In their original report of ACh-evoked DA release, Cachope and colleagues (2012 ) presented an example recording whereby strong optogenetic activation of cholinergic interneurons for several seconds is accompanied by DA elevation in the NAc of a urethane-anesthetized mouse. More recently, we and others showed in mice and rats that the patterns of DA and ACh release in various striatal locations in vivo are strongly correlated on sub-second time-scales in a manner that is consistent with ACh-evoked DA release (Howe et al., 2019; Liu et al ., 2022; Chantranupong et al., 2023; Krok et al., 2023; Mohebi et al., 2023) . Yet DA dynamics in the dorsal and lateral striatum were found to persist even after molecular, pharmacological and optogenetic interference with ACh signaling (Chantranupong et al., 2023; Krok et al., 2023) . A recent paper in eLife (Mohebi et al., 2023) provided some of the most compelling evidence to date that optogenetic activation of channelrhodopsin-expressing striatal cholinergic interneurons drives DA release in the NAc of awake behaving rats, as measured with the redshifted DA sensor RdLight1 (Patriarchi et al., 2020). However, one concern with these experiments is that mApple-based fluorescent sensors 3 including RdLight1 and the GRAB-rDA3 series (Zhuo et al., 2023) 3 may exhibit photoactivation (also known as 8photoswitching9 or 8photoconversion9), a process whereby mApple9s red fluorescence changes in the presence of blue light (Shaner et al., 2008). This phenomenon is one of the main downsides of the R-GECO family of red Ca2+ indicators, which also use mApple and grow brighter independently of Ca2+ for hundreds of milliseconds following brief flashes of blue light, limiting their use with optogenetics (Akerboom et al., 2013; Dana et al., 2016) . In the case of RdLight1, it was previously shown that photoactivation effects are negligible when expressed in cultured kidney cells and imaged using light-scanning confocal microscopy (Patriarchi et al., 2020; but see Zhuo et al., 2023). Whether RdLight1 shows photoactivation in vivo under conditions routinely used in behavioral experiments (i.e. fiber photometry and full-field optogenetic stimulation) has not been investigated.

To determine if blue light modiûes the üuorescent properjes of RdLight1 in the behaving brain, we virally expressed RdLight1 in either the dorsolateral striatum (DLS; N=8) or NAc (N=8) of wild-type mi ce (Figure 1A,B) and imaged RdLight1 üuorescence in vivo using ûber photometry while mice were headûxed on a cylindrical treadmill (Figure 1C). Under standard conjnuous illuminajon condijons ( 565 nm excitajon light, 30-50 µW at the jp of the ûber), we observed transient inc reases and decreases in red üuorescence consistent with established pamerns ofDA release (Howe and Dombeck 2016; Da Silva et al., 2018; Wei et al., 2022; Chantranupong et al., 2023; Krok et al., 2023; Markowitz et al., 2023) , including spontaneous üuctuajons during immobility, movement-related transients during self-paced wheel running and large-amplitude reward-evoked responses (not shown).

Delivering blue light pulses through the same ûber at powers typically used for optogenejc manipulajons (6 ms-long; 9 mW at the jp of the patc h cord) evoked disjnct transients in RdLight1 üuorescence resembling DA release (Figure 1D). In the NAc, these transients averaged 9.3 ± 0.2 % DF/F in magnitude, peaked 112 ± 2 ms aqer light onset an d decayed back to baseline within 1 s (tdecay: 438 ± 24 ms; N=8 mice). In the DLS, blue light-evoked tra nsients were smaller (5.6 ± 0.3 % DF/F) but showed similar kinejcs (jme from light onset to peak: 135 ± 5 ms; tdecay: 467 ± 43 ms; N=8 mice). In both regions, blue light-evoked transients scaled in magnitude with the durajon ( Figure 1E) and intensity of light pulses (Figure 1F). In a separate laboratory, we observed comparable increases in RdLight1 üuorescence in N=3 mice that expressed RdLight1 in the ventrolateral striatum (VLS; Figure 1 3 Figure Supplement 1), conûrming their occurrence across a range of experimental condijons. These delayed üuorescent signals are speciûc to RdLight1, as they are not observed in mice expressing the red üuorescent protein tdTomato (not shown).

