Ontogeny of the electric organ discharge and of the papillae of the electrocytes in the weakly electric fish Campylomormyrus rhynchophorus (Teleostei: Mormyridae)

The electric organ of the mormyrid weakly electric fish, Campylomormyrus rhynchophorus (Boulenger, 1898), undergoes changes in both the electric organ discharge (EOD) and the light and electron microscopic morphology as the fish mature from the juvenile to the adult form. Of particular interest was the appearance of papillae, surface specializations of the uninnervated anterior face of the electrocyte, which have been hypothesized to increase the duration of the EOD. In a 24.5 mm long juvenile the adult electric organ (EO) was not yet functional, and the electrocytes lacked papillae. A 40 mm long juvenile, which produced a short biphasic EOD of 1.3 ms duration, shows small papillae (average area 136 μm2). In contrast, the EOD of a 79 mm long juvenile was triphasic. The large increase in duration of the EOD to 23.2 ms was accompanied by a small change in size of the papillae (average area 159 μm2). Similarly, a 150 mm long adult produced a triphasic EOD of comparable duration to the younger stage (24.7 ms) but featured a prominent increase in size of the papillae (average area 402 μm2). Thus, there was no linear correlation between EOD duration and papillary size. The most prominent ultrastructural change was at the level of the myofilaments, which regularly extended into the papillae, only in the oldest specimen—probably serving a supporting function. Physiological mechanisms, like gene expression levels, as demonstrated in some Campylomormyrus species, might be more important concerning the duration of the EOD.

The mormyrid adult EO is composed of four colums of electrocytes, two dorsal and two ventral ones (Figure 1f). In each column between 21 and 91 disc-shaped electrocytes (number depending on species, Bass, 1986a) (Bass, Denizot, & Marchaterre, 1986;Denizot, Kirschbaum, Westby, & Tsuji, 1978;Denizot, Kirschbaum, Westby, & Tsuji, 1982;Szabo, 1960). Each electrocyte forms multiple finger-like processes, the stalks, which join together to a larger stalk, which receives the innervation by spinal electromotor neurons (Szabo, 1958). The stalks arise either at the posterior or the anterior face of the electrocyte. In some species the stalks penetrate the central axis of the electrocyte (Bass, 1986a(Bass, , 1986b. Each stalk and face of an electrocyte generates an action potential and the summed activity determines the shape of the EOD (Bennett & Grundfest, 1961).
Thus different geometries of the mormyrid electrocytes lead to different EOD types concerning polarity and number of phases (Bass, 1986b). The duration of the EOD, however, was correlated with specializations of the electrocyte's anterior face such as the degree of surface proliferations in the form of tubule-like structures (tubules) and calveoli (Bass et al., 1986;Luft, 1958), which are more pronounced in species with longer EODs (Bass et al., 1986). The increased electrocyte's surface leads to a longer EOD presumably by increasing the membrane capacitance (Bennett, 1970).
F I G U R E 1 Morphology of four developmental stages of Campylomormyrus rhynchophorus (left) (a-d) and the corresponding representative EOD (right). Amplitude values of the EOD are scaled to the same peak-to-peak amplitude. The different phases of the EOD are denoted as p1, first head positive phase; p2 second head positive phase; and n1, first head negative phase. Note differences in time scales. The EOD of the 24.5 mm long specimen (a) represents the activity of the larval EO (for details see text). The EODs of the three larger specimens are the activity of the adult EO. Note the change in the duration of the EOD, in particular between the 40 (b) and 79 mm long specimen (c). A medial sagittal section (e) through the dorsal and ventral columns (vc) of the EO of one side, shows the many individual electrocytes (e) with their main stalks (ms) arranged in parallel in the longitudinal axis of the fish. In f the organization of the EO is shown schematically in oblique view. In figure g two electrocytes are shown schematically in sagittal section with their main structures: stalks (s) at the posterior face (pf), papillae (pa) at the anterior face (af), centrally located myofibrils (my) and the connective tissue (ct) In the mormyrid genus Campylomormyrus there are species with very short EODs (ca. 200 μs duration) (e.g., C. compressirostris [Pellegrin, 1924]) and others with very long EODs (ca. 25 ms duration) (Feulner, Kirschbaum, Mamonekene, Ketmaier, & Tiedemann, 2007;Feulner, Kirschbaum, & Tiedemann, 2008;Lamanna, Kirschbaum, & Tiedemann, 2016). In species with the longest EODs of about 25 ms duration, as in C. numenius (Boulenger, 1898) (Paul et al., 2015) and C. rhynchophorus (Boulenger, 1898) (Nguyen et al., 2020), there are additional large foldings or evaginations on the anterior face of the electrocyte, so called papillae, which increase the surface area of the electrocyte even further. In one of the few ontogenetic studies of weakly electric fish, it was shown that C. rhynchophorus produces a short juvenile EOD which elongates by a factor of about 25 during ontogeny, a process which is accompanied by a proliferation of papillae on the anterior face of the electrocytes (Nguyen et al., 2020).
Since C. rhynchophorus can be bred in captivity (Nguyen et al., 2017), we choose this species for a longitudinal study of the EOD development and a detailed study of the fine structure of the papillae in order to find out, if there is indeed a correlation between the duration of the EOD and the size increase of the papillae.

