ABSTRACT
Enigmatic ctenophores are descendants of one of the earliest branching metazoan lineage. Their nervous systems are equally elusive. The lack of convenient neurogenic molecules and neurotransmitters suggests an extensive parallel evolution and independent origins of neurons and synapses. However, the field is logged behind due to the lack of microanatomical data about the neuro-muscular systems in this group of animals. Here, using immunohistochemistry and scanning electron microscopy, we describe the organization of both muscular and nervous systems in the sea gooseberry, Pleurobrachia bachei, from North Pacific. The diffused neural system of Pleurobrachia consists of two subsystems: the subepithelial neural network and the mesogleal net with about 5000-7000 neurons combined. Our data revealed the unprecedented complexity of neuromuscular organization in this basal metazoan lineage. The anatomical diversity of cell types includes at least nine broad categories of neurons, five families of surface receptors and more than two dozen types of muscle cells as well as regional concentrations of neuronal elements to support ctenophore feeding, complex swimming, escape and prey capture behaviors. In summary, we recognize more than 80 total morphological cell types. Thus, in terms of cell type specification and diversity, ctenophores significantly exceed what we currently know about other prebilaterian groups (placozoan, sponges, and cnidarians), and some basal bilaterians.
1 INTRODUCTION
Ctenophores are enigmatic basal metazoans (Kozloff, 1990; Brusca and Brusca, 2003; Nielsen, 2012), and their nervous systems are equally elusive even in term of their anatomy and functions (Tamm, 1982; Hernandez-Nicaise, 1991; Moroz, 2015). Due to the fragile nature of these jelly-like marine organisms, it is difficult to work with comb jellies outside of their marine habitats. The second challenge is a unique organization (Hernandez-Nicaise, 1991) and remarkably different molecular makeup of ctenophore nervous systems (Moroz, 2015), which significantly limit the use of conventional protocols to label neurons.
These challenges lead to numerous controversies about the early evolution of animals in general, and the origins of neurons and muscles in particular (Moroz, 2014). The exact phylogenetic position of Ctenophora has been extensively debated over the last decade (Nielsen, 2012; Ryan et al., 2013; Moroz et al., 2014; Borowiec et al., 2015; Dunn et al., 2015; Whelan et al., 2015a; Whelan et al., 2015b; Halanych et al., 2016; Telford et al., 2016; Alamaru et al., 2017; Arcila et al., 2017; Cavalier-Smith, 2017; Feuda et al., 2017; King and Rokas, 2017; Shen et al., 2017; Simion et al., 2017; Whelan et al., 2017). However, the most recent models of evolution and interdisciplinary data strongly suggest that Ctenophora is a sister group to all extant Metazoa including nerveless sponges, which are morphologically simpler than comb jellies (Moroz et al., 2014; Arcila et al., 2017; Shen et al., 2017; Whelan et al., 2017).
The combination of genomic, physiological, neurochemical, pharmacological, proteomic and metabolomic data indicate that neural systems and synapses in ctenophores evolved independently from the rest of animals (Moroz et al., 2014; Moroz, 2015; Moroz and Kohn, 2015; 2016). The muscles and mesoderm might also evolve independently in ctenophores and cnidarians/bilaterians (Steinmetz et al., 2012; Moroz et al., 2014). Molecular analyses further confirmed extensive parallel evolution and ctenophore-lineage-specific diversification of ion channels, gap junction proteins, transmitter receptors, neuropeptides, RNA editing, RNA modifications and signaling pathways (Moroz and Kohn, 2016).
But the field is logged behind due to the lack of microanatomical data about the structure of neuro-muscular systems in ctenophores. Here, using immunohistochemistry and scanning electron microscopy, we expand earlier developmental (Norekian and Moroz, 2016) and comparative studies (Jager et al., 2011), and characterize the organization of both muscular and nervous systems in the sea gooseberry, Pleurobrachia bachei, from North Pacific. Our data revealed the unprecedented complexity of neuro-muscular organization in this basal metazoan lineage with diverse types of neurons and muscles as well as regional concentrations of neuronal elements to support ctenophore feeding, complex swimming, escape and prey capture behaviors.
2 MATERIALS AND METHODS
2.1 Animals
Adult specimens of Pleurobrachia bachei were collected from the breakwater and held in 1-gallon glass jars in the large tanks with constantly circulating seawater at 10° C. Experiments were carried out at Friday Harbor Laboratories, the University of Washington in the spring-summer seasons of 2012-2017.
2.2 Scanning Electron Microscopy (SEM)
The details of the protocol have been described elsewhere (Norekian and Moroz, 2016). Briefly, adult animals were fixed in 2.5% glutaraldehyde in 0.2 M phosphate-buffered saline (pH=7.6) for 4 hours at room temperature and washed for 2 hours in 2.5% sodium bicarbonate. Large animals, more than 1 cm in diameter, were then dissected into a few smaller pieces, while small animals were processed as whole mounts. For secondary fixation, we used 2% osmium tetroxide in 1.25% Sodium Bicarbonate for 3 hours at room temperature. Tissues were then rinsed several times with distilled water, dehydrated in ethanol and placed in Samdri-790 (Tousimis Research Corporation) for Critical point drying. After the drying process, the tissues were placed on the holding platforms and processed for metal coating on Sputter Coater (SPI Sputter). SEM analyses and photographs were done on NeoScope JCM-5000 microscope (JEOL Ltd., Tokyo, Japan).
