The discovery of promoter-proximal pausing
Early models of transcriptional regulation assumed that gene expression was controlled primarily at the step of RNA Polymerase II (Pol II) recruitment to gene promoters. If TFIID, TFIIB, and the Mediator complex assembled at a promoter and recruited Pol II, transcription proceeded; if they did not, the gene was silent. This model was adequate for constitutively expressed housekeeping genes but could not explain the kinetics of inducible gene activation — in particular, why some genes respond to stimuli within two minutes while others require an hour or more.
The resolution came from genome-wide Pol II ChIP-seq and GRO-seq (global run-on sequencing) studies in Drosophila and mammals. GRO-seq, which captures nascent RNA from actively elongating Pol II, revealed that a large fraction of transcriptionally silent or low-level genes nevertheless had Pol II positioned 30-60 base pairs downstream of the transcription start site — engaged on the DNA, having initiated transcription, but stopped. This paused Pol II is transcriptionally competent but held in place by a set of negative elongation factors. The seminal characterisations came from work by Lis and colleagues (Drosophila heat shock genes, the hsp70 locus as a model), and from genome-wide analyses by Core, Waterfall, and Lis (2008, Science) and Zeitlinger et al. (2007, Nature Genetics).
The subsequent mechanistic dissection showed that promoter-proximal pausing is an active and regulated step — not a passive failure of elongation — and that its release is a key rate-limiting step in regulated transcription.
The pausing machinery: DSIF and NELF
Two protein complexes impose Pol II pausing after promoter clearance:
DSIF (DRB sensitivity-inducing factor) is a heterodimer of Spt4 and Spt5 (yeast nomenclature; SUPT4H and SUPT5H in humans). Spt5 is a long, multi-domain protein that contacts Pol II through its NusG-like NGN domain (which clamps around the Pol II cleft) and its C-terminal repeat domain (CTR), which recruits capping enzymes, splicing factors, and other elongation regulators. DSIF initially stabilises pausing in concert with NELF, but upon P-TEFb phosphorylation of Spt5 CTR, DSIF converts from a pause-inducing factor to an elongation- stimulating factor — it remains associated with Pol II throughout elongation and promotes processivity.
NELF (negative elongation factor) is a four-subunit complex: NELF-A (WHSC2), NELF-B (COBRA1), NELF-C/D (NELF-C and NELF-D are splice isoforms from the same gene, NELFCD), and NELF-E (RD/RDBP). NELF-E contains an RRM (RNA recognition motif) that contacts the nascent RNA emerging from Pol II at approximately 25 nucleotides of length — this RNA contact is required for stable pausing. NELF-A and NELF-B contact the Pol II surface directly. Structural studies (Vos et al., 2018, Science; Vos et al., 2018, Molecular Cell) have resolved the architecture of the paused Pol II elongation complex at atomic resolution, showing how DSIF and NELF collaborate to clamp Pol II in a conformation that prevents further nucleotide addition.
The paused state is further stabilised by the DSIF-mediated recruitment of the cap-binding complex, which reads the co-transcriptionally added 7-methylguanosine cap on the nascent RNA. The CBC (cap-binding complex, consisting of CBP80 and CBP20) binds the 5-prime cap and, through interactions with NELF-E, reinforces the paused architecture. This is why DRB (5,6-dichlorobenzimidazole riboside), a CDK9 inhibitor, was originally found to affect primarily capped mRNAs — it prevents DSIF and NELF phosphorylation and locks Pol II in the paused state.
P-TEFb: CDK9/Cyclin T1 and the pause-release signal
Positive transcription elongation factor b (P-TEFb) is the kinase that releases Pol II from pausing. It consists of the catalytic subunit CDK9 (cyclin-dependent kinase 9) in complex with Cyclin T1, T2, or K. CDK9 phosphorylates three substrates in the paused complex to trigger elongation: (1) Spt5 CTR (DSIF), converting it from a pause factor to an elongation stimulator; (2) NELF-A and NELF-E, releasing NELF from Pol II; and (3) Pol II CTD Ser2, establishing the elongation-associated CTD phosphorylation code that recruits splicing factors, capping enzymes, and other co-transcriptional processing factors.
