The lac Operon
A single E. coli cell floating in a gut rich with lactose faces a simple economic problem. Should it manufacture the enzymes needed to digest a sugar that might not even be present? Building a protein costs energy and amino acids. Making the wrong enzyme at the wrong time is wasteful. Natural selection punishes this waste. The bacterium solves this problem with a molecular switch. It is one of the most thoroughly dissected control circuits in all of biology: the lac operon. This system matters far beyond one bacterium and one sugar. It was the first example ever worked out of how a gene can be turned on and off in response to a specific small molecule. It gave biology its basic vocabulary for Negative vs Positive Control of transcription — operator, repressor, inducer, operon. Before Jacob and Monod's work in the late 1950s and early 1960s, nobody knew that genes could have dedicated "off switches." These switches are bound by dedicated regulatory proteins. Their operon model was built from genetics rather than biochemistry. It correctly predicted the existence of a repressor protein years before anyone isolated one. The big picture is that the lac operon lets E. coli express three lactose-metabolizing genes as a single unit. This unit is controlled by one regulatory region. All three enzymes appear together only when lactose is around. They also appear only when glucose (the preferred sugar) is scarce. This article works out the negative-control half of that logic in full mechanistic detail. It covers how the repressor finds and binds the operator. It covers how the inducer allosterically releases it. It also covers how the whole circuit was proved experimentally rather than merely inferred. This article delivers the complete negative-control mechanism. It presents the genetic and biochemical evidence that established it. It offers worked numerical examples of operon genetics. Along the way, it flags where positive control and catabolite repression — the other half of the story — pick up.
- The lac operon is a cluster of three genes (lacZ, lacY, lacA) transcribed together from one promoter, controlled by an adjacent operator.
- The lacI gene makes a repressor protein that binds the operator and physically blocks RNA polymerase, keeping the operon off by default.
- Allolactose (made from lactose by trace basal β-galactosidase activity) binds the repressor, changes its shape, and makes it fall off the DNA — this is the inducer.
- Jacob and Monod's PaJaMo experiment and their constitutive/dominance analysis of lacI⁻ and lacI^s mutants proved the repressor is a diffusible, DNA-binding trans-acting protein, not a permanent DNA modification.
- Gilbert and Müller-Hill directly purified the lac repressor and showed by filter-binding and footprinting that it binds operator DNA in vitro and is released by IPTG, confirming the genetic model biochemically.
- This is only the negative-control layer; a second, positive-control layer (CAP–cAMP) integrates glucose availability, previewing lac Operon Positive Control and Catabolite Repression.
The lac operon is switched off by a repressor protein bound at the operator and switched on when an inducer (allolactose or its analog IPTG) binds the repressor and forces it to release the DNA.
The core idea
An operon is a cluster of genes transcribed as a single polycistronic mRNA from one promoter, so that the genes are always expressed together, in the same amounts, under the same conditions. The lac operon contains three structural genes in a row: lacZ (encoding β-galactosidase, which cleaves lactose into glucose and galactose, and also isomerizes some lactose into allolactose), lacY (encoding lactose permease, a membrane transporter that pumps lactose into the cell), and lacA (encoding a transacetylase whose physiological role is less central to the induction story). Upstream of these lies the promoter, where RNA polymerase binds, and overlapping it the operator (lacO), a specific DNA sequence recognized by a regulatory protein. That regulatory protein is the repressor, the product of the lacI gene, which sits just upstream of the promoter and is transcribed from its own weak, constitutive promoter — it is made continuously, independent of lactose. In the absence of lactose, repressor tetramers bind lacO tightly and block RNA polymerase from productively transcribing the structural genes: this is negative control, because the regulatory protein's default action is to repress. Note the terminology carefully, since it is often garbled: a repressor is negative-acting by definition, but negative control does not mean "always off" — it means the regulator's job is to prevent transcription until told otherwise. The molecule that lifts this block is the inducer: allolactose in vivo, or the synthetic non-metabolizable analog IPTG (isopropyl β-D-1-thiogalactopyranoside) used universally in the laboratory because it induces without being consumed. The elegant logic of the system is that it senses the presence of the substrate itself rather than requiring a separate sensing pathway. This is a general theme in inducible catabolic operons, in contrast to repressible biosynthetic operons like the trp operon (previewed in The trp Operon and Attenuation), where the end product of a pathway shuts the pathway down. The lac system is also not the whole story of glucose-versus-lactose logic: a second, independent positive-control mechanism using the CAP protein and cyclic AMP ensures that even with lactose present and the repressor off, transcription stays low if glucose — the preferred carbon source — is also available. That layer is treated separately in lac Operon Positive Control and Catabolite Repression; here we focus purely on the repressor/operator/inducer switch.
