12.6: Epigenetics - Biology

12.6: Epigenetics - Biology

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

The word “epigenetics” has become popular in the last decade and its meaning has become confused. The term epigenetics describes any heritable change in phenotype that is not associated with a change the chromosomal DNA sequence.

Originally it meant the processes through which the genes were expressed to give the phenotype; that is, the changes in gene expression that occur during normal development of multicellular organisms. This includes the change in transcriptional state of a DNA sequence (gene) via DNA or chromatin protein reversible modifications. Thus, DNA methylation and chromatin protein methylation, phosphorylation, and acetylation have been targeted as mechanisms for “heritable” changes in cells as they grow from a single cell (zygote) and differentiate to a multicellular organism. Here, dividing cells commit to differentiate into different tissues such as muscle, neuron, and fibroblast due to the genes that they express or silence. Some genes are irreversibly silenced, through epigenetic mechanisms, in some cell types, but not in others. This doesn’t involve any change in DNA sequence.

Remember, these epigenetic effects are not permanent changes and thus cannot be selectable in an evolutionary context. However, mutations in the genes that regulate the epigenetic effect can be selected.

Definition: Epigenetics

Epigenetics describes any heritable change in phenotype that is not associated with a change the chromosomal DNA sequence.

Some heritable information can be passed on independent of the DNA sequence

More recently however, researchers have found many cases of environmentally induced changes in gene expression that can be passed on to subsequent generations – a multi-generational effect. These cases have also been called “epigenetics”, and probably involve similar reversible changes to the DNA and chromatin proteins.These altered expression patterns represent the diversity of expression for a genome. This “extended” phenotype, the ability to influence traits in the next generation, is a topic of current research and only some examples will be discussed here.

One example comes from the grandchildren of famine victims are known to have lower birth weight than children without a family history of famine. This heritability of altered state of gene expression is surprising, since it appears not to involve typical changes in the sequence of DNA.The term epigenetics is applied here since the apparently heritable change in phenotype is associated with something other than chromosomal DNA sequence.

This change is inherited from one generation to the next and is thus transgenerational. In develpmental epigenetics, the expression state (developmentally differentiated state) is conserved only from one mitosis to the next, but is erased or rest at meiosis (the boundary of one generation to the next). The basis of at least some types of epigenetic inheritance appears to be replication of patterns of histone and DNA methylation that occurs in parallel with the replication of the primary DNA sequence. It is becoming clear that epigenetics is an important part of biology, and can serve as a type of cellular memory, sometimes within an individual, or sometimes across a few generations, at least.

The permanence of this “change” is not the same as changes in the DNA sequence itself. What is clear is that epigenetics is an important part of regulating gene expression, and can serve as a type of cellular memory, certainly within an individual, or across a few generations in some cases.

Imprinting and Parent-of-Origin Effects

For some genes, the allele inherited from the female parent is expressed differently than the allele that is inherited from the male parent. This is distinct from sex-linkage and is true even if both alleles are wild-type and autosomal. During gamete development (gametogenesis), each parent imprints epigenetic information on some genes that will affect the activity of the gene in the offspring. Imprinting does not change the DNA sequence, but does involve methylation of DNA and histones, and generally silences the expression of one of the parent’s alleles. In humans, some genes are expressed only from the paternal allele, and other genes are expressed only from the maternal allele. The imprinting marks are reprogrammed before the next generation of gametes are formed. Thus, although a male inherits epigenetic information from both his mother and father, this information is erased before sperm development, and he passes only one pattern of imprinting to both his sons and daughters. Most examples of imprinting come from placental mammals, and many imprinted genes control growth rate, such as IGF2 (insulin-like growth factor 2).

Imprinting appears to explain many different parent-of-origin effects. For example, Prader-Willi Syndrome (PWS) and Angelman Syndrome (AS) are two phenotypically different conditions in humans that result from deletion of a specific region of chromosome 15, which contains several genes. Whether the deletion results in PWS or in AS depends on the parent-of-origin. If the deletion is inherited from the father, PWS results.Conversely, if the deletion is inherited from the mother, AS is the result. The gene(s) involved in PWS is maternally silenced by imprinting, therefore the deletion of its paternally-inherited allele results in a complete deficiency of a required protein.On the other hand, the paternal allele of the gene involved in AS is silenced by imprinting, so deletion of the maternal allele results in deficiency of the protein encoded by that gene.

Transgenerational inheritance of nutritional influences

Nutrition is one aspect of the environment that has been particularly well-studied from an epigenetic perspective in both mice and humans. People alive today who experienced the Dutch famine of 1944-1945 as fetuses have IGF2 genes that are less-methylated than their siblings. Methylation of IGF2 (and birth rate) is also lower in children of mothers who do not take folic acid supplements as compared those who do.Furthermore, an individual’s phenotype can be influenced by the nutrition of parents or even grandparents.This transgenerational inheritance of nutritional effects appears to involve epigenetic mechanisms.

The mouse agouti gene produces a signaling molecule that regulates pigment-producing cells and brain cells that affect feeding and body weight. Normally, agouti is silenced by methylation, and these mice are brown and have a normal weight. When agouti is demethylated by feeding certain chemicals or by mutating a gene that controls methylation, some mice become yellow and overweight, although their DNA sequence remains unchanged.Methylation of agouti and normal weight and pigmentation of offspring can be restored if their mothers are fed folic acid and other vitamins during pregnancy.

A study of an isolated Swedish village called Överkalix provides an example of transgenerational inheritance of nutritional factors. Detailed historical records allowed researchers to infer the nutritional status of villagers going back to 1890.The researchers then studied the health of two generations of these villagers’ offspring, using medical records. A significant correlation was found between the mortality risk of grandsons and the food availability of their paternal grandfathers.This effect was not seen in the granddaughters. Furthermore, the nutrition of paternal grandmothers, or either of the maternal grandparents did not affect the health of the grandsons. It was therefore proposed that epigenetic information affecting health (specifically diabetes and heart disease) was passed from the grandfathers, to the grandsons, through the male line.

Vernalization as an example of epigenetics

Many plant species in temperate regions are winter annuals, meaning that their seeds germinate in the late summer, and grow vegetatively through early fall before entering a dormant phase during the winter, often under a cover of snow. In the spring, the plant resumes growth and is able to produce seeds before other species that germinated in the spring. In order for this life strategy to work, the winter annual must not resume growth or start flower production until winter has ended. Vernalization is the name given to the requirement to experience a long period of cold temperatures prior to flowering.

How does a plant sense that winter has passed? The signal for resuming growth cannot simply be warm air temperature, since occasional warm days, followed by long periods of freezing, are common in temperate climates. Researchers have discovered that winter annuals use epigenetic mechanisms to sense and “remember” that winter has occurred

Fortunately for the researchers who were interested in vernalization, some varieties of Arabidopsis are winter annuals. Through mutational analysis of Arabidopsis, researchers found that a gene called FLC (FLOWERING LOCUS C) encodes a transcription repressor acting on several of the genes involved in early stages of flowering (Figure (PageIndex{16})).In the fall and under other warm conditions, the histones associated with FLC are acetylated and so FLC is transcribed at high levels; expression of flowering genes is therefore entirely repressed. However, in response to cold temperatures, enzymes gradually deacetylate the histones associated with FLC. The longer the cold temperatures persist, the more acetyl groups are removed from the FLC-associated histones, until finally the FLC locus is no longer transcribed and the flowering genes are free to respond to other environmental and hormonal signals that induce flowering later in the spring. Because the deacetylated state of FLC is inherited as cells divide and the plant grows in the early spring, this is an example of a type of cellular memory mediated by an epigenetic mechanism.


New England Biolabs has called upon its 40 years of expertise in enzymology to develop several solutions for epigenetics research, including NEBNext ® Enzymatic Methyl-seq (EM-seq&trade) and the EpiMark ® suite of products.

A new method for DNA methylation analysis at the single base level, NEBNext Enzymatic Methyl-seq (NEB #E7120), is now available. This enzyme-based technology minimizes damage to DNA and produces high quality libraries that enable superior detection of 5mC and 5hmC from fewer sequencing reads.

Easy-to-use EpiMark kits simplify DNA methylation (5mC) and hydroxymethylation (5hmC) detection and analysis, as well as ChIP, histone and nucleosome analysis. Independently applicable, individual epigenetics reagents also complement the EpiMark ® kits. NEB's methylation- and hydroxymethylation- sensitive or dependent enzymes, DNA methyltransferases and DNA controls are all useful for mapping DNA modifications and methylating DNA at specific sites for gene expression studies. Our protein methyltransferases and recombinant histones perform efficiently in protein modification and characterization studies. Our range of modified and unmodified genomic DNAs can be used as controls for detection of DNA methylation. Our series of human DNA (cytosine-5) methyltransferase (DNMT) antibodies are ideally suited for Western blots and immunoprecipitation.

