Gene regulation in bacteria

A bacterium carries genes for thousands of proteins but needs only some of them at any moment. Making enzymes to digest a sugar that is not present, or to build an amino acid already abundant in the medium, wastes energy and materials that could be spent on growth. Bacteria therefore regulate their genes, mostly at the level of transcription, switching them on when needed and off when not. The study of how they do this, beginning with the lac operon of E. coli, created the vocabulary used for gene regulation in all organisms: repressors and activators, operators and inducers, negative and positive control. This lesson works through the lac and trp operons in detail, models their behaviour, and surveys other bacterial strategies, from riboswitches to quorum sensing and CRISPR.

Jacob, Monod and the lac operon

When E. coli grows on glucose, it makes very little of the enzyme β-galactosidase, which splits the sugar lactose into glucose and galactose. If lactose is the only sugar available, the enzyme’s level rises about 1,000-fold within minutes. In the 1950s and early 1960s, François Jacob and Jacques Monod at the Pasteur Institute in Paris used genetics to explain this induction. Their operon model, published in 1961, earned them a share of the 1965 Nobel Prize with André Lwoff.

The lac operon contains three structural genes transcribed together as one mRNA (lesson 5): lacZ, encoding β-galactosidase; lacY, encoding lactose permease, which brings lactose into the cell; and lacA, encoding a transacetylase whose role is minor. Upstream lie the promoter and the operator, a DNA sequence that overlaps the promoter. Nearby, a separate gene, lacI, is transcribed from its own promoter and encodes the lac repressor.

In the absence of lactose, the repressor, a tetramer, binds tightly to the operator and blocks RNA polymerase from transcribing the genes. When lactose enters the cell, some is converted by β-galactosidase into allolactose, an isomer that acts as the inducer. Allolactose binds the repressor and changes its shape so that it releases the operator, allowing transcription. Because the repressor turns the genes off and the inducer relieves that repression, the lac operon is said to be under negative control. In the laboratory, the synthetic inducer IPTG is used instead of lactose because β-galactosidase cannot break it down, so its concentration stays constant.

Positive control by CAP

Lactose is not the whole story. If both glucose and lactose are present, E. coli uses glucose first and the lac genes stay largely off, a phenomenon called catabolite repression. Glucose is the preferred fuel. The mechanism involves a second regulatory protein, the catabolite activator protein (CAP, also called CRP). When glucose is scarce, levels of the small molecule cyclic AMP (cAMP) rise. cAMP binds CAP, and the CAP–cAMP complex binds a site just upstream of the lac promoter, bends the DNA and contacts RNA polymerase, recruiting it to the promoter. Recall from lesson 5 that the lac promoter matches the consensus poorly: without CAP, it is weak. When glucose is plentiful, cAMP is low, CAP does not bind and transcription remains low even if lactose is present. CAP is therefore a positive regulator, an activator.

The two controls together implement simple logic: transcribe strongly only if lactose is present AND glucose is absent. The code below encodes this logic and adds the detail that the repressor does not shut the operon off completely; a low basal level of transcription allows a few molecules of permease and β-galactosidase to be present, which is essential, because without them lactose could never enter the cell to produce the inducer in the first place.

def lac_output(lactose, glucose):
    repressed = not lactose            # inducer (allolactose) releases the repressor
    cap_bound = not glucose            # low glucose -> high cAMP -> CAP-cAMP binds
    if repressed:
        return 1                        # basal leak (relative units)
    return 1000 if cap_bound else 50   # induced: strong with CAP, weak without

print("glucose  lactose  relative lac expression")
for glucose in (True, False):
    for lactose in (False, True):
        print(f"{str(glucose):8s} {str(lactose):8s} {lac_output(lactose, glucose):6d}")

glucose  lactose  relative lac expression
True     False         1
True     True         50
False    False         1
False    True       1000

Genetic analysis of the operon

Jacob and Monod worked out this scheme largely from mutants. Mutations in lacI that destroy the repressor (lacI⁻) cause constitutive expression: the genes are on all the time, with or without inducer. Mutations in the operator that prevent repressor binding (lacOᶜ, for operator constitutive) have the same effect. To tell them apart, Jacob and Monod built partial diploids, or merodiploids, in which a second copy of part of the lac region is carried on a plasmid-like F′ factor. A normal lacI⁺ gene on the F′ restored regulation to a cell with a lacI⁻ chromosome, because the repressor protein diffuses and can act on any operator in the cell: lacI acts in trans. But a normal operator on the F′ did not restore regulation of genes on a chromosome with lacOᶜ, because an operator controls only the genes physically next to it on the same DNA molecule: lacO acts in cis. This cis–trans test remains a basic tool for distinguishing DNA sites from diffusible products.

