Shark specimens were captured during the Chatham Rise Trawl survey by the R.V. Tangaroa in January 2020 off the coast of eastern New Zealand. The survey used the same eight-seam hoki bottom trawl and survey methodology that was used on previous surveys (Hurst et al., 1992; Stevens et al., 2018). The net has 100 m sweeps, 50 m bridles, 12 m backstrops, 58.8 m groundrope, 45 m headline, and 60 mm codend mesh. The trawl doors were Super Vee type with an area of 6.1 m².

The following depth range information are available: D. licha – mean maximal depth 678 ± 26 m (min-max 443–997 m); E. lucifer – mean maximal depth 542 ± 8 m [min-max 235–1078 m]; E. granulosus – mean maximal depth 903 ± 13 m (min-max 498–1269 m).

A total of 37 D. licha [40.9–138.0 cm total length (TL)], 304 E. lucifer (16.2–53.2 cm TL), and 281 E. granulosus (19.3–75.6 cm TL) were captured on the survey, of which 13 D. licha, 7 E. lucifer, and 4 E. granulosus were used for bioluminescence studies. Each specimen was maintained in a tank with fresh cold sea water in a dark cold room until manipulation. Each shark was sexed, measured, weighed (Supplementary Table 1) and photographed in dim daylight and in dark conditions using Sony α7SII camera before having a full incision of the spinal cord at the level of the first vertebrae, according to the European regulation for animal research handling. Ventral skin of a specimen of S. aliae and I. brasiliensis, collected, respectively, as in Delroisse et al. (2021) and Duchatelet et al. (2020b), were used for dalatiid comparative photophore histology.

Skin patches of 3 cm² were dissected from different locations along the body of D. licha specimens (i.e., rostral, mandibular, pecto-ventral, pectoral, ventral, dorsal, dorsal fin, pelvic, flank, infra-caudal, precaudal, and caudal zones; Figure 1A) to assess photophore presence, size and densities. Skin patches were fixed in 4% formalin at least overnight before being transferred to phosphate buffer saline (PBS). Skin patches were observed and photographed under a transmitted light microscope (Leitz Diaplan, Germany) coupled with a ToupCam camera (UCMOS Series C-mount USB2.0 CMOS camera, ToupTek, Zhejiang, China). Photophore densities (per mm²) and mean diameter (n = 30 or 50 zones per species) were also measured on the two etmopterid species using the same protocol (Supplementary Figure 1).

In parallel, skin patches of D. licha, E. lucifer, E. granulosus, S. aliae, and I. brasiliensis were used to perform histological sections across the photogenic organ. Skin tissues were bathed for 7 days in decalcifying solution (OsteoRAL, Fast decalcifier for Large Anatomical Specimens, RAL Diagnostics, France) with constant agitation and renewal of the solution every 2 days, rinsed in PBS, and placed in PBS with increasing concentrations of sucrose (10% for 1 h, 20% for 1 h, and 30% overnight). Tissues were then embedded in optimal cutting temperature compound (O.C.T. compound, Tissue-Tek, Netherlands) and rapidly frozen at −80°C. Sections of 10 μm were obtained with a cryostat microtome (CM3050S, Leica, Solms, Germany). Sections were placed on coated Superfrost slides (Thermo Scientific) and left overnight to dry. All sections were observed under a transmitted light microscope (Leitz Diaplan) equipped with a ToupCam camera (ToupTek).

In addition to the skin patches used for histology, round skin patches were dissected from the ventral luminous area of each shark using a metal cap driller (6 mm diameter) as described in Duchatelet et al. (2020b). Freshly dissected patches were rinsed and kept in shark saline [292 mmol L^–1 NaCl, 3.2 mmol L^–1 KCl, 5 mmol L^–1 CaCl₂, 0.6 mmol L^–1 MgSO₄, 1.6 mmol L^–1 Na₂SO₄, 300 mmol L^–1 urea, 150 mmol L^–1 trimethylamine N-oxide, 10 mmol L^–1 glucose, 6 mmol L^–1 NaHCO₃; total osmolarity: 1.080 mOsmol; pH 7.7 (Bernal et al., 2005)] at 4°C in dark conditions before being used for pharmacological tests.