Our complete Epigenetics suite is expertly designed for optimized research and discovery.

EpiMark ® is a registered trademark of New England Biolabs, Inc.
EM-seq&trade is a trademark of New England Biolabs, Inc.

Epigenetic adaptation of the placental serotonin transporter gene (SLC6A4) to gestational diabetes mellitus

We tested the hypothesis that gestational diabetes mellitus (GDM) alters the DNA methylation pattern of the fetal serotonin transporter gene (SLC6A4), and examined the functional relevance of DNA methylation for regulation of the SLC6A4 expression in the human placenta. The study included 50 mother-infant pairs. Eighteen mothers were diagnosed with GDM and 32 had normal glucose tolerance (NGT). All neonates were of normal birth weight and born at term by planned Cesarean section. DNA and RNA were isolated from samples of tissue collected from the fetal side of the placenta immediately after delivery. DNA methylation was quantified at 7 CpG sites within the SLC6A4 distal promoter region using PCR amplification of bisulfite treated DNA and subsequent DNA sequencing. SLC6A4 mRNA levels were measured by reverse transcription-quantitative PCR (RT-qPCR). Functional SLC6A4 polymorphisms (5HTTLPR, STin2, rs25531) were genotyped using standard PCR-based procedures. Average DNA methylation across the 7 analyzed loci was decreased in the GDM as compared to the NGT group (by 27.1%, p = 0.037) and negatively correlated, before and after adjustment for potential confounder/s, with maternal plasma glucose levels at the 24th to 28th week of gestation (p<0.05). Placental SLC6A4 mRNA levels were inversely correlated with average DNA methylation (p = 0.010) while no statistically significant association was found with the SLC6A4 genotypes (p>0.05). The results suggest that DNA methylation of the fetal SLC6A4 gene is sensitive to the maternal metabolic state in pregnancy. They also indicate a predominant role of epigenetic over genetic mechanisms in the regulation of SLC6A4 expression in the human placenta. Longitudinal studies in larger cohorts are needed to verify these results and determine to which degree placental SLC6A4 changes may contribute to long-term outcomes of infants exposed to GDM.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.


Fig 1. Location and sequence of the…

Fig 1. Location and sequence of the SLC6A4 promoter region targeted by DNA methylation analyses.

Fig 2. Placental SLC6A4 methylation levels correlate…

Fig 2. Placental SLC6A4 methylation levels correlate with maternal fasting glucose concentrations and placental SLC6A4…

Fig 3. Placental SLC6A4 mRNA levels according…

Fig 3. Placental SLC6A4 mRNA levels according to maternal glucose tolerance status and 5-HTTLPR /…

State of the Art Production Facility

Our Production facility complies with the highest standards in the industry which include ISO, GMP, FDA registration, HACCP and SALSA. To ensure that you receive a product that is of the highest quality, all of our ingredients are rigorously tested – what you’re putting in is the best of the best.

We combine a clean-room production facility with the accuracy and attention to detail of our trained staff. This results in formulas being created in the cleanest environment, meaning they’ll be most effective in your system. We believe in producing only the purest natural quality products for our customers.

Epigenetics – It’s not just genes that make us

Quick look: In its modern sense, epigenetics is the term used to describe inheritance by mechanisms other than through the DNA sequence of genes. It can apply to characteristics passed from a cell to its daughter cells in cell division and to traits of a whole organism. It works through chemical tags added to chromosomes that in effect switch genes on or off.

Researchers studying the microscopic roundworm Caenorhabditis elegans recently discovered a set of mutations that extended the worms’ normal 2-3 week lifespan by up to 30%. This was exciting, not least because discoveries in animals such as roundworms can sometimes help us understand processes like ageing in humans. This was not the end of the story though, as the researchers found that the descendants of the long-lived roundworms could also live longer than normal, even if they only inherited the non-mutated version of the genes from their parents. This doesn’t seem to make sense at first surely characteristics such as hair colour, height and even how long we or a microscopic worm could potentially live are carried in the DNA sequence of the genes that we inherit from our parents. So how can we solve the conundrum of how the roundworms inherited the long lived characteristic, without inheriting the DNA sequence that initially caused it? The answer is epigenetics.

It’s not all in your DNA

In a nutshell, epigenetics is the study of characteristics or “phenotypes” that do not involve changes to the DNA sequence and the long-lived roundworms are just one of many examples. Others, as we will see below, include how queen and worker honey bees can appear so different despite being genetically identical, how starvation in human populations may affect the health and longevity of the next generation, why all tortoiseshell cats are female and even how we all develop from a single cell (a fertilized egg) to end up with bodies containing many different types of specialised cells but which all contain the same genes and DNA sequence.

So what is epigenetics?

Another way of looking at epigenetics is like this while traditional genetics describes the way the DNA sequences in our genes are passed from one generation to the next, epigenetics describes passing on the way the genes are used. To make a computer analogy, think of epigenetics as metadata, information describing and ordering the underlying data. If you own an MP3 player for example, it will contain a lot of data, the MP3 files. Think of these as analogous to genes. But you will also probably have playlists or you may play tracks by artist or genre. This information, playlist, artist, genre etc. is metadata. It determines which tracks are played and in what order, and this is what epigenetics is to genetics. It is a set of processes that effects which genes are switched-on, or “expressed”, as molecular biologists would say.

How does epigenetics work?

So epigenetics is about how genes are expressed and used, rather than the DNA sequence of the genes themselves, but how does this work? Many researchers have been studying epigenetics over the past few decades, and it is currently an area of intense research activity. We know that a part of how epigenetics works is by adding and removing small chemical tags to DNA. You can think of these tags as post-it notes that highlight particular genes with information about whether they should be switched on or off. In fact the chemical tag in question is called a methyl group (see Diagram 1) and it is used to modify one of the four bases or “chemical letters”, A, C, T and G, that makes up the genetic code of our DNA. The letter that is tagged is C or cytosine and when it is modified, or methylated it is called 5-methyl cytosine. Methyl groups are added to DNA by enzymes called DNA methyl transferases (DNMTs).

Diagram 1. Two chemical tags, methyl and acetyl groups that are central to epigenetic phenomena and the chemical structure of cytosine and 5-methyl cytosine in DNA. The pentagonal part of the molecule forms the continuous “backbone” of the DNA . Only one of the two strands of DNA that makes up the familiar double helix is shown.

Queen bee status is partly determined by fewer methyl tags

In most cases, more methylated Cs in the DNA of a gene results in the gene being switched off. Honey bees provide us with a good example of how this can work. Worker bees and the queen have very different bodies the queen is much larger, longer lived, has an enlarged abdomen and lays many thousands of eggs, while the smaller workers are sterile but have complex foraging and communication skills. Despite this, the queen and workers in a hive are female and genetically identical. The clue to how this comes about lies in royal jelly, a secretion that is fed to some developing larvae, and which results in these larvae becoming queens rather than workers. We will come back to royal jelly and its queen-making properties later, but a fascinating piece of research showed that if the amount of the methyl group adding DNMT enzyme was artificially reduced in bee larvae, then the larvae developed into queens, even if they weren’t fed royal jelly. Thus, the switch between queen and worker can be flipped by the abundance of methyl tags on the bee larvae’s DNA. Fewer methyl tags leads to switching on of a special gene or genes in the developing larvae that results in the development of the larvae into queens and not workers.

Tags on tails also operate gene switches

DNA methyl tags are only one part of the story though. In the cells of all plants and animals, DNA is packaged or wrapped up into nucleosomes where the DNA double helix is wrapped around a central core of protein (see Diagram 2). About 150 letters-worth of DNA (or base-pairs) is wrapped around each nucleosome, and this helps package the 3 billion base pairs of genetic code into each of our cells. Nucleosomes are too small to see using conventional microscopes, but biologists use a technique called X-ray diffraction to work out the shape and organisation of objects like nucleosomes, and in 1997 this technique revealed the beautiful structure of nucleosomes at high resolution – see (

Diagram 2. The familiar DNA double helix (blue) is wrapped around nucleosomes (grey cylinders) in cells. The string of nucleosomes can be coiled into a thicker filament, called the 30 nm fibre and this can be further coiled into a still thicker chromatin fibre. When genes are switched on their nucleosomes are more uncoiled like the 10nm fibre.

Nucleosomes are compact, but the ends or “tails” of the proteins that make up the nucleosome, which are called histones, stick out from the otherwise compact nucleosome structure. Like the methyl tags on DNA, small chemical tags can also be added to these histone tails (see Diagram 3). Two of the chemical tags that are added to these tails are acetyl groups and methyl groups. Methyl, acetyl and a few other types of tags can be added to the tails in a large number of combinations and this effects whether an underlying gene is switched on or off. In fact genes can be switched right off (this is called silencing), full on, or somewhere in between by DNA methyl tags and histone tail tags. The combination of DNA and histone tags can also effect how easily a gene is turned on or off.

