Eukaryotic gene regulation & differentiation

Same genome, different cells

Gurdon's frogs, and later Dolly the sheep in 1996, showed that a specialised nucleus still carries the complete genome. With a few exceptions (for example, mature red blood cells in mammals have no nucleus at all), every cell in your body has essentially the same DNA.

What differs is gene expression: which genes are switched on, and how strongly. A liver cell expresses genes for blood-clotting proteins and detoxifying enzymes; a muscle cell expresses genes for actin, myosin and other contraction proteins. The process by which a cell becomes specialised by switching on a particular set of genes is called differentiation.

In bacteria, most control happens at one step: starting transcription. Eukaryotes regulate expression at almost every step on the way from gene to working protein. Let's walk through them in order.

Level 1: chromatin, open or closed

Eukaryotic DNA is wrapped around bundles of proteins called histones, like thread around spools. Each spool with its DNA is a nucleosome, and the whole DNA–protein package is chromatin. When nucleosomes are packed tightly (heterochromatin), RNA polymerase and transcription factors can't reach the DNA, and genes there are silent. Loosely packed chromatin (euchromatin) is accessible and can be transcribed.

Cells change packing by chemically tagging histones. Histones are positively charged, which helps them grip DNA's negatively charged backbone. Histone acetylation, adding acetyl groups to the histone tails, reduces that positive charge, so the DNA is held less tightly. The chromatin opens up and genes become easier to transcribe. Removing the acetyl groups (deacetylation) tightens the packing again.

The DNA itself can also be tagged. DNA methylation adds methyl groups to cytosine bases, usually where a C is followed by a G. Clusters of these CG pairs are common in promoters. A heavily methylated promoter is usually silenced, often because methylated DNA attracts proteins that deacetylate the nearby histones and compact the chromatin.

A dramatic example is X inactivation. Female mammals have two X chromosomes, males one. To even out the dose, early in a female embryo's development each cell shuts down one X at random. The silenced X is coated by an RNA called XIST, condenses into a small dense Barr body, and stays inactive in all of that cell's descendants. A tortoiseshell cat's patchy coat is a map of which X is active in each patch of skin.

Level 2: transcription factors and enhancers

Unlike bacterial RNA polymerase, eukaryotic RNA polymerase II can't bind a promoter on its own. It needs general transcription factors, proteins that assemble at the core promoter (often at a sequence called the TATA box, about 25–35 bases upstream of the start). These factors are needed for every protein-coding gene, so on their own they give only a low, basic level of transcription.

The gene-specific control comes from specific transcription factors. Activators bind DNA sequences called enhancers and increase transcription; repressors bind other sites and decrease it. Enhancers can be thousands of base pairs away from the gene, upstream, downstream, or even inside an intron. Proteins bend the DNA into a loop so the activators bound at the enhancer can touch the proteins at the promoter and help RNA polymerase get started.

Each gene has its own mix of enhancers and binding sites, and each cell type makes its own mix of transcription factors. A gene is switched on only where the right combination is present. This combinatorial control means a limited number of transcription factors can produce a huge variety of expression patterns, much as a few dozen letters can spell thousands of words.

Some transcription factors respond to signals from outside the cell. Steroid hormones such as oestrogen cross the plasma membrane and bind receptor proteins inside the cell, and the hormone–receptor complex acts as a transcription factor that binds DNA and changes which genes are expressed.

Levels 3 and 4: RNA processing and mRNA lifetime

Once a gene has been transcribed, the cell can still steer what it makes. Alternative splicing joins different combinations of exons, so one gene can give different proteins in different tissues.

The cell also controls how long each mRNA lasts. The 5′ cap and poly-A tail protect it, and proteins that bind the untranslated regions at either end can make it more stable or less stable. An mRNA that lasts longer gets translated more times.

MicroRNAs (miRNAs) add another layer. They're tiny RNAs, only about 21–24 nucleotides long, that are made from the genome but never translated. A miRNA joins a protein complex called RISC (the RNA-induced silencing complex). The miRNA base-pairs with matching sequences on target mRNAs, and RISC then blocks their translation or has them broken down.

