Introduction

Genomic imprinting, the expression of alleles of parental origin, occurs in therian mammals (marsupials and eutherians) but in no other vertebrates (Ferguson-Smith 2011; Renfree et al. 2009). This epigenetic process is essential for normal development in therians as disruption of imprinting causes many developmental abnormalities (Ferguson-Smith 2011; Lefebvre et al. 1998; Reik and Walter 2001; Renfree et al. 2009). However, how and why imprinting evolved in therian mammals is not clear. It is certain that genomic imprinting arose very early in mammals, but since there is no imprinting in monotremes, it must have evolved after the split between monotremes and therian mammals (Pask et al. 2009; Renfree et al. 2009). Therefore, to understand the ancestral condition and the evolution of imprinting in mammals, gene loci that are imprinted in both marsupials and eutherians need to be compared amongst all three mammalian groups, including monotremes.

There are more than 200 imprinted genes in eutherians (Santini et al. 2021; Wang et al. 2011; Wang and Clark 2014). In marsupials, 36 genes that are imprinted in eutherians have been examined so far (Stringer et al. 2014; Edwards et al. 2019; Ishihara et al. 2022a, 2022b). However, of these, only 9 genes have been confirmed to be imprinted in marsupials (Ager et al. 2007; Cao et al. 2023; Das et al. 2012; Ishihara et al. 2022a; Killian et al. 2000; O’Neill et al. 2000; Smits et al. 2008; Suzuki et al. 2007, 2005). In addition, there are eighteen marsupial-specific imprinted genes (Cao et al. 2023; Douglas et al. 2014; Grant et al. 2012; Mahadevaiah et al. 2020; Stringer et al. 2012; Suzuki et al. 2018). These common imprinted genes could provide insights into the evolutionary conserved and/or lineage specific imprinting mechanisms. Mesoderm specific transcript (MEST, also known as PEG1) is an important candidate for this comparison between marsupials and eutherians because it is critical for successful reproduction. MEST is thought to be a maternally imprinted (paternally expressed) gene in therian mammals (Das et al. 2012; Eggermann et al. 2012; Huntriss et al. 2013; Li et al. 2015; Mayer et al. 2000; Reule et al. 1998; Riesewijk et al. 1997; Suzuki et al. 2005). However, the marsupial MEST lacks differentially methylated CpG islands between the parental alleles (Das et al. 2012; Suzuki et al. 2005) whereas eutherian MEST has a distinct differentially methylated region (DMR), in which CpG islands are uniquely methylated between the parental alleles, at its promoter (Li et al. 2015; Riesewijk et al. 1997). This suggests that marsupial MEST imprinting may be regulated by another mechanism such as differential histone modification-based mechanisms. However, the previous studies (Das et al. 2012; Suzuki et al. 2005) did not confirm allelic expression of the MEST gene using combinations of mothers homozygous for a different SNP, so the gene does not appear to be imprinted in marsupials. Its imprinting status needs to be clarified.

Although its exact function is unknown, disruption of MEST imprinting causes developmental defects in placental and fetal growth and abnormal maternal behaviour in mice (Hiramuki et al. 2015; Lefebvre et al. 1998; Mayer et al. 2000). The MEST imprinted isoform also regulates mammary gland maturation in mice (Yonekura et al. 2019). Since the gene is associated with many classical mammal-specific traits, characterising the evolution of this gene locus across mammals could shed light on how imprinting evolved in the common ancestor of therian mammals.

There is a clear difference between the MEST locus in eutherians and marsupials: the presence or absence of differentially methylated CpG islands between the parental alleles. This suggests that the regulation of imprinting at the MEST locus may have evolved independently with very different mechanisms or that there may be common factors other than the DMR. Conserved imprinted loci between marsupials and eutherians have been found to associate with lncRNAs that may have been conserved or evolved independently (Smits et al. 2008; Suzuki et al. 2018). In the human MEST gene locus, there is an antisense lncRNA, MESTIT1 (PEG1-AS) (Li et al. 2002; Nakabayashi 2002). This antisense lncRNA is predominantly expressed in the testis and mature spermatozoa (Li et al. 2002). If this lncRNA plays an important part in the establishment of MEST imprinting, it is possible that marsupials also have a conserved antisense transcript. More detailed comparison of the MEST gene locus across mammals would elucidate conserved features as to how MEST became imprinted during mammalian evolution.

In this study, we examined similarities and differences in the MEST gene locus across mammals. First, we asked whether the shorter isoform, which is imprinted in eutherians, is present in all mammalian groups. Next, we examined whether the antisense lncRNA was present in the tammar wallaby. We also examined the gene structure and genomic features of MEST and its neighbouring gene, centrosomal protein 41 (CEP41) in all three mammalian groups. Here, we report that the shorter MEST isoform is conserved across mammals. We confirmed the presence of an antisense lncRNA from the MEST locus in tammar sperm. However, despite the presence of the lncRNA, we found that although MEST was mono-allelically expressed in the tammar placenta, it was not imprinted. In contrast, in eutherians, the evolution of MEST must have been accompanied by the acquisition of differentially methylated CpG islands between parental alleles and imprinting after the marsupial-eutherian split.