| Fish
For reproduction, we used a pair (male of 37 cm, female of 31 cm total length) of C. rhynchophorus, which was originally purchased via a wholesale-dealer in Germany from the African region around Brazzaville from the Congo River. These fish were identified as to their systematic position by genetic analysis (Feulner et al., 2008). They were kept in tap water (temperature between 25 and 28 C, water conductivity 660 μS/cm, pH 7.4). Decrease of conductivity for several weeks down to 224 μS/cm (this led to a pH decrease to 6.5 as well) induced ripening of the gonads (see Nguyen et al., 2017). The ripe fish were injected with gonadotropinreleasing hormone (commercial name Ovaprim: GnRH + Domperidone; Western Chemical Inc., Ferndale, WA 98248) to induce ovulation, to enhance sperm release, and to perform artificial reproduction. One hundered and sixty-two eggs were obtained of which 62 hatched.

| Rearing
Eggs, free embryos and larvae were kept in plastic petri dishes (; = 8.6 cm, h = 1.5 cm) for about 40 days. Freshly hatched Artemia nauplii were offered twice a day and up to a size of about 40 mm, the juveniles were in addition fed small (ca. 2-4 mm) live chironomid larvae. Afterwards, the fish were transferred to small glass tanks (30 x 20 x 20 cm) with a density of about 15-20 larvae per tank. Larger fish were fed as the adults; the adult fish were fed twice daily with live Chaoborus (Corethra) and chironomid larvae or frozen larvae.

| Photography
The morphology of the fish was documented with a digital camera Canon EOS 350D and Canon EOS 100D in anesthetized specimens.

| Oscilloscope recordings
The EODs were recorded by using a Tektronix TDS 3012B digital phosphor oscilloscope (maximum sampling rate: 1.25 GS/s; 9 bit vertical resolution). For larvae with a size of 24.5 mm, we used a Tektronix ADA 400A differential preamplifier (variable gain; 0.1x, 1x, 10x, 100x and bandwidth 100 Hz-1 MHz) and a Tektronix TM 502A differential amplifier (upper bandwidth 3 kHz)-for juvenile and adult fish. All EODrecordings were made in water of a temperature of 25-26 ± 0.5 C and the water conductivity was of about 600-700 μS/cm. The EOD of the fish was recorded in plastic aquaria of different sizes: the recording tank measured 11 × 11 mm for the larvae, 7 × 5 cm for the small juveniles, and 13 × 15 cm for the largest fish. The water level was adjusted according to the size of the fish. The positive electrode was placed near the head of the fish and the negative near the tail. These different recording conditions led to amplitude values, which were not comparable between different size groups of fish (see e.g., Table 1).

| Histological and microscopical techniques
Four specimens (24.5; 40; 79 and 150 mm in total length; individuals 1-4) were euthanized with an overdose of tricaine methanesulfonate  prior to fixation. The tail and the caudal peduncle were cut off and fixed by immersion for 24 hr in 1% paraformaldehyde and 2.5% glutaraldehyd in 0.1 M phosphate buffer. For the semi-thin sections, tissues were postfixed in 1% osmium tetroxide for an additional 1-2 hr and then washed twice with 0.1 M phosphate buffer for half an hour. Then, the tissues were dehydrated in a graded alcohol series from 50 to 100% (30 min each, two changes). Following this, tissues were placed in propylenoxid for 30 min (twice) and then kept in a 50: 50 mixture of epoxy resin  Abbreviations: TD, total duration of EOD; TL, total length of the fish.
100% Durcupan and polymerized for 1 day at 40 C and for 2 days at 60 C (method described by Robinson et al., 1985). After trimming, serial sagittal sections, 1 μm thick, were prepared using an ultramicrotome. The sections were mounted on glass slides and stained with Richardson blue.
The semi-thin sections were studied by light microscopy using a Leica microscope DM4000B to determine the right areas for future ultra-thin sectioning. Ultra-thin sections, of about 70 nm, were cut with a diamond knife on the ultramicrotome in the sagittal plane; about 10 sections per specimen were cut. After cutting, the ultra-thin sections were collected on single slot grids coated with Formvar (Electron Microscopy Sciences) from the distilled water surface. Remaining water was absorbed with filter paper. Finally, the ultra-thin sections were contrasted with uranyl acetate (1%) and lead acetate. The micrographs were obtained on a transmission electron microscope (JEM 1011, manufacturer JEOL).

