Translation and the Genetic Code: Overview

Translation is the direct target of a large fraction of clinically used antibiotics, precisely because bacterial and eukaryotic ribosomes differ enough in structure to allow selective toxicity: aminoglycosides (streptomycin, gentamicin) distort the 30S decoding center and increase misreading; tetracyclines block A-site tRNA accommodation; macrolides (erythromycin) plug the nascent peptide exit tunnel of the 50S subunit; oxazolidinones (linezolid) block A-site tRNA positioning near the PTC; and chloramphenicol directly inhibits peptidyl transfer. Understanding exactly which step of the elongation cycle each drug perturbs — decoding, peptidyl transfer, or translocation — is the basis of rational antibiotic design and of tracking resistance mutations that remodel rRNA at these exact contact points. Kinetic proofreading, first developed conceptually to explain translational fidelity, has since become a general framework applied across biology: DNA polymerase proofreading exonucleases, kinetochore–microtubule attachment error correction, and T-cell receptor antigen discrimination all use structurally analogous multi-step, energy-dissipating discrimination schemes. Ribosome profiling (Ribo-seq), which sequences the mRNA footprints protected by translating ribosomes, has become a standard genome-wide tool for measuring elongation rates, detecting ribosome stalling, pausing at rare codons, and upstream open reading frames, directly building on the mechanistic codon-by-codon stepping picture developed here. Efforts to expand the genetic code — incorporating unnatural amino acids via orthogonal aminoacyl-tRNA synthetase/suppressor-tRNA pairs that read reassigned stop codons — depend on precisely engineering around the termination mechanism and the fidelity checkpoints described above, and are now used to install click-chemistry handles, photocrosslinkers, and post-translational-modification mimics into recombinant proteins for structural biology and therapeutics. Cryo-electron microscopy has, since roughly 2015, resolved successive intermediate states of the translating ribosome at near-atomic resolution, turning what was once inferred kinetically into directly visualized structural snapshots of decoding, accommodation, and translocation intermediates.

Ask the tutor: How exactly does the ribosome's decoding center distinguish a correct codon–anticodon pair from a near-cognate mismatch using only rRNA bases, and what does this reveal about the elongation cycle's mechanics?

A ribosome is built from two subunits that perform conceptually distinct jobs. The small subunit (30S in bacteria, 40S in eukaryotes) contains the decoding center, where the geometry of the ribosomal RNA reads out whether a codon–anticodon minihelix has correct Watson–Crick geometry, independent of which particular bases are paired. The large subunit (50S bacterial, 60S eukaryotic) contains the peptidyl transferase center (PTC) — an active site built entirely from rRNA — and the exit tunnel through which the nascent polypeptide threads as it grows. Between the two subunits run three tRNA-binding sites, A (aminoacyl), P (peptidyl), and E (exit), through which every tRNA passes in strict order: A → P → E → out. The chemistry at the PTC is deceptively simple: it is nothing more than an aminolysis of an ester. The peptidyl group esterified to the P-site tRNA's 3′-OH is transferred onto the free α-amino group of the aminoacyl-tRNA sitting in the A site, extending the peptide by one residue and leaving the growing chain now esterified to the A-site tRNA. The mystery, historically, was where the catalysis came from. Crystal structures of the large subunit (Steitz and Moore, 2000) showed no protein side chain within 18 Å of the reaction center — only ribosomal RNA. The ribosome is the largest known ribozyme, roughly 2.5 megadaltons in bacteria, and its catalytic power comes overwhelmingly from substrate positioning and a proton-shuttle mechanism assisted by the tRNA's own 2′-OH, not from ionizable rRNA side chains acting as general acids/bases in the way a protein enzyme would use His or Asp.

Layered on this RNA core is a set of protein translation factors that use the binding energy of GTP hydrolysis as a molecular stopwatch and a source of directional bias. EF-Tu delivers aminoacyl-tRNA to the A site as an EF-Tu·GTP·aa-tRNA ternary complex; EF-G drives translocation of the tRNA–mRNA complex by exactly one codon after each peptide bond; and in eukaryotes an extensive family of eIFs assembles the initiation complex through a scanning mechanism rather than direct Shine-Dalgarno pairing. None of these factors touch the chemistry of bond formation — their entire job is kinetic and mechanical: to enforce that only correctly paired tRNAs linger long enough to react, and to reset the ribosome geometrically for the next codon.

The intrinsic free energy driving translocation and decoding comes from GTP hydrolysis; the intrinsic free energy driving peptide bond formation comes from the high-energy aminoacyl-tRNA ester bond created earlier by the tRNA and Aminoacyl-tRNA Synthetases charging reaction. The overall energetic accounting per residue added is:

where:

• $\Delta G_{\text{charging}}$ is the cost of activating an amino acid onto its tRNA (formally 2 ATP-equivalents, since AMP + PPi is produced and PPi hydrolysis pulls the reaction forward),
• $\Delta G_{\text{decoding, GTP}}$ is one GTP hydrolyzed by EF-Tu per correctly accommodated aminoacyl-tRNA,
• $\Delta G_{\text{translocation, GTP}}$ is one GTP hydrolyzed by EF-G per translocation step.

Reading: four phosphoanhydride bonds are broken per peptide bond formed, making translation among the most energetically expensive biosynthetic processes per bond formed — more costly than DNA replication per nucleotide. The fidelity of decoding is governed by a two-step kinetic proofreading scheme rather than by equilibrium binding alone:

where:

• $k/K$ is the specificity constant (analogous to $k_{cat}/K_M$) for GTPase activation by a cognate versus near-cognate codon–anticodon pair at the initial selection step,
• the exponent 2 arises because the same discrimination is applied independently at initial selection and again at the proofreading step after GTP hydrolysis and EF-Tu release, before peptide bond formation is allowed to proceed.