2.3 Immunocytochemistry and phalloidin staining
Adult Pleurobrachia were fixed overnight in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) at +4-5° C and washed for 2-6 hours in PBS. The fixed animals were then dissected to get better access to specific organs and processed as described for embryonic and larval Pleurobrachia (Norekian and Moroz, 2016), with minor modifications. To label the muscle fibers, we used well-known marker phalloidin (Alexa Fluor 488 and Alexa Fluor 594 phalloidins from Molecular Probes), which binds to F-actin (Wulf et al., 1979) The tissue was incubated in phalloidin solution in PBS for 4 to 8 hours at a final dilution 1:80 and then washed in several PBS rinses for 6 hours.
For immunocytochemistry, adult animals were fixed the same way in 4% paraformaldehyde in PBS for 8-12 hrs and washed in PBS several times for 6-12 hrs. Sometimes, we used smaller, 2-6 mm, pieces for better access to different areas and better antibody penetration. Tissues were pre-incubated overnight in a blocking solution of 6% goat serum in PBS and then incubated for 48 hours at +4-5° C in the primary antibodies diluted in 6% of the goat serum at a final dilution 1:40. The rat monoclonal anti-tubulin antibody (AbD Serotec Cat# MCA77G, RRID: AB_325003) recognizes the alpha subunit of tubulin, specifically binding tyrosylated tubulin (Wehland and Willingham, 1983; Wehland et al., 1983). Following a series of PBS washes for total 6 hours, the tissues were incubated for 12 hours in secondary antibodies of one of two types of goat anti-rat IgG antibodies: Alexa Fluor 488 conjugated (Molecular Probes, Invitrogen, Cat# A11006, RRID: AB_141373) and goat anti-rat IgG antibody Alexa Fluor 594 conjugated (Molecular Probes, Invitrogen, Cat# A11007, RRID: AB_141374), at a final dilution 1:20. To stain nuclei, the tissues were incubated for 8 hours in SYTOX Green nucleic acid stain from Life Technologies (Cat# S7020) or was mounted in VECTASHIELD Hard-Set Mounting Medium with DAPI (Cat# H-1500).
After washing in PBS, the preparations were mounted in Fluorescent Mounting Media(KPL) on glass microscope slides to be viewed and photographed using a Nikon Research Microscope Eclipse E800 with Epi-fluorescence using standard TRITC and FITC filters, BioRad (Radiance 2000 system) Laser Scanning confocal microscope or Nikon C1 Laser Scanning confocal microscope. Multiple optic sections were taken through the depth of the samples for 3D reconstructions. To test for the specificity of immunostaining either the primary or the secondary antibody were omitted from the procedure. In both cases, no labeling was detected.
2.4 Antibody specificity
Rat monoclonal anti-tyrosinated alpha-tubulin antibody is raised against yeast tubulin, clone YL1/2, isotype IgG2a (Serotec Cat # MCA77G; RRID: AB_325003). The antibody is routinely tested in ELISA on tubulin, and ctenophore Pleurobrachia is listed in species reactivity (manufacturer’s technical information). The epitope recognized by this antibody appears to be a linear sequence requiring an aromatic residue at the C terminus, with the two adjacent amino acids being negatively charged, represented by Glu-Glu-Tyr in Tyr-Tubulin (manufacturer’s information). As reported by Wehland et al. (Wehland and Willingham, 1983; Wehland et al., 1983)-this rat monoclonal antibody “reacts specifically with the tyrosylated form of brain alpha-tubulin from different species.” They showed that “YL 1/2 reacts with the synthetic peptide Gly-(Glu)3-Gly-(Glu)2-Tyr, corresponding to the carboxyterminal amino acid sequence of tyrosylated alpha-tubulin, but does not react with Gly-(Glu)3-Gly-(Glu)2, the constituent peptide of detyrosinated alpha-tubulin”. The epitope recognized by this antibody has been extensively studied including details about antibody specificity and relevant assays (Wehland and Willingham, 1983; Wehland et al., 1983). Equally important, this specific monoclonal antibody has been used before in two different species of both adult and larval Pleurobrachia: the previously obtained staining patterns are very similar to our experiments (Jager et al., 2011; Moroz et al., 2014; Norekian and Moroz, 2016).
3 RESULTS
3.1 Brief Introduction to the Pleurobrachia anatomy
The pelagic Pleurobrachia bachei has the biradial symmetrical body-plan (Kozloff, 1990; Hernandez-Nicaise, 1991), characteristic for the phylum including the morphological features of the canonical cydippid larvae (Fig. 1A). Two symmetrical polar fields are connected to the aboral organ forming the aboral pole of the animal, which is opposite to the mouth or the oral pole. Here, we use the term of the aboral organ to distinguish it from the non-homologous apical organ in bilaterian larvae (Nielsen, 2012).
Together with the mouth (Fig. 1D), the polar fields are making the first symmetry plane of the body, perpendicular to the tentacle plane. The size of polar fields is age-dependent, from less than 1 mm in juveniles to very long and narrow loops in adult animals. A bright single statolith (Tamm, 2015) is visible in the center of the aboral organ (Fig. 1B). The aboral organ performs multiple integrative functions and has a complex ultrastructural organization (Aronova, 1974; Tamm, 1982; Hernandez-Nicaise, 1991; Lowe, 1997; Tamm, 2016b). Eight thin and long ciliated furrows (Fig. 1B) connect the aboral organ with eight swim cilia rows (Fig. 1A, C), which represent the main means of locomotion in the ctenophores. A pair of tentacles (Fig. 1A, E), with sticky colloblasts, is used to collect the food and bring it to the mouth. Upon complete retraction, tentacles withdraw into a pair of tentacle pockets, which extend to the center of the body (Fig. 1A). The mouth (Fig. 1D) opens into the muscular pharynx and then short stomach chamber, which then connects to a system of gastrovascular canals.