Do these blue light-evoked RdLight1 transients reüect a phasic elevajon in extracellular DA? Several lines of evidence suggest that this is not the case. First, our wild-type mice do not express blue lightgated opsins to drive DA release and DA neurons are not thought to be intrinsically sensijve to blue light. Second, blue light-evoked transients show limle trial-by-trial variability in their amplitudesand kinejcs ( Figure 1D,G). Third, blue light-evoked transients do not display short-term facilitajon or depression under a variety of sjmulajon condijons ( Figure 2 and Figure 1 3 Figure Supplement 1C), calling into quesjon their synapjc origin. Fourth, we repeated the above experiments in a subset of mice treated with reserpine, an irreversible antagonist of the transporter required for loading DA into synapjc vesicles ( Figure 1G). Under these condijons, spontaneous üuctuajons in RdLight1 üuorescence vanished in both the DLS (N=4) and NAc (N=4), conûrming the absence of acjvity-dependent DA release in vivo. By contrast, blue light-evoked RdLight1 transients did not disappear and, if anything, grew in amplitude and durajon in both DLS and NAc ( Figure 1G,H), demonstrajng that they do not reüect synapjc release of DA.

Collectively, our results show that RdLight1 displays strong photoactivation following exposure to blue light under conditions routinely used to monitor and manipulate neural activity in vivo. This photoactivation manifests as a prolonged, DA-independent increase in RdLight1 fluorescence that outlasts the blue light pulse and slowly decays back to baseline over hundreds of milliseconds, giving it the appearance of synaptically-released DA. Under our recording conditions, photoactivation remained detectable with as little as 0.3 mW blue light (Figure 1F), indicating that RdLight1 fluorescence should be interpreted with caution when combined with blue light in a variety of experimental conditions, including dual-color imaging with alternating stimulation of green and red fluorophores. In addition, our findings call into question the nature of the RdLight1 fluorescent transients reported in Figure 1 of the study by Mohebi et al. (2023) . Given the similarity of our recordings in terms of response magnitude, timing and dynamics over a variety of stimulation parameters, it is likely that the light-evoked RdLight1 responses reported reflect this photoactivation effect. Although the study used 405 nm illumination to control for changes in fluorophore properties independent of ligand binding (i.e., the so-called isosbestic point), this deep blue wavelength has not been shown to be the isosbestic point for red-shifted fluorophores such as RdLight1. Additional experiments will therefore be needed to determine the conditions under which cholinergic interneurons locally evoke DA release in the striatum of behaving animals. If using RdLight1, experiments should include a mutated sensor that does not bind DA, controls expressing RdLight1 only (i.e., no opsin, to control for the effects of blue light alone) and minimal constant blue light stimulation. Alternatively, green fluorescent protein (GFP)-based DA sensors such as dLight1 (Patriarchi et al., 2018) or GRAB-DA3 (Zhuo et al., 2023) may be used in combination with redshifted optogenetic actuators (Klapoetke et al., 2014; Marshel et al., 2019) , although this configurat ion is not without caveats either, as we have observed that blue light illumination at photometric intensities can cause persistent opsin activation (not shown).

B.L.S. and N.X.T conceived of the project. J.T. and R.M. performed experiments in the NAc and DLS. L.C. and M.J.W. carried out experiments in the VLS. J.T., R.M., L.C., B.L.S and N.X.T. analyzed and interpreted the data, and wrote the manuscript in collaboration with A.M and J.D.B.

The other authors declare no competing interests.

Correspondence should be addressed to N.X.T. (nicolas.tritsch@nyulangone.org)

