Chromatin DANIELA RHODES / MRC LABORATORY OF MOLECULAR BIOLOGY, CAMBRIDGE, UK CAMBRIDGE, BIOLOGY, MOLECULAR OF LABORATORY MRC / RHODES DANIELA

Pulling our strings There is much more to DNA than that elegant double helix. Philip Ball explores the twists and tangles of chromatin

50 | Chemistry World | May 2008 www.chemistryworld.org Why isn’t my arm a leg, or my liver read and transcribed into RNA a kidney? Their cells all have the In short – the first step in the conversion same genes. But genetics isn’t just  Chromosomes are of genetic information to proteins. about what genes you have – it also made up of a fibrous To understand why some genes depends on how you use them. composite of DNA and are silent while others are actively This crucial aspect of our biological protein called chromatin transcribed – and thus what allows identities has been overshadowed by  This coiled network genetically identical cells to become a focus on the identity of our genes, of protein fibres plays differentiated into different tissue and how these differ from those of an important role in types – we need to unravel the other people and other species. But it controlling the activity of packaging. has now become clear that genetics an organism’s genes can’t be fully understood until we  Chromatin can be Wound up genes have a better picture, not just of what changed and remodelled In eukaryotic cells, where the is in our genes, but of how they are with chemical labels DNA is housed in a cell nucleus, labelled and packaged. that provide signals for the chromosomes aren’t pure This evolving view of gene enzymes, indicating DNA. About half of their material packaging is quite different from the whether or not genetic is protein: chromosomes are made classical image of genes as stretches material should be up of a fibrous composite of DNA of pristine double-stranded DNA. transcribed into proteins and protein called chromatin. It Our genetic book has been scribbled  These modifications, now seems that this structure can all over: there are marginal notes called epigenetic determine the activity of genes, and and sections crossed out. Not only changes, are a way of that it therefore plays as big a role in do these textual changes determine changing the activity of the genetic make-up of organisms what the genetic book says, but they genes without changing as the basic sequence of the genes reveal the book’s history, showing the genes themselves themselves. The chemical nature how its environment has modified of chromatin may hold the key to and shaped its message. understanding how a set of genes Perhaps part of the reason for the creates a complex, multicelled relative neglect of this aspect of gene organism like ourselves – and also function is that it is a tremendously to how this process sometimes goes hard problem, to which the answer awry and leads to conditions such as is likely to be messy. It could be cancer. said that molecular biologists have In 1974, Roger Kornberg, then been in denial ever since Francis working at the UK Medical Research Crick and James Watson discovered Council’s Laboratory of Molecular the structure of DNA in 1953. That Biology (LMB) in Cambridge, beautiful double helix, with its proposed that chromatin is like a genetic information written into the string of beads. The DNA double spiral staircase of paired nucleic- helix, he argued, is acid bases, offers such an elegant looped around picture of the chemical principles of barrel-like life and inheritance that everyone units made fell for it. up of This image of DNA is celebrated proteins in the corridors of biology labs the called world over. But when we come face Chromatin’s proposed histones. to face with DNA in the cell, it’s ‘string of beads’ Each of like meeting a movie star whose structure airbrushed publicity photos don’t look at all like the real thing. You these would barely recognise Crick and barrels is Watson’s perfectly-formed molecule made of two in the tangled, twisted and bent copies each spaghetti that is stuffed inside the of four different nuclei of our cells. histone proteins It may be messy, but the packaging – H2A, H2B, H3 and H4. of DNA into the chromosomes seems Segments of DNA about 146 to have a deep logic to it. These rules base-pairs long are spooled determine how available the DNA is around each histone octamer in for being a double loop. Each of these DNA- histone units is called a nucleosome, and chromatin consists of a succession of nucleosomes linked, by short stretches of DNA, into a flexible chain. There is a fifth histone protein, called H1/H5, attached to the nucleosome at the places where the DNA first makes contact with,

and exits from, the protein core. LIBRARY PHOTO SCIENCE / DESIGN LAGUNA www.chemistryworld.org Chemistry World | May 2008 | 51 Chromatin

