You have two copies of every gene your body; one inherited from each of your parents. However, there are circumstances where only the maternally or paternally inherited allele is expressed, while the reciprocal allele is repressed. This is known as imprinting and is usually achieved by changes in the structure of DNA, rather than the code. These alterations are a type of epigenetic modification and are crucially important for controlling gene dosage and expression. Parental specific origin affects in plants, mammals and insects have been documented for over 40 years now but it is only recently that we have been able to decipher some of the mechanisms responsible for imprinting. 

It was crucial embryological research that first showed a functional non-equivalence of the parental genomes. During the early stages of fertilisation the maternal and paternal genomes remain separate in the zygote within defined pronuclei. In the 1980s, the Solter and Surani groups developed techniques to carefully remove the paternal pronucleus from a newly fertilised mouse zygote and replaced it with a second maternal pronucleus. This creates a bimaternal embryo, also known as gynogenetic or parthenogenetic embryos (Figure 1). Reciprocal experiments were also performed in which the maternal pronucleus was removed and replaced with a second paternal pronucleus, resulting in an androgenetic embryo (Figure 1). In both cases, the embryos failed to survive beyond mid-gestation.  

Interestingly the two types of conceptuses resulting from these experiments were morphologically distinct:

  • Parthenogenetic conceptuses developed embryonically but fail to produce any extra-embryonic tissues.
  • Androgenetic conceptuses developed extra-embryonic lineages but had poor embryonic development. 

These contrasting observations indicated that the two parental genomes served both distinct and complementary functions.

Figure 1. Pronuclear Transplant

Autosomal Imprinted Regions

Genetic studies using mice showed that entire regions of non-sex choromosomes, or autosomes, had imprinted expression patterns. A deletion in the Tme locus of chromosome 17 was shown to be lethal when inherited maternally, but not paternally. These experiments were subsequently confirmed and refined to identify a 0.8-1.1Mb locus that had an asymmetric bias of expression towards the maternal chromosome.

The Cattanach, Beechey and Searle groups conducted experiments using mice with specific types of translocation aberrations; they were able to manipulate the parental origin of a particular chromosomal locus. Translocation heterozygotes can produce unbalanced gametes – either deficient in a particular chromosomal region or duplicate. If reciprocal gametes are fused (for example; a translocation duplicate egg and deleted sperm, Figure 2, C) the offspring generated are expected to develop as normal diploid zygotes as the gene dosage is compensated between the sperm and egg. These offspring are described as having uniparental disomy, literally meaning that one parent provides a double genetic complement. Occasionally however, uniparental disomic offspring would be unviable or develop abnormalities in their growth or behaviour. These findings suggested differences in the expression of chromosomal regions dependant on the parental origin. This approach was used extensively over the following years and eventually led to an identification of 13 subchromosomal regions, which need to be inherited both maternally and paternally for normal development to occur.

Figure 2. Uniparental Disomy

Figure 2. Uniparental Disomy:  A. Normal union between maternal and paternal haploid genomes.  Maternally repressed imprinted genea in light blue, paternally and maternally expressed unimprinted geneb in yellow.  B. Sperm cell contains partial double copy of the chromosomal region containing genea and geneb, oocyte cell contains deletion of reciprocal chromosomal region. Resultant somatic cells following fertilisation express normal dose of unimprinted gene b but a double dose of the paternally expressed imprinted gene a.  C. Oocyte cell contains partial double copy of the chromosomal region containing genea and geneb, sperm cell contains deletion of reciprocal chromosomal region. Resultant somatic cells following fertilisation express normal dose of unimprinted geneb but does not express gene a as it is maternally repressed due to imprinting.  Examples B and C would develop abnormally due to issues arising from altered gene dosage. These experiments originally performed by Cattanach, Beechey and Searle labs highlighted the importance of gene imprinting in controlling gene dosage.

