Biology by Bradford
support
One base out of three billion. Sometimes nothing happens. Sometimes a red blood cell folds into a crescent. The difference is worth understanding exactly.
A gene mutation is a structural change to a gene at the molecular level; the base sequence itself is different. Only three things can happen to a base.
One nucleotide is replaced by another. The gene stays the same length, so the reading frame is untouched: at most one codon changes.
One or more nucleotides are added. Unless the number is a multiple of three, every codon downstream is re-read: a frameshift.
One or more nucleotides are lost. Same rule: not a multiple of three means everything after it shifts.
D1.3 is about gene (small-scale) mutations. Whole-chromosome changes (non-disjunction, translocation, inversion, duplication) are a different level of organisation, covered in D2.1 and D3.2.
These are the first twelve codons of HBB, the human beta-globin gene, shown as the template strand: the one RNA polymerase reads, so the mRNA underneath is its complement, with U in place of T. Choose a tool, then tap a base. The mRNA and the polypeptide update, and the verdict tells you which kind of mutation you just made.
Two rules do all the work. A substitution can only reach one codon, so the damage is capped, and the degenerate code means it often does nothing at all. An insertion or deletion moves the frame, so every codon after it is read wrongly and the protein is usually junk from that point on. Try "Insert three" to see the one indel that behaves itself.
Four outcomes, and which one you get depends on the genetic code, not on the size of the change. Every one of these is a single base.
The fourth ending is the quiet one. Most of the genome is non-coding. A substitution there usually changes nothing, unless it lands in a promoter, an enhancer or a splice site, in which case the protein is fine but the cell makes the wrong amount of it, at the wrong time.
64 codons, 20 amino acids. Most amino acids have several codons, usually differing at the third base, so third-base substitutions are often silent. That redundancy is a built-in shock absorber.
A single-nucleotide polymorphism: a base substitution that has been inherited and is common in a population. SNPs are the raw material of genetic fingerprinting, ancestry tests and disease-risk studies.
Swapping one hydrophobic amino acid for another (conservative) often leaves the fold intact. Swapping a charged one for a hydrophobic one (non-conservative) can wreck it, which is exactly what happens next.
Follow a single A→T substitution in HBB all the way from the double helix to a blocked capillary.
Glutamic acid is negatively charged and sits happily on the outside of haemoglobin, in the water. Valine is hydrophobic. A greasy patch on the surface is a sticky patch.
When deoxygenated, the valine patch on one molecule slots into a pocket on another. Haemoglobin polymerises into rigid fibres that deform the cell into a sickle. Sickle cells block capillaries and are destroyed early: pain crises and anaemia.
Heterozygotes (HbA/HbS) have some resistance to malaria. Where malaria is common, natural selection keeps the allele at high frequency despite the cost to homozygotes: the same mutation being harmful and helpful, depending on the environment.
A ribosome cannot see codons. It reads three bases at a time from the start codon and never re-checks. Move one base and it keeps counting in threes: just in the wrong place.
Why indels are usually worse than substitutions. A substitution damages at most one codon. An indel corrupts every codon downstream, and often creates a premature stop, so the polypeptide is both wrong and truncated. The exception is an indel of three (or a multiple of three): one amino acid is added or lost and the rest of the frame survives. Cystic fibrosis is most often caused by exactly that: a three-base deletion, ΔF508.
Two sources: mistakes the cell makes on its own, and damage from outside. Neither aims at anything.
DNA polymerase mispairs a base roughly once in 10⁷ and proofreading catches most of it. What escapes proofreading and mismatch repair becomes a permanent mutation at the next round of replication.
UV makes adjacent thymines bond into a dimer that kinks the helix and blocks replication. Ionising radiation (X-rays, gamma, radon) breaks the sugar-phosphate backbone.
Benzopyrene in tobacco smoke, mustard gas, nitrosamines; they bind to or chemically alter bases so the wrong partner is inserted. Each leaves a characteristic mutational signature in a tumour's genome.
HIV copies RNA to DNA with reverse transcriptase, which has no proofreading. Errors accumulate fast, so the viral population varies enormously, which is why drug resistance and vaccine design are so hard.
D1.3.5: it can happen anywhere. Mutation is random with respect to need; no base is protected because it matters, and none is targeted because a change would help. It can land in a coding sequence, an intron, a promoter, non-coding DNA or mitochondrial DNA. What is not random is what happens next: natural selection removes the harmful and keeps the useful.
The same molecular change has completely different consequences depending on whether the cell will ever become a gamete.
Reach, not severity. A somatic mutation affects one cell and its descendants: a mole, a patch of tissue, or, if it hits a cell-cycle gene, a tumour. A germline mutation is copied into every cell of the offspring and every generation after, which is why it can enter the gene pool and a somatic mutation cannot.
A changed protein usually works worse: cystic fibrosis, Huntington's, or a mutation in a tumour-suppressor gene such as p53 that lets a cell divide unchecked.
Silent substitutions, changes in non-coding DNA, conservative missense in an unimportant part of a protein. They accumulate quietly and are the basis of molecular clocks.
Antibiotic resistance in bacteria; lactase persistence in adult humans; the ApoA-1 Milano variant in Limone sul Garda that protects against atherosclerosis. Beneficial in that environment.
