A clock in mouse cones contributes to the retinal oscillator network and to synchronization of the circadian system

Multiple circadian clocks dynamically regulate mammalian physiology. In retina, rhythmic gene expression serves to align vision and tissue homeostasis with daily light changes. Photic input is relayed to the suprachiasmatic nucleus to entrain the master clock, which matches behaviour to environmental changes. Circadian organization of the mouse retina involves coordinated, layer-specific oscillators, but so far little is known about the cone photoreceptor clock and its role in the circadian system. Using the cone-only Nrl-/- mouse model we show that cones contain a functional self-sustained molecular clockwork. By bioluminescence-combined imaging we also show that cones provide substantial input to the retinal clock network. Furthermore, we found that light entrainment and negative masking in cone-only mice are subtly altered and that constant light displayed profound effects on their central clock. Thus, our study demonstrates the contribution of cones to retinal circadian organisation and their role in finely tuning behaviour to environmental conditions.


Introduction
by cell autonomous molecular oscillators constituted of oscillating auto-regulatory clock 23 transcription factors able to drive gene expression programs, hence cellular physiology. The 24 retina plays a particular role in the circadian system in mammals because it is responsible for 25 the unique photosensory input to ensure entrainment of the clock in the SCN to the LD cycle 26 (Yamazaki, Goto, & Menaker, 1999). 27

28
The retina was the first circadian clock identified outside the SCN, based on the capacity of 29 explanted tissue from hamsters to secrete melatonin in a rhythmic manner (Tosini & Menaker, 30 1996). Since the retinal clock is able to synchronize to the LD cycle in vitro, this tissue 31 constitutes on its own a complete circadian system, with molecular clock machinery, resetting 32 input mechanism and biological outputs (Felder-Schmittbuhl et al., 2018;McMahon, Iuvone, 33 & Tosini, 2014). Besides melatonin synthesis the retina displays a plethora of rhythmic 34 properties, including expression of photopigment genes, processing of light information, 35 phagocytosis of photoreceptor outer segments, metabolism, together contributing to adapt 36 visual function to the LD cycle and ensuring tissue homeostasis (for review (Felder-37 Schmittbuhl, Calligaro, & Dkhissi-Benyahya, 2017)). Given the complexity of the retinal 38 tissue comprising glial cells and six major types of neurons, identification of the cell type(s) 39 constituting its main oscillator has been a matter of debate. Analysis of clock gene expression 40 in vitro and ex vivo suggested that the retina is composed of several layer-specific, coupled 41 oscillators (Dkhissi-Benyahya et al., 2013;Jaeger et al., 2015;Sandu, Hicks, & Felder-42 Schmittbuhl, 2011) but the existence/identity of a main driver remains under question. Several 43 lines of evidence, notably the presence of melatonin synthesis machinery (Gianesini, Clesse, 44 Tosini, Hicks, & Laurent, 2015;Niki et al., 1998), the detection of cycling clock factors (Liu, 45 Zhang, & Ribelayga, 2012), have pointed to cones as a potential retinal clock component but 46 their precise contribution to the network has not been evaluated. 47 The retina possesses a laminar organisation as well as parallel microcircuits processing light 48 information. Photon capture occurs in photoreceptors, highly specialized cells located in the 49 outer retina. Cones respond to bright light (photopic vision) and mediate color vision whereas 50 rods are much more sensitive and function under low intensities (scotopic vision). In mice, 51 most cones (95%) are M-cones which express 2 types of opsins (short wavelength -sws, with 52 maximal sensitivity at 360 nm and middle wavelength -mws, with peak sensitivity at 509 53 nm) and a minority of these cones express either the blue or the green opsin alone respectively 54 in the ventral and dorsal regions of the retina (Applebury et al., 2000;Hughes, Watson, 55 Foster, Peirson, & Hankins, 2013). Investigation of cone properties has been challenging 56 given their low number in retinas of routinely used laboratory mammals, i.e., <3% of total 57 photoreceptors in mice (Jeon, Strettoi, & Masland, 1998) and <1% in rats (Szel & Rohlich, 58 1992). The Nrl -/mouse (Mears et al., 2001), in which absence of the NRL transcription factor 59 totally blunts rod generation, has a cone-only retina with a majority of S-cones, and has been 60 extensively used to study cone properties without the interference from rods (Krigel, Felder-61 Schmittbuhl, & Hicks, 2010;Liu et al., 2012;Wenzel et al., 2007). 