Ecdysis‐related neuropeptide expression and metamorphosis in a non‐ecdysozoan bilaterian

Ecdysis‐related neuropeptides (ERNs), including eclosion hormone, crustacean cardioactive peptide, myoinhibitory peptide, bursicon alpha, and bursicon beta regulate molting in insects and crustaceans. Recent evidence further revealed that ERNs likely play an ancestral role in invertebrate life cycle transitions, but their tempo‐spatial expression patterns have not been investigated outside Arthropoda. Using RNA‐seq and in situ hybridization, we show that ERNs are broadly expressed in the developing nervous system of a mollusk, the polyplacophoran Acanthochitona fascicularis. While some ERN‐expressing neurons persist from larval to juvenile stages, others are only present during settlement and metamorphosis. These transient neurons belong to the “ampullary system,” a polyplacophoran‐specific larval sensory structure. Surprisingly, however, ERN expression is absent from the apical organ, another larval sensory structure that degenerates before settlement is completed in A. fascicularis. Our findings thus support a role of ERNs in A. fascicularis metamorphosis but contradict the common notion that the apical organ‐like structures shared by various aquatic invertebrates (i.e., cnidarians, annelids, mollusks, echinoderms) are of general importance for this process.

Notably, ERNs are not limited to coordinating ecdysisrelated processes. CCAP exerts excitatory effects on heartbeat frequency, oviduct contractions, and the secretion of digestive enzymes in arthropods (Žitňan and Daubnerová 2016), while MIP inhibits muscular contractions, regulates feeding and reproductive behavior, and integrates developmental timing with environmental cues in numerous bilaterians (Shao et al. 2019;Williams 2020). Bursα homodimers contribute to stem cell quiescence in Drosophila (Scopelliti et al. 2016), while Bursα and Bursβ homodimers contribute to neuroendocrine-immune responses of both insects and crustaceans Zhang et al. 2017Zhang et al. , 2020Li et al. 2019). