Transcription, RNA Processing and Translation

A gene is only a promise until it is expressed. In this chapter you follow information from DNA to RNA to protein: how RNA polymerase finds a gene and copies it, how eukaryotic cells cap, cut, splice and tail the transcript, how a triplet code was cracked with synthetic RNAs, and how the ribosome, itself a machine made largely of RNA, reads that code to build a protein. Nearly every mutation, drug and disease you meet later in genetics acts somewhere along this path.

Section 1Transcription in bacteria

Template, coding strand and direction

Transcription is the synthesis of RNA using one strand of DNA as a template. Like DNA polymerase, RNA polymerase joins nucleotides (here ribonucleoside triphosphates: ATP, GTP, CTP and UTP) to the 3' end of the growing chain, so RNA is always made 5' to 3' and the template is read 3' to 5'. Unlike DNA polymerase, RNA polymerase needs no primer; it can start a chain from scratch. It also unwinds DNA itself, so no separate helicase is required.

The DNA strand that is read is the template strand (also called the antisense or noncoding strand). The other strand, the coding strand (sense or nontemplate strand), has the same sequence as the RNA except that the RNA contains uracil where the DNA has thymine. By convention, gene sequences are written as the coding strand, 5' to 3', so that they read like the mRNA. Different genes on the same chromosome may use different strands as template; what matters is the orientation of each gene's promoter. Positions are numbered relative to the first transcribed nucleotide, +1; nucleotides before it, in the upstream direction, have negative numbers, and there is no position 0.

RNA polymerase and sigma factors

Bacteria have a single RNA polymerase that makes all their RNAs. The core enzyme has five subunits, written α₂ββ'ω. The β and β' subunits form the catalytic site and the channels for DNA and RNA; the two α subunits assemble the enzyme and contact some regulators; ω helps assembly. The core enzyme can elongate RNA but cannot find promoters. That job belongs to the sigma (σ) factor, which joins the core to form the holoenzyme. In E. coli, the housekeeping factor σ⁷⁰ directs transcription of most genes. Alternative sigma factors recognize different promoter sequences and switch on whole sets of genes at once: σ³² for heat shock genes, σ^S for stationary phase and stress, σ⁵⁴ for nitrogen metabolism and σ²⁸ for flagellar genes. Bacteria that form spores use a cascade of sigma factors to control development.

The promoter

A promoter is the DNA sequence where the holoenzyme binds and starts transcription. Comparing hundreds of σ⁷⁰ promoters reveals two conserved hexamers on the coding strand. The −35 element has the consensus TTGACA, and the −10 element (the Pribnow box) has the consensus TATAAT. They are separated by a spacer of 16 to 18 bp, with 17 bp optimal, which places both elements on the same face of the helix where σ can contact them. Region 4 of σ binds the −35 element and region 2 binds the −10 element. A consensus sequence is the most common base at each position; few real promoters match it exactly, and in general the closer a promoter is to consensus, the stronger it is. Very strong promoters, such as those of ribosomal RNA genes, also carry an AT-rich UP element upstream of −35 that is bound by the C-terminal domain of the α subunits.

Initiation, elongation and the transcription bubble

Initiation proceeds in stages. The holoenzyme first binds the promoter as a closed complex in which the DNA is still double stranded. It then unwinds about 13 bp from roughly −11 to +2, forming the open complex. The AT-rich −10 element is where melting begins, which is one reason that TATAAT is AT rich. The polymerase starts synthesis at +1, usually with a purine, but it often makes and releases short transcripts of 2 to 10 nucleotides before it breaks free of the promoter, a process called abortive initiation. Once the RNA is about 10 to 15 nucleotides long, the enzyme undergoes promoter escape, σ is usually released, and the core enzyme continues as an elongation complex.

