The Actual Mechanism, Not The Textbook Version
Protein synthesis is transcription followed by translation. That's the entire arc. The molecular biology departments love to draw it as two clean acts with a dramatic intermission, but in practice there are about forty regulatory checkpoints, a bunch of quality control failures, and plenty of things that go wrong if the conditions aren't right. I spent too many semesters watching students memorize the steps without understanding why the cell occasionally skips them entirely. There are really two phases, but each phase has substeps that matter more than people admit. Transcription comes first because the DNA template has to become mRNA before anything else happens. Then translation reads that mRNA and builds the polypeptide chain. That's the skeleton. The details are where the actual work lives. The RNA polymerase holoenzyme binds to the promoter region. In bacteria, this is straightforward - sigma factor recognizes the -10 and -35 boxes. In eukaryotes, you're dealing with TFIID, TBP, and a whole assembly line of general transcription factors before polymerase II even gets recruited. I've seen protocols fail because someone treated eukaryotic transcription like it was prokaryotic. It isn't. The chromatin structure alone can block initiation if histones aren't properly modified.
Once the polymerase is docked, elongation begins. Nucleotides are added 5' to 3'. The emerging RNA strand peels away from the DNA template as the polymerase moves forward. This is where people usually gloss over termination. In prokaryotes, Rho-dependent and Rho-independent mechanisms exist. The hairpin loop mechanism in intrinsic termination is elegant but finicky - a single misfolded sequence and the polymerase just keeps going. I once had a culture where the terminator sequence was one nucleotide off and we got readthrough transcripts that were two kilobases longer than expected. Took three weeks to figure out what happened. Eukaryotic transcription adds another layer. The primary transcript is a pre-mRNA that needs capping at the 5' end, polyadenylation at the 3' end, and splicing to remove introns. The cap isn't decorative. It's required for ribosome binding and nuclear export. Skip the cap and the mRNA gets degraded within minutes in the cytoplasm. The poly-A tail serves a similar protective function and also aids in translation efficiency. Splicing is where things get genuinely complicated. Alternative splicing means a single gene can produce multiple protein variants. The spliceosome makes decisions about which exons to include, and those decisions are regulated by cellular conditions. Stress, developmental stage, tissue type - they all influence splicing patterns.
The mRNA Has To Leave The Nucleus
This isn't a trivial step in eukaryotes. The mRNA has to pass through nuclear pore complexes, and the export machinery checks that the RNA is properly processed before it lets anything through. Unspliced or improperly capped transcripts get retained and degraded. I worked with a construct where the 5' UTR had a secondary structure that looked like it might trigger nuclear retention. It didn't, but it was close enough that we lost about half the yield compared to a clean control. The lesson was that UTR design matters more than most people calculate for. The mRNA lands on a ribosome and the reading frame gets established. The start codon - usually AUG - is recognized by the initiator tRNA carrying methionine. In eukaryotes, the small ribosomal subunit scans from the 5' cap until it finds the first AUG in a good Kozak context. In prokaryotes, the Shine-Dalgarno sequence positions the ribosome directly. If you're engineering expression systems, getting the initiation context right is the single highest-impact decision you'll make. A weak Kozak sequence can reduce protein yield by orders of magnitude compared to an optimal one. Elongation cycles through three sites in the ribosome: A, P, and E. The aminoacyl-tRNA enters at the A site, the peptidyl-tRNA sits at the P site, and the deacylated tRNA exits at the E site. Each cycle adds one amino acid. The ribosome is essentially a molecular machine that hydrolyzes GTP to power translocation. EF-Tu delivers the charged tRNA to the A site in bacteria, and eEF1A does the equivalent in eukaryotes. Proofreading happens here too - incorrect codon-anticodon pairs are rejected before peptide bond formation. This isn't perfect, but it keeps the error rate around one mistake per ten thousand residues, which is remarkably good for something happening at thousands of cycles per second.
Get the Full Details
I should mention that not all proteins start with methionine. In both prokaryotes and eukaryotes, the initiator methionine is frequently cleaved off by methionine aminopeptidase if the second residue is small enough. Alanine, glycine, proline, serine, threonine, and valine all permit this processing. So the final protein may or may not contain that initial Met depending on the sequence context. If you're expressing a recombinant protein and the expected molecular weight doesn't match, check whether post-translational methionine removal is happening.
Termination And Beyond
When a stop codon enters the A site, release factors recognize it instead of a tRNA. In bacteria, RF1 stops at UAA and UAG, RF2 stops at UAA and UGA. In eukaryotes, eRF1 handles all three. The peptidyl transferase center then hydrolyzes the bond between the polypeptide and the tRNA in the P site, releasing the chain. This is where the polypeptide is free, but it's far from done as a functional protein. Folding begins co-translationally. The N-terminal end of the protein starts folding while the C-terminus is still being synthesized. Chaperones like GroEL/GroES in bacteria and Hsp70 in eukaryotes assist with this process. Some proteins fold spontaneously, but many require assistance, especially ones with complex domains or disulfide bonds. I once spent two months trying to get a secreted protein to express in the right form, only to realize the issue wasn't the sequence - it was that the oxidizing environment in the cytoplasm prevented proper disulfide bond formation. Moving the construct to a periplasmic expression system in E. coli solved it almost immediately. The protein itself was fine. The cellular compartment was wrong. Post-translational modifications are where protein synthesis gets genuinely messy. Phosphorylation, glycosylation, ubiquitination, acetylation - these happen after the ribosome finishes and can dramatically alter protein function, localization, and stability. The same polypeptide chain can become fundamentally different molecules depending on which modifications are applied. This is also where most pharmaceutical interventions target, which is why understanding the full synthesis pathway matters for anyone working in drug development or molecular therapeutics.
Quality Control Failures
The nonsense-mediated decay pathway monitors for premature stop codons. If a stop codon appears upstream of an exon-exon junction, the mRNA gets degraded rather than producing a truncated protein. This is a legitimate safety mechanism, but it's also a common reason why expression constructs fail in heterologous systems. Intronless constructs or synthetic genes without proper junction positioning can accidentally trigger NMD. I've seen this kill expression more times than I can count. The workaround is usually adding introns at strategic positions or using expression systems that lack the NMD machinery entirely. Misfolded proteins get targeted for degradation by the ubiquitin-proteasome system. Sometimes this is desirable - it clears out defective products. Sometimes it's your actual protein getting torn apart because it's aggregating in the wrong way. Overexpression tends to worsen this problem because the chaperone capacity gets overwhelmed. Diluting the induction, lowering the temperature, or using weaker promoters can sometimes keep the synthesis rate slow enough that folding keeps up with production. The whole process is also sensitive to energy availability. Both transcription and translation require ATP and GTP. Nutrient starvation, hypoxia, or mitochondrial dysfunction can shut down protein synthesis at multiple points. This isn't theoretical - it's why cancer cells rewire their metabolism to maintain translation rates while normal cells in the same environment slow down. Understanding the energetic constraints of protein synthesis explains a lot about cellular behavior that pure sequence analysis doesn't.
