DNA Structure, Replication and Gene Expression

Every cell in your body carries about two meters of DNA folded into a nucleus a few micrometers wide, and nearly every time a cell divides it copies all three billion base pairs with fewer than one mistake per billion. That same molecule is read thousands of times a day to make the proteins that build and run you. This chapter explains how scientists proved DNA is the genetic material, how its structure makes copying possible, how genes are transcribed and translated into proteins, and what happens when the sequence changes.

Section 1The search for the genetic material and DNA structure

Protein or DNA?

By the 1930s, geneticists knew that genes sit on chromosomes and that chromosomes are made of two kinds of molecule: protein and DNA. Most scientists bet on protein. Proteins are built from 20 different amino acids and come in endless shapes, so they seemed rich enough to carry complex information. DNA, with only four kinds of nucleotides, looked too simple, perhaps just a structural scaffold. A series of experiments, mostly with bacteria and viruses, overturned that view.

Griffith and transformation

In 1928, Frederick Griffith was studying two strains of the bacterium Streptococcus pneumoniae in mice. The S strain made a smooth polysaccharide capsule that protected it from the immune system and was lethal. The R strain lacked a capsule, formed rough colonies and was harmless. Heat killed S cells were also harmless. But when Griffith injected a mixture of heat killed S cells and living R cells, the mice died, and he recovered living S bacteria from their blood. Something from the dead S cells had permanently changed the R cells into S cells, and the change was inherited by their descendants. Griffith called this transformation: a change in genotype and phenotype caused by uptake of external DNA, although he did not yet know DNA was the agent.

Avery, MacLeod and McCarty

In 1944, Oswald Avery, Colin MacLeod and Maclyn McCarty purified the transforming principle from S cell extracts. They treated samples with enzymes that destroy specific molecules. Protein digesting enzymes and RNA digesting enzymes left transformation intact, but DNase, which destroys DNA, abolished it. Their conclusion that the transforming substance was DNA was strong, but many biologists remained skeptical, partly because they suspected tiny amounts of contaminating protein and partly because bacteria were not yet thought to have genes like those of complex organisms.

Hershey and Chase

The decisive experiment came in 1952 from Alfred Hershey and Martha Chase, using bacteriophage T2, a virus that infects Escherichia coli. A phage is little more than a protein coat around DNA, and when it infects a cell it reprograms the cell to make new phages. The question was which part enters the cell and carries the instructions. Hershey and Chase exploited a chemical difference: proteins contain sulfur (in methionine and cysteine) but almost no phosphorus, while DNA contains phosphorus (in its backbone) but no sulfur.

1. One batch of phages was grown with radioactive sulfur (³⁵S), labeling their protein. Another batch was grown with radioactive phosphorus (³²P), labeling their DNA.
2. Each batch infected separate cultures of E. coli. After a few minutes, the cultures were agitated in a blender to shake loose any phage parts outside the cells.
3. Centrifugation spun the heavier bacterial cells into a pellet and left the lighter phage coats in the liquid (supernatant).
4. In the ³⁵S batch, most radioactivity stayed in the liquid. In the ³²P batch, most radioactivity was in the pellet, inside the cells, and some appeared in the next generation of phages.

The phage DNA, not its protein, had entered the cells and directed the production of new viruses. DNA was the genetic material.

Chargaff's rules

Meanwhile, Erwin Chargaff analyzed the base composition of DNA from many species. He found two patterns, now called Chargaff's rules. First, base composition varies from one species to another, which is exactly the kind of variety a genetic material needs. Second, within any species the amount of adenine equals thymine and the amount of guanine equals cytosine. In human DNA, for example, A and T are each about 30 percent and G and C are each about 20 percent. Nobody could explain the second rule until the double helix was discovered.

Building the double helix

A DNA nucleotide has three parts: a five carbon sugar called deoxyribose, a phosphate group and a nitrogenous base. The bases come in two families. The purines, adenine (A) and guanine (G), have a double ring. The pyrimidines, cytosine (C) and thymine (T), have a single ring (in RNA, uracil replaces thymine). Carbons of the sugar are numbered with primes: the base attaches to the 1' carbon, and the phosphate attaches to the 5' carbon.

