Bacterial and Viral Genetics

Bacteria and viruses have no meiosis, yet they exchange genes constantly, and that exchange is why antibiotic resistance spreads through a hospital in months and why a new influenza strain can appear almost overnight. The same tiny genomes gave geneticists their sharpest tools: billions of offspring in a single tube, selection for one rare recombinant among millions, and maps fine enough to resolve neighboring base pairs. This chapter shows how bacteria and phages move DNA, how each route of transfer becomes a mapping method, and how viral genomes, including retroviruses, are built and copied.

Section 1Bacterial genomes, plasmids and mutants

Why bacteria are such powerful genetic organisms

A culture of Escherichia coli in rich medium doubles about every 20 minutes, so a single cell becomes more than a billion cells overnight. Bacteria are haploid, which means every mutation shows its phenotype at once, with no dominant allele hiding a recessive one. And because you can spread a hundred million cells on one Petri plate and allow only a rare type to grow, you can find events that happen once in a million or once in a billion cells.

The bacterial chromosome

Most bacteria carry a single, circular, double stranded DNA chromosome. The E. coli K-12 chromosome is about 4.6 million base pairs (4.6 Mb) and encodes roughly 4,300 genes. It is compacted into an irregular region called the nucleoid. There is no nuclear membrane. Compaction comes from negative supercoiling, introduced by DNA gyrase (a type II topoisomerase), and from nucleoid associated proteins such as HU, H-NS and IHF that bend and bridge DNA.

Bacterial genomes are gene dense: about 88 percent of the E. coli chromosome codes for protein, introns are essentially absent, and functionally related genes are often clustered into operons that are transcribed as one messenger RNA. Replication begins at a single origin, oriC, and proceeds in both directions until the two forks meet in the terminus region opposite the origin.

Plasmids

Plasmids are small DNA molecules, usually circular and double stranded, that replicate independently of the chromosome. Each has its own origin of replication, which also sets its copy number: some plasmids are kept at one or two copies per cell under tight control, while others reach dozens or hundreds. Plasmids are generally not essential for growth under ordinary conditions, but the genes they carry can be decisive when conditions change. An episome is a plasmid that can either replicate on its own or integrate into the chromosome; the F factor is the classic example.

Notation for genotypes and phenotypes

Bacterial geneticists use a compact notation. Genotypes are written as three lowercase italic letters followed by an uppercase letter for the specific gene: leuA, lacZ, trpE. A superscript plus means the wild type allele (leu⁺, able to make leucine) and a superscript minus means a mutant allele (leu⁻, unable to make leucine). str^r means streptomycin resistant. Phenotypes are written in roman type with the first letter capitalized: Leu⁺, Lac⁻, Str^r.

Kinds of bacterial mutants

A prototroph is a wild type strain that grows on minimal medium: water, salts, a carbon source such as glucose and nothing else. From those raw materials it builds every amino acid, nucleotide and vitamin it needs. An auxotroph has lost the ability to make one or more of these and grows only if the missing nutrient is supplied. A his⁻ auxotroph, for example, grows on minimal medium plus histidine but not on minimal medium alone. Complete medium contains everything, so both prototrophs and auxotrophs grow on it.

• Nutritional (auxotrophic) mutants cannot synthesize a nutrient.
• Carbon source mutants cannot use a particular sugar; a Lac⁻ mutant cannot grow when lactose is the only carbon source.
• Resistance mutants survive an antibiotic or a phage that kills the wild type.
• Conditional mutants show the mutant phenotype only under restrictive conditions. A temperature sensitive mutant makes a protein that works at 30 °C (permissive) but unfolds at 42 °C (restrictive). They let geneticists study essential genes.

Finding mutants: selection, screening and replica plating

Resistance mutants are easy to find by direct selection: spread 10⁸ cells on a plate containing the antibiotic, and only resistant cells form colonies. Auxotrophs are harder, because the mutant grows on fewer media than the wild type, not more. Joshua and Esther Lederberg solved this in 1952 with replica plating. Cells are spread on a complete medium master plate, giving a few hundred colonies. A sterile velvet pad pressed onto the master picks up cells from every colony and can then print that exact pattern onto a series of fresh plates. Any colony present on the master and on complete medium but absent from minimal medium is an auxotroph, and its position tells you where to pick it from the master. Printing onto minimal medium supplemented with single nutrients then identifies exactly which nutrient it needs.

