From extraocular photoreception to pigment movement regulation: a new control mechanism of the lanternshark luminescence

The velvet belly lanternshark, Etmopterus spinax, uses counterillumination to disappear in the surrounding blue light of its marine environment. This shark displays hormonally controlled bioluminescence in which melatonin (MT) and prolactin (PRL) trigger light emission, while α-melanocyte-stimulating hormone (α-MSH) and adrenocorticotropic hormone (ACTH) play an inhibitory role. The extraocular encephalopsin (Es-Opn3) was also hypothesized to act as a luminescence regulator. The majority of these compounds (MT, α-MSH, ACTH, opsin) are members of the rapid physiological colour change that regulates the pigment motion within chromatophores in metazoans. Interestingly, the lanternshark photophore comprises a specific iris-like structure (ILS), partially composed of melanophore-like cells, serving as a photophore shutter. Here, we investigated the role of (i) Es-Opn3 and (ii) actors involved in both MT and α-MSH/ACTH pathways on the shark bioluminescence and ILS cell pigment motions. Our results reveal the implication of Es-Opn3, MT, inositol triphosphate (IP3), intracellular calcium, calcium-dependent calmodulin and dynein in the ILS cell pigment aggregation. Conversely, our results highlighted the implication of the α-MSH/ACTH pathway, involving kinesin, in the dispersion of the ILS cell pigment. The lanternshark luminescence then appears to be controlled by the balanced bidirectional motion of ILS cell pigments within the photophore. This suggests a functional link between photoreception and photoemission in the photogenic tissue of lanternsharks and gives precious insights into the bioluminescence control of these organisms.


Results
In vitro characterization of the Opn3 photopigment. In vitro characterization of the Es-Opn3 was performed to examine its ability to perceive light and determine its absorption spectrum. The protein was expressed in COS1 mammalian cells and purified pigments were successfully obtained and characterized as a blue-sensitive pigment with an absorption spectrum ranging from 410 to 490 nm ( Supplementary Fig. S1). Following Terakita et al. (2008), C-terminal truncated construction was also expressed in the mammalian cells 83 . The C-terminal truncation resulted in a more than 2-fold higher yield of the purified Es-Opn3-based pigment, which allowed to determine the maximum absorption of the Es-Opn3-based pigment at 445 nm (Fig. 1g).
Light induced ip 3 modulation within E. spinax photophores. IP 3 is acting on IP 3 receptors to mobilize intracellular Ca 2+ that play a significant role in some photoresponse components. To determine the effect of light stimulations on photogenic skin (i.e. photophore region), putative modulation of IP 3 concentrations after light exposure was evaluated on E. spinax ventral epidermis. Ventral skin patches enlighten with monochromatic light of 415 (bluish-violet), 480 (azure-blue) or 630 (orange-red) nm wavelengths presented a significant variation of the IP 3 intracellular level (P < 0.05). E. spinax skin patches exposed during 15 min revealed an IP 3 concentration level of 1787.7 ± 186.9 pg mL −1 at 415 nm, 2316.4 ± 139.4 pg mL −1 at 480 nm and 1345.6 ± 149.7 pg mL −1 at 630 nm (Fig. 2a). Thirty minutes exposure revealed a concentration of 2234.6 ± 280.5 pg mL −1 at 415 nm, 1728.4 ± 164.5 pg mL −1 at 480 nm and 1195.8 ± 248.8 pg mL −1 in red light (Fig. 2a). Tissues exposed during 45 min under 415 nm light presented an IP 3 concentration of 1695.5 ± 105.5 pg mL −1 ; at 480 nm, a concentration of 1582.8 ± 225 pg mL −1 and at 630 nm, a concentration of 988.2 ± 113.6 pg mL −1 (Fig. 2a). Ventral skin patches preserved in dark condition were used as controls. They presented an IP 3 intracellular level not significantly different from 630 nm experiments (P > 0.05); 15 min exposure led to an IP 3 concentration level of 1115.7 ± 223.9 pg mL −1 ; 30 min, an IP 3 concentration of 643.3 ± 182.3 pg mL −1 and 45 min, an IP 3 concentration of 1000.7 ± 101.9 pg mL −1 (Fig. 2a). All values were expressed per gram of tissue. Therefore, E. spinax ventral skin, full of photophores, react to blue light wavelengths (e.g. similar to shark luminescence) by modulating the intracellular concentration of IP 3 .