62 63 Studies from the last 20 years have led to improved understanding of how information linked 64 to light perception in the eye is conveyed to the SCN and translated into a message reflecting 65 the alternation of day and night, able to entrain the central clock. In particular, a minor, light-66 sensitive population of retinal ganglion cells (RGC) expressing the melanopsin photopigment 67 (intrinsically photosensitive RGC or ipRGC) constitutes the (unique) cellular connection 68 between the retina and the SCN (Goz et al., 2008;Guler et al., 2008;Hatori et al., 2008). 69 Despite the major role played by these blue sensors (peak sensitivity = 480 nm), some data 70 demonstrate a role for rods in synchronisation to the LD cycle at low light intensities (Altimus 71 et al., 2010;Boudard, Mendoza, & Hicks, 2009;Lall et al., 2010) and also for cones (Dkhissi-72 Benyahya, Gronfier, De Vanssay, Flamant, & Cooper, 2007;van Diepen, Ramkisoensing, 73 Peirson, Foster, & Meijer, 2013;van Oosterhout et al., 2012). In addition, recent results in 74 mice suggested that cones also play a role in entrainment mechanisms by perceiving spectral 75 changes characteristic of dusk or dawn (Mouland, Martial, Watson, Lucas, & Brown, 2019;76 Walmsley et al., 2015). However, these functions have not been investigated with gain of 77 function mutants, in particular for cones. 78 Here we investigate the role of murine cones in the circadian system. We show that cones in 79 the Nrl -/retina harbor a functional molecular clock, the elements of which are similar to other 80 central or peripheral clocks. Furthermore, the cone population contributes, together with the 81 inner and ganglion cell layers, to the oscillatory network of the retina. However, light-82 mediated behavior seems to be altered in the cone-only retina from Nrl mutant mice in acute 83 and chronic light exposure conditions, particularly at low light intensities. This suggests that 84 total replacement of rods by cones induces modifications in the global non-image forming 85 visual function of the retina in mice. 86 87

88
A functional clock in cone photoreceptors 89 We first aimed to characterize the cone molecular clock on microdissected photoreceptors 90 isolated from the Nrl KO mice over 24 h in DD ( Figure 1A). We found that all core clock 91 gene transcripts examined, Bmal1,Clock,Per1,Per2,Per3,Cry1,Cry2,Rorβ are 92 expressed in cones ( Figure 1B, top panel). Significant rhythmic levels of expression were 93 determined for Bmal1, Per1, Per2, Per3, Rev-Erbα (Table 1). Interestingly, the expression 94 profile of Bmal1 was in opposite phase in comparison to the profiles of Per transcripts, as 95 described in the SCN (King & Takahashi, 2000;Welsh, Takahashi, & Kay, 2010) and other 96 peripheral tissues such as liver (Noguchi et al., 2010;Oishi, Sakamoto, Okada, Nagase, & 97 Ishida, 1998). 24 h profiles in cones were also similar, at least for Bmal1 and Per1 transcripts, 98 to those reported for mouse whole retinas sampled in DD (Ruan, Allen, Yamazaki, & 99 McMahon, 2008). 100 We also investigated the expression of several well-known or putative target genes of the 101 retinal clock such as S-opsin, Crx,arrestin 3,lower panel). S-opsin, arrestin 3 and Cnga3, expressed in S-cones (Mears et al., 2001), 103 displayed significantly rhythmic profiles (Table 1). 104 To further evaluate the capacity of cones to sustain rhythmicity, we used a vibratome-based 105 sectioning of the retina to isolate photoreceptor (cone-only) layers from the KO mice raised 106 on the Per2 Luc reporter background (Yoo et al., 2004) for real-time bioluminescence 107 recordings. Photoreceptor layers from WT mice were used as control. As previously described 108 (Jaeger et al., 2015) the latter showed robust PER2:: LUC oscillations with a 26.46 ± 0.02 h 109 period ( Figure 1C). Cone layers from the KO retinas also proved robustly rhythmic in culture, 110 but yet with a significantly longer period: 29.07 ± 0.03 h (n = 6 for WT, n = 9 for KO; 111 genotype effect: p = 0.018) ( Figure 1C). 112 Finally, we examined how cone layers oscillate within the context of the whole retina by 113 using in vitro real-time bioluminescence combined with imaging. 