Material and methods

Animals

Tammar wallaby (Macropus eugenii) samples of Kangaroo Island, South Australia origin, were collected from either wild animals or captive animals from our colony maintained by the University of Melbourne. Adults were killed humanely and tissues were collected as previously described (Ishihara et al. 2022b; Renfree 1973a, 1973b; Stringer et al. 2012). Adult testes were snap-frozen immediately after dissection. Sperm was collected post-mortem from adult male wallabies after euthanasia. The cauda epididymis was cut into small pieces in sterile saline and incubated in 15 ml tubes for 1 h at 37 °C to allow the spermatozoa to swim up. The sperm pellets were collected by centrifugation of supernatant and immediately snap-frozen. Both the avascular bilaminar (BOM) and vascular trilaminar (TOM) regions of the choriovitelline placentas were collected from fetuses in the final third of gestation (day 19–25 days of the 26.5 day active pregnancy (n = 29)) from adult females post mortem as previously described (Ager et al. 2007; Renfree 1973b, 1973a; Stringer et al. 2012) and snap-frozen immediately after dissection. Endometrial tissues were collected from adult females and snap-frozen immediately after dissection. Postnatal liver samples were collected from six pouch youngs (PYs) aged approximately 35 days after birth. All tammar animal handling, husbandry and experimental sampling were in accordance with the National Health and Medical Research Council of Australia (2013) guidelines (National Health and Medical Research Council (Australia) 2013) and approved by the University of Melbourne Animal Experimentation Ethics committees. Platypus (Ornithorhynchus anatinus) tissues were collected from wild-caught animals under permits from NSW Parks and Wildlife and ethically approved by the University of Melbourne Animal Ethics committees. Testes of mouse (Mus musculus) were kindly provided by the Pask laboratory at the University of Melbourne, Melbourne, Australia as a secondary use from ethically approved projects.

RNA extraction and cDNA synthesis

Snap-frozen adult testes of each species and the tammar placenta tissues consisting of the separated avascular bilaminar omphalopleure (BOM) and the vascular trilaminar omphalopleure (TOM) were used for RNA extraction using the GenElute Mammalian total RNA Miniprep Kit (Sigma-Aldrich, Missouri, USA) following the manufacturer’s instructions. Snap-frozen tammar sperm pellets were used for RNA extraction using the RNeasy Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. The extracted RNA was treated with the DNA free DNase treatment and removal kit (Thermo Fisher Scientific, Massachusetts, USA) to remove residual genomic DNA. 200 ng of total RNA was used as a template for cDNA synthesis using SuperScript IV First strand Synthesis System (Invitrogen, Carlsbad, USA).

Transcriptome analysis

To identify the potential antisense lncRNA in adult tammar testes, tammar transcriptome data sets derived from adult testis were analysed. Publicly available tammar raw RNA-seq data sets (DRP001145) were downloaded from NCBI SRA (https://www.ncbi.nlm.nih.gov/sra). All RNA-seq reads were trimmed using TrimGalore! (v0.6.5) (https://github.com/FelixKrueger/TrimGalore) with default settings to eliminate adaptor sequences, poor quality reads and very short (<20 bp) reads. The trimmed reads were aligned to the wallaby genome.v3 (https://wallabase.science.unimelb.edu.au) using HISAT2 (v2.1.0) (Kim et al. 2019) with a parameter –rna-strandness FR to reflect the strandedness of sequenced RNA. The mapped reads were assigned to each strand by Samtools (v1.9) (Li et al. 2009). The output file was visualised on Integrative genome viewer (IGV) (Robinson et al. 2011; Thorvaldsdóttir et al. 2013).

5′ and 3′ Rapid amplification of cDNA ends (RACE)

To determine the transcription start site (TSS) of MEST in the tammar and platypus, RACE (rapid amplification of cDNA ends) experiments were performed using the SMARTer RACE 5′/3′ kit (Clontech, California, USA). The TSS of CEP41, a neighbouring gene of MEST, was also confirmed by 5′ RACE reactions in the tammar, platypus and mouse. The transcription start position of a putative antisense novel lncRNA was also confirmed by 5′ RACE reaction in the tammar and mouse. The first round 5′ RACE reaction was performed with adult testis cDNA from each species using SeqAmp DNA Polymerase (Clontech, California, USA) with gene specific primers (Supplementary Table 1). Up to 1 µg of adult testis RNA was used to synthesise the testis cDNA for RACE reactions with the SMARTer RACE 5′/3′ kit. The nested 5′ RACE was performed by GoTaq DNA polymerase (Promega, Wisconsin, USA) and the RACE products were cloned into pGEM-T Easy Vector (Promega, Wisconsin, USA) and JM109 competent cells or Stellar Competent Cells (Clontech, California, USA). Plasmids were extracted using Wizard Plus SV Minipreps DNA Purification System (Promega, Wisconsin, USA) and sequenced by the Sanger sequencing method with M13 primers. To complete the full sequence of a putative antisense novel lncRNA in the tammar, 5′ and 3′ RACE experiments were performed. The first round RACE reaction was performed with adult testis cDNA or placenta tissue (BOM) cDNA using SeqAmp DNA Polymerase (Clontech, California, USA) with gene specific primers (Supplementary Table 1). The nested RACE was performed by GoTaq DNA polymerase (Promega, Wisconsin, USA) and the RACE products were cloned into a pGEM-T Easy Vector (Promega, Wisconsin, USA) and JM109 competent cells. Plasmids were extracted using Wizard Plus SV Minipreps DNA Purification System (Promega, Wisconsin, USA) and sequenced by the Sanger sequencing method with M13 primers (Supplementary Table 1).

