Zoomorphology

, Volume 130, Issue 1, pp 51–84

The nervous system of Neodasys chaetonotoideus (Gastrotricha: Neodasys) revealed by combining confocal laserscanning and transmission electron microscopy: evolutionary comparison of neuroanatomy within the Gastrotricha and basal Protostomia

  • Birgen H. Rothe
  • Andreas Schmidt-Rhaesa
  • Alexander Kieneke
Original Paper

DOI: 10.1007/s00435-011-0123-2

Cite this article as:
Rothe, B.H., Schmidt-Rhaesa, A. & Kieneke, A. Zoomorphology (2011) 130: 51. doi:10.1007/s00435-011-0123-2

Abstract

We present a reconstruction of the nervous system of Neodasys chaetonotoideus Remane, 1927 (Gastrotricha, Chaetonotida) based on different microscopical methods: (1) immunohistochemistry (anti-acetylated α- and β-tubulin-, anti-5-HT- and anti-FMRFamide labelling) and (2) histochemistry (labelling of musculature and nuclei) by the means of confocal laser scanning microscopy (cLSM) and (iii) ultrastructure by means of transmission electron microscopy (TEM). All parts of the nervous system contain structures with an immunoreaction against the used immunohistochemical markers and labelling of histochemical markers. Results of both techniques (cLSM, TEM) reveal that the nervous system of N. chaetonotoideus is composed of a “dumb-bell-shaped” brain and one pair of posterior longitudinal neurite bundles. The brain is made up of a pair of laterally located clusters of neuronal somata, a large dorsal interconnecting dorsal commissure and two tiny ventral commissures in the region of the lateral clusters. From this, it follows that the brain is circumpharyngeal in position. The innervation of the head region is conducted by three pairs of anterior-directed neurite bundles. We describe here the gross anatomy of the nervous system and give additional details of the ultrastructure and the 5-HT and RFamide-like IR components of the nervous system. We compare our newly obtained data with already published data on the nervous system of gastrotrichs to reconstruct the hypothetical ground pattern of the nervous system in Gastrotricha, respectively, in Macrodasyida.

Keywords

Gastrotricha Nervous system (Immuno)histochemistry cLSM Ultrastructure Serotonin Tubulin FMRFamide Evolution 

Introduction

Studies focusing on the anatomy of the nervous system in microscopic and lesser known taxa such as Rotifera (e.g. Hochberg 2006, 2007a), Gnathostomulida (e.g. Müller and Sterrer 2004) and Gastrotricha (e.g. Hochberg 2007b, Rothe and Schmidt-Rhaesa 2008, 2009) are still rare, but by degrees facilitate a broad comparison with the nervous system of different taxa. Since antibody labelling, fluorescence techniques, confocal microscopy and computer-aided 3D analyses have widely been applied to morphological studies of small animals (e.g. Wanninger 2007), these comparisons can now be performed in a very detailed and accurate manner, sometimes uncovering homology of single pairs of nervous cells. One problem in investigations of the nervous system is the absence of an ubiquitous marker to be used for the whole nervous system. Instead, one has to use a variety of neuron-specific markers for visualizing a major fraction of the nervous system, at best in combination with electron microscopy (see Rothe and Schmidt-Rhaesa 2009).

The findings of a study on two species of the marine genus Dactylopodola by Rothe and Schmidt-Rhaesa (2009) compared with data of further gastrotrich species (e.g. Hochberg 2007b) yielded a hypothesis on the construction of the common ancestor’s nervous system in Gastrotricha. According to Rothe and Schmidt-Rhaesa (2009), the gastrotrich stem species might have possessed a bilateral-symmetric brain with somata of neurons arranged laterally on both sides of the strong dorsal commissure. There is an additional, but fine, ventral commissure and a pair of ventral longitudinal neurite bundles that span the whole trunk and fuse posteriorly. Although there already exists a reconstruction of the RFamide-like immunoreactive components in the central nervous system of one species of the marine gastrotrich taxon Neodasys cirritus Evans, 1992 (Hochberg 2007b), we here describe the neuroanatomy of a second species of that taxon, Neodasys chaetonotoideus Remane, 1927. The taxon Neodasys is of important phylogenetic value (see e.g. Hochberg 2005) and possibly is the sister group of one of the traditional high-level taxa of Gastrotricha, the Paucitubulatina, together forming the Chaetonotida (Hochberg and Litvaitis 2000). However, a recent cladistic analysis (Kieneke et al. 2008) revealed Chaetonotida to be polyphyletic and Neodasys as the sister group to all other gastrotrichs united within a common monophylum. This putative basal position makes Neodasys a prior object for evolutionary–phylogenetic estimations, such as ground pattern reconstructions. As demonstrated in a recent study dealing with the reproductive system of Gastrotricha, the newly obtained data on N. chaetonotoideus had been obligatory for the evolutionary inferences drawn by a parsimonious character optimization (Kieneke et al. 2009). Since the only data of the nervous system of Neodasys by means of immunohistochemistry comprises the RFamide-like components in the brain of N. cirritus (Hochberg 2007b), we here try to gather a maximum of signals by using a variety of techniques (i.e. immunolabelling with antibodies against 5-HT, acetylated α-tubulin and β-tubulin and FMRFamides, in addition musculature-labelling with phalloidin, DNA staining with DAPI). Furthermore, we try to extend the ultrastructural data on the nervous system of Neodasys, previous TEM-based studies have been carried out by Ruppert (1982, 1991), Ruppert and Travis (1983) and Travis (1983), but these studies did not focus on the nervous system. All data combine to create a detailed reconstruction of the neuroanatomy of N. chaetonotoideus, while signals of certain cell populations (e.g. 5-HT-positive neurons) enable a broader systematic comparison among different taxa. The use of a second species of Neodasys allows recognition of interspecific differences of closely related species. The main goal of this study is to test and complement the hypothesized nervous system of the gastrotrich stem species and to search for relations in neuroanatomy among possible sister taxa, thus bringing us forward to uncover the phylogenetic position of Gastrotricha, which is still under debate.

Materials and methods

Specimen sampling

The specimens of Neodasys chaetonotoideus Remane, 1927 (Gastrotricha: Chaetonotida) used for transmission electron microscopic (TEM) studies were sampled on April 12, 2005 at the eulittoral zone of the Geniusbank in the north of Wilhelmshaven, Germany (Bay of Jade, uppermost and oxygenated layer of the fine sandy sediment) and on September 4, 2005 at the beach of the island Spiekeroog, Germany (North Sea, medium grain size sediment).

The material of N. chaetonotoideus for immunohistochemistry (IHC) was sampled on April 2007 and May 2008 at the island of Sylt (Germany) from different localities in the vicinity of the Wadden Sea Station in List/Sylt. The specimens were found in high abundance in medium grain size sediment with a high content of detritus and also in coarse grain size sediment with a low content of organic material. The animals were extracted from the intertidal sediment with the seawater–ice method (Uhlig 1964) or by the decantation of sediments after a relaxation with a solution of 7% (w/v) magnesium chloride (MgCl2) in distilled water (Rieger and Ruppert 1978).

Immunohistochemical preparations

The gastrotrichs were relaxed in 7% MgCl2 (w/v) in distilled water for 10 min. Completely relaxed specimens were incubated overnight in 4% paraformaldehyde (PFA) (w/v) in 0.1 M phosphate-buffered saline (PBS according Crittenden and Kimble 1999) (pH 7.3) on ice. After fixation, the samples were washed several times in 0.1 M PBS (pH 7.3) and stored in 0.1 M PBS (pH 7.3) containing 0.05% (w/v) NaN3 at 4°C for several weeks. Prior to immunolabelling, the samples were pretreated with preincubation buffer (0.1 M PBS, 0.2% Triton X-100 (C14H22O(C2H4O)n) (Sigma) and 6% (w/v) goat serum overnight at 4°C. The preparations were incubated for 24–48 h in a solution of the primary antibody in 0.1 M PBS containing 1% Triton X-100 at 4°C. The following antibodies were used: anti-acetylated α-tubulin (Sigma) diluted 1:500, anti-β-tubulin (Sigma) diluted 1:500, anti-5-HT (Sigma) diluted 1/2,000–1/4,000 and anti-FMRFamide (ImmunoStar) diluted 1/800. The host species for the antibodies against the acetylated α-tubulin and the β-tubulin were mice, and the antibodies against the neurotransmitters (5-HT and FMRFamide) were produced in rabbits. After incubation, samples were rinsed several times in 0.1 M PBS. This was followed by the incubation with the secondary antibody solution, anti-rabbit immunoglobulin goat serum conjugated with tetramethylrodamine isothiocyanate (Tritc) (Sigma), fluorescein isothiocyanate (Fitc) (Sigma) or the cyanine dye Cy5 (Jackson) diluted 1/100 and/or anti-mouse immunoglobine goat serum conjugated with Tritc (Sigma) or Fitc (Sigma) diluted 1/100 in 0.1 M PBS containing 1% Triton X-100 at 4°C over night. The labelling was terminated by rinsing again several times in 0.1 M PBS.

In some cases, an additional staining with phalloidin Tritc- or Fitc-conjugate (Sigma) and Dapi (4′, 6–Diamidino–2–phenylindol) (Sigma) was used. A concentration of 2 μl of 3.87.6 μM phalloidin or 2 μl of Dapi (1 mg/ml distilled water) was added to 100 μl of the solution of secondary antibodies. Depending on the fluorescence of these stains, Fitc- or Cy5-labelled secondary antibodies were used. After washing, the gastrotrichs were embedded in Citiflour (Plano) on microscopic slides.

The microscopic investigation took place with a cLSM Leica TCS 2 equipped with a near-UV laser (405 nm wavelength). The postprocessing of the data and the projections with greater focal depth were made by using LCS software (Leica) and Zen lite software (Zeiss).

The specificity controls of the polyclonal primary antibodies were carried out by the suppliers by liquid-phase preadsorbtion. The anti-5-HT antibodies (Sigma) were preadsorbed in 500 μM serotonin or 200 μg/ml serotonin–BSA conjugate, and the diluted anti-FMRFamide antibodies (ImmunoStar) were reabsorbed in 100 μg/ml FMRFamide. After the preabsorption, the labelling was completely inhibited. For the monoclonal primary antibodies (anti-acetylated α- and anti β-tubulin, Sigma), the characterization was also made by the supplier. Specificity of the secondary antibodies was tested by the normal incubation procedure, but omitting the primary antibody.

For histochemistry, 29 adult specimens of N. chaetonotoideus were used for anti-5-HT and 31 specimens for anti-FMRFamide. Anti-5-HT and anti-FMRFamide approaches were made as double-, triple- or quad-labelling with phalloidin, anti-acetylated α-tubulin or anti-β-tubulin (totally five specimens) and Dapi as nuclear counterstain.

Electron microscopic preparations

Prior to fixation, specimens were relaxed for a few minutes by incubation in a 7% (w/v) MgCl2 solution. Animals were fixed with a 2.5% glutaraldehyde, plus ruthenium-red for contrasting extra-cellular matrix according to Luft (1964, 1971a, b), buffered in 0.1 M sodium cacodylat (C2H6AsNaO2 × 3 H2O) solution (pH 7.2) for 60 min at 4°C. All specimens were postfixed for 60 min at 4°C in 1% osmium tetroxide (OsO4) (diluted in 0.1 M sodium cacodylat buffer, pH 7.2), dehydrated in an increasing series of acetone and finally embedded in Araldite (polymerization for 72 h at 60°C). Serial ultrathin sections (70 nm) were cut on a Reichert Ultracut E microtome and stained automatically with uranyl acetate and lead citrate (Leica EM Stain). One cross-sectioned animal and an additional one sectioned sagittally have been investigated on a Zeiss EM 502 transmission electron microscope at 80 kV acceleration voltage. Images were captured with a Dual Scan CCD camera using the iTEM® software (soft imaging systems).

Phylogenetic assessment

In order to reconstruct the neuronal character pattern of the stem species of Gastrotricha, we have sampled data on the neuroanatomy of different taxa using a species–character matrix comprising 22 terminal taxa (species) and 40 unordered characters (Table 1). For data input, we used the program Nexus Data Editor 0.5.0 (Page 2001) that operates with the nexus format (Maddison et al. 1997). Some characters are dependent on others, i.e. several entries of the matrix depend on the existence of a superordinate structure. If the superordinate structure is not present, all following substructures have to be coded as inapplicable states (–). Most character states are linked to a certain taxon (species) to simplify comparability (Kieneke et al. 2008 called this “morphological type concept”).
Table 1

Species–character matrix containing 40 neuronal characters of 12 gastrotrich species and 9 outgroup representatives

Character

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Source

Neodasys chaetonotoideus

1

0

4

0

2

1

1

1

1

0

0

1

2

0

1

1

0

1

1

Present study

Neodasys cirritus

1

?

?

0

?

1

1

?

?

?

0

1

0

0

1

1

?

1

?

Hochberg (2007b)

Dactylopodola baltica

1

0

3

?

?

1

1

1

0

0

0

1

4

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2009)

Dactylopodola typhle

1

0

3

?

?

1

1

?

?

?

0

1

?

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2009)

Xenodasys riedli

1

?

?

0

2

1

1

1

1

0

0

1

1

0

1

1

0

1

0

Hochberg (2007b)

Cephalodasys maximus

1

?

?

?

?

1

2

1

0

0

0

1

5

0

1

1

0

?

?

Wiedermann (1995)

Turbanella ambronensis

1

0

0

?

?

1

1

?

?

?

0

1

0

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cornuta a

1

0

1

?

?

1

1

?

?

?

0

1

0

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cornuta b

?

?

?

?

?

1

1

1

0

0

0

1

1

0

1

1

0

?

?

Teuchert (1977)

Turbanella hyalina

1

0

1

?

?

1

1

?

?

?

0

1

0

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cf. hyalina

1

?

?

0

0

1

1

1

0

0

0

1

0

0

1

1

0

1

0

Hochberg (2007b)

Oregodasys cirritus

1

0

0

0

?

1

1

?

?

?

0

1

5

0

1

1

0

1

0

Rothe and Schmidt-Rhaesa (2010a)

Paucitubulatina

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

 

Caenorhabditis elegans

3

1

2

2

4

1

4

1

0

3

0

1

2

1

0

Different authors (see character list)

Tubiluchus troglodytes

2

1

6

1

3

1

4

1

0

3

1

2

1

5

1

0

Rothe and Schmidt-Rhaesa (2010b)

Gnathostomula peregrina

0

?

?

0

0

1

?

1

0

1

1

0

?

?

0

1

1

0

1

0

Müller and Sterrer (2004)

Gnathostomula paradoxa

0

?

?

?

?

1

0

1

0

1

1

0

1

3

0

1

1

0

1

0

Lammert (1986, 1991)

Pterognathia meixneri

0

?