iris-like structure regulation pathway. IP 3 triggers photophore opening in a calcium dependent manner. To determine the effect of an IP 3 increase on the E. spinax light emission and photophore pigmentation state, application of D-myo-IP 3 on ventral skin patches was performed. In parallel, effect of calcium absence was followed. IP 3 was applied on freshly dissected ventral skin patches at a concentration of 10 −4 M 84 . Alone, IP 3 application does not trigger any light emission (Fig. 3a,b). The absence of light emission was also observed for IP 3 application on skin patches in absence of free calcium (Fig. 3a,c). The pictures of the final pigmentation state of treated tissue samples revealed fully open photophores in the case of the IP 3 treatment in presence of calcium (Fig. 3b). Conversely, the IP 3 treatment in absence of calcium shown dark closed photophores (Fig. 3c). As a control, light emission was efficiently recorded after melatonin application in a calcium-free saline containing a calcium chelator (Fig. 3a,d). In that case, pictures of final pigmentation state of experiments revealed fully open photophores (Fig. 3d).
These data together show that (i) IP 3    www.nature.com/scientificreports www.nature.com/scientificreports/ Calmodulin is involved in the light emission. As already depicted by , melatonin application at 10 −6 M triggers lanternshark luminescence and ILS opening that allow the light to come out of the photophore (Fig. 4a,b) 44 . Conversely, the application of a calmodulin inhibitor (trifluoperazine, 10 −5 M 85 ) followed by a MT treatment with (Fig. 4a,d) or without (Fig. 4a,e) an increase of calcium does not generate any light emission ( Fig. 4a,d,e). The in vivo calcium increase was performed through the application of the calcium ionophore A23187. This result highlights the putative implication of the calmodulin in the regulation of the light emission. Control with an application of the calmodulin inhibitor followed by two applications of shark saline does not trigger any light emission (Fig. 4a,c). The pictures of the final pigmentation state of treated tissue samples revealed fully closed/dark photophores in the case of the calmodulin inhibitor alone or followed by a MT application (with or without calcium increase) (Fig. 4c-e).
So, calcium-dependent calmodulin is, here, demonstrate as essential to open the E. spinax ILS photophore.
Dynein is involved in the light emission. Ciliobrevin D, an inhibitor of the dynein minus-end intracellular motor, was used at a concentration of 10 −5 M to investigate the implication of dynein in the light emission process. Skin patches subjected to a first application of MT followed by a second application of ciliobrevin D emitted light ( Fig. 5a-c). A rapid decrease of the light emission was then observed with a third application of α-MSH consistently with the observations of Claes and Mallefet (2009) (Fig. 5a) 44 . Conversely, an initial application of the ciliobrevin D, prevented the light emission with (Fig. 5a,e) or without (Fig. 5a,d) an increase of the in vivo calcium. The pictures of the skin patch pigmentation at the end of the experiment show (i) open photophores for the MT treatment followed by the ciliobrevin D treatment (Fig. 5b); (ii) closed photophores for the MT treatment followed by the ciliobrevin D and by an α-MSH application (Fig. 5c); (iii) closed photophores for both experiments starting with the ciliobrevin D treatment followed by MT application (Fig. 5d,e), even with an intracellular calcium increase (Fig. 5e). Control treatments with only ciliobrevin D or A23187 applications followed by shark saline applications did not trigger the light emission (data not shown). Therefore, these results highlight (i) the important role of dynein in the pathway regulating the opening of the photophore ILS, (ii) the non-implication of this later protein in the melanocortin regulation pathway.