Do mutations arise before or after selection? The fluctuation test

In the early 1940s it was an open question whether phage T1 caused resistance in the few cells that survived it, or whether resistant mutants arose spontaneously and the phage merely revealed them. Salvador Luria and Max Delbrück answered with the fluctuation test (1943). They grew many small, independent cultures from a few cells each, then plated each entire culture with phage.

The logic is simple. If the phage induced resistance at the moment of plating, every culture would have about the same small chance per cell, and the counts would vary little from culture to culture, like samples from one flask. If mutations arose at random during growth, a mutation early in the history of one culture would leave a huge resistant clone (a "jackpot"), while most cultures would have few or none. The independent cultures showed enormous fluctuation, with many zeros and a few jackpots, while the samples from the single flask were uniform. Mutations are spontaneous and random; selection only reveals them.

Section 2Conjugation and mapping by interrupted mating

The discovery of gene exchange

In 1946 Joshua Lederberg and Edward Tatum mixed two E. coli auxotrophs. One strain was met⁻ bio⁻ thr⁺ leu⁺; the other was met⁺ bio⁺ thr⁻ leu⁻. Neither grew on minimal medium alone, but the mixture produced prototrophic colonies at about 1 in 10⁷ cells. Double reversion could not explain them. Bernard Davis then placed the two strains on opposite sides of a U-tube separated by a filter that let medium and large molecules through but not cells. No prototrophs formed, showing that exchange required cell to cell contact. This contact dependent transfer is conjugation.

The F factor and the three mating types

Conjugation depends on the F (fertility) factor, a plasmid of about 100 kb. A cell carrying F is F⁺; a cell lacking it is F⁻. The tra genes of F encode the sex pilus, which pulls an F⁻ cell close, plus the transfer machinery. Transfer begins when a relaxase nicks one strand at the origin of transfer (oriT). That single strand, 5′ end first, threads into the recipient while the donor replaces it by rolling circle replication; the recipient synthesizes the complementary strand. The result of an F⁺ × F⁻ mating is two F⁺ cells. F spreads rapidly, yet F⁺ cells rarely transfer chromosomal genes.

About once in 10⁴ to 10⁵ F⁺ cells, the F factor integrates into the chromosome by recombination between short homologous sequences (often insertion sequences) on F and on the chromosome. The result is an Hfr (high frequency of recombination) cell. When an Hfr cell conjugates, transfer still starts at oriT, but now the strand that follows is chromosomal DNA. Genes beside the integration site enter first, then genes farther away, in a fixed order and direction set by where and in which orientation F inserted. Transferring the entire 4.6 Mb chromosome takes about 100 minutes at 37 °C, and the mating pair almost always breaks apart before that. The part of F that lies behind oriT is transferred last, so the recipient of an Hfr mating nearly always remains F⁻.

Occasionally F excises from the Hfr chromosome imprecisely and carries a piece of neighboring chromosome with it. This plasmid is an F′ (F prime) factor. An F′ transfers efficiently like F, and the recipient becomes a partial diploid, or merozygote, carrying two copies of the captured genes.

Recombination in the recipient

The donor fragment that enters the recipient is linear and cannot replicate on its own. To be inherited, its genes must be recombined into the recipient's circular chromosome, and that requires an even number of crossovers, at least two, one on each side of the gene. A single crossover would open the circle into a linear molecule that cannot survive. The donor genes that do not recombine in are degraded. The recipient therefore needs a functional RecA protein.

Interrupted mating

In the mid 1950s Élie Wollman and François Jacob turned the ordered transfer of Hfr strains into a mapping method. They mixed an Hfr strain sensitive to streptomycin (str^s) with an F⁻ recipient that carried several auxotrophic markers and was streptomycin resistant (str^r). At intervals they removed samples and broke the mating pairs apart in a kitchen blender, then plated the cells on media containing streptomycin to kill the donors (counterselection) and lacking a particular nutrient to select for recipients that had received a donor allele.

Each donor gene first appeared among recombinants at a characteristic time of entry, and the frequency of recombinants for that gene then rose to a plateau. Genes that entered early reached higher plateaus than genes that entered late, because the longer the transfer runs, the more likely the pair is to break apart spontaneously. This decline is called the gradient of transfer. The time of entry measures position: the E. coli chromosome is mapped in minutes, from 0 to 100, with the thr locus placed at 0. One minute corresponds to roughly 46 kb.

Building a circular map from several Hfr strains

Different Hfr strains have F inserted at different places and in different orientations, so each transfers the chromosome starting from a different point and in one of two directions. No single Hfr can map the whole chromosome, because pairs break apart long before 100 minutes. But the orders from several strains overlap, and when you align them the only arrangement consistent with all of them is a circle.