?

?

?

1

0

1

0

1

1

0

1

3

0

1

1

0

1

0

Lammert (1986, 1991)

Notommata copeus

0

0

5

0

1

1

0

1

0

2

1

1

0

0

1

1

0

1

2

Hochberg (2007a)

Macrostomum hystricinum marinum

0

?

?

?

?

1

0

1

0

0

0

?

?

0

1

1

0

1

0

Ladurner et al. (1997)

Macrostomum pusillum

0

?

4

?

?

1

0

1

0

0

0

?

?

0

1

1

0

1

0

Ladurner et al. (1997)

Convolutriloba longifissura

1

0

4

?

?

1

3

1

0

0

?

?

?

?

0

1

1

0

0

Gaerber et al. (2007)

Y Gastrotricha A

1

0

4

0

?

1

1

1

0

0

0

1

?

0

1

1

0

1

0

 

Y Gastrotricha B

1

0

4

0

?

1

1

1

0

0

0

1

?

0

1

1

0

1

0

 

Y Gastrotricha C

1

0

4

0

?

1

1

1

0

0

0

1

5

0

1

1

0

1

0

 

Y Gastrotricha D

1

0

4

0

?

1

1

1

0

0

0

1

5

0

1

1

0

1

0

 

Y Gastrotricha E

1

0

4

0

?

1

1

1

?

0

0

1

?

0

1

1

0

1

0

 

Y Gastrotricha F

1

0

4

0

?

1

1

1

?

0

0

1

?

0

1

1

0

1

0

 

Y Gastrotricha (strict consensus)

1

0

4

0

?

1

1

1

?

0

0

1

?

0

1

1

0

1

0

 

Y Gastrotricha (semi-strict consensus)

1

0

4

0

?

1

1

1

0

0

0

1

?

0

1

1

0

1

0

 

Character

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

Source

Neodasys chaetonotoideus

0

0

0

1

1

1

0

0

0

1

0

0

0

Present study

Neodasys cirritus

0

0

1

1

?

?

?

?

0

0

0

0

0

0

0

Hochberg (2007b)

Dactylopodola baltica

1

2

0

0

0

0

0

0

0

0

0

0

Rothe and Schmidt-Rhaesa (2009)

Dactylopodola typhle

?

?

0

?

?

?

0

0

0

0

0

0

0

0

Rothe and Schmidt-Rhaesa (2009)

Xenodasys riedli

1

0

0

1

1

?

?

?

?

0

1

0

0

1

1

0

0

0

Hochberg (2007b)

Cephalodasys maximus

0

0

?

?

?

?

?

?

0

1

0

1

0

0

1

0

0

0

Wiedermann (1995)

Turbanella ambronensis

0

0

?

?

?

0

0

?

?

0

?

?

?

0

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cornuta a

0

0

?

?

?

0

0

?

?

0

?

?

?

0

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cornuta b

1

0

1

0

1

?

?

?

?

?

0

0

0

0

0

0

0

Teuchert (1977)

Turbanella hyalina

0

0

?

?

?

0

0

?

?

0

?

?

?

0

0

Rothe and Schmidt-Rhaesa (2008)

Turbanella cf. hyalina

0

0

1

1

?

?

?

?

0

0

0

1

1

0

0

0

Hochberg (2007b)

Oregodasys cirritus

1

1

0

0

0

0

0

0

0

0

0

1

0

?

Rothe and Schmidt-Rhaesa (2010a)

Paucitubulatina

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

 

Caenorhabditis elegans

1

3

1

2

1

1

3

?

?

1

2

1

1

?

?

?

0

?

Different authors (see character list)

Tubiluchus troglodytes

1

3

1

2

1

1

2

2

?

1

2

1

1

?

?

?

0

1

Rothe and Schmidt-Rhaesa (2010b)

Gnathostomula peregrina

?

?

?

?

?

0

?

?

?

?

1

1

0

1

0

1

0

0

0

Müller and Sterrer (2004)

Gnathostomula paradoxa

1

0

0

?

?

?

1

0

?

?

0

0

0

0

0

0

Lammert (1986, 1991)

Pterognathia meixneri

1

0

0

?

?

?

1

0

?

?

0

0

0

0

0

0

Lammert (1986, 1991)

Notommata copeus

1

1

0

0

1

0

1

0

0

0

1

0

0

0

Hochberg (2007a)

Macrostomum hystricinum marinum

1

0

1

2

?

?

?

?

?

0

0

1

0

0

?

1

0

1

1

Ladurner et al. (1997)

Macrostomum pusillum

1

0

1

1

?

?

?

?

?

0

0

1

0

0

?

1

0

1

1

Ladurner et al. (1997)

Convolutriloba longifissura

0

0

1

?

1

0

0

1

1

0

0

0

1

?

Gaerber et al. (2007)

Y Gastrotricha A

0

0

1

1

?

1

?

1

0

?

?

0

0

0

0

0

 

Y Gastrotricha B

0

0

?

?

?

1

?

1

0

?

?

0

0

0

0

0

 

Y Gastrotricha C

?

?

0

?

?

?

?

?

?

0

0

0

0

0

0

0

 

Y Gastrotricha D

?

?

0

0

?

?

?

0

0

0

0

0

0

0

 

Y Gastrotricha E

?

?

0

1

1

?

1

?

1

0

?

?

0

0

0

0

0

 

Y Gastrotricha F

?

?

0

1

1

?

1

?

1

0

?

?

0

0

0

0

0

 

Y Gastrotricha (strict consensus)

?

?

0

?

?

?

?

?

?

0

?

?

0

0

0

0

0

 

Y Gastrotricha (semi-strict consensus)

?

?

0

1

1

?

1

?

1

0

?

?

0

0

0

0

0

 

The ‘Y’ symbolizes the stem species, A–F show the results of the character optimization on the basis of six different tree topologies (see text for details). A dash (–) indicates an inapplicable character state, the question mark (?) stands for unknown character states or equivocal solutions for the stem species, respectively

For parsimonious character optimization, we used MacClade 4.0 (Maddison and Maddison 1989, 2000). To this purpose, we have constructed different trees (Fig. 11a–f) reflecting the internal relationships of the Gastrotricha as inferred by three recent phylogenetic analyses (Fig. 11a, b: Hochberg and Litvaitis 2000, Fig. 11c, d: Todaro et al. 2006, Fig. 11e, f: Kieneke et al. 2008). The twelve gastrotrich species (plus the branch leading to the Paucitubulatina) that are included in our character matrix have been manually assembled to cladograms using the tree window of MacClade 4.0. These thirteen gastrotrich taxa were arranged according to the inferred systematic position of congeneric species (systematic position taken from Todaro et al. 2006 and Kieneke et al. 2008) or the genus they belong to (systematic position taken from Hochberg and Litvaitis 2000). We regard this procedure legitimate since monophyly of macrodasyidan and most paucitubulatinan genera is much likely, i.e. the phylogenetic position of a given species can be inferred from the position of a congeneric one. Furthermore, we tried to consider two major hypotheses on the gastrotrich sister group relationship in our constructed trees for testing their influence on the result of the ground pattern reconstruction: Gastrotricha + Cycloneuralia (e.g. Ahlrichs 1995, Ehlers et al. 1996; Fig. 11a, c, e) and Gastrotricha + (Gnathifera + Plathelminthes) (e.g. Giribet et al. 2000; Fig. 11b, d, f). We either set the one or the other group of cycloneuralian or platyzoan representatives as the direct sister group of Gastrotricha. The aforementioned considerations led to the six cladograms depicted in Fig. 11. According to recent outcomes that reveal the Acoela as the most basal group within Bilateria (e.g. Wallberg et al. 2007), we have used one member of these worms to root the trees.

The ground pattern of Gastrotricha has been explored using the “trace character” function of MacClade 4.0. Since the algorithms of programs like MacClade treat an inapplicable (–) like an unknown (?) state, one has to check the resulting character vector for the stem species carefully since sometimes character states, e.g. concerning the number or shape of a certain structure, are reconstructed although the structure itself is absent. To get a concise result from all considered possibilities, we have assembled a “consensus ground pattern” as performed by Kieneke et al. (2009). Here, we have compiled the consensus in (1) a strict and (2) a semi-strict manner, whereas in the strict proceeding only those character states were retained that have been reconstructed on the basis of each tree. In the semi-strict proceeding, character states that have been reconstructed on the basis of at least 50% of the used trees were put to the consensus ground pattern as well (Table 1).

We renounced to calculate a new phylogenetic analysis of different gastrotrichs and some basal protostome species on the basis of the data presented here on the nervous system. The reasons for this are as follows:
  1. (1)

    We believe phylogenetic inference based on a single character system (i.e. the nervous system or just components of it) is not reasonable. We aim for a ‘total phenotypic evidence’ where a maximum of characters (e.g. morphological data of many organ systems) should be analysed and hopefully yield a consistent systematic pattern.

     
  2. (2)

    The data collection is highly fragmentary. Much information on the nervous system of the different taxa used in this paper is still missing as it can be recognized by the many gaps (?) in the matrix.

     
  3. (3)

    Just focusing on the Gastrotricha, data of important taxa are completely missing. There still is e.g. no information on the nervous system of species from the huge group Paucitubulatina or the macrodasyid family Macrodasyidae.

     
In the following, the defined characters and character states along with some annotations are provided.
  1. 1.

    Braingeneral architecture: (0) compact, with central neuropil and cortex of somata—type Gnathostomula paradoxa Ax, 1956; (1) dumb-bell-shaped, with two bilateral clusters of somata and connecting neuropil—type Neodasys chaetonotoideus; (2) “Scalidophora-type”, anterior and posterior, ring-shaped clusters of somata with connecting, ring-shaped neuropil; (3) “Nematoida-type”, circumpharyngeal neurite ring with distinct anterior and posterior neuronal ganglia.

     
  2. 2.

    5-HT IR cells (brain)symmetry: (0) bilateral symmetry—type Neodasys chaetonotoideus; (1) radial-like bilateral symmetry—type Tubiluchus troglodytes (Rothe and Schmidt-Rhaesa 2010b).

     
  3. 3.

    5-HT IR cells (brain)number of cell pairs: (0) one pair—type Turbanella ambronensis Remane, 1943; (1) two pairs—type Turbanella cornuta Remane, 1925; (2) two pairs plus one unpaired cell (5 cells)—type Caenorhabditis elegans Maupas, 1900 (see Chase and Koelle 2007, Loer-Lab homepage: http://home.sandiego.edu/~cloer/loerlab/5htcells.html) Biogenic amine neurotransmitters in C. elegans.; (3) four pairs—type Dactylopodola baltica Remane, 1926; (4) five pairs—type Neodasys chaetonotoideus; (5) six pairs—type Notommata copeus Ehrenberg, 1834; (6) more than six pairs—type Tubiluchus troglodytes Todaro and Shirley, 2003.

     
  4. 4.

    FMRFamide-like IR cells (brain)symmetry: (0) bilateral symmetry—type Xenodasys riedli Schoepfer-Sterrer, 1969; (1) radial-like bilateral symmetry—type Tubiluchus troglodytes; (2) bilateral symmetry with unpaired cells—type Caenorhabditis elegans (see Schinkmann and Li 1992).

     
  5. 5.

    FMRFamide-like IR cells (brain)number of cell pairs: (0) 5 pairs—type Turbanella cf. hyalina Schultze, 1853; (1) approximately 15 pairs—type Notommata copeus; (2) approximately 25 pairs—type Xenodasys riedli; (3) more than 25 pairs—type Tubiluchus troglodytes; (4) 14–19 cells (paired plus unpaired)—type Caenorhabditis elegans (see Schinkmann and Li 1992).

     
  6. 6.

    Dorsal neuropil (brain)existence: (0) absent; (1) present.

     
  7. 7.

    Dorsal neuropil (brain)shape: (0) compact central neuropil—type Gnathostomula paradoxa (Lammert 1986, 1991); (1) belt-like commissure—type Neodasys chaetonotoideus; (2) massive, belt-like neuropil—type Cephalodasys maximus Remane, 1926; (3) thin neurite bundle—type Convolutriloba longifissura Bartolomaeus and Balzer, 1997; (4) massive belt-like neuropil surrounding the pharynx with equal diameter—type Tubiluchus troglodytes.

     
  8. 8.

    Ventral commissure(s) (brain)existence: (0) absent; (1) present.

     
  9. 9.

    Ventral commissure(s) (brain)number: (0) one ventral commissure (of the brain)—type Dactylopodola baltica; (1) two ventral commissures (of the brain)—type Neodasys chaetonotoideus.

     
  10. 10.

    Ventral commissure(s) (brain)shape: (0) single neurite(s) or thin neurite bundle(s)—type Neodasys chaetonotoideus; (1) “buccal neurites (nerves)” connected to the buccal ganglion—type Gnathostomula paradoxa; (2) mastax-related neurite bundles—type Notommta copeus; (3) massive belt-like neuropil surrounding the pharynx with equal diameter—type Tubiluchus troglodytes.

     
  11. 11.

    Buccal ganglionexistence: (0) absent; (1) present—type Gnathostomula paradoxa.

     
  12. 12.

    Buccal ganglionshape: (0) buccal ganglion—type Gnathostomula paradoxa; (1) mastax ganglion—type Notommata copeus; (2) neck ganglion—type Tubiluchus troglodytes.

     
  13. 13.

    Anterior neurite bundle(s)existence: (0) absent; (1) present.

     
  14. 14.

    Anterior neurite bundle(s)number of pairs: (0) one pair—type Turbanella cf. hyalina; (1) two pairs—type Xenodasys riedli; (2) three pairs—type Neodasys chaetonotoideus; (3) four pairs—type Pterognathia meixneri Sterrer, 1969 (Lammert 1986, 1991); (4) five pairs—type Dactylopodola baltica; (5) numerous pairs (more than 10)—type Cephalodasys maximus.

     
  15. 15.

    Unpaired longitudinal neurite bundle (ventral)existence: (0) absent; (1) present—type Tubiluchus troglodytes.

     
  16. 16.

    Paired longitudinal neurite bundles (ventrolateral)existence: (0) absent; (1) present.

     
  17. 17.

    Paired longitudinal neurite bundles (ventrolateral) pronouncedexistence: (0) absent; (1) present (also coded ‘present’ if the ventrolateral longitudinal neurite bundles are the only longitudinal components of the CNS).

     
  18. 18.

    Paired longitudinal neurite bundles (ventrolateral)number of pairs: (0) one pair—type Neodasys chaetonotoideus.

     
  19. 19.

    Paired longitudinal neurite bundles (ventrolateral)caudal fusion (posterior commissure)existence: (0) absent; (1) present—type Neodasys chaetonotoideus.

     
  20. 20.

