SALMFamide2 and serotonin immunoreactivity in the nervous system of some acoels (Xenacoelomorpha)

Abstract Acoel worms are simple, often microscopic animals with direct development, a multiciliated epidermis, a statocyst, and a digestive parenchyma instead of a gut epithelium. Morphological characters of acoels have been notoriously difficult to interpret due to their relative scarcity. The nervous system is one of the most accessible and widely used comparative features in acoels, which have a so‐called commissural brain without capsule and several major longitudinal neurite bundles. Here, we use the selective binding properties of a neuropeptide antibody raised in echinoderms (SALMFamide2, or S2), and a commercial antibody against serotonin (5‐HT) to provide additional characters of the acoel nervous system. We have prepared whole‐mount immunofluorescent stainings of three acoel species: Symsagittifera psammophila (Convolutidae), Aphanostoma pisae, and the model acoel Isodiametra pulchra (both Isodiametridae). The commissural brain of all three acoels is delimited anteriorly by the ventral anterior commissure, and posteriorly by the dorsal posterior commissure. The dorsal anterior commissure is situated between the ventral anterior commissure and the dorsal posterior commissure, while the statocyst lies between dorsal anterior and dorsal posterior commissure. S2 and serotonin do not co‐localise, and they follow similar patterns to each other within an animal. In particular, S2, but not 5‐HT, stains a prominent commissure posterior to the main (dorsal) posterior commissure. We have for the first time observed a closed posterior loop of the main neurite bundles in S. psammophila for both the amidergic and the serotonergic nervous system. In I. pulchra, the lateral neurite bundles also form a posterior loop in our serotonergic nervous system stainings.

commissural brain shows substantial differences between acoel groups, but it takes one of three main shapes: barrel, rosette, or bridge (Reuter, Raikova, Jondelius, et al., 2001). The number of commissures constituting the brain and the number of longitudinal neurite bundles is a character of use for systematics (Achatz & Martínez, 2012). In most acoels studied so far, three pairs of longitudinal neurite bundles are present, but this number can range from 2 to 6 neurite bundles (Raikova, Reuter, Kotikova, & Gustafsson, 1998). Commissures posterior to the brain are not distinct and are usually restricted to the peripheral plexus (Gaerber, Salvenmoser, Rieger, & Gschwentner, 2007;Raikova et al., 1998).
In acoels, the serotonergic nervous system has been studied most extensively, using antibodies against 5-hydroxytryptamine (5-HT). The majority of studies deals with relatively few genera: several species of the genera Childia and Convolutriloba, and eight species of other genera (see Table 1). Another fairly well explored part of the acoels' nervous system concerns the RFamidergic nervous system, using antibodies against the mollusc FMRFamide, the flatworm GYIRFamide and a simple RFamide (see Table 1).
The neuropeptide S2 is a dodecapeptide with the amino acid sequence Ser-Gly-Pro-Tyr-Ser-Phe-Asn-Ser-Gly-Leu-Thr-Phe-NH 2 (SGPYSFNSGLTFamide), first purified from two starfish species, Asterias rubens and Asterias forbesi (Elphick, Price, Lee, & Thorndyke, 1991). Besides several echinoderms, the S2 antibody has been shown to stain parts of the nervous system of Xenoturbella bocki and the enteropneust Harrimania kupfferi, and it has been interpreted as supporting a deuterostome affiliation of Xenoturbella (Stach et al., 2005).

| Immunocytochemistry
Live animals were relaxed for 5-10 min in 7.14% MgCl 2 *6H 2 O and then fixed at room temperature in 4% formaldehyde in PBS for 1 hr.
The whole-mounts were washed in phosphate buffered saline (PBS) with 0.1% Triton-X100 (PBS-T) for 30 min with multiple changes of solutions. The specimens were blocked for 3 hr in PBS-T with 1% bovine serum albumin (BSA-T) before being incubated in primary antibody diluted in BSA-T at 48C overnight. Several dilutions of the monoclonal primary antibody against S2 (raised in rabbit) (Newman, Elphick, & Thorndyke, 1995) have been tested (1:100; 1:500; 1:800, 1:1,500; 1:1,600,1:2,000; 1:3,200, 1:10,000; 1:20,000), the best results were achieved with a dilution of 1:500. In M. lignano and M. crozieri, dilutions of 1:800, 1:1,600, and 1:3,200, and in the juvenile bivalves a dilution of 1:500 were tested. The third use of the recycled antibody dilution resulted in the best staining. The primary antibody against 5-HT (monoclonal mouse anti-human serotonin clone 5HT-H209 DakoCytomation) was used in a dilution of 1:25 in BSA-T for I. pulchra and S. psammophila. For double stainings, both primary antibodies were applied simultaneously. Afterwards, whole-mounts were washed in PBS-T overnight at 48C and blocked in BSA-T for 2 hr at room temperature and incubated in a 1:300 dilution of the secondary antibody (Alexa Fluor 488 goat anti-rabbit or Alexa Fluor 555 goat anti-mouse) for 1 hr at room temperature. Specimens were washed again for at least 30 min in PBS-T, best results were achieved by washing over three nights at 48C. For observing the samples under the microscope, whole-mounts were mounted in VectaShield (Vector Laboratories). The primary antibody was omitted in negative controls.