Hormones known to trigger or inhibit light emission in luminous elasmobranchs were applied (Claes and Mallefet, 2009c; Duchatelet et al., 2020b). Here, evaluations of the effect of MT, α-MSH and ACTH were conducted for the first time on the dalatiid species, D. licha, and the etmopterid species, E. lucifer, and E. granulosus.

Experiments were first conducted on 10 D. licha specimens. To obtain a dose response curve for MT application, three different concentrations of MT (i.e., 10^–6, 10^–7, 10^–8mol L^–1) were used. Skin patches were immersed in 200 μL of MT solution (either 10^–6, 10^–7, 10^–8mol L^–1). To analyze the effect of α-MSH and ACTH on the light emission of D. licha, another set of skin patches were subjected to an immersion in 100 μL of MT 10^–6 mol L^–1 followed after 5 min by an application of 100 μL of either α-MSH 10^–6 mol L^–1 or ACTH 10^–5 mol L^–1. Luminescence of ventral skin patches subjected to the various treatments was measured using a FB12 tube-luminometer (Titertek-Berthold, Pforzheim, Germany) calibrated as in Duchatelet et al., 2020b. Lights emissions were recorded through FB12- Sirius, multiple kinetics software (Titertek-Berthold) for at least 30 min with a measurement every 58 s. For comparative purposes, similar treatments were performed on seven specimens of E. lucifer (same experiments) and four E. granulosus specimens (MT dose response and α-MSH treatments). In parallel, for D. licha and E. lucifer, photophore aperture and closure were observed after drug application by taking a time-lapse series of pictures (every 10 min) with a Sony α7SII camera mounted on a binocular microscope.

Luminescence measurements were characterized as follows (Duchatelet et al., 2020b): the maximum intensity of light emission [Lmax, in megaquanta per second (Mq s^–1)], the total amount of light emitted during experimentation [Ltot, in Gigaquanta per hour (Gq h^–1)] and the time to reach maximum light intensity [TLmax, in seconds (s)]. Inhibitory actions of α-MSH and ACTH were measured as the total amount of light emitted after the second drug application [Ltot_app, in Gq h^–1]. All light parameters were standardized according to the surface area of each skin patch (in cm²). A second treatment (α-MSH or ACTH) was added to the first one when the light intensity plateau was reached with the MT application, each timing being species-specific. Results of the luminescence decrease were expressed as a percentage of the maximal luminescence value (i.e., plateau MT) measured before the second application.

To evaluate the putative evolutive conservation of the hormonal control of light emission in dalatiids and etmopterids, pharmacological data on shark luminescence were extracted from literature.

All analyses were performed with the software R studio (version 1.1.383, 2009, R Studio Inc., United States). Variance normality and homoscedasticity assumptions were tested by Shapiro-Wilk and Levene’s test, respectively, before running ANOVA which reveals significant differences between skin photophore densities or pharmacological treatments. When these parametric assumptions were not met, a non-parametric Kruskal-Wallis ANOVA was used. Post hoc Tukey’s tests or Wilcoxon tests allowed pair-wised comparison of means, attributing different letters to significantly different values (P-value < 0.05).

A blue glow was observed on the ventral surface of D. licha, E. lucifer, and E. granulosus specimens kept in a fully dark environment (Figures 2A,D,E). D. licha also emit a faint blue glow from the lateral and dorsal areas and at the two dorsal fins (Figure 3 – Mallefet personal observation). Both etmopterids present a more complex pattern of light emission with flank marks, and lateral, dorsal, and rostral patterns (Figure 4; E. granulosus – Mallefet personal observation). Skin patches observed in toto present black round-shaped photophores distributed between placoid scales for all the observed sharks (Figures 1C–E and Supplementary Figure 1). The mean photophore diameters are 83.9 ± 9.5, 122.4 ± 10.8, and 132.3 ± 14.5 μm, for D. licha, E. lucifer, and E. granulosus, respectively. No statistical differences in photophore diameter were observed between the zones presenting large amount of photophores (ventral, pecto-ventral and infra-caudal) (D. licha ANOVA: F(2,183) = 1,1928, P-value = 0,3057; E. lucifer ANOVA: F(2,183) = 0,1014, P-value = 0,9036; E. granulosus ANOVA: F(2,183) = 0.1376, P-value = 0.8716).