Chromosomal Imprinting

The parental-origin-specific effects of imprinting are also known to affect whole chromosomes. The process of X chromosome inactivation occurs in order to regulate gene dosage in female mammals. Mammalian females inherit two X chromosomes, whereas males gain only a single X and a male specific Y chromosome. Therefore females have double the copies of all genes contained on the particularly gene rich X chromosome. Two copies of the X chromosome would lead to an unbalanced gene dosage and therefore, one of the X chromosomes is silenced by imprinting. At the blastocyst stage of development, X chromosome inactivation randomly silences either the maternally or paternally inherited X chromosome by packaging it into transcriptionally silent heterochromatin. This process is irreversible and so all descendants arising from these cells maintain inactivation of the same chromosome. A similar process occurs in marsupial mammals, however the paternal X chromosome is imprinted and therefore always selected for transcriptional silencing. This means that genes on the paternally inherited X chromosome are never expressed in the father’s offspring.



Evolution of Imprinted Genes

Insulin-like growth factor 2 receptor (Igf2r) was the first single gene shown to be imprinted. The gene was targeted as it is nested along with several other genes within the Tme locus of Chromosome 17 (see above).The Igr2r gene is solely expressed from the maternal allele and its translated protein functions as a receptor for insulin like growth factor (IGF2). Remarkably, IGF2 was also shown to be an imprinted gene. Mice harbouring a targeted Igf2 deletion on distal chromosome 7 showed a growth deficiency when inherited paternally but not maternally. Separate studies corroborated this evidence by showing that Igf2 was unexpressed from embryos generated from maternal uniparental duplication and paternal loss of the distal chromosome 7 region.  Therefore, it was concluded that the gene was ordinarily expressed from only the paternal chromosome. This means the Igf2 gene is imprinted in an opposing fashion to its receptor Igf2r. 

The Igf2-Igf2r paradigm becomes particularly interesting when considering the resultant protein function in relation to their imprinted expression pattern. IGF2 (paternal allele expressed) is a factor which potentiates foetal growth and development primarily through IR-A (insulin receptor isoform A) and IGF1R (insulin-like growth factor 1 receptor). Resultantly Igf2 knockout mice weigh up to 60% less than wild type mice, and over expression of the gene leads to overgrowth. IGF2 also binds the clearance receptor IGF2R (maternally expressed). Functionally IGF2R operates as a 'growth limiter', transporting IGF2 to the lysozymes for degradation and therefore reducing its bioavailability. Hence the paternally expressed IGF2 and maternally expressed IGF2R are oppositely imprinted and have reciprocal functions, respectively promoting and attenuating growth. 

Mice can be genetically manipulated to produce a colony harbouring a deletion of the Igf2 placental specific promoter; abolishing Igf2 expression in the placenta only. When this deletion was paternally inherited placental growth was significantly limited and temporally followed by fetal growth restriction (a condition in which fetal growth does meet it's genetically determined potential). This was the first experimental evidence that imprinted genes directly regulated the placenta, an organ which the supply of maternal nutrients to the fetus. Since then many other imprinted genes have been shown to exhibit a profound effect on placental development. As an outcome mammalian imprinting is now believed to have co-evolved with the placenta. Phylogenetically, gene imprinting emerges in concordance with viviparity (live birth) and the number and regulatory complexity of imprinted genes appears to be related to the maternal investment.  


  • Eutherian mammals (including Humans and Mice)
    • Highly developed placenta
    • Huge maternal investment
    • Over 100 imprinted genes currently identified


  • Marsupial mammals (including Kangaroos and Koalas)
    • Diverged from Eutherians 125-148MYA
    • Produce a functional placenta but gestation intrauterine development is greatly diminished in comparison to Eutherians
    • Characteristically underdeveloped young
    • Fewer imprinted gene clusters, many exhibit differential imprinted regulation


  • Monotremes (Echidna and Duck-billed Platypus)
    • Diverged from Eutherians and Marsupials 166MYA
    • Persist in laying eggs
    • No placenta
    • No identified imprinted genes