Nucleotides join through phosphodiester bonds: the phosphate on the 5' carbon of one nucleotide links to the hydroxyl on the 3' carbon of the next. The result is a sugar-phosphate backbone with bases sticking out to the side. A strand has direction. One end has a free phosphate on a 5' carbon (the 5' end) and the other a free hydroxyl on a 3' carbon (the 3' end). Sequences are written 5' to 3' by convention.

Rosalind Franklin, working with Maurice Wilkins, produced X-ray diffraction images of DNA fibers. Her best image showed an X pattern that indicated a helix, and her measurements showed that the helix had a width of about 2 nm and repeating features at 0.34 nm and 3.4 nm. She also concluded that the phosphates lay on the outside. In 1953, James Watson and Francis Crick used her data, Chargaff's rules and model building to propose the structure.

• DNA is a double helix: two strands wound around each other, usually right handed (the B form).
• The sugar-phosphate backbones are on the outside and the bases point inward, stacked like the rungs of a twisted ladder.
• The two strands are antiparallel: one runs 5' to 3' and its partner runs 3' to 5'.
• Bases pair by hydrogen bonds: A with T (two hydrogen bonds) and G with C (three hydrogen bonds). Each pair is one purine plus one pyrimidine, which keeps the width a constant 2 nm and explains Chargaff's equalities.
• Adjacent base pairs are 0.34 nm apart and one full turn contains about 10 base pairs (10.5 in solution) in 3.4 nm.
• The winding creates a major groove and a minor groove, where proteins can read the edges of the bases without unwinding the helix.

Watson and Crick noted that the pairing rules immediately suggest how DNA could be copied: each strand carries the information to rebuild its partner. Because the strands are complementary, knowing one sequence gives you the other: the partner of 5'-ATGCCA-3' is 3'-TACGGT-5', written by convention as 5'-TGGCAT-3'.

Packing DNA into chromosomes

A bacterium such as E. coli has one circular chromosome of about 4.6 million base pairs, which is supercoiled and packed with proteins into a region called the nucleoid. Many bacteria also carry small circular plasmids. A human diploid cell has 46 linear chromosomes totaling about 6 billion base pairs (3.2 billion per haploid set), which would stretch about 2 meters if laid end to end. Eukaryotic DNA is packed in several levels with proteins, together called chromatin.

During interphase, loosely packed euchromatin is accessible and can be transcribed, while densely packed heterochromatin (for example at centromeres, telomeres and the Barr body) is mostly silent. Cells change packing by chemically modifying histones, a topic in Chapter 7.

Section 2DNA replication and repair

Semiconservative replication

Watson and Crick predicted that each strand of the double helix serves as a template for a new partner strand. Under this semiconservative model, each daughter molecule would contain one old strand and one new strand. Two alternatives were possible: a conservative model, in which the parent helix stays intact and a completely new copy forms, and a dispersive model, in which every strand is a patchwork of old and new segments.

In 1958, Matthew Meselson and Franklin Stahl settled the question. They grew E. coli for many generations in a medium containing heavy nitrogen (¹⁵N), so all their DNA was dense. They then moved the cells to a medium with ordinary light nitrogen (¹⁴N) and extracted DNA after each round of replication. Density gradient centrifugation separated DNA by density. After one round, all the DNA formed a single band of intermediate (hybrid) density, which ruled out the conservative model. After two rounds, there were two bands, one hybrid and one light, which ruled out the dispersive model. Only the semiconservative model fit both results.

Getting started: origins and forks

Replication begins at specific sequences called origins of replication. The bacterial chromosome has a single origin, and replication proceeds in both directions until the two forks meet on the opposite side of the circle. Eukaryotic chromosomes are far longer and replication forks move more slowly (about 50 nucleotides per second in humans, compared with about 500 to 1,000 in bacteria), so each chromosome uses hundreds to thousands of origins. A human cell copies its 6 billion base pairs in a few hours this way. At each origin, proteins open a replication bubble, and at each end of the bubble is a Y shaped replication fork, where parental strands are unwound and new strands are made.