3.2 The Aboral Organ and Polar Fields are complex integrative structures with multiple neural elements
Scanning electron microscopy (SEM)
Since ctenophores are composed of >95% of water, SEM imaging is very challenging because the dehydration and fixation cause significant tissue deformation. Nevertheless, we successfully adopted the SEM protocol for both juvenile and adult Pleurobrachia bachei (Norekian and Moroz, 2016) to reveal the microscopic anatomy for most of the tissues including the aboral organ and polar fields (Fig. 2). The SEM sagittal section shows a dome structure over the aboral organ, which consists of long cilia protecting the statolith (Fig. 2A, B).
The aboral end of the digestive tract has two short branches on both sides of the aboral organ (Fig. 2A). Each of these branches opens as an anal pore (a small bulge on the SEM image in the Fig. 2A, C). In fact, the anal pores act as the functional anus similarly to the majority of bilaterian animals (Presnell et al., 2016; Tamm, 2016a).
The horizontal SEM view shows that the polar fields are formed by the short ciliated bands on their outside edges (Fig. 2C, D). The densely ciliated furrows connect the aboral organ with the comb rows and might represent conductive paths (Tamm, 1982; Tamm, 2014).
Immunoreactivity for neurons and cilia
As was shown in previous studies (Jager et al., 2011; Moroz et al., 2014; Norekian and Moroz, 2016), the tubulin antibodies are useful tools for the identification of neurons in ctenophores. In adult Pleurobrachia, the tubulin antibody staining always produced a very low background and a good signal/noise ratio in most parts of the body. Specific labeling occurred in neural elements and different cilia associated with such structures as the aboral organ, two polar fields, eight ciliated furrows and two tentacular nerves approaching the aboral organ (Fig. 3A). The tubulin immunoreactivity (IR) also revealed a diffused neural network covering the entire surface of the body.
In the sagittal view, the aboral organ resembles a horseshoe. The aboral canal of the gastrovascular system branches into two short anal canals and subsequently two openings near the aboral organ (Fig. 3A, B, and 6A). These anal pores can be seen in both sagittal and horizontal planes of the images (Fig. 3A, B, and 6A). Inside the aboral organ, tubulin IR staining revealed four elongated cell-like structures (Fig. 3C, D), which were located under four balances that support the statolith (Tamm, 1982; Tamm, 2015; 2016b). The statolith itself was often washed away during the incubation and washing in numerous solutions.
The bright tubulin-specific fluorescence of aboral organ observed in all preparations (e.g., Fig. 4A, C) was related to the large number of very tightly packed immunoreactive small cells covering the entire area (Fig. 4C, D). The size of these cells varied between 1 and 4 μm in diameter. Some of them revealed elongated processes suggesting their neuron-like nature (Table 1).
Diffused neural network located within the body wall formed two deep foldings on both sides of the aboral organ (Fig. 4A, B, and Video-1 in Supplements). These two network foldings neural had a wide web of direct connections to the aboral organ (Fig. 4B). The areas, where diffused neural network joined the aboral organ, contained two narrow structures stained by phalloidin (F-actin marker), which included groups of small cells and fibers (Fig. 4A, B).
Also, phalloidin labeled four cylindrical structures at the base of each polar field (Fig. 5A, B). These “cylinders” consisted of many parallel, very thin, phalloidin-labeled filaments, and looked like ‘anchors’ to which polar field bands were attached. In the center of the aboral organ, phalloidin labeled four groups of muscle filaments related to fours balancers (Fig. 5B).
Phalloidin is known to be a good marker of muscle cells (Wulf et al., 1979). It stained a group of very thin and short fibers connecting the aboral organ to the walls of two short anal canals in Pleurobrachia (Fig. 6B). There were also several long individual muscle filaments under the aboral organ and around the anal canals (Fig. 6C). Finally, several thick muscle fibers were attached to the polar fields; these fibers presumably mediate the defensive withdrawal of polar fields inside the body. SYTOX green staining revealed multiple nuclei in each muscle filament (Fig. 6D).
Neural cells are located in the polar fields
The length of the polar fields always directly corresponded to the age and the size of the animals (Fig. 7). In very young animals, the polar fields are very short and small, and sometimes cannot even be seen on the surface – they are inside the body-wall invagination where the aboral organ is located (Norekian and Moroz, 2016). After hatching, the polar fields grow in length to form an elongated figure eight structure.
The tubulin antibodies intensely stained two bands of numerous short cilia at outside edges of the polar fields (Fig. 7 and 8A, B), efficiently ‘masking’ neural elements inside the structure. However, using the confocal microscopy for isolating optical sections, we identified densely packed neurons through the entire length of the polar fields (Fig. 8B, C; Table 1). These small cells (less than 3 μm in diameter) had short neurites (Fig. 8D) and formed a narrow band next to the outside edge of the polar fields. Many neuron-like cells were bipolar, while others had three or four elongated processes, primarily oriented to the inner parts of the polar fields.
The epithelial polygonal neural network had numerous connections with the polar fields (Fig. 8C). Nevertheless, we could not visualize a direct contact between the polygonal network and neurons inside the polar field because of the brightly stained cilia on edge.
3.3 Two Neural Subsystems in Pleurobrachia bachei
Tubulin IR revealed two major parts of the diffused nervous system in Pleurobrachia: (i) the surface, subepithelial polygonal neural network, and (ii) a large diffuse meshwork of neurons and fibers in the mesoglea, which spread over the internal gelatinous part of the body.
