COLLOQUIUM PAPER

Angelman syndrome imprinting center encodes a transcriptional promoter Michael W. Lewisa, Jason O. Brantb, Joseph M. Kramerc, James I. Mossd, Thomas P. Yangb, Peter Hansend, R. Stan Williamsc, and James L. Resnicka,1 a Department of Molecular Genetics and Microbiology, University of Florida College of Medicine, Gainesville, FL 32610-0266; bDepartment of Biochemistry and Molecular Biology, University of Florida College of Medicine, Gainesville, FL 32610-0245; cDepartment of Obstetrics and Gynecology, University of Florida College of Medicine, Gainesville, FL 32610-0294; and dDepartment of Animal Sciences, University of Florida, Gainesville, FL 32611-0910

Clusters of imprinted genes are often controlled by an imprinting center that is necessary for allele-specific gene expression and to reprogram parent-of-origin information between generations. An imprinted domain at 15q11–q13 is responsible for both Angelman syndrome (AS) and Prader–Willi syndrome (PWS), two clinically distinct neurodevelopmental disorders. Angelman syndrome arises from the lack of maternal contribution from the locus, whereas Prader–Willi syndrome results from the absence of paternally expressed genes. In some rare cases of PWS and AS, small deletions may lead to incorrect parent-of-origin allele identity. DNA sequences common to these deletions define a bipartite imprinting center for the AS–PWS locus. The PWS–smallest region of deletion overlap (SRO) element of the imprinting center activates expression of genes from the paternal allele. The AS–SRO element generates maternal allele identity by epigenetically inactivating the PWS–SRO in oocytes so that paternal genes are silenced on the future maternal allele. Here we have investigated functional activities of the AS–SRO, the element necessary for maternal allele identity. We find that, in humans, the AS–SRO is an oocyte-specific promoter that generates transcripts that transit the PWS–SRO. Similar upstream promoters were detected in bovine oocytes. This result is consistent with a model in which imprinting centers become DNA methylated and acquire maternal allele identity in oocytes in response to transiting transcription. Angelman syndrome

| imprinting | oocytes

T

he process of genetic imprinting results in the selective silencing of one parent’s alleles. Imprinted genes are often clustered and have an imprinting center (IC) that regulates both allele-specific gene expression and the resetting of parental allele identity between generations. The imprinted domain at 15q11– 15q13 is unusual not only in that it contains both maternally and paternally imprinted genes, but also in that it underlies two clinically distinct syndromes. Angelman syndrome (AS) [Mendelian Inheritance in Man (MIM) 105830] results from the absence of a functional UBE3A gene product transcribed from the maternal allele (1, 2). Several genes are expressed only from the paternal allele at this locus, and mutations in two of these genes are currently implicated in traits associated with Prader–Willi syndrome (PWS) (MIM 17620) (3, 4). Most individuals with either AS or PWS harbor a 5- to 7-Mb deletion spanning the entire imprinted domain, but some cases of AS or PWS are the result of a failure to reset parent-of-origin allele identity between generations. These AS and PWS imprinting defects may be accompanied by much smaller deletions of sequences essential to the imprint resetting process. The smallest regions of overlap common to these deletions indicate sequences necessary for allele identity and thereby define the AS–PWS imprinting center. Thus, the IC at the PWS–AS locus is defined by genetic lesions detected in individuals with AS or PWS imprinting defects, rather than by engineered mutations in mice as has been used at most imprinted domains (5). The PWS–AS imprinting center is www.pnas.org/cgi/doi/10.1073/pnas.1411261111