Levels 5 and 6: translation and protein lifetime

Translation can be throttled too. Initiation needs helper proteins called initiation factors, and when a cell is stressed it can phosphorylate one of them (eIF-2), which blocks the start of translation for many mRNAs at once.

Finally, a finished protein can be switched on or off by chemical modifications such as phosphorylation, or destroyed. Tagging a protein with a small protein called ubiquitin marks it for breakdown by the proteasome, a large barrel-shaped protein complex that chops proteins back into short pieces. Controlling how fast a protein is destroyed controls how much of it is around.

From one cell to many types

Differentiation happens step by step. A fertilised egg is totipotent: it can give rise to every cell type, including the placenta. Early embryonic stem cells are pluripotent: they can become any cell of the body. As development goes on, cells commit to narrower fates, and the chromatin marks that silence unused genes help lock those choices in.

Master regulators sit at the top of these cascades. Homeotic (Hox) genes, for example, code for transcription factors that tell cells which part of the body they're in along the head-to-tail axis. Mutations in them can produce striking results, such as a fly growing legs where its antennae should be.

In 2006 Yamanaka and Kazutoshi Takahashi showed that switching on just four transcription-factor genes (Oct4, Sox2, Klf4 and c-Myc) could turn mouse skin cells called fibroblasts back into pluripotent cells, now called induced pluripotent stem (iPS) cells. Human iPS cells followed in 2007. Adding the right transcription factors was enough to rewrite a cell's identity, which shows how central these factors are.

Eukaryotes vs. bacteria

• Packing: bacterial DNA isn't wrapped in nucleosomes, so chromatin control is a mainly eukaryotic layer.
• Grouping: bacteria often control related genes together in operons with one mRNA. Eukaryotic genes usually have their own promoters and are coordinated by shared transcription factors instead.
• Distance: bacterial regulators mostly bind right next to the promoter. Eukaryotic enhancers can act from thousands of base pairs away.
• Steps: bacteria regulate mostly at transcription initiation. Eukaryotes add splicing, mRNA export and lifetime, miRNAs and more.
• Shared logic: both use proteins that bind specific DNA sequences to activate or repress transcription, often in response to small signal molecules.

Key terms

Differentiation

The process by which a cell becomes specialised by expressing a particular set of genes.

Chromatin / nucleosome

DNA plus its packaging proteins / one length of DNA wrapped around a histone core.

Histone acetylation

Adding acetyl groups to histones, loosening chromatin and making genes easier to transcribe.

DNA methylation

Adding methyl groups to cytosines, usually silencing the gene if the promoter is methylated.

Epigenetic

A heritable change in gene expression that doesn't change the DNA sequence.

Enhancer

A DNA sequence, possibly far from the gene, where activators bind to increase transcription.

Combinatorial control

A gene's expression depends on the particular combination of transcription factors present.

MicroRNA (miRNA)

A ~21–24-nucleotide RNA that, in RISC, blocks translation of or degrades matching mRNAs.

Induced pluripotent stem cell

A specialised cell reprogrammed to pluripotency by switching on a few transcription factors.

Check yourself

Try answering in your head before you open each answer.

• 1.A mutation deletes an enhancer 40,000 base pairs upstream of a muscle gene, but leaves the gene and its promoter intact. Predict the effect in muscle cells and in liver cells.Show answerHide

  In muscle cells, transcription drops sharply, because the activators that normally bind the enhancer and loop over to the promoter can no longer bind. In liver cells there's probably little change, since the gene is barely expressed there anyway. The protein made in muscle, if any, is normal: only the amount changes.

• 2.Two cell types both make transcription factor A. Only one of them makes factor B. Gene X needs A and B; gene Y needs only A. Which genes does each cell express?Show answerHide

  The cell with A and B expresses both X and Y. The cell with only A expresses Y but not X. This is combinatorial control: the same factor (A) contributes to different outcomes depending on its partners.

• 3.A drug causes a certain miRNA to be overproduced. Would the level of its target mRNA go up or down? What about the target gene's transcription rate?Show answerHide

  The target mRNA level (and its protein) would go down, because more RISC–miRNA complexes would find it and either block its translation or break it down. Transcription of the target gene wouldn't be directly affected: miRNAs act after the mRNA is made.