3.3.1 The Subepithelial Neural Net
The subepithelial neural network is a large polygonal mesh of neurites covering the entire body surface under the epithelium, as well as at the surface of the tentacle pockets. This meshwork consisted of individual polygonal units of different sizes (between 20 and 150 μm) and shapes: many units had four ‘corners’, but there were also pentagonal (five corners), hexagonal (six corners), heptagonal (seven corners), and sometimes octagonal (eight corners) configurations (Fig. 9A, B). Higher resolution microscopy identified many tightly packed neuronal processes (Fig. 9B). and direct connections between the subepithelial network and other surface structures such as the polar fields and ciliated furrows (Fig. 9C, D).
The subepithelial network is formed by 4-8 μm neurons with multiple processes, which contribute to the structure of the polygonal units (Fig. 10A, B; Table 1). All neurons had a single nucleus (Fig. 10A, B), but DAPI staining also revealed numerous nuclei inside the meshwork itself suggesting that neuronal somata with little cytoplasm area were tightly packed (Fig. 10C). Some of these cells were located in the nods of a mesh and appeared to be tripolar, while others were bipolar (Table 1).
Our data suggest that the subepithelial network controls contractions of the surface muscles and body wall movements. Indeed, we identified a web of very fine neuronal processes running in parallel to subepithelial muscles and connected to each other slightly below the focal plane of the principal network (Fig. 10G, H). Moreover, higher resolution double labeling confirmed that several individual neural processes are located at the surface of single muscle fibers over their entire length (Fig. 10I), and we traced these processes to some neurons located between polygonal neurites of the subepithelial network.
In addition to the small neurons with well-defined neurites, we identified larger secretory-like cells (5-10 μm in diameter) of a round shape and without processes - glandular or gland cells (Hernandez-Nicaise, 1991); Fig. 10C, D). These secretory cells were distributed randomly or grouped in clusters of 3 to 8 cells. We think these cells were captured at different secretory stages. At their presumed last secretion phase (e.g., after emptying their content), the cells looked like large aperture-shaped structures with a large circular dark center and approximately 710 μm in diameter (Fig. 10D, E).
When SYTOX green was used to stain the cell nuclei instead of DAPI, while tubulin antibody was tagged with a red fluorescent label, the difference between neurons with processes and round-shaped glandular cells became even more obvious. Glandular cells appeared to be orange or yellow, unlike neurons which remained red, apparently as a result of co-localization of tubulin IR and diffuse SYTOX green staining (Fig. 10E). SYTOX green labels not only DNA but also detects a high concentration of RNA in cell bodies, suggesting that secretory cells have much higher RNA concentration (before they empty their content). In some preparations, when we scanned the upper epithelial layer, a slightly auto-fluorescent cloud was observed over the clusters of ‘empty’ secretory cells, thus suggesting that they released their content during fixation, and further supporting our hypothesis that they are indeed secretory cells (Fig. 10F).
3.3.2 Identification of putative sensory cells in the intergument
We identified five types of putative sensory cells in the skin of Pleurobrachia: two types were integrated parts of the subepithelial neural network with predominant tubulin IR, and three types represented the epithelial receptors (Fig. 11 and Table 2). SEM observations also revealed several different classes of sensory cilia on the body surface of Pleurobrachia, with some of them corresponding to these receptor cell types (Table 2 and Fig. 12).
The first type of receptors has a relatively thick and short cilium, which brightly stained by tubulin antibody (Fig. 11A, B, C; very rarely, we observed two or three cilia). A wide base root of the ciliated structure was always labeled by phalloidin suggesting the presence of F-actin. The morphology of these cells was very similar to most neurons described in the previous section (e.g., Fig. 11H). They had cell bodies about 5 μm in diameter and usually two or three processes fully integrated within the subepithelial neural net (Fig. 11A, B, C, I; though sometimes there was one or more than three). Also, we observed with SEM many short surface cilia (Fig. 12F). Some of them could belong to the first receptor type described here. However, there was no sufficient information to establish an irrefutable link between them.
The second receptor type included cells located deeper in the body wall, just under the layer of the subepidermal neural network (presumably proprioreceptors). These cells produced three thin neurites (labeled by tubulin antibody), which spread in the radial directions (Fig. 11D, E) and fused with the rest of the network. Very often these receptors were located next to the joints points (‘corners,’ Fig. 11D), and more rarely, receptors had only two visible processes instead of three. Interestingly, the cell body itself of this receptor type was labeled neither by phalloidin nor tubulin antibody. However, a few short phalloidin-stained structures were always located on the top of their DAPI stained nucleus (Fig. 11D, E).
We recognized the third type of receptors as 5 to 7 μm cells with a compact group of 2-7 cilia labeled by phalloidin only (Fig. 11F). The length of individual cilia in the group could vary, but they all had the same thickness and appearance under SEM (Fig. 12A, B, C). These cilia were 3 to 8 μm long and always originated from one central region close to the nucleus. The base where all cilia from the group were joining each other was also revealed by SEM (Fig. 12C).
The fourth type of receptors had a long thick single cilium (up to 7-10 μm) that was labeled by phalloidin only (Fig. 11H, I and Fig.12D).
In contrast, the fifth type of receptors also had a single long cilium, which was labeled by tubulin antibody only (Fig.11G), but with a small base labeled by phalloidin. The very top of the cilium was always curled and looked very similar in SEM preparations (Fig. 12E). The last two receptor types were less abundant compared to others.
3.3.3 Tentacles and their neuro-muscular system
The main food capturing organs in Pleurobrachia is the highly mobile tentacles with dozens of tentilla (or tentillae, “little tentacles”) (Figs. 1, 13, 14A). The base of each contractile tentacle is located inside the tentacle pocket, which itself is attached to the posterior pharynx forming the single functional unit (Fig. 14A). Both tentacles and tentilla were completely covered with colloblasts (Fig. 13C, D), glue-like cells that capture prey. As a result, the tentacles had a complex neuro-muscular organization with sensory capabilities and innervation.