also atypical in that it consists of two separate DNA elements. The 4.3-kb PWS–smallest region of deletion overlap (SRO) spans the SNURF/SNRPN (henceforth SNRPN) major promoter and exon 1. This element is required for paternal expression of genes implicated in PWS (6). The 880-bp AS–SRO is located ∼35 kb centromeric and upstream of the PWS–SRO. This element is deleted in some individuals with AS imprinting defects and is required for silencing of PWS genes on the maternal chromosome and subsequent expression of UBE3A (7). Here we will use the terms AS–SRO and PWS–SRO to refer to the sequences deleted in individuals with AS and PWS imprinting defects and use AS–IC and PWS–IC to refer to the imprinting activities of these elements. The prevailing model of the roles of these two elements posits that the PWS–IC functions somatically to activate paternal allele transcription. The PWS–IC activates several protein-encoding genes including MKRN3, MAGEL2, NECDIN, and SNRPN. The SNRPN gene is part of complex transcript that is also host to several small nucleolar RNAs and to UBE3A–ATS. UBE3A–ATS is transcribed in antisense direction to UBE3A and silences the paternal UBE3A allele, potentially by a transcriptional collision mechanism (8). The AS–IC functions in female germ cells to epigenetically inactivate the PWS–IC and thereby silence paternally expressed genes on the future maternal allele (9). The absence of maternal UBE3A–ATS transcription on the maternal allele is necessary for UBE3A expression and the avoidance of AS. Targeted mutations in mice confirm that the PWS–IC functions somatically, activates paternal gene expression, and is necessary to silence the paternal Ube3a allele (10–14). However, the absence of a conserved AS–SRO sequence in mice has stymied functional investigations into this element of the imprinting center. Horsthemke and colleagues were, to our knowledge, the first to suggest that the AS–IC uses transcription to inactivate the PWS–IC (15). These investigators noted that the SNRPN transcription unit contains several upstream (U) exons located over 130 kb that splice into SNRPN exon 2, thereby replacing exon 1. Two of these U exons are located within the 880-bp AS–SRO (15–17). Extrapolating from these paternally derived transcripts detected in brain tissue, they speculated the existence of similar

This paper results from the Arthur M. Sackler Colloquium of the National Academy of Sciences, “Epigenetic Changes in the Developing Brain: Effects on Behavior,” held March 28–29, 2014, at the National Academy of Sciences in Washington, DC. The complete program and video recordings of most presentations are available on the NAS website at www.nasonline.org/Epigenetic_changes. Author contributions: M.W.L. and J.L.R. designed research; M.W.L. and J.O.B. performed research; J.M.K., J.I.M., T.P.Y., P.H., and R.S.W. contributed new reagents/analytic tools; M.W.L., J.O.B., T.P.Y., and J.L.R. analyzed data; and M.W.L. and J.L.R. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. E.B.K. is a guest editor invited by the Editorial Board. 1

To whom correspondence should be addressed. Email: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1411261111/-/DCSupplemental.

PNAS Early Edition | 1 of 5

GENETICS

Edited by Eric B. Keverne, University of Cambridge, Cambridge, United Kingdom, and accepted by the Editorial Board September 3, 2014 (received for review July 14, 2014)

transcripts in oocytes that would transit the PWS–SRO and thereby contribute to the imprint-setting process. Although lacking sequence homology, the mouse Snrpn locus also contains several upstream promoter exons that are alternatives to body exon 1 (10, 18). These transcripts become detectable in mouse oocytes just before the application of DNA methylation at the PWS–IC (19, 20). Using BAC transgenes containing both mouse Snrpn upstream and body exons, we recently demonstrated that these oocyte transcripts transiting the PWS–IC are necessary and sufficient to observe both DNA methylation at the PWS–IC and imprinted Snrpn expression (20). These results suggest that imprint setting at the PWS–AS locus is part of a generalized oocyte mechanism in which transcribed CpG islands become methylated (21, 22). Although transcription transiting the human PWS–SRO is likely essential to imprint-setting activity, the molecular functions of the human AS–SRO remain unclear. We have now investigated potential activities of the AS–SRO in human oocytes. In contrast to the patterns observed in brain, we find that few oocyte SNRPN transcripts initiate at the major promoter at exon 1. Instead, the AS–SRO is a frequent start site for oocyte SNRPN transcripts. Upstream initiation likely ensures that transcripts cross the complete PWS–SRO and contribute to setting the maternal imprint. Results Previously Unidentified SNRPN Transcripts in Human Oocytes. The SNRPN transcription unit has a complex pattern of transcription initiation and processing in somatic tissues. The University of California at Santa Cruz (UCSC) Genome Browser and Fig. 1 demonstrate that somatic SNRPN transcripts most commonly initiate at exon 1, and nearly all remaining transcripts have 5′ ends at either exon U1B or U1A, located 130 and 100 kb upstream,

