Promoter-specific expression and imprint status of marsupial IGF2

The admixed brushtail possum genome reveals invasion history in New Zealand and novel imprinted genes

Combining genome assembly with population and functional genomics can provide valuable insights to development and evolution, as well as tools for species management. Here, we present a chromosome-level genome assembly of the common brushtail possum (Trichosurus vulpecula), a model marsupial threatened in parts of their native range in Australia, but also a major introduced pest in New Zealand. Functional genomics reveals post-natal activation of chemosensory and metabolic genes, reflecting unique adaptations to altricial birth and delayed weaning, a hallmark of marsupial development. Nuclear and mitochondrial analyses trace New Zealand possums to distinct Australian subspecies, which have subsequently hybridised. This admixture allowed phasing of parental alleles genome-wide, ultimately revealing at least four genes with imprinted, parent-specific expression not yet detected in other species (MLH1, EPM2AIP1, UBP1 and GPX7). We find that reprogramming of possum germline imprints, and the wider epigenome, is similar to eutherian mammals except onset occurs after birth. Together, this work is useful for genetic-based control and conservation of possums, and contributes to understanding of the evolution of novel mammalian epigenetic traits.

Placentation in Marsupials

It is sometimes implied that marsupials are "aplacental," on the presumption that the only mammals that have a placenta are the eponymous "placental" mammals. This misconception has persisted despite the interest in and descriptions of the marsupial placenta, even in Amoroso's definitive chapter. It was also said that marsupials had no maternal recognition of pregnancy and no placental hormone production. In addition, it was thought that genomic imprinting could not exist in marsupials because pregnancy was so short. We now know that none of these ideas have held true with extensive studies over the last four decades definitively showing that they are indeed mammals with a fully functional placenta, and with their own specializations.

The role of mammalian foetal membranes in early embryogenesis: Lessons from marsupials

Across mammals, early embryonic development is supported by uterine secretions taken up through the yolk sac and other foetal membranes (histotrophic nutrition). The marsupial conceptus is enclosed in a shell coat for the first two-thirds of gestation and nutrients pass to the embryo through the shell and the avascular bilaminar yolk sac. At around the time of shell rupture, part of the yolk sac is trilaminar and supplied with blood vessels. It attaches to the uterus and forms a choriovitelline placenta. Rapid growth of the embryo ensues, still supported by histotrophe as well as exchange of oxygen and nutrients between maternal and foetal blood vessels (haemotrophic nutrition). Few marsupials have a chorioallantoic placenta and the highly altricial newborn is delivered after a short gestation. Eutherian embryos pass through a similar sequence before there is a fully functional chorioallantoic placenta. In most orders, there is transient yolk sac placentation, but even before this, nutrients are transferred through an avascular yolk sac. Yolk sac placentation does not occur in rodents or catarrhine primates. Early embryonic development in the mouse is nonetheless dependent on histotrophic nutrition. In the first trimester of human pregnancy, uterine glands open to the intervillous space and secretion products are taken up by the trophoblast. Transfer of nutrients to the early human embryo also involves the yolk sac, which floats free in the exocoelom. Marsupials can therefore inform us about the role of foetal membranes and histotrophic nutrition in early embryogenesis, knowledge that can translate to eutherians.

Identification of a novel antisense noncoding RNA, ALID, transcribed from the putative imprinting control region of marsupial IGF2R

Background

Genomic imprinting leads to maternal expression of IGF2R in both mouse and opossum. In mouse, the antisense long noncoding (lnc) RNA Airn, which is paternally expressed from the differentially methylated region (DMR) in the second intron of Igf2r, is required to silence the paternal Igf2r. In opossum, however, intriguingly, the DMR was reported to be in a different downstream intron (intron 11) and there was no antisense lncRNA detected in previous analyses. Therefore, clarifying the imprinting mechanism of marsupial IGF2R is of great relevance for understanding the origin and evolution of genomic imprinting in the IGF2R locus. Thus, the antisense lncRNA associated with the marsupial DMR can be considered as the ‘missing link’. In this study, we identified a novel antisense lncRNA, ALID, after detailed analysis of the IGF2R locus in an Australian marsupial, the tammar wallaby, Macropus eugenii, and compared it to that of the grey short-tailed opossum, Monodelphis domestica.

Results

Tammar IGF2R showed maternal expression and had a maternally methylated CpG island (CGI) in intron 12 as well as a promoter CGI without differential methylation, but none in the second intron. Re-analysis of the IGF2R of opossum detected the CGI in intron 12, not intron 11, as previously reported, confirming that the DMR in intron 12 is conserved between these marsupials and so is the putative imprinting control region of marsupial IGF2R. ALID is paternally expressed from the middle of the DMR and is approximately 650 bp long with a single exon structure that is extremely short compared to Airn. Hence, the lncRNA transcriptional overlap of the IGF2R promoter, which is essential for the Igf2r silencing in the mouse, is likely absent in tammar. This suggests that fundamental differences in the lncRNA-based silencing mechanisms evolved in eutherian and marsupial IGF2R and may reflect the lack of differential methylation in the promoter CGI of marsupial IGF2R.

Conclusions

Our study thus provides the best candidate factor for establishing paternal silencing of marsupial IGF2R without transcriptional overlap, which is distinct from the Igf2r silencing mechanism of Airn, but which may be analogous to the mode of action for the flanking Slc22a2 and Slc22a3 gene silencing in the mouse placenta.

Electronic supplementary material

The online version of this article (10.1186/s13072-018-0227-8) contains supplementary material, which is available to authorized users.