Mechanism
Step 1 — Basal repression in the absence of lactose. lacI is transcribed and translated constitutively at a low level (roughly 10 tetramers per cell). The repressor protein is a homotetramer; each monomer has an N-terminal DNA-binding domain (a helix-turn-helix motif) and a C-terminal domain responsible for tetramerization and inducer binding. The tetramer engages DNA as a dimer of dimers, allowing it to bind two operator-like sequences simultaneously and loop the intervening DNA — an unusually cooperative arrangement, discussed further below. The primary operator, lacO1, overlaps the transcription start site and part of the promoter; two auxiliary weaker operators, lacO2 (inside lacZ) and lacO3 (upstream of the promoter, inside lacI), are not individually essential but greatly increase repression efficiency by allowing DNA looping. When repressor sits on lacO1, it sterically occludes RNA polymerase from initiating productive transcription; the operon is transcribed only at a very low basal rate, enough to make perhaps 1–2 molecules of β-galactosidase per cell per generation.
Step 2 — Entry of lactose and generation of the true inducer. When lactose becomes available, a small amount enters the cell through whatever few permease molecules are present at basal levels (a bootstrap problem — solved because a handful of leaky permease molecules is enough to start the process). Inside the cell, basal-level β-galactosidase converts some lactose not to glucose+galactose but, via a transglycosylation side reaction, into allolactose (a 1,6-linked isomer of lactose). Allolactose, not lactose itself, is the physiological inducer.
Step 3 — Allosteric inactivation of the repressor. Allolactose binds a cleft in the C-terminal core domain of each repressor subunit, far from the DNA-binding domain. This binding triggers a conformational change transmitted through the protein: the DNA-binding domains rotate and their affinity for operator DNA drops roughly 1,000-fold. The repressor does not fall off DNA entirely in an all-or-none sense at the population level; rather, its equilibrium binding constant shifts so that, thermodynamically, it spends most of its time dissociated. This is a textbook case of allostery: a small molecule binding at one site changes affinity at a distant site, exactly as in hemoglobin's response to oxygen, but here repurposed for gene control.
Step 4 — Transcription proceeds. With the operator unoccupied, RNA polymerase holoenzyme (core enzyme plus σ⁷⁰ factor) binds the promoter's −10 and −35 elements and initiates transcription of a single polycistronic mRNA spanning lacZ-lacY-lacA. Because the mRNA is polycistronic, all three proteins are made from the same transcript, each with its own ribosome-binding site, ensuring coordinate expression — β-galactosidase, permease, and transacetylase levels all rise and fall together.
Step 5 — Positive feedback and shutoff. More permease pumps in more lactose; more β-galactosidase makes more allolactose; the system is self-reinforcing once triggered (this underlies the "all-or-none" induction behavior seen at the single-cell level in bistability experiments). When lactose is exhausted, allolactose production ceases, existing allolactose is metabolized away, the equilibrium shifts back toward repressor–operator binding, and the operon returns to its low, repressed baseline. The following two figures summarize the operon's physical layout and the allosteric switch itself.
The evidence
The lac operon model was not read off a DNA sequence — DNA sequencing did not exist yet — it was inferred entirely from bacterial genetics, and only later confirmed biochemically. This makes it one of the great examples in biology of a correct molecular mechanism being deduced from pure genetic logic.
Jacob and Monod's constitutive mutants and the cis/trans test (1959–1961). François Jacob and Jacques Monod, working at the Pasteur Institute with Arthur Pardee, isolated two classes of mutants that made β-galactosidase constitutively — that is, all the time, whether or not lactose was present. The first class mapped to a gene they called lacI; the second class mapped to a site immediately adjacent to the structural genes, which they called the operator (lacO^c^, for operator-constitutive). Crucially, these two classes of constitutive mutants behaved completely differently when tested in partial diploid (merodiploid) bacteria carrying an episome with a second copy of the lac region.
- lacI⁻ mutants (loss of repressor) were recessive to lacI⁺: a cell with one lacI⁻ chromosome and one lacI⁺ episome behaved normally, inducible, because the good copy supplied functional repressor that diffused through the cytoplasm and acted on both copies of the operator. This proved lacI encodes a diffusible, trans-acting product — a protein, not a fixed structural feature of one particular DNA molecule.