We identified four major nerves in Pleurobrachia - two on each side of the animal. All of them are associated with the tentacles and, therefore, were named as the tentacular nerves (Hernandez-Nicaise, 1991). The base of each contractile tentacle is located inside the tentacle pocket, which itself is attached to the posterior pharynx forming the single functional unit (Fig. 13A).
Two symmetrical aboral tentacular nerves can connect tentacles and the aboral complex (Fig. 3A, 7C, 14B, C), where nerves began to branch, and gradually merged with the subepithelial network (Fig. 7C). On the opposite side, after crossing half of the body, the tentacle nerves ran into the ipsilateral tentacle pocket (Fig. 14B, C, D), and eventually innervated the base of each tentacle (Fig. 14B, C, D and Video-2 in Supplements).
Two oral tentacular nerves originated from the surface network near the mouth. These nerves ran between comb rows to the ipsilateral tentacle pocket, where they branched, and ultimately merged with the subepithelial neural network at the opening of the pocket (Fig. 15A, B, C). Unlike the aboral tentacular nerves, the oral tentacular nerves did not run through the entire tentacle pocket (Fig. 15A, B, C).
Each tentacular nerve was localized within the same layer as the subepithelial network and had numerous interconnections with it through its entire length. Thus, the tentacular nerves appeared to be an integrated part of the entire surface network, possibly collecting information from the mouth, the aboral organ, body surface and tentacles to control the feeding, locomotion and defense responses, which are tightly correlated with tentacle movements.
The tentacle pockets themselves contained a dense subepithelial network (Fig. 15D). In several SEM preparations (n=7), we identified neurons formed different polygonal units with clearly visible neural cell bodies and their processes (Fig. 16).
Each tentacle itself contained a central nerve (Fig. 17B) with numerous neuronal processes, which ran parallel to the muscles through the entire length (Fig. 17C). In addition to numerous colloblasts, the surface of the tentacles was covered with ~5 μm receptors with a single cilium; each receptor sent a neural processes to the central nerve (Fig. 17C, D; Table 2).
Phalloidin staining confirmed the muscular natures of tentacles, although tentilla had different types of double strain muscles (Fig. 17A).
3.3.4 Mesogleal neural system
The entire space between the skin and inner gastroderm in Pleurobrachia is occupied by mesoglea, a jelly-like, fully transparent tissue, characteristic of the phylum. Both phalloidin and tubulin antibody labeled very diverse systems of muscles and neurons in the mesoglea; many of these cells had processes, which crossed the mesoglea in all directions (Fig. 18). The vast system of mesogleal neurons represented the second major part of the nervous system in Pleurobrachia.
There were approximately 2 to 3 thousand mesogleal neurons in the entire adult Pleurobrachia according to our estimates (Fig. 19A). All cell bodies had similar sizes between 3 μm and 7 μm in diameter (for the 1-1.5 cm animals) and could be separated into three types (Table 1).
First, and the most numerous type was represented by cells with a thin processes of similar length projecting in all directions, which gave them a star-like shape (Fig. 19B). The second type of cells was represented by bipolar neurons with two processes on opposite sides (Fig. 19B). The third type included cells with long thin processes crossing the large areas of the mesogleal space (Fig. 20D, E, F).
Often, the neuronal-like processes in mesoglea were attached to the surface structures such as the aboral organ and polar fields, the mouth area and regions between the comb rows (Fig. 20) as well as tentacle pockets (Fig. 20B, C). The terminals always branched at the very end significantly increasing the contact area (Fig. 20C and 21A).
We detected connections of mesogleal neuronal processes both to the surface polygonal neural network (Fig. 22A) and to the subepithelial nerve net in the tentacle pockets (Fig. 22B). Some of the neural-like terminals, however, ended shortly before making visible contact with the network (Fig. 22B) and the branching of the filaments occurred in the center of the mesoglea (Fig. 21B and 18C).
In selected SEM experiments, the Pleurobrachia body was intentionally broke open after SEM drying process to get access to the inside organs. Cells and their processes (Figs. 23, 24) were preserved very well, while the mesogleal gelatinous tissue was washed away during the procedure. For example, there were many processes that at the point of contact with the pharynx surface had a different morphology than the muscle fibers (Fig. 23) and looked more like tubulin IR neurites described above (Fig. 24A, B). Unlike muscle fibers, which were firmly attached to the surface of the pharynx (Fig. 23C, D), these terminals produced several very thin and long branches that were barely connecting to the surface and were not capable of providing strong mechanical binding (Fig. 24C, D). These processes could be a part of the neural-like integrative system.
Using SEM, we detected a few mesogleal neuron-like cells associated with muscles (Fig. 25). The largest of these cells were 5-7 μm in diameter and had long thin processes (Fig. 25B, E, F). Some of the cells were multipolar, while others were bipolar with extensively branching processes (Fig. 25E, F). The mesoglial cells were found not only in the space among the muscle fibers, but also attached to the surface of the pharynx (Fig. 25C).
3.4 Identification of muscular and neural elements in comb rows
Swim cilia in ctenophores are one of the longest cilia in the animal kingdom with eight comb rows mediating complex locomotion of Pleurobrachia. SEM confirms that the swim cilia are fused together into individual plates - combs or ctenes (Fig. 26A). Each comb row is connected to the aboral organ (the major gravity sensor and integrative center (Tamm, 1982; Tamm, 2014)) by highly specialized ciliated furrows (Figs. 1A, B; 2A, C, D and 27A,B). Two neighboring or “sister” furrows fused near the aboral organ (Figs. 3A, 27A). At the earlier stages of development, the animals have only four ciliated furrows and comb rows, but on the 7-10th day, each furrow and each comb row begin to split into two “sister” structures, making eight homologous ciliated rows in adults (Norekian and Moroz, 2015).