During elongation, the polymerase maintains a transcription bubble of about 12 to 14 unpaired base pairs, within which the newest 8 to 9 nucleotides of RNA are base paired to the template as an RNA:DNA hybrid. DNA rewinds behind the bubble and the RNA peels away. Bacterial RNA polymerase adds roughly 40 to 80 nucleotides per second. Its error rate, about 1 in 10⁴ to 10⁵, is much higher than that of DNA replication; a limited form of proofreading, in which the enzyme backs up and cleaves off a misincorporated nucleotide, improves it somewhat. Errors in RNA matter less than errors in DNA because each RNA is one of many short-lived copies.

In bacteria there is no nucleus, so ribosomes begin translating an mRNA while its 3' end is still being made. This coupling of transcription and translation is important for regulation (as in attenuation, Chapter 9). Bacterial mRNAs are often polycistronic, carrying the coding sequences of several genes of an operon, each with its own ribosome binding site.

Termination

Bacteria end transcription in two ways. At an intrinsic (Rho-independent) terminator, the DNA contains a GC-rich inverted repeat followed by a run of about 7 to 8 A's on the template strand. As the inverted repeat is transcribed, the RNA folds back on itself into a stable hairpin (stem-loop) that pulls on the RNA in the polymerase exit channel. The run of U's that follows forms an RNA:DNA hybrid held only by weak A:U pairs, and the combination of hairpin and weak hybrid causes the RNA to dissociate. Rho-dependent termination requires the protein Rho, a ring-shaped hexameric helicase. Rho binds the RNA at a rut site (Rho utilization site), a stretch rich in C and poor in G that is not covered by ribosomes, and uses ATP to move 5' to 3' along the RNA. When the polymerase pauses at a termination site, Rho catches up and pulls the RNA out of the complex.

Section 2Transcription in eukaryotes

Three nuclear RNA polymerases

Eukaryotic nuclei contain three RNA polymerases, each a complex of 12 or more subunits, and each responsible for a different class of genes. Plants have two more (Pol IV and Pol V) that make small RNAs for gene silencing, and mitochondria and chloroplasts have their own simpler polymerases. The three nuclear enzymes were first distinguished by their sensitivity to α-amanitin, a cyclic peptide toxin from the death cap mushroom.

Unlike bacterial RNA polymerase with sigma, none of the eukaryotic enzymes can recognize a promoter by itself. Each depends on a set of general transcription factors that bind the promoter first and then recruit the polymerase. Pol I and Pol III use their own factor sets (for example, TFIIIA, TFIIIB and TFIIIC for Pol III, some of which bind internal promoters located inside the transcribed region of tRNA and 5S rRNA genes). Because protein-coding genes are the focus of most genetics, the rest of this section concentrates on Pol II.

The core promoter and the general transcription factors

The core promoter is the minimal region, roughly −40 to +40, where the transcription machinery assembles. Its best-known element is the TATA box, consensus TATA(A/T)A(A/T), located about 25 to 30 bp upstream of the start site. Only about 10 to 20 percent of human promoters have a recognizable TATA box. Others depend on an initiator (Inr) element spanning +1, a downstream promoter element (DPE) near +30, or a TFIIB recognition element (BRE) just upstream of the TATA box. Many housekeeping genes have TATA-less promoters embedded in CpG islands. Just upstream of the core, proximal promoter elements such as the GC box (bound by Sp1) and the CAAT box (bound by NF-Y) raise the rate of transcription.

On a TATA-containing promoter, the general factors assemble in a well-studied order to form the preinitiation complex (PIC):

1. TFIID binds first. Its subunit TBP (TATA-binding protein) grips the TATA box through the minor groove and bends the DNA by about 80 degrees; the other subunits, the TBP-associated factors (TAFs), recognize Inr and DPE and contact activators. TBP is also used by Pol I and Pol III, making it a universal eukaryotic factor.
2. TFIIA stabilizes TBP binding, and TFIIB binds TBP and the BRE, forming a platform that positions the polymerase and helps select the start site.
3. Pol II arrives bound to TFIIF.
4. TFIIE joins and recruits TFIIH, the last factor. TFIIH has two critical activities: an ATP-dependent DNA translocase (the XPB subunit) that opens the promoter DNA, and a kinase (CDK7) that phosphorylates the polymerase's tail.