    Paired longitudinal neurite bundles (ventrolateral)caudal fusion (posterior commissure)shape: (0) arc-like—type Dactylopodola baltica; (1) arc-like with posterior projections of longitudinal neurite bundles—type Neodasys chaetonotoideus; (2) arc-like with posterior projections of longitudinal neurite bundles and neural loop—type Notommata copeus.

     
  21. 21.

    Paired longitudinal neurite bundles (ventrolateral)additional ventral commissuresexistence: (0) absent; (1) present.

     
  22. 22.

    Paired longitudinal neurite bundles (ventrolateral)additional ventral commissuresnumber: (0) one—type Gnathostomula paradoxa; (1) two—type Notommata copeus; (2) four—type Dactylopodola baltica.

     
  23. 23.

    Paired/unpaired longitudinal neurite bundles (ventrolateral/ventral)additional dorsal commissuresexistence: (0) absent; (1) present—type Macrostomum pusillum Ax, 1951.

     
  24. 24.

    Paired/unpaired longitudinal neurite bundles (ventrolateral/ventral)additional dorsal commissuresnumber: (0) one (at least)—type Turbanella cornuta b; (1) two—type Macrostomum pusillum; (2) three—type Macrostomum hystricinum marinum Rieger, 1971; (3) more than three—type Tubiluchus troglodytes.

     
  25. 25.

    Paired/unpaired longitudinal neurite bundles (ventrolateral/ventral)associated cell somataexistence: (0) absent; (1) present.

     
  26. 26.

    Paired/unpaired longitudinal neurite bundles (ventrolateral/ventral)associated cell somataFMRF-like IRexistence: (0) absent; (1) present—type Turbanella cf. hyalina; (2) present—type Tubiluchus troglodytes (scattered along the neurites).

     
  27. 27.

    Paired/unpaired longitudinal neurite bundles (ventrolateral/ventral)associated cell somata5-HT IRexistence: (0) absent; (1) present—type Convolutriloba longifissura.

     
  28. 28.

    Posterior neuronal somata (=“Caudal ganglion”?)existence: (0) absent; (1) present—type Neodasys chaetonotoideus.

     
  29. 29.

    Posterior neuronal somata (=“Caudal ganglion”?).shape: (0) unpaired, median structure—type Gnathostomula paradoxa; (1) paired structure—type Neodasys chaetonotoideus; (2) unpaired—type Tubiluchus troglodytes; (3) preanal ganglion—type Caenorhabditis elegans.

     
  30. 30.

    Posterior neuronal somata (=“Caudal ganglion”?)5-HT IR cellsexistence: (0) absent; (1) present—type Neodasys chaetonotoideus; (2) present—type Tubiluchus troglodytes.

     
  31. 31.

    Accessory longitudinal neurite bundle (ventro-median, unpaired)existence: (0) absent; (1) present—type Gnathostomula peregrina.

     
  32. 32.

    Longitudinal neurite bundles (dorsal)existence: (0) absent; (1) present.

     
  33. 33.

    Longitudinal neurite bundles (dorsal)number of pairs: (0) one pair—type Xenodasys riedli; (1) two pairs—type Convolutriloba longifissura; (2) more than four pairs—type Tubiluchus troglodytes.

     
  34. 34.

    Longitudinal neurite bundles (lateral)existence: (0) absent; (1) present—type Cephalodasys maximus.

     
  35. 35.

    Longitudinal neurite bundles (lateral)number of pairs: (0) one pair—type Gnathostomula peregrina; (1) more than two pairs—type Tubiluchus troglodytes.

     
  36. 36.

    Longitudinal neurite bundles (dorso-median, unpaired)existence: (0) absent; (1) present—type Gnathostomula peregrina.

     
  37. 37.

    Longitudinal neurite bundles (ventral, paired)existence: (0) absent; (1) present—type Macrostomum pusillum.

     
  38. 38.

    Longitudinal neurite bundles (ventral, paired)number of pairs: (0) one pair—type Macrostomum pusillum.

     
  39. 39.

    Peripheric neurite plexusexistence: (0) absent; (1) present—type Macrostomum pusillum.

     
  40. 40.

    Central nervous systemorthogonal pattern (longitudinal plus circular bundles of neurites)existence: (0) absent; (1) present.

     

Notes on terminology

We are fully aware that there exist numerous definitions for a nerve. For instance, according to Bullock and Horridge (1965), a nerve is a bundle of affarent and/or efferent neurites, while Romer and Parson (1991) further define a nerve by presence of a common sheath of connective tissue that surrounds the bundle. Because afferent and efferent neurites are not readily distinguishable, and connective tissue sheaths are not known from all invertebrate taxa (Abbot 1995), we choose to use the term “neurite bundle” to distinguish a common collection of neurites regardless of their structure and pathway. In general, we will follow the terminology for neuroanatomical structures in invertebrates by Richter et al. (2010).

Notes on immunohistochemistry

For the description of the distribution of the neurotransmitters, we want to state, that all the results depend on the specificy of the antibodies, which were used. We have used polyclonal antibodies raised against 5-HT and FMRFamide. The observed immunoreactivity (IR) may be perceived for anti-5-HT as a serotonin IR or 5-HT IR and for the anti-FMRFamide as an RFamide-like IR, unless this is not stated explicitly. The term RFamide-like IR is due to the recognition of the conservative RF-sequence by the used antibody, the outcome of this is that the antibody is labelling the members of the RFamide family, or the RFamide-related (neuro)peptides (FaRPs), in general (e.g. Price and Greenberg 1989; Maule et al. 1999).

We will not use the terms serotonergic or FMRFamidergic for IR structures, because this implies a testing of the function of 5-HT and FMRFamide as neurotransmitters in 5-HT and RFamide-like IR cells. With our method, we are not able to verify this; we are only able to show the presence of the respective antigen, but not the presumed function.

Results

First, we describe the general pattern of the nervous system of Neodasys chaetonotoideus, as reconstructed from all labelling and observation techniques (i.e. antibody labelling of 5-HT, FMRFamide, tubulin and histochemical staining of DNA and musculature in combination with CLSM; serial ultrathin sections in combination with transmission electron microscopy). In the description of the brain (=“cerebral ganglion” sensu Hochberg 2007b), we use percentage length units in relation to the pharynx (UP). Such percentage length units are routinely used in descriptions of gastrotrichs for the entire antero-posterior axis, e.g. Todaro and Hummon (2008). We refer here to the total length of the pharynx as 100 units (the club-shaped mouth tube not considered). However, the positions of certain structures are approximations; distinct distances and measurements as well as morphological details are given in the subsequent chapters. There might be slight variations in these morphometric values between the different specimens examined for this study. The terms positive immunoreaction and immunoreactive are abbreviated in the text as IR.

Gross neuroanatomy of N. chaetonotoideus

The nervous system of N. chaetonotoideus consists of a commissural brain with a strong dorsal and two weak ventral commissures in anterior and posterior position relative to the dorsal commissure. The dorsal commissure and the posterior ventral commissure form a more or less ring-like structure of neurites around the pharynx at UP 60 (Figs. 1b, 3a, 4). Bilateral clusters of neuronal cell somata are present lateral and dorso-lateral of the pharynx (UP 35 to UP 85) (Figs. 1b, 2a, b, 3a, 7b, c, 8b–e). However, cell bodies are not homogenously distributed along the longitudinal axis of each hemisphere (Fig. 7b, c, n). The highest density of somata of neurons is situated in the region posterior to the dorsal commissure (dco). Here, both hemispheres of the brain approach each other, thus surrounding the pharynx nearly completely leaving open the ventral sector only (Fig. 8e). Ventrolaterally, two posteriorly directed bundles of neurites arise from the commissure of the brain (Figs. 1b, 5g). The pair of ventrally positioned longitudinal neurite bundles (ln) spans the whole animal up to the caudal pedicles. At the caudal pedicles, each of the bundles branches into numerous fine fibres (Figs. 1a, 5q, s). A few micrometres frontal to this branching, there is one single pair of somata with anti-5-HT immunoreactivity closely associated with the longitudinal neurite bundles. This pair of somata is called here the 5-HT IR posterior somata (pss) (Figs. 3b, 6a, h, i, l). Serially, there are thin lateral and dorsal neurites branching from each longitudinal neurite bundle (Figs. 1a, b, 5q, u). Each such neurite is connected with a ciliated receptor cell within the epidermis. The lateral ones are associated with the lateral adhesive organs (adhesive tubes) (Fig. 5q). The number varies considerably among the investigated specimens.
Fig. 1

Neodasys chaetonotoideus. Schematic reconstruction of the nervous system, compiled from different techniques and labellings, dorsal view. a Habitus showing the gross neuroanatomy, b anterior trunk (pharyngeal region) with details of the brain. an-Ian-III anterior neurite bundles of the brain I–III, asc anterior sensory cells, avco anterior ventral commissure, bc buccal cavity, br brain, dco dorsal commissure, dcoa anterior neurite bundle of the dorsal commissure, dcop posterior neurite bundle of the dorsal commissure, dsc dorsal sensory cells, dpn dorsal pharyngeal neurite bundle, in intestine, lat lateral adhesive tubes, lsc lateral sensory cells, ln longitudinal neurite bundle, lpn lateral pharyngeal neurite bundle, mdn mediodorsal neurite(s) of the brain, ph pharynx, pl pharyngeal lumen, pss posterior 5-HT IR soma (=“caudal ganglion”?), pvco posterior ventral commissure. Black arrow: pharyngeal commissure connecting unpaired dpn and paired lpn. Black triangles indicate sectional planes for the cross-sections of Figs. 2 (schemes) and 7 (TEM-sections)

Fig. 2

Neodasys chaetonotoideus. Schematic cross-sections of the nervous system at different levels (sectional planes indicated in Fig. 1b). a Anterior part of the brain with different frontal longitudinal bundles of neurites. b Brain at the level of the dorsal commissure. c Posterior part of the pharyngeal region with the pair of posterior longitudinal neurite bundles accompanying the longitudinal muscle strands and cell bodies of the epidermis. an-Ian-III anterior neurite bundles of the brain I–III, br brain, cut cuticle, dco dorsal commissure, dpn dorsal pharyngeal neurite bundle, epi enwrapped cells of the ventral epidermis, glc glia-like cells, lci locomotory cila, lm longitudinal muscle strands, ln longitudinal neurite bundle, lpn lateral pharyngeal neurite bundle, ph pharynx

Fig. 3

Neodasys chaetonotoideus. Schematic reconstruction of the anti-5-HT (serotonin) IR nervous system. a Anterior trunk (pharyngeal region) with 5-HT IR somata and neurites associated to the brain. b Posterior trunk region. afs anterior longitudinal 5-HT IR fibre, avcos anterior ventral commissure, br brain, dass dorsal anterior 5-HT IR cell, dcoas anterior 5-HT IR fibres of the dorsal commissure, dcops posterior 5-HT IR fibres of the dorsal commissure, \( dps_{s}^{d} \) dorsal posterior 5-HT IR soma (more dorsally), \( dps_{s}^{v} \) dorsal posterior 5-HT IR soma (more ventrally), in intestine, lfs longitudinal 5-HT IR fibres, \( ls_{s}^{d} \) lateral 5-HT IR soma (more dorsally), \( ls_{s}^{v} \) lateral 5-HT IR soma (more ventrally), pncs posterior neurite bundle connection, pss posterior 5-HT IR soma (=“caudal ganglion”?), pp posterior pedicle

Frontal to the dorsal commissure, there are three pairs of longitudinal bundles of neurites spanning the head region (Figs. 1a, b, 2a). The most dorsal pair, anterior longitudinal bundle of neurites I (an-I), starts in the region of the dorsal commissure (UP 60) with numerous fine branches (Figs. 1b, 5b, 8a), some of which fuse with the neuropil of the dorsal commissure (Fig. 5d). Frontally, an-I runs up to UP 10 where it branches again into fine fibres that obviously innervate the anterior sensory cells. Around UP 40, a thin commissure ventral of the pharynx connects the neurite bundle an-I of both sides (Fig. 1b). Slightly more ventrolateral than an-I is the second pair of “cephalic” neurite bundles, the anterior longitudinal bundle of neurites II (an-II) (Figs. 1b, 5b, g). It starts at UP 70, also branches into fine fibres, parts of which are fused with the dorsal commissure. Frontally, an-II runs up to UP 15 where it branches into fine fibres innervating the anterior sensory cells (Fig. 8a). A third pair of frontally directed neurites, anterior longitudinal bundle of neurites III (an-III), arise directly from the lateral region of the dorsal commissure at UP 60 (Figs. 1b, 5b, g). Each bundle runs frontally a few micrometres until it fuses with an-II around UP 45. In addition to the paired bundle an-I to an-III, there is an unpaired dorsal neurite(s) (mdn), which has its origin in a pair of nervous cells positioned around UP 70 within the medial side of each hemisphere of the brain (Figs. 1b, 5b, h). Short fibres of both cells fuse at approximately UP 65, the unpaired bundle of neurites could be traced up to UP 30.

Apart from these major bundles of neurites, we detected several neurite bundles accompanying the pharynx of Neodasys chaetonotoideus. As ultrastructural investigation has shown (see below for details), all three pharyngeal neurite bundles have a basi-epithelial position between the myoepithelial cells of the foregut (Figs. 2a–c, 9a). There is an unpaired dorsal pharyngeal neurite bundle (dpn) and one pair of lateral pharyngeal neurite bundles (lpn) (Fig. 2c). All of them range from the region of the pharyngeo-intestinal junction (at UP 100) to the region of the buccal cavity (UP 10). The unpaired, mediodorsal dpn branches into two fine neurites in the area of the buccal cavity (Figs. 1b, 5i). Here, two ciliary sensory cells are associated with the dorsal pharyngeal neurite bundle. Along its entire length, we observed up to nine immunoreactive (anti-acetylated α-tubulin and some of them additionally anti-RFamide-like positive) swellings. Studying the serial ultrathin sections with TEM revealed that these swellings are monociliary receptor cells (Fig. 10a–d). Furthermore, there are at least three pairs of ciliary sensory cells associated with the paired lpn. All of these sensory cells are situated in the anterior half (UP 15 to UP 40) of the pharynx; they are monociliary cells (Fig. 9g).

General comments on immunolabelling of the nervous system

First, we have to point out that we have used antibodies against acetylated α-tubulin and against β-tubulin. The results were quite similar between the two members of the tubulin family, but the labelling effort was better by the use of the antibodies against acetylated α-tubulin. Antibodies against β-tubulin produce more intracellular background in non-nervous tissue than the antibodies against acetylated α-tubulin, but in general, we found no significant differences between the labelled nervous structures.