Diagram 3. Chemical tags can be added to the “tails” of the histone proteins that make up nucleosomes. Grey cylinder, nucleosome curved black lines, histone tails green circles, methyl tags red triangles, acetyl tags mauve hexagons, other types of tag.

When cells divide

When cells divide, the entire DNA sequence from the original cell (3 billion base pairs contained in 23 pairs of chromosomes in a human cell) is duplicated so that both daughter cells receive an exact copy. What, you might ask, happens to all those epigenetic tags? We have known for some time that the DNA-methyl tags are copied too, so that both daughter cells have the same pattern of DNA methylation. We now know that the pattern of histone tags is also mostly duplicated as cells divide, although this is currently less well understood. Nevertheless, cell division is also a time when epigenetic tags can most easily be changed.

Return of the long-lived worm

Right at the beginning we came across the story of the long-lived microscopic worms thatpassed on their longevity to their offspring even if the individual offspring did not inherit the variant gene (mutation) that originally caused the extended lifespan. We are now in a position to explain this apparently strange result. In most cases genes contain the information to make a protein molecule, and the protein molecules might be enzymes that carry out chemical reactions in the cell, or parts of the structure of the cell itself. It turns out that the genes that were mutated in the worm study make proteins that work together to add a methyl tag to nucleosomes. This tag is an on-switch. When one or more of the genes were mutated this tag was absent and several genes that should be on, including some involved in ageing, were switched off and the worms had a longer lifespan. The unexpected thing is that the epigenetic tags were thought to be completely erased or reset during the formation of sperm and egg, and so unlike the genes themselves they shouldn’t be passed on to the next generation. But this result and other research that shows that this is not always the case and that sometimes, the pattern of epigenetic tags are passed on.

How to make a queen

Whether a larval honey bee becomes a worker or a queen depends on an epigenetic switch, and this switch seems to be “flipped” by royal jelly. But what is it about royal jelly that leads a larva that would otherwise grow up to be a worker, to become a queen? The answer lies in understanding that the individual chemical tags that are added to the histone tails of nucleosomes are constantly being revised by the cell. Acetyl tags are added by enzymes called histone acetyl transferases and they are removed or erased by a second group of enzymes called histone deacetylases (HDACs). Both of these enzymes are present in most cells and this allows genes to be switched on or off over time.

More acetyl tags help deliver queen bee status

Recently, researchers set out to identify compounds in royal jelly that could alter this process, and what they found was something known as an HDAC inhibitor. This was a relatively simple chemical compound that is present in royal jelly and that stops the action of HDAC enzymes that normally remove acetyl tags from histones. This results in a build-up of acetyl tags in the cells of the bee embryos, and like the reduction in DNA-methyl groups described previously, this is thought to switch on key genes required for development of a queen. Without the HDAC inhibitor in the royal jelly, the larvae follow a “default” set of genetic instructions and develop into workers.
HDAC inhibitors are not only important to queen bees, but are also part of a small but growing number of medically useful drugs that target epigenetic tags and which are useful in treating some kinds of cancer. Furthermore HDACs also have a role in the way our brains form memories, and novel drugs that affect histone acetylation may have a role in the future in treating memory impairment in elderly patients.

The environment and epigenetics

We have seen how the difference between a queen and worker bee is determined by exposure to a chemical that directly alters epigenetic tags such as acetyl groups but are there examples where nutrition or other aspects of the environment affect human populations in a way that can be explained by epigenetics? Obviously we can’t do experiments on human populations as we can on microscopic worms or bees, but sometimes human history or natural phenomena do it for us. One such example is what is known as the Dutch Hunger Winter. In the last year of the Second World War in Europe, a food embargo imposed by occupying German forces on the civilian population of the Netherlands resulted in a severe famine, coinciding with a particularly harsh winter. About 20,000 people died from starvation as rations dropped to below 1000 kilocalories per day. Despite the chaos of war, medical care and records remained intact allowing scientists to subsequently study the effect of famine on human health. What they found was that children who were in the womb during the famine experienced a life-long increase in their chances of developing various health problems compared to children conceived after the famine. The most sensitive period for this effect was the first few months of pregnancy. Thus, something appears to happen early in development in the womb that can affect the individual for the rest of their lives.

Epigenetic effects can sometimes pass to grandchildren

Even more surprisingly, some data seems to suggest that grandchildren of women who were pregnant during the Hunger Winter experience some of these effects. From what we have already discussed, this strongly suggests an epigenetic mechanism. In fact, research with the Dutch Hunger Winter families continues, and a recent study looking at a gene galled IGF2 found lower levels of the methyl tag in the DNA of this gene in individuals exposed to the famine before birth. Although IGF2 may not itself be involved in the increased risk of poor health in these people, it shows that epigenetic effects (i.e. reduction of the number of methyl tags on particular genes) that are produced before birth can last for many decades. Studies in animals have also found that the diet of the mother can have effects on her offspring. For example, feeding sheep a diet lacking the types of food required to make methyl groups leads to offspring with altered patterns of DNA methylation and which have higher than expected rates of certain health problems.

Epigenetics and imprinting, why genes from Mum and Dad are not always equivalent

We all have 23 pairs of chromosomes in our cells. For each pair, one came from mother and one from father. Thus, we inherit one copy of each gene from each parent and we generally assume that the function of the gene does not to depend on which parent it came from. However, for imprinted genes things are different. For these genes, either the maternal or paternal copy of the gene is active, while the other one is kept silent. There are at least 80 imprinted genes in humans and mice, many of which are involved in growth of the embryo or the placenta. How can one copy of a gene be switched off, while the other copy in the same cell is switched on? The answer is epigenetics. Probably the most studied imprinted gene is IGF2(see above). One part of IGF2 operates as a switch. If the DNA is methylated here the IGF2 gene can be expressed. The switch is only methylated in Dad’s copy of the gene and so only this copy is expressed, while the maternal copy is silent. This switch is thought to be set up in the gametes (eggs and sperm) so right from the start, genes received from Mum and those from Dad are labelled differently with epigenetic tags and so are not equivalent.

Imprinting and mental disorders

Angelmann and Prader-Willi syndromes are two distinct genetic conditions with different symptoms, both caused by loss of a part of chromosome 15. Children who inherit one copy of this faulty chromosome develop either Angelmann or Prader-Willi syndrome, despite having a normal copy of the chromosome from their other parent. So how does the same mutation (loss of part of chromosome 15) lead to these two different conditions? The answer lies in the discovery that this particular piece of chromosome 15 contains a number of genes that are imprinted, so only the paternal or maternal copy of these gene are expressed which of the two syndromes appears depends on whether the deletion was in the maternal or paternally inherited chromosome. When the faulty chromosome is inherited from Dad, there is no functional copy of the imprinted genes that are switched off on the maternal chromosome 15 and the result is Angelmann syndrome and vice versa for Prader-Willi syndrome. This is quite unlike most genetic conditions such as cystic fibrosis, where an effect on development or health is only seen when a mutated gene or genes is inherited from both parents.

Boys versus Girls, how to switch off a whole chromosome

A bit of genetics that most of us know about is what makes a boy a boy, and a girl a girl. It’s the X and Y chromosomes. At the very beginning of our existence each of us received one X chromosome from our Mums via the egg, and while the girls received another X chromosome from their dads, via the sperm, the boys got a Y chromosome. The Y chromosome in the cells of a male embryo directs it to develop into a boy, while with two X and no Y chromosome the female embryo develops into a girl. Now, you might notice that there is an imbalance here. We all have two each of all the other chromosomes, but for the sex chromosomes (X and Y) the girls have two Xs while the boys only have one X (and a Y). While the Y chromosome contains few genes, mostly involved in “maleness”, the X chromosome contains quite a few genes involved in important processes such as colour vision, blood clotting and muscle function. In order to even up the “dosage” of X chromosome genes between male and female cells, one entire X chromosome is switched off in female cells. This is called X-chromosome inactivation and happens very early in the womb. In this process cells randomly switch off either the paternal or maternal X chromosome, so that when a girl baby is born her body is a mixture or chimera of cells where either the maternal or paternal X-chromosome is switched off. The way that this happens involves the type of epigenetic tags that we have discussed and it has been known for decades that female cells contain one very compact X chromosome called the Barr body that can be seen under the microscope, and this is the inactive X chromosome.

