The Genetic Code and Translation

Change one base in a gene and the protein may not change at all. Add one base and every amino acid after it can change.

Both facts follow from how the code is written. In this lesson you will see how the code was worked out, who reads it inside the cell, and how to predict what a mutation does to the protein. You already know the twenty side chains from amino acids, and that an mRNA is made by transcription and, in eukaryotes, trimmed by RNA processing. A codon is a group of three bases in mRNA that specifies one amino acid or a stop.

How many letters per word? Crick and Brenner

Four bases must spell twenty amino acids. Pairs give only 16 words, too few. Triplets give 64, more than enough. That count argued for at least three, but it was arithmetic, not evidence.

The evidence came from phage genetics. In 1961 Francis Crick, Leslie Barnett, Sydney Brenner and Richard Watts-Tobin studied mutants in the rIIB gene of phage T4, the gene you met in complementation and pathways. Many mutants were made with proflavin, a dye they argued adds or removes one base. A frameshift is an insertion or deletion that is not a multiple of three, so every codon after it is read out of register. They sorted the mutants into two classes, called plus and minus. A plus with a nearby minus often restored function. Two pluses did not. Three pluses, or three minuses, restored wild or nearly wild growth in the combinations they tested.

Think of a sentence of three-letter words: THE FAT CAT ATE THE RAT. Insert one letter and the rest is nonsense; delete one nearby and the sense returns after the second mark. Insert three and only the words between them are wrong.

Their conclusion was careful: the code is read from a fixed starting point, in nonoverlapping groups of three bases or a multiple of three, with no punctuation between words. They did not say which triplet meant which amino acid, and they did not claim to have proved three.

Breaking the code in a test tube

The dictionary came from biochemistry. In 1961 Marshall Nirenberg and J. Heinrich Matthaei, at the National Institutes of Health, made an E. coli cell-free system: broken cells that still make protein when given an RNA. They added poly(U), a synthetic RNA of uridine only, plus radioactive amino acids. Phenylalanine went into an acid-insoluble product with many properties of polyphenylalanine. The reaction needed ribosomes and the soluble fraction. The experiment alone could not say how many U residues make one codeword; with the phage result, the likely answer was that UUU means Phe.

Random copolymers of two bases gave codon compositions but not orders. Two methods supplied the orders. In 1964 Nirenberg and Philip Leder showed that a single trinucleotide, mixed with ribosomes, holds the matching aminoacyl-tRNA on the ribosome: UUU bound Phe-tRNA, AAA bound Lys-tRNA, CCC bound Pro-tRNA, and dinucleotides did not work. Ribosome complexes stick to a filter, so each triplet could be tested on its own. Har Gobind Khorana’s laboratory made RNAs of exact repeating sequence and translated them. Poly(UC), read as UCU CUC UCU, gave alternating serine and leucine.

The finished table has 64 codons: 61 sense codons and 3 stops (UAA, UAG, UGA). AUG means Met both at the start and inside a gene. Most amino acids have several codons, a property called degeneracy, and synonymous codons usually differ at the third base. Crick’s wobble hypothesis of 1966 explains why: the third codon base can pair loosely with the first anticodon base, so one tRNA can read more than one codon. The table is nearly universal. Vertebrate mitochondria read UGA as Trp, some ciliates read UAA and UAG as Gln, and a UGA can encode selenocysteine when the mRNA carries a special signal.

Adaptors: tRNA and the synthetases

A codon cannot recognize an amino acid by shape. Crick had argued that an adaptor must carry each amino acid to the template, and transfer RNA is that adaptor. Each tRNA has an anticodon that pairs with the codon, and a 3’ CCA end where the amino acid is attached.

The matching is done by aminoacyl-tRNA synthetases, one family of enzymes for each amino acid in most organisms. Charging has two steps:

amino acid + ATP → aminoacyl-AMP + PPi aminoacyl-AMP + tRNA → aminoacyl-tRNA + AMP

Pyrophosphate is then hydrolyzed, so charging spends two high-energy phosphate bonds. Some synthetases edit, hydrolyzing a wrong amino acid before or after it reaches the tRNA. The ribosome checks only codon against anticodon. If the synthetase puts the wrong amino acid on a tRNA, the ribosome inserts it.

The ribosome at work

A ribosome has two subunits: 30S and 50S making a 70S ribosome in bacteria, 40S and 60S making an 80S ribosome in the eukaryotic cytosol. Svedberg values measure sedimentation, so they do not add. Three tRNA sites span the subunits: the A site (aminoacyl), the P site (peptidyl) and the E site (exit).