The ciliated furrows were densely covered with cilia (Figs. 26B, 27B). Only tubulin antibody (but not phalloidin) did label the cilia in the furrows and ctenes. Tubulin IR also revealed the polygonal neural net between individual comb plates (Fig. 29B).
Phalloidin labeled three distinct structures in the comb rows: (i) the modified muscle-like (non-contractive) fibers connecting neighboring comb plates (Fig. 28B, C, and 29A, B), (ii)-the basal cushion or polster at the base of each comb plate (Fig. 28B, C and 29A), (iii) and true muscles attached to the comb rows (Fig. 28D and 29C, D). The muscle-like fibers connecting the plates have many branches or extension within the same orientation (Fig. 28B, C and 29A). Most of them in the center mediated a direct mechanical connection between plates (Fig. 28B, C and 29A).
Hundreds of very long individual muscle cells were attached to both sides of comb rows and extended deep into the mesoglea (Figs. 28D, 29C, D, and SEM in Fig. 26C, D). Each muscle cell was 20-30 μm wide with branches close to the cilia row (Figs. 26D and 29D). These numerous muscles were evenly distributed through the entire length of each comb and could retract the comb rows as a defensive withdrawal response in Pleurobrachia.
3.5 Organization of the digestive system: muscles, cilia, and innervation
Pleurobrachia is an active predator frequently preying on different copepods: sometimes two or even three whole copepods could be seen inside the pharynx (Fig. 31C). Phalloidin clearly labeled muscles in copepods.
The mouth in Pleurobrachia is slightly elongated and covered with hundreds of short cilia (Fig. 30A, B), which facilitate movements of food particles. The oral complex continues into the ectodermal stomodaeum or pharynx (Fig. 1) with folded walls for efficient digestion (Fig. 30C). The pharynx has many muscles, attached to it from its mesogleal side (Fig. 30C, D). The pharynx opens into a short stomach chamber (Fig. 30D, E). These two digestive compartments are separated by a narrow sphincter, which isolates the pharynx from the canal system. The sphincter area is densely covered by long cilia (Fig. 30E). The thin-walled stomach extends into the aboral canal that then continues to the aboral complex, where it branches into two shorter canals with two muscular anal pores used for excretion (Fig. 2A and 30F).
Neural plexus that covered the entire body of Pleurobrachia continues to the “lips” area (Fig. 31A) with two furrows inside the mouth. The mouth was outlined by numerous cilia, which brightly stained by phalloidin (Fig. 31B, E and 32B), but not tubulin antibody. A large group of muscles encircled the mouth (Fig. 32A, B) and mediated its closure. In contrast, fewer muscle filaments had radial orientation, presumably leading to the opening of the mouth (Fig. 32B).
Like the mouth, the elongated pharynx was covered with numerous short cilia (Fig. 31D, E and 32C, D). The pharynx was connected to the short stomach chamber via a small constriction, or sphincter, which could isolate the pharynx from the thin-walled stomach (Fig. 31D, E and 32C, D). It appears that this sphincter allows only partially digested food or small particles to go through the system, while large parts like chitinous remains of the copepods would remain in the pharynx and be ejected via the mouth. The sphincter was also covered with cilia stained by phalloidin (Fig. 32C, D) similarly to oral parts of the digestive system.
The short, thin-walled stomach had the opening into the whole gastrovascular canal system. The first part includes the aboral canal that extended toward the aboral organ (Fig. 31E). The wall of the aboral canal contained a layer of thin longitudinal muscles (Fig. 13A, 32C and 33D). The stomach chamber also connected to two short transverse canals, which turned into two tentacular and four interradial canals, running in radial directions through the mesoglea (Hernandez-Nicaise 1991). Each interradial canal branches into two adradial canals and reaching every comb rows subsequently forming the meridional canals (Fig. 34A, B).
The interradial and adradial canals were lined up with cilia (Fig. 35C) facilitating the nutrient flow along the gastrovascular system. These cilia were specifically labeled by tubulin antibody (Fig. 35A, B), but not phalloidin.
Tubulin antibody also stained several individual pores located along each meridional canal (Fig. 34C), called ciliated rosettes (Hernandez-Nicaise 1991). These pores were found through the entire length of every canal under the comb rows facing the mesogleal area of the body away from the combs. SYTOX green staining, which was used to label the nuclei in the cells, revealed that each pore consisted of at least 10-12 cells forming a circle (Fig. 34D). Numerous long cilia attached to every single rosette from the mesoglea side (Fig. 35D). Some of the pores were completely closed, while others were wide open in our fixed preparations. These pores provided a connection between the gastrovascular canal system and mesogleal area of the body.
Using SYTOX green or DAPI for nuclei staining had an unusual side-effect in our studies providing a nice visualization of some anatomical structures. For example, SYTOX green in our preparations brightly stained parts of the reproductive system (Fig. 36). Pleurobrachia is a hermaphrodite, like most of the ctenophores, with bands of male and female gonads residing in the walls of the meridional canals: the ovary is located on one side and the testis on the other (Fig. 36B, C).
3.6 The diversity of muscle systems in Pleurobrachia
Table 4 summarizes the information about more than a dozen muscle types in Pleurobrachia. In fact, the diversity of muscles and muscle type cells (frequently labeled by phalloidin) exceed diversity of neuronal or receptor subtypes (see Tables 1 and 2) and comparable with the diversity of ciliated cells (Table 3).