The case of the tortoiseshell cat

We are probably all familiar with tortoiseshell cats and their mottled coats with patches of orange and black fur. What you might not know is that almost all cats with this type of coat are female! The reason for this is that a gene for coat colour is located on the cat’s X chromosome. There are two versions of this gene, called “O” and “o” one gives ginger fur and the other black. Two copies of the same version in a female cat results in ginger or black fur respectively, but one copy of each gives a tortoiseshell effect. This is down to X-chromosome inactivation. The skin of these cats is composed of patches of cells where either the maternal or paternal X chromosome is inactivated. This results in skin with the O gene switched on and o silenced in some patches (orange fur) and o gene on and O silenced in other patches (black fur), hence the tortoiseshell pattern. Since the male cats only have one X chromosome, and no X-chromosome inactivation, they are either orange or black all over.

Epigenetic inheritance, can epigenetic states be passed from one generation to the next?

As we have seen from the roundworm example, epigenetic effects (in this case extended lifespan) can sometimes be passed from one generation to the next, although the effects only seem to last for a few generations. Are there examples where epigenetic effects carry over to subsequent generations in humans or other mammals? There is some evidence that the effects of the Dutch Hunger Winter affected grandchildren of women who were pregnant during the famine. Similarly, in a study of a 19th century northern Swedish population who underwent cycles of famine and plenty, the amount of food available appears to have affected the health and longevity of the next generation.

Hair colour in mouse can be determined by an epigenetic effect

Perhaps the best known example of transgenerational epigenetic effects is provided by the mouse Agouti gene. This gene controls hair colour, and is switched on at just the right time in hair follicle cells to produce a yellow stripe in the otherwise dark hairs, resulting in what is called an agouti coat. But mice with a particular variant of the Agouti gene called Avy have coats that are anywhere between yellow and the normal dark (agouti) pattern of wild-type mice. The yellow mice also become obese and suffer other health problems. So the Avy gene seems to have a variable effect (in fact the Avy stands for Avariable yellow). How this works has puzzled geneticists for years, but we can now recognise this as an epigenetic effect. The yellow fur occurs because Avy version of the Agouti gene has faulty controls and is switched on all the time. However, methyl tags are often added to the faulty control DNA sequence and this tends to switch the gene off, resulting in mottled or dark agouti fur in individual mice. Pups born to dams with the Avy gene range in colour from yellow to dark, but the proportion depends on the coat colour of the mother litters of dark (agouti) females are more likely to contain dark pups. Furthermore, a higher proportion of dark offspring is observed if both the mother and the grandmother have the dark colouration. So the agouti colouration, which is determined epigenetically (by the number of methyl tags on the Avy gene) can to some extent, carry through from one generation to the next.

Eggs and sperm do not usually ‘carry over’ epigenetic effects

Although we can find cases where epigenetic effects apparently last from parents to offspring, this is not usually the case and almost all of the epigenetic switches or marks are reset in germ cells (eggs and sperm) and in the very earliest stages of development of an embryo. In fact if this wasn’t the case, the amazing development of a fertilised egg into a fully formed creature would be impossible.

Getting from a fertilized egg to a fully formed human, it’s all in the (epi) genome

So far we have described some specific cases of epigenetic regulation, but we now know that epigenetics in its broad sense, (how genes are expressed and used, rather than the DNA sequence of the genes themselves) is central to how a fertilised egg can eventually give rise to a whole organism and how cells of, let’s say your skin, remain skin cells and are different from your brain cells, despite containing exactly the same genes. Shortly after fertilisation, a developing human embryo consists of a ball of cells called embryonic stem cells. Each of these cells has the capacity to give rise to any of the types of cells in the body as the embryo grows (for example, brain cells, skin cells or blood cells). By contrast, 9 months later when a baby is born, most of the cells making up his or her body are committed to be a specific type of cell with specific functions. So as the cells divide, the ball of embryonic stem cells gradually develops into all the cell types and structures of the baby at term. For this to happen, thousands of genes must be switched on or off at just the right times and in the right cells as an embryo grows. For example, genes that make the fibrous keratin protein that gives our skin its strength, are only switched on in skin cells and not in the developing brain and genes required for brain cells to develop and make their interconnections are on in the brain but not in the skin.

During development genes have to be switched ‘on’ and ‘off’. Epigenetic tags help with this

A very big area of research today concerns how all this gene switching on an off works, and a large part of this process uses the epigenetic chemical tags, especially acetyl and methyl histone tags. In order for those embryonic stem cells to be able to give rise to all of the other types of cells, their epigenetic switches are (almost) completely reset compared to adult cells. I have put “almost” in brackets as we know from imprinted genes and transgenerational epigenetic inheritance that there are exceptions.

Epigenetics, Dolly the sheep and friends

In February 1997, a sheep called Dolly became the most famous example of her species, briefly even becoming a TV celebrity. The reason for her fame is that she was the first mammal to be “created” by a process called somatic cell nuclear transfer, or in other words the first man-made clone (man-made to be distinct from identical twins, who are natural clones). The process leading to her birth required a mature oocyte (a unfertilised egg) from one female sheep and an ordinary cell from the udder of a second sheep. First the nucleus (the part containing the DNA) was removed from the oocyte. This was done using a special microscope as although oocytes are quite big compared to other cells, they are still too small to see with the naked eye. Then the nucleus from the udder cell was inserted into enucleated oocyte. Thus, Dolly had three “mothers”: the donor of the oocyte, the donor of the udder cell and the sheep that carried the developing embryo to term. No father was involved. Although this process was, and remains, very inefficient it was the first proof that the genes from an adult mammalian cell can be “epigenetically reprogrammed” back to the state of the embryonic stem cells that can develop into any other type of cell. Subsequently the same process has been applied to other species and may have medical uses in generating cells that could repair tissues damaged by injury or disease.

Summary: the epigenome and the ENCODE project – the “Large Hadron Collider” of Biology

Whereas the term “genome” refers to the entire DNA sequence of an organism (three billion letters of it for humans), the epigenome refers to the entire pattern of epigenetic modifications across all genes, including methyl DNA tags, methyl histone tags, acetyl histone tags and other chemical tags that we have not mentioned, in each cell type of an organism. This represents an almost unimaginable amount of information, dwarfing even the human genome project. Nevertheless, knowledge of the epigenome is essential to fully answer some of the biggest questions in biology such as: how do we develop from a ball of identical cells into a whole organism? why do we age? and how can we better understand diseases such as cancer? Not surprisingly then, epigenetics and the epigenome is a big area of research. Some of the research in this field is encompassed by the ENCODE (Encyclopedia of DNA Elements) project, an ongoing venture to identify patterns of epigenetic tags in many different types of cells for the entire human genome ( The ENCODE project is sometimes likened to the Large Hadron Collider or LHC in Switzerland. The LHC is the largest piece of scientific equipment ever built and the experiments physicist conduct with it aim to probe the fundamental details of the matter that makes up our Universe. Although biologists don’t have (or need) such a spectacular piece of kit for their research, the effort to examine the intricacy of the human epigenome has been likened to the LHC project because of its scale, complexity and the amount of information being created.

Epigenetic errors

Epigenetics is an area where our scientific knowledge is rapidly increasing. One thing that scientists have discovered is that epigenetic errors are common in diseases such as cancer and in ageing cells. As a result, scientists are developing medicines that target faulty epigenomes and one of the first examples is the use of HDAC inhibitors, similar to the compound found in royal jelly. From the study of strange patterns of inheritance such as genetic imprinting, the yellow/agouti Avy mouse, the all-female tortoiseshell cat population and other related phenomena biologists have uncovered a whole new layer of information that lies “on top” of the DNA sequence of our genes. These new discoveries explain these previous puzzling observations, but also have great potential for new understanding and treatments for human disease.


Cytosine DNA methylation (5-methylcytosine (5mC)) is an epigenetic mark that is widespread in both animals and plants, and appears to play important roles in various biological processes, such as gene silencing and imprinting. Recently, studies have shown that embryonic stem cells (ESCs) and Purkinje neurons contain high levels of 5-hydroxymethylcytosine (5hmC) [1, 2]. Human TET1, a 2-oxoglutarate- and Fe(II)-dependent enzyme, has been shown to catalyze the conversion of 5mC to 5hmC both in vitro and in vivo [1]. Subsequently, all mouse Tet proteins, Tet1, Tet2 and Tet3, were shown to be able to convert 5mC to 5hmC [3]. Disruption in human TET1 and TET2 is associated with diseases such as MLL-associated leukemia [4] and myeloproliferative disorders [5]. Studies have suggested that 5hmC inhibits the methyl-CpG-binding protein MeCP2 from binding DNA [6]. In addition to the exclusion of methyl-CpG-binding proteins, 5hmC may recruit unknown 5hmC binding protein(s). Moreover, because the DNA methyltransferase DNMT1 binds poorly to 5hmC [1, 7], it is possible that 5hmC plays a role in excluding DNMT1 from methylating cytosines and thus may promote DNA demethylation. Importantly, 5hmC diminishes as embryonic stem cells (ESCs) differentiate, suggesting that 5hmC may play specific roles in ESCs. Indeed, mouse Tet1 has been shown to be required for ESC maintenance [3]. The function of 5hmC in mammals remains poorly understood. To further understand the role of 5hmC, it is necessary to understand where 5hmC localizes in the genome. Very recently, a genome-wide map of 5hmC was reported in mouse cerebellum [8]. 5hmC was chemically tagged and affinity enriched, and the purified DNA was sequenced. The authors found that 5hmC is enriched over genes and is positively correlated with expression levels [8].