100 µm transversal sections 114 were cut using the vibratome technique illustrated in Figure 1D, transferred on a 115 semipermeable membrane, then cultured and imaged for several days in a temperature 116 controlled microscope chamber. PER2 bioluminescence signal emerged from all layers, with 117 higher intensity in ganglion and inner cell layers and weaker signal in the outer, photoreceptor 118 layer ( Figure 1E). Moreover, the PER2 signal was rhythmic in all layers ( Figure 1F), with 119 distinct free-run periods (24.28 ± 0.26 h for ganglion cell layers, 27.79 ± 0.20 h for inner 120 nuclear layers and 26.80 ± 1.19 h for photoreceptor layers; Figure 1G Table 1    Clock properties in a cone-only retina 181 We then examined how the unique presence of cones affects the overall circadian function in 182 the mouse retina. Previous bioluminescence studies demonstrated that mouse retina displays 183 oscillatory expression of PER2 (Jaeger et al., 2015;Ruan et al., 2008). To evaluate how a 184 cone-only photoreceptor population impacts on the retinal clock we performed real-time 185 bioluminescence recordings of whole retinal explants from Nrl -/-Per2 Luc mice, compared to 186 WT controls ( Figure 1H). Sustained circadian oscillations of PER2 bioluminescence were 187 observed for several days, with no difference in period (p = 0.557), relative rhythmic power (p 188 = 0.273) and damping (p = 0.583) comparing to the wild-type littermates ( Figure 1I). 189 However, the amplitude of the oscillations was significantly reduced in the mutants by 35% (p 190 = 0.013). Moreover, a significant reduction of the baseline levels was observed in mutant 191 mice as compared to wild-type (p = 0.015) ( Figure 1H and 1I). were measured at ZT4 with respect to ZT16 in the wild-type animals but not in Nrl mutants 222 (2-way ANOVA post hoc analysis: p = 0.003 and p = 0.07, respectively) indicating that the 223 response to light is altered in the KO retinas, as previously described for rodless mice. Results 224 are represented as mean ± SEM. **: p < 0.01. ***: p < 0.0001. 225

SCN-driven rhythms are preserved in the cone-only mouse
226 SCN explants from Nrl -/-Per2 Luc mice produced autonomous and sustained PER2::LUC 227 oscillations for at least 6 days in vitro similar to those from WT ( Figure 2A). The robustness 228 of rhythms was similar between genotypes, based on the relative rhythmic power (p = 0.344), 229 indicating that in the cone-only mutant the master clock is not impaired ( Figure 2B). 230 Moreover, there was no effect on the phase of the oscillations (p = 0.938) and on the 231 amplitude (p = 0.476), period (p = 0.944) and damping (p = 0.09) ( Figure   Wheel-running activity of additional WT and KO mice also used for the analyses presented in 288 Figure 2H,I. Note that in a few cases some LD data are missing due to technical recording 289 problems. 290

292
To evaluate behavioral response to acute light exposure, WT and KO mice were first exposed 293 to a phase-resetting protocol. Thus, animals received a 15 min light pulse with different 294 intensities, at the beginning of the constant dark period. Light pulses of high (170-220 lux), 295 medium (14-20 lux) or low (1.0-1.3 lux) intensities provided at projected ZT15 induced a 296 phase-delay in the onset of activity of both WT and Nrl mutant mice ( Figure 3A). The 297 ANOVA shows a light intensity effect on phase-shifts (2-way ANOVA, p = 0.006), but not a 298 genotype effect (p = 0.251) nor an interaction between light intensity and genotype (p = 299 0.485) ( Figure 3B). This indicates that the response to the 15 min light pulse was not altered 300 in the absence of rods, at least not down to 1 lux light. 301 Secondly, in the negative masking protocol, a 3 h light pulse applied 2 h after the lights off 302 inhibited locomotor activity ( Figure 3C) with a significant light intensity effect (p < 0.001) 303 and an interaction between light intensity and genotype (p = 0.004). Mutant animals indeed 304 showed reduced masking effect at lowest light intensities (p < 0.01 at <1 lux and p < 0.001 at 305 1-10 lux) ( Figure 3D). 306 to the total activity during the preceding night (p < 0.001) and showed significant interaction 319 between light intensity and genotype (p = 0.004) (WT n = 6, KO n = 7). Post hoc analysis 320 shows significant differences between genotypes for the two lowest stimuli (p < 0.01 at <1 lux 321 and p < 0.001 at 1-10 lux). Results are represented as mean ± SEM. *: p < 0.01. Grey shading 322 indicates darkness. 323 supplement 1). WT animals were able to entrain to each shifted LD cycle at different light 329 intensities while the KO mice needed longer time to entrain at 1 lux (jet-lag 3, p = 0.026) and 330 were not able to entrain at 0.1 lux (jet-lag 4, p < 0.001) even after 50 days ( Figure 4B). 331 Subsequent exposure to total darkness (DD, 22 days) confirmed that almost all mutant 332 animals were free-running in the previous condition ( Figure 4A and Figure 4-figure 333 supplement 1; data not shown). When animals were subsequently exposed to LD at 100 lux 334 during the light phase, animals from both genotypes were able to re-entrain ( Figure 4B

406