One of the major theories to explain the complex driving force underpinning the evolution of imprinting is called the parental fitness conflict which arises from the divergent needs of the mother and father. Essentially females can bear and raise offspring by multiple fathers; the mother is related to all offspring however any one father is related only to a subset of the mother’s total offspring. Hence there is an issue of genetic fitness: the father seeking to exploit the resources of the mother for the sake of his own offspring even if this damaging the mother; and the mothers need to defend herself against the parasite-like demands of the offspring so that she may carry more young in the future. These conflicts are thought to be played out on the genetic level and the hypothesis is indeed substantiated by the fact that many paternally expressed imprinted genes encourage placental and fetal growth (such as Igf2) while conversely maternally expressed genes act to counteract these affects and suppress the development of the fetus (including IGF2R). Similarly as previously discussed parthenogenetic conceptuses develop mostly embryonically whereas androgenetic conceptuses are embryonically weak but have well differentiated extra-embryonic lineages, observations which fit well with the conflict hypothesis.



Epigenetic Control of Imprinted Genes

So far we have discussed the monoallelic expression of genes and the importance of this trait for normal development. However, how are genes marked for parental specific regulation? Definitively genomic imprinting must fulfil four key mechanistic principles:

  • The imprint must influence transcription
  • The imprint must be heritable to all somatic daughter cells deriving from the zygote so that the parental origin is permanently maintained throughout the life of the organism.
  • The imprint must be defined during gametogenesis or immediately following fertilisation when the paternal and maternal chromosomes are kept within separate nuclei. 
  • The inherited imprint must be deleted in germ line cells and then re-established with the appropriate pattern.

A process known as DNA methylation fulfils all these criteria. In mammals a cytosine base followed by a guanine is known as a CpG dinucleotide (the p represents the phosphate backbone). These motifs are sensitive to enzyme driven modification where the cytosine component is converted to 5-methylcytosine. This modification profoundly changes the ability of many DNA binding proteins to interact with DNA. As a result, gene regulatory regions rich in this modification are usually transcriptionally silent. These DNA methylation marks are also known to be heritable; the enzyme DNA methyltransferase 1 (DNMT1) performs a maintenance role by propagating the DNA methylation marks from the old methylated strand to the new DNA molecule during DNA replication. 

Soon after the first imprinted genes were identified, it was shown that these loci had specific DNA methylation marks that were different between the parental chromosomes. These regions are called differentially methylation regions (DMRs) of which there are two types:

  1. Germline DMRs; which become differentially methylated within the separate parental germlines.
  2. Secondary (or somatic) DMRs; which are acquired after fertilisation. The manifestation of secondary DMRs are often tissue specific and require the presence of a germline DMR to be marked. 

Targeting these regions for deletion in the germline cells of mice showed that DMRs were in fact special regions of DNA that regulate whole gene clusters. These regions are known as imprinting control regions (ICRs) and form the mechanistic switch responsible for the monoallelic (from one allele only) gene expression.



Igf2-H19 Imprinted Cluster Model

The H19 gene, is situated only 90kb downstream of Igf2 and expresses an evolutionarily conserved micro RNA (miRNA). miRNAs are short non-coding stands of RNA, which specifically bind to complementary sequences of specific messenger RNAs (mRNA) and repress their translation though silencing or degradation. They are essential regulators of mRNA translation and are often associated with imprinted gene clusters. The H19 gene is monoallelically expressed from the maternally inherited allele, therefore having the opposing imprint to the neighbouring Igf2 gene. The contrasting pattern of imprinted regulation of both Igf2 and H19 was found to be due to a differentially methylated ICR and secondary DMRs which affects binding of an insulator protein CTCF. CTCF regulates the function of downstream enhancers, allowing them to elicit transcription from the Igf2 or H19 promoters in a parental inherited allele specific manner. 

The Igf2-H19 cluster has since become an established model of our understanding of genetic imprinting control (Figure 3).

Figure 3. The Igf2-H19 Imprinted Gene Cluster

Figure 3 Igf2-H19 Schematic: Schematic shows how imprinted DNA methyation controls the expression of Igf2 and H19 genes via downstream enhancers. On the maternally inherited copy the unmethylated imprinting control region (ICR) is bound by CTCF. CTCF acts as an insulator preventing downstream enhancer regions from driving  transcription of Igf2.  The enhancers can however access H19 and elicit its transcription. Conversely the paternally inherited copy has a methylated ICR. CTCF cannot bind methylated DNA and so the enhancers are able to promote Igf2 transcription. A secondary DMR (blue circle) lies at the H19 promoter and represses its transcription. Resultantly Igf2 is only expressed on the paternal allele and H19 only from the maternal allele.