Analyses of photophore density along the D. licha body show an increasing dorso-ventral repartition of photophores reaching up to 20.14 ± 4.01 photophores per mm² at the ventral side of the shark (Figures 1A,B,F). The lowest densities were observed for the caudal and dorsal areas with a mean density of 2.85 ± 1.11 and 4.85 ± 1.70 photophores per mm², respectively (Figure 1A). Statistical differences [ANOVA: F(11,514) = 128.64, P-value < 2.2 × 10^–16] in photophore densities are illustrated Figure 1F and Supplementary Table 2A. The scales of the rostral area were leaf-like in shape while the scales of the remaining body parts were pavement-like in shape (Figures 1C,D).

For both studied etmopterids, a high density of photophores was observed at the pectoral zone with 34.00 ± 6.20, and 15.63 ± 2.50 photophores per mm² for E. lucifer and E. granulosus, respectively. E. granulosus also have a high density of photophores at the infra-caudal and caudal zones. Conversely, for both species, only a few photophores were spread within the dorsal epidermis. Both species have a well-defined flank mark with photophores. All the remaining photophore densities and their respective statistical differences [E. lucifer ANOVA: F(9,490) = 263.39, P-value < 2.2 10^–16; E. granulosus ANOVA: F(10,298) = 175.16, P-value < 2.2 10^–16] are reported in Supplementary Figure 1 and Supplementary Tables 2B,C. Both etmopterids present needle-shaped placoid scales in all the studied zones (Supplementary Figure 1).

Histological sections across photogenic skin highlight the structure of D. licha photophores. Each light organ is embedded in the stratified squamous epidermis and is composed of a cup-shaped pigmented sheath containing a unique photocyte, topped by a lens cell with a few diffuse pigmented cells between the photocyte and lens cell (Figure 2A). This structural organization is similar to that found in S. aliae and I. brasiliensis photophores (Figures 2B,C).

Photophore morphologies of E. lucifer and E. granulosus are consistent with those already described for other etmopterids (i.e., E. spinax, Etmopterus molleri, and Etmopterus splendidus) (Figures 2D–F). They are composed of a cup-shaped pigmented sheath embedded with luminous cells and topped a with lens. They are similar to dalatiid photophores, but they harbor a higher number of photocytes, a larger iris-like structure area, and more lens cells (up to 3) (Figures 2D–F).

The effect of MT on D. licha, E. lucifer, and E. granulosus was tested through a dose-dependent response. For the studied species, MT 10^–6 mol L^–1 triggered a long-lasting light emission, significantly different from the MT 10^–8 mol L^–1 application (P-value < 0.05; Figures 5A–C and Supplementary Tables 3A, 4), while MT 10^–7 mol L^–1 triggered an intermediate light emission and Ltot value (Figures 5A–C and Supplementary Tables 3A, 4). All treatments were significantly different from the shark saline control, except for the MT 10^–7 and 10^–8 mol L^–1 treatments of E. granulosus (P-value < 0.05; Supplementary Tables 3A, 4). Although the total amount of light emitted under MT 10^–6 mol L^–1 treatment was significantly different [Kruskal-Wallis χ²(2) = 10.14, P-value = 0.0063], E. lucifer produced a mean total amount of light during the experiment 2.5 and 5 times higher than D. licha and E. granulosus, respectively. Similar patterns of bioluminescence were observed for the three species (Figures 2, 5).