Initiation differs most between the two systems. In bacteria, a purine-rich Shine-Dalgarno sequence a few bases upstream of the start codon pairs with the 3’ end of 16S rRNA and places the start codon on the small subunit. The initiator is fMet-tRNA (formylmethionine) in most bacteria, and it can start at GUG as well as AUG. In eukaryotes, the 40S subunit with initiation factors and Met-tRNAi loads at the 5’ cap and scans 5’ to 3’ to an AUG, usually the first in good Kozak context (a purine at -3 and a G at +4). In Kozak’s preproinsulin reporter, changing bases around the AUG changed protein yield over a 20-fold range. Either way the initiator ends in the P site, and the large subunit joins.

Elongation repeats three steps. EF-Tu (eEF1A in eukaryotes) with GTP delivers an aminoacyl-tRNA to the A site; if it matches, GTP is hydrolyzed. The large subunit transfers the growing chain from the P-site tRNA to the amino group on the A-site amino acid. EF-G (eEF2) with GTP moves the ribosome one codon, so the empty tRNA goes to the E site and leaves. The chain grows from N terminus to C terminus. The catalyst of peptide bond formation is rRNA, the 23S rRNA in bacteria, which is why the ribosome is called a ribozyme.

cost per peptide bond ≈ 2 (charging) + 1 GTP (EF-Tu) + 1 GTP (EF-G) = 4 high-energy phosphate bonds

Termination happens when a stop codon enters the A site. No tRNA reads it. A release factor does: in bacteria RF1 reads UAA and UAG, RF2 reads UAA and UGA, and in eukaryotes eRF1 reads all three. The bond to the P-site tRNA is hydrolyzed, the protein leaves, and the ribosome is recycled. How fast all this runs, and when it is throttled, is the subject of mTOR and AMPK; how the chain folds is in protein 3D structure.

Reading frames, mutations and NMD

A single RNA strand has three reading frames, starting at base 1, 2 or 3. A double-stranded DNA has six, three on each strand. The frame that matters is set by the start codon.

Worked problem. An mRNA reads 5’-GCCAUGGCUUGGAAACCUGACUAGCC-3’. Use the standard table and assume translation starts at the first AUG.

1. Find the AUG: bases 4 to 6.
2. Split into triplets from there: AUG GCU UGG AAA CCU GAC UAG CC.
3. Look each up and recheck: AUG Met, GCU Ala, UGG Trp, AAA Lys, CCU Pro, GAC Asp, UAG stop.
4. Answer: Met-Ala-Trp-Lys-Pro-Asp, then stop. Six residues; the stop adds none.

The other two frames give different words. From base 2: CCA UGG CUU GGA AAC CUG ACU AGC (Pro-Trp-Leu-Gly-Asn-Leu-Thr-Ser). From base 3: CAU GGC UUG GAA ACC UGA (His-Gly-Leu-Glu-Thr-stop).

Now predict four mutations, each checked codon by codon:

• GCU to GCC: GCC is Ala. Met-Ala-Trp-Lys-Pro-Asp. Silent.
• GCU to ACU: ACU is Thr. Met-Thr-Trp-Lys-Pro-Asp. Missense, Ala to Thr.
• UGG to UGA: UGA is stop. Met-Ala, then stop. Nonsense.
• Insert G after the AUG: AUG GGC UUG GAA ACC UGA, so Met-Gly-Leu-Glu-Thr, then stop. Frameshift: every codon after the insertion changes, and a stop that was out of frame now comes early.

Silent does not always mean harmless. In human cell lines, synonymous changes can alter mRNA stability while the protein sequence stays the same.

A premature stop also changes the mRNA’s fate. In mammalian cells, splicing leaves an exon junction complex about 20 to 24 bases upstream of each exon-exon junction. A ribosome on the first round of translation displaces these complexes. If it stops more than about 50 to 55 bases upstream of the last junction, a complex remains downstream, and the mRNA is targeted by nonsense-mediated decay (NMD). A stop in the last exon, or within about 50 bases of the last junction, usually escapes and may yield a truncated protein. The rule is a tendency with exceptions, and it is drawn from mammalian cells, not yeast or bacteria. For how such mutations arise, see mutation.

Design problem

In an E. coli cell-free system, poly(AG) (AGAGAG…) directs a polypeptide of alternating Arg and Glu. Poly(AAG) (AAGAAG…) directs three homopolymers: polyLys, polyArg and polyGlu. Assign the codons AGA, GAG, AAG and GAA as far as the data allow, state each assumption, and name the experiment that settles what is left.