Muscle cells attached to the cilia rows were among the largest and, perhaps, the most powerful muscle groups in Pleurobrachia (Figs. 26C, D and 29D). These multinucleated cells were flat, 20-30 μm wide (Fig. 33A), and extend into the mesoglea. Although Pleurobrachia does not change the body shape, these flat muscles, as well as body wall muscles, could support the maintenance of the hydroskeletal and mediated local movements such as defense withdrawal of comb rows.
The body wall of the Pleurobrachia contained a number of external parietal muscles, which formed a loose rectangular network of fibers (Figs. 33B and 37A). Most of them were running in the oral-aboral direction, while some filaments were localized in the perpendicular direction (Fig. 37A). Near the aboral complex, the pattern of muscles was different. Muscle fibers were running in parallel to each ciliated furrow and then were ‘bent’ towards the neighboring ciliated furrow (Fig. 37B). Phalloidin revealed many muscles in the wall of each tentacle pocket (Fig. 33C). Contraction of these muscles might help to extend the tentacles. There were also many distinct groups of muscles in the wall of the stomach and circular muscles in the aboral canal (Fig. 33D and 37C).
We also identified a broad spectrum of muscle filaments crossing the mesogleal region and attached to the body wall, pharynx or tentacle pockets (Fig. 18).
There were also many distinct groups of muscles in the wall of the stomach and circular muscles in the aboral canal (Fig. 33D and 37C) and other parts of the digestive system as revealed by SEM microscopy (Fig. 23).
Deciphering tissue organization and a cell type diversity in the mesoglea was challenging. Numerous muscle fibers identified in the mesoglea, including the ones attached to the pharynx, were stained by phalloidin only and not labeled by tubulin antibody. They were usually thicker and different than neuronal filaments. In contrast, the mesoglea also contained many very thin neuronal filaments that were labeled only by tubulin antibody and were not stained by phalloidin.
However, a small fraction of the cells with long processes that were labeled by tubulin antibody also showed phalloidin staining in double-labeling experiments (Fig. 38A, B). In the freshly stained preparations (within two days of processing), when the concentration of phalloidin was at the highest end, and the antifade mounting medium was used, the proportion of filaments stained by both phalloidin and tubulin antibody could reach 20-30%. Even a few cells (but not all) with neuron-like morphology were labeled by phalloidin under these conditions. In contrast, the neurons within the subepidermal neural net were never labeled by phalloidin.
4 DISCUSSION
Origins and early evolution of neural systems is a highly debated topic, and ctenophores are one of the key reference species (Striedter et al., 2014) in any reconstruction of the genealogy of cell types in Metazoa (Moroz, 2018). Recent advantages of comparative neuroscience and single-cell genomic point out to insufficient knowledge about the cellular diversity in basal metazoans in general, and ctenophores in particular. In fact, recent single-cell RNA-seq profiling of two ctenophores (Moroz, 2018; Sebe-Pedros et al., 2018) indicated that many cells could not be reliably recognized using conventional approaches. The absence of cell-specific molecular markers for ctenophore neurons and muscles is a noticeable bottleneck in the field. As a first step, a systematic reference microscopic atlas is needed for Pleurobrachia and related species. This work is based on earlier studies (Hernandez-Nicaise, 1991; Jager et al., 2011; Jager et al., 2013; Moroz et al., 2014; Norekian and Moroz, 2016) and provide the detailed atlas of the neuro-muscular organization of Pleurobrachia bachei.
Using two markers (tubulin IR and phalloidin), we conservatively recognized at least 45 distinct morphological types of cells related to sensory, neuronal and effector systems (Tables 1–4) as well as male and female gonads, colloblasts, pores in the canals, epithelial cells, balances, phalloidin stained cells in aboral organ, secretory cells, etc. Combined with electron microscopical observations (Horridge, 1965; Hernandez-Nicaise, 1968; 1973b; a; c; Hernandez-Nicaise, 1991), we estimate that at least 80 morphologically and ultrastructurally distinct cell types present in ctenophores, including at least 9 types of neurons and neuronal-like cells, 6 types of putative receptors and a large diversity of muscles and muscle-like elements – many of them described here for the first time.
The overall patterns of neuro-muscular systems in Pleurobrachia bachei is similar to those reported in a related species (Jager et al., 2011), but here, we recognized novel types of muscular, ciliated and neuronal-like cells as well as six distinct receptors (vs. one or two types described previously – see also (Hernandez-Nicaise, 1991)). These data have been complemented by scanning electron microscopy, which is challenging, if not impossible, on such fragile species as Mnemiopsis or Bolinopsis.
There is an unexpectedly complex organization of the tentacular nerves and innervation of the tentacle pockets, which might reflect the role of tentacles in prey detection and feeding including a stereotyped rotation behavior, characteristic for this species (Tamm, 1982).
Importantly, we identified connections between the subepithelial and mesogleal networks, suggesting their functional coupling (see also (Hernandez-Nicaise, 1968; Hernandez-Nicaise, 1991)), despite remarkable differences in 3D organization and cellular composition. The subepithelial network was never labeled by phalloidin, but some mesogleal neuron-like elements could be stained by both the tubulin antibody and phalloidin with at least three distinct morphological types. Furthermore, we detected connections between the subepithelial nerve net and the polar fields as well as ciliated conductive tracts – which provides strong morphological support for the integration of all parts of the neuro-muscular system.