Identification and expression analysis of the antisense lncRNA MESTIT1 in tammar sperm

Tammar MESTIT1 specific primers were designed based on the sequences determined by RACE reactions using Primer3Plus (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi) (Supplementary Table 1). PCR reactions were performed using adult tammar testis cDNA with GoTaq Green master mix (Promega Corporation, USA) to confirm PCR amplification with the MESTIT1 specific primers and sequencing. To determine whether the lncRNA was present in the tammar sperm, a PCR reaction was performed on both sperm cDNA and DNase treated RNA (n = 3).

Comparative analysis of mammalian MEST and CEP41 gene locus

DNA sequences of human MESTA, human MESTB, mouse Mesta, mouse Mestb, mouse Cep41, platypus putative MEST and platypus putative CEP41 were obtained from NCBI (https://www.ncbi.nlm.nih.gov). Amino acid sequences retrieved from DDBJ/EMBL/GenBank/RefSeq database were used for generating alignments using CLC sequence viewer 8 (https://clc-sequence-viewer.software.informer.com). Accession numbers: Homo sapiens MESTA, NM_177525.2; H. sapiens MESTB, NM_177524.2; M. musculus Mesta, NM_008590.2; M. musculus Mestb, NM_001252293.1; M. musculus Cep41 : NM_031998.3; O. anatinus putative MEST : XM_001511283.6; O. anatinus putative CEP41 : XM_001511256.6. Genomic sequences of the MEST-CEP41 flanking region for each mammalian species (human, mouse, African savanna elephant (Loxodonta africana), tammar and platypus) were obtained from either NCBI (https://www.ncbi.nlm.nih.gov) or Wallabase (https://wallabase.science.unimelb.edu.au/). CpG island predictions were performed by MethPrimer (http://www.urogene.org/cgi-bin/methprimer/methprimer.cgi) (Li and Dahiya 2002). Graphs of CpG dinucleotide ratio were made by GENETYX software (https://www.genetyx.co.jp). Genomic elements within the intergenic region between mammalian MEST and CEP41 genes were investigated by BLAST searches (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Genomic DNA extraction

Snap-frozen BOM, TOM, and endometrial tissues were used for genomic DNA (gDNA) extraction. BOM and TOM were the fetal gDNA source and endometrium was the maternal gDNA source. DNA extraction was performed with Wizard Genomic DNA purification kit (Promega, Wisconsin, USA) following the manufacturer’s instructions.

Allelic expression analysis

Extracted genomic DNAs were used as a template for PCR amplification to find a combination of maternal homozygous and fetal heterozygous single nucleotide polymorphisms (SNPs). PCR reactions were performed using gene specific primers (Supplementary Table S1) with Ex-Taq polymerase (Takara, Shiga, Japan) or Go-Taq polymerase (Promega, Wisconsin, USA) under the following cycle conditions: 95 °C 30 s, 65 °C 30 s, and 72 °C 45 s. To analyse sequences of the transcript, cDNA of placenta tissues (BOM and TOM) was used as a template for PCR amplification under the following cycle conditions: 95 °C 30 s, 65 °C 30 s, and 72 °C 45 s or 3 min. After performing gel electrophoresis, confirmed PCR products from gDNA and cDNA were extracted and directly sequenced by Sanger sequencing to confirm SNP sites and allele specific expression.

Results

Re-evaluating the transcription start site of tammar MEST

Although MEST imprinting has been characterised in both the South American grey short-tailed opossum (Monodelphis domestica) (Das et al. 2012) and the tammar (Suzuki et al. 2005), the transcription start sites (TSSs) of the MEST isoforms are not yet well defined in marsupials. To characterise the tammar MEST gene locus, the putative tammar MEST gene was searched using the wallaby genome database (Wallabase: https://wallabase.science.unimelb.edu.au/) and compared with mouse Mest protein (Accession number: NP_032616.1). 2859 bp of putative tammar MEST was identified (Fig. 1A). Since the MEST gene has isoform dependent imprinted expression in eutherians (Kosaki et al. 2000; Li et al. 2015; Reule et al. 1998; Riesewijk et al. 1997; Yonekura et al. 2019), we examined isoforms of MEST by 5′RACE experiments using the adult tammar testis. In eutherians, the imprinted MEST isoform, MESTA, shares the protein coding sequence with the non-imprinted longer isoform, MESTB (Fig. 1A: green coloured boxes), 5′ RACE reaction was performed using primers designed for the protein coding sequences of the putative tammar MEST (Fig. 1A: green coloured boxes). After sequencing the 5′ RACE products, we confirmed that there are three isoforms of MEST (Fig. 1B, C). The longer isoform of MEST was expressed from more upstream CpG islands that was distinct from the other two isoforms (Fig. 1C). We renamed the shortest isoform of MEST as MESTA (DDBJ accession number: LC747011) and the longer isoform of MEST as MESTB (DDBJ accession number: LC747012) in accordance with eutherian MEST isoforms. The intermediate size isoform was renamed MESTC. Although each isoform had a different TSS, the translation start sites were common to each other.