The proteins at the fork

Why replication is lopsided: leading and lagging strands

Every DNA polymerase has two firm limits. It cannot start a strand from scratch; it can only add nucleotides to an existing 3' hydroxyl, which is why a primer is needed. And it can add nucleotides only to the 3' end, so new DNA always grows in the 5' to 3' direction. The energy for each addition comes from the incoming nucleoside triphosphate (such as dATP), which loses two phosphates as pyrophosphate when it joins the chain; hydrolysis of the pyrophosphate makes the reaction strongly favorable.

Because the two parental strands are antiparallel, these rules force the two new strands to be made differently. On one template, polymerase can follow the fork as it opens, adding nucleotides continuously after a single primer. This new strand is the leading strand. On the other template, synthesis in the 5' to 3' direction means moving away from the fork. The polymerase must make the strand in a series of short pieces, each started from its own primer as the fork exposes more template. This is the lagging strand, and its pieces are Okazaki fragments, about 1,000 to 2,000 nucleotides long in bacteria and 100 to 200 in eukaryotes. DNA polymerase I then replaces each RNA primer with DNA, and DNA ligase joins the fragments into a continuous strand.

Eukaryotic differences

Eukaryotes use the same logic with different enzymes. A complex of DNA polymerase α and primase starts each strand or fragment; DNA polymerase ε extends the leading strand, and DNA polymerase δ extends the lagging strand. Primers are removed by an RNase and a flap endonuclease rather than by a polymerase I, and the gaps are filled and sealed in the same way. Behind the fork, histones are rapidly reloaded so the new DNA is packed into nucleosomes.

Proofreading: how replication stays accurate

Base pairing alone would allow about one error in every 10⁴ to 10⁵ nucleotides. Replicative DNA polymerases have a 3' to 5' exonuclease activity that lets them back up, remove a mismatched nucleotide they have just added and try again. Proofreading improves accuracy about a hundredfold. After replication, mismatch repair enzymes scan the new strand for remaining errors, recognize which strand is new (in E. coli because it is not yet methylated) and replace the wrong nucleotide. Together these steps bring the final error rate to about one in 10⁹ to 10¹⁰ nucleotides, roughly one mistake per human cell division.

The end replication problem and telomerase

Linear chromosomes have a problem at their ends. When the last RNA primer on a lagging strand is removed, there is no upstream 3' end from which polymerase can fill the gap. Each round of replication therefore leaves the daughter chromosome slightly shorter. Eukaryotic chromosome ends are capped by telomeres, repeated noncoding sequences (TTAGGG repeated hundreds to thousands of times in humans) bound by protective proteins. Telomeres act as buffers, so losing some sequence costs no genes, and they prevent cells from mistaking chromosome ends for broken DNA.

Telomeres shorten with each division in most somatic cells, which contributes to a limit of roughly 50 divisions for many human cells in culture. In germ cells and many stem cells, the enzyme telomerase restores telomere length. Telomerase is a reverse transcriptase that carries its own RNA template; it extends the 3' overhang of the chromosome with new repeats, after which primase and polymerase fill in the complementary strand.

DNA damage and repair

Beyond replication errors, DNA is constantly damaged. Each human cell suffers thousands of spontaneous lesions every day: bases lose amino groups (cytosine deaminates to uracil), purines fall off the backbone, reactive oxygen species oxidize bases, ultraviolet light fuses adjacent pyrimidines into thymine dimers, and ionizing radiation breaks both strands. Because the information is stored twice, once on each strand, most damage can be repaired by cutting out the damaged section and using the undamaged strand as a template. Cells have well over a hundred repair enzymes.