Recently, commercial antibodies specific to 5hmC have become available. While these antibodies specifically recognize 5hmC, it is important to note that they tend to prefer densely 5-hydroxymethylated sites to single 5hmC sites (Figure S1 in Additional file 1). Here we generated genome-wide maps of 5hmC in human ESCs (hESCs) by performing hydroxymethyl-DNA immunoprecipitation followed by massively parallel sequencing with an Illumina Genome Analyzer (hmeDIP-seq). As did Song et al. [8], we found that a large fraction of 5hmC peaks were enriched over genes. However, we also found that 5hmC is enriched over predicted hESC enhancers, further suggesting a potential role of 5hmC in gene regulation. Moreover, we observed enrichment of 5hmC peaks with transcription binding sites such as those of pluripotency factors OCT4 and NANOG. In addition, we found that 5hmC regions correspond to genomic regions that are GC-skewed.


Non-ionising radiations, electromagnetic fields (EMF) such as radiofrequency (RF), or power frequency radiation have become very common in everyday life. All of these exist as low frequency radiation which can come from wireless cellular devices or through electrical appliances which induce extremely low frequency radiation (ELF). Exposure to these radioactive frequencies has shown negative affects on the fertility of men by impacting the DNA of the sperm and deteriorating the testes [2] as well as an increased risk of tumor formation in salivary glands. [3] [4] The International Agency for Research on Cancer considers RF electromagnetic fields to be possibly carcinogenic to humans, however the evidence is limited. [5]

Radiation and medical imaging Edit

Advances in medical imaging has resulted in increased exposure of humans to low doses of ionizing radiation. Radiation exposure in pediatrics has been shown to have a greater impact as children's cells are still developing. [2] The radiation obtained from medical imaging techniques is only harmful if consistently targeted multiple times in a short space of time. Safety measures have been introduced in order to limit the exposure of harmful ionizing radiation such as the usage of protective material during the use of these imaging tools. A lower dosage is also used in order to fully rid the possibility of a harmful effect from the medical imaging tools. The National Council on Radiation Protection and Measurements along with many other scientific committees have ruled in favor of continued use of medical imaging as the reward far outweighs the minimal risk obtained from these imaging techniques. If the safety protocols are not followed there is a potential increase in the risk of developing cancer. This is primarily due to the decreased methylation of cell cycle genes, such as those relating to apoptosis and DNA repair. The ionizing radiation from these techniques can cause many other detrimental effects in cells including changes in gene expression and halting the cell cycle. However, these results are extremely unlikely if the proper protocols are followed. [1] [4]

Target theory concerns the models of how radiation kills biological cells and is based around two main postulates:

  1. "Radiation is considered to be a sequence of random projectiles
  2. the components of the cell are considered as the targets bombarded by these projectiles" [6]

Several models have been based around the above two points. From the various proposed models three main conclusions were found:

  1. Physical hits obey a Poisson distribution
  2. Failure of radioactive particles to attack sensitive areas of cells allow for survival of the cell
  3. Cell death is an exponential function of the dose of radiation received as the number of hits received is directly proportional to the radiation dose all hits are considered lethal [7]

Radiation exposure through ionizing radiation (IR) affects a variety of processes inside of an exposed cell. IR can cause changes in gene expression, disruption of cell cycle arrest, and apoptotic cell death. The extent of how radiation effects cells depends on the type of cell and the dosage of the radiation. Some irradiated cancer cells have been shown to exhibit DNA methylation patterns due to epigenetic mechanisms in the cell. In medicine, medical diagnostic methods such as CT scans and radiation therapy expose the individual to ionizing radiation. Irradiated cells can also induce genomic instability in neighboring un-radiated cells via the bystander effect. Radiation exposure could also occur via many other channels than just ionizing radiation.

The basic ballistic models Edit

The single-target single-hit model Edit

In this model a single hit on a target is sufficient to kill a cell [7] The equation used for this model is as follows:

Where k represents a hit on the cell and m represents the mass of the cell.

The n-target single-hit model Edit

In this model the cell has a number of targets n. A single hit on one target is not sufficient to kill the cell but does disable the target. An accumulation of successful hits on various targets leads to cell death. [7] The equation used for this model is as follows:

Where n represents number of the targets in the cell.

The linear quadratic model Edit

The equation used for this model is as follows: [7]

where αD represents a hit made by a one particle track and βD represents a hit made by a two particle track and S(D) represents the probability of survival of the cell.

The three lambda model Edit

This model showed the accuracy of survival description for higher or repeated doses. [7]

The equation used for this model is as follows:

The linear-quadratic-cubic model Edit

The equation used for this model is as follows: [7]

Sublesions hypothesis models Edit

The repair-misrepair model Edit

This model shows the mean number of lesions before any repair activations in a cell. [7]

The equation used for this model is as follows:

where Uo represents the yield of initially induced lesions, with λ being the linear self-repair coefficient, and T equaling time

The lethal-potentially lethal model Edit

This equation explores the hypothesis of a lesion becoming fatal within a given of time if it is not repair by repair enzymes. [7]

The equation used for this model is as follows:

T is the radiation duration and tr is the available repair time.

The saturable repair model Edit

This model illustrates the efficiency of the repair system decreasing as the dosage of radiation increases. This is due to the repair kinetics becoming increasingly saturated with the increase in radiation dosage. [7]

The equation used for this model is as follows:

n(t) is the number of unrepaired lesions, c(t) is the number of repair molecules or enzymes, k is the proportionality coefficient, and T is the time available for repair.

Radiation hormesis Edit

Hormesis is the hypothesis that low levels of disrupting stimulus can cause beneficial adaptations in an organism. [8] The ionizing radiation stimulates repair proteins that are usually not active. Cells use this new stimuli to adapt to the stressors they are being exposed to. [9]

Radiation-Induced Bystander Effect (RIBE) Edit

In biology, the bystander effect is described as changes to nearby non-targeted cells in response to changes in an initially targeted cell by some disrupting agent. [10] In the case of Radiation-Induced Bystander Effect, the stress on the cell is caused by ionizing radiation.

The bystander effect can be broken down into two categories, long range bystander effect and short range bystander effect. In long range bystander effect, the effects of stress are seen further away from the initially targeted cell. In short range bystander, the effects of stress are seen in cells adjacent to the target cell. [10]

Both low linear energy transfer and high linear energy transfer photons have been shown to produce RIBE. Low linear energy transfer photons were reported to cause increases in mutagenesis and a reduction in the survival of cells in clonogenic assays. X-rays and gamma rays were reported to cause increases in DNA double strand break, methylation, and apoptosis. [10] Further studies are needed to reach a conclusive explanation of any epigenetic impact of the bystander effect.

Formation of ROS Edit

Ionizing radiation produces fast moving particles which have the ability to damage DNA, and produce highly reactive free radicals known as reactive oxygen species (ROS). The production of ROS in cells radiated by LDIR (Low-Dose Ionizing Radiation) occur in two ways, by the radiolysis of water molecules or the promotion of nitric oxide synthesis (NOS) activity. The resulting nitric oxide formation reacts with superoxide radicals. This generates peroxynitrite which is toxic to biomolecules. Cellular ROS is also produced with the help of a mechanism involving nicotinamide adenosine dinucleotide phosphate (NADPH) oxidase. NADPH oxidase helps with the formation of ROS by generating a superoxide anion by transferring electrons from cytosolic NADPH across the cell membrane to the extracellular molecular oxygen. This process increases the potential for leakage of electrons and free radicals from the mitochondria. The exposure to the LDIR induces electron release from the mitochondria resulting in more electrons contributing to the superoxide formation in the cells.

