RNA processing in eukaryotes: editing the message before use

In the last lesson, RNA polymerase copied a gene into a fresh RNA strand and let it go. If you were a bacterium, that strand would be ready to use almost immediately. A ribosome could grab it and start building protein, sometimes before the copy was even finished. But you are not a bacterium. You are a eukaryote, meaning your cells keep their DNA locked inside a nucleus, and that changes everything about what happens next. In a eukaryote, the fresh RNA strand is not a finished message. It is a rough draft, and the cell will not let it out of the nucleus until it has been edited in three specific ways. This lesson is about those edits, and about the strange and useful payoff hiding in the biggest of them.

We have a name for that rough draft. The raw strand that comes straight off the gene is the primary transcript, often called pre-mRNA (the "pre" meaning it is not yet a working messenger). Our job here is to turn pre-mRNA into a mature mRNA the cell can actually use.

The gene comes in pieces

Here is the fact that forces the whole editing step to exist. In a eukaryotic gene, the instructions for a protein are not written in one clean, continuous run. They come in segments, and between the useful segments sit long stretches that are not part of the final message at all.

The kept segments are called exons. A useful mnemonic: exons are the parts that are expressed, the parts that exit into the mature message. The interrupting stretches are called introns, the parts that lie intervening, inside the gene, between the exons. When RNA polymerase transcribes the gene, it copies everything faithfully, exons and introns alike, straight into the pre-mRNA. The rough draft contains both.

Now you can see the problem. If the cell tried to translate that pre-mRNA as written, the ribosome would read straight through the intron sequence and produce nonsense, almost certainly hitting a premature stop and truncating the protein. The introns are noise wedged between the signal. They have to come out, and the exons that remain have to be joined together in their original order, with no seam and nothing lost.

Splicing: cut the introns, weld the exons

That cut-and-join operation is called splicing. It is carried out by a machine called the spliceosome, which, fittingly, is built largely from RNA working together with protein.

Splicing has to be exact to the single letter. Remove one nucleotide too many or too few at a joint and everything downstream shifts, wrecking the protein (you will meet reading frames in the next lesson, but for now just trust that being off by one letter here is catastrophic). So how does the spliceosome know precisely where each intron begins and ends? The elegant answer is that the intron marks its own borders. In the great majority of introns, the sequence begins with the letters GU and ends with the letters AG, with a few more signal letters nearby. The spliceosome reads those marks, loops the intron out into a lariat, snips it free at both ends, and welds the two flanking exons directly together.

Introns out. Exons joined in order. What remains is a continuous coding message: the mature mRNA.

Two more edits, one at each end

Splicing is the dramatic one, but two smaller edits happen to the ends of the message, and both are about survival and recognition.

The front end (the 5 prime end) gets a 5 prime cap: a single modified guanine nucleotide stuck onto the very start, in an unusual backward linkage. The cap does two jobs. It shields the front of the message from enzymes that chew RNA from its ends, and it acts as a tag that the ribosome recognizes as "this is a real, finished message, translate me."

The back end (the 3 prime end) gets a poly-A tail: a long run of adenine (A) nucleotides, often a hundred or more, added on by a dedicated enzyme. Notice something important here. Those A's are not copied from the DNA. There is no run of T's in the gene templating them. The enzyme simply appends them. The tail also does two jobs: it protects the back end from being degraded, and its length works like a slow timer. As the message ages, the tail is gradually trimmed, and once it is worn down the message is marked for destruction. The tail is, in effect, the message's remaining shelf life.

The programmer's model: this is a build step

Here is the mental picture to carry. Think of the gene as a source file that contains sections wrapped in conditional-compilation guards, the kind of code that is present in the source but stripped out before the program actually runs. The introns are those conditionally-stripped sections. Splicing is the build or preprocess step that removes them and assembles the runnable message from the raw source. The cap and the tail are like the wrapper a linker adds so the runtime will accept and load the result.

This model is genuinely useful. It tells you that the mature mRNA is a processed artifact, not the raw source, and that the processing is a distinct stage with its own machinery.

Alternative splicing: one source, many builds

Now the payoff. If the introns are conditionally-stripped sections, then some exons can be treated as conditional too. The cell does not have to keep every exon every time. It can include a given exon in one version of the message and skip it in another, or choose between two mutually exclusive exons, or keep or drop a whole block.

This is called alternative splicing, and the programmer's version of it is feature flags. One source file, a set of toggles, and out come several different builds, each with a slightly different feature set. In the cell, one gene, spliced different ways, yields several different mature mRNAs, which translate into several different but related proteins. The variants a single gene can produce this way are called isoforms.

Take that seriously and a famous puzzle dissolves. The human genome (the complete DNA sequence) holds only around 20,000 protein-coding genes, far fewer than people once guessed. Yet the human proteome (the complete set of proteins the body can build) is much larger than that. How does a small parts list make a big catalog? Alternative splicing is a major part of the answer. The great majority of human genes with multiple exons are spliced in more than one way. One gene is not one protein. It is a small factory that can assemble a family of related products depending on context.

Key terms

pre-mRNA (primary transcript)

The raw RNA strand straight off the gene, still containing introns and not yet capped or tailed, so not yet usable.

exon

A segment of a gene that is kept in the mature message, the parts that are expressed and joined together to code the protein.

intron

A non-coding segment that interrupts the exons in a eukaryotic gene and is cut out of the RNA before translation.

splicing

The precise cut-and-join step that removes introns from the pre-mRNA and welds the remaining exons together in order.

5 prime cap

A modified guanine added to the front of the message that protects it and marks it to the ribosome as a real message to translate.

poly-A tail

A long run of A nucleotides added to the back of the message that protects it and acts as an aging timer for its lifetime.

alternative splicing

Choosing different combinations of exons from the same gene to make several different mature mRNAs, and thus several proteins.

isoform

One of the several related protein variants a single gene can produce through alternative splicing.

A note on the word editing

This lesson's title says the message is edited, and that is a fair plain-English description of capping, splicing, and tailing together. Be aware, though, that biologists reserve the narrower term RNA editing for a different phenomenon: enzymes that chemically change individual bases in an RNA after it is made, for example converting one base so the ribosome reads it as another. That base-changing kind of editing is real, is important in the nervous system, and can make a protein that the DNA sequence alone would never predict. It is not the same as splicing. When you see "RNA editing" as a technical term later, it means that narrow base-changing sense, not the whole processing pipeline. Keeping the two apart will save you confusion.