Does pyruvate kinase link glucose metabolism to circadian output via TARANIS and PDF?

Testing glycolysis enzymes one at a time

The study asks whether reduced glucose metabolism inside clock cells changes circadian timing, and if so through which molecule. The authors used RNA interference to lower three rate-limiting glycolytic enzymes — Hexokinase-C (Hex-C), Phosphofructokinase (Pfk) and pyruvate kinase (Pyk) — specifically in Drosophila clock cells, using a timeless (tim) promoter-driven GAL4 driver. Multiple RNAi lines were tested per gene to reduce off-target effects.

The comparison that matters is between enzymes: only Hex-C and Pyk knockdown changed period length, lengthening and shortening it respectively, and quantitative RT-PCR confirmed the matching mRNA reductions. Pfk knockdown also produced period changes, but its mRNA was not reduced, which the authors treat as a possible off-target effect rather than a genuine Pfk result. That internal check is why the paper's argument rests on Pyk, not on glycolysis as a whole.

Why adult-only knockdown matters

Continuously lowering a glycolytic enzyme from development onward could disturb how the clock neuronal network forms, which would make any adult rhythm change hard to interpret. To separate this, the authors combined tim-GAL4 with a temperature-sensitive tubulin-GAL80ts repressor: flies raised at 20 °C suppress the RNAi, and shifting to 30 °C induces knockdown in adults.

Under this adult-specific knockdown (AKD), Hex-C no longer changed period, while Pyk still shortened it, matching the continuous-knockdown result. Overexpressing Pyk in the knockdown background restored both period and Pyk mRNA, supporting specificity of the 35218 RNAi line used afterward. Adult-specific Pyk overexpression alone did not significantly alter period or rhythmicity. Metabolomics of fly heads at ZT2 and ZT14 confirmed the intended biochemical effect: elevated phosphoenolpyruvate, reduced pyruvate, and reduced NADH, while ATP was unaffected.

A short period that resembles PDF loss

Pyk AKD flies showed reduced PER intensity in clock neurons including LNvs, DN1s and LNds at ZT2 and ZT14, and reduced PDF staining in small ventral lateral neuron (sLNv) somata and dorsal projections. Behaviorally they had weakened rhythmicity with a shorter free-running period, plus an advanced evening activity onset under light-dark cycles — a profile the authors note closely resembles Pdf knockdown flies.

This is where the paper's interpretation becomes explicit rather than measured. Prior work cited by the authors associates reduced PER with lengthened periods and reduced PDF with shortened periods. Because the Pyk phenotype looks like PDF loss, the authors propose that the PDF effect dominates over the PER effect, producing net period shortening. That ordering is a model built from the comparison, not a direct measurement of competing forces.

PDF levels track Pyk activity reversibly

A key control addresses whether Pyk knockdown simply damages the clock neurons. Flies were kept at 30 °C for 5 days of knockdown, then moved to 20 °C for 5 days to restore Pyk expression. PDF levels, significantly reduced during knockdown, returned to control levels after the shift, and PER recovered similarly. The sLNvs therefore retained functional integrity, and PDF responds dynamically to Pyk activity rather than reflecting permanent cell loss.

The reduction in PDF protein was much larger than the change in Pdf mRNA, which was modest and not statistically significant. The authors read this mismatch as evidence that PYK regulates PDF mainly after transcription. This distinction matters: a protein-level change of that size cannot be explained by the small transcript change alone.

TARA is required between PYK and PDF

The authors first tested VRILLE (VRI), a clock component known to support PDF expression, but VRI levels at ZT14 were similar between control and Pyk AKD flies, so VRI abundance did not track the PDF loss. They then turned to TARANIS (TARA), a transcriptional coregulator previously reported to enhance VRI activity and whose mutants show reduced PDF.

Pyk AKD lowered tara mRNA and TARA protein, while Pyk overexpression left them unchanged. Adult-specific tara overexpression in the Pyk knockdown background rescued the shortened period and reduced rhythmicity, and substantially restored PDF levels. tara overexpression alone did not raise PDF, which the authors interpret as TARA and PYK acting in the same pathway rather than additively. Critically, PER levels were not rescued by TARA, and Pdf mRNA stayed low even when PDF protein was restored — pointing to post-transcriptional control of PDF in this hypometabolic context.

A temperature-sensitive tara effect

One result complicates a simple reading of TARA as a period lengthener. Overexpressing tara in LNvs lengthened period significantly at 25 °C but not at 30 °C, even though TARA::GFP fusion protein levels were similar at both temperatures. Because the rescue experiments were run at 30 °C, the temperature sensitivity is not explained by differing protein abundance, and the authors report it as an unresolved feature of TARA biology rather than a settled mechanism.

The paper also does not find direct contact between the two proteins: co-immunoprecipitation assays failed to detect PYK–TARA interaction in vitro or in flies. The proposed link is therefore indirect, with PYK influencing tara transcription — supported by the drop in tara mRNA — rather than by binding TARA.

Two separable branches from PYK to rhythm

Putting the pieces together, the authors propose a PYK–TARA–PDF axis: Pyk knockdown lowers tara transcription, which reduces VRI activity, which lowers PDF and shortens the period. Separately, Pyk knockdown lowers NADH, and the authors suggest this redox change reduces PER through a TARA-independent route, citing mammalian work in which NADH and sirtuin1 modulate CLOCK:BMAL1 and PER stability. The NADH–PER link is presented as a hypothesis consistent with the data, not as a demonstrated mechanism in this study.

The two branches explain why TARA rescue restores period and PDF but leaves PER low. It also means the shortened period should not be read as a single downstream consequence of low PER; in this system the PDF branch appears to set the direction of the period change.

What the fly result does and does not cover

All of the evidence comes from Drosophila clock cells, with behavioral periods measured in individual flies and protein levels quantified by immunostaining of specific clock neuron groups. The authors connect the result to aging and neurodegeneration, where brain hypometabolism and disrupted rhythms co-occur, and note that aged Drosophila show reduced PDF. But they explicitly state that future studies in aging and neurodegeneration models are needed to test that connection; the paper does not measure those conditions.

The practical priority is therefore the separation of outputs rather than a general claim that glycolysis sets the clock. When glycolytic flux falls in clock cells, the TARA-dependent drop in PDF is the branch that shifts period direction, while PER falls through a route TARA cannot rescue. Any attempt to predict circadian consequences of metabolic impairment should track these two outputs separately, because a treatment that restores PDF signaling would not be expected to restore PER.

Sources used in this answer