Stages Of Protein Synthesis

Transcription and Translation, Actually

Protein synthesis is two separate events that happen in different places. The first part copies DNA into mRNA. The second part reads that mRNA and chains amino acids together into a protein. That's it. Most textbooks split it into three stages for each part, but the reality is messier than that diagram you saw in high school. Here's how it actually looks when you're dealing with it in a lab or trying to troubleshoot why a recombinant protein isn't expressing.

Stages Of Protein Synthesis: What Actually Happens

Stage 1 — Initiation. RNA polymerase binds to the promoter region on the DNA. In eukaryotes this requires transcription factors to help it find the right spot. Once that complex forms, the DNA double helix unwinds locally and the enzyme starts building an RNA strand complementary to the template strand. For translation initiation, the ribosome small subunit binds the 5' cap of the mRNA in eukaryotes, scans to the first AUG codon, and then the large subunit joins. In prokaryotes it's different — the Shine-Dalgarno sequence on the mRNA pairs with the 16S rRNA in the ribosome, positioning the start codon directly in the P site. If you're working with cloned genes, getting this ribosome binding site right is where most people mess up. Put a eukaryotic RBS into a prokaryotic system and nothing happens. Stage 2 — Elongation. RNA polymerase marches along the DNA, adding nucleotides at roughly 20 to 40 bases per second in bacteria, maybe slower in eukaryotes. The RNA peels away from the DNA as it goes and the helix re-forms behind it. Meanwhile the original DNA strand stays intact and can be transcribed again. For translation, the ribosome has three sites: A, P, and E. A new aminoacyl-tRNA enters at the A site, matching its anticodon to the mRNA codon. The ribosome catalyzes peptide bond formation between the amino acid in the P site and the one in the A site. Then translocation happens — the ribosome shifts one codon downstream, pushing the now-empty tRNA out through the E site and moving the peptidyl-tRNA from A to P. Each step takes maybe 15 to 20 milliseconds per amino acid in bacteria. That's about 15 to 20 amino acids per second. The ribosome doesn't just walk randomly either. It proofreads during elongation, rejecting mismatched tRNAs through kinetic discrimination. The error rate is roughly one mistake per 10,000 to 100,000 codons incorporated, which is surprisingly low for something moving this fast. Stage 3 — Termination. RNA polymerase hits a terminator sequence. In bacteria this is either a rho-dependent signal that chases down the polymerase, or a rho-independent hairpin structure that causes the enzyme to stall and release the RNA transcript. In eukaryotes the story is more complicated — the pre-mRNA gets cleaved and a poly-A tail is added, and polymerase continues transcribing for hundreds of bases past the cleavage site before eventually falling off. For translation termination, a stop codon (UAA, UAG, or UGA) enters the A site. There's no tRNA for these. Instead release factors bind, the ribosome hydrolyzes the bond between the finished polypeptide and the tRNA in the P site, and the whole complex falls apart. The newly freed protein then needs to fold, and that's where things get interesting because folding isn't part of the synthesis machinery itself.

I ran into a problem last year where we were expressing a membrane protein in E. coli and the protein kept precipitating. We assumed the issue was with the gene sequence. It wasn't. The real problem was that the protein was being synthesized so fast that hydrophobic regions exposed themselves before chaperones could engage. Slowing down the translation rate by optimizing the codon usage — swapping rare codons for more common ones in the hydrophobic transmembrane segments — actually improved solubility dramatically. Fast isn't always better when the protein needs time to fold co-translationally.

Post-Transcriptional Stuff Nobody Talks About Enough

The mRNA you get from transcription isn't ready to go. Eukaryotic pre-mRNA has introns that need splicing out, a 5' cap added, and a poly-A tail tacked on. This processing happens co-transcriptionally, meaning while RNA polymerase is still working. Spliceosomes recognize specific sequences at exon-intron boundaries and cut out the introns, ligating the exons back together. Alternative splicing means one gene can produce multiple different mRNA variants, which means multiple different proteins from the same genetic template. Your genome has roughly 20,000 protein-coding genes but estimates suggest the proteome contains over 100,000 distinct proteins, and alternative splicing is a huge reason why the numbers don't match up. RNA editing is another layer. The APOB gene in humans produces two different proteins from the same mRNA. In the liver the full-length apoB-100 is made, but in the intestine a cytidine deaminase converts a cytosine to uracil at a specific position, creating a premature stop codon that yields the shorter apoB-48. This isn't theoretical — mutations in that editing enzyme are linked to abetalipoproteinemia, a serious metabolic disorder. There's also RNA interference to consider. Small RNAs like miRNAs and siRNAs can bind to complementary mRNA sequences and trigger degradation or block translation. This is a major regulatory layer that most introductory courses gloss over. If you're doing anything with gene expression analysis, you need to account for miRNA activity or your results will be confusing. We spent weeks troubleshooting why a knockdown experiment was partially working in one cell line and not at all in another, and the difference came down to endogenous miRNA expression patterns between the two cell types.

Common Pitfalls

People often think transcription and translation are coupled in all organisms. They're not. In bacteria they can happen simultaneously because there's no nuclear membrane — the ribosome latches onto the mRNA while it's still being transcribed. In eukaryotes transcription happens in the nucleus, the mRNA gets processed, and only then does it get exported to the cytoplasm for translation. This spatial separation gives eukaryotes additional regulatory checkpoints that prokaryotes simply don't have. Another assumption that causes trouble is that the genetic code is universal. It basically is, but there are exceptions. Mitochondria use a slightly different code. UGA codes for tryptophan instead of being a stop codon in mammalian mitochondria. Some protozoans and bacteria have rearranged stop codons too. If you're cloning a mitochondrial gene into a bacterial expression system, or vice versa, you need to check the codon table for that specific organelle or organism. The biggest practical issue I see is people not accounting for mRNA stability. The half-life of mRNA varies enormously depending on sequence, structure, and cellular context. Bacterial mRNAs typically last only a few minutes. Eukaryotic mRNAs can range from 30 minutes to several days. If you're designing an expression construct and wondering why your protein levels are lower than expected, the problem might not be transcription or translation efficiency — it might be that your mRNA is being degraded faster than you think. AU-rich elements in the 3' UTR are a common culprit for rapid decay. Adding or removing these elements can change protein output by an order of magnitude without touching the coding sequence at all.

Quantitative Reality Check

Here's what the numbers actually look like in a typical bacterial cell. A single mRNA molecule might be translated by 50 to 100 ribosomes simultaneously, forming what we call a polysome or polyribosome. Each ribosome produces a complete protein in roughly 1 to 2 seconds for a small protein. So one mRNA can yield hundreds of protein copies per minute. Meanwhile the cell is constantly degrading both mRNA and protein, so you're dealing with a dynamic steady state, not a one-time production run. In eukaryotes the throughput is lower per ribosome but the regulation is far more sophisticated. Ribosome profiling experiments have shown that ribosomes don't move uniformly along an mRNA. They pause at certain codons, slow down at structural obstacles, and sometimes even back up. These pauses are functionally important — they give the nascent polypeptide time to fold properly or recruit necessary cofactors. Artificially removing these natural pause sites can actually reduce functional protein yield even though the total amount of protein synthesized goes up. More protein doesn't mean more useful protein. If you're studying this for an exam, focus on the mechanistic steps and the differences between prokaryotic and eukaryotic systems. If you're actually doing research with it, pay attention to mRNA stability, codon usage optimization, and co-translational folding. Those are the variables that make or break your experiment.