| Determination of the size of the papillae
For the determination of the number and size of the papillae, we choose a 200 μm long stretch of three different electrocytes in the three specimens-number 2 (40 mm total length), 3 (79 mm total length) and 4 (150 mm total length), respectively. The length of the papillae was measured from the central alignment of the myofibrils on and the width was determined as the maximum width of the papillae.
From these values, the area of the papillae was calculated.

| Ontogeny of morphology and EOD
A juvenile of 24.5 mm total length ( Figure 1a) lacks the uniform black pigmentation in the unpaired fins and the elongated snout typical of adult C. rhynchophorus. The EOD is a biphasic discharge with a large head positive peak (p1) and a short head negative phase (n1) (Figure 1a) of about 1 ms total duration. This is a typical EOD of the larval EO (Werneyer & Kramer, 2006;Westby & Kirschbaum, 1977) that is located in the deep lateral muscle (Kirschbaum, 1981). The treatment of the larval EO is beyond the scope of the present article. At this 24.5 mm stage, the adult EO, that is located in the caudal peduncle, is not yet functional. In a 40 mm long juvenile ( Figure 1b  3.2 | Ontogeny of the electrocytes at the light microscopical level 3.4 | Overview of the fine structure of the electrocyte at four developmental stages In Figure 4, the electrocytes of four developmental stages (24.5 to 150 mm total length) are shown at the electron microscopical level at low magnification. In all stages, centrally located myofibrils are present, whose striation becomes only apparent at higher magnification (see Figures 5 and 7). Furthermore, in addition to cell nuclei, numerous    (Bass et al., 1986;Luft, 1958), is of particular interest. These structures are only rarely observed in the material from the smallest specimen (24.5 mm fish, Figure 5a).
In contrast, the posterior face and the stalks of the electrocytes of the next developmental stage (specimen of 40 mm length, Figure 5b) feature tubules and calveoli that are depicted at higher magnification in the insert (Figure 5b11).
The tubules and calveoli remain prominent features in the electrocytes of older specimens (Figure 5c,d).