Due to the occurrence of tubulin, especially in an accumulation of tubulin in the neurites as neurotubuli, we are able to describe here the general innervation pattern of N. chaetonotoideus. We found massive concentrations of anti-tubulin IR material (bundles of neurites) in the anterior part of the animal, starting from the area of the brain and projecting in an anterior direction and a large paired bundle alongside the whole trunk in a ventrolateral position, the main lateral longitudinal fibre bundles (ln). Those are obviously based on the location associated with epidermal cells with motile cilia (Figs. 2c, 5a, q, s-u, 8f) and the ventral longitudinal musculature (Figs. 2c, 5h, 9c, d). Furthermore, anti-tubulin IHC is labelling the entirety of ciliary structures, this includes the ventral motile cilia (Fig. 5a, q, t), the sensory cilia (Fig. 5q, u), the axonemata of the spermatids (Fig. 5q, r) and the adhesive tubes (Fig. 5q).

In general, the nervous system of N. chaetonotoideus shows a strong IR against 5-HT and FMRFamide antisera. We have found a 5-HT immunoreaction in several somata of the brain and multiple longitudinal fibres posterior and anterior of the brain (Figs. 3a, b, 6a, j). The 5-HT-positive somata are not exclusively located in the area of the brain; additionally two 5-HT-positive somata have been found at the posterior end of the trunk associated with the longitudinal nervous fibres (Figs. 3b, 6a, h, i, q). In the case of anti-FMRFamide labelling, the nervous system of N. chaetonotoideus shows a strong anti-RFamide-like IR within the two lateral clusters of neuronal somata of the brain and in several longitudinal fibres, posterior and anterior of the brain (Figs. 4, 7a). The IR somata are exclusively located in the area of the brain (Fig. 7a–c, n), with the exception of some RFamide-like IR sensory cells of the pharynx (Fig. 7h, i, k). In addition, some (weak) IR soma-like structures could be found associated with the posterior longitudinal fibres (Fig. 7b), as shown by Hochberg (2007b) in Neodasys cirritus. Comparable structures could be found only in freshly prepared material (storage under 1 week) from samples of sediment enriched with detritus. In this connection, it has to be mentioned that the fluorescence was very low. This resulted an adequate documenting of those cells by cLSM, because the signal quickly bleached. Moreover, we have not found any hint of neuronal somata in this position laterally of the main neurite bundle by TEM.
Fig. 4

Neodasys chaetonotoideus, anterior trunk (pharyngeal region). Schematic reconstruction of the RFamide-like IR nervous system of Colouration in slade blue indicates anti-RFamide-like IR structures, and in green indicates anti-tubulin IR structures. admsRF RFamide-like IR soma anterior dorsomedian of the dorsal commissure, afRF anterior RFamide-like IR fibre(s), an-Ian-III anterior neurite bundles of the brain I to III, anbsRF anterior RFamide-like IR soma of the anterior neurite bundles, br brain, dcoaRF anterior bundle of RFamide-like IR fibres of the dorsal commissure, dcopRF posterior bundle of RFamide-like IR fibres of the dorsal commissure, dscpRF anteriodorsale RFamide-like positive sensory cell of the pharynx, dtsbr dorsale tubulin IR somata of the brain, dpnRF RFamide-like IR fibre(s) of dpn, in intestine, lfRF posterior RFamide-like IR fibres, mo mouth opening, pdsRF posteriodorsal RFamide-like positive soma of the brain, pdmsRF RFamide-like IR soma posterior dorsomedian of the dorsal commissure, ph pharynx, ppcRF RFamide-like IR posterior pharyngeal (sensory?) cell(s), pvcoRF RFamide-like IR fibre of the posterior ventral commissure (pvco) of the brain

The brain

The brain consists of the interconnections among the lateral clusters of neuronal somata, the dorsal commissure (dco) and the fine ventral commissure(s). All components show an IR against 5-HT, RFamide and tubulin. TEM studies strongly support the characteristic bilateral-symmetric, dumb-bell-like appearance of the brain.

The dorsal commissure

The dorsal commissure (dco) of the brain is a very large mass of neurites (dorsal of the pharynx 5 μm high and with a width of more than 12.5 μm) (Fig. 2b). Not all of the neurites sectioned for TEM show a strictly orthogonal course from one hemisphere to the other (Figs. 8d, 9a, b). There are as well several neurites, which are cross-sectioned, indicating a longitudinal or diagonal course within the commissure. However, para-median sections of a sagittally sectioned specimen show that neurites in the centre of the dorsal commissure indeed stretch from one-half of the brain to the other. Mitochondria are visible in some neurites, and there is a varying density of small vesicles (50 nm in diameter) and even smaller circular structures (25 nm in diameter), obviously neurotubuli. Due to the high number of sectioned neurites and the differing orientation of filaments, we suppose a certain degree of neuronal networks within the dorsal commissure.

The dorsal commissure (dco) is well labelled by immunohistochemistry (Fig. 5h, m). The density of fibres is heterogeneous; it seems to have a maximum in the anterior part of the dorsal commissure (dcoa) and a second maximum (dcop) at the posterior margin (Figs. 1b, 5b). The diameter of the anterior dense region is ~1.5 μm, the one of the posterior 0.6–0.8 μm. In the space between these dense accumulations of fibres, additional fibrous material is more or less loosely distributed. At the lateral sides, where the commissural fibres run into the cluster of neuronal somata, the distribution becomes more homogeneous in density. It seems that the majority of the fibres of the anterior fibre accumulation runs posterior in a lateral position and merges with the posterior dense fibre accumulation. The resulting bundle of fibres (diameter ~3 μm) runs in posterioventral direction and constitutes the lateral longitudinal neurite bundles (Fig. 5g).
Fig. 5

Neodasys chaetonotoideus. Visualization of anti-tubulin (acetylated α-tubulin) IR (cLSM). a, b, g, i, q, s, t Anti-tubulin IR colour coded by depth (CCD-projections), a with an overlay of the transmission channel. cf Single optical sections of an overlay of a double labelling of anti-tubulin (green) and a nuclei counterstain (DAPI) (white). h, r, u Simulated fluorescence projection (SF-projections) of the anti-tubulin IR (green), h with a double labelling of tubulin (green) and a muscular staining (phalloidin) (red). j–p Orthogonal projections (OP) of h. Anti-tubulin IR (green), musculature staining (phalloidin) (red) and nuclei (DAPI) (white). jo cross-sections; p sagittal section. a Overview of anti-tubulin IR components within brain, longitudinal bundles of neurites, ciliation and testes. b Overview of the brain from dorsal with projections of the brain. c, f Details of area of the brain. g Overview of the brain from a lateral view with projections of the brain; the black arrowheads indicate the connections between an-I and an-II. h, i Anterior part from dorsal; dashed lines indicate planes of orthogonal projections, in i, white arrowheads indicate branching of ventrolateral neurite bundles of the pharynx. j–o Anterior trunk region, cross-sections of. Blue asterisk ventrolateral pharyngeal bundles of neurites, white asterisk dorsal pharyngeal bundle, yellow asterisk dorsal bundle of neurites of the brain, yellow dashed circle cluster of nuclei of the ventrolateral ciliated somata, yellow arrowheads sensory cells of the pharynx. p Sagittal section of the anterior trunk region; yellow arrowheads sensory cells of the pharynx. q Posterior trunk region with allospermatophore. r Detail of allospermatophore. s Detail of the posterior neurite bundle connection, ventrolateral view. t Detail of the association between lateral nervous neurite bundles and ventral bands of motile cilia. u Detail of lateral nervous neurite bundle and innervation of lateral adhesive tubes and associated sensory cells; white arrowheads lateral neurite bundles between main longitudinal neurite bundle and laterally sensory cells.an-Ian-III anterior neurite bundles of the brain I–III, br brain, dco dorsal commissure, dcoa anterior bundle of neurites of the dorsal commissure, dcop posterior bundle of neurites of the dorsal commissure, dtsbr dorsal tubulin IR cell of the brain, dpn dorsal pharyngeal neurite bundle, epi nuclei of epidermis cells, fcor fronto-caudal organ, lat lateral adhesive tube, lci ventral locomotory cilia, ln longitudinal neurite bundle, lsc lateral sensory cell, lpn lateral pharyngeal neurite bundle, mdn mediodorsal neurite(s) of the brain, pat posterior adhesive tubes, ph pharynx, pnc posterior connection of the neurites of ln, ppc posterior pharyngeal cell, sph spermatophore, te testes, vlm ventral longitudinal musculature

Within the dorsal commissure, the serotonin-expressing fibres are located in the posterior half of the neuropil (Fig. 6m). The 5-HT-positive fibres of the dorsal commissure (dcos) form a solid arc on the upper half of the pharynx. The diameter of the 5-HT-positive fibre bundle amounts on average 2.5–3.5 μm. This bundle shows a compartmentalization in an anterior wider part (diameter ~2 μm) and a posterior fine fibre (diameter ~1 μm). An additional diffuse anti-5-HT signal occurs in some specimens anterior of this region (Fig. 6b, d), but this does not form a continuous trans-pharyngeal fibre. The solid part of the dorsal commissure covers the upper third of the pharynx (Fig. 6k); the arc length was estimated to be approximately 30 μm. At both lateral ends, the more or less solid fibre arrangement branches into an anterior and a posterior part; the posterior one runs in ventrolateral direction to the ventral longitudinal neurite bundle (Fig. 6r), while the anterior fibre runs in anteroventral direction. By means of this course, the two connecting 5-HT IR fibres form a nearly right angle in a lateral view (Fig. 6c). At the point where the anterior connecting fibre joins the ventral longitudinal fibre (afs), a fine ventral 5-HT-positive commissure (avcos) originates and connects the lateral 5-HT-positive longitudinal fibres (Figs. 3a, 5a); this fibre shows an anti-tubulin IR signal, too (Fig. 6o). The diameter of the fibre was estimated to be 0.5 μm, but approximately 3 μm from the connection to the longitudinal fibre a 5-HT-positive thickening of the fibre occurs, with a diameter of up to ~2.5 μm (Fig. 6b). Because of the small diameter of the fibre, we were not able to detect it in all specimens, but the thickening was detectable in nearly all specimens.
Fig. 6

Neodasys chaetonotoideus. Visualization of anti 5-HT IR (cLSM). a, b, h Anti-5-HT IR colour coded by depth (CCD-projections), in a with an overlay of the transmission channel. c, d Simulated fluorescence projection (SF-projections) of the anti-5-HT IR (grey). eg Single optical sections of an overlay of a double labelling of anti-5-HT IR (red) and nuclei counterstain (DAPI) (white). j, l Maximum projection of an overlay of double labelling of anti-5-HT IR (white) and staining of the musculature (phalloidin) (red). k Orthogonal projection of j. m–r Single optical plane of a triple labelling of 5-HT (red), acetylated α-tubulin (green) and nuclei (DAPI) (white). a Overview with labelled anti-5-HT IR components within brain and longitudinal neurites. b Overview of the brain from dorsal with dorsal commissure, 5-HT-positive somata of the brain and 5-HT-positive ventral commissure in front of the brain. c Lateral view of left hemisphere of 5-HT-positive components of the brain. d Dorsal view of 5-HT-positive structures of the brain. e Detail of the pair of 5-HT-positive dorso-posterior somata of the brain. f Detail of 5-HT-positive dorso-anterior soma of the brain. g Detail of 5-HT-positive lateral somata of the brain. h Overview of anti 5-HT IR structures at the posterior end, with 5-HT-positive posterior soma and posterior neurite bundle connection. i Detail of one 5-HT-positive posterior neuronal soma. j Overview of entire 5-HT-positive nervous system (white) in relation to musculature (red); lateral longitudinal fibres are closely related to ventrolateral longitudinal muscles. k Detail of dorsal commissure, which runs arc-like over the pharynx. l Detail of 5-HT-positive nervous system (white) at the posterior end in relation to musculature (red). m Overview of anterior trunk region, indicating distribution of 5-HT-positive components esp. at the posterior part of the brain. Yellow asterisk strong anti-tubulin IR cells at terminal end of dorsal bundle of neurites of pharynx, white dashed circle area of lateral cluster of neuronal somata. n Detail of 5-HT-positive fibre distribution within the lateral longitudinal nervous neurite bundle at posterior end. White arrowheads exact position of the single 5-HT-positive fibres, yellow arrowheads innervation of lateral adhesive tubes and associated sensory cilia. o Detail of 5-HT-positive ventral commissure and co-localized anti-tubulin IR. p Brain at a median plane, detail. Ventral somata of dorso-posterior somata of the brain. q Detail of 5-HT distribution in the dorsal commissure, lateral view, dashed circle area of dorsal commissure. r Detail of 5-HT-positive components (red) of the brain in relation to tubulin IR innervation pattern (green), lateral view. afs anterior longitudinal 5-HT IR fibre, an-Ian-III anterior neurite bundles of the brain I–III, avco anterior ventral commissure between an-I, avcos 5-HT IR fibre of the anterior ventral commissure between an-I, br brain, brs 5-HT IR components of the brain, dass dorsal anterior 5-HT IR soma, dco dorsal commissure, dcos 5-HT IR components of the dorsal commissure, \( dps_{s}^{d} \) dorsal posterior 5-HT IR soma (more dorsally), \( dps_{s}^{v} \) dorsal posterior 5-HT IR soma (more ventrally), epi nuclei of epidermis cells, in intestine, lat lateral adhesive tube, lci ventral locomotory cilia, lfs longitudinal 5-HT IR fibre(s), ln longitudinal neurite bundle of neurites, \( ls_{s}^{d} \) lateral 5-HT IR soma (more dorsally), \( ls_{s}^{v} \) lateral 5-HT IR soma (more ventrally), lsc lateral sensory cell, ph pharynx, plfis inner posterior longitudinal 5-HT IR fibre, plfos outer posterior longitudinal 5-HT IR fibre, pncs posterior 5-HT IR neurite connection, pss posterior 5-HT IR soma (=“caudal ganglion”?), te testes, vlm ventral longitudinal musculature

The RFamide-like IR components of the dorsal commissure consist of two well-separated strands of IR fibres, a fine anterior and a broader posterior bundle (Fig. 7c, f). The posterior fibres extend 4–5 μm in anterior–posterior direction; the anterior fibre bundle has a diameter of approximately 1–2 μm (Fig. 7m). The posterior bundle consists of two, in the dorsalmost part well separated, sub-bundles (Fig. 7l). The gap between the RFamide-like IR bundles is approximately 2 μm. In the dorsalmost part of the neuropil arc, anterior and posterior of the commissure, no neuronal somata are located; approximately 12 μm of the anterior and 7–10 μm of the posterior part of the dorsal commissure are free of neuronal somata. We did not find here any associated RFamide-like positive somata in vicinity of this part of the dorsal commissure.
Fig. 7

