Sigma Factors and Promoter Recognition

Bacterial RNA polymerase exists in two functionally distinct forms. The core enzyme (α₂ββ'ω in E. coli) contains the full catalytic machinery for RNA synthesis — it can elongate an RNA chain on a template and even bind DNA nonspecifically, but it cannot recognize a promoter sequence or initiate transcription accurately. Add one more subunit, sigma factor, and the core becomes the holoenzyme (α₂ββ'ωσ), which now binds specifically to promoter DNA at the −35 element (consensus TTGACA) and the −10 element (consensus TATAAT), melts the DNA locally to form the open complex, and begins RNA synthesis. Sigma is the sequence-reading module; core is the chemistry module. The crucial biological trick is that a cell keeps a repertoire of different sigma factors, each with a distinct DNA-binding specificity, and swaps them in and out of a shared pool of core enzyme depending on conditions. Because promoter recognition is entirely delegated to sigma, changing which sigma is bound to core is equivalent to changing which promoters — and hence which genes — are transcribed. This is regulon-level control: dozens to hundreds of genes scattered around the chromosome, all preceded by the same sigma-specific promoter consensus, are switched on or off together as one sigma factor is exchanged for another. It is a fundamentally different regulatory strategy from operon-specific repressors and activators (like the lac repressor), which act gene by gene; sigma exchange acts genome-wide in a single step. Sigma factors fall into two structurally unrelated families. The σ⁷⁰ family (named for the E. coli housekeeping sigma, 70 kDa) includes most alternative sigmas and shares four conserved regions (σ1.1–σ4) that fold around core and contact the −10 and −35 elements and the discriminator/extended −10 sequence. The σ⁵⁴ family (σ^N, ~54 kDa) is structurally and mechanistically distinct: σ⁵⁴-holoenzyme forms a stable closed complex at its promoters (consensus centered near −24/−12) that cannot spontaneously isomerize to the open complex — it requires ATP-hydrolyzing activator proteins bound at distant enhancer sites, contacted through DNA looping, an arrangement that structurally anticipates eukaryotic enhancer action.

Step 1 — Core enzyme alone is promiscuous and inaccurate. Core RNA polymerase binds DNA with modest affinity nonspecifically, at almost any sequence, and can even initiate RNA synthesis at incorrect sites or on single-stranded nicks in vitro. It has no intrinsic mechanism for recognizing the −10/−35 spacing and sequence that define a real bacterial promoter.

Step 2 — Sigma binding creates the holoenzyme and unmasks the true DNA-binding surface. Sigma docks onto core primarily via the β' subunit (an interaction described as a "clamp" that also allosterically repositions core's own nonspecific DNA-binding determinants). Sigma's own conserved regions become properly positioned only in the context of core: region 4.2 is presented to contact the −35 hexamer, region 2.3–2.4 to contact the −10 hexamer and to help melt it, and (for σ⁷⁰) region 3.0 to contact an "extended −10" TG motif in some promoters lacking a strong −35.

Step 3 — Closed complex formation. Holoenzyme locates a promoter and forms the closed complex (RPc): double-stranded promoter DNA is bound but not yet melted. This step is driven by direct base-specific contacts of sigma regions 4.2 and 2.4 to the −35 and −10 hexamers respectively, plus σ1.1 and the polymerase's αCTDs contacting an UP element upstream when present.

Step 4 — Isomerization to the open complex. The DNA strands separate over roughly 13 base pairs around the −10 element and the start site, forming the open complex (RPo), often called the "transcription bubble." Two conserved aromatic residues in sigma region 2.3 flip out and stack against flipped-out bases of the nontemplate strand at the −10 element, nucleating and stabilizing the melted state — a direct sigma contribution to strand separation, not merely to recognition.

Step 5 — Abortive initiation and promoter escape. RNA polymerase begins synthesizing short RNAs (2–9 nucleotides) that are frequently released — abortive initiation — because sigma region 3.2 occupies the RNA exit channel and clashes with the growing transcript. Once the nascent RNA reaches roughly 9–12 nucleotides, it displaces σ3.2, the enzyme escapes the promoter, and sigma is typically released from core (the "sigma cycle"), freeing it to associate with a fresh core molecule and reinitiate elsewhere. Some sigma can remain transiently associated during early elongation, but stable elongation complexes are generally sigma-free.

Step 6 — Sigma competition sets the transcriptional program. Because core enzyme is limiting relative to the sum of all sigma factors, different sigmas compete for a shared core pool. Whichever sigma factor is most abundant, most activated, or bound with highest affinity at a given moment captures core and directs it to its own promoter class. Cells regulate which sigma "wins" by controlling sigma synthesis, sigma degradation, sigma sequestration by anti-sigma factors, and release of sigma from anti-sigma factors by anti-anti-sigma factors — a layered switch architecture seen repeatedly (e.g., σ³² is normally short-lived and translationally repressed; heat shock stabilizes it and relieves repression, so it surges and redirects core to heat-shock promoters).