respectively (23). Transcripts initiated at U1A or U1B may also include one or more internal upstream exons, numbered U2–U8 distally to proximally, before splicing into SNRPN body exon 2. These primary transcripts thus transit across the entire PWS– SRO (17). Previous work indicates that imprint setting on the maternal allele at the mouse PWS–AS locus is dependent on Snrpn transcripts transiting the PWS–IC in mouse oocytes (20). To better understand the role that transcription plays in imprint setting at the human PWS–AS locus, we applied 5′ rapid amplification of cDNA ends (5′ RACE) to human oocyte RNA. To ensure that our results would not be limited to a rare pattern unique to one oocyte or a single donor, we characterized RNA obtained from pools composed of six oocytes retrieved from at least three donors. A primer located in SNRPN body exon 3 was used to initiate cDNA synthesis, and a second reverse primer in SNRPN exon 2 was used for PCR amplification before cloning. 5′ RACE detected five previously undocumented SNRPN transcription start sites in oocytes, all within or near the AS–SRO, the smallest region of overlapping deletions present in some individuals with imprinting defects (Fig. 1B). Most RACE clone 5′ ends fell within the 880-bp AS–SRO, but several were located just 40 bp downstream of the proximal boundary of the AS–SRO in a previously unidentified exon named “U6.5.” In brain and testis, the rare transcripts may have 5′ ends even further upstream than UIB (17), but RACE using reverse primers in U2 failed to detect additional upstream 5′ ends in oocytes. We were surprised by the absence of SNRPN exon 1-derived transcripts among the oocyte RACE products. To confirm these findings, we performed RT-PCR using forward primers in previously known or in newly discovered exonic sequence paired with a reverse primer in SNRPN body exon 3. The products were then Southern blotted using a probe containing body exon 2 to identify transcripts that had transited across the PWS–SRO.

Fig. 1. SNRPN upstream exon transcripts in human oocytes and brain. (A) Map of the human AS–PWS-imprinted domain. Genes shown in green are expressed from the paternal (pat) allele. UBE3A is expressed from the maternal (mat) allele. The 880-bp AS–SRO and 4.3-kb PWS–SRO are shown in red and blue, respectively. Not all products of the locus are shown, and the map is not to scale. (B) Structure of transcripts identified by 5′ RACE performed on human oocyte RNA. Exons located within the AS–SRO are shown in red. Other U exons are shown in orange. Open boxes are exonic sequences not previously documented in the UCSC Genome Browser. Numbers to the right indicate the number of clones with the indicated structure. Arrows identify the locations of RT-PCR primers used below. (C) RT-PCR analysis of U exons in whole brain and oocytes. The location of forward primers used in each lane is indicated in B above. In each case, the reverse primer was located in body exon 3. SNRPN products were identified with an exon 2 probe. Multiple bands in a lane likely represent inclusion of additional exons in the product.

2 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1411261111

Lewis et al.

Characterization of the Bovine SNRPN Locus. As the AS–SRO se-

quence is not conserved from humans to mice, we sought to identify another species in which to test its transcription initiation function. The National Center for Biotechnology Information BLAST tool does identify an 880-bp sequence located 40 kb upstream of the bovine SNRPN gene with 68% overall identity (Fig. S2). Similar to the mouse, exon 1 of the bovine SNRPN gene is hypermethylated on the maternal allele, with methylation acquisition in the oocyte being related to oocyte size (24, 25). To date, no bovine SNRPN upstream exons appear in the UCSC Genome Browser, and application of 5′ RACE to bovine oocyte or hypothalamus RNA identified 5′ ends only in the vicinity of body exon 1 (Fig. 3A). However, GENSCAN (26) identified three potential exons near the conserved AS–SRO sequence of the bovine genome. RT-PCR using forward primers in these putative