Meiosis and sexual reproduction shuffle alleles; they cannot invent one. Every allele that exists began as a mutation. No mutation, no variation, no evolution.
To find out what a gene does, make it inoperative and see what goes wrong. Deliberate mutation as an experimental method.
The control is the whole experiment. Both groups are the same strain, the same age and kept in the same conditions; the only difference is that one has both copies of p53 disabled. The knockouts are born and develop normally, so p53 is not needed to build a mouse, but tumours, most often thymic lymphoma, appear within months. That contrast is the evidence that p53's normal job is to halt the cell cycle when DNA is damaged, and it is why a mutated p53 turns up in so many human cancers.
Gene knockout uses genetic engineering to remove or inactivate one specific gene. The organism that develops without it is compared with normal controls; the difference is evidence of what the gene was doing.
Mice with p53 knocked out develop normally but grow tumours very early. That is how the gene's role as a tumour suppressor, halting the cell cycle when DNA is damaged, was established.
For mice, zebrafish, Drosophila, Arabidopsis and yeast, libraries of knockout strains exist; one gene disabled in each. A researcher can order the strain rather than make it.
Knockouts often show nothing. That does not mean the gene is useless: another gene may be redundant with it, or the effect may only appear under stress. Absence of a phenotype is weak evidence.
CRISPR is a bacterial memory of past viruses. Cas9 is the enzyme that acts on it. Give Cas9 a guide RNA of your own design and it will cut wherever you like. Tap a component.
Tap Cas9, the single guide RNA, the PAM sequence, the double-strand cut, the unwound region or the target DNA.
The cut is the easy part. What the cell does with the broken ends is what decides whether you have disabled a gene or rewritten it.
Non-homologous end joining glues the ends back together and is error-prone: it usually loses or gains a few bases. That indel causes a frameshift and the gene is dead, which is how CRISPR makes a knockout.
Supply a DNA template with the sequence you want and homology-directed repair copies it in. This is how a mutation is corrected or a new gene inserted: precise, but far less efficient than NHEJ.
Bacteria store fragments of past viral DNA as spacers between CRISPR repeats. Transcribed into crRNA, these guide Cas9 to destroy that virus if it returns. An adaptive immune system in a prokaryote.
A 20-base guide can partially match sequences elsewhere in a genome of billions of bases. Off-target cuts put indels in genes nobody intended to touch: the central safety problem, and the reason newer base and prime editors avoid cutting both strands.
You met this mutation in section 02. In 2023 a CRISPR therapy for it was approved in the UK and the US. It never touches the mutation.
The mutation is still there. Casgevy does not repair HBB. The A→T is untouched, the gene still codes valine, and the patient still makes HbS. What the edit breaks is a switch: an enhancer that turns on BCL11A, the protein that shuts down fetal haemoglobin shortly after birth. Break the enhancer and the fetal gene comes back on, and HbF is enough to stop HbS polymerising. A disease caused by a mutation, treated by adding another one on purpose.
Blood stem cells are taken from the patient’s own bone marrow, cut by Cas9 in the lab, and returned. Somatic and one patient only: nothing is inherited, because no gamete is ever involved.
Correcting HBB would need HDR, which is far less efficient in stem cells. NHEJ knockouts work. So the designers went looking for a target where breaking something is the cure, and found the off-switch for fetal haemoglobin.
The marrow has to be cleared with chemotherapy before the edited cells are put back, with everything that carries. Add a price of over two million dollars and you have the ethics discussion below, in one real case.
Nobody will ask you for BCL11A. They may ask for an example of gene editing used therapeutically, and why somatic editing is permitted where germline is not. This is that example, and the detail is only here so the answer isn’t hand-waving.
Tap a claim to see the strongest reply from the other side. Exam answers that only give one column don't reach the top band.
Somatic editing treats one patient and stops there: Casgevy, approved in 2023, edits a patient's own bone-marrow cells to treat sickle cell disease. Germline editing changes embryos, so the change is inherited by everyone after. Most jurisdictions permit the first and prohibit the second.
Disease-resistant and higher-yield crops; animal models of human disease; microbes engineered to make drugs, enzymes and fuels. Gene drives could suppress malarial mosquitoes, and could not be recalled.
Compare the same gene across species and some stretches are nearly identical in a mouse, a fly and you. Two hypotheses compete to explain why, and you are asked for both.
The sequence codes for something essential, so almost any mutation is harmful or lethal and its carrier leaves fewer offspring. The variant is removed as fast as it appears: purifying selection. Histone genes, rRNA genes and homeobox genes are the standard examples.
Some regions simply mutate less often: local sequence context, chromatin state and more efficient repair all lower the rate. The sequence looks conserved without selection having done the work.
Compare the region with a nearby non-functional stretch of similar composition. If both are unchanged, the mutation rate is low. If only the functional one is unchanged, selection is doing it.
Conserved non-coding sequences are usually regulatory. Researchers use conservation to decide which regions are worth knocking out or editing: the two techniques above depend on this one.
Tap a node to light up its links. The pink dashed link is the sentence that connects this whole topic to evolution.
Drag a term into a gap, or tap a term and then tap a gap. Two terms belong nowhere.
Single best answer. You get the reasoning as soon as you commit. HL items are marked.