Bryozoans represent a large lophotrochozoan phylum with controversially discussed phylogenetic position and in group relationships. Developmental processes during the budding of bryozoans are in need for revision. Just recently a study on a phylactolaemate bryozoan gave a comprehensive basis for further comparisons among bryozoans. The aim of this study is to gain more insight into developmental patterns during polypide formation in the budding process of bryozoans. Particular focus is laid upon the lophophore, also its condition in adults. For this purpose we studied organogenesis during budding and lophophoral morphology of the ctenostome bryozoan Hislopia malayensis.
Polypide buds develop on the frontal side of the developing cystid as proliferation of the epidermal and peritoneal layer. Early buds develop a lumen bordered by the inner budding layer resulting in the shape of a two-layered sac or vesicle. The hind- and midgut anlagen are first to develop as outpocketing of the prospective anal area. These grow towards the prospective mouth area where a comparatively small invagination marks the formation of the foregut. In between the prospective mouth and anus the ganglion develops as an invagination protruding in between the developing gut loop. Lophophore development starts with two lateral ridges which form tentacles very early. At the lophophoral base, intertentacular pits, previously unknown for ctenostomes, develop. The ganglion develops a circum-oral nerve ring from which the tentacle nerves branch off in adult zooids. Tentacles are innervated by medio-frontal nerves arising directly from the nerve ring, and medio-frontal and abfrontal nerves which originate both from an intertentacular fork.
We are able to show distinct similarities among bryozoans in the formation of the different organ systems: a two-layered vesicle-like early bud, the ganglion forming as an invagination of the epidermal layer in between the prospective mouth and anal area, the digestive tract mainly forming as an outpocketing of the prospective anal area, and the lophophore forming from two lateral anlagen that first fuse on the oral and afterwards on the anal side. Future studies will concentrate on cyclostome budding to complement our knowledge on developmental patterns of bryozoans.
The Bryozoa represent a large lophotrochozoan phylum consisting of sessile filter-feeders comprising over 6000 extant species. The phylum consists of three large clades: The Phylactolaemata, the Stenolaemata and the Gymnolaemata (Ctenostomata and Cheilostomata) . The relationship in between the different clades and also to other phyla remains controversially discussed . The Phylactolaemata represent a small group of freshwater inhabiting species. From a phylogenetic perspective they are interesting, since they are often regarded as the most basal bryozoans and show several morphological characters that distinguish them from all remaining bryozoans, such as an epistome and body wall musculature [1, 3]. In particular their sexual development, however, is heavily altered, probably as an adaptation to freshwater habitats, and therefore impedes comparisons to other phyla and bryozoans. Within the Gymnolaemata, the Ctenostomata are a group of uncalcified, comparatively simple species that are currently regarded as paraphyletic with the species-richest bryozoan group Cheilostomata as well as Cyclostomata being ingroups [4–7]. Consequently, they represent an important clade for addressing phylogenetic questions of bryozoans.
As previously mentioned by Nielsen , budding in bryozoans, in particular organogenesis, is only poorly known. Schwaha et al.  recently studied the organogenesis in the budding process of the phylactolaemate Cristatella mucedo and established a first comprehensive study for further comparative purposes. Detailed investigations on the polypide development during the budding process of ctenostome bryozoans were only carried out by Davenport  for Paludicella articulata. Soule  studied several species, but only gave generalized and short descriptions with poor documentation. Accordingly, ctenostome budding requires new data to gain more insight into general trends and patterns in bryozoan budding. This study focuses on the organogenesis in the budding process of the ctenostome Hislopia malayensis, a species occurring in freshwater habitats of South East Asia . Since it shows many ancestral traits among ctenostomes , it represents a suitable species for the current study. With the lophophoral base being the most complex organ of the polypide, we laid special focus on its formation and differentiation, but also on its condition in adults.
Specimens of Hislopia malayensis Annandale, 1916 were collected from the pond of the Faculty of Fisheries of the Kasetsart University in Bangkok (see ). Colony pieces were fixed in 1.5 % glutaraldehyde in 0.01 M sodium cacodylate buffer (pH 7.4) for about 1 hour. Specimens were afterwards rinsed three times for 20 minutes in the buffer. Until further preparation in Vienna, specimens were stored in the buffer. Postfixation was conducted with 1% Osmium tetroxide solution in distilled water for 1-2 hours, followed by rinsing in distilled water for about 1 hour. Specimens were afterwards dehydrated with a graded alcohol series prior to embedding the samples into Agar Low-viscosity resin using acetone as intermedium. Eight colony pieces each containing 1-2 buds and several adult zooids were taken for sectioning. Ribbons of serial semi-sections (1 μm thickness) were conducted as described by Ruthensteiner . Sections were stained with toluidine blue and afterwards analysed and photographed with a Nikon DS5M-U1 digital camera mounted on a Nikon Eclipse E800 light microscope. Image stacks from the serial section micrographs were enhanced in contrast, converted to greyscales and imported with an image size of 1024 × 768 into the 3D reconstruction software Amira 4.1 (Mercury Computer Systems, Chelmsford, MA, USA). Alignment of the image stacks was conducted with the AlignSlice tool of Amira. Segmentation of different structures was conducted manually with a brush. A surface for each structure was generated followed by iterated steps of triangle reduction and smoothing (see ). Snapshots were taken with the Amira software.