Discussion Results obtained primarily in mice point to a multistep pathway by which parental imprints are reset between generations. Parent-of-origin information, in the form of DNA methylation at imprinting control regions, is first applied during gametic development. These germline differentially methylated regions DMRs (gDMRs) are retained despite a widespread demethylation of both parental genomes during early cleavage divisions. Although maintained during somatic development, differential methylation at gDMRs is erased during fetal germ-cell development, particularly as primordial germ cells complete their migration to the fetal gonad (27). In female embryos, DMR methylation first appears shortly after birth in a DNMT3A- and DNMT3L-dependent process that is tied to oocyte growth (25, 28, 29). How specific regions are selected for methylation in oocytes is poorly understood, but substantial recent work indicates the essential role of transcription in setting methylation imprints in the female germline. Many maternal gDMRs are intergenic sites overlying promoters for transcripts that, in turn, silence genes in cis (22). Chotalia et al. (30) first demonstrated the necessity for transcription-transiting maternal DMRs in oocytes to effect imprinting at the GNAS locus. At this locus, truncation of oocyte transcripts before the gDMR prevents acquisition of the methylation imprint. At the murine PWS–AS locus,

Fig. 2. RNA seq analysis of human oocyte RNA. RNA-Seq data of SNRPN upstream exon transcripts in human oocytes and brain tissues. Data are displayed as custom tracks in the UCSC Genome Browser. (A) SNURF/SNRPN gene locus, including upstream transcripts. (B) Location of the upstream exons U3–U8 and body exons 1 and 2. Exons U5 and U6 are located within the 880-bp AS–SRO. (C) RNA-Seq data for human oocytes (raw data obtained from GEO studies GSE36552 and GSE44183). (D) RNA-Seq data from human brain tissue. Tracks in C and D are scaled from 0 to 200.

Lewis et al.

PNAS Early Edition | 3 of 5

COLLOQUIUM PAPER

exons paired with a reverse primer in exon 3 on bovine oocyte and hypothalamus RNA is shown in Fig. 3B. As predicted by the 5′ RACE and in contrast to humans, bovine oocytes use exon 1 as a transcription start site. Transcripts from within the AS–SRO conserved sequence were not detected in either oocytes or hypothalamus. However, similar to human oocytes, bovine oocytes use a transcription start site located just downstream of the AS– SRO sequence. Performing RACE with primers in either exons B or C did not lead to the identification of additional upstream exons. Taken together, these data suggest that the conserved AS–SRO sequence functions as a promoter to generate oocytespecific transcripts that are necessary to establish the maternal imprint at the PWS–AS locus.

GENETICS

Results in Fig. 1C corroborate the 5′ RACE results and show that exon 1 is rarely used in human oocytes. This is in stark contrast to its use in the brain as the major promoter. These results also illustrate the nearly reciprocal patterns of SNRPN transcription initiation in brain and oocytes. Although the AS– SRO is largely unused in brain, it is a focal point for transcription initiation in oocytes. Also notable is the absence of transcripts that initiate at U1B or U1A in oocytes, the other commonly used transcription start sites in brain. Our Southern blot results confirmed their use in the brain, but did not detect their utilization in oocytes. We did detect transcripts initiating at UIB in one pool of oocytes— suggesting that there may be small discrepancies between individual oocytes or donors—but larger distinctions between brain and oocyte transcription are still evident. During the course of this work, two human oocyte RNA-seq datasets became available in the Gene Expression Omnibus. Query of these two datasets confirms the main findings above. In oocytes, nearly all SNRPN transcripts initiate at the AS–SRO or at U6.5 in strong preference to exon 1 (Fig. 2). Also, as found by RT-PCR, RNA-Seq did not detect use of the upstream somatic promoters U1A and U1B by oocytes (Fig. S1).

Fig. 3. SNRPN upstream exon transcripts in bovine oocytes and hypothalamus. The location of the 880-bp conserved AS–SRO sequence is shown as an open box 39.5 kb upstream of exon 1. The location of new upstream exons (red) relative to body exon 1 is indicated. SNRPN body exons are in green. (A) 5′ RACE was performed with anchor primers in SNRPN body exons 2 and 3. Most RACE products in hypothalamus and oocytes had 5′ ends at body exon 1. “H” and “O” refer to RACE products identified in the hypothalamus or oocyte, respectively. A new variably sized exon (white) located between body exons and 1 and 2 was found in several RACE products. This exon would be expected to disrupt expression of SNURF (45). (B) Southern blot of RT-PCR products amplified with forward primers in putative exons suggested by GENSCAN and reverse primer in exon 3. The blot was probed with body exon 2.

