i) Systematics:
Diagnoses of the Spengelidae and its genera:
Class Enteropneusta [42] .
Family Spengelidae [43].
Glandicipitidae [44].
The family Spengelidae is distinguished by a short proboscis; the presence of a circular muscle layer in the proboscis; stomochord with a vermiform process; absence of collar roots, very wide and well-developed muscles in the perihaemal cavities; presence or absence of hepatic sacs, lack of a lateral septum, and small egg size [21] [43] [45].
ii) The etymology of the species name: Schizocardium karankawa. sp. nov. other Schizocardium species are named according to their location of collection, including Schizocardium braziliense [46] from Brazil, S. peruvianum [46] from Peru, and S. californium from California [21]. S. karankawa were first collected along the South Texas coast, home of the Karankawa First Nations people, from who it gets its name. Its known range extends to Mississippi where it was collected at 12.5 meters depth by [22] and referred to as S. cf. brasiliense. It is likely common throughout the subtidal Gulf of Mexico in fine mud.
iii) Experimental manipulations
To explore this animal’s development and increase the efficacy of fertilization events, three small experiments were done to increase sperm motility, fertilization, and developmental success. In some animals, the neutral NH3 crosses the sperm membrane resulting in an alkaline interior that results in sperm motility. To test this, 100 µm of 100 mM NH4Cl was added to 1 ml of immobile sperm. Sperm motility was not induced at that concentration. On the other hand, sperm motility increases with time in sperm that were removed from the testis and diluted in pasteurized seawater. At zero minutes sperm motility was less than 1%. At 20 minutes sperm motility was less than 20%, and 30 min less than 40% and by 50 minutes sperm motility was high at about 80%. In a second experiment, motile sperm were added to oocytes that were either 3 hours post dissected (and rinsed) or freshly dissected (and rinsed). Fertilization was only observed in the freshly dissected oocytes. These two experiments suggest that the presence of sperm in seawater may induce female spawning, and that male to female proximity may be distant in the wild. A third experiment was done to determine when germinal vesicle breakdown (GVBD) occurs in the absence of sperm. Twenty minutes post dissection (and rinse) 0/200 eggs showed GVBD; at 1h 5 min 5/200 oocytes showed GVBD, at 1h 50 min 5/200 showed GVBD, at 2h 35 min 8/200 showed GVBD and at 3h 35 min 17/200 showed GVBD. In the presence of sperm, GVBD occurs almost immediately. These results indicate that sperm induces GVBD in S. karankawa oocytes.
An additional set of small experiments were done to determine the best seawater recipe for developing embryos, to determine if dechorionated embryos would develop normally, and to estimate the production of gut enzymes in early-stage and late-stage tornaria. The first experiment consisted of rearing fertilized eggs in filtered and pasteurized seawater or in the artificial seawater Jamarin U (Jamarin La. Co., Osaka, Japan). Those in Jamarin U showed a higher fertilization success and after 4 hours embryos in Jamarin U had developed more synchronously. A second set of experiments was done to dechorionate embryos and to determine if they would develop normally. Trypsin and protease did not weaken the chorion, nor negatively affected development. Further embryos were passed through nitex meshes of 100 µm and 75 µm but the chorion passed through intact, and development was normal. Dechorionation was achieved using a pair of fine pointed tungsten needles at the 2-cell, 4-cell, 8-cell, 16-cell stages, and development was normal, indicating that this species is suitable for experimental manipulations including microsurgery, and cell injection investigations. No difference in chorion strength was apparent when dechorionated at newly fertilized versus 2-cell stage embryos. For the third experiment, 80-hour embryos were incubated in an enzyme medium for esterase and alkaline phosphatase and no staining was apparent. This shows that early tornaria either do not produce these digestive enzymes or the protocol was done too early before the enzymes are produced. Seventeen-day old larvae showed a strong alkaline phosphatase reaction after incubation for 10 hours. Dark purple staining showed in the gut, pillae of the ectoderm and the apical plate. At this same developmental time-period, no esterase activity was detected.
