Sixth sense in the deep-sea: the electrosensory system in ghost shark Chimaera monstrosa

Animals that continually live in deep sea habitats face unique challenges and require adaptive specializations solutions in order to locate and identify food, predators, and conspecifics. The Ampullae of Lorenzini are specialized electroreceptors used by chondrichthyans for important biological functions. Ampullary organs of the ghost shark Chimaera monstrosa, a deep-sea species commonly captured as by-catch in the bottom trawl fishery, are here described for the first time using macroscopic, ultrastructural and histological approaches. The number of ampullary pores in C. monstrosa is about 700, distributed into the whole cephalic section of C. monstrosa, and organized in12 pore clusters and they are arranged into different configurations and form a distinct morphological pattern for this species, showing some anatomical peculiarities never described before in others cartilaginous fishes and may constitute an evolutionary adaptation of this ancient chondrichthyan species to the extreme environmental conditions of its deep sea niche.


Results
Ampullary pore distribution. The number of ampullary pores in our specimens of Chimaera monstrosa has an average of 702 (± 103). They are distributed in the entire cephalic portion. We recognized 12 types of pore clusters ( Fig. 2): nine of them are double (one on each side of the head), while the ethmoid (EAF), nasal (NP) and mandibular (MP) are single clusters, that are settled in the medio-ventral part of the head. The total number and average pore size of every cluster type is summarized in Table 1.
In particular: density-the pores are mainly concentrated in 3 clusters (suprarostral STP, supra-rostrolateral SRL, and subrostral SUB) which include almost the 65% of the whole ampullary pores and which have a medium or small pore diameter (Table 1 and Fig. 3); size-the range diameter is between 0.2 and 12 mm and the various clusters reveal difference also in the dimension of the pores summarized in Table 1 and Fig. 4.

Gross morphology.
Observations of the gross morphology show that in both sexes the electroreceptive organs of C. monstrosa are organized into clusters and consist of a canal and dilated ampullary portion that are filled with a gel and surrounded by connective tissue (Fig. 5a). The ampullary portion of the sensory organ is formed by eight or nine alveolar sensory chambers (Fig. 5b). The proximal end of the ampullary portion is connected to one afferent nerve (Fig. 5c).
Ultrastructure. SEM observations allowed us to analyze the superficial structure of the alveolar portion of the Ampullae of Lorenzini and the ampullary canal, revealing no differences among the sexes. The alveolar portion of the chambers converge into a common area (Fig. 6a) and are composed of a single columnar epithelium that contains supporting and sensory cells (Fig. 6b), However, the canals are distinguished from the alveolus by their composition of simple squamous epithelium (Fig. 6c).
Histology. Histological analysis confirms the general organization of the ampullary organs and the histological structure of the canal and the alveolar portion (organized in chambers), both surrounded by connective tissue composed mainly of collagen (Fig. 7a). There is a clear histological separation between the epithelia of these two parts divided by a medial zone of cuboidal-shape cells (Fig. 7b).  www.nature.com/scientificreports/ The canal of the Ampullae of Lorenzini is, in fact, covered by a simple squamous epithelium (Fig. 7c), also presenting mucous flattened cells showing Alcian positive gel protrusions (Fig. 7d); while, on the contrary, the ampullary portion appears to consist of a simple columnar epithelium, formed by supporting and sensory cells (Figs. 7e, 8). The central part of the ampullary portion is formed by the same columnar epithelium. Bundles of afferent nerve fibers reach the base of each alveolar portion, below the central part (Fig. 7f) and nerve terminals were observed at the base of the sensory cells only (Fig. 8).