transcripts derived from several promoters located upstream of the PWS–IC first appear in perinatal oocytes. These transcripts are necessary for methylation and silencing of Snrpn on the maternal allele (20). The AS–SRO is defined as the smallest region of overlap common to deletions associated with AS-imprinting defects. U5 and U6, the exons encoded within the 880-bp AS–SRO, have both consensus splice acceptor and donor sequences, consistent with being spliced into transcripts that initiate upstream at U1B or U1A (17). Here we have explored the transcriptional activity of the AS–SRO. We find that the AS–SRO is the primary transcription start site for SNRPN transcripts in human oocytes. These transcripts use the U5 or U6 splice donor sequences to splice into SNRPN body exon 2. We interpret these results to indicate that the major function of the AS–SRO in oocytes is to drive transcripts across the PWS–IC and thereby set the imprint. We infer that imprint-setting transcripts are reduced or absent from genomes lacking the AS–SRO, resulting in AS-imprinting defects. The mouse lacks a conserved AS–SRO sequence. Instead, the mouse has nine highly repeated upstream promoters located over 530 kb (31). Although all nine repeats are unlikely to be active promoters, it is clear that some are used in both oocytes and somatic cells (10, 18–20, 32). Removal of the three proximal promoters leads to a weak partial imprinting defect, consistent with functional complementation by the remaining upstream promoters (33). Conversely, BAC transgenes bearing only the three proximal promoters generate sufficient transcripts for efficient imprinting (20). How many times the transcription apparatus must transit the PWS–IC to set the imprint is unknown. However, the partial imprinting defect seen upon deletion of the three proximal promoters points to an increased risk of an imprinting defect with declining numbers of transcripts, in turn suggesting that multiple transits of the PWS–IC are necessary for reliable imprint setting in the mouse. Curiously, a polymorphism 4 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1411261111

located within the AS–SRO is associated with an increased risk of AS-imprinting defects but is also present in unaffected individuals (34). Assuming that this polymorphism affects the frequency of transcript initiation, multiple transits of the PWS– SRO are also likely for reliable imprint setting in humans. We examined the transcription initiation potential of AS– SRO-like sequences in the cow. Here the AS–SRO region also has transcription initiation activity, consistent with the conserved sequence harboring oocyte-specific promoter elements. Also similar to humans, bovine oocytes use a transcription start site just downstream of the 3′ border of the AS–SRO. This work identifies SNRPN upstream exons that had not been previously documented in somatic tissues. The upstream exon U6.5 appears to be robustly used in oocytes and is located only 40 bp downstream of the AS–SRO. Several explanations are possible for its location outside of the AS–SRO. First, RACE may not accurately identify the 5′ ends of U6.5 initiation sites that could lie farther upstream within the AS–SRO. Alternatively, the AS–SRO may generate sufficient