Neodasys chaetonotoideus. Visualization of anti-RFamide-like IR by cLSM in. a–c Anti-RFamide-like IR colour coded by depth (CCD-projections), in a with an overlay of the transmission channel. d–g, p Single optical sections of an overlay of a triple labelling of anti-(FM)RFamide (blue), anti-acetylated α-tubulin (green) and a nuclei counterstain (DAPI) (white). h, n, q Simulated fluorescence projection (SF-projections) of the anti-(FM)RFamide-like IR (blue), in h as double labelling with anti-acetylated α-tubulin (green), in q with double labelling of musculature (red). i–k, l, m Orthogonal projection of h. o Orthogonal projection of pharyngeal area posterior of the brain (comparable to k), shown as triple labelling: RFamide-like IR (blue), anti-acetylated α-tubulin IR (green) and musculature (red). a Overview with anti-RFamide-like positive components within brain and longitudinal neurites. b Overview of anterior region. Note roundish RFamide-like positive structures (white arrowheads) laterally of longitudinal fibres. c Detail of the brain. dg Single sections of dorsal hemisphere of the brain, red dashed line RFamide-like IR posteriolateral somata of the brain. h Detail of the brain, white asterisks indicate anterior RFamide-like IR somata of anterior neurite bundles (anbsRF). i Cross-section anterior of the dorsal commissure with detail of RFamide-like positive dorsal sensory cell of the pharynx (indicated by yellow arrowhead cilium, red arrowhead RFamide-like IR soma. j Cross-section of dorsal commissure. Note dorsal pharyngeal bundle of neurites ventral of the dorsal commissure and mediodorsal bundle of neurites (mdn) on dorsal side. k Cross-section posterior of the dorsal commissure with details of dorsal pharyngeal bundle of neurites and mediodorsal bundle of neurites. l Median sagittal section of the brain with anterior (yellow dashed circle) and posterior (white dashed circle) RFamide-like positive bundle of neuropile fibres of the dorsal commissure. Note rarely occurring somata anterior and posterior of dorsal commissure. m Paramedian sagittal section of the brain with anterior (yellow dashed circle) and posterior (white dashed circle) RFamide-like positive bundle of fibres of dorsal commissure neuropile. Note dorsal sensory cells (yellow arrowheads) of the pharynx. Anterior of the dorsal commissure somata of sensory cells show RFamide-like positive signal (red arrowhead), which is not shown bysoma of sensory cell posterior of the dorsal commissure (blue arrowhead). n Lateral view of the brain, area of RFamide-like IR cluster of somata (green dashed line). Clearly visible are two anterior RFamide-like IR somata (white asterisks) of anterior neurite bundles. o Cross-section posterior of dorsal commissure. Note cilium of the dorsal pharyngeal sensory cell (yellow arrowhead) within pharyngeal lumen and dorsal bundle of neurites of the pharynx embedded within pharyngeal musculature. p Detail of the ventral part of the anterior trunk region with ventral commissure anterior of the brain connecting both an-I and the RFamide-like positive commissure of the brain. q Overview of posterior trunk region. The longitudinal RFamide-like fibres are closely associated to the ventral longitudinal musculature. admsRF RFamide-like IR anterior dorsomedian soma of the brain, afRF anterior longitudinal Rfamide-like IR fibre, an-Ian-II anterior neurite bundles of the brain I–II, anbsRF RFamide-like IR somata of the anterior neurite processes of the brain, avco anterior ventral commissure between an-I anterior of the brain, brRF RFamide-like IR of the brain, dcoRF RFamide-like positive fibres of the dorsal commissure, dcoaRF anterior RFamide-like positive bundle of the dorsal commissure, dcopRF posterior RFamide-like positive bundle of the dorsal commissure, dlm dorsal longitudinal muscle strand, dtsbr dorsomedian tubulin IR somata of the brain, dpn dorsal pharyngeal neurite bundle, dpnRF RFamide-like IR fibre of the dorsal pharyngeal neurite bundle, lci locomotory cilia, lfRF RFamide-like IR fibre(s) of the ventral longitudinal neurite bundle (ln), lpn lateral pharyngeal neurite bundle, mdn mediodorsal bundle of neurites of the brain, pdsRF RFamide-like IR posterior dorsal soma of the brain, pdmsRF RFamide-like IR posterior dorsomedian soma of the brain, pdmsRF1 RFamide-like IR more anterior posterior dorsomedian soma of the brain, pdmsRF2 RFamide-like IR more posteriorventral posterior dorsomedian soma of the brain, ph pharynx, pp posterior pedicle, ppcRF RFamide-like IR posterior pharyngeal cells, pvcoRF Rfamide-like IR fibre of pvco, scm somatic circular musculature, vlm ventral longitudinal musculature

The lateral clusters of neuronal somata

On both sides of the dorsal commissure is a cluster of neuronal somata (Figs. 2a, b, 8b–e, 9a) with the highest abundance of somata slightly posterior to the dorsal commissure. However, somata belonging to the brain can be found along more than 50% of the total length of the pharynx (approximately from UP 35 to UP 85). The nervous cells vary slightly in dimensions (3–4 μm in diameter) and contain big, globular and highly active nuclei (2 μm in diameter). There are numerous mitochondria in the periphery of the cells; the cytoplasm is rather electron-lucent with granular content and small vesicles (50 nm in diameter). Within these clusters, several somata show RFamide-like IR (Fig. 4) and few 5-HT IR (Fig. 3a).
Fig. 8

Neodasys chaetonotoideus. Ultrathin cross-sections (TEM) through the nervous system. Levels of sectional planes in Fig. 1b. a Branching of anterior longitudinal neurite bundles (asterisks) at the level of the buccal tube. b Main anterior longitudinal neurites. c Anterior part of the brain with all three anterior longitudinal neurite bundles. d Section of the brain at the level of the dorsal commissure. e Posterior part of the brain with longitudinal bundle of neurites. f Posterior longitudinal neurite bundle at the level of the pharyngeo-intestinal junction. anIan-III anterior neurite bundles of the brain I–III, bc buccal cavity, br brain, dco dorsal commissure, epi cell bodies of the enwrapped epidermis, glc glia-like cells, lci locomotory cilia, lm longitudinal muscles, ln longitudinal neurite bundle, nc nerve cell(s), ph pharynx, ph/mig pharyngeo-intestinal junction (transition between pharynx and midgut), sci sensory cilia

Fig. 9

Neodasys chaetonotoideus. Ultrastructure (TEM) of the nervous system. a Section of the brain at the level of the dorsal commissure. Note the pharyngeal neurite bundles. b Close-up of the dorsal commissure with numerous neurites filled with vesicles and neurotubuli (see, e.g. cells marked by asterisks). c, d Sections of the longitudinal bundle of neurites at different levels. c Pharyngeal region, d mid-trunk region. e, f Sections of the myoepithelial pharynx at different levels showing the there pharyngeal bundles of neurites. g Close-up of a ciliary receptor of the pharynx. Note collar of microvilli surrounding the cilium. anIan-II anterior neurite bundles of the brain I–II, br brain, cr ciliary rootlets, cut cuticle, dco dorsal commissure, dpn dorsal pharyngeal neurites, epi epidermis, glc glia-like cells, in intestine, ln longitudinal neurite bundle, lpn lateral pharyngeal neurites, mc muscle cells, sc sensory cilium, ph pharynx, phl pharyngeal lumen

The brain contains five pairs of 5-HT IR somata in a bilateral-symmetric arrangement (Fig. 6b, c, d). All cells are connected with the dorsal commissure in a more or less lateral region of the neuropil. Three pairs are located in a dorsolateral position, posterior of the dorsal commissure. One of them, the most anterior pair (dass) is situated directly at the lateral border of the 5-HT-positive fibres of the dorsal commissure, where the fibres branch into an anterior and a posterior fibre. These fibres are connecting the dorsal commissure with the 5-HT-positive longitudinal fibres twice; anterior and posterior of the dorsal commissure (Fig. 6c, d). The soma of the 5-HT-positive anterior dorsal cell (dass) appears roundish with a diameter of approximately 3–4.5 μm (Fig. 6c, f, r). The somata of the two other pairs of the dorsal somata (dpss) are more posteriolaterally located (Fig. 6e, r, p). The somata appear drop shaped (Fig. 6c, d, p). Their position to each other is in a dorsoventral orientation, we can distinctly separate a dorsal soma (\( {\text{dps}}_{\text{s}}^{\text{d}} \)) and a more ventral one (\( {\text{dps}}_{\text{s}}^{\text{v}} \)) (Fig. 6c, e). In comparison with the other 5-HT IR cells, the somata are sizeable. The measurement from one specimen, in lateral orientation, shows in optical sagittal sections a size of 5.5 μm × 8.9 μm (longitudinal × dorsoventral dimension) and 5.5 μm × 7 μm for the upper pair. The more ventral one was measured with 8.9 μm × 7.1 μm and 5.3 μm × 7.1 μm. Both pairs send a process to the dorsal commissure in an anterior direction and are connected to the dorsal commissure in the area of the anterior dorsal soma (dass). The 5-HT-positive dorsal soma \( {\text{dps}}_{\text{s}}^{\text{d}} \) reaches the dorsal commissure directly in the range of the soma of the anterior dorsal soma (Fig. 6c); it is not possible to decide whether these somata are directly connected or whether the dorsal posterior one is connected with the neuropil beneath the anterior cell. The process of the more ventral located soma \( {\text{dps}}_{\text{s}}^{\text{v}} \) runs clearly to the neuropil laterally of the anterior soma dass (Fig. 6c, d).

Lateroventral of the dorsal commissure are two pairs of 5-HT-positive soma (lss) that occur within the angle of the anterior and posterior 5-HT-positive fibres (Fig. 6c). These somata seem to be connected with the dorsal commissural fibres via the 5-HT-positive longitudinal fibres (lfs) (Fig. 6b, c). We are confident that these structures are somata, because nuclei were recognizable within the IR swelling (Fig. 6g). The somata are in a dorsoventral orientation, so we can distinguish a 5-HT IR dorsolateral soma (lssd) and a ventrolateral one (\( {\text{ls}}_{\text{s}}^{\text{v}} \)) (Fig. 6c). The ventrolateral soma lies close to the ventral 5-HT-positive longitudinal fibre; the dorsal lateral cell follows in direct contact to the ventral one. The soma \( {\text{ls}}_{\text{s}}^{\text{d}} \) is connected with a fine process to the dorsal commissure in the area of the branching fibres laterally of dass.

In general, the neuronal somata within the brain are not indicated by a well-expressed anti-tubulin IR, except for one pair of strongly IR somata (Fig. 5b, h, i, 7h). This bilaterally arranged pair of anti-tubulin IR somata of the brain (dtsbr) is located approximately 10–13 μm posterior of the dorsal commissure, and the single somata are in a dorsolateral position; the distance between them averages 13 μm. The somata are drop shaped (6 μm in length and 5 μm in width). Each soma bears a strongly IR neurite (diameter between 0.5 and 0.7 μm) at the tapered anterior side; these are running in an obtuse angle to a dorsomedian position. The length of the solitary neurites is between 7 and 8 μm, before they approach each other and run in anterior direction. When the fibres have met, the single fibres become undistinguishable. The diameter of the resulting fibre (or two fibres?), the mid-dorsal neurite(s) of the brain (mdn), is between 1.2 and 0.5 μm. It is oriented orthogonally on the dorsal commissure (Fig. 5m, 7j) and runs further in anterior direction. The fibre lies distally and appears surrounded by the nuclei of the epidermal cells (Fig. 5c). In some specimens, the fibre is slightly displaced from the mid-dorsal position (Fig. 5c, 7d), but in other individuals the position is totally mid-dorsal and parallel to the dorsal bundles of neurites of the pharynx (Fig. 5b, c, 7j, k). In total, the fibre(s) (mdnbr) runs approximately 50–55 μm in anterior direction, next to the area of the dorsal commissure the fibre(s) lies directly on the top of the musculature of the pharynx (Fig. 5k, l, 7o). The fibre(s) end directly beneath a pair of mid-dorsal cilia of the epidermis at UP 30–35. The position (and occurrence) of these cilia is confirmed by IHC (anti-tubulin labelling) and SEM investigations of N. chaetonotoideus (not shown here).

The general distribution of the RF-like IR structures in the brain, with a concentration of the neuronal somata in a lateral position to the dorsal commissure, is similar to the results of the 5-HT IR pattern. However, the number of RFamide-like IR somata is much higher. The brain consists of a bilateral cluster of RFamide-like positive somata connected by the dorsal commissure with RFamide-like IR fibres (Fig. 4, Fig. 7b, c, n). From a dorsal view, the brain appears horseshoe-like with the open side in posterior direction (Fig. 7b, c) and from lateral it has an ovoid appearance (Fig. 7n). This appearance is caused by the massive concentration of RFamide-like IR somata posterior of the dorsal commissure. The anterior and posterior dorsalmost part of the commissure is completely free of RFamide-like positive somata (for approximately 7–10 μm). The overall number of RFamide-like IR somata is obviously higher, as in the case of 5-HT-positive somata. We have attempted to approximate the number by counterstaining of the nuclei. Totally, we were able to count the number of cells in seven specimens. The lowest counted number was 22 cells per hemisphere of the brain and the highest number was 28 per hemisphere; we only count somata where the nucleus was completely embedded within the IR signal, to avoid the counting of non-IR soma, that are only close to RFamide IR somata located. Depending on this procedure, we expect rather an underestimation than an overestimation of the absolute number. The median number was 48 RFamide-like IR cells in the entire brain. Despite the variation of the total number (44–56), we were able to recognize within individual specimens consistent patterns. Those patterns are the following:
  1. (1)

    In the dorsalmost part of the brain lie two IR somata directly next to the fibres at the posterior side of the dorsal commissure (Fig. 7c, h). These posterior dorsomedian RFamide-like IR somata of the commissure (pdmsRF) are partly embedded in the fibrous material of the dorsal commissure. They occur always as a pair of somata at each hemisphere, but the relative position between the two somata was not always constant. In most cases, one soma was found in a dorsalmost position (pdmsRF1) (Fig. 7d) and the associated second soma (pdmsRF2) was located slightly more posterioventral (Fig. 7e), but in some cases the two somata were arranged laterally to each other.

     
  2. (2)

    At the anterior margin of the dorsal commissure one soma is located, the paired anterior dorsomedian RFamide-like positive soma (admsRF) (Fig. 7h). These cells are the median RFamide-like immunoreactive somata of the brain anterior of the dorsal commissure.