| Myofibrils in the electrocyte and in the striated muscle
The myofibrils run parallel to the center of the electrocyte in all stages investigated. The bundles, are often, but not regularly, oriented parallel to each other. In addition, numerous mitochondria are arranged near the bundles or in between in all age stages (Figures 4, 5, 8). In the electrocytes of the smallest specimen (total length 24.5 mm), the myofibrils are organized in thick, wide bundles, which sometimes branch into thin layers running parallel to each other (Figures 6a and   7a (Figure 8c,d). Tubules and calveoli are present in all developmental stages, and appeared less dense in the youngest fish than at later age.
The location of these structures is shown highly simplified in The main aim of our study was the investigation of a supposed correlation between the size of the papillae and the duration of the EOD.
In mormyrids, the degree of surface area proliferation of the electrocyte, in particular of its anterior face, is positively correlated with EOD duration (Bass et al., 1986). This feature is likely related to an increase in membrane capacitance, which will delay spike initiation (Bass et al., 1986;Bass & Volman, 1987;Bennett, 1970). We examined this possible correlation based on a developmental study in C. rhynchophorus.
There was no linear correlation between EOD duration and papillary size, when comparing the three developmental stages of 40, 79, and 150 mm total length (Figure 3). The large increase in duration of the EOD from 1.3 ms (40 mm long fish) to 23.2 ms (79 mm long fish) was accompanied by only a small increase in the size of the papillae (Table 2). In contrast, the large increase in papillary size from the 79 mm fish towards the 150 mm fish was only correlated to a slight increase in EOD duration (23.2 to 24.7 ms). Therefore, the size of the papillae can only to a minor degree be correlated with the duration of the EOD, which is in contrast to the suggestion of, for example, Schwartz, Pappas, and Bennett (1975) and Bass et al. (1986) in studies on weakly electric fish.
Apart from papillary size, the degree of membrane invaginations, such as density and size of tubule-like structures and calveoli, is expected to play a role in shaping EOD duration (Bass et al., 1986).
Qualitatively, there was a clear increase in density of these structures both in the anterior and posterior face in the smallest specimen with a functional EO (40 mm long fish) as compared with the earlier stage with a nonfunctional EO (24.5 mm long fish), but differences among the older stages were not obvious in our material. Since we did not perform serial sections and 3D reconstruction, it cannot be excluded that the observed relative small increase (present study) or relative drop in EOD duration (Nguyen et al., 2020) of the largest fish is related to subtle ultrastructural changes of the tubular network such as shorter or fewer tubules.
Additional (physiological) factors, such as a change in gene expression level, might be responsible for the EOD elongation during development. Gallant, Hopkins, and Deitcher (2012) studied gene expression in a mormyrid fish and found 28 genes which were differently expressed when muscle and EO tissues were compared. More specifically, Nagel, Kirschbaum, and Tiedemann (2017) studied voltage gated ion channel genes in C. compressirostris, which produces a very short EOD of about 164 μs duration compared with C. tshokwe with an EOD of 3.7 ms duration. These authors found an upregulation of all voltage gated ion channel genes in in the EO of C. tshokwe compared with that of C. compressirostris, but also in C. tshokwe when EO and muscle tissue were compared. Neither species develops papillae (Paul et al., 2015) as found in C. rhynchophorus. More interestingly, the mormyrid Gnathonemus petersii possesses papillae (El Arbani, 2017) but produces only a short EOD of about 250 μs duration .
A recent study on gene expression in several Paramormyrops operational taxonomic units (Losilla, Luecke, & Gallant, 2020) compared specimens with long and short duration EODs. They did not find evidence of large changes in expression of sodium channels between short duration and long duration species, but they found upregulation in short EOD samples of a voltage sensitive outwardly rectifying potassium channel and of a regulatory subunit of a shalltype voltage-gated potassium channel. The most prominent differences in gene expression between the EOD duration phenotypes comprise genes that code for cytoskeletal, sarcomeric, and lipid metabolism proteins. In a comparable study on gene expression levels between EO and muscle tissue and between the EOs of three mormyrid species (G. petersii, C. compressirostris, and C. tshokwe) Lamanna et al. (2016) conclude that in general terms the most diverging functional classes of genes across the species in the EO include metabolism of fatty acids and ion transport. Although, these investigations are first attempts to understand differences in gene expression between species of different EO characteristics it seems to emerge that each mormyrid genus has followed its own evolutionary pathway to change the duration of the EOD.

| Papillae in weakly and strongly electric fish
Papillae in weakly electric fish are usually present at the anterior face (Bass, 1986b;Bass et al., 1986;Schwartz et al., 1975). However, there are some exceptions as in Eigenmannia sp., where papillae are found at the posterior face of the electrocyte; or they occur on both faces as in Sternopygus sp. (Schwartz et al., 1975). In strongly electric fish, papillae either are absent (Luft, 1956) or are found at the innervated posterior or at both, the anterior and posterior faces of the electrocyte (Luft, 1956;Machado, de Souza, Benchimol, Attias, & Porter, 1980;Mermelsstein et al., 1997;Stuart & Kamp, 1934). Raia batis possesses papillae at the posterior face (Stuart & Kamp, 1934). In the little skate Leucoraja erinacea papillae are also located at the posterior face (Morson & Morrissey, 2007). In Electrophorus electricus, large numbers of fine papillae with minute rodlets are found at both faces. However, the papillae at the anterior face are much larger and thicker than the papillae at the posterior face (Luft, 1956).
The function of the papillae in strongly electric fish is not quite clear (Stuart & Kamp, 1934). Papillae in weakly electric fish are considered to be responsible for the elongation of the EOD duration (Bass et al., 1986;Bennett, 1971;Nguyen et al., 2020;Schwartz et al., 1975); this, however, is in contradiction to the results of this article.

| Surface invaginations in weakly and strongly electric fish
Increase in surface area is not only obtained by surface evaginations, like papillae, but also by surface invaginations, like tubules and calveoli. They have been described in many species of both weakly and strongly electric fish (Bass, 1986b;Bass et al., 1986;Bennett, 1971;Bruns, 1971;Denizot et al., 1978;Luft, 1956Luft, , 1958Machado et al., 1980;Machado, de Souza, Cotta-Pereira, & Oliveira-Castro, 1976;Mermelsstein et al., 1997;Nguyen et al., 2020;Schwartz et al., 1975;Schwassmann, Assunção, & Kirschbaum, 2014). They are also found at both faces of the electrocyte in C. rhynchophorus and increase in density during ontogeny. They increase the surface area of the electrocyte and thus, probably, also contribute to the elongation of the duration of the EOD.