| Documentation
Specimens were photographed with a Leica DM 5000B compound microscope equipped with a Leica DFC 490 digital camera and with a Leica TCS SP5 IIT confocal laser scanning microscope. Further image processing was performed with Fiji and Adobe Photoshop CS2. Drawings were produced in Adobe Illustrator CS2. Stainings are shown in false colours to accommodate colour-blindness.

| Immunoreactivity to an S2 antibody
In Symsagittifera psammophila, Isodiametra pulchra, and Aphanostoma pisae (Figure 1), the nervous system shows positive immunoreactivity with the antibody raised against S2 (Figures 2, 3a,b,e,f, 4, and 5). S2 immunoreactivity can be observed in the commissural brain of all three species. In I. pulchra, only the dorsal posterior commissure of the commissural brain and the posterior, but not the anterior lobe shows immunoreactivity to S2 ( Figure 3B). This is in contrast to S. psammophila and A. pisae which have two labelled commissures, between which the statocyst is located (Figures 2b and 4). Both in A. pisae and S. psammophila, but not in I. pulchra, the frontal ring (comprised of ventral anterior commissure and dorsal anterior commissure, see Sprecher et al., 2015) is labelled (Figures 2, 4, and 5). The dorsal, the lateral, as well as the ventral neurite bundles can be observed in all three species (Figures 2a, 3b, and 4). In I. pulchra, the medio-ventral neurite bundle is stained as well and delicate anterior processes can be observed (Figure 3b). We  Between the median neurite bundles, the commissure slightly sinks to the ventral side. We termed this commissure the collum commissure.
A subepidermal nerve net could not be observed in any of the three acoels. The S2amidergic nervous system is schematised in  3c,g, 5a, and 6a,b). Only in I. pulchra, the medio-ventral neurite bundles can be observed (Figures 3c,g, 5b). In the posterior body-part of S. psammophila the dorsal, lateral and ventral neurite bundles each form a loop, interconnected by several short longitudinal processes (Figures 5a and 6a,c). In both species, the peripheral nerve plexus is only stained by an antibody against serotonin, but not by an antibody against S2.
In I. pulchra serotonergic staining can also be observed in the frontal ring, the anterior lobe and the dorsal anterior commissure which are parts of the commissural brain (Figures 3d and 5b). Also, the posterior loop of the lateral neurite bundles can only be seen with serotonin.
In the posterior body-part of S. psammophila, the three loops of the main neurite bundles and their interconnections are less pronounced in the serotonergic immunoreactivity (Figure 6c). The collum commissure is specific to the S2 immunoreactivity and missing in the serotonergic nervous system (Figures 2d, 5a, and 6a). Finally, the serotonergic pattern around the statocyst is weaker than the S2 immunoreactivity (Figures 2e and 5a) in S. psammophila.

Our serotonergic nervous system stainings do not readily reveal indi-
vidual neurons and the number of their processes, while this is possible in S2 stainings of I. pulchra (but not of S. psammophila).
Negative controls omitting the primary antibody did not show any staining. A comparative overview of both S2 and serotonin stainings in all three acoel species is provided (Figures 3f,g and 5).

