Development of the foregut in Katharina tunicata (Mollusca; Polyplacophora)

As a highly diverse phyla, Mollusca is both intriguing to study and difficult to define. No single common feature distinguishes the phyla, but the most common and recognizable is the radula. Throughout Molluscan history, there have been many developments to the radular structure. Evolutionary development of unique radular structures may have been made possible by developmental modularity. This histological study examines the development of polyplacophoran internal gut development. Here, I describe the developmental sequence of chiton feeding structures for the first time. Feeding structures are present by 10 days post hatch. Future research should examine larval stages prior to this in order to determine whether developmental modularity exists in polyplacophorans.


Introduction 19
With over 100,000 described species, Mollusca is one of the largest and most successful 20 metazoan phyla (Seed, 1983). Although no single character defines the phylum, the radula is one 21 of the most distinctive and easily identifiable features, and is present in members of all 22 molluscan classes except Bivalvia (Runnegar and Pojeta, 1985;Salvini-Plawen, 1988;Trueman 23 and Clarke, 1988;Purchon, 2013). Fossil records date Mollusca back ~500 My to the Cambrian, 24 but the abrupt presence of 15 genera in the Phanerozoic records suggest a Precambrian presence 25 that may yet await discovery (Runnegar and Pojeta, 1985;Trueman and Clarke, 1988;Todt et 26 al., 2008). The fossil record includes trace fossils from Ediacaran strata, radula-like fossils from 27 the mid-Cambrian, and definitive radular evidence from the Ordovician (Scheltema et al., 2003;28 Todt et al., 2008). The fossil record suggests that radular rasping may be the primitive molluscan 29 feeding mode (Seed, 1983). The presence of this unique feeding structure throughout molluscan 30 history is suggestive of its importance within the clade, and possible presence in the molluscan 31 common ancestor. 32

33
The polyplacophoran radula is unique among molluscs encompassing key features from each of 34 the other major types; these features allow them to simultaneously excavate and sweep food from 35 the substratum (Steneck and Watling, 1982). With robust buccal muscles and biominerally 36 hardened denticles, the chiton radula is capable of generating immense force (Steneck and 37 Watling, 1982). Bilaterally symmetric ribbons of transparent proteinaceous material hold 25 to 38 150 rows of teeth in a lateral arrangement of 17 teeth per row, occupying up to 1/3 of the body 39 length (Schwabe and Wehrtmann, 2009). Four teeth per row are used in grazing, two dominant 40 and two marginal (Steneck and Watling, 1982). Dominant teeth are often tricuspid and heavily 41 mineralized for sediment excavation; x-ray diffraction shows the presence of magnetite (an iron-42 containing biomineral), goethite, lepidocrocite, carbonate apatite, and magnetite occurs in 43 varying degrees per species (Lowenstam, 1962;Steneck and Watling, 1982). 44

45
The ancestral life-history and evolution of the molluscan life cycle are still extensively debated. 46 Extant molluscs' broadly fall under two of the three main types of life-history patterns: 47 planktonic feeding larvae (planktotrophic), and planktonic non-feeding larvae (lecithotrophic), 48 with isolated instances of the third type (direct development). The two major hypotheses 49 regarding life history evolution of the molluscan ancestor are the 'larval-first' hypothesis and the 50 'intercalation' hypothesis (Page, 2009). According to the larval-first hypothesis, or "Trochaea 51 theory", the ancestral molluscan condition was holoplanktonic. Later, these organisms adopted a 52 benthic lifestyle later in development and postponed sexual maturation to the benthic stage. 53 Eventually, the initial planktonic stage of the life history, which must have been capable of 54 feeding, became the planktotrophic larvae of a life history that was otherwise mostly benthic. 55 The intercalation theory starts with a holobenthic organism, which later acquiring a planktonic 56 juvenile stage. Because the planktonic stage is added secondarily, structures for planktonic 57 feeding did not exist and the larva was therefore lecithotrophic. Subsequent development of 58 larval feeding is then a derived state under the intercalation hypothesis (Haszprunar et al., 1995). 59 60 Basal clades may preserve the ancestral state, so mapping life history on to phylogeny can help 61 resolve the ancestral state and polarity of change for molluscan life history evolution. Within 62 Gastropoda, Patellogastropoda and Vetigastropoda are generally considered the basal clades; 63 within Mollusca, Polyplacophora is considered basal. All three of these clades produce 64 lecithotrophic larvae. Furthermore, among extant gastropods, the smallest species almost always 65 exhibit direct development or lecithotrophic larvae and the fossil record suggests that the first 66 molluscs were very small (Chaffee and Lindberg, 1986). The notion that the first molluscs did 67 not produce planktotrophic lavae is also corroborated by the fossil record (Nützel et al., 2006). 68 Finally, comparative studies of cleavage patterns during early development of spiralians have 69 also been interpreted as supporting the notion of ancestral lecithotrophy rather than 70 planktotrophy (Guralnick and Lindberg, 2001). Clades of gastropods with planktotrophic larvae 71 show the largest diversity and complexity in feeding systems (Page and Hookham, 2017 Resolving the molluscan phylogeny has been regarded as "one of the greatest challenges in 76 invertebrate evolution" (Todt et al., 2008;Sigwart et al., 2013;Schrödl and Stöger, 2014). The 77 two major competing hypotheses are Aculifera/ Conchifera and Testaria/ Conchifera; both with 78 Polyplacophora as a basal clade (Schrödl and Stöger, 2014) Serial 1 μm thick sections were obtained using a manual rotary MT5000 Sorvall Ultra 91 microtome. One cross-sectional series was prepared of a ten days-post-hatch (dph) specimen; all 92 other sections were longitudinal. Four longitudinal section series were completed on larvae at the 93 following stages: 10 dph, 13 dph, 17 dph (X2); another four were juvenile specimens: 1 day post-94 metamorphosis (dpm), 14 dpm, 35 dpm (X2). Sections were stained using Richardson's stain and 95 mounted on glass slides (Richardson et al., 1960 Results 105