- lacO^c^ mutants (defective operator) were dominant only in cis</strong>: a lacO^c^ lacZ⁺ chromosome paired with an lacO⁺ lacZ⁻ episome gave constitutive expression of the lacZ gene physically linked to the mutant operator, but the other, wild-type-linked lacZ copy remained inducible. This proved the operator is a cis-acting DNA site, not a diffusible product — it can only affect genes physically adjacent to it on the same DNA molecule.
This cis/trans asymmetry — one class of mutation acting only in cis, the other acting equally well in trans — was the central logical proof that gene regulation involves a diffusible protein (the repressor) acting on a fixed DNA target (the operator). It is often called the PaJaMo experiment (Pardee, Jacob, Monod, 1959), which used bacterial conjugation (Hfr mating) to introduce a lacI⁺ gene into a lacI⁻ recipient and showed that β-galactosidase synthesis, initially ongoing, was rapidly shut off within minutes as new repressor protein was synthesized from the incoming gene — direct kinetic proof that repressor actively silences an already-transcribing operon, and that repression is due to an actively made protein rather than a pre-existing barrier.
Dominance of super-repressor mutants. Jacob and Monod also found lacI^s^ ("super-repressor") mutations that made the operon permanently un-inducible even with lactose present. These mapped to the lacI gene itself and were dominant to wild type in trans — a single copy of lacI^s^ repressed both operators in a merodiploid, even in the presence of inducer. This fit perfectly: lacI^s^ mutants make repressor protein that can still bind operator DNA but has lost the ability to respond allosterically to inducer, so it clamps down on all operator copies in the cell regardless of lactose. This was independent genetic confirmation that induction works through modifying a repressor protein's DNA-binding activity, not through destroying it.
Direct biochemical confirmation: Gilbert and Müller-Hill (1966). The operon model made a strong, falsifiable prediction: a specific protein must exist that binds specifically to lac operator DNA and is released by inducer. Walter Gilbert and Benno Müller-Hill isolated the actual lac repressor protein — a technical triumph, since it comprises only about 10 copies per cell — using an lacI^s^ strain overproducing the mutant protein and tritium-labeled IPTG to track it through purification via equilibrium dialysis. They showed: 1. The protein bound radiolabeled IPTG with high affinity (confirming it is the inducer receptor). 2. The purified protein bound specifically to lac DNA (from λplac phage) but not to unrelated DNA, retaining DNA–protein complexes on nitrocellulose filters (a technique now called the filter-binding assay). 3. Adding IPTG to the DNA–repressor complex caused the complex to dissociate, releasing free DNA — the biochemical reenactment of induction. Gilbert later refined this with DNase footprinting, showing the exact stretch of DNA (roughly 26 base pairs) physically protected by bound repressor, which matched precisely the operator location predicted from lacO^c^ mutant mapping. Sequence analysis confirmed operator DNA has a partial inverted-repeat (near-palindromic) symmetry, consistent with a symmetric tetrameric protein binding across both strands. Together, the genetics (Jacob–Monod–Pardee) and the biochemistry (Gilbert–Müller-Hill) form one continuous, mutually reinforcing body of evidence — genetics predicted the players, biochemistry proved they existed and behaved exactly as predicted.
Worked examples
Example 1 — Predicting phenotypes in a merodiploid genetics cross. Suppose you construct partial diploid E. coli strains and score β-galactosidase (lacZ product) activity with and without IPTG. Genotypes are written chromosome/episome. Fill in the phenotype using the rules above: lacI⁺ is recessive to lacI⁻ (i.e., wild-type repressor dominates, since a functional trans-acting repressor "wins"); lacO^c^ is cis-dominant only.
| Genotype | – IPTG | + IPTG | Reasoning |
|---|---|---|---|
| I⁺ O⁺ Z⁺ / I⁻ O⁺ Z⁻ | off | on | I⁺ supplies diffusible repressor in trans; both operators controlled normally; inducible |
| I⁻ O⁺ Z⁺ / I⁺ O⁺ Z⁻ | off | on | Same logic — trans-acting lacI⁺ rescues the lacI⁻ chromosome; Z⁺ copy is regulated normally |
| I⁺ O^c^ Z⁺ / I⁺ O⁺ Z⁻ | on | on | O^c^ is cis-only: it deregulates the Z⁺ gene linked to it on the same DNA molecule, regardless of repressor abundance |
| I⁺ O⁺ Z⁺ / I⁺ O^c^ Z⁻ | off | on | The functional Z⁺ gene is linked to the wild-type operator, so it stays properly regulated; the O^c^-linked copy is Z⁻ so it doesn't matter |
| I^s^ O⁺ Z⁺ / I⁺ O⁺ Z⁻ | off | off (locked) | I^s^ repressor cannot respond to inducer and is trans-dominant; it silences both operator copies even with IPTG present |
The key asymmetry to internalize: whether a Z⁺ gene turns on with IPTG depends only on (a) whether some functional, inducible repressor exists anywhere in the cell (an I question, answered in trans), and (b) whether that particular Z⁺ gene sits next to a functional operator on its own DNA molecule (an O question, answered only in cis). This table is exactly the logic Jacob and Monod used to deduce protein-versus-DNA-site identity of lacI and lacO.