Condition of the lophophoral base in adult zooids
Origin of the budding layers
Like in all other bryozoans, the polypide in H. malayensis develops from two budding layers; the inner budding layer originates from the epidermis and the outer budding layer from the peritoneum . Some previous investigations described the outer budding layer to form from proliferating epidermal cells [11, 16]. More recent observations [17, 18] found both layers of the body wall directly involved in the formation of buds. Although this study did not focus on the early bud formation, we never found any peritoneal cells derive from the epidermal layer. In addition, it should be mentioned that the peritoneal layer of the body wall in H. malayensis is always inconspicuously thin, even in adult zooids. As a consequence, it is more reasonable to assume the sole thin peritoneal layer to form the outer budding layer. Whether coelomocytes liberated from the peritoneal layer, as observed in the current study in early buds, participate in the formation of the outer budding layer remains unanswered. Different kinds of coelomocytes within the body cavity have been reported in representatives of all bryozoan clades. In adult zooids, they possibly act in phagocytosis of excretory substances . Their role during budding could be similar in accumulation of metabolic waste created during budding. On the other hand, coelomocytes may be involved in the formation of peritoneally derived tissues, such as muscles. A similar function has been indicated for phylactolaemate coelomocytes .
Formation of the lophophore
The initial lophophore anlage develops as two lateral ridges in H. malayensis. A similar formation has been described for all other bryozoan clades (Cyclostomata: , Cheilostomata: e.g. [18, 20], Ctenostomata: [10, 11], Phylactolaemata: ). In Paludicella articulata the lateral ridges first unite at the oral side, while the tentacles on the anal side are the last to form . In the current study on H. malayensis, a stage showing a U-shaped arrangement of the developing tentacles was not encountered. However, in our budding stage 3 of H. malayensis, the lophophoral ridges bulge slightly inward on the oral side, whereas they abruptly end on the anal side as described in P. articulata. A similar formation of the lophophore is described for the cheilostome Membranipora membranacea  and the phylactolaemate Cristatella mucedo . The Phylactolaemata, however, show some differences regarding the formation of the lophophore which are also reflected in their adult condition. The lateral lophophoral ridges or more precisely bulges form the large lophophoral arms giving this clade the typical horse-shoe shaped lophophore. At first these do not carry tentacles in phylactolaemates  as seen in members of the Ctenostomata  and the Cheilostomata [18, 20]. In contrast to the remaining, pre-dominantly marine clades, the oral tentacles are the first to be formed in the Phylactolaemata [9, 21, 22]. However, these differences are again reflected in the condition of the coelomic compartments of the adults. In the Phylactolaemata the ring-canal on the oral side of the lophophore base is comparatively short supplying only few tentacles , whereas the ring canal in the Ctenostomata (H. malayensis, this study) and Cheilostomata (Cryptosula pallasiana, ) encompasses almost the entire lophophoral base. Accordingly, two major patterns in the development of the lophophore can be recognized from the currently available data: 1. it starts with paired lateral anlagen that first close on the oral side and later also on the anal side and 2. the first tentacles arise on the area of the prospective ring canal with the most medial ones on the oral side to appear last.
Intertentacular pits of the lophophoral base
As stated by Gordon , the lophophoral base represents the most complex structure of the bryozoan polypide. Surprisingly, his detailed study of the organization in the cheilostome Cryptosula pallasiana currently remains the only study for bryozoans. In H. malayensis we describe intertentacular pits at the lophophoral base for the first time in a ctenostome. A similar, most probably homologous structure is present in C. pallasiana. In the latter they are called ciliated pits and measure 25 - 30 μm in length . In adult specimens of H. malayensis they are approximately two times longer than in C. pallasiana and consequently much more noticeable. In C. pallasiana the pits are covered by a cuticle and the lining cells possess cilia projecting into the lumen of the pit. In H. malayensis a thin acellular layer, most likely cuticle, lines the intertentacular pits as well, but the confirmation of the presence of cilia would require electron microscopic examination. Conspicuous intertentacular pits also occur in several other ctenostome bryozoans (Schwaha, unpublished data). Consequently, it seems likely to expect them in more cheilostomes as well and thus might be a synapomorphy for these two clades. As in C. pallasiana , we currently can give no indication about the function of the intertentacular pits.