The phosphorylation of Ser2 by CDK9 is perhaps the most consequential: Ser2-phosphorylated CTD (pSer2-CTD) recruits the spliceosomal U2 snRNP and later the 3-prime processing machinery (including CstF and CPSF cleavage factors), coupling elongation to co-transcriptional splicing and polyadenylation. Loss of CDK9 activity therefore simultaneously disrupts elongation, splicing, and 3-prime processing.
P-TEFb itself is regulated at two levels. First, the 7SK snRNP sequesters P-TEFb in an inactive complex: the non-coding RNA 7SK (a 331-nucleotide structured RNA) acts as a scaffold for HEXIM1 (or HEXIM2), which binds CDK9 and inhibits its kinase activity. The LARP7 protein stabilises 7SK RNA; MEPCE (methyl phosphate capping enzyme) adds a gamma-methyl cap to 7SK's 5-prime end. When cells are activated by growth signals, inflammatory stimuli, or stress, HEXIM1 dissociates from CDK9, and free, active P-TEFb is released to elongate paused Pol II. Second, BRD4 recruits free P-TEFb to chromatin through its PID domain, concentrating CDK9 activity at super-enhancers and highly active gene promoters — this BRD4-P-TEFb axis is the mechanism through which SE-driven transcription is coupled to elongation (see the companion post on super-enhancers).
Immediate-early genes and the paused Pol II reservoir
The evolutionary logic of promoter-proximal pausing becomes clear when considering immediate- early (IE) genes — genes that must be induced within minutes of a stimulus without requiring new protein synthesis. The canonical examples are c-FOS (the AP-1 transcription factor subunit), EGR1 (early growth response 1), and JUN-B. These genes have paused Pol II constitutively positioned at their promoters; the machinery for transcription is pre-loaded and requires only P-TEFb-mediated release to produce a burst of mRNA. This is analogous to having a cocked but uncocked firearm — the activation energy for rapid gene induction is dramatically reduced.
GRO-seq analyses in unstimulated cells show that c-FOS has one of the highest ratios of promoter-proximal to gene-body Pol II signal in the human genome — a quantitative measure called the pausing index (PI). Upon serum stimulation, the PI collapses as Pol II floods into the gene body, producing a wave of c-FOS mRNA within 2-3 minutes. The speed of this response would be impossible without the pre-loaded paused Pol II reservoir.
NF-kappaB-driven inflammatory genes and the Pol II pausing model
The paused Pol II paradigm applies with particular force to NF-kappaB target genes. A 2013 study by Hargreaves et al. in Cell demonstrated that LPS stimulation of macrophages activates a two-wave transcriptional response: primary response genes (PRGs, activated within 30 minutes) are pre-loaded with paused Pol II and are induced without requiring new transcription factor synthesis or extensive chromatin remodelling; secondary response genes (SRGs, activated 1-4 hours later) require de novo synthesis of transcription factors (often the PRG products themselves) and more extensive chromatin opening.
The PRGs include canonical NF-kappaB targets: TNF-alpha, IL-1B, CXCL1, CXCL2, and several others. Their rapid induction depends on NF-kappaB (p65/p50 heterodimer) binding to kappaB sites in their promoters/enhancers, recruiting the Mediator-BRD4-P-TEFb axis to release the pre-positioned paused Pol II. Critically, chromatin immunoprecipitation experiments have shown that these genes have paused Pol II in unstimulated macrophages, and that LPS causes rapid CDK9 recruitment and Ser2-CTD phosphorylation — not new Pol II recruitment — as the primary activation event.
IL-6 is a somewhat different case — it does require NF-kappaB-mediated enhancer activation and chromatin remodelling before full induction, but P-TEFb recruitment is still a critical rate-limiting step. The IL-6 promoter contains a paused Pol II in many cell types, demonstrating that the pausing mechanism operates across the full spectrum of inflammatory gene regulation.
Where spirulina connects: attenuating the pause-release signal
Spirulina's anti-inflammatory effects converge on the Pol II pausing machinery at two mechanistically distinct points:
First, NF-kappaB suppression by phycocyanin/PCB via IKK-beta inhibition reduces the transcriptional activation signal that drives P-TEFb recruitment to paused Pol II at inflammatory gene promoters. If NF-kappaB nuclear translocation is reduced, p65 occupancy at kappaB sites is lower, BRD4-p65 co-activation is reduced, and P-TEFb is not efficiently delivered to the paused Pol II complex at TNF-alpha, IL-1B, and IL-6 promoters. The net effect is that paused Pol II remains paused — the burst of inflammatory mRNA production is attenuated, not abolished. This is an important distinction: constitutive genes without NF-kappaB dependence would continue to be elongated normally, because their P-TEFb recruitment relies on BRD4-SE interactions that are not NF-kappaB-dependent.