Section 3Transcription and RNA processing

From genes to proteins

How does a DNA sequence produce a trait? Early clues came from inherited metabolic diseases and from George Beadle and Edward Tatum, who studied mutant strains of the bread mold Neurospora that could not make particular nutrients. Each mutant lacked one enzyme in a biosynthetic pathway, which led to the one gene, one enzyme hypothesis. It was later refined to one gene, one polypeptide, since many proteins are not enzymes and many proteins have several polypeptide chains, and refined again once it became clear that some genes make functional RNAs that are never translated, and that one gene can give several polypeptides through alternative splicing.

The flow of genetic information is summarized by the central dogma: DNA is copied to DNA (replication), DNA is transcribed into RNA, and RNA is translated into protein. Retroviruses such as HIV add a reverse step, using reverse transcriptase to copy RNA into DNA. RNA is the intermediate because it lets a cell make many working copies of a gene without exposing the master copy, and because it allows each gene's output to be regulated separately.

Template strand and coding strand

For any given gene, only one strand of DNA is read. This is the template strand (also called the noncoding or antisense strand). The RNA is complementary to it and antiparallel to it, and is built 5' to 3' as the polymerase reads the template 3' to 5'. The other DNA strand, the coding strand (sense strand), therefore has the same sequence as the RNA, except with T in place of U. When geneticists publish a gene sequence, they write the coding strand, so you can read codons directly from it. Different genes on the same chromosome may use different strands as templates.

RNA polymerase and the stages of transcription

RNA polymerase pries the two DNA strands apart and links ribonucleotides in the 5' to 3' direction. Unlike DNA polymerase, it needs no primer, and it proofreads much less, so transcription makes more errors than replication; this is tolerable because each RNA is temporary and made in many copies. Transcription has three stages.

1. Initiation. RNA polymerase binds a promoter, a DNA sequence just upstream of the gene that marks where to start and which strand to read. The first nucleotide transcribed is called +1; positions before it are given negative numbers.
2. Elongation. The polymerase unwinds about 10 to 20 base pairs at a time, forming a transcription bubble, and adds about 40 nucleotides per second in eukaryotes. Behind it, the DNA rewinds and the new RNA peels away. Several polymerases can transcribe one gene at once, one behind another.
3. Termination. The polymerase stops and releases the RNA.

Bacteria compared with eukaryotes

Eukaryotic RNA processing

In eukaryotes, the initial transcript from a protein coding gene is a pre-mRNA (primary transcript). Before it leaves the nucleus, enzymes modify both ends and remove interior segments.

• 5' cap. Soon after transcription starts, a modified guanine nucleotide (7-methylguanosine) is attached to the 5' end through an unusual 5' to 5' triphosphate linkage.
• Poly-A tail. After the transcript is cleaved downstream of the polyadenylation signal, an enzyme adds 50 to 250 adenine nucleotides to the 3' end.

The cap and tail help export the mRNA from the nucleus, protect it from degradation by exonucleases, and help ribosomes attach to the 5' end in the cytoplasm. The mature mRNA also keeps untranslated regions (5' UTR and 3' UTR) at each end that are not translated but influence stability and translation.

Splicing

Most eukaryotic genes are interrupted. The segments that are expressed in the final mRNA are exons; the intervening sequences removed from the transcript are introns. A typical human gene has about eight introns, and introns make up a large share of the length of most genes. RNA splicing removes the introns and joins the exons. It is carried out by the spliceosome, a large complex of proteins and small nuclear ribonucleoproteins (snRNPs), each containing small nuclear RNA. The snRNAs base pair with conserved sequences at the intron boundaries, almost always GU at the 5' end of the intron and AG at the 3' end, and with a branch point adenine inside the intron. The intron is released as a looped structure called a lariat, and the exons are joined. The snRNAs themselves catalyze key steps, which makes the spliceosome a ribozyme, an RNA molecule that acts as an enzyme. Some introns even splice themselves out without any proteins.