transcripts to set an imprint in the absence of U6.5. Finally, the downstream border of the AS–SRO is defined by a single case of an AS-imprinting defect, as all other deletions in the area likely impinge on U6.5 (5, 7). Whether the absence of U6.5 contributed to the imprinting defect in these other cases is unknown. Mosaicism is detected in about 30% of individuals diagnosed with AS-imprinting defects (35, 36). These are presumably cases in which the imprint was lost postfertilization. Deletions impinging on the AS–SRO have been detected in about 15–20% of imprinting defect cases. This implies that in about half of individuals with AS-imprinting defects, transcripts from the AS– SRO were not generated or the imprint at the PWS–SRO was not set in response to transiting transcription. How does transcription lead to DNA methylation in oocytes? Oocyte availability is a challenge to molecular studies, but Kelsey and Feil (37) have speculated that the act of transcription results in a constellation of chromatin modifications that are conducive to interaction of DNMT3A and DNMTL, whereas other transcribed regions might be protected from methylation by CXXCdomain proteins. The relative abundance of bovine oocytes compared with those of humans or even rodents may make this system especially useful in the molecular testing of these ideas. Materials and Methods Human oocytes were collected from consented women attempting conception through in vitro fertilization, as well as from oocyte donors, at the University of Florida Reproductive Medicine clinic. In general, patients underwent ovarian stimulation using standard long Lupron suppression or GnRH antagonist protocols. Oocyte retrieval was scheduled for 36 h after administration of an ovulation trigger. Oocytes used in this study were derived from those that failed to fertilize following conventional insemination or from immature oocytes not included for intracytoplasmic sperm injection. Oocytes were briefly rinsed in Hepes-buffered medium (Cooper Surgical) and snap-frozen at −80 °C. Human oocyte donation was approved by the University of Florida Institutional Review Board, Project 356–2012. Bovine cumulus–oocyte complexes were obtained as previously described (38), using a scalpel to slice 2- to 10-mm follicles on the surface of ovaries obtained from a local abattoir. Cumulus cells were removed from the oocytes by vortexing in a solution of 1,400 units/mL hyaluronidase in synthetic oviduct fluid (39) buffered with 10 mM Hepes in a 2-mL microcentrifuge tube for 3–4 min at 38 °C. The oocytes were then washed three times in Dulbecco’s PBS containing 0.1% (wt/vol) polyvinylpyrollidone (DPBS/ PVP), placed in warm drops of 0.1% (wt/vol) protease from Streptomyces griseus for 5 min to remove zona pellucidae and attached debris, rinsed three times in DPBS/PVP, and snap-frozen in liquid nitrogen. Animal tissue collection methods were approved by the University of Florida Animal Use and Care Committee. Oocyte RNA was isolated using the PicoPure RNA kit (Life Sciences catalog # KIT0204). Oocytes were harvested into extraction buffer provided by the kit, combined into groups of 6 (human) or 10 (cow), and processed into RNA following the manufacturer’s guidelines. Human brain RNA was generously