Formation of the central nervous system and adult condition
In H. malayensis the central nervous system forms by an invagination of the inner budding layer (epidermal layer) in between the prospective mouth and anus, thus being identical to all other bryozoan classes in this respect [8, 9]. However, in contrast to the Phylactolaemata, the ganglion in H. malayensis never contains a lumen and thus is compact in late budding stages (stage 5) and adults.
The nervous system, in particular at the lophophoral base and the tentacle innervation, has been subject of several studies in phylactolaemates and gymnolaemates [1, 23–31]. So far, the central nervous system in cyclostome bryozoans remains unstudied. Tentacle innervation, however, is briefly mentioned by Nielsen & Riisgard . All studied bryozoans possess a circum-oral/pharyngeal nerve ring. Tentacles are innervated by 4-6 nerves. Four of these tentacle nerves (the abfrontal, frontal and the paired latero-frontal nerves) are located subepidermally, while the remaining two are located subperitoneally. Only the phylactolaemate Asajirella gelatinosa shows a slightly different configuration of the subepidermal nerves . In the current study on H. malayensis, we were only able to locate the four subepidermal nerves and not the paired subperitoneal ones. Only in the cheilostome Cryptosula pallasiana, the full set of six tentacle nerves was detected . In the cyclostome Crisia eburnea  and the cheilostome Electra pilosa  the four subepidermal nerves were confirmed, whereas only the two subperitoneal and the latero-frontal tentacle nerves were found in Flustrellidra hispida, Membranipora membranacea , Farrella repens and Alcyonidium sp. . In phylactolaemate bryozoans radial nerves extend from the nerve ring in between the tentacles, in the intertentacular membrane. Towards the tentacles, the radial nerves bifurcate within the intertentacular membrane and branch off the tentacles nerves [1, 27]. This intertentacular origin of the tentacles nerves resembles the abfrontal and latero-frontal tentacle nerves of H. malayensis. However, the medio-frontal tentacle nerves branch off directly from the circum-oral nerve ring in H. malayensis. In the ctenostomes Flustrellidra hispida , Alcyonidium sp., Farrella repens  and the cheilostomes Membranipora membranacea  and Electra pilosa  only one pair of tentacle nerves were found to originate from an intertentacular origin, i.e. the latero-frontal nerves. In E. pilosa the abfrontal nerve extends directly from the circum-oral nerve ring, whereas it was not detected at all in the other aforementioned species - most likely a result of methodological problems of vital staining of nervous tissue. Nonetheless, summing up the little information available the following trend seems to be present in bryozoans: In the Phylactolaemata all tentacle nerves are of intertentacular origin, while gymnolaemates subsequently branch off tentacle nerves directly from the nerve ring, first the medio-frontal nerve (Ctenostomata: H. malayensis, this study) and then the abfrontal nerve as well (Cheilostomata: E. pilosa ). This trend coincides with current opinions of bryozoan phylogeny, with the Phylactolaemata as most basal branch and the Ctenostomata being paraphyletic as ancestors of the Cheilostomata. However, a broader range of taxa (including the neglected Cyclostomata) need to be studied to confirm this trend. Additionally, the basic number of tentacle nerves needs to be clarified. In several Phylactolaemata [1, 34] and the cheilostome E. pilosa , 'subperitoneal' or 'enclosed' peritoneal cells that are topologically identical to the position of the subperitoneal nerves described for C. pallasiana  and several ctenostomes [28, 29] were described. Consequently, it seems probable that these cells represent nerves and that they can be expected in most if not all bryozoans, also in H. malayensis, but their detection requires detailed electron microscopic or state-of-the-art immunocytochemical studies.
Formation of the gut and the esophagus-cardia length
The mid- and hindgut in H. malayensis form as anal outpocketing as described for the ctenostomes Flustrellidra hispida , Paludicella articulata  and Pottsiella erecta . As recently summarized by Schwaha et al. , diverging descriptions of gut formation have been provided. Some authors claim an oral outpocketing to give rise to these parts of the digestive tract [11, 37]. Considering the similarities in the formation of all other organ systems during budding of bryozoans, it appears more probable that the mid- and hindgut of bryzoans generally develop from an anal outpocketing. Ultimately, reinvestigating and increasing the number of species in all clades needs to be conducted to confirm this suggestion.