Animals. Procedures were performed in accordance with protocols approved by the NYU Grossman School of Medicine (NYUGSM) and Harvard Medical School (HMS) Insjtujonal Animal Care and Use Commimees. Wild type mice (C57Bl6/J; Jackson Laboratory strain #000664; 12318 weeks of age) were housed in group before surgery and singly aqer surgery under a reve rse 12-hour light-dark cycle (dark from 6 a.m. to 6 p.m. at NYUGSM and 10 a.m. to 10 p.m. at HMS) with ad libitum access to food and water. Stereotaxic surgery. Mice were prepared for intracranial infecjons of adeno-associated viruses (AAVs) as before (Chantranupong et al., 2023; Krok et al., 2023) . Brieüy, mice were anaesthejzed with isoüurane, administered Ketoprofen (10 mg/kg, subcutaneous) or Carprofen (5 mg/kg, subcutaneous) and placed in a stereotaxic apparatus (Kopf Instruments), where a small craniotomy was drilled above the NAc (from Bregma: AP +1.0 mm, ML +0.75 mm), DLS (from Bregma: AP +0.5 mm, ML +2.5 mm), or VLS (from Bregma: AP +0.6 mm, ML +2.3 mm). 300 nL of AAV2/9.Syn.RdLig ht1 (CNP Viral Vector Core at the CERVO Research Center contribujon) was injected (100 nL/min) at a depth of 3.7 mm below dura for NAc, 2.3 mm for DLS or 3.2 mm for VLS using a microsyringe pump (KD Scienjûc; Legato 111) connected to a pulled glass injecjon needle (100 µm jp; Drummond Wiretrol II). Fiber opjcs (NAc and D LS: 400 ¿m diameter core, 0.5 NA; RWD Life Science Inc; VLS: 200 ¿m diameter core, 0.48 NA; Doric) were implanted 100 µm above the injecjon site and cemented to the skull using C &B metabond (Parkell) along with a custom jtanium headpost placed over lambda to allow for head ûxajo n during recordings. Mice were allowed to recover in their home cage for 234 weeks before head-ûxajon and treadmill habituajon, and recordings. Fiber Photometry. RdLight1 photometry recordings were carried out by feeding constant, low-power yellow-green excitajon light (565 nm LED, 30350 µW at the jp of the patch cord; Thorlabs M565F3) to a üuorescence mini cube (FMC5_E1(460-490)_F1(500-540) _E2(555-570)_F2(580-680)_S; Doric) connected to the mouse9s ûber opjc implant via a 0.48 NA patc h cord (NAc and DLS: MFP_400/460/9000.48_2m_FCM-MF1.25; VLS: MFP_200/220/900_2m_FCM-MF1 .25; both from Doric). The red light emimed by RdLight1 was collected through the same patch cord and routed via the üuorescence mini cube and a second ûber opjc (MFP_600/630/LWMJ-0.48_0.5m_ FCM-FCM; Doric) to a photoreceiver (Newport 2151) to produce a voltage that is proporjonal to t he intensity of the emimed light. Voltages were digijzed at 2 kHz with either a Najonal Instruments acquisij on board (NI USB-6343) or a Labjack (T7) and saved to disk as 8trials/sweeps9 lasjng 5320 s in durajon ea ch using Wavesurfer soqware (Janelia). To character ize the photoacjvajon behavior of RdLight1, we delivere d brief pulses (136 ms in durajon) of blue light (N Ac, DLS: 470 nm LED, Thorlabs M470F3; VLS: 470 nm laser, Optoengine) to the brain via the same üuorescence mini cube and patch cord used for photometry. We tested a range of blue light powers (measured at the jp of the patch cord) frequently used for optogenej c manipulajons in vivo. The jming, durajon and intensity of blue light pulses were controlled digitally using Wavesurfer, with each sjmulajon paramet er repeated at minimum 10 jmes per mouse/recording sit e. RdLight1 photoacjvajon responses are extremely stable over jme and were reliably seen fo r the durajon of 1 h-long recording sessions. Fiber Photometry Analysis. Photometry signals were processed and analyzed oÿine using custom code in MATLAB (Mathworks) and Igor Pro 6.02A (Wavemetrics). Raw voltages collected from the photoreceiver &" = were down sampled to 20 Hz and converted to 8percent changes in üuorescence9 using the equajon " "#"!, where ! is the mean baseline üuorescence, computed for each trial/sweep during a 1.532 s baseline "! window preceding the blue light sjmulus. Blue excit ajon light led to an instantaneous arjfact in the r ed channel for the exact durajon of the blue light pul se, which was blanked in display panels. For each experiment, 10312 replicates were performed. In ûgures, gray traces represent single trials, while colored traces represent the mean of 10312 trials, with standard error of the mean (SEM) shown as a shaded area. The properjes of RdLight1 photoacjvajon transients [peak amplitude, latency to peak (i.e., from blue light onset to RdLight1 photoacjvajon peak) and dec ay jme constant (i.e., jme from peak to 37% of peak)] were measured for each mouse using averaged waveforms. Data are reported in the text and ûgures as mean ± SEM. N-values represent the number of mice.