Example 2 — Quantifying induction: fold-repression and enzyme kinetics. In a typical laboratory strain, basal (uninduced) β-galactosidase activity is around 1–2 units per milligram of cell protein, while fully induced activity (saturating IPTG, no glucose) reaches roughly 1,000–2,000 units per milligram — a ~500- to 1,000-fold induction ratio. Suppose an uninduced culture has a basal specific activity of 1.5 units/mg, and after adding 1 mM IPTG for one full generation (doubling time 30 minutes at 37 °C), activity rises to 1,200 units/mg. Fold-induction = 1200 / 1.5 = 800-fold. If β-galactosidase mRNA has a half-life of about 3 minutes (typical for bacterial mRNA) and translation begins immediately on ribosome loading, the delay between operator opening and detectable enzyme rise is dominated by transcription/translation kinetics, not mRNA stability — explaining why induction is detectable within 2–3 minutes of adding inducer in Pardee-style kinetic experiments, but reaches steady state only after several minutes of continued transcription, translation, and protein folding. This rapid kinetics — repression lifted, mRNA made, protein made, all within a few minutes — is exactly what allowed Pardee, Jacob, and Monod to time repressor action to the minute in the original PaJaMo mating experiment, since new repressor synthesized from a newly introduced lacI⁺ gene shut the operon back down within about 1–2 minutes of its own synthesis.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "Negative control means the gene is always off." | Negative control means a repressor's default action is to block transcription; the operon still turns on robustly when inducer removes the repressor. |
| "Lactose itself is the inducer." | The true inducer is allolactose, a lactose isomer made by trace basal β-galactosidase activity; IPTG is a synthetic inducer analog that is not metabolized. |
| "The repressor destroys or modifies the operator DNA." | The repressor binds reversibly and non-covalently; DNA sequence is unchanged. Induction is a conformational, allosteric event in the protein, not a change in DNA. |
| "lacO^c^ mutants act in trans like lacI mutants." | lacO^c^ is cis-acting only — it is a DNA sequence, not a diffusible molecule, and only affects genes physically linked to it. |
| "More repressor bound means 'more off'; the switch is graded like a dimmer." | At the single-cell level, induction of the wild-type operon is closer to bistable/switch-like (all-or-none) due to positive feedback via permease, though population averages look graded. |
| "All three lac genes are separately regulated." | They share one promoter/operator and are transcribed as a single polycistronic mRNA — regulation is coordinate, not gene-by-gene. |
The scientist behind it
Origins · 1 minFrançois Jacob and Jacques Monod and how this was discovered
Jacob, trained as a surgeon before turning to bacterial genetics after being wounded in World War II, and Monod, a biochemist and resistance fighter turned enzymologist, formed one of the most productive partnerships in the history of molecular biology, working alongside Arthur Pardee and later André Lwoff. Their genius was methodological: instead of trying to purify an unknown, vanishingly rare regulatory protein directly, they used the logic of bacterial genetics — conjugation, merodiploids, and mutant complementation — to infer its existence, its properties, and its mode of action with near-total precision, years before anyone could isolate it in a test tube. Monod, already famous for describing enzyme induction and diauxic growth (bacteria consuming glucose before lactose — the phenomenon underlying catabolite repression), brought deep intuition about allostery, later formalized with Jean-Pierre Changeux and Jeffries Wyman in the MWC model of allosteric proteins, a framework the lac repressor itself perfectly exemplifies.
Where this matters today
The lac operon is not merely historical; its parts are now standard biotechnology hardware. The lac promoter/operator system is one of the most widely used inducible expression switches in recombinant DNA technology: genes of interest (insulin, vaccine antigens, industrial enzymes) are commonly cloned downstream of a lac-derived promoter in expression vectors, and protein production is triggered on demand simply by adding IPTG to a culture — exactly the synthetic inducer discovered in these original experiments, chosen because bacteria cannot metabolize it away. The blue/white screening method for identifying recombinant plasmids, ubiquitous in undergraduate cloning labs, exploits lacZ (α-complementation with X-gal) directly. Beyond biotechnology, the operon remains the pedagogical prototype for understanding regulatory logic in every genome: the general concepts of cis-regulatory elements, trans-acting factors, allosteric ligand-controlled DNA-binding proteins, and inducible-versus-repressible circuits — the same vocabulary used to describe steroid hormone receptors, bacterial two-component systems, and eukaryotic transcription factors — all trace back to the operator/repressor logic first defined here. Synthetic biologists building genetic circuits (toggle switches, oscillators like the repressilator) still use lac repressor–operator modules as one of their core interchangeable parts, precisely because its biochemistry is so completely characterized.