In the present study we used different approaches to determine the role of cones in the 407 circadian system. We show that photoreceptor layers lacking functional rods but having 408 normal cone and cone-pathway contain a molecular machinery characteristic of a functional 409 clockwork and likely contribute, together with the inner and ganglion cell layers, to the 410 overall clock rhythmicity in the retina. We bring evidence that the Nrl -/retina also displays 411 novel distinctive properties regarding light impact on the central clock, providing new insight 412 into the role of cones in the circadian system. 413 Rhythmic functions in mammalian cones have been only poorly documented (Bobu, Sandu, 414 Laurent, Felder-Schmittbuhl, & Hicks, 2013;Liu et al., 2012;Sakamoto, Liu, Kasamatsu, 415 Iuvone, & Tosini, 2006;Storch et al., 2007;von Schantz, Lucas, & Foster, 1999), likely 416 because of the scarcity of this cell type in nocturnal rodents (Jeon et al., 1998;Szel & 417 Rohlich, 1992). To circumvent this limitation, we used the Nrl -/animal model in which all 418 rods are replaced by cones (Akimoto et al., 2006;Mears et al., 2001). These photoreceptors 419 were previously shown to have major characteristics of native blue cones regarding 420 morphology, molecular content, nuclear architecture and light response (Akimoto et al., 2006;421 Daniele et al., 2005;Mears et al., 2001;Nikonov et al., 2005)  The literature agrees on a main contribution from the inner retina (Jaeger et al., 2015;Ruan et 430 al., 2008) and several reports exclude rod-type photoreceptors from the circadian network 431 (Baba et al., 2018;Liu et al., 2012;Ruan, Zhang, Zhou, Yamazaki, & McMahon, 2006) 432 although the presence of sustained clock gene rhythms in rods has been suggested elsewhere 433 (Dkhissi-Benyahya et al., 2013;Jaeger et al., 2015;Sandu et al., 2011;Tosini, Davidson, 434 Fukuhara, Kasamatsu, & Castanon-Cervantes, 2007). Upon immunofluorescence analysis of 435 clock protein factors, cones appeared the most evident cell-autonomous clock in the mouse 436 retina (Liu et al., 2012). In agreement with this study, we here describe robust rhythms in 437 expression of clock genes from the main (Bmal1, Per1, Per3) and secondary (Rev-Erbα) 438 loops of the well described molecular machinery (Takahashi, 2017) in Nrl -/photoreceptor 439 layers laser-microdissected throughout the 24 h cycle in constant dark condition. However, 440 unlike what was described for immunostained clock factors, their mRNAs show distinct 441 phases, as observed at the level of the whole retina (Ruan et al., 2008), which might be due to 442 the enrichment in S-versus M-cones in the KO retinas or suggest post-transcriptional 443 regulation of clock factors. We previously described that cones are the photoreceptor site of 444 robust oscillations in Aanat (the enzyme responsible for melatonine rhythm) expression by 445 using a diurnal, cone-rich rodent, Arvicanthis ansorgei (Bobu et al., 2013;Gianesini et al., 446 2015). Besides, circadian rhythms in cone-specific genes have essentially been investigated in 447 chicken (Haque et al., 2010;Pierce et al., 1993) and zebrafish (P. Li et al., 2008). In 448 particular, robust rhythms in phototransduction genes in zebrafish cones appear driven by key 449 transcription factors (Neurod, Crx) themselves regulated by the clock (Laranjeiro & 450 Whitmore, 2014). In our study, cones express major phototransduction elements in a rhythmic 451 manner with high amplitudes but we did not detect any rhythm in Crx expression, indicating 452 that in mammalian cones phototransduction elements retain clock regulation but with 453 mechanisms distinct from the zebrafish. Importantly, when isolated by vibratome-sectioning 454 of fresh retinas, cone layers express sustained rhythms with a specific period, distinct from the 455 period measured in photoreceptor layers from control mice. This observation probably reflects 456 the differences in clock machinery and associated signalling occurring in rods (97% of 457 photoreceptors in WT) versus cones. It might also reflect a difference in coupling strength 458 within the respective photoreceptor populations, as previously described in the retina (Jaeger 459 et al., 2015). Communication through gap junctions might be reduced in the S-cone enriched 460 photoreceptor layers of the KO, since expression of connexion 36 was shown to be absent in 461 this cone population in mammals (W. Li & DeVries, 2004). This might be responsible for the 462 increased period in the KO (Jaeger et al., 2015). Taken together with our demonstration of 463 rhythmic phagocytosis of cone outer segments (Krigel et al., 2010)  The involvement of cones in circadian functions has been substantially documented. A role in 483 synchronisation of the SCN has been demonstrated for both green cones (Dkhissi-Benyahya 484 et al., 2007) and S-cones (Provencio & Foster, 1995;van Diepen et al., 2013;van Oosterhout 485 et al., 2012;Walmsley et al., 2015). However, the contribution of cones to the effects of white 486 ambient lighting on circadian properties and more generally non-image forming vision, has 487 been evaluated with a limited variety of visually impaired mouse models. We used a battery 488 of behavioural tests (Hughes, Jagannath, Hankins, Foster, & Peirson, 2015) Lall 497 et al., 2010). Indeed, no defect was detected in Nrl -/mice under the phase shift paradigm (as 498 also seen in (Calligaro et al., 2019)). By contrast, our model rather displays some features 499 typical of rodless animals, such as reduced capacity to shift at low light intensity (1 or 0.1 lux) 500 in a jet-lag experiment, as previously observed with the Gnat1 -/model . 