As mentioned by Rogick , the gut terminology of bryozoans is in a 'nice state of confusion'. Only Silen  attempted to give a general terminology to the various parts of the digestive tract by considering all bryozoan classes. Two valve-like constrictions, one at the end of the foregut and a second before the intestine (or rectum), are important criteria for assigning terms for specific gut regions . The valve at the end of the foregut is commonly termed cardiac valve or esophageal valve and represents the border between the esophagus and the cardia. As seen in the current study on H. malayensis, previous studies on the Phylactolaemata  as well as the Cyclostomata  this valve develops at the border of the two anlagen assembling the gut during budding. In the Phylactolaemata and the Stenolaemata the digestive tract from the caecum to the mouth opening is short, while it is usually elongated in gymnolaemates. Based on the distal position of the cardiac valve in the latter, this elongated tube was considered to be a result of the elongation of the cardia. Consequently, an esophagus was stated to be absent in gymnolaemates, because no proper differentiation towards the pharynx is given . While this might be true for several cheilostome species (e.g. Membranipora: ; Bugula: ; Cryptosula: [16, 40]; Electra: ; Hippothoa: ; Lageneschara: , also see [1, 43]), ctenostome bryozoans show a larger variation concerning this feature of the gut. Our results on H. malayensis show that the cardiac valve is situated far proximally and most of the tube-like elongation develops and consists of the foregut, the esophagus. Only a comparatively small part of the tube is composed of the cardia distally of the muscular proventriculus. An identical arrangement is present in the hislopiids H. corderoi  and Echinella placoides . In contrast, in the only ctenostome superfamily showing similar flattened box-shaped zooids, the Alcyonidioidea, the esophagus is negligibly small and the cardiac tube elongated [46, 47]. An elongated esophagus is generally considered to be present in 'stoloniferan' ctenostomes  in which the polypide bud becomes dislocated into an elongated peristome that later is separated from the remaining 'stolon' [49, 50]. However, in other ctenostomes with elongated peristomes but lacking true stolons, like the Victorellidae, the esophagus and cardia are both present as almost equally long tubes. In addition, the relative size, particularly of the cardiac tube, is affected by the state of its contraction . It seems worthwhile to investigate whether the differences in the morphology of the gut prove to be valuable for drawing phylogenetic inferences. Comparative data is currently sparse, because the location of the cardiac valve is only given for very few species. Since the cardiac valve hinders reflux of food particles from the cardiac stomach, it seems more reasonable that the anatomy of the gut is influenced by the diet and the mode of digestion.
Compared with the recent study of the phylactolaemate Cristatella mucedo  and older studies, we are able to show that the development of the polypide shows distinct similarities in the formation of the different organ systems. These include the early polypide bud formation as a proliferation of epidermal cells bulging towards the peritoneal layer of the bud, a two-layered vesicle-like early bud, the central nervous system or ganglion forming as an invagination of the epidermal layer in between the prospective mouth and anal area, the digestive tract mainly forming from an outpocketing of the prospective anal area that grows towards a comparatively small anlage of the foregut (pharynx and esophagus), and the lophophore forming from two lateral anlagen that first fuse on the oral and afterwards on the anal side. These similarities found between phylactolaemates and ctenostomes thus support the monophyly of Bryozoa.
The site where the anlage of the mid/hind-gut and the foregut meet is represented in adult zooids by the cardiac valve. A comparison of different bryozoan species and superfamilies shows that its location is not identical in gymnolaemates which always possess an elongated tube-shaped gut connecting the pharynx with the caecum. With the current paucity of comparative data, it is more appropriate to consider the diet and the mode of digestion to be decisive on the variable location of the cardiac valve.
At the complex lophophoral base of adult zooids intertentacular pits of unknown function are described for the first time in a ctenostome. Similar structures were only reported in the cheilostome C. pallasiana . It is likely that they are present in more if not all gymnolaemate species, but have escaped the attention of previous investigators. Along with structures of the nervous system at the lophophoral base and the tentacle innervation, these characters appear promising for further analysis for comparative phylogenetic purposes on bryozoans.
With the polypide development of the Phylactolaemata  and Ctenostomata (this study) studied in more detail with modern visualisation techniques, the Cyclostomata remain an essential taxon for further study. Organogenesis in the budding process of the later was only studied by Borg  and Nielsen , but is only poorly documented. In cyclostome bryozoans the polypide is formed first and the cystid later. This formation of buds is also found in the basal Phylactolaemata, in contrast to budding of the Cteno- and Cheilostomata where the cystid is formed first and the polypide later. Accordingly, future studies should concentrate on cyclostome budding to complement our knowledge on developmental patterns of bryozoans.
We would like to thank the staff of the Department of Environmental Sciences of the Kasetsart University of Bangkok and especially Jukkrit Mahujchariyawong, Patana Anurakpongsatorn, and Ratcha Chaichana, for their support. TS trip to Thailand was supported by the KWA-scholarship of the University of Vienna. TS is currently supported by FWF project P 22696-B17 granted to Andrey Ostrovsky (University of Vienna).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.