All-or-none induction

Something interesting happens at low inducer concentrations. Because the permease encoded by lacY imports more inducer, the operon is part of a positive feedback loop: more expression brings in more inducer, which causes more expression. With weak inducers at intermediate concentrations, individual cells tend to be either fully on or fully off, rather than all partly on. Aaron Novick and Milton Weiner described this all-or-none behaviour in 1957, and it is now understood as bistability, a property of positive feedback loops that is important in development and cell-fate decisions. The model below, a pair of simple differential equations solved step by step, shows how a cell’s outcome depends on its starting state when the external inducer is at an intermediate level.

import numpy as np

def simulate(y0, inducer_out, hours=10, dt=0.01):
    """y = permease level (relative). Internal inducer depends on permease;
    expression rises steeply (Hill n=2) with internal inducer; dilution by growth."""
    y = y0
    for _ in range(int(hours / dt)):
        inside = inducer_out * (0.05 + y)                 # leak plus active import
        expression = 0.02 + inside**2 / (1 + inside**2)   # basal + induced
        y += dt * (expression - 0.7 * y)                  # synthesis - dilution
    return y

for ext in (0.5, 1.5, 4.0):
    low, high = simulate(0.02, ext), simulate(1.2, ext)
    print(f"external inducer {ext:3.1f}: start OFF -> {low:4.2f}   start ON -> {high:4.2f}")

external inducer 0.5: start OFF -> 0.03   start ON -> 0.03
external inducer 1.5: start OFF -> 0.07   start ON -> 1.10
external inducer 4.0: start OFF -> 1.41   start ON -> 1.42

At low inducer, cells end low whatever their start; at high inducer, cells end high. At the intermediate level, the outcome depends on the starting state: cells that were already induced stay on, and uninduced cells stay off. This memory, called hysteresis, explains why a population grown at intermediate inducer can contain two kinds of cells.

The trp operon: repression and attenuation

The tryptophan (trp) operon contains five genes for enzymes that make the amino acid tryptophan. Here the logic is reversed: the genes should be on when tryptophan is scarce and off when it is plentiful. The trp repressor is made in an inactive form. When tryptophan is abundant, it binds the repressor as a corepressor, and the repressor–tryptophan complex binds the operator and blocks transcription. This is repression of a biosynthetic operon, still negative control, but triggered by the end product rather than a substrate.

Charles Yanofsky at Stanford discovered a second, more subtle layer of trp control: attenuation. The trp mRNA begins with a leader sequence of about 160 nucleotides, containing a short coding region for a 14-amino-acid leader peptide with two tryptophan codons in a row, and four segments (1 to 4) that can pair in alternative ways. Segments 3 and 4 can form a terminator hairpin followed by a run of U (lesson 5). Segments 2 and 3 can instead form an antiterminator hairpin, which prevents the terminator from forming. Which structure forms depends on where the ribosome translating the leader peptide sits. If tryptophan is plentiful, charged tryptophan tRNA is abundant, the ribosome moves quickly through the tryptophan codons and covers segment 2, so segments 3 and 4 pair and transcription terminates. If tryptophan is scarce, the ribosome stalls at the tryptophan codons on segment 1, leaving segment 2 free to pair with 3, forming the antiterminator, and transcription continues into the genes.

Attenuation depends on the coupling of transcription and translation in bacteria and so could not work the same way in eukaryotes, where the processes are separated by the nuclear envelope. Repression alone reduces trp operon expression about 70-fold, and attenuation adds about another 8- to 10-fold, giving roughly 600-fold total control.

Other bacterial strategies

Riboswitches are segments of mRNA, usually in the 5′ untranslated region, that bind a small molecule directly and change shape, switching transcription or translation on or off without any protein. A riboswitch that binds thiamine pyrophosphate, for example, controls genes for thiamine synthesis in many bacteria, and similar switches control genes for other vitamins and amino acids. Because they act without proteins, riboswitches may be relics of an RNA world (lesson 2), and some are being explored as antibiotic targets.

Small regulatory RNAs (sRNAs), typically 50 to 300 nucleotides, pair with target mRNAs to block or promote translation or trigger degradation, often with help from the chaperone protein Hfq. Two-component systems let bacteria sense their environment: a membrane sensor kinase detects a signal, phosphorylates itself and transfers the phosphate to a response regulator protein that changes gene expression. E. coli has about 30 such systems, sensing things from osmotic pressure to phosphate levels.

Quorum sensing allows bacteria to coordinate behaviour with population density. Each cell secretes a small signal molecule, an autoinducer; when enough cells are present, the signal reaches a threshold and switches on genes, often for light production, biofilm formation or virulence. The marine bacterium Aliivibrio fischeri produces light only at high density inside the light organ of the Hawaiian bobtail squid, where it helps camouflage the squid against moonlight. The opportunistic pathogen Pseudomonas aeruginosa, a major cause of lung infections in people with cystic fibrosis, uses quorum sensing to control toxin production and biofilms, and blocking quorum sensing is being explored as a way to disarm pathogens without killing them, which may slow resistance.

CRISPR: an adaptive immune system

Many bacteria and most archaea carry CRISPR loci: clustered regularly interspaced short palindromic repeats, in which short repeated sequences alternate with spacers derived from the DNA of viruses and plasmids that previously attacked the cell or its ancestors. The locus is transcribed and processed into short CRISPR RNAs, each carrying one spacer. Bound to Cas (CRISPR-associated) proteins, the RNA guides the complex to matching foreign DNA or RNA, which is then cut. New spacers are added when the cell survives an infection, so CRISPR gives bacteria a heritable memory of past infections, a form of adaptive immunity. Studies of yogurt-making Streptococcus thermophilus by the food company Danisco showed in 2007 that bacteria gaining spacers from a phage became resistant to it. In 2012, Jennifer Doudna and Emmanuelle Charpentier showed that one CRISPR protein, Cas9, could be programmed with a designed guide RNA to cut any chosen DNA sequence, the basis of genome editing (lesson 12). They received the 2020 Nobel Prize in Chemistry.