Back

A Far-Red FRET biosensor for AMPK enables multiplexed imaging of single-cell bioenergetic homeostasis

DeCuzzi, N. L.; Hardy, M.; Cabel, M. B.; Hu, J. Y.; Abbate, C. C.; Pargett, M.; Albeck, J. G.

2026-05-16 systems biology
10.64898/2026.05.13.723899 bioRxiv
Show abstract

Metabolic homeostasis has been studied primarily at the tissue and organism level, identifying molecular control mechanisms such as the energy charge-sensing kinase AMPK. Feedback loops involving AMPK and other regulators align cellular ATP generation and consumption, determining energetic balance. Recent work has demonstrated surprising oscillatory dynamics in AMPK activity, revealing unidentified kinetic modulation in single-cell homeostatic behaviour. However, probing the kinetic mechanisms of intracellular feedback requires simultaneous observation of multiple energetic parameters, and such experiments are precluded by the shared wavelength band occupied by most metabolic biosensors. We have overcome this obstacle by constructing a red-shifted FRET-based AMPK activity biosensor, RAMPKAR2, that is comparable to existing FRET-based AMPK activity biosensors. Multiplexed imaging of RAMPKAR2 with PercevalHR, which detects ATP/ADP ratio, confirmed that the kinetics of AMPK activity and ATP/ADP ratio are tightly coupled, with a lag of less than 6 minutes at the single-cell level. Pairing of RAMPKAR with HYlight, which detects the glycolytic intermediate fructose 1,6-bisphosphate (FBP), revealed that glycolytic activity co-oscillates with AMPK, shifted by [~]1.5 hours, and that these oscillations are suppressed by sustained AMPK activity. Together these data advance a model in which temporally offset increases in glycolytic ATP supply and AMPK deactivation contribute to single-cell oscillations.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.