Kinesin is involved in the light emission. Conversely to the dynein, the kinesin is a plus-end intracellular motor which carries pigment vesicles from the nucleus periphery to the cell extremity. SUK4 antibody was used to inhibit the action of kinesin 86,87 . Starting with MT application, light emission was observed during the first 15 min (Fig. 6a). Luminescence was maintained after a second application with MT ( Fig. 6a,b) or shark saline (Fig. 6a,c). Conversely, a rapid decrease of the light emission was shown with the secondary application of the SUK4 antibody (Fig. 6a). This rapid decrease of luminescence was not observed in the control treatment (i.e. secondary application of glycerol) (Fig. 6a,d).
Finally, a decrease of the emitted light was recorded for all experiments after a third drug, α-MSH, application (Fig. 6a). With the α-MSH last application, inhibiting the light emission, closure of photophores was expected ( Fig. 6b-d). However, photogenic skin patch pre-treated with the kinesin inhibitor, SUK4, presented fully open photophores (Fig. 6e) similarly to the ones observed during the melatonin application alone (Fig. 4b). Control treatments with only SUK 4 or glycerol application followed by applications of shark saline did not trigger any light emission (data not shown). www.nature.com/scientificreports www.nature.com/scientificreports/ Thus, kinesin is, here, demonstrate as main actor in the translocation of pigmented granules leading to the closure of the ILS photophore. Moreover, this protein is showed to also be involved in the light emission regulation.

Discussion
By deciphering the molecular and cellular processes underlying the lanternshark counterillumination behaviour using molecular and pharmacological assays, this work highlights the interlinked pathways of bioluminescence, photoreception and pigmentation control within photophores. Our work, demonstrated the implication of (i) Es-Opn3 and (ii) actors involved in both MT and α-MSH/ACTH pathways on the velvet belly lanternshark bioluminescence and ILS cell pigment motions. Our results revealed the implication of Es-Opn3, MT, IP 3 , intracellular calcium, calcium-dependent calmodulin and dynein in the ILS cell pigment aggregation. Conversely, our results also highlighted the implication of an α-MSH/ACTH pathway involving kinesin to regulate the ILS cell pigment dispersion.  www.nature.com/scientificreports www.nature.com/scientificreports/ Based on previous pilot studies revealing an abundant and singular presence of Es-Opn3 within the photogenic ventral skin tissue 77,78 , the opsin absorbance spectrum was measured and revealed a blue-green absorption wavelength: between 410 and 490 nm, with a maximum value of 445 nm. This result is consistent with recent studies that have determined and characterized photoreceptive properties for several vertebrate Opn3 homologs ( Table 1) such as fishes 88 or other vertebrates 88,89 . The mosquito Opn3 (Mos-Opn3) was shown to form a bistable photopigment, absorb blue-green light, and activate Gi/o protein in a light-dependent manner [90][91][92] . Despite an increasingly number of newly-discovered opsin 3 homolog sequences both in invertebrates and vertebrates, the molecular properties of these proteins remain elusive with only a few studies describing the Gi/Go activation and the light-dependent cAMP modulation 90,92 . The measured Es-Opn3 absorption spectrum strongly suggests that lanternsharks can detect both environmental and their own light at the photophore level since a clear overlap with the intrinsic luminescence emission spectrum is observed.
Our results show that deep blue and blue-green light exposures cause a clear modification of the IP 3 intracellular level, adding evidence of a blue light extraocular photoreception at the level of the ventral skin (Fig. 2). No significant IP 3 concentration change was observed for the red light or dark condition. No cAMP concentration change was observed during light exposure experiments neither in blue nor in red or dark conditions. These results are contradictory to what is known for the Mos-Opn3 literature as an efficient decrease of cAMP was observed for cells expressing Mos-Opn3 homolog 90,92 . Even if previous studies depicted an inefficient activation of Gq protein in Mos-Opn3 expressing cells 90 , our results suggested that Es-Opn3 would rather be linked to a Gq than a Gi/o protein, or that β/γ subunit of the Gi/o protein might be able to trigger an IP 3 intracellular modulation and do not modify cAMP intracellular level. Therefore, Es-Opn3 is assumed to directly act through its specific G protein on the inositol lipid signalling system and trigger the activation of G protein-regulated phospholipase C. Further studies need to be conducted to precisely describe the downstream pathways of this deep-sea lanternshark specific opsin.
