -/- at 4 DAP (G), 7 DAP (H) and 15 DAP (I). White arrowheads show small shriveled seeds at an early stage. (J-M) Images of cleared seeds of wild type (J and L) and of heteroallelic seeds from a cross between female ssrp1-3+/- and male ssrp1-2-/- (K and M) at 4 DAP (J and K) and 6 DAP (L and M). (N-T) Fluorescence images of pFWA::∆FWA-GFP in heterozygous spt16-1 (N and O), heterozygous spt16-2 (P and Q), heterozygous ssrp1-4 (R and S), and homozygous spt16-2 (T) mutants, before fertilization (N, P, R, and T) and after fertilization (O, Q, and S). Shrinkage of the embryo sac at the micropylar end and absence of central cell nuclei (white arrowhead) are visible in (Q). These aborted ovules are GFP negative. Bars = 200 µm (A-D and G-I), 50 µm (E, F and J-T). See also Figure S2 and Tables S1 and S2. 26.

(27) Figure 3. Expression and methylation status of imprinted genes (A-C) qRT-PCR analysis of well known imprinted genes in ovules that inherited either the wild type allele and the ssrp1-3 allele before fertilization (A) or at 3 DAP with wild type pollen (B). Expression patterns of the selected imprinted genes from a recent genome wide analysis of the ssrp1-3 ovules are displayed in (C). The transcript levels were normalized against UBQ10 expression for (A-B) and ACT11 for (C). Error bars represent s.d.; n = 3 biological replicates. (D) (Top) Central cells with or without GFP were released, and collected under a microscope. Bars = 50 µm. (Bottom) qRT-PCR analysis of FWA, FIS2 and DME in the isolated wild type, dme-1 and ssrp1-3, central cells. The transcript level was normalized against ACT11 expression. Error bars represent s.d.; n = 3 biological replicates. (E) Percent methylation of the 5’ SINE-related repeat of FWA in the WT embryo and endosperm or after maternal inheritance of ssrp1-3. Methylation levels for each fraction were determined by bisulfite sequencing. n = number of sequenced amplicons. See also Figure S3.. Figure 4. Genetic interactions of DNA methyltransferase (met1), DNA demethylase (dme) and the HMG domain gene (ssrp1) (A) The frequencies of non-fluorescent and fluorescent ovules in plants with different genotypes; wild type, heterozygotes for met1-3, dme-1 or ssrp1-3, 27.

(28) double heterozygotes for dme-1;met1-3, or ssrp1-3;met1-3. The proportions (%) of ovules with strong fluorescence (green box), weak fluorescence (yellow-green box) or non-fluorescence (cream box) were determined during endosperm development after fertilization with wild type pollen. n = number of counted ovules. (B) Proportion (%) of fluorescent ovules in the absence of fertilization. In ssrp1-3 heterozygotes (left), a 1:1 segregation of fluorescence was consistently observed from 0 to 3 days after maturation of the female gametophyte (DAM) regardless of nuclear division in the autonomous endosperm. (C–N), Fluorescence images of pFWA::ΔFWA-GFP in the double heterozygote for ssrp1-3 met1-3 before fertilization (C-E), in ovules 1 DAP (F-H), 3 DAP (I-K) and 2 DAM of the female gametophyte without fertilization (I-N). The white arrowheads indicate autonomous endosperm nuclei. Bars = 50 µm. See also Figure S4.. 28.

(29)