The nucleosomes are themselves packed together to form chromatin fibres about 30nm wide – but no one knows quite how. One idea is that the fibres might be supercoiled into a solenoid, or one-start structure, as happens if you twist the helical cables of old telephones. In this structure, each pair of successive nucleosomes is connected by a bent piece of linking DNA. But it has also been suggested that adjacent nucelosomes along the chain could be connected by straight segments, creating twinned stacks of nucleosome helices – two-start structures (see right). In 2004, Timothy Richmond’s research team at ETH in Zurich, Switzerland, argued that chromatin 1 fibres must have a two-start form. MOZZICONACCI JULIEN But last year, Daniela Rhodes and her colleagues at the LMB challenged coiling, as well as that of the DNA Tangled web: single types of chromatin in the nucleus this idea.2 They chopped up and path around the fibre axis.’ supercoiled or one- of a eukaryotic cell, at least during reconstituted chromatin with It all depends on how many start structure (A) and the interphase period – when it is varying lengths of DNA linking the base pairs there are between one twinned stacks or two- not about to divide. Most is in the nucleosomes – from 10 to 70 base nucleosome and the next, which can start structures (B,C) of form of euchromatin, a fairly open, pairs (bp). They found that, whereas vary in different parts of the genome chromatin fibres gel-like fibrous tangle. The rest is two-start structures should have a and under different conditions. heterochromatin, which is much fibre width that increases gradually In both the centres and the tips of denser and confined to a few small with increasing linker length, in chromosomes, for example, the patches. Because DNA wound fact the fibres fall into two distinct average repeat length is different onto histones and packed tightly classes. For linkers of up to around from the rest. The repeat length also into fibres can’t be accessed by the 40 bp, the fibres have a diameter of seems to change when chromosomes molecular machinery that controls 33nm and contain 11 nucleosomes become compact as a cell prepares transcription of genetic material in every 11nm of length. But with to divide. To accommodate such into RNA, there is a complex array linkers of more than 50 bp, the changes, Mozziconacci thinks that of enzymes involved in unpacking fibres are 44nm wide and have 15 the H1/H5 histone plays the role and repacking (‘remodelling’) nucleosomes per 11nm length. They of an adjustable clip that alters the chromatin, which is vital to the proposed that these results could angle at which the DNA meets and readout of our genes. be understood with yet another exits from the nucleosome core, so It’s tempting to imagine that structure, in which nucleosomes as to make for a tidy fit in the packed- euchromatin is unpacked and in adjacent helical arrays are up fibre. accessible for transcription, while interdigitated. This shape-shifting nature heterochromatin is inactive, The argument isn’t yet settled. could provide a mechanism for like a compressed data file. But Julien Mozziconacci of the controlling gene function. ‘I think it’s not that simple. A lot of the Université Pierre et Marie Curie that the histones evolved so that DNA in euchromatin never gets in Paris, France and his colleagues they can use the helical shape of transcribed – so it’s not obvious why have tried to sort it out by combining DNA to acquire this polymorphism,’ it should be left ‘open’. Conversely, Rhodes’ results with molecular says Mozziconacci. ‘Shuffling chromosomes containing a large modelling.3 They have built of nucleosomes to alter the amount of heterochromatin can computer models of chromatin repeat length might be a way of be transcriptionally active. Some fibres that include every single regulating transcription and other researchers think that euchromatin atom, to see whether the proposed chromosomal processes.’ is actually just a blanket term structures will fit without any atoms It’s still an open question how all that hides structural subtleties of getting in each others’ way. They of this, which stems mostly from ‘Coming face to chromatin we don’t understand. say that chromatin fibres can adopt work on chromatin isolated in test face with DNA ‘How heterochromatin and many different structures, including tubes, carries over to living cells, euchromatin structures differ is the supercoiled solenoid and various but it seems clear that there’s much in the cell is unclear,’ says chemical biologist two-start helices – meaning that more to chromatin structure and like meeting a Michael Grunstein of the University almost all of the theories could be packaging than the 30nm fibres. of California at Los Angeles. correct. ‘I believe there are definite higher- movie star – the To fully understand transcription ‘We assumed that nucleosomes order structures of chromatin in the airbrushed – what determines if genes are ‘on’ or are stacked into columns which are, cell, above the level of these fibres,’ ‘off’ – we must therefore get to grips in turn, coiled to form 35 or 45nm says Mozziconacci. ‘I’d speculate publicity photos with what chromatin remodeling fibres,’ says Mozziconacci. ‘We that there are motifs such as kinks look nothing entails. It is done by various big identified 15 different ways to do and crosslinkings which result in an protein complexes, which distort this, depending on the number of ordered architecture.’ like the real the contacts between histones columns and the handedness of their For one thing, there are in fact two thing’ and DNA or move nucleosomes 52 | Chemistry World | May 2008 www.chemistryworld.org DANIELA RHODES / MRC LABORATORY OF MOLECULAR BIOLOGY, CAMBRIDGE, UK CAMBRIDGE, BIOLOGY, MOLECULAR OF LABORATORY MRC / RHODES DANIELA