That tail is the C-terminal domain (CTD) of the largest Pol II subunit, a series of repeats of the seven amino acids Tyr-Ser-Pro-Thr-Ser-Pro-Ser (52 repeats in humans, 26 in yeast). The polymerase that enters the PIC has an unphosphorylated CTD. Phosphorylation of serine 5 by TFIIH helps the polymerase break its contacts with the promoter factors and escape. Many genes then pause about 20 to 60 nucleotides downstream until the kinase P-TEFb (CDK9) phosphorylates serine 2, releasing the polymerase into productive elongation. The changing pattern of CTD phosphorylation acts as a code read by RNA processing factors: serine 5 phosphorylation recruits the capping enzymes, and serine 2 phosphorylation recruits splicing and 3'-end processing factors as the gene is transcribed. This is why processing is cotranscriptional.

Enhancers, activators and chromatin

The general factors alone support only a low, basal level of transcription. High-level, regulated transcription needs activators, sequence-specific DNA-binding proteins that bind enhancers. Enhancers are short regulatory regions that can lie upstream, downstream or within introns, sometimes tens or hundreds of kilobases from the promoter, and still work when their orientation is reversed. DNA looping brings an enhancer near its promoter, and the large Mediator complex carries the signal from activators to the PIC. Activators also recruit enzymes that modify or move nucleosomes, because promoter DNA wrapped in nucleosomes is largely inaccessible. Silencers bind repressors that do the opposite. Chapter 9 develops this regulation in detail.

Termination of Pol II transcription

Pol II has no fixed termination sequence. Instead, termination is coupled to 3'-end processing (Section 3). When the polymerase transcribes the polyadenylation signal AAUAAA, the processing machinery binds it from the CTD, and the RNA is cleaved downstream. The polymerase keeps going, but the cleavage leaves an uncapped 5' end on the RNA still emerging from it. A 5' to 3' exonuclease, Xrn2 in humans, attacks that end, degrades the RNA faster than it is made and dislodges the polymerase, often hundreds to thousands of bases past the poly(A) site. This is the torpedo model. Pol III, by contrast, terminates at a simple run of T's in the coding strand, and Pol I uses a terminator-binding protein.

Section 3RNA processing: capping, splicing, polyadenylation

A bacterial mRNA can be translated as it is made. A eukaryotic transcript must first be converted from pre-mRNA into mature mRNA, then exported from the nucleus. Three main modifications occur, all while transcription is still under way: a cap is added to the 5' end, introns are removed by splicing, and the 3' end is cut and given a poly(A) tail. The collection of unprocessed nuclear transcripts is sometimes called heterogeneous nuclear RNA (hnRNA).

The 5' cap

When the RNA is only about 20 to 30 nucleotides long, three enzymes, recruited by the serine 5 phosphorylated CTD, add the cap. An RNA triphosphatase removes the outermost phosphate of the first nucleotide; a guanylyltransferase adds GMP from GTP in an unusual 5' to 5' triphosphate linkage; and a methyltransferase methylates the guanine at position N7 using S-adenosylmethionine. The result, 7-methylguanosine (m⁷G) joined by a 5'-5' triphosphate bridge, is called cap 0. In vertebrates the 2'-OH of the first one or two transcribed nucleotides is also methylated (cap 1, cap 2), which helps the innate immune system distinguish the cell's own RNA from viral RNA. The cap protects the mRNA from 5' to 3' exonucleases, is needed for efficient splicing of the first intron and for export, and is recognized by eIF4E during translation initiation.

Introns and exons

Most eukaryotic genes are interrupted by introns, intervening sequences that are transcribed but removed, separated by exons, the sequences that remain in mature RNA. Introns were discovered in 1977 when adenovirus mRNA hybridized to its DNA left loops of unpaired DNA. A typical human protein-coding gene has about 8 to 9 introns. Exons average roughly 150 nucleotides, while introns average several thousand, so introns make up most of the length of many genes. Exons include untranslated regions as well as coding sequence: the first exon usually contains the 5' untranslated region (5' UTR) and the last contains the 3' UTR.