Iv) Development:
Worms were collected starting on February 7, 2001, but gravid animals and successful fertilization was not achieved until February 20, marking the start of that reproductive season. The last fertilizations gotten, marking the end of the season was April 16. These were not gravid: the oocyte nuclei were mostly central, and many were small because yolk had not entered the vitelline envelope. The largest oocytes were 140 µm. On February 19 more worms were collected, a small incision made into the dorso-lateral gonads, releasing oocytes or sperm. Oocytes were 138 µm, some were slightly oblong and only a few were very early oocytes with little vitelline space. Spermatid heads were 3.6 µm and tails were 40.0 µm and very few were actively swimming. On Feb 20 mature sperm and oocytes were obtained. Oocytes were rinsed and fertilized with diluted sperm. The jelly layer was visualized with India ink at 300 µm in diameter. Within 4 minutes the fertilization membrane expanded with the germinal vesicle still present. By 30 minutes post-fertilization the nucleus migrated from a central location to a side of the oocyte. First and second polar bodies were observed 1.5 hours post-fertilization (Fig. 1A). The embryo was 117 µm and the expanded envelope 159 µm. The first cleavage event began at 2 h and completed at 2h 30 min with oblong (66 µm by 90 µm) blastomeres (Fig. 1B). Nuclei were apparent at 2h 20 min. Four cell stage was complete at 3h 40 min (Fig. 1C), 8 cell at 3h 55min (Fig. 1D), 16 cell at 5h with polar bodies atop (animal pole) four tiers of blastomeres including two central macromeres, vegetal and animal mesomeres which then divide to form two tiers of micromeres (Fig. 1E). A blastula (Fig. 1F) was formed at 6h 15 min, and gastrulation commenced at 9h. The gastrula soon took a hemispherical form (maximum 163 µm wide) with a large blastopore and no cilia (Fig. 1G). At 11h 45 min post-fertilization short cilia begin to rotate the embryo in the vitelline space at the speed of 35 seconds per revolution. Three tori shaped tiers of cells then develop around the blastopore reducing its size. Hatching occurred at the developmental period where the archenteron appeared as a dense mass (Fig. 1H).
At 22 h 30 min post-fertilization the apical end of the early larva flattened and thickened (pre-apical tuft) and the ectoderm at the site of the mouth decreased in width. There was no feeding or telotroch-ciliated bands (Fig. 1I). At 34h 30 min post-fertilization the larva had a well-developed apical tuft included flagella with a length of 58 µm, an apical retractor muscle, buccal cavity oesophagus, stomach, intestine, and the site of the feeding and telotroch bands were apparent but the cilia were uniformly 14 µm in length (Fig. 1J). Coeloms were absent. At 50 h post-fertilization the apical plate showed some pigment, and a few mesenchyme cells just below it. A well-formed pericardium pulsated once every 10 seconds and was in direct contact with the protocoel. Podocytes were apparent on the protocoel duct that connected to an excretory pore (Fig. 1J). No mesocoels or metacoels were present. External cilia were 23 µm in length, uniform, with a predominantly downward beat. At 58 h post-fertilization larva had well developed feeding bands, larval rotation was clockwise when viewed from the apical pole and the algae Nanochloropsis was ingested. At 72 hours post-fertilization a pair of dark eyespots were apparent. Table 1 shows the developmental time periods of Schizocardium karankawa from February 25 and April 10, 2001, at 19oC and 23.5oC, respectively.
More worms were collected, and fertilizations done between March 13 to April 12, 2013, and reared through tornaria (Fig. 1K), and metamorphosis to the juvenile worm stage (Figs. 1L to 1O). We do not know the start of the 2013 reproductive season, but the last worms collected on April 12 included 8 females and 9 males. Very few to no oocytes were found marking the end of the 2013 season, four days earlier than the 2001 season. These embryos were maintained in culture through tornaria (Fig. 1K), late-tornaria (Figs. 1L and 2A), though metamorphosis (Figs. 1N and 2B), to the 6-gill pore stage (Figs. 1O). Metamorphosis was asynchronous with the earliest of the sibling tornaria settled at 50 days post-fertilization (Fig. 1L). Others had not metamorphosed or settled 85 days post-fertilization, when our observations ended. A cavity in the anterior proboscis shows the site of the larval apical tuft (Fig. 2C). The tornaria anus was ciliated and a vertical band of cilia connected the mouth to the telotroch, then the trochophore to the anus (Fig. 2A and D). Like other indirect developing acorn worms, metamorphosis was not catastrophic, but a gradual process that incorporated larval tissue into the adults, through elongation and regional (proboscis, collar, trunk) elongation, with the telotroch maintained into the early juvenile stage (Fig. 2B). The cilia of the larval apical organ were lost in the worm, but the pigment cells were maintained. This most apical point appears to maintain a sensory function into the early juvenile worm stage (Fig. 1M). It projects forward and actively explores the surface. In SEM is appears as a pit, though this may be a dehydration artifact (Fig. 2C). It lacked the long cilia of the larval apical tuft. It is not clear whether neurons of the larval apical organ, which is sensory, are maintained as the animal transitions from the plankton to the benthos.