Fig. 1: Identification of MEST isoforms in the tammar wallaby.
figure 1

A Exon structures of eutherian MESTs and putative tammar MEST. In both human and mouse, the protein coding region was common between the two isoforms, MESTA and MESTB. The highly conserved protein coding regions are represented by green coloured boxes. White and grey coloured boxes represent protein-coding exon and UTRs, respectively. B Identification of the TSS of tammar MEST isoforms. 5′RACE and nested RACE primers are represented by the black arrows. Asterisks represent RACE products encoding partial tammar MEST sequences. Boxes represent exons identified by sequencing of the RACE products. Black boxes represent CpG islands determined by the Methprimer programme.

Full size image

Identification of two isoforms of MEST in monotremes

To ask whether the short isoform of MEST is therian mammal specific, the presence of MEST isoforms in monotremes was examined by 5′RACE experiments using adult platypus testis. To characterise the monotreme MEST gene locus, the platypus orthologue of MEST was searched by NCBI Blast (https://blast.ncbi.nlm.nih.gov/Blast.cgi) with the tammar MEST amino acid sequences, and 2374 bp of putative platypus MEST (Accession number: XM_001511283.6) was identified (Fig. 2A). Next, 5′ RACE reactions were performed using primers designed for the protein coding region of the putative platypus MEST (Fig. 2A, B: green coloured boxes). After sequencing the RACE products, we confirmed that platypus has two isoforms of MEST (Fig. 2B, C). Their TSSs were located at different CpG islands (Fig. 2C). We renamed the shortest isoform of MEST as MESTA (DDBJ accession number: LC747014) and the longer isoform of MEST as MESTB (DDBJ accession number: LC747015) in accordance with eutherian and marsupial MEST isoforms. Although each isoform had a different TSS, their translation start sites were common.

Fig. 2: Identification of MEST isoforms in platypus.
figure 2

A Exon structures of eutherian MESTs, the tammar MESTs and putative platypus MEST. In human, mouse and the tammar wallaby, the protein coding region was common between the two isoforms, MESTA and MESTB. The highly conserved protein coding regions are represented by green coloured boxes. White and grey coloured boxes represent protein-coding exon and UTRs, respectively. B Identification of the TSS of the platypus MEST isoforms. 5′RACE and nested RACE primers are represented by the black arrows. Asterisks represent RACE products encoding partial platypus MEST sequences. Boxes represent exons identified by sequencing of the RACE products. Black boxes represent CpG islands determined by the Methprimer programme.

Full size image

Identification of an orthologue of MESTIT1 in the tammar wallaby

Since the human lncRNA MESTIT1 is expressed from the DMR at the promoter of human MESTA (Li et al. 2002; Nakabayashi 2002), we asked whether similar antisense transcripts are present around the TSS of the tammar MESTA by analysing stranded transcriptome data sets. In adult tammar testis transcriptome data, antisense mapping reads were present in the CpG island near the TSS of the tammar MESTA (Fig. 3A). After isolating a partial transcript by PCR with an antisense transcript candidate-specific primer, 5′ and 3′RACE experiments were performed to confirm the full length of the transcript (Fig. 3B). While the 5′RACE reaction yielded one band, the 3′RACE reaction resulted in three different bands. Of these, two 3′RACE products were non-coding transcripts with an alternative poly A signal and distinct poly A tail (DDBJ accession numbers; MESTIT1 isoform1: LC746974; MESTIT1 isoform 2: LC746975) (Fig. 3B: black asterisks). Surprisingly, the largest 3′RACE product encoded a transcript spliced into the neighbouring gene of MEST, CEP41 (later renamed CEP41B, DDBJ accession number: LC747013) (Fig. 3B: red asterisks). The tammar lncRNA had 39.5% sequence identity with the human MESTIT1. However, the neighbouring genes of the putative MESTIT1 orthologue in the tammar showed synteny to corresponding genes in human genome, suggesting that the MESTIT1 gene is conserved between the tammar and the human genomes.