The production of ROS in high quantity in cells results in the degradation of biomolecules such as proteins, DNA, and RNA. In one such instance the ROS are known to create double stranded and single stranded breaks in the DNA. This causes the DNA repair mechanisms to try to adapt to the increase in DNA strand breaks. Heritable changes to the DNA sequence have been seen although the DNA nucleotide sequence seems the same after the exposure with LDIR. [11]

Activation of NOS Edit

The formation of ROS is coupled with the formation of nitric oxide synthase activity (NOS). NO reacts with O2 − generating peroxynitrite. The increase in the NOS activity causes the production of peroxynitrite (ONOO-). Peroxynitrite is a strong oxidant radical and it reacts with a wide array of biomolecules such as DNA bases, proteins and lipids. Peroxynitrite affects biomolecules function and structure and therefore effectively destabilizes the cell. [11]

Mechanism of oxidative stress and epigenetic gene regulation Edit

Ionizing radiation causes the cell to generate increased ROS and the increase of this species damages biological macromolecules. In order to compensate for this increased radical species, cells adapt to IR induced oxidative effects by modifying the mechanisms of epigenetic gene regulation. There are 4 epigenetic modifications that can take place:

  1. formation of protein adducts inhibiting epigenetic regulation
  2. alteration of genomic DNA methylation status histone interactions affecting chromatin compaction
  3. modulation of signaling pathways that control transcription factor expression

ROS-mediated protein adduct formation Edit

ROS generated by ionizing radiation chemically modify histones which can cause a change in transcription. Oxidation of cellular lipid components result in an electrophilic molecule formation. The electrophilic molecule binds to the lysine residues of histones causing a ketoamide adduct formation. The ketoamide adduct formation blocks the lysine residues of histones from binding to acetylation proteins thus decreasing gene transcription. [11]

ROS-mediated DNA methylation changes Edit

DNA hypermethylation is seen in the genome with DNA breaks at a gene-specific basis, such as promoters of regulatory genes, but the global methylation of the genome shows a hypomethylation pattern during the period of reactive oxygen species stress. [12]

DNA damage induced by reactive oxygen species results in increased gene methylation and ultimately gene silencing. Reactive oxygen species modify the mechanism of epigenetic methylation by inducing DNA breaks which are later repaired and then methylated by DNMTs. DNA damage response genes, such as GADD45A, recruit nuclear proteins Np95 to direct histone methyltransferase's towards the damaged DNA site. The breaks in DNA caused by the ionizing radiation then recruit the DNMTs in order to repair and further methylate the repair site.

Genome wide hypomethylation occurs due to reactive oxygen species hydroxylating methylcytosines to 5-hydroxymethylcytosine (5hmC). [13] The production of 5hmC serves as an epigenetic marker for DNA damage which is recognizable by DNA repair enzymes. The DNA repair enzymes attracted by the marker convert 5hmC to an unmethylated cytosine base resulting in the hypomethylation of the genome. [14]

Another mechanism that induces hypomethylation is the depletion of S-adenosyl methionine synthetase (SAM). The prevalence of super oxide species causes the oxidization of reduced glutathione (GSH) to GSSG. Due to this, synthesis of the cosubstrate SAM is stopped. SAM is an essential cosubtrate for the normal functioning of DNMTs and histone methyltrasnferase proteins.

ROS-mediated post-translation modification Edit

Double stranded DNA breaks caused by exposure to ionizing radiation are known to alter chromatin structure. Double stranded breaks are primarily repaired by poly ADP (PAR) polymerases which accumulate at the site of the break leading to activation of the chromatin remodeling protein ALC1. ALC1 causes the nucleosome to relax resulting in the epigenetic up-regulation of genes. A similar mechanism involves the ataxia telangiectasia mutated (ATM) serine/threonine kinase which is an enzyme involved in the repair of double stranded breaks caused by ionizing radiation. ATM phosphorylates KAP1 which causes the heterochromatin to relax, allowing increased transcription to occur.

The DNA mismatch repair gene (MSH2) promoter has shown a hypermethylation pattern when exposed to ionizing radiation. Reactive oxygen species induce the oxidization of deoxyguanosine into 8-hydroxydeoxyguanosine (8-OHdG) causing a change in chromatin structure. Gene promoters that contain 8-OHdG deactivate the chromatin by inducing trimethyl-H3K27 in the genome. Other enzymes such as transglutaminases (TGs) control chromatin remodeling through proteins such as sirtuin1 (SIRT1). TGs cause transcriptional repression during reactive oxygen species stress by binding to the chromatin and inhibiting the sirtuin 1 histone deacetylase from performing its function. [11]

ROS-mediated loss of epigenetic imprinting Edit

Epigenetic imprinting is lost during reactive oxygen species stress. This type of oxidative stress causes a loss of NF- κB signaling. Enhancer blocking element CCCTC-binding factor (CTCF) binds to the imprint control region of insulin-like growth factor 2 (IGF2) preventing the enhancers from allowing the transcription of the gene. The NF- κB proteins interact with IκB inhibitory proteins, but during oxidative stress IκB proteins are degraded in the cell. The loss of IκB proteins for NF- κB proteins to bind to results in NF- κB proteins entering the nucleus to bind to specific response elements to counter the oxidative stress. The binding of NF- κB and corepressor HDAC1 to response elements such as the CCCTC-binding factor causes a decrease in expression of the enhancer blocking element. This decrease in expression hinders the binding to the IGF2 imprint control region therefore causing the loss of imprinting and biallelic IGF2 expression. [11]

After the initial exposure to ionizing radiation, cellular changes are prevalent in the unexposed offspring of irradiated cells for many cell divisions. One way this non-Mendelian mode of inheritance can be explained is through epigenetic mechanisms. [11]

Ionizing radiation and DNA methylation Edit

Genomic instability via hypomethylation of LINE1 Edit

Ionizing radiation exposure affects patterns of DNA methylation. Breast cancer cells treated with fractionated doses of ionizing radiation showed DNA hypomethylation at the various gene loci dose fractionation refers to breaking down one dose of radiation into separate, smaller doses. [15] Hypomethylation of these genes correlated with decreased expression of various DNMTs and methyl CpG binding proteins. LINE1 transposable elements have been identified as targets for ionizing radiation. The hypomethylation of LINE1 elements results in activation of the elements and thus an increase in LINE1 protein levels. Increased transcription of LINE1 transposable elements results in greater mobilization of the LINE1 loci and therefore increases genomic instability. [11]

Ionizing radiation and histone modification Edit

Irradiated cells can be linked to a variety of histone modifications. Ionizing radiation in breast cancer cell inhibits H4 lysine tri-methylation. Mouse models exposed to high levels of X-ray irradiation exhibited a decrease in both the tri-methylation of H4-Lys20 and the compaction of the chromatin. With the loss of tri-methylation of H4-Lys20, DNA hypomethylation increased resulting in DNA damage and increased genomic instability. [11]

Loss of methylation via repair mechanisms Edit

Breaks in DNA due to ionizing radiation can be repaired. New DNA synthesis by DNA polymerases is one of the ways radiation induced DNA damage can be repaired. However, DNA polymerases do not insert methylated bases which leads to a decrease in methylation of the newly synthesized strand. Reactive oxygen species also inhibit DNMT activity which would normally add the missing methyl groups. This increases the chance that the demethylated state of DNA will eventually become permanent. [16]

Epigenetic affects on a developing brain Edit

Chronic exposure to these types of radiation can have an effect on children from as early as when they are fetuses. There have been multiple cases reported of hindrance in the development of the brain, behavioral changes such as anxiety, and the disruption of proper learning and language processing. An Increase in the cases of ADHD behavior and autism behavior has been shown to be directly correlated with the exposure of EMF waves. The World Health Organization has classified RFR as a possible carcinogen for its epigenetic effects on DNA expression. The exposure to EMF waves on a consistent 24hr basis has shown to lower the activity of miRNA in the brain affecting developmental and neuronal activity. This epigenetic change causes the silencing of necessary genes along with the change in expression of other genes integral for the normal development of the brain. [2]

MGMT- and LINE1- specific DNA methylation Edit

DNA methylation influences tissue responses to ionizing radiation. Modulation of methylation in the gene MGMT or in transposable elements such as LINE1 could be used to alter tissue responses to ionizing radiation and potentially opening new areas for cancer treatment.

MGMT serves as a prognostic marker in glioblastoma. Hypermethylation of MGMT is associated with the regression of tumors. Hypermethylation of MGMT silences its transcription inhibiting alkylating agents in tumor killing cells. Studies have shown patients who received radiotherapy, but no chemotherapy after tumor extraction, had an improved response to radiotherapy due to the methylation of the MGMT promoter.

Almost all human cancers include hypomethylation of LINE1 elements. Various studies depict that the hypomethylation of LINE1 correlates with a decrease in survival after both chemotherapy and radiotheraphy.

Treatment by DNMT inhibitors Edit

DMNT inhibitors are being explored in the treatment of malignant tumors. Recent in-vitro studies show that DNMT inhibitors can increase the effects of other anti-cancer drugs. Knowledge of in-vivo effect of DNMT inhibitors are still being investigated. The long term effects of the use of DNMT inhibitors are still unknown. [16]


We thank Elin Grundberg for providing the probe-specific h 2 estimates from [9], and Maria Landi (senior author of Shi et al. [23]) for sharing with us their estimates of the proportion of methylation variance explained by neighboring SNPs. Cases and their vital status were ascertained through the Victoria Cancer Registry (VCR) and the Australian Institute of Health and Welfare (AIHW), including the national Death Index and the Australian Cancer Database. We thank Adam Gillum (Baylor College of Medicine) for assistance with the figures.