Discussion
While teleosts have successfully colonised the deep sea, chondrichthyans are uncommon 5 . The species who live there rely on their sensory systems to gather information about their environment and to guide their behaviour 30 . Despite species of the order Holocephali regularly dwell in deep sea habitats, little is yet known about their  www.nature.com/scientificreports/ biology and ecology 7 , and the information regarding their sensory systems are very limited 10,31,32 . Here we describe for the first time the Ampullae of Lorenzini in the rabbit fish Chimaera monstrosa, providing morphological details and attempting to clarify their functional role. Although electroreception in shallow-water have been the subject of a number of studies 12 , which have started to investigate also the correlation between the morphology and the lifestyle 18 , this sensory system has been more rarely investigated in deep sea species 11 , even less so in the Holocephalans 10,21,22 . Compared to elasmobranchs 12 , and according to previous data on the other genus of chimaerids, Hydrolagus 21 , C. monstrosa shows a lower number of electroreceptive ampullary pores. This might be related to the feeding habits of this species: it is an opportunistic suction-feeder that actively searches the prey near the sea bottom 24,26,27,33 , and then uses its tooth plates to crush the food 29 . In fact, relatively few pores and low electrosensory resolution are mainly observed in cartilaginous fishes which feed with an indiscriminate suction or ram-feeding methods of prey capture 12 .
The location of pores determines the spatial representation and direction of the electrosensory field around the head 34 . In C. monstrosa, more pores are located on the anterior frontal sides of the head and ventrally near the mouth, while the pores with the largest diameter are located dorsoventral to the eyes. In this area we describe the supra-rostrolateral (SRL) cluster, a group of ampullary pores never described before in Holocephalans. Most of the pore clusters fields are located facing forward, which allows C. monstrosa to detect potential prey that swims along the bottom. The highest density of pores is found near the mouth because the primary function of electroreception is to detect prey and correctly position the subterminal mouth during the final strike on prey 35 . Therefore, pore number and location correlate with the foraging strategy [36][37][38] . The positions of the clusters with fewer but largest pores might allow the fish to detect possible the presence of possible predators during navigation and prey searching: C. monstrosa is actively preyed upon by larger deep-sea dogfishes 28,39 . This functional hypothesis is supported by the fact that chondrichthyans feeding on the sea bottom, like C. monstrosa, have limited routes of escape, being partially blocked by the substrate. They usually position their anti-predatory countermeasures, such as tail, fin spines and additional electrosensory pores, along the dorsal and posterior body surfaces 12 . www.nature.com/scientificreports/ Ampullary organs have been classified into three different types based on the size and the length of the canals: (1) macro-ampullae, (2) micro-ampullae, and (3) mini-ampullae 40 . C. monstrosa presents the so-called "macro-Ampullae", characterized by large, visible pores and long canals. As the others cartilaginous fishes, electroreceptors in C. monstrosa are embedded in a gelatinous substance and organized in clusters containing several functional sensory units, the ampullary organs. The gross anatomy of the ampullary portion it is not organized in alveoli usually arranged in a grape-like formation 12 , but it is constituted by numerous independent fingershaped sacs. This suggests a morphological pattern typical for C. monstrosa or for chimaerids 21 . This anatomical feature might enhance the sensory surface in relation of deep sea environment. It was previously observed that both the number of alveolar chambers and the overall size of the Ampullae significantly increase with depth, and species inhabiting deeper regions of the water column possess higher numbers of receptor structures, in order to improve feeding capability in a relatively prey-poor environment 15,16 . In addition, unlike what is observed in some derivate elasmobranchs, like the catshark Scyliorhinus canicula 41 , in C. monstrosa the size of ampullary organs and the number of finger-shaped sacs are the same in both sexes, and it might be related with the more primitive phylogenetic position of this species 42 .
The histological organization and innervation of the ampullary organs of C. monstrosa are similar to those already described for some coastal benthic elasmobranchs 20,[43][44][45] . The presence of very abundant Alcian positive gel both in the canal and in the ampullary portion suggests a very important role of this mucopolysaccharidic substance also for C. monstrosa. It has been demonstrated, in fact, that ampullary gel plays a fundamental role as a semiconductor with temperature-dependent conductivity and thermoelectric behavior, as a simple ionic conductor with the same electrical properties as the surrounding seawater, and as proton-conductor as well [46][47][48][49] . The extreme scarcity of light and the limited probability to find food typical of deep sea environments require adaptations of the chondrichthyan sensory systems 35 , which might also include enhanced efficiency in the transmission of stimuli mediate by this mucopolysaccharidic substance. Another histological characteristic is that the central part of the ampullary portion seems to be constituted by the same sensory epithelium as the alveolar finger-shaped sacs. This contrasts with what was observed in others chondrichthyans, where the central region appears to be formed by typical cells, called "brush cells" 19,40,44 .
Variations in alveolar morphology exist within the different ampullary groups, and they have been classified into five types, based on alveolar arrangement: "single-alveolate", "multi-alveolate", "branched alveolate", "centrum cap" and "club-shaped" 39 . The majority of sharks, including deep sea species, possess the "centrum cap" type 15 . The ampullary structure observed in C. monstrosa is similar to the so called "central cap" type already suggested for other species 44 . This organization is beneficial for detecting a variety of stimuli, therefore representing a possible evolutionary adaptive improvement of this phylogenetically conserved sensory system.
There are over 1300 extant species of cartilaginous fishes 50 that arose sometime during the group 400 million years evolution, with the main living families having first appeared sometime between the Permian and www.nature.com/scientificreports/ Jurassic Periods 51 . Currently, they occupy a very broad range of habitats worldwide, and play a key role in the functioning of ecosystems 52 . Under selective pressure, the chondrichthyan sensory systems, including olfaction and electroreception, have been differentially optimised to function in various habitats 33,53 : in particular, recent researches revealed how electrosensory system is adapted to suit the lifestyle or environmental niche of a species, also through discrete molecular and biophysical modifications 54 . The electroreception in C. monstrosa reveals here some anatomical peculiarities never described before in others cartilaginous fishes, which might constitute www.nature.com/scientificreports/ an evolutionary adaptation of this ancient chondrichthyan species to the extreme environmental conditions of its deep sea niche 18 . In this view, Holocephali could be a very interesting model for future comparisons to the study of chondrichthyan ampullary system because : (1) it is a small and phylogenetically isolated group constituted by different closely related species, occupying different deep sea habitats and with also different ecological features 8,54 , which makes them a very interesting group in which to examine the relationship between structure and function; (2) they reside mostly in deep-sea, a light-limited environment where non-visual senses, like electroreception is most likely to be important [55][56][57] .

