Transcription (biology)

• 1. Introduction

  ›

  Transcription is the process of duplicating a segment of DNA into RNA, and it happens inside your cells right now, over and over, faster than you can read this sentence. RNA polymerase, the central enzyme, can add roughly 10 to 100 nucleotides every second. A single gene can be read by multiple polymerases at once, producing many RNA molecules in rapid succession. This is how a fixed strand of DNA becomes the moving, working instructions of life. But the process raises questions worth following. How does the cell decide which genes to read, and when? Why does an enhancer sitting tens of thousands of nucleotides away matter to a gene it never touches directly? How does a copying machine that makes more mistakes than DNA replication still keep an organism alive? And why do some viruses, and some of our own cells, run this process backwards, turning RNA into DNA?

• 2. Reading One Strand

  ›

  Only one of the two DNA strands serves as a template for transcription. The antisense strand is read by RNA polymerase from the 3' end to the 5' end, while the complementary RNA grows in the opposite direction, 5' to 3'. RNA polymerase can only add nucleotides to the 3' end of the growing chain, which fixes this directionality. Reading just the 3' to 5' strand carries a hidden economy. It eliminates the need for the Okazaki fragments seen in DNA replication, and it removes the need for an RNA primer to start synthesis. The new RNA differs from DNA in a small but consistent way. Wherever thymine would appear in a DNA complement, the RNA carries uracil instead. The non-template strand is called the coding strand, because its sequence matches the new RNA transcript, except for that uracil-for-thymine swap. By convention, this is the strand written down when a DNA sequence is presented. A protein-coding transcription unit holds more than the coding sequence itself. It includes regulatory regions, the five prime untranslated region upstream and the three prime untranslated region downstream, that direct and regulate the protein's synthesis.

• 3. Action At A Distance

  ›

  An activated enhancer can drive up to a 100-fold increase in a gene's transcription, and the enhancer may sit in a DNA region far from the gene it controls. Enhancers are among the genome's major gene-regulatory elements, and they often work by looping through long distances to come into physical proximity with their target promoters. There are hundreds of thousands of enhancer regions, but for any one type of tissue only specific enhancers are brought close to the promoters they regulate. In a study of brain cortical neurons, researchers found 24,937 loops connecting enhancers to their target promoters. The loop itself is held in place by a dimer of a connector protein, such as CTCF or YY1, with one member anchored to the enhancer and the other to the promoter. Several transcription factors bind specific motifs on an enhancer, and a small combination of them, once a loop brings them near a promoter, governs the gene's level of transcription. A human cell holds about 1,600 transcription factors. Between the enhancer-bound factors and the polymerase sits Mediator, a complex usually made of about 26 proteins, which passes regulatory signals directly to RNA polymerase II at the promoter. Active enhancers do something surprising of their own. They are transcribed from both DNA strands at once, with polymerases moving in two directions, producing two enhancer RNAs. An activated enhancer begins transcribing its own RNA before it activates transcription of messenger RNA from its target gene.

• 4. The Methyl Switch

  ›

  About 60% of promoters have their transcription controlled by methylation of cytosines within CpG dinucleotides, sites where a 5' cytosine is followed by a 3' guanine. The marker is 5-methylcytosine, an epigenetic flag found predominantly at CpG sites. The human genome contains about 28 million CpG dinucleotides, and in most mammalian tissues, on average, 70% to 80% of CpG cytosines are methylated. Unmethylated cytosines tend to cluster in groups called CpG islands at active promoters. About 60% of promoter sequences carry a CpG island, while only about 6% of enhancer sequences do. When a promoter's CpG island is methylated, transcription of that gene can be reduced or silenced. Methylation does its silencing work through methyl binding domain proteins, including MeCP2, MBD1, and MBD2. These proteins bind most strongly to highly methylated CpG islands, and they carry both a methyl-CpG-binding domain and a transcription repression domain. Once bound, they guide protein complexes that remodel chromatin or modify histones, generally creating a repressive environment. One transcription factor stands out in this story. EGR1 is important for regulating CpG island methylation, with about 12,000 binding sites in the mammalian genome, roughly half in promoters and half in enhancers. Its binding is insensitive to whether the cytosine nearby is methylated. Only small amounts of EGR1 appear in unstimulated cells, but one hour after stimulation by growth factors, neurotransmitters, hormones, stress, or injury, its production rises sharply. In the brain, activated neurons up-regulate EGR1, which recruits pre-existing TET1 enzymes already abundant there. TET enzymes can demethylate 5-methylcytosine, so EGR1 brings TET1 to methylated promoters and clears the marks, letting those genes switch on. Hundreds of genes change their expression in neurons this way after activation.

