Stolons are individuals with specialized structures for swimming and reproduction
In this study, the differences between stocks and stolons were firstly revealed by detailed morphological and histological observations in M. nipponica (Fig. 7). As expected, and in agreement with previous literature on the morphology of syllid stolons in other species [6, 40], the stolons in M. nipponica do not have a functional digestive tube subdivided into a pharynx, proventricle and caeca, while they have a uniform non-functional digestive tube. Stolons of M. nipponica are dicerous [7], which means that they bear two pairs of enlarged eyes and two pairs of short antennae (Fig. 2). Anterior structures of stolons have been proposed to reveal phylogenetic information [12]. These characteristics of M. nipponica agree with those revealed in other Megasyllis species [41]. As for the inner structures, a stolon possesses a ganglion in a dorsal head region of the anterior-most segment, showing a similar arrangement to that of a stock brain (Fig. 3). This suggests that stolons have their own central nervous system independent from the stock. The enlarged eyes of a stolon appear to be functional in terms of photoreception, and the short antennae are suggested to be used for the reception of pheromones from opposite-sex stolons. The anterior nervous structures in the head of stolons agree with previous descriptions in other species as well [18, 21, 31]. In contrast to stocks, stolons lack some additional sensory organs such as palps and ciliated nuchal organs that directly project from the brain [42, 43]. Stolons appear to have simple and specialized sensory organs adaptive for their swimming and spawning behavior.
Gonad development precedes stolon head formation
In this study, furthermore, the developmental stages during stolonization were defined, along which gene expression profiles were investigated (Fig. 8). Histological observations during the development revealed that, at Stage 1, tissues stained with hematoxylin were found on both lateral sides of segments posterior to the kinked gut (Fig. 5). Generally, annelids have a pair of gonad primordia abutting on a septum [44, 45], so that the tissues observed at Stage 1 (Fig. 5b, c) are thought to be the reproductive primordia, indicating that the gonad development seems to start at this stage. At Stage 2, female individuals were found with oocytes and males with developing testes (Fig. 5d–g), indicating that sexes have been determined before this stage. At Stage 3, stolon eyes were formed at the head part. Nerve tissues were observed interior to the stolon eyes at Stage 4, and then, at Stage 5, they enlarged at the dorsal side surrounding the gut (Fig. 5h–s), suggesting that the onset of nervous modifications follows the gonadal development.
Analyses of gene expression of germ-cell markers showed differences in gene expression profiles (Fig. 6d). Namely, vasa and piwi were up-regulated in the posterior region toward Stage 2 and then gradually decreased, while the nanos expression increased toward Stage 4 and 5. The expression patterns of vasa and piwi were similar to those in Capitella teleta, where immature oocytes showed up-regulation of vasa, nanos and piwi, whereas mature oocytes showed no expression of these [46, 47], which is consistent with the lack of expression of germ-line markers in mature females of Typosyllis antoni [26]. No significant differences in the expression patterns of vasa and piwi were detected between the sexes, suggesting that the two sexes share the earlier process of gametogenesis. Generally, in annelids, vasa, nanos and piwi are also expressed in the Posterior Growth Zone (PGZ) and are known to be involved in posterior growth [48–51]. In M. nipponica, secondary-tail regeneration of the stock occurs at Stages 4 and 5, when nanos expression is up-regulated. Therefore, the late expression of nanos, in comparison with vasa and piwi, could be involved in the later process of gametogenesis, or the PGZ formation of the regenerating stock tail.
Following the vasa and piwi expressions, the head-identification genes, i.e., six3, otx and pax6, were up-regulated, simultaneously with the tissue changes in the nervous system alterations during stolonization (Fig. 8). This may suggest that the expression of head-identification genes in the stolon head could be triggered by the gonad development, leading to the neuronal modifications.