around. One possibility is that these genes without changing the genes Daniela Rhodes proposed language of life. chromatin-remodelling complexes themselves. Such behaviour is called a chromatin structure in Geneticist Bryan Turner of the (CRCs) simply loosen up the epigenetic. In effect, epigenetic which helical arrays are University of Birmingham, UK, chromatin so that nucleosomes can processing enables cells to select interdigitated who first proposed it,5,6 admits slide along the DNA chain. But it from the complete genome just that the idea of an epigenetic code now seems likely that their activity those genes that it needs to serve ‘is still very controversial’. ‘We do is more specific. Last October, a particular function – which can not yet have the finding that will Gernot Längst of the University change with time. Epigenetic prove or disprove the epigenetic of Regensburg found that several modifications are reversible, so that code hypothesis,’ he says. ‘Histone CRCs position nucleosomes quite a cell’s active genetic resources are modifications may be involved in precisely by looking for specific being continually manipulated. all sorts of processes. If there is a DNA sequences.4 And since there Some researchers believe that heritable code, then we need to clear are hundreds or even thousands epigenetic modifications of away modifications due to all these of different CRCs in the nuclei of chromatin should be regarded as a other processes before we can find eukaryotic cells – at concentrations kind of code that complements the it.’ that may provide as many as one basic genetic code of DNA, and is Whether it’s a code or not, the CRC for every ten or so nucleosomes even capable of over-riding it. repertoire of histone modifications – it seems inevitable that they must Others say that code is not the is beginning to be catalogued. One have rather specialised functions. right word – even if chemical of the most common changes is Since this remodeling can alter the alteration of histones can rearrange acetylation – where an acetyl group accessibility of genes, determining chromatic structure and turn is substituted in place of a hydrogen whether they are active or silent, genes on and off. ‘A code needs atom on the amino group of a it seems clear that CRCs must play to be consistent, predictive and lysine residue. This can occur on a crucial role in defining a cell’s combinatorial,’ says Grunstein. several of the lysines of all four core genetic status. (Morse code is a true code in this histones, particularly H3 and H4, so sense because, whereas one dot some nucleosomes can be multiply Read the label means ‘e’, two dots mean ‘i’ rather acetylated. Acetylation of lysine Both remodeling of chromatin by than ‘ee’.) ‘Epigenetic modifications removes its positive charge, reducing CRCs and transcription of DNA are none of these.’ ‘The shape- the tendency of the histone to stick are governed by chemical labels It seems like a semantic point, to negatively charged DNA. That stuck to the histones or to the DNA but at its core is the question of shifting nature in turn can open up the chromatin itself. Specialised enzymes add or how information is encoded in the of chromatin structure and activate genes for chop bits off the proteins or DNA, working genome: a true epigenetic transcription. providing signals for other enzymes code would be more than just a could provide Enzymes that add a methyl group that indicate whether the genetic series of chemical on-off switches, a mechanism to arginine and lysine residues – the material should be read out or but would in some ways be as process of methylation – can be ignored. These modifications are fundamental as the genetic code for controlling responsible for either activating or a way of changing the activity of of DNA base pairs – a part of the gene function’ silencing genes. Transcriptionally www.chemistryworld.org Chemistry World | May 2008 | 53 Chromatin silent heterochromatin often ‘This right soon enough, however, that can reproduce sexually, epigenetic contains nucleosomes in which lead to breakdowns in the cell cycle modifications of chromatin are histone H3 has been methylated at a undermines and the onset of cancer – which is generally cleared during the site to which the enzyme HP1, which the classical why labels