The existence of sigma was discovered, not postulated, through careful biochemical fractionation. In 1969, Richard Burgess, Andrew Travers, Ekkehard Bautz, and Michael Chamberlin purified E. coli RNA polymerase and found that during purification the enzyme's activity changed character: highly purified "core" polymerase could still incorporate nucleotides into RNA using nicked or randomly-sheared DNA templates, but it initiated transcription at many nonspecific sites and could not distinguish a promoter. A minor polypeptide, roughly 70 kDa, separated from the enzyme during phosphocellulose chromatography. When this polypeptide (christened sigma) was added back to the core enzyme, the reconstituted holoenzyme now initiated RNA synthesis preferentially and accurately at the correct start sites on double-stranded, intact phage DNA templates (they used bacteriophage T4 DNA), producing RNA that matched the pattern expected from correct promoter usage rather than random initiation. This single reconstitution experiment established the central logic of the field: specificity is a separable, addable property, contributed by a dissociable subunit, not an intrinsic property of the catalytic core. Subsequent work sharpened the picture in three ways.

Filter-binding and footprinting located sigma's job precisely. Using DNase I footprinting (protection assays) on defined promoters, researchers (building on Mark Ptashne's and David Pribnow's contemporaneous promoter-sequencing work) showed that holoenzyme, but not core alone, protects a stretch of DNA spanning roughly −40 to +20, encompassing the −35 and −10 hexamers — direct physical evidence that sigma-containing holoenzyme is what makes the sequence-specific contacts, and that core alone gives a much less defined, promiscuous footprint.

In vitro transcription with purified components showed selectivity is quantitative and sigma-dependent. Comparing initiation rates of core alone versus holoenzyme on templates carrying strong versus mutant promoters showed that holoenzyme discriminates by orders of magnitude in favor of consensus-like −10/−35 sequences, while core shows little such discrimination — confirming that promoter strength, not merely promoter choice, is dictated through sigma's contacts.

The discovery of alternative sigmas showed that swapping sigma reprograms the whole cell. Studies of Bacillus subtilis sporulation in the 1970s–80s (notably by Richard Losick and colleagues) found that as vegetative cells commit to sporulation, the predominant sigma factor associated with core changes in a defined temporal sequence (σ^A → σ^H, then compartment-specific σ^F, σ^E, σ^G, σ^K in the forespore and mother cell). Purifying RNA polymerase from sporulating cells at different time points and testing it on cloned sporulation-gene promoters in vitro showed directly that holoenzyme containing the new sigma, and only that holoenzyme, transcribed the sporulation genes efficiently — while the original vegetative holoenzyme could not. This was decisive proof that sigma exchange, not new core enzyme, drives the developmental switch in gene expression, and it generalized the E. coli discovery into a broad principle: differentiation programs in bacteria can be executed by a cascade of sigma factors, each sigma both transcribing the genes needed for the current stage and, among them, the gene encoding the next sigma in the cascade. Later structural biology (X-ray crystal structures of Thermus aquaticus holoenzyme bound to promoter DNA, from Seth Darst's and Richard Ebright's laboratories in the early 2000s) visualized these contacts atomically, confirming the aromatic-residue base-flipping mechanism at the −10 element and the physical clamp of σ4.2 around the −35 element predicted from decades of biochemistry and genetics.

Example 1 — Predicting the outcome of a sigma swap experiment. Suppose you purify E. coli core enzyme and separately purify σ⁷⁰ and σ³² (heat-shock sigma). You set up two in vitro transcription reactions, each containing core plus one sigma, and offer both reactions a mixed DNA template pool containing: (a) the rpoH-independent housekeeping promoter of the lacZ operon (a classic σ⁷⁰ promoter, strong match to TTGACA/TATAAT), and (b) the promoter of groEL (a classic σ³²-dependent heat-shock promoter, with sequence matching the σ³² consensus, distinct from σ⁷⁰'s). Reasoning: sigma alone determines which promoter sequence is recognized; core is shared and catalytically identical in both tubes. So:

If you now mix both sigmas with a limiting amount of core (mimicking the real cell, where core is limiting relative to total sigma), the fraction of core enzyme captured by each sigma — and hence the relative output from the two promoters — depends on the relative concentration and core-binding affinity of σ⁷⁰ versus σ³² in the mix. This is exactly the competitive logic that operates in vivo during heat shock: σ³² is normally scarce and rapidly degraded by the protease FtsH, keeping its regulon off; a temperature upshift stabilizes σ³² (partly because chaperones DnaK/DnaJ that normally sequester and help degrade σ³² are titrated away onto newly accumulated misfolded proteins), so σ³² concentration rises roughly 15-fold within minutes, out-competing σ⁷⁰ for core at heat-shock promoters and inducing chaperone and protease genes without any new core enzyme being made.