Section 3Transformation and transduction

Transformation: taking up naked DNA

Transformation is the uptake of free DNA from the surroundings and its incorporation into the recipient's genome. It was the first form of bacterial gene transfer discovered. In 1928 Frederick Griffith found that heat killed virulent (smooth, encapsulated) Streptococcus pneumoniae could convert living harmless (rough) cells into virulent ones inside a mouse. In 1944 Oswald Avery, Colin MacLeod and Maclyn McCarty purified the "transforming principle" and showed that it was destroyed by DNase but not by protease or RNase. That experiment identified DNA as the genetic material.

Some species are naturally competent: they switch on a set of genes, usually in response to crowding or stress, that builds a DNA uptake machine. Streptococcus pneumoniae, Bacillus subtilis and Neisseria gonorrhoeae are classic examples. One strand of bound DNA enters and pairs with the homologous region of the chromosome with the help of RecA, forming a heteroduplex, and mismatch repair or the next round of replication resolves it into one recipient chromosome and one transformed chromosome. Because the entering strand is linear, at least two crossovers are again needed, one on each side of the marker.

E. coli is not naturally competent, so laboratories make it artificially competent, either by chilling cells in calcium chloride and giving a brief heat shock, or by electroporation, a short high voltage pulse that opens transient pores.

Mapping by cotransformation

A transforming fragment is only a small piece of the chromosome, typically a few tens of kilobases. Two genes can be transferred together on one fragment, an event called cotransformation, only if they lie close together. The frequency of cotransformation therefore measures linkage: the closer two genes, the more often they travel together. Two tests are used.

• Compare double transformants with the product of single rates. If genes a and b are far apart, a cell must take up two separate fragments, and the double transformation frequency is roughly the product of the single frequencies. If they are close, double transformants are far more common than that product.
• Order three genes by pairwise cotransformation. The two genes that are cotransformed least often are the outside genes; the third lies between them.

Transduction: phages as couriers

Transduction is gene transfer from one bacterium to another carried inside a phage particle. Zinder and Lederberg discovered it in 1952 with phage P22 in Salmonella: transfer crossed a U-tube filter and resisted DNase. There are two types, and they differ in what the phage packages.

Generalized transduction

In generalized transduction, any part of the host chromosome can be transferred. Phages such as P1 (E. coli) and P22 (Salmonella) degrade the host chromosome into fragments during a lytic infection and package DNA into heads by a headful mechanism: they fill the head with as much DNA as fits. Occasionally, about once in every thousand or so particles, a head is filled with a piece of host DNA instead of phage DNA. Such a transducing particle contains no phage genes. It injects its bacterial DNA into the next cell it infects, and that cell survives because no phage genes enter. If the donor fragment recombines into the recipient chromosome by two crossovers, the result is a stable transductant (complete transduction). A fragment that persists without recombining or replicating gives abortive transduction.

A P1 head holds about 100 kb, roughly 2 percent of the E. coli chromosome, or about 2 minutes of the map. Two genes can be cotransduced only if they lie within that distance, and the closer they are, the more often they are.

Specialized transduction

In specialized transduction, only genes next to a prophage's integration site can be transferred. The classic example is phage lambda (λ), which integrates at a single att site between the gal and bio genes of E. coli. When a lysogen is induced, the prophage normally excises precisely. Rarely (about once in 10⁵ to 10⁶ excisions), the excision cut is in the wrong place, and the circle that loops out contains some host DNA on one side (gal or bio) and lacks a corresponding piece of phage DNA on the other. The resulting particle, written λdgal (defective, carrying gal) or λdbio, can inject the host gene into a new cell. A lysate that contains such particles transduces gal or bio at high frequency, but never transduces genes elsewhere on the chromosome.

Mapping by cotransduction

Cotransduction frequency is used exactly like cotransformation. You select for one donor marker and score how many of the selected transductants also inherited a second, unselected marker. Close genes are cotransduced often; genes more than one transducing fragment apart are never cotransduced. For P1, Wu derived a useful approximation relating cotransduction frequency C to map distance d (in minutes) when the fragment length L is 2 minutes:

C = (1 − d / L)³

Three factor transduction crosses resolve gene order without any distance formula. Select for one marker and score the other two among the transductants. The class that requires four crossovers rather than two will be by far the rarest, and it identifies which gene is in the middle.