In addition to Opn3, the hormones controlling bioluminescence (MT, α-MSH, ACTH) are also shown to regulate skin pigmentation [6][7][8][9][10][11][12][13][14][15][16] . For this reason, we investigated the possible links between these proteins. Going further in the transduction pathways, pharmacological results demonstrated the interlinking of both bioluminescence, photoreception, and pigmentation in the light emission control. Our results (i) showed that melatonin does not require calcium to trigger light emission and (ii) highlight the involvement of the Ca 2+ -binding protein calmodulin and minus-end cellular motor dynein in the MT action way (Figs. 3-5). The classical MT pathway is related to the inhibition of the adenylate cyclase activity through the activation of a Gi α subunit protein [93][94][95][96] . The β/γ subunits of the Gi protein, release from the α subunit after the melatonin receptor activation, also display various effector actions such as the opening of calcium channels or the activation of specific phospholipases 97,98 .
Pharmacological assays also highlighted that IP 3 acts at the level of the ILS melanophore-like cells and is needed to aggregate ILS cell pigments. IP 3 is already known as leading the aggregation of pigmented vesicles in fish and amphibian chromatophores 84,99 . However, IP 3 increase does not directly trigger the luminescence of E. spinax (Fig. 3). In the same manner, our results demonstrate the necessity of calcium and the calmodulin activity to trigger the aperture of the ILS cells (Figs. 3, 4). Calcium is involved in the organelle motility in a huge variety of cell types, including fish chromatophores in which an increase of intracellular Ca 2+ level triggers the aggregation of pigmentary organelles in the perikaryon [100][101][102] . Furthermore, studies demonstrated that the regulation of pigment motion involving Ca 2+ is mainly provided by intracellular storage [103][104][105] . Other studies also highlighted the action of calmodulin activity on the cell pigmentation in various species (i.e. melanophores and erythrophores) where an increase of the Ca 2+ -binding protein calmodulin activity leads to the aggregation of pigmented vesicles in the nucleus periphery 103,106,107 . Although not studied in this work, the role of calcineurin has also been demonstrated in fish pigment vesicle aggregation and the protein is assumed to act in the IP 3 transduction pathway regulating the aperture of the photophore ILS cells. Calcineurin mediates pigment aggregation in fish melanophores as a Ca 2+ /Calmodulin-stimulated phosphatase which dephosphorylates a 57 kDa protein, freeing cytoplasmic dynein 85 . Through the calcineurin phosphatase activity, previous studies depicted the involvement of melanophore-located cytoplasmic dynein as a minus-end cellular motor that carries pigmented granules/vesicles toward the nucleus periphery leading to more lighter melanophore cells and skin 12,108-110 . Following the pathway, our results demonstrated the necessary involvement of dynein to aggregate pigment and open the lanternshark photophores (Fig. 5). Combining all these pharmacological data with the literature supports a first model into which the ILS aperture occurs through the activation of dynein and the movement of pigmented granules allowing light to go out of the photophore. A potential pathway is, hence, suggested for the ILS aperture: (i) through the Es-Opn3 luminescence perception and activation, intracellular level of IP 3 increases leading to (ii) the release of stored intracellular Ca 2+ which activates (iii) the Ca 2+ -dependent calmodulin activity, which in turn, (iv) triggers the activity of calcineurin phosphatase stimulating (v) the cytoplasmic dynein. This later could, then, transport melanin granules from the cellular periphery toward the nucleus periphery and allow the outward light emission (Fig. 7). www.nature.com/scientificreports www.nature.com/scientificreports/ Our data also decipher a second pathway involved in the photophore ILS closure and the inhibition of E. spinax light emission. Firstly, results clearly support the independency between the α-MSH pathway and dynein activity since ciliobrevin D dynein inhibition does not block the photophore closure by α-MSH application (Fig. 5c). By opposition with the first described cascade, this second pathway involves the cellular plus-end motor kinesin which is demonstrated to be needful to close the photophore but also to trigger the light emission (Fig. 6). Interestingly, when blocking kinesin activity with the SUK4 