Pigment granules that marked the larval feeding bands were maintained on the juvenile proboscis and anterior collar (Fig. 1M), whereas those of the telotroch were easily identifiable into the 8-gill slit stage (compare Figs. 1N and 1O). The pharyngeal gills developed in the late tornaria where they did not obviously serve a function, because the ectodermal pores had not opened to the exterior until the juvenile worm stage. About 1-month post-metamorphosis at the 6 gill-pore stage, gill bars were apparent as were protuberances on the trunk, which contain CaCO3 ossicles in adult acorn worms (Figs. 1P and 2E). The gill bars at this stage have well developed primary gill bars with lateral cilia (Fig. 2F). Elongation by cell proliferation is most evident in the post-trochophore trunk.
We performed confocal microscopy on actin (phalloidin) labeled, late stage tornaria (Fig. 3A), early and 4-gill pore stage juvenile worms (Figs. 3B and C). The 50-day post-fertilization stage tornaria had a well-developed apical retractor strand that connects the contractile pericardial sac to the apical tuft (Fig. 3A). In living animals, it frequently, rapidly contracted to form a temporary dimple in the animal pole. Another set of contractile muscles connected the pericardium with the oesophagus. The oesophagus was muscular dorsally (Fig. 3A). Together, these muscles were responsible for a ‘coughing’ behaviour that would clear the mouth of algae. At this stage muscle cells began to develop in the early mesocoels, and an anal sphincter muscle was evident (Fig. 3A). Newly metamorphosed worms had limited circular muscle fibres in the proboscis. Little muscle was detected in the collar. The trunk has well developed anterior to posterior muscle bands (Fig. 3B). These began to form more discrete, paired dorsal and ventral-lateral bundles around the 30-days post-settlement stage (Fig. 3C). The dorsal pair are more densely packed than the ventral-lateral muscles, though the ventral ones are more extensive, as they are in adults. The paired buccal and perihaemal muscles of adult worms, which develop from anterior extensions of the trunk coeloms, are not apparent at this stage. Well defined muscle cells form a sphincter around each of the gill pores (Fig. 3C).
V) External Adult Features:
Total body length of the type specimen after fixation was 60.7 mm. The proboscis length was 5.1 mm by 4.6 mm wide. The collar was 2.3 mm long and 3.4 mm wide, and the trunk 53.3 mm long (branchial region 25.6 mm long, the hepatic region was 20.6 mm, and the caudal region was 7.1 mm), while the anterior width of the trunk was 3.4 mm, and posterior width was 2.1 mm. This is generally smaller than Schizocardium peruvianum and Schizocardium californicum [21]. The first paratype whole body length was 66 mm, proboscis length was 8 mm, and the width was 5 mm. The collar was 5 mm long and 3 mm wide. The first third of the hepatic region sacs had one pair then two pairs of sacs, then mid-hepatic region the sacs were arranged in rows of three to four pairs (Fig. 4). The posterior third of the hepatic region had one or two pairs, the outer more well developed than the inner which were small and oval, unique to this species. The synapticula that bridge the primary and secondary gill bars are medium developed. Gill pores were diminutive but numerous, estimated at 120–150 pairs. The gill bars were long, extending almost the depth of the pharynx, from dorsal to ventral, except for a ventral hypobranchial ridge, characteristic of the genus. The very small pores and large pharynx may be an adaptation to deposit feeding on very fine sediment. We found no evidence of a filter feeding current through the pharynx.
Vi) Internal Adult Features:
The proboscis ciliated epidermis was 400–500 µm thick. The proboscis cavity was filled with diffusely arranged longitudinal muscle fibres surrounded by a layer of a circular muscle of almost the same thickness as the epidermis (Fig. 5A), with a well-developed nerve fiber layer between them. The proboscis coelom was filled with a connective tissue arranged in a small plate, and the center of the coelom was divided into distinct left and right portions via a thin muscle plate, each with a cardiac vesicle tube and glomerulus enveloping the anterior end of a bifurcated stomochord from which the genus gets its name (Fig. 5B). The right-side cardiac vesicles extended more anteriorly than the left. Each cardiac vesicle has a narrow coelom that separated it from the glomerulus that surrounds it. Each glomerulus has a ring shape anteriorly, and crescent shape with a thick ventral-lateral wall posteriorly (Fig. 5B). Posteriorly, the glomeruli and the cardiac vesicles connect in conjunction with the appearance of the stomochord. The vermiform process of the stomochord was short. A single proboscis pore was located on the left of the dorsal midline of the proboscis neck, like Schizocardium peruvianum and S. californicum [21]. At this point, a thick dorsal septum connected the stomochord with the dorsal wall of the proboscis, forming a deep dorsal groove in the proboscis (Fig. 5C). A limited ventral septum may attach to the ventral wall of proboscis but was broken in transverse sections (Fig. 5C). The stomochord radius was greatest and variable shaped due to intrusions of the surrounding basal lamina sheath, forming diamond, and star shapes, with a horizontal mesh of tissue in its central part. The stomochord lacked a lumen anteriorly, but posteriorly a small one appeared and disappeared sequentially. The stomochord lacked blind pouches. At this level of proboscis, the thickness of the epidermis and the circular muscle were about equal.