501 Physiological features of rodless retina are also reflected in dopamine metabolism (Figure 1-502 figure supplement 2D), with the loss of daily rhythm of DOPAC generation in KO retinas 503 (Nir & Iuvone, 1994;Perez-Fernandez et al., 2019) as previously described for the rds strain. 504 The discrepancy between the results from light pulse and the jet-lag experiments might be due 505 to the fact that the threshold levels required for entrainment constitute a more sensitive test of 506 deficit in entrainment than phase shift following a light pulse (Mrosovsky, 2003). 507 508 Rats or mice with outer retinal impairment were repeatedly reported to exhibit total loss of 509 positive masking by light and (consequently) enhanced inhibition of locomotor activity 510 (negative masking), especially at low light intensities (Mrosovsky, Foster, & Salmon, 1999;511 Thompson et al., 2010;Thompson et al., 2011). By contrast, melanopsin phototransduction 512 appears indispensable for negative masking (Mrosovsky & Hattar, 2003). Using 513 monochromatic light, Thompson et al. also provide evidence that cones (short-and medium-514 wavelength sensitive) contribute to negative masking and influence its dynamic range 515 (Thompson, Foster, Stone, Sheffield, & Mrosovsky, 2008). In the present study, the Nrl -/-516 animals show reduced negative masking behaviour specifically at low light intensities 517 (between 0.5 and 10 lux), despite a normal ipRGC population and unlike most rodless mice. 518 The discrepancy between this result and the literature might be explained by the fact that we 519 used global activity recordings and not wheel running activity. Indeed, positive masking 520 might be more pronounced when using wheel running activity and hence introduce a 521 confounding effect (increased negative masking in rodless animals) at low light intensity. 522 Furthermore, some data also indicate that rods contribute, at least transiently, to negative 523 masking at light intensities too low to excite ipRGC (Butler & Silver, 2011). Thus, the 524 behaviour triggered in the Nrl -/animals by acute light stimulation probably reflects the 525 combined absence of rods and integrity of ipRGC. 526 527 Increase of the endogenous period in constant light has also been partly attributed to rod 528 signalling Lall et al., 2010) and requires the integrity of ipRGC (Goz et 529 al., 2008). In our experiments we observed that the free running periods in LL were first 530 increased to a similar extent for both the WT and mutant mice, suggesting that mechanisms 531 distinct of the rod-pathway are involved. However, periods then decreased, with WT reaching 532 a mean value around 24 h and the mutants rather lower periods (23.25 h on average). In 533 addition, KO mice exhibit particularly high (around 8-fold increase with respect to the WT) 534 level of activity, indicating loss of masking by constant light, a feature which is also shared 535 with mice devoid of ipRGC (Goz et al., 2008). However, upon re-exposure to a standard 12 536 h:12 h LD cycle after the LL, both WT and KO mice re-entrained very rapidly, suggesting 537 that there was no major impairment of the circadian photosensitivity. Short free-running 538 period values have been rarely described in LL, except in Per2 clock gene mutants of 539 different backgrounds (Pendergast, Friday, & Yamazaki, 2010;Spoelstra & Daan, 2008;540 Steinlechner et al., 2002). The phenotype in the Nrl -/mice could be explained by distinct 541 hypotheses: 1, their high wheel running activity in LL might feedback on the clock and induce 542 period shortening (Edgar, Martin, & Dement, 1991); 2, cone abundance could trigger another, 543 yet unknown signalling towards the central clock. Identification of the mechanisms by which 544 excess of cones alters properties of the circadian system will require further investigation. 545 546 In conclusion, by using the Nrl -/cone-only mouse model we provide compelling evidence 547 that cones contain a circadian clock part of the retinal oscillating network. Although Nrl -/-548 mice do not exhibit overt dysfunction of circadian behaviour, their exposure to specific 549 experimental paradigms highlights their particularities, namely properties induced by the 550 absence of rods or, importantly, specific to the enlarged cone population and revealed by 551 Experiments were performed on both males and females unless otherwise stated. In most 572 cases, no a priori estimation of sample size was performed. Our groups were based on 573 previous or preliminary data and tried to conform to the 3R rule. 574