Fig. 3: Identification of the marsupial lncRNA in adult testis and mature sperm.
figure 3

A Identification of an antisense transcript at the TSS of the tammar MESTA with transcriptome data. Tammar testis transcriptome data was visualised with IGV in a strand specific manner. There was an antisense transcript candidate at the TSS of MESTA in the tammar testis. B Characterisation of the full-length sequence of the antisense lncRNA by RACE reactions. 5′ and 3′ RACE and nested RACE were performed using adult testis with primers for the putative antisense transcript. 5′ RACE generated a single amplicon whereas 3′RACE reactions resulted in three different products. The two 3′RACE products (black asterisks) encoded a non-coding transcript with a polyadenylation signal (red letters) and poly-A tail (green letters). The largest RACE product encoded an isoform of CEP41 with a protein coding sequence (909 bp). C RT-PCR analysis of the marsupial lncRNA in sperm. Sperm samples were collected from three adult males, #1, #2, and #3, and total RNAs were purified from them. cDNAs were synthesised with (RT+) or without reverse transcription (RT−) using the oligo(dT) primer, and PCR was performed with primer pairs for the tammar lncRNA. Presence of the transcript in sperm was confirmed in all biological replicates. M Molecular marker, P Positive control (adult testis), WT Water.

Full size image

To ask whether the lncRNA MESTIT1 is present in the tammar sperm as seen in human, RT-PCR analysis was performed. PCR amplification was observed in all sperm samples only after reverse-transcription (Fig. 3C).

Tammar has two isoforms of CEP41

During the RACE reaction to identify MESTIT1 in tammar, an isoform of CEP41 was identified. In the wallaby genome database (Wallabase: https://wallabase.science.unimelb.edu.au), there was a putative CEP41, but its exon structure was not the same as the CEP41B identified in this study (Fig. 4A). The putative CEP41 and CEP41B shared several exons. To examine the TSS of CEP41 in the tammar, 5′RACE reactions were performed using adult testis cDNA with primers designed for the common exons of CEP41B and the putative CEP41 (Fig. 4B: green coloured boxes). After sequencing the 5′RACE products, two TSSs were identified for the tammar wallaby CEP41. CEP41B shared a CpG island with marsupial MESTA. The other isoform was found to share a CpG island with MESTB. Because of the different exon structures, the newly identified CEP41 isoform was named CEP41A. Furthermore, we confirmed that the possible protein encoded by CEP41A differs in its C-terminal region from the amino acid sequence encoded by CEP41B (Fig. 4C).

Fig. 4: Identification of two isoforms of CEP41 in the tammar wallaby.
figure 4

A Exon structures of CEP41B, which was identified by our 3′RACE experiment, and a putative CEP41 gene obtained from the wallaby genome database (Wallabase). These two isoforms shared some exons but had different gene structures. B Identification of TSS of the tammar CEP41 isoforms. 5′RACE and nested RACE primers are represented by the black arrows. Asterisks represent race products coding partial tammar CEP41 sequences. The longest isoform had the same sequence with the CEP41B. The smallest isoform encoded a different transcript with two additional exons. Since the two transcripts had a different exon structure from each other, we renamed the newly identified transcript as CEP41A. C Protein alignment of the two CEP41 isoforms, CEP41A and CEP41B. The putative CEP41A protein had additional amino acids in its C-terminal region.

Full size image

CEP41 B isoform is not present in either mouse or platypus

To ask whether the CEP41 isoform, CEP41B, is a marsupial specific isoform, the presence of CEP41 isoforms in mouse and platypus was examined by 5′RACE experiments using adult testes (Supplementary Fig. 1). First, the sequence of the mouse CEP41 gene was obtained from NCBI (Accession number: NM_031998.3). Since CEP41B shared several protein-coding exons with CEP41 A in the tammar, 5′ RACE reactions were performed using primers against the conserved region of mammalian CEP41 (Supplementary Fig. 1A: green coloured boxes). A single transcript was isolated from adult mouse testis. This transcript was identical to the known mouse CEP41 and contained the two upstream exons from the common exons, which is similar to the tammar CEP41A (Supplementary Fig. 1A). Similar experiments were performed with adult platypus testes after obtaining a putative platypus CEP41 gene from NCBI (Accession number: XM_001511256.6). A single transcript was isolated from an adult platypus testis. The transcripts also contained the two upstream exons from the common exons, which is similar to the tammar CEP41A (Supplementary Fig. 1B).

Genomic analysis of the MEST and CEP41 flanking regions

To investigate CpG island locations and CEP41 and MEST isoforms in mammals, the genomic structures of the MEST and CEP41 flanking region in human, tammar and platypus were compared with each other. In both tammar and platypus, two large domains of CpG islands in close proximity (Fig. 5A: yellow highlighted regions). The two major MEST isoforms, MESTA and MESTB, were expressed from the two large CpG island domains, respectively (Fig. 5A). However, there is expansion of the intergenic region between CEP41 and MEST in the human genome, and the human MESTB did not share the same CpG island with human CEP41 (Fig. 5A). NCBI BLAST searches of the human MEST and CEP41 flanking region identified a LINE1 ORF1 in the vicinity of MEST. Similar LINE1 elements were also found in the mouse and elephant genomes by NCBI BLAST searches (Fig. 5B). However, the elephant LINE1 element was located close to CEP41 (Fig. 5B).