Stolon lacks body-trunk identity
Annelids exhibit homonymous metamerism, wherein the internal and external morphological features are repeated in each body segment [52]. In stolonization, however, gonads mature only in the segments which develop into a stolon, and the construction of the nervous system that forms the stolon ganglion occurs only in the first segment of the stolon. Therefore, it is suggested that positional information should be required to identify at least the first and the subsequent segments of a stolon to be differentiated. Metazoans generally share a similar body plan along the anteroposterior axis that is determined by some specific conserved genes such as Hox genes and head-identification genes (six3, otx, pax6, nk2.1), so it was predicted that the stolon was formed based on ectopic expression of these genes. If the stolon has a similar anterior-posterior axis to that of a stock, it is expected that these genes would be expressed ectopically at the region where the stolon is formed.
In fact, some head-identification genes (six3, otx, pax6) were up-regulated in the posterior part of the body during Stages 4 and 5 (Fig. 6b), i.e., when neural tissue develops at the anterior end of the stolon, suggesting that these genes provide identity to the anterior end as the stolon head. On the other hand, the expression of the anterior Hox genes (Hox1, Hox2, Hox4) was highly localized in the anterihageegion of a stock with a developing stolon, while those of the posterior Hox genes (Lox4, Lox2, Post2) were higher in the posterior region (Fig. 6a), as shown in other annelids [53–55]. Also, temporal changes of the expression patterns of these Hox genes were not detected during stolonization. These results suggest that, unexpectedly, the body-trunk identities specified by Hox genes are not altered during stolon formation, and the head region is determined in the middle of the existing AP axis of the body. It is possible that the head identification genes might be co-opted downstream of positional genes such as Hox that determine the middle region, for the formation of the stolon head. The results are also consistent with the fact that stolons lack the digestive tract components such as the pharynx, proventricle, and caeca, which the stock has anteriorly, and that stolons have repeated body segments except for the anteriormost segment (prostomium) and the posteriormost segment (pygidium) (Fig. 3). In terms of organization and gene expression, although the stolon head seems incomplete in comparison with that of a stock, stolons have a tail and a head, but lack the body trunk identity. It is particularly unique, in this phenomenon, that the stolon head formation is launched in the segment with the posterior body-trunk identity. Further studies will reveal what mechanisms underlie the induction of head-identification genes expression in the posterior region of the body trunk.
Hormonal control of stolonization
In this study, since it is well known that juvenile hormone (JH) and ecdysone are generally involved in molting and metamorphosis in arthropods [38, 39], we hypothesized that they may also be involved in stolonization. Therefore, the expression patterns of genes involved in the up- and downstream pathways were also investigated (Fig. 6c). The results showing that FAMeT, one of the juvenile hormone synthases, was mainly expressed in the anterior part of the body, peaked at Stage 3, and decreased thereafter, suggest that the JH pathway is involved in stolonization in M. nipponica. As for the downstream factors of JH, while males did not show significant expression changes, in females, Krh1 was gradually up-regulated toward Stage 5, following the FAMeT peak. This epistatic relationship between molecular pathways is similar to that of insects [38, 39].
In Platynereis dumerilli, which exhibits epitoky, MF, a JH precursor, is suggested to promote posterior growth while inhibiting yolk formation [33]. In females at Stage 5 in M. nipponica, when Krh1 expression is maximal, the oocytes are fully mature and the stock tail is regenerating, suggesting that the JH pathway may cause an energy shift from yolk formation to posterior growth. This result is consistent with the fact that MF inhibits yolk formation [33], while it contradicts a previous finding that the expression of a downstream gene of MF (MTr) is upregulated in individuals with mature stolons [37]. Since the expression of ecdysone receptors was also upregulated in Stage 5, it is possible that they are involved in morphogenetic processes such as the development of chaetae, as in the metamorphosis in insects.
Overall, juvenile hormone and ecdysone may be involved in the progression of stolon formation, although they could not trigger stolon formation, because their expression changed in the later stages of stolon formation. Future studies will reveal more comprehensively the endocrine mechanism in stolonization by analyzing other hormones, such as gonadotropin-releasing hormone, inhibin, and BMP, which are hypothesized to be involved in sexual maturation in annelids [56, 57], and comparing them with the results obtained in this study.