The spliceosome and the chemistry of splicing

Nuclear pre-mRNA introns carry short conserved signals. The 5' splice site begins the intron with GU (consensus exon...AG | GURAGU), the 3' splice site ends it with AG, and a branch point containing a critical adenosine lies about 18 to 40 nucleotides upstream of the 3' end, followed by a polypyrimidine tract. The nearly invariant GU and AG give the GU-AG rule.

Splicing is carried out by the spliceosome, a large complex of five small nuclear ribonucleoproteins (snRNPs), called U1, U2, U4, U5 and U6, plus more than 100 proteins. Each snRNP contains a short snRNA that base pairs with the pre-mRNA or with other snRNAs. U1 pairs with the 5' splice site; the protein U2AF binds the polypyrimidine tract and the AG; and U2 pairs with the branch point, bulging out the branch adenosine. The U4/U6.U5 tri-snRNP then joins, and a series of ATP-driven rearrangements releases U1 and U4. U6 replaces U1 at the 5' splice site and pairs with U2, forming an RNA catalytic center that binds two magnesium ions. The spliceosome is therefore a ribozyme, an RNA enzyme.

The chemistry is two transesterification reactions. In step 1, the 2'-OH of the branch-point adenosine attacks the phosphate at the 5' splice site. Exon 1 is released with a free 3'-OH, and the 5' end of the intron becomes linked to the branch A by an unusual 2'-5' phosphodiester bond, forming a looped lariat. In step 2, the 3'-OH of exon 1 attacks the phosphate at the 3' splice site, joining exon 1 to exon 2 and releasing the intron as a lariat, which is debranched and degraded. Each step exchanges one phosphodiester bond for another, so no net energy is needed for the chemistry itself; ATP is spent on the rearrangements that ensure accuracy.

Not all introns need a spliceosome. Group I introns, found in some rRNA, organelle and phage genes, splice themselves using a free guanosine as the attacking nucleophile; Thomas Cech's discovery of self-splicing in Tetrahymena rRNA was the first demonstration that RNA can be a catalyst. Group II introns, in organelles and bacteria, also self-splice but form a lariat, as the spliceosome does, and are thought to be its evolutionary ancestors. Introns in tRNA genes are removed by protein enzymes. A rare minor spliceosome removes a small class of introns with different consensus sequences.

Alternative splicing

One gene can produce several mRNAs by alternative splicing. About 95 percent of human multi-exon genes are alternatively spliced. The main patterns are exon skipping (a cassette exon is included or left out), mutually exclusive exons, alternative 5' or 3' splice sites, and intron retention. Choices are controlled by short sequences in exons and introns, splicing enhancers bound by SR proteins and splicing silencers bound by hnRNP proteins, whose levels differ among tissues. Classic examples include sex determination in Drosophila, in which female-specific Sex-lethal protein sets off a cascade of alternative splicing decisions, and calcitonin/CGRP, in which thyroid cells make the hormone calcitonin and neurons make a neuropeptide from the same gene.

Polyadenylation

Except for replication-dependent histone mRNAs, which end in a stem-loop, nearly all Pol II mRNAs end in a poly(A) tail. The signal is AAUAAA (or a close variant), located 10 to 30 nucleotides upstream of the cleavage site, plus a GU- or U-rich downstream element. The factor CPSF (cleavage and polyadenylation specificity factor) binds AAUAAA and CstF binds the downstream element, both recruited from the CTD. The RNA is cleaved, usually just after a CA, and poly(A) polymerase adds about 200 to 250 adenosines in mammals without any template. Nuclear poly(A)-binding protein controls the tail length. The tail, bound in the cytoplasm by poly(A)-binding protein (PABP), promotes export, stabilizes the mRNA and enhances translation; gradual shortening of the tail is a main route to mRNA decay. Many genes have more than one poly(A) site, so alternative polyadenylation can produce mRNAs with longer or shorter 3' UTRs and different stabilities.