Our mapping suggests that neurons in the subepithelial net can perform at least three canonical functions: (a) be motoneurons for muscles (Fig. 11I, with glutamate as a neuromuscular transmitter (Moroz et al., 2014)), (b) be a primarily sensory cells with at least two, or more modalities (e.g. mechanosensation with different strength, and a possible chemoreception (Kass-Simon and Hufnagel, 1992)), (c) be interneurons since we identified many cells with branched neurites as well as connections to the mesogleal net, polar fields, ciliated furrows, combs, tentacle nerves, and mouth. It is also possible that a simpler nerve net in ctenophores and their neurons might combine multiple functions (Hernandez-Nicaise, 1974); and be physiologically as well as chemically highly heterogeneous recruiting different neuropeptides and non-conventional low molecular weight transmitters (see discussion in (Moroz, 2015; Moroz and Kohn, 2015).
Control of comb beating occurs under three mechanisms (i) neuronal inputs; (ii) cilia-based coupling and conductance; and (iii) coordination, conduction, and coupling by non-contractive muscle cells including those cells between comb plates described in the current study (Figs. 27 B, C and 28 A, B). These intercomb bridges can be an analog to the Purkinje fibers, the electrical conductive system in the mammalian heart when muscle derivatives ‘lost’ contractility and ‘become’ neuroid-like elements mediating synchronous conduction and coupling of the entire organ. The similar conjecture can be applied to enigmatic and quite diverse mesoglea cell populations with mixed patterns of phalloidin/tubulin IR.
Tubulin labeling and distinct morphology releal the presence of at least four types of neurons related to the subepithelial net with connections to combs, the aboral organ, tentacles and other parts of the feeding system such as pharynx. We also classified at least three types of tubulin IR mesogleal neurons, some of them shared morphological characteristics with phalloidin (only)-stained muscles. Mesogleal muscles might act as part of the body hydroskeleton and mediate retractions of combs or support functions of the pharynx or other organs. However, there are some musclelike/neuron-like cells, not only with the shared morphology but also can be double labeled both with both tubulin AB and the F-actin marker - phalloidin. In other words, these cells have a ‘dual’ nature and, perhaps, shared origins.
Could neurons evolve from muscle-type ancestor lineages? An instant hypothesis: the mesogleal neurons and some other ‘neuroid’ populations are derived from muscle-like cells in ctenophores. Moreover, over the course of evolution, ‘true’ contractive muscle cells might be transformed into neuronal-like elements similar to fibers between combs or fibers in the mesoglea. Equally possible, that muscle and neuronal cells might have a common ancestral cell lineage in evolution and, therefore, share a core gene regulatory network. Considering that mesoderm and muscles in ctenophores can evolve independently from those in bilaterians (Moroz et al., 2014) - the described diversity of cell types in ctenophores can be a fertile ground for testing hypotheses of convergent evolution on the cellular level.
It is equally possible that neural-like elements might share the common ancestry with ciliated cells, which reach their utmost diversity of forms and functions in Ctenophora - more than in any other clade of the animal kingdom. We treat these evolutionary reconstructions, not as an alternative, but rather complementary hypotheses of the convergent evolution of cell phenotypes in Metazoa. Moreover, the hypothesis of convergent evolution of neurons and muscles in ctenophores is congruent with any outcome of the ongoing discussion about the Ctenophora-sister (Whelan et al., 2015b; Arcila et al., 2017; Shen et al., 2017; Whelan et al., 2017) vs. Porifera-sister (Pisani et al., 2015; Feuda et al., 2017; Simion et al., 2017) phylogenetic hypotheses. Indeed, many genes involved in the neuronal and muscular specification and functions had not been identified in the sequenced ctenophore genomes or, if such gene orthologs are present, they are not dependable neuronal or muscular markers (Moroz et al., 2014; Moroz, 2015; Moroz and Kohn, 2015; 2016).
Life of ctenophores is based on cilia (Tamm, 2014), and ctenophores took full advantage to explore cilia to achieve countless innovations in structures and functions including signal transmission as in the ciliated furrows. In summary, both ciliated cells and muscles have greater overall morphological diversity than neurons. However, we suspect that there is a hidden molecular diversity among all cell types and mostly in ctenophore neurons, which can be revealed in the ongoing and future systematic comparisons of cell-specific genomic, proteomic and metabolomic studies (Moroz, 2018). The presented atlas of the neuro-muscular organization in Pleurobrachia would be one of many reference platforms to understand the complex life of comb jellies better.
Conflict of interest
None of the authors has any known or potential conflict of interest including any financial, personal, or other relationships with other people or organizations within three years of beginning the study that could inappropriately influence, or be perceived to influence, their work.
Role of the authors
All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. TPN and LLM share authorship equally. Research design: TPN, LLM. Acquisition of data: TPN, LLM. Analysis and interpretation of data: TPN, LLM. Drafting of the article: TPN, LLM. Funding: LLM.
Acknowledgments
We thank FHL for their excellent collection and microscope facilities including the new Nikon Laser Scanning confocal microscope; Dr. Victoria Foe for the use of the BioRad confocal microscope and Dr. Adam Summers for the use of SEM. We also thank Dr. Claudia Mills and Dr. Billie Swalla for useful ctenophore discussions.
This work was supported by the United States National Aeronautics and Space Administration (grant NASA-NNX13AJ31G), the National Science Foundation (grants 1146575, 1557923, 1548121 and 1645219) and National Institute of Health (grants R01GM097502, R01MH097062-01A1).
Footnotes
Grant Acknowledgments: This work was supported by the United States National Aeronautics and Space Administration (grant NASA-NNX13AJ31G), the National Science Foundation (grants 1146575, 1557923, 1548121 and 1645219), and Human Frontiers Research Program and National Institute of Health (R01GM097502).