Establishment of the Imprint

Imprinting is suspected to arise in the germline, when the maternal and paternal genomes are separate. During development the methylation patterns in primordial germ cells are identical to those found in somatic cells. Once they migrate fully to the genital ridge, the genome is globally demethylated in a process believed to be intrinsically linked to chromatin architecture. Once erased the parental specific pattern of methylation is remarked, however this process is markedly different between the sexes. In males the process occurs during prospermatogonia and hence occurs in the period between mitotic arrest and birth. Females on the other hand do not impart their imprint until after birth, specifically during pre-ovulatory oocyte growth phase (Figure 4).

The de novo DNA methylation imprint itself is imparted by DNA methyltransferase 3 (DNMT3) enzymes.  There are three DNMT3 enzymes; DNMT3a and DNMT3b, which have DNA methyltransferase activity, and DNMT3L which is catalytically inactive but serves a purpose as a regulatory factor in germ cells. All DNMT enzymes including DNMT1 are essential for embryonic development, thus underscoring the pivotal role DNA methylation plays in mammalian development. Almost all imprinted genes tested have a requirement of DNMT3b and DNMT3L to mark their ICRs.

The DNA methylation mechanism is believed to have been co-opted from pre-existing molecular defence system. The host defence hypothesis postulates that gene imprinting developed from molecular machinery which used DNA methylation as a way of silencing foreign DNA motifs which were inserted into out genome.  This hypothesis is bolstered by the finding that DNMT3L is a key regulator of methylation driven silencing of retrotransposons in the male germ cells. 



Maintenance of the Imprint

Once the imprint has been established in the primordial germ cells, the DNA methylation marks must be faithfully maintained throughout fertilisation and development (Figure 4). The paternal pronucleus undergoes active demethylation, under the control of a family of enzymes known as TET enzymes. Whereas the maternal pronucleus is subject to passive but just as pervasive demethylation of the maternal pronucleus through DNA replication in the early embryo. In both circumstances however the parental imprint must be protected. One factor recently shown to play a role in this paradigm is a KRAB zinc finger protein known as ZFP57. Murine mutations of the Zfp57 gene in the mother and zygote cause loss of DNA methylation at the ICRs of several known imprinted genes. This results in maternal and zygotic lethality. It may be that ZFP57 is required to protect ICRs from global demethylation and hence maintain the imprinted DNA methylation patterns where non-imprinted methylation marks are lost. Zfp57 as well as other factors implicated to be important in protecting imprinted domains are currently under intense investigation to establish their unambiguous role.

Figure 4. Cycle of DNA Methylation and Imprinting



Bartolomei M S, Ferguson-Smith A C. Mammalian genomic imprinting. 2011. Cold Spring Harb Perspect Biol. Jul 1;3(7).

Renfree M B, Hore T A, Shaw G, Graves J A, Pask A J. Evolution of genomic imprinting: insights from marsupials and monotremes. 2009. Annu Rev Genomics Hum Genet. 10:241-62.

Original Articles

McGrath J, Solter D. Completion of mouse embryogenesis requires both the maternal and paternal genomes. 1984. Cell. 37(1):179-83.

Barlow D P, Stöger R, Herrmann B G, Saito K, Schweifer N. The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus. 1991 Nature. 3;349(6304):84-7.

Bell AC, Felsenfeld G. Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene. 2000. Nature. 25;405(6785):482-5.

Wossidlo M, Nakamura T, Lepikhov K, Marques CJ, Zakhartchenko V, Boiani M, Arand J, Nakano T, Reik W, Walter J. 5-Hydroxymethylcytosine in the mammalian zygote is linked with epigenetic reprogramming. 2011. Nat Commun.2:241.


Fastbleep © 2019.