Fig. 5: The evolution of MEST-CEP41 flanking region in mammals.
figure 5

A Comparison of the genomic structures between MEST and CEP41. The blue graphs show CpG contents in the genomic sequences and the yellow highlight represent CpG islands predicted by the Methprimer programme. B The timing of the LINE1 element emergence during mammalian evolution. Red boxes represent the location of the LINE1 element. In eutherians such as elephant, human and mouse, MEST and CEP41 are physically separated by the LINE1 element. However, the location of the LINE1 element in mouse and human is completely different from that of the elephant. In both marsupials and monotremes, CEP41 and MEST are not physically separated.

Full size image

Neighbouring transcripts of MESTA are not imprinted in the tammar placenta tissues

Since human MESTIT1 is imprinted (Nakabayashi 2002), it was possible that the tammar MESTIT1 is also imprinted. To confirm whether MESTIT1 is imprinted in the tammar wallaby, allelic expression of the gene was performed using tammar placenta tissues. First, SNP sites were examined by direct sequencing of genomic DNA with RT-PCR (Fig. 6A: black arrows). After examining 29 samples, one SNP site was identified in the common region of the two isoforms (Fig. 6A). In the shorter isoform, specific primers could not be designed because the SNP site was too close to the poly-A tail. However, for the longer isoform, we could detect it from cDNA using the same primers as used for the SNP search (Fig. 6A). Using these primers, the imprinting status of the tammar lncRNA was determined by direct sequencing of the PCR products that contained the SNP site (Fig. 6A). 29 samples were examined, and three samples were a combination of maternal homozygous and fetal heterozygous SNPs (Fig. 6B). In contrast to the gDNA PCR data in which the two peaks at the SNP site have almost the same signal strength, the cDNA PCR products showed that the lncRNA is preferentially expressed from either one of the two alleles (Fig. 6B). Individuals #1 and #2 showed paternally skewed expression, but individual #3 showed maternally skewed expression (Fig. 6B). Therefore, in the tammar placenta, the lncRNA was not imprinted.

Fig. 6: Allelic expression analysis of the tammar MESTIT1 and CEP41B.
figure 6

A Isoforms of MESTIT1 and primers for allelic expression analysis. A SNP site was found in the common region of the two isoforms (arrowhead). In the shorter isoform, specific primers could not be designed because the SNP site was close to the poly-A tail. However, primers used for the SNP search could detect the longer isoform. B Allelic expression of the longer isoform of MESTIT1 in the tammar placentas. The imprinting status of the tammar MESTIT1 was determined by direct sequencing of PCR products that contained a SNP site. In contrast to the genomic DNA data in which double peaks at the SNP site have almost the same signal strength, the cDNA data clearly showed that the lncRNA is predominantly expressed from either one of the two alleles. C Exon structures of the two CEP41 isoforms. Orange and aqua coloured boxes represent the CEP41B-specific and CEP41A-specific exons, respectively. Blue arrows represent primers used for detecting the genomic SNP site. D Primer design for the allelic expression analysis of the CEP41B transcript. Allelic expression analysis of CEP41B was performed using CEP41B specific forward primer and a reverse primer designed for the 3′ UTR common amongst CEP41 isoforms. Arrowheads represent SNP sites. E Allelic expression of the tammar CEP41B by direct sequencing with PCR amplification. In animal #1, skewed maternal expression was observed in the placenta tissues. However, in animal #2, skewed paternal expression was observed in the TOM tissue. Due to low expression levels, it was not possible to determine allelic expression of CEP41B in the BOM of Animal #2.

Full size image

Although CEP41 is not imprinted in mouse (Yamada et al. 2002), the tammar CEP41B shares the same CpG island with the shorter MEST isoform, MESTA. In eutherians, MEST imprinting is under the control of a CpG island at the TSS of MESTA isoform (Li et al. 2015; Riesewijk et al. 1997). It was, therefore, possible that marsupial CEP41B is imprinted with MESTA. To confirm whether CEP41B is imprinted in the tammar wallaby, allelic expression of the gene was examined in the tammar placenta tissues. Since each isoform has a unique TSS and unique exons, allelic expression analysis was performed using CEP41B specific primers (Fig. 6C, D). First, SNP sites in 3′UTR were examined by direct sequencing of PCR products generated from genomic DNA (Fig. 6C: blue arrows). After examining 18 samples, two SNP sites were identified in the 3′UTR of the tammar CEP41B (Fig. 6D: Arrowheads). Allelic expression was performed on placenta cDNA samples using CEP41B specific primers (Fig. 6D). Of these 18 samples, two animals were a combination of maternal homozygous and fetal heterozygous. Fortunately, the maternal homozygous SNPs in these two samples were different from each other (Fig. 6E). In animal #1, CEP41B was preferentially expressed from the maternal allele (C) in the BOM and TOM tissues. However, in animal #2, CEP41B was preferentially expressed from the paternal (C) allele in the TOM tissues (Fig. 6E). CEP41B was not detectable in the BOM of animal #2 so we could not determine its allelic expression (Fig. 6E). Therefore, in the tammar placenta, CEP41B was not imprinted.