Reading: because discrimination is applied twice with an irreversible (GTP hydrolysis) step in between, the overall error rate is roughly the square of the single-step discrimination ratio — this is why translation achieves accuracy far beyond what simple base-pairing thermodynamics alone would predict. The rate of translation elongation sets the overall protein synthesis rate and can be written per ribosome as:

where:

• $k_{\text{elong}}$ is the codon-addition rate (residues per second per ribosome), typically 15–20 s⁻¹ in E. coli at 37 °C and slower (~3–8 s⁻¹) in eukaryotic cytosol,
• $N_{\text{codons}}$ is the length of the message in codons, giving the time to synthesize one complete protein as $N_{\text{codons}}/k_{\text{elong}}$.

Example 1 (foundational): energetic cost of synthesizing a 100-residue protein. Take a protein of 100 amino acids. Initiation consumes 1 GTP (fixed, once per chain, via IF2/eIF2). Elongation adds 99 peptide bonds (the first residue is already present as initiator fMet-tRNA, so 99 further residues must be added by elongation, each costing 2 ATP-equivalents in charging plus 2 GTP in decoding/translocation — the initiator itself was charged too, so its charging cost is a 100th "charging unit"). Explicitly: Charging cost: 100 residues × 2 ATP-equivalents = 200 ATP-equivalents (irreversible via PPi hydrolysis). Elongation GTP cost: 99 peptide bonds × 2 GTP (1 EF-Tu + 1 EF-G) = 198 GTP. Initiation GTP cost: 1 GTP (fixed overhead). Termination GTP cost: 1 GTP (RF3-mediated release factor recycling) + ribosome recycling factor/EF-G contributes ~1–2 more GTP for subunit dissociation.

This matches the per-residue figure of 4 derived above, with initiation/termination overhead becoming negligible for longer proteins (a 1000-residue protein would show a per-residue cost even closer to exactly 4.0, since the fixed ~4-bond overhead is diluted 10-fold). As a limiting case: as chain length $N \to \infty$, the per-residue cost converges to exactly 4, confirming the fixed initiation/termination costs are an overhead term, not a per-residue term.

Example 2 (applied): how long does it take to translate a 450-codon mRNA (bacterial), and what limits the rate of a polysome? E. coli elongation proceeds at roughly $k_{\text{elong}} \approx 17\ \text{s}^{-1}$ at 37 °C (a commonly cited value; ranges of 12–21 s⁻¹ appear across studies and conditions). For a message encoding a 450-residue protein (450 sense codons plus a stop), translation time per ribosome is:

But in vivo, mRNAs are simultaneously translated by many ribosomes arranged as a polysome, spaced roughly every 80 nucleotides (about 27 codons, close to the ribosome's physical footprint of ~30 codons including its two subunits and the mRNA path through the channel). The number of ribosomes that can simultaneously occupy a 450-codon message is approximately:

so the overall rate of protein production from this one mRNA is not limited by the 26.5 s per-ribosome transit time but by the initiation rate — how often a new ribosome can load onto the 5′ end and begin scanning/pairing — since transit and loading proceed in parallel across the polysome. If initiation occurs every ~2 seconds (typical for a moderately efficient bacterial mRNA), the steady-state protein output rate is:

which is far higher than $1/26.5\ \text{s}^{-1} \approx 0.038$ proteins/s that a single ribosome alone could produce — demonstrating explicitly why polysomes exist: they parallelize elongation so that initiation frequency, not elongation speed, becomes rate-limiting for total protein output from an mRNA.

Example 3 (exam/competition level): quantifying the fidelity gain from proofreading given experimental rate constants. Suppose kinetic measurements (broadly consistent with published EF-Tu decoding kinetics, e.g. Pape, Wintermeyer & Rodnina, 1999) give, for the initial selection step:

giving a first-step discrimination ratio:

and for the proofreading step (after GTP hydrolysis, before peptidyl transfer, governed by relative rates of accommodation versus rejection/dissociation of EF-Tu·GDP):

The proofreading discrimination ratio compares the partitioning fraction toward acceptance (accommodation) for near-cognate vs cognate:

Overall error frequency:

This is comfortably consistent with (indeed somewhat better than) the canonically quoted in vivo translational error rate of order $10^{-4}$, confirming that two multiplicatively combined, imperfect (~100- to ~200-fold) discrimination events are sufficient to build up to $10^{4}$–$10^{5}$-fold overall fidelity — a direct numerical demonstration of the proofreading-squaring argument, using genuinely asymmetric (non-round) rate constants rather than an idealized symmetric toy case.

Har Gobind Khorana, Marshall Nirenberg, and Robert Holley shared the 1968 Nobel Prize for cracking the genetic code itself, but the mechanistic understanding of how the ribosome actually catalyzes peptide bond formation — the central biochemical question this article addresses — owes most directly to Thomas A. Steitz, whose crystallographic determination of the large ribosomal subunit structure in 2000 settled a decades-old argument by showing unambiguously that the peptidyl transferase center is composed of RNA, with no protein within striking distance of the catalytic center. Steitz, working with Peter Moore, Nenad Ban, and Poul Nissen at Yale, solved the 2.4 Å structure of the Haloarcula marismortui 50S subunit bound to transition-state analogs, definitively establishing the ribosome as a ribozyme rather than a conventional protein enzyme — a finding that also reinforced the RNA World hypothesis for the early evolution of life, since a catalytic core built from RNA suggests translation itself is an evolutionary relic of an RNA-dominated biochemistry that predates the widespread use of protein enzymes.