575
Six week-old Nrl -/males (n = 30) reared in LD were exposed to constant dark (dark/dark, 576 DD; no dim red light). After 36 h in DD mice were euthanized within the following 24 h in 577 DD in a CO2 (up to 20%) airtight chamber at the following projected ZT time points: 0, 4, 8, 578 12, 16, 20 (n = 5, randomly allocated, per time point). Eyes were enucleated, embedded in 579 Tissue-Tek OCT compound (Sakura Finetek USA, Torrance, CA), frozen on dry ice and 580 stored at −80°C until use. Animal handling and eye sampling were performed by using night 581 vision goggles (ATN NVG-7, ATN-Optics, Chorges, France). 582 20 µm thick eyeball sections were cut on cryostat and placed on polyethylene naphthalate 583 (PEN) Membrane Frame slides (Life Technologies, Grand Island, NY). Three to four slides (4 584 sections/slide) were prepared from a single eye specimen. Each slide was stored at −80°C in a 585 50 mL nuclease-free tube (pre-chilled on dry ice) and used for laser microdissection within a 586 week. 587 Frozen slides were thawed at room temperature for 30 s. Sections were stained with cresyl 588 violet (1% cresyl violet acetate in 70% ethanol) for 30 s, then dehydrated through a series of 589 ethanol solutions: 2 x 75% for 30 s, 95% for 30 s, 100% for 30 s and 100% for 2 min. Slides 590 were air dried at room temperature for 1 min then completely dehydrated in a vacuum 591 chamber for 1 h before microdissection. The whole procedure was performed in RNase free 592 conditions. 593 Laser microdissection was performed using the Veritas Microdissection Arcturus system and 594 software (Arcturus Bioscience, Inc. Mountain View, CA, USA) immediately after complete 595 dehydration of the slides. The cone photoreceptor areas of interest were selected under 596 microscope (20x magnification) and transferred on CapSure Macro LCM Caps (Life 597 Technologies, Grand Island, NY) by the combined use of the infrared (power 70-80 mW, 598 pulse 1500-3500 µs) and UV (low power 2-4) lasers (See also Figure 1A). A total cone 599 photoreceptor area of 3 mm 2 was collected per eye. In order to prevent RNase reactivation 600 and RNA degradation, the microdissection was carried out within maximum 60 min for each 601 slide. 3-4 caps/eye were collected into the same reaction tube which contained RLT + lysis 602 buffer (Qiagen, Hilden, Germany) and stored at −80°C. Transcript levels were determined by quantitative PCR as described (Sandu et al., 2011), with 620 PCR reactions run in duplicates. The purity of the microdissected samples was verified by the 621 absence of detection by qPCR, of transcripts for tyrosine hydroxylase (Th) gene and 622 metabotropic glutamate receptor 6 (mGluR6) gene, as markers for the inner nuclear layer. 623 Transcript levels were normalized to the levels of Tbp and Hprt which showed constant 624 expression in the isolated cones over the 24 h (data not shown). Transcript levels in whole 625 retinas were normalized to the levels of Gapdh and Hprt which did not vary between 626 genotypes (data not shown). All TaqMan probe-based assays were purchased from Applied 627 Biosystems (Applied Biosystems, Foster City, CA, USA) and designed to span exon 628 boundaries (Table 2). Data was quantified using the ΔCq method, modified to take into 629 account gene-specific amplification efficiencies and multiple reference genes, and the qBase 630 software (free v1.3.5) (Hellemans, Mortier, De Paepe, Speleman, & Vandesompele, 2007). In 631 microdissected cones, log transcript levels were calculated relative to the transcript levels 632 measured in a WT photoreceptor sample which were rescaled to one. We used Excell 633 software to detect outliers which were removed for the final statistical analysis (n = 1 for 634 Per1, Per2 and Per3 quantification). 635 retinas from the same animal are considered as independent, biological replicates. We here 665 analysed n = 12 (8 mice) for WT and n = 12 (7 mice) for KO. 666 667 Photoreceptor layer explant cultures 668 Retinas were dissected as described above. Photoreceptor layers were isolated using the 669 vibratome technique and cultured as reported previously (Jaeger et al., 2015). WT (n = 6 670 samples, 6 mice) and KO (n = 9 samples, 8 mice) photoreceptor explants were recorded for at 671 least 5 days and the photons were integrated for 112 s every 15 min. Exceptionally, when 672 layers of insufficient size were collected, samples from both retinas were cultured together (

SCN bioluminescence recordings