| Comparison of nervous system patterns
The present study documents the staining of parts of the nervous system in the three acoel species Symsagittifera psammophila, Isodiametra pulchra, and Aphanostoma pisae using a monoclonal antibody raised against the echinoderm neuropeptide S2, with the specific peptide sequence used for antibody production being KYSGLTFamide (Newman et al., 1995). Additionally, we have stained the serotonergic (5-HT) nervous system in S. psammophila and I. pulchra. No staining of the serotonergic nervous system of A. pisae was performed as this has been recently published (Zauchner et al., 2015). In A. pisae, the main differences between the S2amidergic and the serotonergic pattern of the nervous system lie in the position of the anterior processes, where the serotonergic processes are situated towards the midline, and the S2amidergic processes are situated more laterally (Zauchner et al., 2015; Figures 4 and 5c). Also, the peripheral nerve plexus typical for serotonin is missing in the S2 stainings. Even if the stainings of the nervous system with antibodies raised against S2 and serotonin show a similar pattern, they are not identical and do not co-localise either in S.
psammophila or in I. pulchra (Figures 2e and 3e). This is especially apparent in the dorsal anterior commissure of S. psammophila, which is located more anteriorly in the serotonergic pattern than in the S2 stainings ( Figure 5). Also, individual neurons could only be identified in S2 stainings of I. pulchra and A. pisae, but not in the respective 5-HT stainings, whereas in S. psammophila, neither S2 nor 5-HT stainings revealed individual neurons. Possibly, the smaller size of the isodiametrid species allows for easier analysis of the nervous system. FMRFamide and 5-HT were shown to not co-localise in I. pulchra (Achatz & Martínez, 2012), and in other acoels, GYIRFamide and 5-HT stainings were also found to be similar, but not to co-localise (Raikova et al., 2004a). We have  (Figures 4a and 5a,c), which are also evident in the respective S2 stainings (Figures 2a,c,d and 5a).
In sea stars, S2 was found to be an important agent to relax the cardiac stomach (Elphick et al., 1991;Newman et al., 1995). While no functional studies in acoels have been performed to date, the overall distribution of the nervous system stained with the S2 antibody suggests a broader role of an S2-like neuropeptide in acoels.
We have identified a commissure posterior of the brain, spanning between the lateral neurite bundles in S. psammophila, which we termed the collum commissure (Figures 2d and 5a). This commissure could only be detected in the S2amidergic, but not in the serotonergic nervous system. In S. roscoffensis, a similar (unnamed) commissure was found, disjointedly connecting the main neurite bundles, both in the serotonergic and synaptotagminergic nervous systems (Gavil an et al., 2016;Semmler et al., 2010). Apart from the commissural brain, commissures between the lateral neurite bundles are rare in acoels and have been detected in Aphanostoma pisae, but not in the closely related Isodiametra pulchra (Zauchner et al., 2015). In the nemertodermatid Meara stichopi, such commissures occur at regular intervals in the serotonergic nervous system, especially in the anterior half of the body (Raikova et al., 2000). The size of the animals does not seem to play a role in the absence or presence of these commissures, as they can be found both in rather large (Symsagittifera, several mm in length) and small (A. pisae, less than a mm long) species.
4.3 | Homology of the nervous system of the three observed acoel species All three acoel species used here belong to the large taxon Crucimusculata, encompassing all acoels except Solenofilomorphidae, Hofsteniidae, Paratomellidae and Diopisthoporidae (Jondelius, Wallberg, Hooge, & Raikova, 2011). While both, Isodiametra pulchra and Aphanostoma pisae, are members of the same family, Isodiametridae, Symsagittifera psammophila belongs to another large and well-studied family, the Convolutidae (Jondelius et al., 2011). This implies that the nervous system of I. pulchra and A. pisae are likely to be more similar to each other, than to S. psammophila.
pisae, we have only detected three pairs of longitudinal neurite bundles ( Figure 4).
Three pairs of longitudinal neurite bundles are reported in Convolutriloba longifissura (dorsal, dorsal, median in Gaerber et al., 2007), Symsagittifera roscoffensis (median, lateral, lateral in Semmler et al., 2010), and in Symsagittifera psammophila (dorsal, ventral, lateral, this study). Given these species, there seems to be a higher variability in the location of the neurite bundles along the dorso-ventral axis, as the second neurite bundle in C. longifissura is described as being ventral, in S. psammophila it is dorsal.
It is unclear which longitudinal neurite bundles in Isodiametridae are homologous to the ones found in Convolutidae and given the variability in the Convolutidae, the position along the dorso-ventral axis is probably not a good indicator of homology.
According to the seminal study of Achatz and Martínez (2012), the dorsal posterior commissure behind the statocyst and the frontal ring anterior to the statocyst are likely homologous between members of the Crucimusculata. The frontal ring is also apparent in Aphanostoma pisae ( Figure 5), although it was not named as such in the description of the serotonergic nervous system (Zauchner et al., 2015).

| Specificity of the S2 antibody
In all three acoel species, parts of the nervous system can be stained by an antibody raised against echinoderm S2. This antibody has been demonstrated to work in a variety of echinoderms, a hemichordate and Xenoturbella bocki, but not in chordates, lophotrochozoans, crustaceans, and cnidarians (Stach et al., 2005). This is also corroborated by our own failed staining attempts in two flatworm species and in juvenile bivalves. With more studied species and groups, it will become clearer if positive immunoreactivity to the S2 antibody is truly restricted to ambulacrarians and xenacoelomorphs. For example, the neuropeptides SSLFFamide in the mollusc Lottia gigantea and SGLFFamide in the annelids Capitella teleta and Helobdella robusta match the Asterias SALMFamide pattern 'S? [LF]?FG' (Veenstra, 2011) and are prime candidates for possible affinity to the S2 antibody, even though two other annelids and a mollusc did not show positive S2 immunoreactivity (Stach et al., 2005 and our own findings).