Example 2 — Reasoning through an anti-sigma factor mutant phenotype. The alternative sigma factor σ²⁸ (FliA) in E. coli and Salmonella directs transcription of late flagellar genes (the genes for flagellin and motor components). σ²⁸ is held inactive by binding to an anti-sigma factor, FlgM, until the earlier-stage flagellar basal body/hook structure is complete; FlgM is then secreted out of the cell through the completed export channel, releasing σ²⁸ to activate late flagellar genes. This creates a checkpoint: late genes are transcribed only after the structure that will use their products already exists. Consider a mutant strain lacking flgM entirely. Predict the phenotype, then check the logic:

• Without FlgM, σ²⁸ is never sequestered.
• σ²⁸ therefore competes for core and activates late flagellar promoters immediately, regardless of whether the hook–basal body is finished.
• Predicted result: premature, checkpoint-bypassing expression of flagellin and other late genes, even in cells whose basal body isn't yet built — producing nonfunctional or prematurely assembled flagellar components and, experimentally, elevated σ²⁸-dependent transcription measured by reporter fusions (e.g., a fliC-*lacZ fusion shows constitutively high β-galactosidase activity in a flgM* mutant compared to wild type, where activity rises only after hook completion).

This matches the actual experimental finding (Hughes, Gillen, Semon, and Karlinsey, 1993, and subsequent work): flgM null mutants show constitutive, checkpoint-independent activation of class 3 (late) flagellar promoters, directly demonstrating that an anti-sigma factor's job is to make sigma activity conditional on an upstream developmental event — the same architectural logic used again in sporulation (anti-σ factors SpoIIAB, RsbW) and in phage development.

Sigma factor was identified through the unglamorous but decisive work of protein purification — separating RNA polymerase into its component parts and asking which piece did what. Richard R. Burgess, then a graduate student with Michael Chamberlin at Berkeley, ran E. coli RNA polymerase through phosphocellulose columns and noticed that the flow-through fractions behaved differently from the bound fractions in transcription assays: one had catalytic activity but poor template selectivity, the other conferred specificity when added back. Working in close contact with Andrew Travers and Ekkehard Bautz — who were independently characterizing the same phenomenon in polymerase preparations used to study phage T4 transcription — the two groups converged on the same conclusion and co-published, cementing sigma as a genuine, separable polypeptide rather than an artifact of purification.

Antibiotic development and persistence. Alternative sigma factors, especially σ³⁸ (RpoS, stationary phase/general stress) and σ³² (heat shock), are implicated in bacterial persistence — a dormant, drug-tolerant state that contributes to chronic and relapsing infections. Because RpoS-controlled genes help cells survive stress including antibiotic exposure, sigma factor activity is an active target area for adjuvant therapies meant to resensitize persister cells.

Virulence gene regulation. Many pathogens use dedicated sigma factors to switch on virulence regulons only inside a host. Mycobacterium tuberculosis encodes over a dozen sigma factors; several (σ^E, σ^H, σ^F) are induced by host-relevant stresses (oxidative stress, heat, envelope stress) and are required for full virulence in animal models, making them of interest as attenuation targets for live vaccine strains.

Sporulation and biotechnology. The Bacillus subtilis sporulation sigma cascade (σ^H → σ^F/σ^E → σ^G/σ^K) is a textbook model system, but it is also directly exploited industrially: Bacillus spores are used as probiotics and as robust chassis for enzyme and vaccine production, and engineering sigma factor timing controls when and how efficiently spores form.

Phage biology and synthetic biology. Bacteriophages often encode their own sigma factors or anti-sigma factors (e.g., phage T4's gene 55 product acts as a phage-specific sigma-like factor; phage SPO1 sequentially replaces host sigma) to commandeer host RNA polymerase for the phage's own transcriptional program — a strategy synthetic biologists now borrow, engineering orthogonal sigma/promoter pairs to build layered, insulated genetic circuits in bacteria that don't cross-talk with the host's native regulons.

Understanding <a class="wiki-link" href="/wiki/bio-molec-rna-polymerase-and-the-transcription-cycle" data-track="wiki_article.body.wiki_link">RNA Polymerase and the Transcription Cycle</a> and general transcription control. Sigma factor logic — a shared catalytic core plus exchangeable specificity factors — is the conceptual bridge to eukaryotic transcription, where general transcription factors (TFIID, TFIIB, etc.) play a partly analogous role in reading core promoter elements, and to combinatorial control more broadly, making sigma a essential stepping stone toward Transcription Initiation and Promoter Escape and comparative studies of gene regulation across domains of life.

Ask the tutor: How does sigma factor competition for limiting core enzyme quantitatively determine the relative transcription rates of two regulons in a cell experiencing simultaneous stresses?