Ask the tutor: How does CAP–cAMP positive control combine with repressor-based negative control to produce the diauxic growth curve on glucose plus lactose?
Why is the lac repressor described as acting by "negative control"?
Negative control means the regulatory protein's role is to actively block transcription, and gene expression proceeds only once that block is removed. The lac repressor binds the operator and physically prevents RNA polymerase from productively transcribing the operon; when an inducer removes the repressor, transcription resumes. This is the opposite logic to positive control, where an activator protein must bind before transcription can occur at all (as CAP does at the lac promoter). The terms describe the direction of the regulator's default effect, not whether the gene ends up on or off in any particular condition.
Why does the cell need trace amounts of permease and β-galactosidase even before induction?
Induction requires allolactose, which is made from lactose by β-galactosidase, and lactose must first get inside the cell via permease. If the operon were absolutely, perfectly off with zero protein made, lactose could never enter and could never be converted into the very molecule needed to trigger induction — a chicken-and-egg problem. The solution is "leaky" basal expression: even a fully repressed operon is transcribed at a low but non-zero rate, producing just enough permease and β-galactosidase (a handful of molecules) to import a little lactose and generate a little allolactose, which then triggers full induction.
What is the difference between lacI⁻ and lacO^c^ mutants, and why does the distinction matter?
Both give constitutive (inducer-independent) expression, but for entirely different molecular reasons. lacI⁻ destroys the repressor protein itself, so no functional repressor exists — this defect is recessive and acts in trans, because a good copy of lacI elsewhere in the cell supplies diffusible repressor that fixes the problem. lacO^c^ changes the operator DNA sequence so repressor can no longer bind it, but the repressor protein itself is fine — this defect is dominant only in cis, since it only affects genes physically linked to that particular altered operator. This distinction was the historical proof that regulation involves a diffusible protein acting on a fixed DNA site.
Is lac operon induction graded or switch-like?
At the level of a bulk liquid culture, induction looks graded and dose-dependent on IPTG concentration, since different cells respond to different degrees. But at the single-cell level, using intermediate, sub-saturating IPTG concentrations, individual cells tend to be either fully OFF or fully ON, rarely at an intermediate state — a phenomenon called bistability. This arises because permease import and allolactose-driven induction reinforce each other in a positive feedback loop: a cell that induces even slightly imports more lactose, which drives further induction. The population-level "graded" curve is really an average over individually switch-like cells.
How does IPTG differ from lactose or allolactose, and why is it preferred experimentally?
IPTG (isopropyl β-D-1-thiogalactopyranoside) binds the lac repressor and induces the operon just like allolactose does, but its thioglycosidic bond is not cleaved by β-galactosidase, so it is never metabolized or depleted by the cell. This means researchers can add a fixed, precisely known concentration of IPTG and expect stable, reproducible, long-lasting induction, unlike lactose or allolactose, whose concentrations change continuously as they are consumed. This property makes IPTG the standard inducer in recombinant protein expression systems and classroom cloning experiments.
Does the lac repressor bind DNA as a single site interaction, or is it more complex?
It is more complex: the repressor tetramer engages DNA at the primary operator lacO1 plus auxiliary operators lacO2 and lacO3, binding two operators simultaneously as a "dimer of dimers" and looping the DNA between them. This looped, multivalent binding increases repression efficiency far beyond what a single operator site could achieve, though lacO1 alone is sufficient for substantial repression. This detail, discovered after the original Jacob-Monod model, refines but does not overturn their core logic of a diffusible repressor acting at a cis DNA operator.
How does the lac operon relate to positive control and catabolite repression?
Everything described here is negative control: whether the repressor sits on the operator or not. But E. coli also prefers glucose over lactose, and even with lactose present and the repressor released, transcription stays weak unless a second, independent activator protein called CAP (catabolite activator protein), bound to cyclic AMP, also binds near the promoter and helps recruit RNA polymerase. Glucose lowers cAMP levels, keeping CAP inactive and the operon weak even when lactose is around — this is catabolite repression, and it operates as a layer of positive control stacked on top of the negative control described in this article, explored fully in lac Operon Positive Control and Catabolite Repression.