The effect of α-MSH was evaluated for the three species after reaching the Lmax triggered through MT 10^–6 mol L^–1 application. Application of α-MSH 10^–6 mol L^–1 induced a rapid decrease of light emission for the studied luminous sharks (Figures 6A–C and Supplementary Table 3B). After MT-induced bioluminescence, Ltot_app values of α-MSH were statistically significant compared with the MT 10^–6 mol L^–1 control (Figures 6A–C and Supplementary Tables 3B, 5).

The effect of ACTH was evaluated on D. licha and E. lucifer bioluminescence. Similar to the results obtained for α-MSH, ACTH 10^–5mol L^–1 applications rapidly induced a decrease in light emission (Figures 6A,B and Supplementary Table 3B). Each Ltot_app value of ACTH 10^–5 mol L^–1 was not significantly different from those of α-MSH 10^–6 mol L^–1, respectively, but were statistically different from the MT 10^–6 mol L^–1 control (Figures 6A,B and Supplementary Tables 3B, 5). Mean values of Lmax, TLmax, Ltot, Ltot_app are presented in Supplementary Table 3.

The time-course of light emission in D. licha under MT 10^–6 mol L^–1 stimulation revealed a concomitant opening of the photophore ILS within 15 min of luminescence, in which the ILS stayed open for the next 30 min while the light level remained high (Figure 7A). In the case of E. lucifer MT-induced luminescence, a rapid opening of the photophore ILS was observed within 8 min followed by a slow decrease during which a closure of the ILS was visible (Figure 7B). Aperture and closure of photophores showed pigment movements concomitant to light emission.

The light emission pattern observed in D. licha is similar to that observed in previously studied dalatiids i.e., S. aliae and I. brasiliensis (Claes et al., 2012; Delroisse et al., 2021). The dorso-ventral gradient and the relative homogeneity in ventral photophore densities suggest the luminescence is used for counterillumination. The luminous pelvic zone of D. licha reveals a sexual dimorphism but, contrary to E. spinax and Etmopterus molleri (Claes and Mallefet, 2010b; Duchatelet et al., 2020c), it is not brighter than the rest of the ventral body, suggesting it is less important for sexual signaling. The kitefin shark D. licha, like other dalatiids, does not have flank marking or specific dorsal patterns. The lack of these luminescent patterns, previously suggested to be used as conspecific signaling for group aggregation, swimming, or hunting in etmopterids (Claes et al., 2015; Duchatelet et al., 2019b), rules out this function in D. licha. The aposematic function described for etmopterids (Claes et al., 2013; Duchatelet et al., 2019b) is also ruled out for D. licha luminescence due to the absence of dorsal fin defensive spines. Nevertheless, D. licha is the first shark with fully luminous dorsal fins (Figures 1A, 3), which raises questions about its luminescence function.

The light emission patterns of E. lucifer and E. granulosus, are similar to that of previously studied etmopterids. The dorsal photophores, flank markings, and brighter pectoral fin and claspers are likely to be used for intraspecific communications while the ventrally emitted light is likely to be used for counterillumination. These functions have been documented for E. spinax (Claes and Mallefet, 2009a; Claes and Mallefet, 2010a), Etmopterus molleri (Claes and Mallefet, 2015), and Etmopterus splendidus (Claes et al., 2011b). However, a bioluminescence aposematic function through specific spine-associated photophores (Claes et al., 2013; Duchatelet et al., 2019b) was not documented for E. lucifer and E. granulosus.

Reif (1985) postulated that a trade-off exists between the space occupied by placoid scales and luminous organs, and that four different types of placoid scales have evolved to allow this trade-off: pavement, cross-, bristle/needle-, and hook-shaped placoid scales. A new type of squamation with overlapping leaf-shaped placoid scales is present in the luminous rostral area of D. licha. This new bioluminescent-associated squamation was observed in the somniosid, Zameus squamulosus, which is assumed to be luminous (Straube et al., 2015). This new type of bioluminescence-associated placoid scale needs to be highly translucent or possess specific physical characteristics to allow efficient light transmission. The use of Reif placoid scale types to assess the bioluminescent status of a shark species is not a decisive character as shown by a recent study of Ferrón et al. (2018); highlighting the presence of bioluminescent-like squamation in a galeomorph shark, Apristurus ampliceps, a species not known to be luminous.