TEV was supported by training grant award RP 140113 from the Cancer Prevention & Research Institute of Texas (CPRIT). CC was partially supported by CPRIT grant RP170005. The ENID Trial was jointly funded by the UK Medical Research Council (MRC) and the Department for International Development (DFID) under the MRC/DFID Concordat agreement (MRC Programme MC-A760-5QX00). MNR, YYG, SEM, and ZH were supported by the Bill and Melinda Gates Foundation Grand Challenge “Achieving Healthy Growth” scheme (grant OPP1 066947). ZH and AG are supported by grants from the Institut National du Cancer (INCa, France) and Association pour la Recherche sur le Cancer (ARC, France). Gustave Roussy, Villejuif, Cohort recruitment was funded by VicHealth and Cancer Council Victoria. This work was supported by the Australian National Health and Medical Research Council (NHMRC) [grants 1088405and 1074383]. MCCS cohort recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further supported by Australian NHMRC grants 209057 and 396414 and by infrastructure provided by Cancer Council Victoria. The nested case -control methylation studies were supported by the NHMRC [grants 1011618, 1026892, 1027505,1050198, 1043616]. MCS is a NHMRC Senior Research Fellow. JLH is a NHMRC Senior Principal Research Fellow. RAW was supported by grants from the USDA (CRIS 3092-5-001-059), CPRIT (RP170295), and UK-MRC (MR/M01424X/1). None of the funding bodies played a role in study design or data collection, analysis, or interpretation.

Availability of data and materials

Regarding the ENID season of conception analysis, the complete HM450 data sets have been deposited in GEO (GSE99863) [30]. Regarding the Melbourne Collaborative Cohort Study, the complete data set used in this analysis (ESS and negative control probe M values for all seven case-control cohorts, and associated covariates) are available at[32] and on Figshare at [79].

Materials and methods

Microarray design

The DNA tiling array was designed in conjunction with Roche NimbleGen to tile a portion of the P. falciparum genome with a targeted median probe spacing of 12 bp. The 3D7 assembly (PlasmoDB v5.5) was used as the reference sequence. The design targeted all of chromosomes 2, 4, 7 and 9 and partial chromosomes 3:106138-147339, 5:947885-end, and 12:start-66805 (Additional file 1). The three chromosomes tiled in their entirety were selected at random. However, the four partially tiled chromosomes represent regions of particular interest given our hypothesis that lncRNAs may be involved in chromatin remodeling and clinically important parasite processes. Probes were variable length Tm-matched long oligonucleotides, averaging 55 bp each. Probe sequences were screened for excessive cross hybridization to 3D7 sequence: any probes with more than five close Sequence Search and Alignment by Hashing Algorithm (SSAHA) matches were eliminated [70, 71]. The final design filled 366,479 probes on the array, 96.81% of which are unique. The raw and normalized data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus and are accessible through accession [GEO:GSE27937].

Parasite culture, RNA preparation, and cDNA labeling for microarray hybridizations

A clone of P. falciparum strain 3D7 was cultured using standard methods [72, 73] and total RNA isolated from each sample as described [74]. Total RNA was cleaned up with an RNeasy column (Qiagen, Valencia, CA, USA) and concentrated in a Microcon YM-30 centrifugal filter (Millipore, Billerica, MA, USA). Total RNA (1 μg) was then subjected to poly-A selective amplification using Message Amp II (Ambion, Foster City, CA, USA), substituting biased dNTP/NTP mixes (2A/T/U:1C/G) for the solutions provided. The resulting aRNA was labeled with Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) using random hexamers, and either Cy3- or Cy5-dUTPs (GE Healthcare, Piscataway, NJ, USA) for 2 hours at 42°C following a 10-minute primer annealing step at 65°C. The reaction was concentrated on a Microcon YM-30 column and subjected to array hybridization per standard NimbleGen protocol.

Data normalization and quality control

Raw data from each sample was quantile normalized [75] and log2 transformed prior to prediction of TARs along the P. falciparum genome. We based our pre-processing pipeline on established quality control metrics: removal of non-biological variation and a strong correlation between raw and normalized data [76]. Additional file 15 shows log2(intensity) distributions of each sample before and after quantile normalization, log2(intensity) boxplots of each sample after quantile normalization, and all pair-wise correlation scatterplots of data before and after normalization. Pearson correlation is equal to 1.0 between raw and normalized data in all matched samples. Additional file 16 shows that, for each sample, we observed only a minimal increase in median intergenic probe hybridization intensity with number of G+C bases. Given that we make no absolute or quantitative expression comparisons of transcripts (only relative expression comparisons of the same transcript across time-points) and we confirm no G+C content bias in predicted lncRNAs, we deemed this inconsequential. Normalized data were median centered (at zero) prior to expression profiling and data browsing in Integrated Genomics Viewer [77].

Detection of TARs from tiling arrays

We wrote and implemented an iterative sliding window algorithm to scan each sample's normalized probe hybridization intensity values for statistically significant TARs. Specifically, we used a single-step maxT permutation procedure (1,000 permutations) to transform the mean probe intensity score 'T' calculated in each of the approximately 366,000 possible window slides along normalized data into a multiple-hypothesis adjusted P-value [29–31]. We then discarded windows with adjusted P-values greater than 0.05 to control the family-wise error-rate of windows predicted to be significant at 5%. We repeated this procedure a total of 28 times using window sizes of 5, 10, 15, 20, 25, 30, and 40 probes. Next, we intersected all significant windows with PlasmoDB v6.5 gene annotations, and merged overlapping annotated windows to define the boundaries of 1,229 protein-coding TARs and 8 ribosomal RNA TARs. Similarly, we merged overlapping windows that did not overlap any known or predicted gene to define the boundaries of 123 un-annotated TARs. BEDTools v4 was used for all data intersections and unions [78]. Additional file 2 lists predicted TAR coordinates.

Filtering un-annotated TARs

We filtered un-annotated TARs by setting a minimum length criterion of 200 bp and ensuring no BLASTX predicted coding potential. Out of 123 predicted un-annotated TARs, 46 were under 200 bp in length (Additional file 5). We retrieved FASTA sequence for the remaining 77 from PlasmoDB v6.5, and used the NCBI BLASTX web server to search for any significant protein matches [79]. Default BLASTX settings (BLOSUM62, word size 3, low complexity filtering, and so on) were used except the Expect threshold for reporting match significance (that is, coding potential) was set at 0.01. Seventeen sequences with an Expect score < 0.01 were categorized as putative novel P. falciparum genes or pseudogenes and were excluded from further lncRNA characterization (Additional file 6). We searched both the Swissprot and Non-redundant protein sequence (nr) databases with the following organism queries: all organisms, Plasmodium, Plasmodium falciparum, and Plasmodium falciparum strain 3D7.

Gene Ontology term analysis of stage-specific genes

Gene Ontology (GO) term analysis of stage-specific genes was performed using GOstat with default settings [80] and Sanger GeneDB P. falciparum gene annotations. Stage-specific genes were determined by intersection of PlasmoDB v6.5 gene annotations with protein-coding transcripts maximally expressed in each time-point. We looked for overrepresented GO terms in stage-specific genes versus the 1,360 protein-coding genes covered by the array. Additional file 4 lists all genes covered by the array, stage-specific genes, and the top four most overrepresented GO terms in each time-point.

Evolutionary sequence conservation

FASTA sequence for the 60 putative lncRNAs and 8 ribosomal RNA transcripts was retrieved from the 3D7 reference sequence (PlasmoDB v7.1). We also downloaded genomic FASTA sequence from PlasmoDB v7.1 representing all eight sequenced or partially sequenced Plasmodium species (P. falciparum, P. reichenowi, P. gallinaceum, P. knowlesi, P. vivax, P. berghei, P. yoelli, and P. chaubadi). We searched for sequence conservation using BLASTN (WU-BLAST 2.0 MP-WashU (4 May 2006)) using the same low-complexity filtering and context parameters as the PlasmoDB v7.1 BLAST server (-filter seg -ctxfactor 2.00) and setting the Expect threshold for significance to 0.01. We recorded the lowest BLASTN P-value within each species (Additional file 7, columns s to z).

The broad conservation of lncRNA-TARE-4L across all eight Plasmodium species was determined to be significant by null permutations. We chose 600 random intergenic regions from the 3D7 reference genome (based on v.7.1 annotation). These intergenic regions were sized to match the length distribution of the 60 putative lncRNAs, and were included in the WU-BLAST search. Out of 600 random intergenic regions, we found only 27 to be conserved across all 8 species, yielding an empirical P-value of 0.045.