     
  3. (3)

    Another interspecific constant pattern was represented by the RFamide-like IR posterior dorsal somata (pdsRF), at least two pairs in each hemisphere of the brain (Fig. 7e). They are special for several reasons in comparison with the majority of the RFamide-like IR somata. These somata are obviously larger than the others, the position and the shape are reminiscent of the 5-HT-positive posterior dorsal somata. Additionally, these are some of the few RFamide-like positive cells, where the neurites are clearly visible (Fig. 7e). The somata are located approximately 13 μm posterior of the dorsal commissure in a dorsolateral position. The two somata of each hemisphere lie laterally to each other more or less solitary. They are not in close contact to other anti-RFamide-like IR cells. The somata are drop shaped, with a length of ~6–8 μm and a width of 4–6 μm. The anterior part of the soma appears cuspidate, depending on the emerging neurite. The process of the soma runs to the dorsolateral part of the dorsal commissure and enters there into the neuropil, thereafter the neurite becomes undistinguishable from the other RFamide-like IR fibres. Medially of these two pairs of somata are the paired anti-tubulin IR somata of the brain (dtsbr) (Fig. 7e). It seems that at least the somata of the inner pair of the two RFamide-like IR pairs are in contact with the somata of the anti-tubulin IR somata (Fig. 7e).

     
  4. (4)

    At the posteriolateral part of the brain, we found a consistent arrangement of almost 4, sometimes 5, RFamide-like positive somata in a row, we name these the group of RFamide-like IR posteriolateral somata of the brain (Fig. 7e, f). This represents the posteriolateral border of the RFamide-like IR somata of the brain.

     
  5. (5)

    In all studied specimens, an RFamide-like positive process runs out of the brain in anterior direction at the most anteriorly part of the cluster of somata. It seems that at this process always an RFamide-like IR soma (anbsRF) is related (Fig. 7f, g, h).

     

The posterior ventral commissure

At the posterior part of the brain, approximately 10 μm behind the dorsal commissure (dco), a fine anti-tubulin IR fibre (diameter less than 0.5 μm) connects the two clusters of cerebral somata at the ventral side. Depending on the origin of the fibre, directly at the posterior end of the brain, we name this structure the posterior ventral commissure of the brain (pvco) (Fig. 1b, 7p). At least one fibre is also anti-RFamide IR (Fig. 4, 7b, p); in that case we named it pvcoRF. We regard this bundle of neurites as a part of the brain, because it has its origin in the area, where the RFamide-positive fibres of the longitudinal neurite bundle leave the brain.

Glia-like cells of the brain

The study of ultrathin sections uncovered a conspicuous tissue that probably is associated to the (central) nervous system of Neodasys chaetonotoideus: a layer of two or three strongly flattened cells (with widths between 0.1 and 0.5 μm), which surround and line the neurons and neurites of the brain laterally and dorsally (Fig. 8d, e). Medially, nervous cells of the brain directly adjoin visceral muscle cells of the pharynx (Fig. 8 b–e, 9a, b). Although poor in structural content and with very electron-lucent cytoplasm, we suspect a supportive role of this glia-like tissue. Due to the similar appearance, these putative supportive cells may derive from the rest of the mesodermal parenchyma tissue. Further posterior, we could not support the presence of glia-like cells along the longitudinal bundles of neurites, although some of the sectioned cells may indeed be such supportive ones.

Anterior projections of the brain

The innervation of the region anterior of the dorsal commissure, emerging from the brain, shows a complex pattern of several distinctly separated bundles of neurites. The findings of the anti-tubulin immunohistochemistry are well matched by the TEM data. We describe here only additional results of the pattern. The dorsalmost bundle of neurites (an-I) originates approximately 15 μm posterior of the dorsal commissure (Fig. 5b, g), next to the posterior 5-HT-positive dorsalmost soma (\( {\text{dps}}_{\text{s}}^{\text{d}} \)). The bundle shows several branches, especially in the posterior part (Fig. 5b). The diameter of an-I reaches ~0.7 μm in the area of the dorsal commissure. Between an-I and the neuropil occur two connecting bundles (Fig. 5e, g); both are in an anteriolateral position of the commissure. The more posterior one is obviously additionally connected with the second cephalic bundle of neurites (an-II) (Fig. 5g). Furthermore, some very fine fibres (one to two per side) occur approximately 6 μm in front of the dorsal commissure and seem to connect an-I with the mediodorsal neurites (mdn). Approximately 15 μm anterior of the dorsal commissure, both bundles of an-I are interconnected by a ventral commissural fibre (diameter between 0.8 and 1 μm); we name this the ventral commissure of an-I (avco) (Figs. 1b, 6o, 7p). This commissure shows at least one 5-HT IR fibre (avcos) (Fig. 6o).

The second bundle of neurites (an-II) located more ventrolateral from the dorsal commissure has two connections with the neuropil of the dorsal commissure. Approximately 10 μm in front of the commissure, the bundle branches and the descendent bundle runs to the anterior part of the neuropil of the dorsal commissure (Fig. 6r). More posteriorly, a second branch occurs, and these fibres run to the posterior part of the commissure.

The immunohistochemical data support the branching pattern of these bundles (anIan-III) of neurites in the frontal area, laterally of the end of the pharynx (Fig. 5a, i). We infer from this an innervation of the anteriorly located sensory cells by an-Ian-III.

The lateral longitudinal neurite bundles (ln)

The pair of lateral longitudinal neurite bundle arises in the posterior part of the clusters of neuronal somata of the brain and emerges obviously directly from the posterior part of the dorsal commissure (Fig. 1b, 5g). In contrast to the innervation of the anterior region, where several neurite bundles (for example an-Ian-III) comprise the innervation, the longitudinal neurite bundles innervate the posterior trunk region exclusively; no additional longitudinal neurites has been observed. The position of the neurite bundle is dorsolateral of the paired ventral strands of the motile cilia (Fig. 5a, q, s–u), and they are in close contact with the ventrolateral longitudinal musculature (Fig. 7q). The fibrous material itself seems to be free of nuclei, as we did never observe nuclei within the longitudinal neurite bundle, neither with TEM nor with nuclei staining. Nuclei may only sometimes be present at the surface of the bundle. The diameter of the neurite bundle decreases slightly during its course to the posterior end, but the diameter varies in general along the course. The density of the fibres of the neurite bundle changes significantly. In some areas, the fibres are highly condensed, thereby minimizing the diameter of the neurite bundle (ca. 1.3–1.5 μm); in other areas, the fibres are more loosely packed, and the diameter is therefore larger (ca. 4–5 μm).

Ultrastructurally, the longitudinal neurite bundles are more clearly separated from surrounding tissues (i.e. neurons of the brain in the region frontal to the dorsal commissure and parenchyma cells further frontally) than the anterior neurites (an-Ian-III) (Fig. 8b, c). This is obviously due to a more abundant branching of the frontal bundles of neurites, even within anterior regions of the brain (Fig. 1b, 5b). The presence of only a single pair of longitudinal neurite bundles (ln) could be supported by numerous cross-sections at different regions (pharyngeal region, testicular region, ovarial region, region of the accessory reproductive organs) (Fig. 9c, d). Counting of neurites per bundle revealed approximately 25, whereas the number slightly decreases from anterior to posterior (ca. 27 neurites per bundle in the pharyngeal region, 22 neurites posterior to the pharyngeo-intestinal junction). The presence of regularly arranged serial somata along the ventral pair of the longitudinal neurite bundles could not be demonstrated. Organelles such as mitochondria and vesicles are present inside the cross-sectioned neurites of the longitudinal fibres (an-Ian-III, ln). In addition to the approximately 50-nm-wide vesicles, microtubule-like structures are present.

At the posterior end of the trunk, the neurite bundles from both sides are connected (Fig. 5q). In this region, the diameter of the neurite bundle is more constant than in the remaining parts, between 0.8 and nearly 2 μm. At the beginning of the posterior connection, a pair of posterior 5-HT-positive cells is present.

Posterior of the brain, two paired 5-HT-positive longitudinal fibres are visible (Fig. 6a, c, h); these fibres have a diameter in the range of approximately 0.5–3 μm; this depends on the “perl necklace”-like appearance of the single fibres. This indicates the heterogeneous distribution of 5-HT along entire length of the fibre. The fibres of one body side run close to each other in posterior direction, oriented directly at the outside of the lateroventral longitudinal musculature (vlm) (Fig. 6j). The paired 5-HT-positive fibres are completely embedded within the fibres of the lateral longitudinal neurite bundle (Fig. 6n). The origins of the single fibres of the double-fibre arrangement are different; one of them comes downward from the dorsal commissure (Fig. 3a, 6c, r) and represents the posterior connecting fibre between the longitudinal fibres and the dorsal commissure. The residual fibre comes from an anterior direction, in a ventrolateral position, and represents the 5-HT IR connection to the anterior trunk region (Figs. 3a, 6c). Anteriorly of the dorsal commissure, the single fibre runs along the pharynx for approximately 30 μm until the fibre reaches the point where the anterior 5-HT-positive connective fibre of the dorsal commissure meets the ventrolateral longitudinal bundles of neurites. It seems that the two fibres do not fuse, rather they run together closely associated for approximately 50 μm in anterior direction. The anterior 5-HT-positive fibres (afs) extend in anterior direction up to the buccal tube (Fig. 6a). For a detailed description of the interconnection within the area of the brain, see above.

Posteriorly, a paired 5-HT-positive thickening is present; the thickenings represent nuclei according to our staining (Fig. 6i): the 5-HT-positive posterior somata (pss) (Fig. 6a, h). These cells are located in adults approximately 20–30 μm anterior of the posterior end, the measurements were taken from the posterior side of the cells orthogonally to the base of the posterior adhesive organs. Both longitudinal fibres reach the area of the 5-HT-positive somata (Fig. 6h). The somata are located clearly dorsal of the longitudinal muscles (Fig. 6l). Obviously, only one longitudinal fibre is connected with the somata, it appears that this is in most cases the inner longitudinal fibre (Fig. 6h), but in some specimens the condition could not be resolved and we cannot exclude an inverted state sometimes. At the posterior side of these somata, a 5-HT-positive fibre is visible again. These inner fibres (ilfs) form a loop-like connection between the two ganglia-like cell somata and constitute the closed 5-HT-positive posterior fibre connection (pfcs) (Fig. 6h, j). The other posterior fibre, the outer longitudinal 5-HT-positive fibre (olfs), follows the lateroventral longitudinal muscles (vlm) for approximately 20–23 μm until the ventral longitudinal muscles reach the base of the posterior adhesive organs (Fig. 6a, h, l).

During its course, thin processes of anti-tubulin IR fibres run out of the neurite bundle in dorsolateral direction in a quite regular pattern (Fig. 5u). The calibre of the fibres is quite thin, in the range of 0.5 μm; due to this, we conclude that they are single neurites. These fibres end distally in the lateral region of the trunk where the lateral adhesive tubes are located. The single fibres seem to terminate in sensory cilia of ~15 μm length, associated with the adhesive tubes (Fig. 5q, u). In most cases, the tubes are indicated by a diffuse anti-tubulin signal (Fig. 5q). An IR structure, ca. 10 μm long by ca. 3 μm wide is present at the distal end of the neurites. This structure does not include nuclei within the tubulin-immunoreactive part, so we conclude that the secretory part of the adhesive organ is responsible for the signal (Fig. 5q). Interestingly, not all somata of the sensory cilia are in a distal position, where the cilium reaches the surface, rather some of them are in a more proximal position. This is indicated as a swelling of the neurite that includes a nucleus. The absolute position seems to be variable; in some cases, the sensory soma is directly beneath the adhesive tube, in other cases it is up to 10 μm proximally of them.

In addition to the nervous structures, anti-tubulin antibodies have labelled some non-nervous structures. Among others, the motile ventral cilia show a strong immunoreaction against tubulin, but only the cilia and not the somata of these cells are stainable (Fig. 5a, q, s–u). The two bands of motile cilia run in a ventrolateral position from anterior up to U 85. Furthermore, the testes are indicated by the labelling of the axonemata of the spermatozoa in the area of the midgut from U 25–30 up to U 60 (Figs. 5q, 6m). Further, in some cases a spermatophore with allosperms is labelled at the end of the ventral bands of cilia (Fig. 5r), in the area of the “frontocaudal” organ (Fig. 5q). Beyond that, the nuclei of the spermatozoa are clearly recognizable by DAPI-staining, due to the conspicuous thin elongated form (Fig. 6m). In the specimens that bear a spermatophore, the posterior part of the “frontocaudal” organ shows a tubulin IR, too. Here, a band of fibrous material forms a triangle posterior of the spermatophore, one apex is directed in posterior direction and one side lies directly posterior of the allosperm. Within the centre of the triangle, only a few immunopositive fibres are located in an unordered way. The fibrous matter appears to consist of densely packed filaments; the two posteriorly directed sides seem to consisting of fibrous matter with a diameter around 3 μm of the strand (Fig. 5q). The side of the spermatophore shows the same pattern, but in anterior direction the material is extended in form of loosely packed fibrous matter and range to the allosperms. We can exclude that spermatozoa cause this signal, because we never observed the typical elongated nuclei within that area, those nuclei only occurred at the allosperms and within the testes (Fig. 6m).

The innervation pattern of the pharynx

Innervation and sensory cells of the pharynx

Three bundles of pharyngeal neurites are confirmed at the ultrastructural level: one unpaired dorsal bundle (dpn) and one pair of lateroventral longitudinal bundles (lpn) (Fig. 9a, e, f). These neurite bundles are less than 0.5 μm in diameter and have a basi-epithelial position between the myoepithelial cells of the pharynx (Fig. 9a). However, in some sections the lateral pharyngeal bundles show a more intra-epithelial position (Fig. 9f). Each bundle consists of a maximum of five single neurites. The additional unpaired neurite bundle (mdn) that lines the pharynx dorsally but exterior to the epithelial cells and visceral muscles could not be detected unambiguously by TEM. It was not possible to separate such a small fibre from sections of small muscle cells. All pharyngeal neurite bundles are associated with sensory cells.