687 Animals (9 month-old, WT n = 5, KO n = 7, Per2 Luc background) were killed by cervical 688 dislocation and brains were rapidly removed and placed in ice-cold HBSS. One 500μm 689 coronal section of the SCN region was obtained using a stainless steel adult mouse brain slicer 690 matrix (ZIVIC Instruments, Pittsburgh, USA), then trimmed to 1 × 1 mm. Each SCN explant 691 (containing both nuclei) was cultured onto a Millicell culture membrane (Merck Millipore 692 Ltd, Tullagreen, Ireland) in a 35-mm culture dish with 1 mL of DMEM (Sigma-Aldrich) 693 supplemented with 0.35% D(+)-glucose, 0.035% sodium bicarbonate, 10 mM HEPES, 2% 694 B27, antibiotics (25U/mL penicillin and 25mg/mL streptomycin) and 0.1 mM beetle luciferin. 695 Culture dishes were sealed with vacuum grease. The bioluminescence was recorded using the 696 LumiCycle for 112 s in 15 min intervals and during at least 6 days. 697 Bioluminescence data analysis 698 Whole retina and SCN explant PER2::LUC raw data were subtracted with a 24 h running 699 average (removal of the baseline drift) using the LumiCycle analysis software (Actimetrics, 700 Wilmette, IL, USA). The first cycle was removed and the analysis was performed on the 701 following 4 (retina) or 5 (SCN) cycles. The robustness of the rhythms (relative rhythmic 702 power (Klarsfeld, Leloup, & Rouyer, 2003)) and the phase were also calculated using the 703 LumiCycle analysis software. The period, amplitude and damping rate were determined using 704 a cosinor derived sine wave function: f = y0 + a * exp (-x/d) * sin [2 * π * (x + c) / b] where a 705 is the amplitude (counts/s), b is the period (h), c is the phase-related term (h) and d is the 706 damping rate (days) and assuming that damping follows an exponential pattern. Baseline for 707 each individual peak in retinal samples was estimated as the baseline from LumiCycle 708 analysis taken at the peak time. 709 Photoreceptor layer data were analyzed as previously described, on 4 successive cycles 710 (Jaeger et al., 2015). 711 Bioluminescence data from whole retinas and photoreceptor layers were obtained over several 712 series of recordings: samples for which activity did not exceed the background of Lumicycle 713 were excluded from the study 714 715 Transversal retinal images were analyzed with ImageJ (open source software 716 https://imagej.nih.gov/). A median 3D filter was applied to remove the hotspots. The ganglion 717 cell layer (GCL), inner nuclear layer (INL) and photoreceptor layer (PRL) were defined as 718 regions of interest (ROI) and the bioluminescence levels (grey levels) were measured and 719 exported for the analysis of rhythmicity. The periods were determined using the cosinor 720 derived sine wave function: f = y0 + a * exp (-x/d) * sin [2 * π * (x + c) / b] as above. 721 722 Retina whole-mount immunohistochemistry 723 Immunohistochemical staining was performed on whole retinas obtained from 6-8 week-old 724 Nrl -/-Per2 Luc mice (WT n = 5, KO n = 6). Eyes were sampled by enucleation from mice 725 euthanized between ZT3 and ZT6 by cervical dislocation and immediately fixed in 4% 726 Aldrich) were diluted in the same mobile phase in order to obtain a 3-point standard curve for 752 each standard for the quantification of the samples. 753 754 Electroretinography 755 Electroretinography was used to assess visual sensitivity of Nrl -/mice in the visible spectrum 756 using the RETI port / scan 21 setup (Stasche & Finger GmbH,Roland Consult,Brandenburg,757 Germany) as previously reported (Ait-Hmyed Hakkari et al., 2016). All recordings were 758 obtained around the middle of the animal's light phase between projected ZT5 and ZT7. Dark-759 adapted mice (n=5 for WT, n=3 for KO) were anesthetized by subcutaneous. injection of 760 ketamine (50 mg/kg; Imalgène 1000; Merial, Lyon, France) and xylazine (10 mg/kg; Rompun 761 2%; Bayer, Puteau, France). Pupils were dilated with 0.5% Tropicamide (Ciba Vision 762 Ophthalmics, Blagnac, France). Animals were then placed on a warming plate to maintain a 763 constant body temperature, and ground, reference, and corneal electrodes (thin gold wire with 764 a 2-mm ring end) were placed accordingly. Eyes were kept moist with eye drops (Ocry-Gel; 765 TVM Lab, Lempdes, France). Mice were then exposed to a rod-saturating white light 766 background (40 cd/m 2 ) inside the Gansfeld bowl. After 10 min of light-adaptation, single-767 flash photopic ERG recordings were performed successively at specific wavelengths 455 and 768 525 nm and then under white light, at 1, 3 and 10 cd.s/m 2 (6 flashes per intensity). Amplitudes 769 of a and b-waves were analyzed off-line: only the b-wave was measurable for photopic ERG. Per2 Luc knock-in allele) were housed in individual standard cages equipped with a 10-cm-774 diameter stainless steel running wheel (Mendoza, Graff, Dardente, Pevet, & Challet, 2005) or 775 with infrared detectors placed above the cage and linked to an automated recording system 776 (CAMS, Circadian Activity Monitoring System, Lyon, France) as previously described 777 (Salaberry, Hamm, Felder-Schmittbuhl, & Mendoza, 2019). Data were collected in 5 min bins 778 and analysed with the ClockLab Software (Actimetrics, Wilmette, IL, USA). Locomotor 779 activity data were double-plotted in actograms. 780