Second, AMPK activation has been shown to suppress CDK9 activity at specific targets. AMPK phosphorylates HEXIM1 at Ser279 (reported in Srisook et al., 2012 context and related work), which promotes HEXIM1 incorporation into 7SK snRNP and sequesters P-TEFb. Under conditions of high AMPK activity (low energy charge), more P-TEFb is in the 7SK-HEXIM1 inactive pool, reducing the free P-TEFb available for pause release at inflammatory genes. This represents a metabolic gating of Pol II elongation — cells under energy stress (AMPK active) limit inflammatory gene induction by sequestering P-TEFb, a logical physiological response given the metabolic cost of sustained inflammatory gene transcription. Spirulina's AMPK activation would be predicted to partially engage this 7SK-mediated P-TEFb sequestration.
It is important to be clear about the evidence hierarchy here. Spirulina's NF-kappaB suppression and AMPK activation are well-documented in the primary literature. The mechanistic connections to paused Pol II — that reduced NF-kappaB reduces P-TEFb recruitment to paused Pol II, and that AMPK activation promotes 7SK-mediated P-TEFb sequestration — are well-established biochemistry. The specific claim that spirulina supplementation alters Pol II pausing indices or P-TEFb recruitment kinetics at inflammatory gene loci in human subjects has not been demonstrated by any published study.
Mechanistic resolution: why this matters more than simply blocking transcription
The Pol II pausing framework explains a pharmacological distinction that has practical importance. A treatment that simply blocked NF-kappaB nuclear translocation completely would abolish all NF-kappaB-driven transcription — inflammatory genes, but also pro-survival genes (BCL2, c-FLIP, cIAP1/2), immune cell identity genes, and wound-healing genes. This explains why complete NF-kappaB inhibitors cause immunosuppression and are clinically limited.
A treatment that partially reduces the transcriptional activation signal — enough to attenuate paused Pol II release at highly NF-kappaB-dependent inflammatory genes, without eliminating P-TEFb recruitment at constitutively-expressed housekeeping genes — would produce preferential anti-inflammatory effects without global immunosuppression. The pausing model predicts that such partial inhibition is mechanistically feasible because of the concentration-dependence of P-TEFb recruitment: highly activated inflammatory genes (which require the strongest NF-kappaB signal to achieve maximal P-TEFb recruitment and release) would be most sensitive to partial NF-kappaB attenuation, while constitutive genes (which require only basal CDK9 activity for their lower pause index) would be largely unaffected.
This provides a mechanistic rationale for spirulina's anti-inflammatory effects without causing immunosuppression — an observation consistent with the clinical literature, where spirulina supplementation reduces inflammatory markers (CRP, TNF-alpha, IL-6) without increasing infection susceptibility.
Summary
RNA Polymerase II pausing at 30-60 bp downstream of transcription start sites, stabilised by the DSIF (Spt4/Spt5) and NELF (NELF-A/B/C/E) complexes, is the rate-limiting step for rapid inflammatory gene induction. P-TEFb (CDK9/Cyclin T1), recruited by BRD4 and regulated by 7SK/HEXIM1 sequestration, releases pausing by phosphorylating Spt5-CTR, NELF-A/E, and Pol II CTD Ser2. NF-kappaB-driven inflammatory genes (TNF-alpha, IL-1B, IL-6) use pre-loaded paused Pol II for rapid induction that requires P-TEFb release rather than de novo Pol II recruitment. Spirulina's NF-kappaB suppression attenuates the transcriptional activation signal driving P-TEFb recruitment to paused Pol II at inflammatory promoters; AMPK activation by spirulina may promote 7SK-mediated P-TEFb sequestration, further reducing elongation at stimulus-responsive inflammatory genes. These mechanisms, grounded in established Pol II biology, provide a mechanistically satisfying account of spirulina's anti-inflammatory effects at the level of transcriptional elongation control.