Larval development 106
Hatched larvae were egg-shaped, measuring 290 µm long by 230 µm wide (Fig. 2). Larvae 107 hatched with 125 µm cilia extending anteriorly (the apical tuft), surrounded by a collar of short 108 15 µm long microvilli, and a locomotory band of 70 µm long cilia encircling the body at its 109 greatest girth (prototroch). As observed by Watanabe and Cox (1974), the larvae also had a short 110 tuft of cilia (30 µm) protruding from the teleotrochal field at the posterior end of the body. At 2 111 dph, larvae had grown more linearly but were still egg-shaped with a wider pretrochal region and 112 a tapered postrochal region (Fig. 3). Length was 340 µm and max width was 250 µm. The entire 113 larval body was highly opaque. 114 At 4 dph, larval eyespots (diameter = 15 µm) and mouth were visible postrochally (Fig. 4). The 116 larvae did not exhibit much growth in body size at this stage (360 µm x 250 µm), but the body 117 appeared more barrel-shaped with the pre-and postrochal regions approaching similar width. 118 The larva also appeared to have a small prototrochal bulge (~15 µm). Although the ventral side 119 was still opaque and dull, the dorsal surface was beginning to reflect light, and the periphery was 120 semi-transparent. A fine covering of microvilli extended from the rostral tip, across the ventral 121 surface, and around the caudal tip; microvilli did not extend across the mantle field (dorsal side 122 from caudal tip to just above prototroch; Fig. 5; Leise, 1984). 123 124 By 6 dph, the larvae had elongated, giving it a narrower profile with a near pill-shaped form. A 125 pit was observed on the anterior tip where the apical tuft extended from the surface. The apical 126 tuft and prototroch both remained near hatching length (140 µm, 75 µm respectively). The shell 127 gland anlagen were beginning to take form on the dorsal surface as narrow ridges and grooves. 128 Larval eyespots were observed, and the girdle surrounding the periphery was becoming nearly 129 transparent. 130 131 Spicules and six shell valves were visible by 8 dph as thin mineral deposits. Five to six rows of 132 spiniferous cells and their associated intracellular spicules could be seen lining the mantle field. 133 The larval eyes were situated directly below (ventrally) the spicule band and behind (caudally) 134 the prototroch (Fig. 6; Henry et al., 2004;Leise, 1984). At this point, large cells appear on the 135 dorsal surface, particularly in the shell gland anlagen and the foot begins delineation (Fig. 7, 8). 136 The first larval stage that was sectioned, 10 dph, was observed to have eyespots (pigment only), 138 buccal mass, radular teeth, radular cartilages, esophagus, neuropil, and spicules (Fig. 8). The 139 mantle field spicules could be seen in a linear arrangement running dorsocaudally in the 140 postrochal region, and a few pretrochally on the dorsal edge ( Fig. 9; (Henry et al., 2004;Leise, 141 1984). In whole mounts, the foot could be seen along the ventral side as a large v-shaped 142 shadowy structure (Fig. 10). By this stage, the prototroch was visible in greater detail and I could 143 view the double layer of circumferential trochoblasts (Fig. 11). 144 145 Longitudinal sections through the flanks of 13 dph specimens showed large, clearly visible 146 eyespots with pigment surrounding a colourless center (Fig. 12). By 13 dph the stomodeum and 147 radular cartilages were much more apparent and differentiated (Fig. 13). 148 149 By 17 dph the larvae were elongated and dorsoventrally flattened with translucent periphery and 150 highly delineated foot. The radular teeth could be seen extruding from the buccal mass as they 151 were added to the radular ribbon (Fig. 14). At this stage the larvae still had apical tuft and 152 prototroch. 153 154 Metamorphosis 155