antibody 86 , (i) light emission was suppressed and (ii) ILS melanophore-like cell pigments remained fully aggregated even with later α-MSH treatment (Fig. 6). Here, the observed light emission suppression leads to assume the involvement of the kinesin to translocate granules containing potential bioluminescent reaction compounds or accessory activators of light reaction (i.e. cofactors) from the nucleus side of the photocyte to the other side (i.e. from vesicular to granular area 71,72 ). Indeed, luminous reactions can involve either a luciferase/luciferin or a photoprotein in which a co-factor such as an ion is needed. Concerning the ILS pigment motion regulation, previous studies demonstrated the implication of various molecules in the α-MSH/ACTH and MCR pathway such as Gαs, adenylate cyclase and cAMP in the luminescence control of E. spinax 44,48,76 . Studies on metazoan melanophore granule dispersion highlighted the effector role of the cAMP-dependent protein kinase (PKA) in the MSH/MCR pathway in which the second messenger cAMP increases due to the adenylate cyclase up-regulation leading to the release of the catalytic subunit of PKA 7,88,[111][112][113]. This kinase was shown to directly phosphorylate a granule-bound 53-57 kDa protein which leads to the activation of the specific heterotrimeric kinesin II as microtubule motor responsible for the pigment dispersion 7,114,115 . In addition to these previous researches, our data strongly suggest that this second pathway occurs in the ILS cells to close the photophore and prevent light to go out (Fig. 7). Therefore, data demonstrate that ILS cells pigment motion is under a dual and antagonist control by IP 3 /Ca 2+ and cAMP pathways controlling www.nature.com/scientificreports www.nature.com/scientificreports/ the bidirectional granule movement in the melanophore-like cells, co-opted from skin pigmentation regulation in metazoan 103,116 .

Maximum absorption wavelength (nm) References
Lanternsharks have a range of ways to control their luminescence and remained cryptic to avoid being spotted by underneath swimming predators. The mechanism allowing shallow water organisms to hide from predators through physiological colour change [1][2][3][4][6][7][8][9][10][11][12][13][14][15][16] , is here demonstrated to regulate the light emission of counterilluminating lanternsharks. This new luminescence control mechanism might also occur in other bioluminescent marine organisms displaying light organ associated melanophores for the counterillumination regulation such as cephalopods 117,118 and bony fishes 119,120 . conclusion Using molecular and pharmacological assays, our results help to decipher some of the molecular pathways underlying the lanternshark counterillumination behaviour. This work highlights the functional interconnexion of bioluminescence, photoreception and pigment motion control within the photophores. Our data allow us to propose two interlinked pathways that regulate the pigment movement in specific melanophore-like cells within the ILS of the photophore and regulate the amount of light emitted by the lanternshark E. spinax. With clear overlapping between the lanternshark light emission wavelength and the Es-Opn3 absorption spectra, associated with the illumination assay and the pharmacological data, one of the transduction cascades starting with the absorption of blue-green light (probably the emitted bioluminescence) and resulting in the ILS pigment aggregation and, at the photophore level, the light passage is proposed. On the other side, α-MSH/ACTH known to inhibit light emission is described as acting on the transduction cascade leading to the pigment dispersion and, at the photophore level, avoidance of light emission. Finally, data also demonstrated that MT is involved in the ILS-cell pigmentation regulation by firstly downregulating α-MSH/ACTH transduction cascade, and secondly is implicated in the aperture of the photophore allowing light to go out of the photophore.
Evidence of the link between photoreception and photoemission in lanternshark is presented and highlights the implication of Es-Opn3 photopigment as a regulator of light emission through ILS cell pigment aggregation regulation. Although many other transduction cascades leading shark bioluminescence remain to be described, the present study highlight pathway parts regulating mechanically, through pigment bidirectional motion regulation, the light emission supervising the bioluminescence use for camouflage by counterillumination in the velvet belly lanternshark, E. spinax.