Section 4Bacteriophage genetics and fine structure mapping

Phage life cycles

Bacteriophages (phages) are viruses that infect bacteria. A phage such as T4 has a protein head holding its double stranded DNA genome (about 169 kb) and a tail that attaches to receptors on the cell surface and injects the DNA. In the lytic cycle, the phage takes over the cell: early genes shut down host functions and degrade host DNA, middle and late genes replicate phage DNA and build heads and tails, the DNA is packaged, and a lysozyme-like enzyme bursts the cell. At 37 °C, a T4 infection lasts about 25 minutes and releases a burst of roughly 100 to 200 new phages.

Temperate phages such as lambda can choose. After infection, lambda either enters the lytic cycle or enters the lysogenic cycle, in which its 48.5 kb genome integrates into the host chromosome at att by site specific recombination, catalyzed by phage integrase. The integrated DNA is a prophage, and the host cell, now a lysogen, divides normally and copies the prophage along with its own DNA. The lambda cI repressor keeps the lytic genes silent and also makes the lysogen immune to superinfection by another lambda. When the host DNA is damaged (for example by ultraviolet light), activated RecA stimulates self cleavage of the cI repressor, the prophage excises (integrase plus excisionase), and the lytic cycle begins: this is induction.

Plaques: how phage genetics is scored

Phages are counted by the plaque assay. Phages are mixed with a large excess of bacteria in soft agar and poured over a plate. Bacteria grow into a cloudy lawn, but wherever a single phage infected a cell, successive rounds of infection and lysis clear a small circular zone, a plaque. Each plaque comes from one phage, so plaques measure the titer:

titer (plaque forming units per mL) = number of plaques / (volume plated in mL × dilution)

Phage mutants change plaque appearance or host range. Rapid lysis (r) mutants of T2 and T4 make large, sharp edged plaques, while wild type (r⁺) makes small, fuzzy plaques. Host range (h) mutants can infect bacterial strains that resist the wild type. Phage crosses use mixed infection: each cell receives both genotypes, which recombine as they replicate.

Benzer and the fine structure of the rII region

In the 1950s, Seymour Benzer asked how finely recombination could divide a gene. He used the rII mutants of phage T4. The trick was a pair of hosts:

• On E. coli B, rII mutants grow well and form large r plaques.
• On E. coli K12(λ), a strain lysogenic for lambda, rII mutants cannot grow at all, while wild type T4 grows normally.

That asymmetry created an extraordinarily sensitive selection. After a cross between two rII mutants, plating on K12(λ) reveals only the rare wild type recombinants, while plating on B counts every progeny phage. Because one plate can hold progeny from a whole lysate, Benzer could detect one wild type recombinant among hundreds of millions of progeny, enough to measure recombination between mutations only one or a few base pairs apart.

Two different tests: complementation and recombination

Benzer used two tests that are easy to confuse, so keep them separate.

The complementation test asks whether two mutations are in the same functional unit. Cells of K12(λ) are coinfected with two rII mutants. If the two mutations are in different genes, each phage supplies the product the other lacks, the infection succeeds and the culture lyses: they complement. If both mutations are in the same gene, neither phage makes a working product and nothing happens. This is a trans test (the mutations are on different DNA molecules). Its control is the cis arrangement (both mutations on one molecule, the other molecule wild type), which always gives wild type function. Benzer called the unit defined by the cis-trans test a cistron, and found that the rII region contains two: rIIA and rIIB.

The recombination test asks whether two mutations are at different sites, and how far apart. Cells of E. coli B (on which all rII mutants grow) are coinfected, the progeny are collected, and the lysate is plated on both hosts. Wild type recombinants are one of two reciprocal products; the other (the double mutant) cannot be seen on K12(λ), so the frequency of wild type is doubled:

RF = 2 × (titer on K12(λ)) / (titer on B)

Mutations at the identical site never give wild type recombinants, and mutations in the same cistron can still recombine if they are at different sites.

Deletion mapping

Rather than crossing thousands of mutants pairwise, Benzer used deletion mutants, which are recognizable because they never revert. A point mutation crossed with a deletion can give wild type recombinants only if the point mutation lies outside the deleted segment; if the site is missing from the deletion phage, there is no wild type copy of that stretch to recombine in. A panel of overlapping deletions therefore divides the region into segments, and a single spot test against the panel assigns each new mutation to a segment. Fine mapping by recombination is then needed only among the few mutations within that segment.

What the rII work showed

Benzer mapped more than 2,000 rII mutations to about 300 sites. Recombination could separate mutations at neighboring sites, so the gene is not indivisible: the units of mutation and recombination are much smaller than the gene, ultimately single base pairs. He also found mutational hotspots. The map was linear, like DNA.