would be abundant in merging of germline cells (sperm is involved in repression of genes the chromatin of a person at risk of and egg) in the process of meiosis. and packaging of heterochromatin, rules of genetic developing it. Researchers hope that Azim Surani of the University of binds. It was once thought that inheritance’ these histone marks can therefore Cambridge and his coworkers methylation is irreversible, but in offer an early-warning system for have argued recently that the 2004 a team at Harvard Medical prognoses of cancer in clinical stripping of methyl tags from School in the US discovered an screening. DNA during this reprogramming enzyme that demethylates lysine.7 may involve enzymes involved in Several others are now known. Rewriting evolution DNA repair.12 But such wiping of Other histone ‘marks’ include Perhaps most dramatically of all, epigenetic changes doesn’t always phosphorylation, ubiquitination it seems that some epigenetic seem to happen – sometimes (attachment of the small protein changes can be inherited: a kind of these modifications get passed ubiquitin, usually used as a marker of Lamarckian inheritance of acquired on. The classic example is called old or damaged characteristics called epigenetic paramutation, a type of gene proteins inheritance, which undermines the silencing in which the gene variant that need classical rules of genetic inheritance. (allele) acquired from one parent’s chewing up by One way in which this can happen DNA affects the other allele in the proteasomes) is by chemical marking of DNA progeny cell. First identified in and itself, rather than histones. plants in the 1950s, paramutation citrullination Between 2 and 7 per is still not well understood, but it (the cent of the cytosine is thought to involve epigenetic transformation bases in our DNA chemical modification of either of arginine (and as much as DNA or histones in a way that is into the amino 30 per cent in mediated by RNA. acid citrulline). some plants) have ‘While it is clear that chromatin Individual nucleosomes may methyl groups states are stable and can be contain several marks at once. attached, which preserved through cell division, Last year, Bradley Bernstein of the alters the activity the precise mechanistic roles Massachusetts General Hospital of the associated played by the histone modifications in Boston, US, and his coworkers genes. Some themselves remain obscure,’ says reported that developmental DNA-methylating Bernstein. All the same, there’s now genes in mouse stem cells carry enzymes, such as no doubt that the neo-Darwininan both ‘activating’ and ‘silencing’ Dnmt 1, tend to picture of genetics and inheritance histone marks.8 This, the stick methyl groups is incomplete, and we won’t fully researchers think, is a way of onto CG-GC base pairs understand either how evolution fine-tuning the genes during that already have them works or what makes us human until the developmental process, for – and so methylation we have done some more unraveling example keeping them inactive can be inherited by a of the subtle chemistry of chromatin. but poised to go into action once replicated strand. In the stem cells become committed 2003, for example, Philip Ball is a science writer based in to developing into a specific a team at the London, UK tissue type. University of Sydney, Some epigenetic marks are Australia, showed References associated with cancer cells. that mice could inherit 1 B Dorigo et al, Science, 2004, 306, 1571 2 P J J Robinson et al, Proc. Natl. Acad. Sci. For example, Grunstein and a kinked tail caused by USA, 2006, 103, 6506 his former student Siavash epigenetic methylation 3 H Wong et al, PLoS ONE, 2007, 2, e877 Kurdistani have found that of the DNA in a gene 17849006 certain patterns of acetylation called AxinFu.11 It’s 4 K Rippe et al, Proc. Natl. Acad. Sci. USA, 2007, 104, 15635 and demethylation of residues in also possible that 5 B M Turner, Cell, 1993, 75, 5 the H3 and H4 histones seem histones attached 6 B M Turner, Cell, 2002, 111, 285 to appear in men with prostate to replicating DNA 7 Y Shi et al, Cell, 2004, 119, 941 cancer.9 A link between histone might pass on their 8 B E Bernstein et al, Cell, 2006, 125, 315 9 D B Seligson et al, Nature, 2005, 435, 1262 marks and cancer would be modifications to 10 J A Downs et al, Nature, 2007, 447, 951 understandable, because those on the daughter 11 V K Rakyan et al, Proc. Natl. Acad. Sci. USA, these marks are often used strands. 2003, 100, 2538 to label damaged DNA so that In organisms that 12 P Hajkova et al, Nature, 2008, DOI: 10.1038/ it can be identified for repair nature06714 by specialised enzymes.10 A break in the double helix can, for example, lead to phosphorylation of many H2 histones in that region within seconds, in effect putting up a distress flag for DNA-repair enzymes. Specific CRCs are thought to have a role in repairing such

NATURE fractures. If the damage isn’t put 54 | Chemistry World | May 2008 www.chemistryworld.org