Material and methods
fish & tissue collections. Twenty-four adult specimens of E. spinax were caught alive during a field session in May 2018 by longlines lowered at 220 m depth in the Raunefjord, Norway (60°169′ N; 05°089′ E) 44,48,71,72,121 . Specimens were directly placed in a 1 m 3 tank filled with running fresh seawater (6 °C) and kept in a dark cold room at Espegrend Marine Station (Bergen University, Norway) 44,48,71,72,121 . Animal procedures were conducted in compliance with the Belgian national guidelines and in agreement with the European directive 2010/63/UE, under the approval of the Animal Ethics Committee of the Université catholique de Louvain in Louvain-la-Neuve. Sharks were treated according to the European regulation for animal research handling and euthanized following the local rules for experimental vertebrate care 73,76,77,121 . Sharks were sexed, measured and weighed before experimentation took place. Following Duchatelet et al. (2020), round shape piece of skin (6 mm diameter; ±100 mg of fresh tissue) were dissected from the ventral skin of sharks 44,48 . These skin patches, full of photophores, were rinsed in shark saline (292 mmol l −1 NaCl, 3.2 mmol l −1 KCl, 5 mmol l −1 CaCl 2 , 0.6 mmol l −1 MgSO 4 , 1.6 mmol l −1 Na 2 SO 4 , 300 mmol l −1 urea, 150 mmol l −1 trimethylamine N-oxide (TMAO), 10 mmol l −1 glucose, 6 mmol l −1 NaHCO 3 ; total osmolarity: 1.080 mosmol; pH 7.7 122 ) during half a day at 4 °c before pharmacological tests and light exposure tests happen. Sixteen sharks were used for the pharmacological experiment and eight sharks were used for the light exposure assay.

Opn3 absorbance measurement.
Recent studies highlight the presence of an opsin 3 colocalized with E. spinax photophores 77,78 . Following this research, the opsin 3 mRNA sequence was in silico extracted from the transcriptome, and cDNA was created. To increase the purification efficiency of the pigment, an Es-Opn3 deletion mutant was constructed having a shorter C terminus 83,90,123 . The cDNA of the C-terminal-truncated lanternshark Opn3 was generated from full-length cDNA. The C-termini of full length and C-terminal-truncated Es-Opn3 were tagged with the rho 1D4 epitope sequence (ETSQVAPA) 124 . The tagged Opn3 cDNAs were inserted between the Hind III and Eco RI sites of the pcDNA3.1 expression vector (GenScript USA inc.). The expression and purification of the Es-Opn3 were performed as previously described 123,125,126 . Opsin expression vectors were transfected into COS1 cells (a human embryonic kidney cell line) using the polyethyenimine (PEI) transfection method 123,125 . Transfected cells were collected two days post-transfection 123,125 . Opsin-based pigments were extracted with 1% dodecyl-ß-D-maltoside (DM) in HEPES buffer (pH 6.5) containing 140 mM NaCl and 3 mM MgCl 2 (buffer A), after addition of 11-cis retinal. Extracted pigments were bound to 1D4-agarose gels, washed with 0.02% DM in buffer A and eluted with buffer A containing 0.02% DM and c-terminal peptide of bovine rhodopsin as described 123,125 . The absorption spectra of the opsin-based pigments were recorded at 4 °C by using a Shimadzu UV2450 spectrophotometer 123,125 . www.nature.com/scientificreports www.nature.com/scientificreports/ to mimic the shark light emission (~486 nm), deep blue light (415 nm) was selected to see if shark skin patches can perceive shorter wavelength, and, finally, red light (630 nm) was selected to see the specificity of the shark Opn3. Two skin patches of each shark were maintained in shark saline in a fully dark condition to display dark state control. Skin patches were exposed to light during 15, 30 or 45 min. Thus, two patches of one shark are subjected to one wavelength during one period. After each period, skin patches were removed and directly frozen at −80 °C. One skin patch was devoted to inositol triphosphate (IP 3 ) concentration level assay and the second, to the cyclic adenosine monophosphate (cAMP) concentration level assay. Each skin patches were weighted before to be homogenized