Section 5Viral genomes and retroviruses

What a virus is, genetically

A virus is a genome wrapped in a protein coat (the capsid), sometimes surrounded by a lipid envelope taken from the host membrane. Viruses have no ribosomes of their own, but their genomes vary enormously. Viral genomes can be DNA or RNA, single stranded or double stranded, linear or circular, in one piece or split into segments. They range from about 2 kb to more than a megabase.

The Baltimore classification

David Baltimore grouped viruses by the route their genome takes to messenger RNA (mRNA, defined as the plus strand), which predicts which enzymes each virus must carry.

Note that the naked genome of a plus strand RNA virus is infectious by itself, because host ribosomes can translate it directly, whereas minus strand and double stranded RNA viruses must carry their polymerase inside the virion. RNA dependent RNA polymerases and reverse transcriptases generally lack proofreading, so RNA viruses mutate far faster than DNA organisms, roughly 10⁻⁴ to 10⁻⁵ errors per nucleotide per replication compared with about 10⁻⁹ to 10⁻¹⁰ for cellular DNA replication. Coronaviruses (about 30 kb) are an exception: they encode a proofreading exonuclease.

Viral recombination and reassortment

Viruses exchange genes in two ways. When two related viruses infect the same cell, their polymerases can switch templates, producing recombinant genomes. Viruses with segmented genomes can also swap whole segments, a process called reassortment. Influenza A has eight minus strand RNA segments, including the ones encoding the surface proteins hemagglutinin (HA) and neuraminidase (NA). Point mutations in HA and NA cause antigenic drift (hence yearly vaccines); reassortment between, say, human and avian strains gives a new HA or NA, antigenic shift, the source of pandemics.

The retrovirus genome

Retroviruses reverse the usual flow of genetic information: their RNA genome is copied into DNA, which integrates permanently into the host chromosome. Each virion carries two copies of a plus strand RNA genome of about 7 to 10 kb (HIV-1 is about 9.7 kb), so retroviruses are effectively diploid. Every retrovirus has three core genes:

• gag (group specific antigen): structural proteins of the matrix, capsid and nucleocapsid.
• pol: the enzymes reverse transcriptase, integrase and protease.
• env: the envelope glycoproteins that bind receptors and fuse membranes (gp120 and gp41 in HIV).

HIV also carries regulatory and accessory genes such as tat, rev and nef.

The retroviral life cycle

1. Entry. Env binds the receptor (CD4 plus the coreceptor CCR5 or CXCR4 for HIV), the envelope fuses with the cell membrane, and the core enters the cytoplasm.
2. Reverse transcription. Reverse transcriptase, an RNA dependent and DNA dependent DNA polymerase with an RNase H domain that degrades the RNA in RNA DNA hybrids, makes a double stranded DNA copy of the genome. It is primed by a host transfer RNA (tRNA^Lys3 for HIV). Two template "jumps" generate identical long terminal repeats (LTRs) at the two ends of the DNA. Switching between the two RNA copies also causes recombination.
3. Integration. The viral DNA enters the nucleus, and integrase cuts the host DNA with a staggered cut and joins the viral ends to it. Repair of the gaps creates a short target site duplication (5 bp for HIV) flanking the provirus. The provirus is now a permanent part of that cell's genome and of all its descendants.
4. Expression. Host RNA polymerase II transcribes the provirus from a promoter in the 5′ LTR; the 3′ LTR supplies the polyadenylation signal. Spliced transcripts make Env; full length RNA makes Gag and Gag-Pol and serves as new genome.
5. Assembly and maturation. Two genome RNAs, Gag and Gag-Pol assemble at the membrane and bud off with an envelope. The viral protease then cleaves the polyproteins into mature proteins, converting the particle into an infectious virion.

Retroviruses, oncogenes and our own genome

Retroviruses gave cancer genetics its first oncogene. Rous sarcoma virus, which causes tumors in chickens, carries v-src, a gene that the virus captured long ago from a normal chicken gene, c-src, a proto-oncogene that encodes a tyrosine kinase. The viral copy is mutated and overexpressed, so it drives uncontrolled growth. Other retroviruses carry no oncogene but cause cancer by insertional activation, integrating beside a proto-oncogene such as MYC so that the strong LTR promoter or enhancer drives it.

When a retrovirus infects a germ line cell, the provirus can be inherited like any other gene. About 8 percent of the human genome consists of such endogenous retroviruses, almost all inactivated by mutation. Related LTR retrotransposons move within the genome by the same copy, reverse transcribe and integrate mechanism.

Words worth knowing