| The fine structure of the stalk
In mormyrids the stalks, in general, originate at the posterior face of the electrocyte and may, in some species, traverse the electrocyte's body via so called penetrations; in some species the stalks stay on the anterior face (single penetration) or go back to the posterior face (double penetrations) (Bass, 1986b). This geometry is accompanied by different EOD shapes. In C. rhynchophorus the stalks arise at the posterior face and do not penetrate the electrocyte. The emerging EOD is a biphasic EOD (Figure 1). During ontogeny, the stalks increase in size, the number of mitochondria increase as well as that of tubules and calveoli (Figure 4). Single nuclei are found in the stalk. A similar fine structure has been described in several Gnathonemus species (Bruns, 1971;Schwartz et al., 1975) and in Pollimyrus isidori (Denizot et al., 1982). The innervation is found at the tip of the main stalk (Bennett, 1971;Denizot et al., 1982;Szabo, 1958). We have not studied details of the innervation in C. rhynchophorus.

| Myofibrils in the electrocyte and in striated muscle
The central part of the electrocytes of mormyrid fish consists of myofibrils (e.g., Bass, 1986b;Bass et al., 1986;Bruns, 1971;Paul et al., 2015;Schwartz et al., 1975). They are also present in the electrocytes of C. rhynchophorus. In contrast to the authors mentioned above, we were able to describe the ontogeny of these myofibrils: they are already found in electrocytes, which are not yet functional ( Figure 7a,b). They increase in size and density during ontogeny ( Figure 7a-f). We, therefore, believe that these myofibrils are not just remnants of the myogenic origin of the electrocytes (Denizot et al., 1982), instead they are structural elements stabilizing the electrocyte. This view is supported by the fact that the myofibrils extend into the large papillae in the course of ontogeny (Figure 2). This has never been described for the papillae found in other mormyrid fish (Bass, 1986b;Bass et al., 1986;Bruns, 1971;Schwartz et al., 1975), except for C. numenius (Paul et al., 2015), the sister species of C. rhynchophorus (Lamanna et al., 2016).
The fine structure of the myofibrils of the electrocytes (Figure 7) indicates that they are not functional myofibrils as in striated muscle,

| CONCLUSION
This study explores for the first time the ontogeny of the fine structure of the electrocytes in a mormyrid fish with an EOD of very long duration. It emerges that there is only a very weak correlation between the duration of the EOD and fine structural features like papillae, tubules, and calveoli in contrast to earlier assumptions. The ontogenetic investigation also suggests that the myofibrils which are a remnant of the myogenic origin of the electrocytes serve a structural/stabilizing function. It would be interesting to perform a similar study with the sister species of C. rhynchophorus, C. numenius, which also produces as adult an EOD of very long duration and possesses papillae.

ACKNOWLEDGMENTS
This study received financial support from the Friedrich Naumann Foundation for Freedom in the course of a scholarship for international PhD students. We would like to thank H. Höft for technical assistance and S. Abelt for the production of the histological preparations and for her support and cooperation during the preparation of semi-thin sections. We thank Prof. Dr. Baumann for the possibility to work in his laboratory during the preparation of semi-thin sections.
Thanks go to C. Prietzel and S. Rüstig for their technical support during our investigation on the electron microscope and their interest in our work. We thank Prof. Dr. Krause for his willingness to act as first supervisor of the PhD project of Yevheniia Korniienko. Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST
There is no conflict of interest to declare.

AUTHOR CONTRIBUTIONS
Yevheniia Korniienko made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data/ drafted the manuscript/agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved/gave final approval of the version to be published. Ralph Tiedemann revised the manuscript critically for important intellectual content/ gave final approval of the version to be published. Marianne Vater made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data/she revised the manuscript critically for important intellectual content/gave final approval of the version to be published. Frank Kirschbaum made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data/he revised the manuscript critically for important intellectual content/gave final approval of the version to be published.
He agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

ETHICS APPROVAL
This research was performed in accordance with the German Protection of Animal Act and with Animal Welfare Act and its subsequent amendments. All experiments during this work were conducted with efforts to minimize the number of utilized animals and their suffering.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/cne.25003.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.