in phosphate buffer saline with 0.02% Triton X-100 (Sigma) thanks to a grinder (T10 Basic UltraTurrax, IKA) on ice, and then centrifuged at 15000 rpm for 5 min at 4 °C. Supernatants were recovered, aliquoted and stored at −20 °C until experiments took place. cAMp assay. cAMP concentrations were measured on the second skin patches of each shark treated according to light and time of exposure thanks to cAMP-Glo Assay kit (V1501, Promega, Madison, WI, USA) following Duchatelet et al. 48 . Briefly, supernatants of each treatment were placed in a well of a 96-wells plate and treated according to the kit manufacturer's instructions. Luminescence induced during the assay was recorded thanks to a microplate luminometer (Berthold MPL12/Orion; Pforzheim, Germany) coupled with the Berthold simplicity software (http://www.titertekberthold.com/). According to cAMP-Glo Assay kit linear regression manufacturer standard, results are expressed in cAMP concentration and refine standardized in cAMP concentration per square centimetre (µM cm −2 ). pharmacological strategies. To decipher the different steps of the pathway involved in the extraocular light perception and the hormonal control of the light emission, luminometric assay after drug application blocking or triggering step by step transduction events were done. Investigations were done to decipher the implication of (i) IP 3 and calcium; (ii) Ca 2+ -dependent calmodulin; (iii) cytoplasmic dynein; and (iv) kinesin in the transduction pathways leading to light emission control. The following drugs were employed: melatonin (M5250, Sigma) activates the light production in lanternshark 44 ; α-MSH (M41135, Sigma) inhibit the light production by lanternshark photophores 44 ; D-myo-Inositol 1,4,5-tris-phosphate (I9766, Sigma) serves to manually increase the IP 3 level in the skin patches; BAPTA (1,2-Bis(2-aminophenoxy)ethane-N,N,N' ,N'-tetraacetic acid, 14513, Sigma) is a calcium chelator; calcium ionophore A23187 (C7522, Sigma) trigger an increase of calcium in the light organ cells; trifluoperazine hydrochloride (T2000000, Sigma) inhibit the Ca 2+ -dependent calmodulin activity; cytoplasmic dynein inhibitor, ciliobrevin D (250401, Sigma) acts as a reversible and specific blocker of AAA + ATPase motor cytoplasmic dynein; and antibody against kinesin heavy chain (SUK4, Developmental studies, Hybridoma Bank, University of Iowa) inhibit the kinesin driven microtubule motility 86 . Glycerol was used as control for SUK4 experiments as the antibody is diluted in this substrate. Used concentrations were extracted from literature and are summarized in Table 2.

The implication of Es
Patches of ventral skin were placed in 96-well plates and subject to various treatments (Table 3). A total of three applications of drugs occurred during one experiment. The first application was done at the beginning of the experiment, the second application at 15 min and the last one at 30 min. Data were recorded for 45 min thanks to a microplate luminometer (Berthold MPL12/Orion; Pforzheim, Germany) coupled with the Berthold simplicity software. pigmentation visualization. At the end of each lumino-pharmacologic assay, a picture of each treated ventral skin patch has been taken to evaluate the pigmentation state. Pictures were taken thanks to a Lumix DMC-FZ300 camera (Panasonic Corporation, Osaka, Japan).

Drugs
Concentration (mol l −1 ) References www.nature.com/scientificreports www.nature.com/scientificreports/ Statistical analyses. All analyses [ANOVA, post-hoc Tukey tests] were performed with the software JMP pro v.14 (SAS Institute Inc., Cary, NC, 1989NC, -2007. The Gaussian distribution (Shapiro test) and the homoscedasticity (Levenee test) were obtained for all analyses after logarithm transformation allowing the use of parametric tests. ANOVA was used to show significant differences between groups while post-hoc Tukey tests allowed the different clusters to be distinguished.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.  www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/