Re-evaluating allelic expression of MEST in the tammar placenta and PY tissues

Since CEP41B and MESTIT1 were not imprinted in the tammar placenta even though they shared the same CpG island with the MEST gene, we re-evaluated MEST imprinting in the tammar placenta. It is impossible to characterise MESTC allelic expression as there is no MESTC specific exon. However, since the two major isoforms, MESTA and MESTB, identified in this study have unique TSSs and unique exons (Fig. 7A), allelic expression analysis was performed using each isoform specific primer and the shared reverse primer (Fig. 7B). Our RACE experiments could not identify the isoform previously described (Suzuki et al. 2005) as the isoform might be an intron retaining transcript which could be longer than the detectable length of our RACE experiments. However, using the same primers as Suzuki et al. (2005), we re-examined its allelic expression (Fig. 7A, B). The reverse primer was designed to detect the previously described C/A SNP site at the 3′UTR region (Suzuki et al. 2005) (Fig. 7B: Arrowhead). Seventeen samples were examined, and six animals had a heterozygous SNP. The mothers of four of the six animals were not homozygous, but fortunately, the maternal homozygous SNPs in the remaining two samples were different from each other (Fig. 7C and Supplementary Fig. 2). In animal #1, all isoforms were expressed from the paternal allele (A), exclusively. However, in animal #2, all isoforms were expressed from the maternal (A) allele (Fig. 7C). Therefore, all isoforms in the two animals were not imprinted but all were mono-allelically expressed from the A allele in the tammar TOM tissues (Fig. 7C). If this mono-allelic expression was due to PCR bias based on differences in primer amplification efficiency between the two alleles, each isoform should have a mono-allelic expression even in different tissues. To clarify this, we examined allelic expression of each isoform in a further six pouch young liver tissues. Of these, only three PY had the C/A SNP and none of the mothers had a homozygous SNP at this site. Therefore, we could not confirm parental origin specificity. However, MESTB showed mono-allelic expression in the PY livers as well as the placenta tissues (n = 2), whereas MESTA showed an evident bi-allelic expression in the PY livers (n = 3) (Fig. 7C). We concluded that the mono-allelic expression of MESTA is tissue specific and reflects the number of templates in the sample, but not a PCR bias between the two alleles caused by the primers.

Fig. 7: Allelic expression of the MEST isoforms in the tammar placenta.
figure 7

A Exon structures of the two MEST isoforms identified in this study and the previously described isoform. Orange, aqua and green coloured boxes represent MESTA-specific, MESTB-specific and the previously described isoform specific exons, respectively. B Primer design for the allelic expression of the MEST isoforms. The previously confirmed SNP site (Suzuki et al. 2005) was used for this study. An arrowhead represents the SNP site. Allelic expression analysis of each isoform was performed using each isoform specific forward primer and a common reverse primer. C Allelic expression of tammar MEST isoforms by direct sequencing with PCR amplification. In animal #1, clear paternal expression was observed in the TOM tissue. However, in animal #2, all MEST isoforms were expressed mono-allelically from the maternal genome in TOM tissue. MESTA was not sufficiently highly expressed in BOM tissue to determine its allelic expression.

Full size image

Discussion

By characterising MEST isoforms in marsupials and monotremes, this study confirmed the presence of the short isoform of MEST in all mammalian groups. We identified a conserved antisense lncRNA, MESTIT1, in the tammar. This antisense transcript was present in tammar sperm as seen in humans (Li et al. 2002). Further comparison of the MEST and CEP41 gene loci amongst mammals showed that there is a marsupial specific CEP41 isoform sharing a CpG island with the shorter isoform of MEST. The MEST-CEP41 flanking region in eutherians acquired differential methylated CpG islands between the parental alleles and a retrotransposon insertion. In contrast to previous studies which suggested that MEST was imprinted in two marsupials (Das et al. 2012; Suzuki et al. 2005), it was not imprinted in the tammar placenta (this study). Although MEST, CEP41B, and MESTIT1 shared the same CpG island, only MEST had monoallelic expression in the placenta. This study suggests that MEST is imprinted only in eutherian mammals with the acquisition of a DMR at its promoter CpG islands. Characterisation of monotreme MEST allelic expression is not yet known but was beyond the scope of this study because obtaining maternal-young material in monotremes is extremely difficult.