• 5. Getting Started

  ›

  Transcription begins when RNA polymerase and one or more general transcription factors bind a DNA promoter to form a closed complex, in which the promoter DNA is still fully double-stranded. The polymerase then unwinds approximately 14 base pairs of DNA to form an open complex, exposing single-stranded DNA known as the transcription bubble. Inside that bubble, the polymerase selects a transcription start site, binds an initiating and an extending nucleotide, and catalyzes the first bond to yield an initial RNA product. The machinery differs across the domains of life. In bacteria, the RNA polymerase core enzyme has five subunits, two alpha, one beta, one beta prime, and one omega, and it binds a single transcription factor, the sigma factor, to form the holoenzyme. The first nucleotide of bacterial mRNA is not capped with a modified guanine. It instead bears a 5' triphosphate, a feature that can be used for genome-wide mapping of where transcription starts. Archaea and eukaryotes need more helpers. Archaea use three general transcription factors, TBP, TFB, and TFE. Eukaryotic RNA polymerase II requires six, TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH, with TBP being the first to bind DNA and TFIIH the last to arrive. In these organisms the closed complex is called the preinitiation complex. Across all of this, activators and repressors, and sometimes coactivators or corepressors, tune the formation and function of the initiation machinery.

• 6. Escape And Elongation

  ›

  After the first bond forms, RNA polymerase has to escape the promoter, and at first it tends to fail. It releases the transcript and produces truncated fragments, a behavior called abortive initiation that is common in both eukaryotes and prokaryotes. This continues until an RNA product reaches a threshold length of about 10 nucleotides, at which point promoter escape occurs and a transcription elongation complex forms. The escape itself is powered by DNA scrunching, which provides the energy to break the contacts between the holoenzyme and the promoter. In eukaryotes at an RNA polymerase II promoter, TFIIH phosphorylates serine 5 on the carboxy terminal domain of the polymerase, which recruits the capping enzyme. Once moving, the polymerase traverses the template strand and uses base pairing to build an RNA copy that grows as it goes. The product is an exact copy of the coding strand, with thymines replaced by uracils, and built on a ribose sugar rather than the deoxyribose found in DNA. Speed comes at a price in fidelity. Transcription has some proofreading mechanisms, but they are fewer and less effective than the controls for copying DNA, so transcription has lower copying fidelity than replication. In eukaryotes, nucleosomes act as major barriers during elongation, and the pausing they cause can be regulated by elongation factors such as TFIIS. Those pauses can also let RNA editing factors bind. There is a darker side to all this activity. Double-strand breaks in actively transcribed regions are repaired by homologous recombination during the S and G2 phases of the cell cycle. Because transcription makes DNA more accessible to chemicals and metabolites that cause damaging lesions, recombination of a transcribed sequence can be strongly stimulated by the act of transcription itself.

• 7. Knowing When To Stop

  ›

  Bacteria end transcription using two strategies, Rho-independent and Rho-dependent termination. In the Rho-independent route, the new RNA forms a G-C-rich hairpin loop followed by a run of uracils, and the mechanical stress of the hairpin breaks the weak rU-dA bonds, pulling the poly-U transcript out of the polymerase. In the Rho-dependent route, the protein factor Rho destabilizes the bond between template and mRNA, releasing the new transcript from the elongation complex. Eukaryotic termination is less well understood, but it involves cleaving the new transcript and then adding adenines to its 3' end in a process called polyadenylation. Sometimes transcription must stop for reasons unrelated to reaching the end of a gene, such as DNA damage or a collision with an active replication fork. In bacteria, the Mfd ATPase removes a polymerase stalled at a lesion by prying open its clamp, and it recruits nucleotide excision repair machinery to fix the damage. Mfd is also proposed to resolve conflicts between replication and transcription. Eukaryotes use the ATPase TTF2 to suppress RNA polymerase I and II during mitosis, preventing errors in chromosomal segregation, while archaea may use the Eta ATPase for a similar job. The danger here is real and constant. Each cell sustains an estimated tens to hundreds of thousands of DNA damages every day, and transcription is a major source, because it creates single-strand DNA intermediates that are vulnerable to harm.

• 8. Running It Backwards

  ›

  Some viruses can transcribe RNA into DNA, and HIV, the cause of AIDS, is one of them. HIV carries an RNA genome that is reverse transcribed into DNA, and that DNA can be merged into the host cell's own genome. The key enzyme is reverse transcriptase, which synthesizes a complementary DNA strand from the viral RNA. After that, ribonuclease H digests the RNA strand, reverse transcriptase builds a second DNA strand to form a double helix, and the enzyme integrase splices the result into the host genome. The hijacked cell then makes viral proteins that reassemble into new particles. In HIV, the host T cell afterward undergoes programmed cell death, or apoptosis, though in other retroviruses the cell survives as the virus buds out. Our own cells run reverse transcription too, through an enzyme called telomerase. Telomerase carries an RNA template and uses it to add a telomere, a repeating DNA sequence, to the ends of linear chromosomes. This matters because every time a linear chromosome is duplicated it is shortened, and the telomere lets that shortening eat away non-essential repeats rather than protein-coding sequence. Cancer turns this protection into a weapon. Telomerase is often activated in cancer cells, letting them duplicate their genomes indefinitely without losing protein-coding DNA, which may be part of how they become immortal. This telomere-lengthening route has been shown to occur in 90% of carcinogenic tumors in vivo. The remaining 10% rely on a different path called ALT, or Alternative Lengthening of Telomeres.