Methods
All experimental protocols followed the recommendations of the Committee for the Animal Welfare of the Stazione Zoologica Anton Dohrn (https:// www. szn. it/ index. php/ en/ who-we-are/ organ izati on/ commi ttee-forthe-animal-welfa re). Moreover, all methods were carried out in accordance with relevant guidelines and regulations of the European Union. Adult specimens of Chimaera monstrosa were collected as by-catch in the Ligurian Sea (North-West Mediterranean Sea) and in the North-East Atlantic (Fig. 1) Ocean by professional bottom trawlers operating at the depth range of −600 and −800 m. They were brought already dead on board and the taxonomical identification was rapidly carried out on board 22 . After anesthesia with 0,01% MS-222 (tricainemethanesulfonate; Argent, Redmond, WA, USA; dilution 1:1000 in sea water), the rostral parts of their heads were excised immediately and fixed in 4% p-formaldehyde solution in phosphate buffered saline, pH 7.4 (PBS). Alternatively, some specimens were frozen at -30° C. In the laboratory, ampullary clusters were removed from the fixed rostral part of heads, isolated and prepared for macroscopical analysis or microscopical observations for both sexes; frozen specimens were used for ampullary pore distribution description.
Ampullary pore distribution. After defrosting, heads were severed from the specimens in the transverse plane at the first gill because no ampullary pores are located caudal to this position. Pores were counted for each cluster. The number and diameter of pores were measured by direct observation through a dissecting microscope or via ImageJ 58 for areas of high pore density. The average number of pores (+ /− SD) were calculated across all specimens. Pore diameter was measured in multiple images for each specimen and an average diameter (+ /-SD) was calculated. The nomenclature used to describe ampullary pore distribution following Didier 21 . The average pore diameters from different clusters were tested using ANOVA and Tukey HSD test. The graphical representations of the data were performed using R software 59  Histomorphological, histochemical and immunohistochemical methods. After washing in PBS, the samples were dehydrated, Paraplast embedded (Bioptica, Italy) and 6 µm sectioned. Dewaxed and rehydrated sections were alternatively haematoxylin-eosin stained or treated with Alcian blue pH 2,5-PAS method for carbohydrate detection. In order to highlight the presence of nerve endings, Na + /K + ATPase immunoreactivity was assessed by pre-treating the slides with Bovine Serum Albumine to block non-specific antibody binding. Next a mouse monoclonal antibodyspecific for the α subunit of chicken Na + /K + ATPase (α5, supernatant, 0.9 mg/ ml, DSHB, USA) and then an anti-mouse FITC conjugated antiserum (1:400 in PBS, Agilent DAKO, USA) was used. Negative controls were performed by omission of the primary antibody or by using the NS1 hybridoma culture supernatant (DSHB, USA) as primary antibody. Alternatively, the immunoreactivity was detected using the Agilent Dako EnVision + Kit with HRP in order to increase sensitivity and to minimize non-specific background staining. In this case the incubation with the primary antibody was preceded by the using the blocking reagents for endogenous peroxidase and nuclei were counterstained using Hematoxylin. Figure 8. Histological sections of ampullary electroreceptors of C. monstrosa. Na + /K + -ATPase immunohistochemisty (brown) and Hematoxylin. a) The sensory epithelium lining the chambers (asterisk) presents basally Na/K-ATPase-like immunoreactivity (arrows). b) The epithelium is made up by at least two types of cells, the supporting cells (arrowhead) and the sensory cells. The Na/K-ATPase-like immunoreactivity in the basal part of sensory cells could highlight the nerve terminals.