The paired ventrolateral ciliary receptor cells of the pharynx are situated in the upper edges of the triangular pharyngeal lumen (Fig. 9g). The sensory cell contains a short cilium that projects into a bulge of the pharyngeal cuticle. Short, 0.5-μm-long microvilli with electron-dense content surround the single cilium. Two 0.5-μm-long cross-striated ciliary rootlets are attached to the basal body of the cilium (Fig. 9g). An accessory centriole has not been observed. The cell is densely filled with 50- to 100-nm-wide vesicles. The receptor cells associated with the unpaired dorsal pharyngeal neurite bundle (dpn) are positioned medially between the adjacent myoepithelial cells of the foregut. When viewed with higher magnifications, one can observe cellular junctions, obviously belt desmosomes (zonulae adhaerens) at the apical sides of the receptor cells of dpn. With TEM, we have also observed a striated pattern basal to the belt desmosomes that could indicate the presence of additional septate junctions. Like the paired lateral receptor cells of the pharynx, the dorsal ones possess a single cilium projecting into a bulge of the pharyngeal cuticle (Fig. 10a). There is also a circle of short, 0.5- to 1-μm-long microvilli (not penetrating the cuticle) (Fig. 10b). Attached to the basal body of the cilium, there is a whole bunch of cross-striated ciliary rootlets projecting deep into the basal cytoplasm of the cell (Fig. 10b–d). An accessory centriole could not be detected. Like the paired ventrolateral pharyngeal receptor cells, the unpaired dorsal ones are filled with 50- to 100-nm-wide vesicles. Basal to the dorsal receptor cells of the pharynx, two or three sections of the dpn could be observed in close contact to the cell (Fig. 10b–g). These neurites are also filled with small (around 50 nm in diameter) vesicles. Mitochondria are abundant in the dorsal receptor cells of the pharynx. Due to the projection of their cilia into the pharyngeal lumen, we suspect a mechano-receptive function of the pharyngeal sensory cells.
Fig. 10

Neodasys chaetonotoideus. Ultrastructure (TEM) of the dorsal ciliary receptor cells of the pharynx. a Cross-section of the myoepithelial foregut, note the triradiate lumen and the receptor cell indicated by the dashed white line. b–d Sections of a dorsal ciliary receptor cell at different levels (anterior to posterior). bb basal body, cil cilium of the receptor cell, cr ciliary rootlets, cut cuticle of the pharynx lumen, dpn dorsal pharyngeal neurites, lpn lateral pharyngeal neurites, mf myofilaments, mv microvilli of the receptor cell, phl pharyngeal lumen, spm subpharyngeal musculature (circular and helicoidal muscle strands), asterisks: neurites of dorsal pharyngeal neurite bundle, black triangles: cellular junctions

Within the pharynx, we were able to detect several bundles of neurites by anti-tubulin IHC: the paired ventrolateral bundles of neurites of the pharynx (lpn) (Fig. 5f, j), the unpaired dorsal one (dpn) (Fig. 7j, k, o), and some associated sensory devices as well. These data are in account with the ultrastructural data. We will describe here only some additional findings. We do not find IR structures against 5-HT or RFamide within the pharynx, with the exception of a weak immunoreaction against RFamide in the dorsal longitudinal pharyngeal bundle of neurites (dpn) and the associated sensory cells (Fig. 7c, i).

First, the dorsal bundle of neurites of the pharynx, the dorsal pharyngeal neurite bundle (dpn), extends from the anterior margin of the pharynx to the posterior transition area between pharynx and intestine (Fig. 5a) and is present under the neuropil of the dorsal commissure (Fig. 5e). It seems that this neurite bundle is extending to the transition between posterior pharynx and the midgut. At the posterior end, a unique arrangement of obviously sensory structures (ppc) is observable. When the dorsal bundle of neurites (dpn) has left the pharynx posteriorly, it ends in a strongly tubulin IR structure (ppc) (Fig. 5i). From the horizontal plane, this structure appears like a triangle, with the apices in posterior and lateral position (Fig. 5a, i). The distance between the lateral apices was measured between 16 and 18 μm, and these apices run far between the pharyngeal epidermis. The structure is extending in the anterioposterior axis approximately 8 μm. The posterior apex runs in a medial position. We were able to demonstrate the occurrence of nuclei within the tubulin IR material. Directly at the point where the dorsal pharyngeal bundle of neurites (dpn) reaches the triangle, in most cases two nuclei were observable in nuclei-counterstaining (Fig. 5o). In some stainings, we could not distinguish two separate nuclei. In those cases, we had the appearance of only one larger nucleus at the base of the triangular structure.

The dorsal pharyngeal bundle (dpn) bifurcates approximately 8 μm posterior of the mouth opening, at the approximate position of the beginning of the buccal tube (Fig. 5i). Two branches extend to the dorsal rim of the pharynx. Here, these separated bundles innervate two dorsolateral sensory cells of the pharynx (Fig. 1b). During its course, the unpaired bundle of neurites is additionally associated with ~5 sensory cells (adscp) anterior of the dorsal commissure and ~4 sensory cells (pdscp) posterior of the dorsal commissure (Figs. 1b, 5p, 7i). These somata are indicated by the occurrence of a cilium that is extending in the lumen of the pharynx (Figs. 5l, n, p, 7i, o). Furthermore, these dorsal somata of pharyngeal sensory cells show an RFamide-like IR (Fig. 7i), whereas the ventrolateral ones do not show such a signal. The arrangement and distribution of the dorsal sensory cells is continuous along the entire pharynx, by contrast the distribution of obviously sensory somata in the ventral part of the pharynx is more heterogeneous.

The ventrolateral bundles of neurites of the pharynx (lpn) seem to end at the posterior border of the pharynx; obviously, they do not extend to the transition area between pharynx and midgut (Fig. 1b). Nevertheless, it is sometimes quite difficult to determine exactly the ending of these bundles of neurites, depending on the position close to the ventral ciliation and the longitudinal neurite bundles. In most specimens, a commissural connection between the ventral and dorsal neurite bundles was detectable in front of the dorsal commissure, in a distance of approximately 25 μm (Fig. 1b). The fine commissural fibre, the diameter is below 0.4 μm, is not closed at the ventral side between the ventral bundles of pharyngeal neurites, we name this fibre the commissure of the pharynx (in Fig. 1b indicated by black arrow). Only in some preparations, one to two additional commissural fibres were observed between the pharyngeal commissure and the dorsal commissure of the brain; those fibres are much thinner than those we have found at the pharyngeal commissure. We cannot exclude that those fibres were not visualizable in the remaining preparation.

The lateroventral bundles of neurites show a branching of the fibres, too. Approximately 35 μm posterior of the anterior end of the pharynx both fibres split in anterior direction (Fig. 5i). Posterior of the splitting, the diameter of the bundle was measured to be around 1 μm. One branch runs further in a ventrolateral position within the pharynx as before, but the diameter is only around 0.5 μm. The other branch runs in an angle of ~45° to a more ventromedial position (Fig. 5j), the distance between the two braches is ~4 μm and the inner bundle runs for about 11–20 μm in parallel to the outer one. The diameter of the inner fibre(s) was (were) estimated to be ca. 0.5 μm. It seems that the inner branch ends anteriorly at the first ventral sensory somata; sometimes, the fibre(s) make a short curve to a slightly more distal position before they end at the area of the sensory cell. In addition to this first pair of ventral sensory cells, at least 3–4 further ventral sensory cells were observable in the anterior part of the pharynx. Mostly, one pair occurs just at the level of the branching and in some specimens a further pair is observable more posteriorly. Not all of these somata occur as paired, sometimes only a single sensory cell at the right or at the left side occurs. In most cases, only the cilium gave a clear signal and the associated soma was only diffusely labelled.

Reconstruction of the stem species

The parsimonious character optimizations carried out based on six different phylogenetic scenarios (tree topologies) representing three competing ingroup relationships and two different ideas for the sister group relationship of Gastrotricha yields the following character pattern for the nervous system of the stem species of Gastrotricha:
  • A dumb-bell-like, commissural brain consisting of a bilateral cluster of neuronal somata; five pairs of which show a positive 5-HT IR and a belt-like dorsal neuropil (dorsal commissure) connecting both hemispheres of the brain.

  • A thin ventral commissure connecting both clusters of neurons.

  • Dorsal and ventral commissure together form a closed ring of neurites around the foregut.

  • Frontally projecting, paired anterior neurite bundles. However, the exact number of pairs could not be reconstructed unambiguously.

  • One pair of posteriorly projecting longitudinal neurite bundles in a ventrolateral position.

  • Both longitudinal neurite bundles fuse posteriorly in an arc-like posterior commissure.

Using the semi-strict approach, the following characters can be amended to the nervous system of the stem species of Gastrotricha:
  • Serially arranged RFamide-like IR neuronal somata that are associated with the longitudinal neurite bundles.

  • Posterior neuronal somata (=“caudal ganglion” =“anal ganglion” sensu Remane 1936) that shows (at least) a positive 5-HT IR.

Using different tree topologies for the ingroup relationships affected the reconstructed states for some characters. In nearly all instances, those differences turned out to be reconstructed states versus equivocal states rather than two different reconstructed character states. However, 27 character states were reconstructed identically among all six used scenarios (Table 1). Comparably, the varied positions of outgroup taxa (compare Fig. 11a with b, c with d, e with f) had no major effects on the reconstructed stem species.
Fig. 11

Phylogenetic trees used as basis for the parsimonious character optimization. Only taxa that are contained in the character matrix (Table 1) are arranged according to the results of different phylogenetic analyses. a, b Tree topology as inferred by Hochberg and Litvaitis (2000). c, d Tree topology as inferred by Todaro et al. (2006). e, f Tree topology as inferred by Kieneke et al. (2008). Left column (a, c, e): representatives of Cycloneuralia set as direct sister group of Gastrotricha. Right column (b, d, f): representatives of partial Platyzoa set as direct sister group of Gastrotricha. Chaet. Chaetonotida, Cyc. Cycloneuralia. The black circle refers to the stem species of Gastrotricha

Discussion

What is known about Neodasys?

The only published data based on the IHC of the nervous system of a species of Neodasys are from N. cirritus on RFamide-like IR (Hochberg 2007b). Our results confirm these previous findings in general.

Both species have a brain in a position slightly more posterior compared to other gastrotrichs and the gross anatomy of the brain appears quite similar. This applies to the shape and the dimensions of the RFamide IR structures, e.g. the tripartite dorsal commissure and to the position of the three single RFamide-positive fibre bundles. Hochberg (2007b) discriminates between three single RFamide IR dorsal commissures, our results of anti-tubulin labelling show that all three RFamide IR fibre bundles are within the broad dorsal tubulin IR commissure. Neodasys chaetonotoideus shows two distinct RFamide IR fibre bundles in the dorsal commissure, and the posterior part can be subdivided in two laterally fused sub-bundles. Furthermore, the number of IR somata within the cerebral ganglia is comparable (~50 in N. cirritus and ~48 in N. chaetonotoideus); among these are interspecifically homologous somata. For example, the large drop-shaped somata posteriolateral of the dorsal commissure (here pdsRF) (Figs. 10a, b, 11a in Hochberg 2007b), as well as the most anterior IR somata (here anbsRF, anc in Hochberg 2007b), occur in both species. Further consistencies are the anterior RFamide IR projections of the brain. In detail, the projections in N. chaetonotoideus extended further in anterior direction or have been better traceable. Finally, the RFamide IR longitudinal fibres that emerge in the posterior part of the brain show the same pattern of two ipsilateral RFamide-positive fibres in the two species. Hochberg (2007b) concludes that four longitudinal cords are present in N. cirritus, whereas the double-labelling IHC of RFamide and tubulin in N. chaetonotoideus clearly shows only one paired longitudinal cord. Each cord contains two RFamide-positive fibres. The posterior connection of the longitudinal cords is congruent between the two species. Further details on the number of longitudinal neurite bundles are found below.

Lateral RFamide-like IR somata associated with the longitudinal neurite bundles: a fact?

The most serious difference between the two Neodasys species is the occurrence of (weak) RFamide-positive somata along the longitudinal neurite bundles in N. cirritus (see Figs. 9b, c, 10d, 11a, b in Hochberg 2007b). The intensity of the fluorescence seems to be comparatively low; the fluorescence level of the IR somata of the brain appears much higher. Such a pattern was only observable in some individuals of N. chaetonotoideus. All of these specimens have been collected from detritus enriched sediment and the fluorescence was only detectable in fresh material (storage shorter than 2 weeks). Due to the scattered occurrence of the fluorescent signal in N. chaetonotoideus, it appears questionable to us whether this is caused by a specific RFamide IR. We present here an alternative hypothesis. It is known that Neodasys sp. possesses haemoglobin-containing cells (Kraus et al. 1981; Ruppert and Travis 1983; Colacino and Kraus 1984). Haemoglobin molecules do show only weak intrinsic fluorescence (Hirsch et al. 1994); the excitation maximum is in the ultraviolet wavelength range. On the other hand, haemoproteins have other optical characters, such as distinct absorption spectra. Colacino and Kraus (1984) have measured the haemoglobin content of cells in Neodasys sp. by means of mikrospectroscopy. The optical density (OD) of the haemoglobin from Neodasys sp. has the wavelength maximum between 520 and 580 nm, depending on the state of oxygenation. We propose a so-called pseudofluorescence (after van der Ploet and van Dujin 1979) as a source for the unspecific detection of haemoglobin during IHC with TRITC-labelled secondary antibodies. This effect is based on the reflection of light in the range of the absorbance peak of the chromophore, not on real fluorescence. The reNIRS (refection near infrared spectroscopy) has recently become a common method for haemoglobin measurement (e.g. Plesnila et al. 2002). In our preparations, the lateral IR cells of the longitudinal cord have been observable only when using TRITC conjugated antibodies, but the majority of labellings have been done with CY 5-labelled antibodies, due to double or triple labelling. Additionally, we do not observe these IR in material from well-oxygenated sediment, where probably the animals have lower haemoglobin content. In many organisms, the haemoglobin content is triggered by the oxygenation of the habitat (e.g. Czeczuga 1961 for Chironomus annularis larvae [Insecta], Zeis et al. 2003 for Daphnia magna [Crustacea]). The alternative to this explanation is the existence of a row of several lateral RFamide IR somata in N. cirritus and in some (sub)-populations of N. chaetonotoideus. By TEM, such neuronal somata were not found. However, Hochberg (2007b) as well reports somata of neurons associated with the longitudinal neurite bundles in Turbanella cf. hyalina and Xenodasys riedli. Furthermore, the existence of such serial neuron somata in Turbanella cornuta is highly probable (see Fig. 14B “peripheral neurons” in Teuchert 1977). Given this and the fact that neuron somata associated with the main longitudinal bundles (paired or unpaired, respectively) occur in some outgroup taxa too (see Table 1) could indicate that they also are a character of the stem species of Gastrotricha though this could only be reconstructed following the semi-strict approach. Also the data basis concerning these cells is far too fragmentary at present.

Other differences between N. chaetonotoideus and N. cirritus

Contrary to the results of Hochberg (2007b), a ventral subpharyngeal commissure is present in N. chaetonotoideus posterior to the dorsal commissure.

Further differences could be found in the innervation of the pharynx. Contrary to N. cirritus, the pharynx of N. chaetonotoideus contains a dorsal RFamide-like IR neurite bundle accompanied by associated sensory cells in the dorsal side of the triradiate pharyngeal lumen. The ultrastructural data on the pharynx are in accordance to the previous findings by Ruppert (1982), concerning the number and position of neurite bundles, position of sensory cells, as well as the presence of paired sensory cells in the anterior most part of the pharynx.

Posteriorly located 5-HT IR somata have not been reported from Neodasys species to date.