Lewis et al.

For RT-PCR, random primed cDNA was synthesized using SuperScript II reverse transcriptase (Life Technologies) and amplified by RT-PCR under normal conditions. Products were separated on a 1.4% (wt/vol) agarose gel, Southern blotted, and hybridized with either human or bovine SNRPN exon 2. All primers used in this study are listed in Table S1. RNA-Seq data were downloaded from the Gene Expression Omnibus (GEO) database in Short Read Archive (SRA) format. These SRA files were converted to fastq format using an SRA toolkit. Sequences were aligned to the human genome (hg19) using Tophat2 (40), and alignments were merged from biological replicates (n = 3) within each study. To generate the UCSC Genome Browser tracks, a bedGraph file was generated from aligned bam files using bedtools (41). Human oocyte data were obtained from GEO studies GSE36552 and GSE44183 (42, 43). The human brain RNA-Seq track is described in Wang et al. (44), available as a UCSC Genome Browser track.

1. Kishino T, Lalande M, Wagstaff J (1997) UBE3A/E6-AP mutations cause Angelman syndrome. Nat Genet 15(1):70–73. 2. Matsuura T, et al. (1997) De novo truncating mutations in E6-AP ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nat Genet 15(1):74–77. 3. Cassidy SB, Schwartz S, Miller JL, Driscoll DJ (2012) Prader-Willi syndrome. Genet Med 14(1):10–26. 4. Schaaf CP, et al. (2013) Truncating mutations of MAGEL2 cause Prader-Willi phenotypes and autism. Nat Genet 45(11):1405–1408. 5. Horsthemke B, Wagstaff J (2008) Mechanisms of imprinting of the Prader-Willi/ Angelman region. Am J Med Genet A 146A(16):2041–2052. 6. Ohta T, et al. (1999) Imprinting-mutation mechanisms in Prader-Willi syndrome. Am J Hum Genet 64(2):397–413. 7. Buiting K, Lich C, Cottrell S, Barnicoat A, Horsthemke B (1999) A 5-kb imprinting center deletion in a family with Angelman syndrome reduces the shortest region of deletion overlap to 880 bp. Hum Genet 105(6):665–666. 8. Meng L, et al. (2013) Truncation of Ube3a-ATS unsilences paternal Ube3a and ameliorates behavioral defects in the Angelman syndrome mouse model. PLoS Genet 9(12):e1004039. 9. Brannan CI, Bartolomei MS (1999) Mechanisms of genomic imprinting. Curr Opin Genet Dev 9(2):164–170. 10. Bressler J, et al. (2001) The SNRPN promoter is not required for genomic imprinting of the Prader-Willi/Angelman domain in mice. Nat Genet 28(3):232–240. 11. Chamberlain SJ, Brannan CI (2001) The Prader-Willi syndrome imprinting center activates the paternally expressed murine Ube3a antisense transcript but represses paternal Ube3a. Genomics 73(3):316–322. 12. DuBose AJ, Smith EY, Johnstone KA, Resnick JL (2012) Temporal and developmental requirements for the Prader-Willi imprinting center. Proc Natl Acad Sci USA 109(9): 3446–3450. 13. Johnstone KA, et al. (2006) A human imprinting centre demonstrates conserved acquisition but diverged maintenance of imprinting in a mouse model for Angelman syndrome imprinting defects. Hum Mol Genet 15(3):393–404. 14. Yang T, et al. (1998) A mouse model for Prader-Willi syndrome imprinting-centre mutations. Nat Genet 19(1):25–31. 15. Dittrich B, et al. (1996) Imprint switching on human chromosome 15 may involve alternative transcripts of the SNRPN gene. Nat Genet 14(2):163–170. 16. Färber C, Dittrich B, Buiting K, Horsthemke B (1999) The chromosome 15 imprinting centre (IC) region has undergone multiple duplication events and contains an upstream exon of SNRPN that is deleted in all Angelman syndrome patients with an IC microdeletion. Hum Mol Genet 8(2):337–343. 17. Wawrzik M, Spiess AN, Herrmann R, Buiting K, Horsthemke B (2009) Expression of SNURF-SNRPN upstream transcripts and epigenetic regulatory genes during human spermatogenesis. Eur J Hum Genet 17(11):1463–1470. 18. Landers M, et al. (2004) Regulation of the large (approximately 1000 kb) imprinted murine Ube3a antisense transcript by alternative exons upstream of Snurf/Snrpn. Nucleic Acids Res 32(11):3480–3492. 19. Mapendano CK, et al. (2006) Expression of the Snurf-Snrpn IC transcript in the oocyte and its putative role in the imprinting establishment of the mouse 7C imprinting domain. J Hum Genet 51(3):236–243. 20. Smith EY, Futtner CR, Chamberlain SJ, Johnstone KA, Resnick JL (2011) Transcription is required to establish maternal imprinting at the Prader-Willi syndrome and Angelman syndrome locus. PLoS Genet 7(12):e1002422. 21. Hanna CW, Kelsey G (2014) The specification of imprints in mammals. Heredity 113(2): 176–183. 22. Smallwood SA, et al. (2011) Dynamic CpG island methylation landscape in oocytes and preimplantation embryos. Nat Genet 43(8):811–814.