RNA editing, export and quality control

RNA editing changes the nucleotide sequence of an RNA after transcription. In the human intestine, the enzyme APOBEC1 deaminates one cytosine in the apolipoprotein B mRNA, converting a CAA (glutamine) codon to UAA, a stop codon. The intestine therefore makes the short apoB-48 for chylomicrons, while the liver, which does not edit, makes the full-length apoB-100 for VLDL and LDL from the same gene. ADAR enzymes deaminate adenosine to inosine, which is read as guanosine; editing of a glutamate receptor mRNA in neurons changes a glutamine codon (CAG) to one read as arginine (CGG), altering the channel's calcium permeability. In trypanosome mitochondria, guide RNAs direct the insertion and deletion of hundreds of uridines.

Mature mRNAs are exported through nuclear pores as protein-bound particles. During splicing, an exon junction complex (EJC) is deposited about 20 to 24 nucleotides upstream of each exon-exon junction. If a ribosome reaches a stop codon while an EJC still lies more than about 50 to 55 nucleotides downstream, the stop is judged premature and the mRNA is destroyed by nonsense-mediated mRNA decay (NMD). NMD protects cells from truncated proteins, and it explains why many nonsense mutations produce almost no mRNA at all.

Section 4The genetic code

A triplet, nonoverlapping code

With four bases and twenty amino acids, a code of single bases (4 words) or pairs (4² = 16 words) is too small; triplets give 4³ = 64 combinations, enough for twenty amino acids with room to spare. The genetic evidence came in 1961 from Francis Crick, Sydney Brenner and colleagues, who used the acridine dye proflavin to make single-base insertions (+) and deletions (−) in the rIIB gene of phage T4. Any single + or − mutation destroyed function, but combining a + with a nearby − often restored it, because the frame was shifted only in the short stretch between them. Most strikingly, three + mutations together (or three −) also restored function, while two did not. Only a code read in groups of three, starting from a fixed point and without punctuation between codons, explains that result. The same experiments showed that the code is nonoverlapping (each base belongs to one codon in a given frame) and suggested that it is degenerate, since a garbled stretch between compensating mutations rarely produced a stop.

Cracking the code

The biochemical solution began the same year. Marshall Nirenberg and Heinrich Matthaei added a synthetic RNA made only of uridine, poly(U), to a cell-free extract of E. coli with radioactive amino acids. The extract made polyphenylalanine, so UUU codes for phenylalanine. Poly(A) gave polylysine and poly(C) gave polyproline. Random copolymers, made from mixtures of two nucleotides in known proportions, revealed which bases (but not their order) were in other codons. Two further methods assigned the rest. In the triplet binding assay of Nirenberg and Philip Leder, single trinucleotides were mixed with ribosomes and charged tRNAs; only the tRNA matching the trinucleotide bound to the ribosomes and was trapped on a filter. Har Gobind Khorana synthesized RNAs with repeating sequences, whose products in each reading frame confirmed the assignments. By 1966 all 64 codons were known.

The codon table

The table shows the code's main properties. It is degenerate (redundant): 61 sense codons specify 20 amino acids, so most amino acids have more than one codon, called synonymous codons. Leucine, serine and arginine have six; methionine and tryptophan only one. It is unambiguous: each codon specifies only one amino acid. Synonymous codons usually differ only at the third position, and amino acids with similar properties tend to have similar codons (codons with U in the second position encode hydrophobic amino acids), so many point mutations are silent or conservative. AUG is the start codon, encoding methionine at the start and inside proteins. UAA, UAG and UGA are stop codons (nicknamed ochre, amber and opal); no normal tRNA reads them. The code is read 5' to 3', and a single strand can be read in three reading frames, so a double-stranded DNA has six. An open reading frame (ORF) is a long stretch that begins with AUG and has no stop codon in frame.