At the posterior proboscis and anterior collar, the ventral septum disappeared while the dorsal mesentery continued. The proboscis skeleton had a prominent keel and chondroid tissue (Fig. 5D). The skeletal keel was short, concave dorsally and convex ventrally (Fig. 5D). The dorsal collar mesentery continued and a ventral one was lacking. More posterior, in the anterior collar, the epidermis was enlarged (Fig. 5E). The skeleton bifurcated into paired cornua and a wide anterior collar neuropore was found. The collar nerve cord had no lumen like the other species of Schizocardium [21]. A dorsal mesentery extended from anterior to mid-collar with no evidence of the ventral one. More posteriorly, at the termini of the skeletal cornua a ventral mesentery began in conjunction with the dorsal one. The buccal cavity of the posterior collar had a long and large epibranchial ridge that continued into the trunk pharynx lumen (Figs. 5E, F, G, H, and I). There was a posterior collar neuropore. The pharynx epibranchial ridge had six zones of cell types, like the epibranchial ridge of S. brasiliense [26] (Fig. 5G). All zones were ciliated. Zone one or the medial zone of cells were light stained and transparent. The adjacent pair of cells, or zone two were slightly biconcave, nuclei present in a dark pink color with trichrome stain. The margin of zone three was biconvex shaped resulting in a pair of grooves, with darked stained cells with long cilia. The cells of zone four had a tapered base containing a large vesicle and distally positioned nuclei. The cells of zone five had large granular vesicles. Cells of zone six were comprised of columnar cells with apical nuclei. Under these zones was a thick epithelial nerve layer (Fig. 5G).
The branchial region of the trunk had left and right collar canals. The dorsal nerve cord of the trunk had small lacunae that appeared and disappeared anteriorly, but none posteriorly (Fig. 5H). The pharynx had a ventral mesentery. The boundary of the pharynx lumen was demarcated by primary and secondary (or tongue) gill bars. The secondary bars had peripharyngeal cavities (Figs. 5H, I). There was a well-developed dorsal blood vessel. Anteriorly, the pharyngeal trunk longitudinal muscles were thickest laterally and tapered dorsally and ventrally (Fig. 5F), whereas in the posterior pharynx region the thickness was less variable (Figs. 5H, I). The ventral digestive pharynx narrowed from anterior to posterior until it was a greatly reduced hypobranchial strip. The posterior pharynx had well-developed dorsal and ventral blood vessels and dorsal lateral gonads (Fig. 5I). The posterior end of the trunk was damaged, but the intestine had a thick and winding wall, and a few gonads were found at the ventral side of the trunk. This region of the trunk had well-developed dorsal and ventral blood vessels.
Vii) Phylogeny:
The full-length 16S sequence obtained from the Texas coast Schizocardium karankawa transcriptome differed from the 16S sequence obtained from the Mississippi coast Schizocardium cf. brasiliens mitochondrial genome (MH841936.1) at only 7/1362 shared positions (excluding a short stretch of 12 Ns in the Schizocardium cf. brasiliens sequence not considered in the analysis), resulting in an uncorrected p-distance of 0.005. For this reason, and their proximity, we regard these two populations from the Gulf Coast of the USA as Schizocardium karankawa. For comparison, 16S of Schizocardium karankawa and Schizocardium californicum differed at 59/1365 shared positions, resulting in an uncorrected p-distance of 0.043. Phylogenetic analysis of 16S recovered Schizocardium monophyletic with Schizocardium karankawa sister to Schizocardium cf. brasiliens, albeit with moderate support (bs = 82) and Schizocardium sister to the other sampled genus of Spengelidae, Glandiceps, with maximal support (Fig. 6A).
Our phylogenomic pipeline retained 7,948 OrthoGroups totaling 1,504,053 amino acids in length with 58.5% missing data after concatenation. Phylogenetic analysis of this matrix yielded a topology consistent with the current understanding of hemichordate phylogeny (Fig. 6B). Except for Ptychoderidae, all hemichordate families were recovered monophyletic with maximal support. As observed in previous studies, Torquaratoridae was nested within Ptychoderidae. Spengelidae was recovered as the sister taxon of the Ptychoderidae-Torquaratoridae clade with maximal support. Schizocardium karankawa was recovered as sister to Schizocardium cf. brasiliens with maximal support. Here we regard them as the same species, S. karankawa, based on the US Gulf coast population locations, and the short phylogenetic distance between them.