Comparison of the nervous system of N. chaetonotoideus to other Gastrotricha

Phylogenetic relation within Gastrotricha remains uncertain; nevertheless, two large sister groups can be characterized based on morphological data: Macrodasyida and Chaetonotida (Fig. 11a, b). Within the Chaetonotida the Multitubulatina (consisting only of the genus Neodasys) represents the sister group of the Paucitubulatina (including all remaining, tenpin-shaped chaetontids) (e.g. Travis 1983; Ruppert 1991; Hochberg and Litvaitis 2000; Ax 2001). Kieneke et al. (2008) concluded from a re-evaluation of morphological data that the genus Neodasys represents the sister taxon of all other gastrotrichs (Fig. 11e, f). On the other hand, the phylogeny of Todaro et al. (2003, 2006) based on molecular data shows a position of Neodasys as an ingroup of the Macrodasyida (Fig. 11c, d). We discuss the data on N. chaetonotoideus in this context and give some hypotheses about the nervous system of both Macrodasyida and Gastrotricha.

General brain architecture of Gastrotricha

The overall architecture of the brain of N. chaetonotoideus fits well the “dumb-bell-like” organization as proposed by Hochberg (2007b) for the ground pattern in Gastrotricha. The “dumb-bell-shaped” brain is composed of a paired lateral cluster of neuronal somata connected by a dorsal neuropil (dorsal commissure). This shape is described for Neodasys cirritus (Hochberg 2007b) and several macrodasyidan gastrotrichs (e.g. Xenodasysriedli, Turbanella cf. hyalina by Hochberg 2007b; Dactylopodola baltica and D. typhle by Rothe and Schmidt-Rhaesa 2009 and in Oregodasys cirritus by Rothe and Schmidt-Rhaesa 2010a) and has also been reconstructed by the parsimonious character optimization (Table 1). Previous descriptions favoured a slightly different brain composition. Based on light microscopy, Remane (1926, 1927, 1936) concluded that the brain of the Gastrotricha shows an organization with a pronounced lateral concentration of neuronal somata (“two lateral masses” Remane 1926). This means that a central dorsal neuropil is more or less equally surrounded by neuronal somata (anterior, posterior and lateral, but not dorsal or ventral of the dorsal commissure), which is more comparable to the conditions in many lophotrochozoans (e.g. Platyhelminthes, see Morris et al. 2007 for Macrostomum lignano with a detailed description of the compact brain). Nevertheless, Remane (1926, 1927, 1936) described a massive dorsal commissure and excluded the presence of a ventral commissure. Remane’s view has been extended and interpreted in part differntly by TEM surveys of the brain of different gastrotrichs by Teuchert (1977), Ruppert (1991) and Wiedermann (1995). These authors either depict the brain as a compact mass of neuronal somata with a central neuropil or as a pattern with an anterior and posterior concentration of neuronal somata like in Turbanella cornuta (Teuchert 1977). Furthermore, the ultrastructural data support a circumpharyngeal organization of the brain, with a massive dorsal and a thin ventral commissure in e.g. T. cornuta (Teuchert 1977), Cephalodasys maximus (Wiedermann 1995) and D. baltica (Ruppert 1991). This pattern (massive dorsal and thin ventral commissure) is confirmed by IHC in several other macrodasyian species (Hochberg 2007b; Rothe and Schmidt-Rhaesa 2008, 2009) and for N. chaetonotoideus as a member of the Chaetonotida (this study). We conclude that a circumpharyngeal organization of the brain belongs to the ground pattern of the Gastrotricha, despite the lack of data from members of the Paucitubulatina. This could undoubtedly be supported by the results of the matrix-based character optimization under the parsimony criterion (Table 1). A circumpharyngeal brain, with an unequal diameter of the surrounding neuropil during its course, is proposed to be the plesiomorphic state for the protostomian taxa (see Nielsen 2001).

The longitudinal neurite bundles of Gastrotricha

It is often mentioned that multiple paired nerve cords (here longitudinal neurite bundles) are a plesiomorphic character of basal gastrotrichs, especially Dactylopodola baltica and Neodasys sp. are often cited in this context. But, for Dactylopodola baltica, Rothe and Schmidt-Rhaesa (2009) have shown the existence of only one pair of lateral longitudinal neurite bundles, an additional paired medial branching 5-HT IR fibre is restricted only to the anterior trunk region. The citation of Travis (1983) as a source for a description of multiple longitudinal nerve cords is a misinterpretation, as Travis (1983) clearly shows in Fig. 16 (p. 65) only one pair of longitudinal cords in Neodasys sp. and furthermore mentions that “in the other Macrodasyida and Neodasys spec. the paired cords are lateral” in contrast to “the two pairs of Dactylopodola baltica lie in a ventral and lateral position” (p. 64 in Travis 1983). This implies one paired cord in all gastrotrichs, except for D. baltica. In addition, Ruppert and Travis (1983) described three pairs of anterior and one pair of posterior neurite bundles emerging from the brain in Neodasys sp. The closed posterior loop of the paired longitudinal cord has been mentioned in several gastrotrich species investigated by IHC with the focus on the nervous system so far. This character has been found in T. cornuta (Joffe and Wikgren 1995; Rothe and Schmidt-Rhaesa 2008), Turbanella ambronensis, T. cf. hyalina (Rothe and Schmidt-Rhaesa 2008), D. baltica and D. typhle (Rothe and Schmidt-Rhaesa 2009) and Oregodasys cirritus (Rothe and Schmidt-Rhaesa 2010a). Considering the same arrangement in N. chaetonotoideus, as well as in many outgroup taxa (see Table 1), yielded an arc-like caudal fusion of the longitudinal neurite bundles, a posterior commissure, to be a ground pattern feature of Gastrotricha.

Hypotheses about the homology of single neurons in Gastrotricha

If we take a closer look on the IHC pattern of the brain of N. chaetonotoideus, we can try to find neuronal homology between single neurons in different gastrotrich species. Neurons can be considered homologous when the criteria of homology, e.g. the same neurotransmitter expression, position, shape and neuronal connectivity are satisfied (see Duham-Scheel and Patel 1999; Hirth and Reichert 2007). Excellent candidates for such a homology search are the 5-HT IR cells, due to the good resolution of the single cells and data present for several species. First candidates for such a comparison are the two paired 5-HT IR dorsoposterior somata (\( {\text{dps}}_{\text{s}}^{\text{d}} \), \( {\text{dps}}_{\text{s}}^{\text{v}} \)) of N. chaetonotoideus. Similar cells are described from Turbanella cornuta (Joffe and Wikgren 1995; Rothe and Schmidt-Rhaesa 2008), Macrodasys caudatus, Dolichodasys elongatus, (Hochberg and Livaitis 2003), T. ambronensis, T. cf. hyalina (Rothe and Schmidt-Rhaesa 2008), D. baltica (Hochberg and Litvaitis 2001; Rothe and Schmidt-Rhaesa 2009), D. typhle (Rothe and Schmidt-Rhaesa 2009) and Oregodasys cirritus (Rothe and Schmidt-Rhaesa 2010a). The similarity is shown by (1) the 5-HT expression, (2) the comparable position (dorsolateral and posterior to the dorsal commissure), (3) the size and the all most similar shape (comparable large soma, drop-shape) and the connectivity (the soma is projecting in neuropil of the dorsal commissure in the dorsolateral area). Differences are present in detail in the number (one or two pairs) and sometimes in shape (additional processes in T. cornuta). Due to the lack of a coherent phylogeny of Gastrotricha and the absence of IHC data on species of the Paucitubulatina, we hypothesize that the ground pattern of the Macrodasyida can be defined in part by the presence of 1–2 pairs of dorsolateral, unipolar, 5-HT IR cells that send a projection to the dorsal commissure.

Apart from the apparent homology of the posterior dorsolateral 5-HT IR somata, there occur additional 5-HT IR somata within the brain of N. chaetonotoideus. Among previous descriptions of the 5-HT-positive components in the brain of species of Gastrotricha, only members of the genus Dactylopodola have comparable somata. The dorsal anterior 5-HT IR soma (dass) in N chaetonotoideus can be compared with the dorsal anterior 5-HT IR soma (dacs) in D. baltica (Fig. 1c, f, h, i in Rothe and Schmidt-Rhaesa 2009). In both species, this soma is located anterior to the dorsoposterior ones, directly posteriolateral to the neuropil of the dorsal commissure. In contrast to D. baltica, N. chaetonotoideus has in dass no anteriorly directed process. Rothe and Schmidt-Rhaesa (2009) could not exclude a direct connection between the posterior dorsolateral 5-HT IR somata and dass, but such a connection definitely does not exist in N. chaetonotoideus. Ventrolaterally of the dorsal commissure are in both species auxiliary 5-HT IR somata, there are two pairs in N. chaetonotoideus (\( {\text{ls}}_{\text{s}}^{\text{d}} \) and \( {\text{ls}}_{\text{s}}^{\text{v}} \)) and one pair (mcs) in D. baltica (Fig. 1f, g, i in Rothe and Schmidt-Rhaesa 2009). The position of the cells is comparable, but the number is different, and the connectivity is not comparable, because Rothe and Schmidt-Rhaesa (2009) do not report a connection between this soma and other cells or a connection to the neuropil as the somata in Neodasys.

The data permit a hypothesis about the presence of posterior dorsolateral 5-HT IR cells in the ground pattern of Macrodasyida, and depending on the position of Neodasys this may account for all Gastrotricha. Furthermore, we try to depict a hypothesis about the 5-HT IR components of the brain in the gastrotrichs based on the presumption of a basal position of Dactylopodola within the macrodasyidan gastrotrichs and a basal position of Neodasys within the Gastrotricha. In that scenario, the occurrence of additional 5-HT IR somata seems reliable and represents a character of the ground pattern of Gastrotricha; a secondary loss in the more derived members of the Macrodasyida must have been occurred in this case. Future investigations should focus on the nervous system of several members of the Paucitubulatina, to get a more holistic picture of nervous system evolution within the gastrotrichs. The reconstructed pattern of five pairs of 5-HT IR somata (when treating those cells as a whole, see Table 1) in the brain of the stem species of Gastrotricha is due to the identical character states of Neodasys chaetonotoideus, Macrostomum pusillum (Plathelminthes) and the only acoel species included in the matrix, Convolutriloba longifissura. However, it has to be tested carefully whether those five cell pairs are homologous among the three taxa, and this needs a broader basis of data.

Finally, we focus on the occurrence of posterior 5-HT IR somata associated with the longitudinal neurite bundles in N. chaetonotoideus. Such somata have not been reported from IHC investigations of other gastrotrich species until today, but data by IHC of the nervous system of members of the Paucitubulatina are completely lacking. Remane (1936) reported, partly based on earlier observations by Zelinka (1889), the presence of an “anal ganglion” in the posterior end of members of the Chaetonotida and Macrodasyida. Within macrodasyidan gastrotrichs, we never found RFamide-like or 5-HT IR somata at the posterior end of the longitudinal bundles, but its presence in Paucitubulatina species should be checked. Given the data contained in the data matrix (especially derived from several outgroup taxa) and following the semi-strict account for compiling the “consensus ground pattern”, posterior neuronal somata (=“caudal ganglion”) might already have been present in the stem species of Gastrotricha (Table 1). However, the homology among taxa such as Gnathostomulida, Rotifera, Cycloneuralia and Gastrotricha requires further data such as their quantity, shape, connectivity and neurotransmitter expression patterns.

The stem species of Gastrotricha

Reconstructions of ground pattern features of Gastrotricha based on a comprehensive species–character matrix and using the criterion of parsimony have by now been carried out for the myoepithelial pharynx (Ruppert 1982), the body musculature (Hochberg and Litvaitis 2001; Hochberg 2005), the protonepridial system (Kieneke et al. 2007) and the reproductive organs (Kieneke et al. 2009). This by degrees improves our knowledge of the last common ancestor of the phylum. The stem species is the hypothetical starting point when tracing the evolutionary pathways of certain organ systems but also serves as a reference in a broader phylogenetic sense.

Carrying out the character optimization based on different competing phylogenetic hypotheses and compiling a “consensus ground pattern” from those results has been done before (e.g. Kieneke et al. 2009). We think this procedure enables us to reconstruct ground pattern features with a high “likelihood” of being plesiomorphic and conserved since the differences among the used trees do not affect the reconstruction results. The hypothesized neuroanatomy of the stem species of Gastrotricha—a dumb-bell-shaped brain with a belt-like dorsal and thin ventral commissure, anterior neurite bundles and one pair of posterior longitudinal neurite bundles that fuse caudally—confirms the former ideas on this topic (see Hochberg 2007b; Rothe and Schmidt-Rhaesa 2009). Although some of these results seem trivial, the computerized character optimization has the advantage of being transparent, reproducible and can be amended if new data are available. Some gaps in our knowledge of the neuroanatomy of the gastrotrich ancestor, e.g. the number of anterior neurite bundles, rely on the still scattered nature of data related to the nervous system (see Table 1).

We recommend for investigations of the nervous system in the future not only the use of markers for neurotransmitters, e. g. antibodies against 5-HT or FMRFamide, but we also advocate the additional use of a marker for a specific but overall labelling of neurites, for example antibodies against acetylated α-tubulin or tyrosinated α-tubulin. This makes the discrimination, whether a single neurite IR for a specific neurotransmitter runs isolated or within a bundle of fibres, easier. Furthermore, the combination of different microscopic methods, for example, “advanced” lightmicroscopy (e.g. cLSM) and transmission electron microscopy (TEM), should be used wherever applicable, because such a dual approach combines the advantages of both methods in an excellent way. The nervous system reconstruction can be made by cLSM in a holistic way with the combined use of several markers and details can be confirmed by TEM on single ultrathin sections from the region of interest. This methodology gives the possibility of a fast reconstruction with a high reliability. For the future, a close connection between the two methods will be represented by the use of immunotechniques for TEM in combination with cLSM, to get the same impression on the ultrastructure level at the TEM as with IHC methods in light microscopy.

Acknowledgments

Many thanks to the people at the Wadden Sea Station (AWI) in List/Sylt, especially to Werner Armonies. BHR and ASR were supported by a grant of the Deutsche Forschungsgemeinschaft (DFG) (SCHM 1278/8-2) within the frame of the focal program Deep Metazoan phylogeny (SPP 1174). Two anonymous referees are also acknowledged for offering suggestions that greatly improved the manuscript.

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • Birgen H. Rothe
    • 1
  • Andreas Schmidt-Rhaesa
    • 1
  • Alexander Kieneke
    • 2
  1. 1.Biozentrum Grindel und Zoologisches Museum, Universität HamburgHamburgGermany
  2. 2.Forschungsinstitut und Naturmuseum Senckenberg, Deutsches Zentrum für Marine BiodiversitätsforschungWilhelmshavenGermany

Personalised recommendations