Wobble

Each codon is read by base pairing with the anticodon of a tRNA, and the two pair antiparallel: the first codon base (5') pairs with the third anticodon base (3'). Cells have fewer kinds of tRNA than the 61 sense codons, because, as Crick proposed in his wobble hypothesis, the first (5') base of the anticodon can pair with the third (3') base of the codon in nonstandard ways. The first two codon positions pair strictly. Inosine (I), a modified purine made from adenosine at the wobble position, is especially flexible.

A nearly universal code

The same code is used by bacteria, archaea, plants and animals, which is why a human gene can be expressed in E. coli to make insulin. There are exceptions. In vertebrate mitochondria, UGA codes for tryptophan, AUA for methionine, and AGA and AGG are stops. Some Mycoplasma read UGA as tryptophan, and some ciliates read UAA and UAG as glutamine. Two rare amino acids are added by recoding stop codons in special contexts: selenocysteine at certain UGA codons when a downstream stem-loop (the SECIS element) is present, and pyrrolysine at UAG in some archaea.

Section 5Translation and protein processing

Charging tRNAs: the second genetic code

Transfer RNAs are adaptors, typically about 76 nucleotides long, that fold into a cloverleaf secondary structure and an L-shaped tertiary structure. The acceptor stem ends in the sequence CCA at the 3' end, and the amino acid is attached to the terminal adenosine. The opposite end of the L carries the anticodon loop. tRNAs contain many modified bases, such as dihydrouridine, pseudouridine (ψ) and inosine.

The link between an amino acid and its anticodon is made by aminoacyl-tRNA synthetases, usually one for each amino acid. Each enzyme works in two steps: amino acid + ATP → aminoacyl-AMP + PP_i, then aminoacyl-AMP + tRNA → aminoacyl-tRNA + AMP. Hydrolysis of the pyrophosphate makes the reaction irreversible, so charging costs two high-energy phosphate bonds. The resulting ester bond is itself high energy and later powers peptide bond formation. Synthetases recognize their tRNAs by the anticodon and by features of the acceptor stem, and many have a separate editing site that hydrolyzes wrong amino acids, such as valine on tRNA^Ile. Accuracy here is critical because the ribosome checks only the codon and anticodon, not the amino acid. This was shown by chemically converting cysteine already attached to tRNA^Cys into alanine: the ribosome inserted alanine at cysteine codons.

Ribosomes

Ribosomes are made of two subunits, each composed of rRNA and proteins. Sizes are given in Svedberg units (S), which reflect sedimentation rate and do not add.

The ribosome has three tRNA binding sites. The A site (aminoacyl) accepts each incoming charged tRNA, the P site (peptidyl) holds the tRNA carrying the growing chain, and the E site (exit) holds the uncharged tRNA before it leaves. The small subunit contains the decoding center, where 16S rRNA bases inspect the minor groove of the codon-anticodon helix to confirm correct pairing at the first two positions. The large subunit contains the peptidyl transferase center, and high-resolution structures published in 2000 showed that no protein lies within reach of the active site. Peptide bond formation is catalyzed by 23S rRNA, so the ribosome is a ribozyme.

Initiation

In bacteria, initiation factor IF3 keeps the subunits apart so a free 30S subunit can bind mRNA. The 30S subunit finds the start codon by base pairing between the Shine-Dalgarno sequence (consensus AGGAGG, about 5 to 10 nucleotides upstream of the AUG) and a complementary sequence near the 3' end of 16S rRNA. Because each gene in a polycistronic mRNA has its own Shine-Dalgarno sequence, ribosomes can start at several internal AUGs. IF2, bound to GTP, brings the initiator tRNA carrying N-formylmethionine (fMet-tRNA^fMet) directly to the P site, while IF1 blocks the A site. The 50S subunit then joins, IF2 hydrolyzes its GTP, the factors leave, and a 70S initiation complex is ready with fMet in the P site and an empty A site.