Why genes are in pieces

Introns may seem wasteful, but split genes provide two major advantages. First, alternative RNA splicing lets one gene produce different mRNAs, and therefore different polypeptides, by including or skipping particular exons. More than 90 percent of human multi-exon genes are alternatively spliced, which helps explain how roughly 20,000 protein coding genes yield a much larger number of proteins. Second, exons often encode separate structural and functional regions of a protein called domains. Introns make it easier for crossing over to shuffle exons between genes (exon shuffling), which can create new proteins during evolution.

Section 4The genetic code and translation

Cracking the code

Four nucleotide bases must specify 20 amino acids. A code of one base per amino acid gives only 4 words, and two bases give only 16 (4²). Three bases give 4³ = 64 combinations, more than enough. Experiments confirmed that the code is read in nonoverlapping triplets called codons. In 1961, Marshall Nirenberg and Heinrich Matthaei added an artificial mRNA made only of uracil (poly-U) to a cell free protein synthesis system and obtained a polypeptide made only of phenylalanine, so UUU must code for phenylalanine. By the mid 1960s, all 64 codons had been deciphered.

The code has a handful of properties worth knowing precisely.

• Triplet and nonoverlapping: each codon is three bases, and successive codons do not share bases.
• Start and stop: AUG codes for methionine and also serves as the usual start codon. UAA, UAG and UGA are stop codons that code for no amino acid. That leaves 61 sense codons for 20 amino acids.
• Redundant (degenerate) but not ambiguous: most amino acids have more than one codon (leucine, serine and arginine have six each), but each codon specifies only one amino acid. Synonymous codons usually differ in the third base.
• Read in a fixed reading frame: the ribosome reads codons in groups of three from the start codon, so shifting the frame by one base changes every downstream codon.
• Nearly universal: the same code operates in bacteria, plants and animals, with minor variations in mitochondria and a few microorganisms. This is why a human gene inserted into bacteria can be expressed to make human insulin, and it is strong evidence that all life shares a common ancestor.

The players in translation

Transfer RNA (tRNA) molecules are the adapters that link codons to amino acids. Each tRNA is a single strand of about 80 nucleotides that folds, through internal base pairing, into a cloverleaf in two dimensions and an L shape in three. At one end of the L is the anticodon, a triplet that base pairs with a complementary codon in the mRNA; at the other end, the 3' end ending in CCA, is the attachment site for the amino acid. A tRNA with anticodon 3'-AAG-5' pairs with the mRNA codon 5'-UUC-3' and carries phenylalanine.

Each amino acid is joined to its correct tRNAs by a specific aminoacyl-tRNA synthetase. There are 20 of these enzymes, one per amino acid, and they spend ATP to form the bond; the resulting aminoacyl-tRNA (a "charged" tRNA) carries the energy that later forms a peptide bond. The synthetases are the real translators of the code, because they match the language of nucleic acids to the language of proteins, and they proofread their own work.

Cells have only about 45 kinds of tRNA, not 61, because of wobble: the pairing rules between the third base of a codon and the first base (5' end) of the anticodon are relaxed. For example, a tRNA whose anticodon has the modified base inosine (I) at the wobble position can pair with codons ending in U, C or A.

Ribosomes bring mRNA and charged tRNAs together and catalyze peptide bond formation. Each ribosome has a large and a small subunit, made of ribosomal RNA (rRNA) and proteins. Bacterial ribosomes are 70S (50S plus 30S subunits) and eukaryotic cytoplasmic ribosomes are 80S (60S plus 40S); S units measure sedimentation rate, so they do not add. A ribosome has an mRNA binding site and three tRNA binding sites.

• The A site (aminoacyl) holds the tRNA carrying the next amino acid.
• The P site (peptidyl) holds the tRNA carrying the growing polypeptide chain.
• The E site (exit) is where discharged tRNAs leave the ribosome.

Initiation

Initiation brings together the mRNA, the first tRNA and both ribosomal subunits. In bacteria, the small subunit binds the mRNA at a purine rich Shine-Dalgarno sequence just upstream of the start codon, which pairs with the 16S rRNA. In eukaryotes, the small subunit, already carrying the initiator tRNA, binds the 5' cap and scans down the mRNA until it reaches the first AUG in a favorable sequence context. The initiator tRNA carries methionine (a formylated methionine, fMet, in bacteria) and is the only tRNA that enters directly into the P site. Proteins called initiation factors help, using energy from GTP, and then the large subunit joins, completing the initiation complex with the A site empty and ready.