Histology revealed an evolutive conservation of photophore morphology across each family. Kitefin shark photophores are larger (mean diameter 83.9 μm) than those observed in S. aliae and I. brasiliensis [i.e., 50 and 56 μm, respectively (Claes et al., 2012; Delroisse et al., 2021)] while the internal structure of typical dalatiid photophores is conserved. Here, D. licha photophores are depicted as morphologically similar to those of S. aliae, S. laticaudus and I. brasiliensis (Seigel, 1978; Delroisse et al., 2021).

E. lucifer and E. granulosus showed typical etmopterid photophore histology (Claes and Mallefet, 2009b, 2015; Claes et al., 2011b; Renwart et al., 2014; Duchatelet et al., 2020b). These observations provide further insights on the evolutive conservation of light organ morphology across luminous squaliform radiation (Straube et al., 2015).

The effect of hormones on light emission in D. licha, E. lucifer and E. granulosus are consistent with increasing literature on light emission control in sharks (Claes and Mallefet, 2009c; Claes et al., 2012; Duchatelet et al., 2020b,d): MT, and α-MSH/ACTH, have been demonstrated as the main triggering and inhibiting agents of shark luminescence, respectively. Similar to observations of E. spinax and Etmopterus molleri photophores (Claes and Mallefet, 2010a; Duchatelet et al., 2020b), aperture and closure of D. licha and E. lucifer photophores involved pigment motion within the ILS cells. Simultaneities of curves kinetics and pigment motions highlight the evolutive conservation of hormonally controlled pigment motion regulating luminescence. These data strongly suggest that luminous etmopterids and dalatiids share a common luminescence control mechanism, involving at least MT, and α-MSH/ACTH hormones. This control is assumed to have been successfully and evolutionary co-opted from shark melanophore pigment motion control by a common ancestor of these two squaliform families. For both families, luminescence appears to be dually controlled at the level of (i) the photocyte, site of luminescent reaction, and (ii) the ILS cells, acting as a diaphragm capable of occluding light produced by the photocytes, via melanophore-associated pigment movements (Duchatelet et al., 2020b,d). This was recently demonstrated within ILS cells of the lanternshark, E. spinax (Duchatelet et al., 2020d) i.e., transduction pathways that activate cellular motors such as dynein and kinesin, leading pigment movements within ILS melanophores. The bioluminescence control mechanisms in the two studied etmopterids, as well as in D. licha, might share common features. Moreover, the involvement of extraocular photoreception events in the light emission control of photophores (Duchatelet et al., 2020d), remains to be deciphered for these sharks. Further research are necessary to fully demonstrate the evolutive conservation of luminescence control within etmopterids and dalatiids.

The question remains concerning bioluminescence in the largest luminous vertebrate; why does D. licha emit light ventrally to counterilluminate when it has few or no predators? Pinte et al. (2020), analyzed the swimming speed of several New Zealand deep-sea sharks, and found that D. licha possesses one of the slowest cruise swimming speeds ever measured in sharks. Conversely, this species is assumed to possess a high burst capability (Pinte et al., 2020). Stomach content analyses have revealed that this shark species hunts and eats etmopterids, which have a higher cruise swimming speed. Therefore, there are two hypotheses which might explain the ventral luminescence of this holobenthic species: luminescence might be used (i) to illuminate the ocean floor while searching and hunting for prey; or (ii) to stealthily approach toward prey, using counterillumination camouflage, before striking fast when close enough (Zintzen et al., 2011), allowing them to predate etmopterids. In both cases, the principle of counterillumination would have been distorted to serve as a predation tool instead of an avoidance mechanism, a hypothesis already proposed for the cookie cutter shark, I. brasiliensis (Widder, 1998). However, to validate such hypotheses for these dalatiid species, in vivo observations and behavioral studies are essential.