Larvae of K. tunicata metamorphosis is indicated by loss of the prototroch and apical ciliary tuft. 156
At 1 dpm the foot was clearly delineated by the pallial groove and seven shell valves were 157 clearly evident and beginning to extend towards each other longitudinally (Fig. 15, 16A, B). The 158 juveniles were near-round and pale with white shell valves and hairs surrounding the margin of 159 the girdle (Leise, 1984). By 3 dpm, valves were 50 µm wide and juveniles were 700 µm long by 160 450 µm wide (Fig. 16C). 161 162 At 14 dpm, the dorsal side of the juveniles was completely covered by white shell valves (Fig.  163   16D). The radula was well developed with magnetite-capped denticles and the esophagus could 164 be traced dorsally to the stomach and intestine (Fig. 17). Cuticle was also clearly visible 165 proximal to shell valves and the foot had an abundance of pedal glands (Fig. 15B). 166 167 By 36 dpm, there were still only 7 valves present, but they were each significantly wider (75 -168 100 µm), extending towards each other and beginning to articulate (Fig. 16F). The valves also 169 appeared to be differentiating into their separate regions (tegmentum, articulamentum). By this 170 stage the radular denticles and their magnetite caps were highly differentiated and the radular 171 cartilages were arranged in a linear order as opposed to the spherical structure observed at earlier 172 stages. The esophagus could easily be seen lined with cilia and extending dorsocaudally. The 173 observed developmental sequence is summarized in Table 1 (Appendix II). 174 175 Discussion 176 Major developmental stages were observed to occur in the same order as previous research (± 177 10 days; Table 2). The overall developmental sequence observed in K. tunicata was hatching, 178 larval eyes, spicules, shell plates, settlement, and metamorphosis. This developmental list 179 focuses mainly on external morphology and behaviour because most previous developmental 180 studies on polyplacophorans have not provided information on internal events of morphogenesis. 181 Although Leise (1984) examined spicule formation via histological sections, the same was 182 accomplished here without sectioning (Fig. 6). I was, however, unable to pinpoint the exact 183 timing and location of spicule formation, unlike Watanabe and Cox (1974) who note that spicule 184 formation began first in the pretrochal region. 185 186 This study went beyond the external morphological development to examine the timing of 187 foregut ontogenesis. By examining histological sections of K. tunicata at 10, 13, and 17 dph, I 188 was able to determine that the buccal mass, radular teeth, radular cartilages, and esophagus had 189 all begun formation by 10 dph (Fig. 10). By 13 dph the cartilages and buccal mass had a more 190 distinct appearance, and by 17 dph the radula could be seen beginning to extrude from the buccal 191 mass. Magnetite caps on the radular teeth, a unique chiton feature, was evident in juveniles at 36 192 days after metamorphosis. The development of these structures has not been previously 193 described in Polyplacophora. 194 195 This study attempted to narrow the timing of foregut development but was unable to determine 196 the start of formation. Because there was already evidence of foregut structures in 10 dph larvae, 197 examination of earlier larval stages is imperative. Also, because eyespots were evident by 4 dph, 198 histological sections as early, or earlier, than 4 dph would have been beneficial to examine. 199      Scale bar represents 50 µm; pr, prototroch; tr, trochoblast; e/eye, larval eye; at, apical tuft. Figure 12. Transverse histological sections, 1 m thick, of 13 dph K. tunicata larvae. Scale bar represents 50 m; pro/pr, prototroch; ne, neuropil, at, apical tuft; mt, mouth; pg, pedal gland.  with evidence of radular rudiment. Scale bar represents 50 m; bm, buccal mass; pr, prototroch; mg, midgut; mt, mouth; ne, neuropil; rr, radular rudiment.