Expression profiling

To profile each predicted TAR, we calculated its expression in each time-point as the mean hybridization intensity of probes tiling within or up to 25 bp on either side of the predicted TAR start and stop coordinates. The expression profile of each TAR was then mean centered across time-points and visualized using a non-hierarchical clustering dimension reduction algorithm. Specifically, we used non-metric multi-dimensional scaling (nMDS) as implemented in the R-project 'NeatMap' package to order rows and preserve data topology. In development and validation of the 'NeatMap' package, Rajaram and Oono [81] have similarly applied nMDS to visualize yeast cell cycle expression data. We point the reader to Figure 1d of Rajaram and Oono [81] and Figure 4 of Taguchi and Oono [82] for examples and thorough discussion of the utility of nMDS in determining relational patterns of gene expression. Notably, because nMDS is a non-linear numerical optimization technique, multiple ordinations were run to select the optimal solution.

We conducted a detailed comparison of lncRNA and protein-coding expression, finding lncRNA candidates to be expressed on par with protein-coding transcripts. Included in Additional file 3 are additional visualizations of protein-coding transcript versus putative lncRNA expression, including standard heatmaps and nMDS ordinated heatmaps without mean centering across time-points. Additional file 3 also provides a histogram of the maximum expression values for lncRNA candidates and protein-coding transcripts. Notably, we found 30 lncRNA candidates (50%) to be induced by greater than two-fold across our time course samples (Additional file 7, column k). By comparison, 309 of 1,229 protein-coding transcripts (25%) match this criterion.

Nearest-neighboring genes

Nearest neighboring genes to the set of 60 putative lncRNAs were extracted using the Cistrome Analysis Pipeline and PlasmoDB v6.5 gene coordinates.

Correlation analysis

To infer putative lncRNA splicing or UTR relationships with neighboring coding genes, we measured the Pearson correlation between putative lncRNA and neighboring coding gene expression profiles (Additional file 7, column j). We conservatively defined the neighboring coding gene to be the highest correlated, expressed gene to either side of each putative lncRNA locus. We also examined a null distribution of correlations from adjacent pairs of coding genes. We found the 60 candidate lncRNAs to be enriched for high correlation to neighboring genes. 40 of these candidates were highly correlated (r > 0.9), whereas only 10 should be highly correlated as demonstrated by our null distribution (Additional file 9). We then further investigated the expression profiles of lncRNA candidates with r < 0.9 to ensure correlation values reflected biologically meaningful variation. We defined biologically meaningful variation as a greater than 0.5-fold change across time-points.

Mapping of homologous lncRNA-TARE sequences

We used the PlasmoDB v6.5 BLASTN web server to record coordinates for homologous lncRNA-TARE sequences based on the predicted lncRNA-TARE-4L sequence. We then retrieved FASTA sequence of the most telomere-proximal 50,000 bp on each chromosome end from PlasmoDB v6.5, and used JDotter [83] software to create DNA dotplots mapping the telomeric repeats, TAREs 1 through 5, Rep20, and the first predicted gene on each end. We placed each predicted lncRNA-TARE gene onto the dotplot maps to confirm that lncRNA-TARE maps to TARE 2 and the sequence between TARE 2 and TARE 3 on 22 chromosome ends. We then used Geneious to cluster (ClustalW) and investigate the conservation of lncRNA-TARE sequences.

Parasite culture for qRT-PCR analysis

Two independent biological replicate time courses were performed to validate and investigate lncRNA-TARE expression in more detail. For each time course, a freshly thawed P. falciparum strain 3D7 clone was cultured using standard methods [72] in human red blood cells at 4% hematocrit. RPMI-HEPES medium was supplemented with 5% human serum (O+) and 5% Albumax II (Invitrogen, Carlsbad, CA, USA). Cultures were initially synchronized using two 5% sorbitol solution treatments [73] spaced by 16 hours. To then obtain highly synchronized cultures, newly formed ring-stage parasites were selected for using 5% sorbitol solution treatments during the subsequent two re-invasion generations. Highly synchronized cultures were expanded and harvested at stage-specific time-points. Each harvested culture was centrifuged at 2,400 rpm in a Sorvall RT6000B, and packed red blood cells lysed using a 0.05% (final concentration) saponin solution. Liberated parasites were washed using phosphate-buffered saline (pH 7.4), pelleted at 13.2 rpm in a microcentrifuge, resuspended in 1 ml TRIZOL reagent, and stored at -80°C prior to RNA extraction.

RNA preparation for qRT-PCR and RACE analysis

TRIZOL-chloroform extraction was performed and the aqueous layer applied to an RNeasy column (Qiagen). On-column DNAse digestion was carried out for 30 minutes to remove genomic DNA. Eluted RNA was also treated with TURBO DNase (Ambion) and cleaned up on a second RNeasy column (Qiagen) to yield high-purity RNA samples.

QRT-PCR analysis

RNA (1 μg) from each time course sample was reverse transcribed using a random priming strategy (Applied Biosystems cDNA High Capacity Reverse Transcription kit Carlsbad, CA, USA) along with a minus reverse transcriptase control reaction for each sample to confirm genomic DNA removal. qPCR reactions were carried out using 800 nM of primers and Roche FastStart SYBR Green Master mix (Indianapolis, IN, USA). Primer annealing and extension (55°C/60 seconds) was carried out for 40 cycles on an Applied Biosystems 7900 instrument.

We used PCR Miner software [84] to calculate both the cycle threshold (Ct) of each qPCR reaction and the amplification efficiency of each primer pair. We then calculated the relative expression of each lncRNA-TARE gene in each time course sample by averaging technical replicates and using the reference gene PF08_0085 and reference time-point T30 (trophozoite) for normalization. The error of normalized expression ratios was calculated using the delta method, based on a truncated Taylor series expansion, to account for technical variability in both the target and reference gene measurements. Biological replicate experiments were analyzed in isolation and then normalized expression measurements were averaged. We used a Taylor limited expansion method to determine how error propagated in the average expression value.

QRT-PCR primer design

Primer pairs to amplify predicted lncRNA-TARE genes and the SPE2-binding protein PfSip2 (PFF0200c) were designed using Premier Biosoft International AlleleID 7.6 software (Palo Alto, CA, USA). AlleleID primer design software carries out highly specific primer design by BLAST searching sequences and masking redundant regions prior to primer design. We also independently verified primer specificity using BLASTN on the PlasmoDB v6.5 website, and ensured single amplicon melting curves and no primer dimer formation. We required primer pair amplification efficiency, as calculated by PCR Miner software [84], to be at least 90% to ensure reproducible results. We used the previously described housekeeping gene P08_0085 (ubiquitin conjugating enzyme 1) [74] to calculate all normalized relative gene expression ratios. lncRNA-TARE and PfSip2 primer sequences are listed in Additional file 11.

Rapid amplification of cDNA ends

We employed RNA ligase-mediated RACE following manufacturer specifications (Ambion) and using 10 μg of T40 ± 3 hpi RNA mixed 1:1 from two independent time course extractions. We used Premier Biosoft International AlleleID 7.6 software to design primers targeting 20 lncRNA-TARE loci (Additional file 11). To map the putative 5' cap, we used a nested priming strategy with primers spaced roughly 350 bp antisense to the target sequence. To map 3' termini, we used a semi-nested priming strategy using a single antisense primer to the target sequence and nested primers corresponding to the 3' adapter sequence. Notably, the 5' RACE outer primer is the reverse complement of the 3' RACE primer, ensuring capture of contiguous transcripts. Minus reverse transcriptase control reactions were included for 3' RACE.

Outer and inner 5' RACE PCR cycling was performed using SuperTaq Plus polymerase (Ambion) and the following cycling conditions: 94°C for 3 minutes, 5 cycles of 94°C for 30 seconds, 60°C for 30 seconds, 68°C for 3 minutes, 35 cycles of 94°C for 30 seconds, 55°C for 30 seconds, 68°C for 3 minutes, and a final extension for 10 minutes at 68°C. 3' RACE PCR cycling was analagous except denaturation was performed at 94°C for 15 seconds and extension was performed at 68°C for 8 minutes. PCR products were gel excised, purified using Qiagen MinElute Gel Extraction Cleanup columns, and cloned into the pCR-2.1TOPO vector (Invitrogen).

We sequenced 27 and 10 colonies corresponding to 3' and 5' RACE products, respectively, using Genewiz services and Geneious analysis software [85]. A total of 12 different lncRNA-TARE loci were unambiguously represented in sequenced 3' RACE products. The original chromosome and syntenic terminus coordinates on the left end of chromosome 4 for each sequenced RACE product are included in Additional file 12 along with a graphical alignment of each sequenced RACE product to the left end of chromosome 4 in Additional file 13. RACE products were trimmed to exclude any low-quality base calls and vector sequence beyond the first four bases prior to alignment.

Watch the video: What is Epigenetics? (August 2022).