Post-natal imprinting: evidence from marsupials

Genomic imprinting has been identified in therian (eutherian and marsupial) mammals but not in prototherian (monotreme) mammals. Imprinting has an important role in optimising pre-natal nutrition and growth, and most imprinted genes are expressed and imprinted in the placenta and developing fetus. In marsupials, however, the placental attachment is short-lived, and most growth and development occurs post-natally, supported by a changing milk composition tailor-made for each stage of development. Therefore there is a much greater demand on marsupial females during post-natal lactation than during pre-natal placentation, so there may be greater selection for genomic imprinting in the mammary gland than in the short-lived placenta. Recent studies in the tammar wallaby confirm the presence of genomic imprinting in nutrient-regulatory genes in the adult mammary gland. This suggests that imprinting may influence infant post-natal growth via the mammary gland as it does pre-natally via the placenta. Similarly, an increasing number of imprinted genes have been implicated in regulating feeding and nurturing behaviour in both the adult and the developing neonate/offspring in mice. Together these studies provide evidence that genomic imprinting is critical for regulating growth and subsequently the survival of offspring not only pre-natally but also post-natally.

Postnatal epigenetic reprogramming in the germline of a marsupial, the tammar wallaby

Background

Epigenetic reprogramming is essential to restore totipotency and to reset genomic imprints during mammalian germ cell development and gamete formation. The dynamic DNA methylation change at DMRs (differentially methylated regions) within imprinted domains and of retrotransposons is characteristic of this process. Both marsupials and eutherian mammals have genomic imprinting but these two subgroups have been evolving separately for up to 160 million years. Marsupials have a unique reproductive strategy and deliver tiny, altricial young that complete their development within their mother's pouch. Germ cell proliferation in the genital ridge continues after birth in the tammar wallaby (Macropus eugenii), and it is only after 25 days postpartum that female germ cells begin to enter meiosis and male germ cells begin to enter mitotic arrest. At least two marsupial imprinted loci (PEG10 and H19) also have DMRs. To investigate the evolution of epigenetic reprogramming in the marsupial germline, here we collected germ cells from male pouch young of the tammar wallaby and analysed the methylation status of PEG10 and H19 DMR, an LTR (long terminal repeat) and a non-LTR retrotransposons.

Results

Demethylation of the H19 DMR was almost completed by 14 days postpartum and de-novo methylation started from 34 days postpartum. These stages correspond to 14 days after the completion of primordial germ cell migration into genital ridge (demethylation) and 9 days after the first detection of mitotic arrest (re-methylation) in the male germ cells. Interestingly, the PEG10 DMR was already unmethylated at 7 days postpartum, suggesting that the timing of epigenetic reprogramming is not the same at all genomic loci. Retrotransposon methylation was not completely removed after the demethylation event in the germ cells, similar to the situation in the mouse.

Conclusions

Thus, despite the postnatal occurrence of epigenetic reprogramming and the persistence of genome-wide undermethylation for 20 days in the postnatal tammar, the relative timing and mechanism of germ cell reprogramming are conserved between marsupials and eutherians. We suggest that the basic mechanism of epigenetic reprogramming had already been established before the marsupial-eutherian split and has been faithfully maintained for at least 160 million years and may reflect the timing of the onset of mitotic arrest in the male germline.

Selected imprinting of INS in the marsupial

BACKGROUND

In marsupials, growth and development of the young occur postnatally, regulated by milk that changes in composition throughout the long lactation. To initiate lactation in mammals, there is an absolute requirement for insulin (INS), a gene known to be imprinted in the placenta. We therefore examined whether INS is imprinted in the mammary gland of the marsupial tammar wallaby (Macropus eugenii) and compared its expression with that of insulin-like growth factor 2 (IGF2).

RESULTS

INS was expressed in the mammary gland and significantly increased, while IGF2 decreased, during established milk production. Insulin and IGF2 were both detected in the mammary gland macrophage cells during early lactation and in the alveolar cells later in lactation. Surprisingly, INS, which was thought only to be imprinted in the therian yolk sac, was imprinted and paternally expressed in the liver of the developing young, monoallelically expressed in the tammar mammary gland and biallelic in the stomach and intestine. The INS transcription start site used in the liver and mammary gland was differentially methylated.

CONCLUSIONS

This is the first study to identify tissue-specific INS imprinting outside the yolk sac. These data suggest that there may be an advantage of selective monoallelic expression in the mammary gland and that this may influence the growth of the postnatal young. These results are not consistent with the parental conflict hypothesis, but instead provide support for the maternal-infant co-adaptation hypothesis. Thus, imprinting in the mammary gland maybe as critical for postnatal growth and development in mammals as genomic imprinting in the placenta is prenatally.

GRB10 imprinting is eutherian mammal specific

GRB10 is an imprinted gene differently expressed from two promoters in mouse and human. Mouse Grb10 is maternally expressed from the major promoter in most tissues and paternally expressed from the brain-specific promoter within specific regions of the fetal and adult central nervous system. Human GRB10 is biallelically expressed from the major promoter in most tissues except in the placental villus trophoblast where it is maternally expressed, whereas the brain-specific promoter is paternally expressed in the fetal brain. This study characterized the ortholog of GRB10 in a marsupial, the tammar wallaby (Macropus eugenii) to investigate the origin and evolution of imprinting at this locus. The protein coding exons and predicted amino acid sequence of tammar GRB10 were highly conserved with eutherian GRB10. The putative first exon, which is located in the orthologous region to the eutherian major promoter, was found in the tammar, but no exon was found in the downstream region corresponding to the eutherian brain-specific promoter, suggesting that marsupials only have a single promoter. Tammar GRB10 was widely expressed in various tissues including the brain but was not imprinted in any of the tissues examined. Thus, it is likely that GRB10 imprinting evolved in eutherians after the eutherian-marsupial divergence approximately 160 million years ago, subsequent to the acquisition of a brain-specific promoter, which resides within the imprinting control region in eutherians.