Elongation

Elongation adds amino acids one at a time in a three step cycle, each cycle moving the ribosome one codon toward the 3' end of the mRNA.

1. Codon recognition. An aminoacyl-tRNA, escorted by an elongation factor bound to GTP, pairs its anticodon with the codon in the A site. GTP hydrolysis increases accuracy by giving the ribosome a moment to reject a wrong pairing.
2. Peptide bond formation. The rRNA of the large subunit catalyzes a bond between the amino group of the new amino acid in the A site and the carboxyl end of the growing chain in the P site. The chain transfers to the tRNA in the A site. Because rRNA, not protein, does this chemistry, the ribosome itself is a ribozyme, and the catalytic activity is called peptidyl transferase.
3. Translocation. A second elongation factor, using GTP, moves the ribosome forward one codon. The tRNA carrying the chain shifts to the P site, the empty tRNA shifts to the E site and leaves, and the A site opens for the next codon.

The polypeptide grows from its amino (N) terminus to its carboxyl (C) terminus, and exits through a tunnel in the large subunit. Each amino acid added costs about four high energy phosphate bonds: two to charge the tRNA (ATP to AMP) and two GTPs during elongation. A bacterial ribosome adds about 15 to 20 amino acids per second; a eukaryotic ribosome adds a few per second.

Termination and what happens next

When a stop codon reaches the A site, no tRNA matches it. Instead a protein called a release factor binds the stop codon and causes a water molecule to be added to the chain instead of an amino acid. This hydrolyzes the bond between the polypeptide and the tRNA in the P site, freeing the polypeptide; the subunits and mRNA then separate, using more GTP. A single mRNA is typically translated by many ribosomes at once, spaced along it like beads, forming a polyribosome (polysome). In bacteria, ribosomes attach to an mRNA even while it is still being transcribed.

A new polypeptide begins folding as it emerges, often helped by chaperone proteins, and many undergo post-translational modifications: addition of sugars, lipids or phosphate groups, removal of the first methionine, or cutting into smaller pieces (insulin is made as a single chain, proinsulin, that is cut to give two linked chains). Proteins destined for secretion, for membranes or for lysosomes begin with a signal peptide of about 20 hydrophobic amino acids. As it emerges, a signal recognition particle binds it, pauses translation and docks the ribosome on the rough ER, where the growing chain is threaded into the ER lumen or membrane. Proteins made on free ribosomes stay in the cytosol or are imported into the nucleus, mitochondria or other organelles.

Section 5Mutations and their effects

What a mutation is

A mutation is a permanent change in the nucleotide sequence of DNA. Mutations range from a single base pair to large chromosomal rearrangements (Chapter 5). They are the ultimate source of all new alleles and therefore of the genetic variation on which evolution acts, yet they are also the cause of inherited diseases and of cancer. Where a mutation occurs matters. A germline mutation arises in cells that give rise to gametes and can be passed to offspring, so it will be present in every cell of a child who inherits it. A somatic mutation arises in a body cell; it affects only that cell's descendants and is not inherited, but it can contribute to cancer.

Point mutations: substitutions

A point mutation changes a single nucleotide pair. In a base pair substitution, one pair is replaced by another. Substitutions are classified chemically as transitions (a purine replaced by a purine or a pyrimidine by a pyrimidine, such as A to G) or transversions (a purine replaced by a pyrimidine or the reverse); transitions are more common. More useful for predicting consequences is the effect on the protein.

• A silent mutation changes a codon to another codon for the same amino acid, usually at the third position (AAG to AAA still codes for lysine). The protein is unchanged, although rarely the change affects splicing or translation speed.
• A missense mutation changes a codon so that it specifies a different amino acid. The effect depends on the substitution. A conservative change swaps an amino acid for a chemically similar one (leucine for isoleucine) and may be harmless. A nonconservative change at a critical position can destroy function.
• A nonsense mutation changes an amino acid codon into a stop codon, ending translation early and producing a truncated, usually nonfunctional protein. Cells often destroy such mRNAs through a surveillance process called nonsense mediated decay.