23. Pruitt KD, Tatusova T, Maglott DR (2005) NCBI Reference Sequence (RefSeq): A curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res 33(database issue):D501–D504. 24. Lucifero D, et al. (2006) Bovine SNRPN methylation imprint in oocytes and day 17 in vitro-produced and somatic cell nuclear transfer embryos. Biol Reprod 75(4):531–538. 25. O’Doherty AM, O’Shea LC, Fair T (2012) Bovine DNA methylation imprints are established in an oocyte size-specific manner, which are coordinated with the expression of the DNMT3 family proteins. Biol Reprod 86(3):67. 26. Burge C, Karlin S (1997) Prediction of complete gene structures in human genomic DNA. J Mol Biol 268(1):78–94. 27. Messerschmidt DM, Knowles BB, Solter D (2014) DNA methylation dynamics during epigenetic reprogramming in the germline and preimplantation embryos. Genes Dev 28(8):812–828. 28. Hiura H, et al. (2007) DNA methylation imprints on the IG-DMR of the Dlk1-Gtl2 domain in mouse male germline. FEBS Lett 581(7):1255–1260. 29. Lucifero D, Mann MR, Bartolomei MS, Trasler JM (2004) Gene-specific timing and epigenetic memory in oocyte imprinting. Hum Mol Genet 13(8):839–849. 30. Chotalia M, et al. (2009) Transcription is required for establishment of germline methylation marks at imprinted genes. Genes Dev 23(1):105–117. 31. Landers M, et al. (2005) Maternal disruption of Ube3a leads to increased expression of Ube3a-ATS in trans. Nucleic Acids Res 33(13):3976–3984. 32. Le Meur E, et al. (2005) Dynamic developmental regulation of the large non-coding RNA associated with the mouse 7C imprinted chromosomal region. Dev Biol 286(2): 587–600. 33. Wu MY, et al. (2006) Mouse imprinting defect mutations that model Angelman syndrome. Genesis 44(1):12–22. 34. Zogel C, et al. (2006) Identification of cis- and trans-acting factors possibly modifying the risk of epimutations on chromosome 15. Eur J Hum Genet 14(6):752–758. 35. Buiting K, et al. (1998) Sporadic imprinting defects in Prader-Willi syndrome and Angelman syndrome: Implications for imprint-switch models, genetic counseling, and prenatal diagnosis. Am J Hum Genet 63(1):170–180. 36. Nazlican H, et al. (2004) Somatic mosaicism in patients with Angelman syndrome and an imprinting defect. Hum Mol Genet 13(21):2547–2555. 37. Kelsey G, Feil R (2013) New insights into establishment and maintenance of DNA methylation imprints in mammals. Philos Trans R Soc Lond B Biol Sci 368(1609): 20110336. 38. You J, et al. (2012) Treatment with the proteasome inhibitor MG132 during the end of oocyte maturation improves oocyte competence for development after fertilization in cattle. PLoS ONE 7(11):e48613. 39. Fields SD, Hansen PJ, Ealy AD (2011) Fibroblast growth factor requirements for in vitro development of bovine embryos. Theriogenology 75(8):1466–1475. 40. Kim D, et al. (2013) TopHat2: Accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol 14(4):R36. 41. Quinlan AR, Hall IM (2010) BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26(6):841–842. 42. Xue Z, et al. (2013) Genetic programs in human and mouse early embryos revealed by single-cell RNA sequencing. Nature 500(7464):593–597. 43. Yan L, et al. (2013) Single-cell RNA-Seq profiling of human preimplantation embryos and embryonic stem cells. Nat Struct Mol Biol 20(9):1131–1139. 44. Wang ET, et al. (2008) Alternative isoform regulation in human tissue transcriptomes. Nature 456(7221):470–476. 45. Gray TA, Saitoh S, Nicholls RD (1999) An imprinted, mammalian bicistronic transcript encodes two independent proteins. Proc Natl Acad Sci USA 96(10):5616–5621.

Lewis et al.

COLLOQUIUM PAPER

provided by Stormy Chamberlain at the University of Connecticut. Bovine hypothalamus RNA was purified using the RNA-Bee reagent (Tel-Test, Inc.) following the manufacturer’s instructions. 5′ RACE was performed using a RACE kit, version 2.0, as recommended by the manufacturer (Life Technologies). Brain and oocyte RNA was reverse transcribed into cDNA using a reverse primer in SNRPN exon 3 followed by purification and tailing with terminal deoxynucleotidyl transferase. The cDNA was amplified using a reverse primer in SNRPN exon 2 and a poly(C) tail-specific primer provided by the kit. Human transcripts could be further amplified using anchored tail primers provided by the manufacturer and the exon 2 reverse primer. Bovine transcripts could be further amplified using anchored tail primers provided by the manufacturer and a nested reverse primer exon 2-R2 or reverse primers located in regions upstream of SNRPN as mentioned. PCR amplifications were performed under the following conditions: 1 min at 94°, followed by 35 cycles of 30 sec at 94°, 30 sec at 55°, and 1 min at 72°. All RACE products were cloned using the TOPO TA kit (Life Technologies) and subjected to Sanger sequencing.

PNAS Early Edition | 5 of 5

GENETICS

ACKNOWLEDGMENTS. We thank Drs. Bernhard Horsthemke, Edward Braun, and Emily Smith for helpful advice. This work was supported by Award NS068456 from the National Institute of Neurological Disorders and Stroke.