In eukaryotes, the small subunit with several initiation factors and eIF2-GTP-Met-tRNA_i forms a 43S complex. The eIF4F complex (cap-binding eIF4E, the scaffold eIF4G and the helicase eIF4A) binds the 5' cap; eIF4G also binds poly(A)-binding protein, which circularizes the mRNA and links the tail to efficient translation. The 43S complex is recruited to the cap and scans 5' to 3', using ATP to unwind structure, until it finds an AUG in a favorable context, the Kozak sequence (gccRccAUGG, with a purine at −3 and G at +4 being most important). GTP hydrolysis by eIF2 releases the factors and the 60S subunit joins. Some viral and cellular mRNAs bypass the cap by using an internal ribosome entry site (IRES). Phosphorylation of eIF2 by stress-activated kinases blocks its recycling and shuts down most translation, a central control point in viral infection and cellular stress.

Elongation

Each elongation cycle has three steps. (1) Decoding: EF-Tu bound to GTP delivers an aminoacyl-tRNA to the A site. If codon and anticodon match, the decoding center triggers GTP hydrolysis and EF-Tu-GDP leaves (EF-Ts regenerates EF-Tu-GTP); a mismatched tRNA usually leaves before it can be used. (2) Peptide bond formation: the amino group of the A-site amino acid attacks the ester bond that holds the peptide to the P-site tRNA, transferring the entire chain onto the A-site tRNA, which is now one residue longer. (3) Translocation: EF-G bound to GTP moves the mRNA and tRNAs by exactly one codon, shifting the peptidyl-tRNA from A to P and the uncharged tRNA from P to E, from which it leaves. Bacterial ribosomes add about 15 to 20 amino acids per second and eukaryotic ribosomes about 5, with an error rate of roughly 1 in 10⁴. Many ribosomes can translate one mRNA at once, forming a polysome.

Termination and the energy budget

When a stop codon enters the A site, no tRNA matches. Instead a release factor binds: in bacteria RF1 recognizes UAA and UAG and RF2 recognizes UAA and UGA; in eukaryotes eRF1 recognizes all three. A conserved GGQ motif of the release factor positions a water molecule in the peptidyl transferase center, which hydrolyzes the bond between the polypeptide and the P-site tRNA, releasing the protein. RF3 (or eRF3) and GTP help remove the release factor, and recycling factors split the ribosome so it can start again.

Protein targeting, folding and modification

Most proteins need help after synthesis. Molecular chaperones such as Hsp70 bind exposed hydrophobic stretches of nascent chains, and chaperonins (GroEL with GroES in bacteria, TRiC in eukaryotes) provide an enclosed chamber where some proteins fold. In bacteria the formyl group of fMet is removed, and in both domains the initial methionine is often trimmed off.

Where a protein goes depends on signals in its sequence. Proteins destined for secretion, the plasma membrane, or the endoplasmic reticulum (ER), Golgi or lysosomes begin with a signal sequence of about 15 to 30 amino acids with a hydrophobic core. As it emerges from the ribosome, the signal recognition particle (SRP), which contains the Pol III product 7SL RNA, binds it and pauses translation. SRP then docks on its receptor in the ER membrane and hands the ribosome to the Sec61 translocon, through which the chain is threaded as it is made (cotranslational translocation). Signal peptidase removes the signal sequence. In the ER, proteins acquire N-linked glycosylation on asparagine in the sequence Asn-X-Ser/Thr and form disulfide bonds; further glycosylation and sorting take place in the Golgi. Other proteins are made on free ribosomes and carry signals for the nucleus (nuclear localization signals), mitochondria or peroxisomes.

Many proteins are activated by cleavage. Insulin is made as preproinsulin; removing the signal peptide gives proinsulin, and removing the connecting C peptide in secretory granules gives mature two-chain insulin. Phosphorylation, acetylation, lipid attachment and other post-translational modifications regulate activity, and attachment of chains of ubiquitin marks proteins for destruction by the proteasome.

The differences between 70S and 80S ribosomes are the basis of many antibiotics. Mitochondrial ribosomes resemble bacterial ones, which helps explain some side effects, such as the hearing loss caused by aminoglycosides in people with a particular mitochondrial rRNA variant.

Words worth knowing