Immunohistochemistry. Mice were deeply anesthejzed with isoüurane and p erfused transcardially with 4% paraformaldehyde in phosphate buûered saline (PBS). Brains were post-ûxed for 24 h and secjoned coronally (100 ¿m in thickness) using a vibratome (Leica; VT1000S). Brain secjons were mounted on superfrost slides and coverslipped with ProLong anj fade reagent with DAPI (Molecular Probes). RdLight1 üuorescence was not immuno-enhanced. Whole secjons were imaged with an Olympus VS120 slide scanning microscope.

Reagents. To inhibit DA vesicular transport and prevent vesicular release of DA, mice were injected intraperitoneally with the irreversible vesicular monoamine transporter inhibitor reserpine (5 mg/kg) for 24 h prior to RdLight1 photometry.

Figure 1 3 Figure supplement 1: Blue light evokes RdLight1 photoacOvaOon in VLS. (A) Injecjon and ûber implantajon scheme for mice expressing RdLight 1 in the VLS (leq). Images of sensor expression from a representajve mouse (right; scale bar: 800 µ m). (B) Mean (± SEM) RdLight1 photoacjvajon signal in the VLS in response to 1, 2, 4, or 6 ms single pulses of blue light sjmulajon (10 mW) couple d with constant 565 nm illuminajon (30 µW). Colored v erjcal bars indicate laser sjmulajon arjfacts which were removed. Mean maximal RdLight1 photoacjv ajon signals recorded under the indicated sjmulajon widths are displayed. Each dot represents the average signal from a single mouse, and error bars denote the standard deviajon (N = 3 mice). (C) Mean (± SEM) RdLight1 photoacjvajon signal in the VLS in response to a pair of 4 ms blue light pulses (10 mW) separated by the indicated inter-pulse intervals. Mean paired pulse rajos are displayed (m agnitude of pulse #2/pulse #1), with each dot represenjng a single mouse, and the error bars deno jng the standard deviajon (N = 3 mice). (B) Mean (± SEM) RdLight1 photoactivation in response to 1, 4, 16, 32, or 64 blue light pulses (9 mW, 4 ms width, 16 Hz frequency) in the NAc of a representative mouse. Green bar illustrates constant 565 nm illumination. Blue bars show when blue light is delivered. (C) Same as (B) for pairs of blue light pulses (9 mW, 4 ms width) separated by 2.5, 5 and 10 s.(D) Mean paired-pulse ratio (magnitude of pulse #2/pulse #1) across three inter-pulse intervals in each of N=8 mice.

Mohebi A, Pe€bone JR, Hamid AA, Wong JT, Vinson LT, Patriarchi T, Tian L, Kennedy RT and Berke JD (2019). Dissociable dopamine dynamics for learning and mojvajon. Nature 570(7759): 65-70. Patriarchi T, Cho JR, Merten K, Howe MW, Marley A, Xiong WH, Folk RW, Broussard GJ, et al. (2018). Ultrafast neuronal imaging of dopamine dynamics with designed genejcally encoded sensors. Science 360(6396).

Patriarchi T, Mohebi A, Sun J, Marley A, Liang R, Dong C, Puhger K, Mizuno GO, et al. (2020). An expanded paleme of dopamine sensors for muljplex im aging in vivo. Nat Methods 17(11): 1147-1155. Shaner NC, Lin MZ, Mckeown MR, Steinbach PA, Hazelwood KL, Davidson MW and Tsien RY (2008). Improving the photostability of bright monomeric orange and red üuorescent proteins. Nat Methods 5(6): 545-551.

Sulzer D, Cragg SJ and Rice ME (2016). Striatal dopamine neurotransmission: regulajon of release and uptake. Basal Ganglia 6(3): 123-148.

Threlfell S, Lalic T, Plam NJ, Jennings KA, Deisseroth K and Cragg SJ (2012). Striatal dopamine release is triggered by synchronized acjvity in cholinergic in terneurons. Neuron 75(1): 58-64.

Wang L, Shang S, Kang X, Teng S, Zhu F, Liu B, Wu Q, Li M, et al. (2014). Modulajon of dopamine releas e in the striatum by physiologically relevant levels of nicojne. Nat Commun 5: 3925.

Wei W, Mohebi A and Berke JD (2022). A spectrum of jme horizons for dopamine signals. BioRxiv.