Understanding how neural networks generate complex behaviour is one of the great challenges of neuroscience. NeurotransmiLers and neurom odulators dynamics modulate neuronal circuits, and understanding their func-on is ke y to unravelling their role in behaviour. Many diseases are associated with dysfunc-onal neuromodula tory signaling. Especially the neuromodulator serotonin (5-HT) plays an important role in a variety of physiological func-ons, cogni-on and behavior. 5-HT seems to be invo lved in regula-ng emo-onal and social behaviors and disturbances in brains serotonin levels are associated with psychiatric disorders, such as major depressive disorder (MDD). Although the func-on of 5-HT has been studied for many years, the lack of appropriate tools to measure dynamics on diûerent -mescale with high spa-al resolu-on prevented the detail ed elucida-on of underlying disease mechanisms. A special feature of neuromodulators, such as serotonin (5-HT), is that they act not only as classical neurotransmiLers at synapses on a mill isecond -mescale, but also through volume transmission on the scale of seconds to minutes. These longer -mescale dynamics of serotonin are thought to underlie internal brain states or behavioral states, such as sleep, foraging or escape 1-3.

In recent years, an increasing number of gene-cally encod ed ûuorescent biosensors have been published that allow the measurement of neurotransmiLer rel ease with high temporal and spa-al resolu-on in vitro and in vivo 4-7. All of this intensity or intensiometric biosensors have in common that upon binding of the ligand to the sensing moiety, conforma-onal changes occur that can either quench or enhance ûuorescence. A change in ûuorescence can be mediated by several factors, such as shi`ing the protona-on equilibrium of the chromophore, changing the quantum yield or the ex-nc-on coeûcient 8.

However, intensity-based sensors all suûer from several limita-ons, such as their dependence on expression levels, excita-on power, photobleachin g and sensi-vity to pH changes 9, which will limit the possible interpreta-ons that can be deriv ed from the intensity readout. A more robust readout may be ûuorescence life-me informa-on ( FLIM), which is not aûected by expression level, excita-on power, photobleaching or pH 10,11. Only tools that will enable the precise measurement of serotonin dynamics independent from expression levels of the sensor and excita-on power will make it possible to compare se rotonin levels across animals, brain areas, disease states and -mepoints.

Previously it has been shown that some, but not all GPCR based sensors for neuromodulators, exhibit changes in their life-me upon binding of their respec-ve ligand and can be u-lized to compare neuromodulator dynamics across animals, brain regions, disease models and over -me 12. Although Ma et al. reported that also GRAB-5-HT showed a change in life-me upon satura-ng serotonin concentra-ons, detected changes were quite small (<0.1 ns), hindering the ability of GRAB5-HT to be u-lized as FLIM based sensor.

We hypothesized that sDarken, a GPCR-based intensiometric biosensor that decreases its ûuorescence upon binding of serotonin, would also show changes in ûuorescence life-me upon applica-on of 5-HT ( Fig. 1a). To test our hypothesis, we applied increasing concentra-ons of serotonin to human embryonic kidney (HEK) cells transiently expressing sDarken. We observed that the ûuorescence life-me is decreased w ith increasing concentra-ons of 5HT (Fig. 1b-c). Life-me was signiûcantly diûerent from baseline star-n g at a concentra-on of 80 nM (p< 0.0021, One Way ANOVA). It is noteworthy that changes in FLIM are only observed within the dynamic range of sDarken (Kd = 127 nM; Kubitschke et al. 2023) . The dynamic range of sDarken, as characterized in Kubitschke et al., is between 100 nM and 1 µM, which corresponds well to a decrease of FLIM within this concentra-on range. We observed that the dynamic range (0.529 ns) of the life-me between baseline (t=2.689± 0.0022 ns )and the maximum concentra-on of 2.560 µM ( t=2.160±0.055 ns) is rela-vely large. The doseresponse curve displays a sigmoidal shape (Fig. 1d) and an IC50 of 221.4 nM. To verify our hypothesis, that changes in FLIM will only be observable with the dynamic range of the u-lized biosensor we repeated our experiments with L-sDarken a low aûnity variant with 1.000 fold lower aûnity for serotonin (KD =45 µM, Kubitschke et al. 2022) . In deed changes of the life-me could only be observed between 10 µM and 1 mM (Fig.1 e-g), well in accordance with the reported dynamic range of L-sDarken. Fluorescence life-me is decreased with increasing concentra-ons of 5-HT ( Fig. 1e-f). Life-me was signiûcantly diûerent from baseline star-ng at a concentra-on of 50 µM (p< 0.021, One Way ANOVA). As for sDarken, also L-sDarken had a remarkable large dynamic range of 0.462 ns, between the baseline life-me (mean t=2.680 ±0.0022 ns) and a maximum concentra-on of 1 mM ( mean t=2.218±0.0047 ns). For control we used a nullmutant of sDarken and a cytosolic expressed GFP. The nullmutant of L-sDarken combines several muta-ons in the binding pocket of the 5-HT1A receptor the sensing domain of sDarken: A50V, D82N, D116N, S198A and T199A, that have been shown to be important for the binding and ac-va-on of the 5-HT1A receptor 13,14. Both controls did not show any signiûcant changes in their life-me to diûerent concentra-ons of serotonin (Fig.2 a-b).

The decrease in life-me is blocked by the applica-on o f the 5-HT1A antagonist WAY (Fig.2 c-d). Life-me is signiûcantly decreased with the applica-on of 5-HT (-5-HT: mean t=2.615±0.0011 ns; + 5-HT: mean t=2.502±0.0020 ns; one-way ANOVA p=0.003), whereas in the presence of WAY no change in life-me could be observed (- 5-HT: me an t=2.615±0.0011 ns; WAY+5-HT: mean t=2.596 ±0.00125 ns; one-way ANOVA p=0.7832).

In conclusion, sDarken, a gene-cally encoded serotonin sensor, exhibits a concentra-ondependent change in ûuorescence life-me. To date, sDar ken has the largest reported life-me change for GPCR based neuromodulator sensors. For sDarken life-me changes are observed at submicromolar 5-HT concentra-ons and across micromolar concentra-ons for L-sDarken, the low aûnity variant of sDarken. We suggest that in the futu re, FLIM measurements of the sDarken family could be a valuable complement to intensity-based measurements of serotonin. As been shown for GRABACh3.0 15, sDarken could provide addi-onal informa-on about tonic levels of serotonin during diûerent internal states 12. FLIM measurements of sDarken have the poten-al to quan-ta-vely assess seroton in levels, also over long periods of -me making it an ideal tool to inves-gate diûerences in brain serotonin levels in health and disease. Another advantage would be that FLIM as a readout could presumably be used to make beLer comparisons of serotonin levels across brain regions, animals, and disease states12.