Imprinted expression can only evolve when gene expression levels affect inclusive fitness in the ancestor. It is thought that imprinted genes might have evolved from genes with pre-existing gene dosage sensitivity, which led to the feature of parental origin specific expression (Haig 2000; Patten et al. 2014). In this way, marsupials and the MEST gene locus may offer an opportunity to study the key steps involved in how imprinting evolved from a normal bi-allelic expressed gene to a dosage sensitive/mono-allelically expressed gene to an imprinted gene. MEST appears to be involved in the maturation of mammary gland (Yonekura et al. 2019) and the formation of white adipose tissue, which is critical for mammary gland development (Couldrey et al. 2002; Kamei et al. 2007; Takahashi et al. 2005). In the tammar wallaby, the neonate is altricial and essentially an exteriorised fetus, and so requires a nutrient supply tailor-made for the differentiation and growth of the young. This is achieved by a changing maternal milk composition through a long, complex, and physiologically sophisticated period of lactation (Green et al. 1983; Green and Renfree 1982; Nicholas et al. 2012; Trott et al. 2003). Marsupials have effectively exchanged the umbilical cord for the teat (Renfree 2010). It is, therefore, evident that marsupial mammary glands have been subjected to different selection pressures compared to that of eutherians. However, since the short isoform of MEST was confirmed even in platypus, it is possible that the evolution of the MEST isoform occurred with the evolution of the mammary gland in the common ancestor of mammals. The subsequent evolution of MEST imprinting only in eutherians may be due to different selection pressures in each lineage. However, the exact function of the MEST short isoform is currently unknown in any mammal. To better understand its conserved role in mammary gland maturation, it would be interesting to characterise transcript localisation of the isoform in developing marsupial or monotreme mammary glands in subsequent studies.

Detailed comparisons of the flanking regions of MEST-CEP41 have provided insight into the evolution of this locus in mammals. For example, there was a conserved antisense lncRNA, MESTIT1, in the tammar. This antisense lncRNA is expressed from the known maternally methylated CpG islands in which the MESTA transcription occurs in humans. These data suggest that the CG rich region has conserved bidirectional promoter activity in therian mammals. Importantly, the tammar MESTIT1 shares the same CpG island and transcriptional orientation with a marsupial specific CEP41, CEP41B. However, in eutherians, there was a retrotransposon insertion between CEP41 and MEST so that this insertion physically separated these two genes in eutherians. Our RACE experiment could not detect any CEP41B-like transcript in mice. Even in the platypus, despite the physical proximity of MEST and CEP41, we were unable to identify a transcript similar to the tammar CEP41B by the 5′ RACE reaction. Thus, the bidirectional promoter activity at the CG rich region where MESTA transcription occurs must have evolved in the common ancestor of therian mammals.

Previously, allelic expression of tammar MEST was examined but combinations of mothers homozygous for a different SNP were not available (Suzuki et al. 2005). In the present study, we used a reciprocal SNP and analysed allelic expression of the tammar MEST in placenta. The results confirmed that tammar MEST is not imprinted in the placenta, although it is clearly mono-allelically expressed. In the opossum study, the MEST transcript was mono-allelically expressed but there was no parental origin information (Das et al. 2012). Loss of imprinting normally leads to either bi-allelic expression or complete inactivation of imprinted genes (Holm et al. 2005). The clear maternal mono-allelic expression in our data indicates that the MEST expression was not caused by loss of imprinting but was due to random monoallelic expression in the tammar placenta. In our data, the MEST isoforms were preferentially expressed from the allele containing an “A” SNP in the placenta. Similar preferential expression is observed at the tammar DIO3 gene in the placenta (Edwards et al. 2008). As DIO3 is also known to be imprinted in eutherians but not in marsupials (Edwards et al. 2008), this preferential expression may reflect a similar gene regulatory mechanism between MEST and Dio3. Since our study and that of DIO3 used placenta tissues for the analysis, the data represent allelic expression of the pooled placental cells. The leaky expression from the opposite allele in the placenta might reflect random mono-allelic expression in the cell population, or the switching of allelic expression by individual cells, as seen for the mouse Ddc gene (Bonthuis et al. 2022).

It is currently unclear how the random mono-allelic expression of MEST is regulated in the tammar. This needs to be clarified in subsequent studies. A transcriptome-based analysis using grey short-tailed opossums showed that there is bi-allelic expression of opossum MEST (Cao et al. 2023). However, the short read-based approach is not sufficient to conclude imprinting status of isoform-dependent imprinted genes like eutherian MEST as each read lost its isoform specificity by fragmentation. Long-read transcriptomics in homogeneous sorted-cell populations would provide an unequivocal answer.

This study demonstrates that monoallelic expression of the MEST gene occurs in the tammar placenta without imprinting. More detailed analyses of the expression mechanism and function are needed to clarify the differences between genes that are imprinted and those that are not. In marsupials, (paternal) X chromosome inactivation is regulated by imprinting but by a different gene (RSX) from that seen in eutherians (Xist) (Cooper et al. 1989, 1971; Cooper 1971; Grant et al. 2012; Mahadevaiah et al. 2020). In contrast, in eutherians, X chromosome inactivation is random (except in the extraembryonic tissues) (Sahakyan et al. 2018). Loss of differential DNA methylation at the MEST gene locus would cause developmental defects due to the loss of its imprinting. Imprinting of this locus has therefore undergone different selection pressures and evolved differently after the eutherian-marsupial split with the acquisition of maternally methylated CpG islands and imprinting in eutherians but not marsupials.