Insertions, deletions and frameshifts

Adding (insertion) or removing (deletion) nucleotide pairs is usually more damaging than a substitution. Because mRNA is read in triplets, inserting or deleting a number of nucleotides that is not a multiple of three shifts the reading frame, a frameshift mutation. Every codon downstream of the change is read in the wrong groups of three, so the amino acid sequence becomes garbled and a stop codon usually appears soon. Inserting or deleting exactly three nucleotides (or a multiple of three) keeps the frame and adds or removes amino acids. The most common cystic fibrosis allele, ΔF508, is such an in frame deletion: three base pairs are missing and the protein lacks a single phenylalanine at position 508, which causes it to misfold and be destroyed before reaching the cell surface.

Repeat expansions

Some genes contain short tandem repeats that can lengthen when DNA polymerase slips during replication. In Huntington disease, a CAG repeat inside the coding sequence of the huntingtin gene, normally fewer than about 27 copies, expands to 36 or more, adding a long stretch of glutamine that makes the protein toxic to neurons. In fragile X syndrome, the most common inherited cause of intellectual disability, a CGG repeat in the 5' UTR of the FMR1 gene expands to more than 200 copies, becomes heavily methylated and shuts the gene off. Repeats tend to grow from one generation to the next, so these disorders often begin earlier or become more severe in later generations, a pattern called anticipation.

Causes of mutations

Spontaneous mutations arise from normal processes: rare mispairing during replication that escapes proofreading, slippage at repeats, loss of purine bases and deamination of cytosine to uracil. Mutations caused by external agents, mutagens, are called induced mutations.

• Base analogs resemble normal bases but pair incorrectly; 5-bromouracil resembles thymine but sometimes pairs with guanine.
• Chemical modifiers such as alkylating agents add groups to bases and change their pairing.
• Intercalating agents such as ethidium bromide and acridine dyes slip between stacked base pairs, distort the helix and cause insertions or deletions (frameshifts).
• Ultraviolet light creates thymine (pyrimidine) dimers that block replication and lead to errors if unrepaired; this is the main cause of skin cancers.
• Ionizing radiation (X-rays, gamma rays) breaks DNA strands, including double strand breaks.
• Many chemicals become mutagenic only after liver enzymes modify them, as with benzo[a]pyrene in tobacco smoke, which binds guanine and causes characteristic mutations in the p53 gene of lung cancers.

Because most carcinogens are mutagens, chemicals are screened with the Ames test. It uses mutant Salmonella bacteria that cannot make the amino acid histidine. The bacteria are exposed to the test chemical, usually mixed with a liver extract to mimic metabolism, and plated on medium lacking histidine. Only bacteria in which a new mutation has restored the histidine gene can grow, so the number of colonies measures the chemical's mutagenic strength.

Mutations, function and dominance

Most mutations in coding regions that change a protein reduce or abolish its function. These loss of function mutations are usually recessive, since one working allele often makes enough product (Chapter 5). A gain of function mutation gives a protein a new or increased activity and is usually dominant; a single changed copy of the RAS gene that locks its protein in the "on" state can drive a cell toward cancer. A dominant negative mutation makes a faulty protein that interferes with the normal one, as when mutant collagen chains spoil the triple helices in some forms of osteogenesis imperfecta. In cancer biology, gain of function mutations turn growth promoting proto-oncogenes into oncogenes, while loss of function mutations disable tumor suppressor genes such as p53 and BRCA1.

Most new mutations, however, have little or no effect, because they fall outside genes, are silent or cause conservative changes. Each human is born with a few dozen new mutations not present in either parent. A small fraction of mutations are beneficial in a particular environment, and these are the raw material that natural selection can spread through a population.

Words worth knowing