781
To determine the daily and circadian rhythm of locomotor activity in Nrl mutant mice, 5-6 782 month-old mice (WT n = 4, KO n = 7, Per2 Luc background) were initially maintained for 12 783 days under LD 12:12 and then 19 days under constant darkness (DD). Total activity and rho-784 and alpha-phase activity levels were calculated during LD and the endogenous period (Chi-785 square Periodogram method) was determined over 10-day interval after 7 days from the 786 transition to DD. 787 Behavioural phase-shifts to light pulses 788 To evaluate phase shifting in response to light pulses 5-6 week-old mice (WT n = 5, KO n = 789 8) were initially maintained in LD 12:12 (100 lux) and then challenged by 3 alternating DD 790 (9-14 days) -LD (14-18 days) cycles. On the day before each light-pulse, the room lights 791 went off at ZT12. On the following day a 15 min light pulse (LP) was applied at the projected 792 ZT15. Then lights remained off for at least 9 days before re-exposing animals to LD 793 condition. The intensity of the light pulses decreased one order of magnitude as indicated in 794 Figure 3A, B. To determine phase changes in control and Nrl mutant mice, a linear regression 795 analysis of the activity onsets was performed by projecting the onset phase of the free run in 796 DD back to the mean onset phase under LD condition (ClockLab). 797

798
To evaluate the negative masking response to light, 3-6 month-old mice (WT n = 6, KO n = 7, 799 Per2 Luc background) mice adapted to 12:12 LD cycle were housed in individual cages into a 800 ventilated cabinet (Charles River Laboratories, France) equipped with broad spectrum white 801 light lamp (MASTER TL-D Super 80 lamp, Philips). The masking effect of light was tested 802 by exposing the animals to light for 3 hours from ZT14 to ZT17 at successive light intensities 803 as follows: day 1 (baseline) -standard 12:12 LD; day 2 -ZT14-17 at <1 lux; day 4 -ZT14-17 804 at 1-10 lux; day 6 -ZT14-17 at 10-50 lux; day 8 -ZT14-17 at 200-400 lux; days 3, 5, 7 -805 standard 12:12 LD. Locomotor activity was monitored with infrared cage top motion sensors 806 connected to the CAMS data acquisition system (Circadian Activity Monitoring system, 807 INSERM, Lyon, France) (Dkhissi-Benyahya et al., 2007). The percent of activity during the 3 808 h light pulse was calculated relative to the 12 h activity of the preceding standard night. 809 Re-entrainment to 6-h light-dark cycle delay 810 1.5-3 month-old mice (WT n = 5, KO n = 5) were kept for 23 days in LD at 100 lux (LD1) 811 and then challenged with 4 successive 6-h phase delays, mimicking a jet-lag (JL) or cycle 812 change across six time zones, combined with reduction of light intensity: JL1 (21 days, 100 813 lux), JL2 (22 days, 10 lux), JL3 (51 days, 1 lux) JL4 (50 days, 0,1 lux). At the end of the last 814 JL exposure, animals were transferred to DD (22 days) and then re-exposed 25 days to LD at 815 100 lux (LD2). The phase angle of entrainment was determined by calculating the difference 816 between the time of lights off and the time of activity onset (ClockLab). 817 818 We tested the effects of constant light exposure (light/light, LL) on cone-only animals by 819 assessing wheel running activity in 6 month-old mice (WT n = 6, KO n = 7, Per2 Luc 820 background). Thus, after 10 days in LD 12:12 animals were transferred to LL for 70 days at 821 200 lux. Total activity per cycle, period and relative rhythmic power were measured by using 822

Exposure to constant light
ClockLab. Mice were then exposed to a second LD cycle (LD2: 10 days) to evaluate if 823 entrainment and locomotor activity returned to baseline levels. 824

825
Results are expressed as means ± SEM, except for qPCR data. Statistical analyses were 826 performed by using SigmaPlot 12 software (Systat Software, San Jose, CA, USA). 827 Comparison of two groups was performed by using the Student's t test. Comparison of several 828 groups was performed by using 1-way or 2-way ANOVA for independent and repeated 829 measures, followed by post hoc test (Holm-Sidak test). 830 Data from qRT-PCR over 24 h in DD were also analyzed by nonlinear least-square fitting of a 831 24 h sinusoid (cosinor analysis) f = a + [b*cos(2* π *(x -c) ⁄ 24)] (Nelson, Tong, Lee, & 832 Halberg, 1979). A posteriori Power analysis was also performed and is presented in Figure 1-833 figure supplement 1. 834 A statistically significant difference was assumed with p values less than 0.05. 835 836 837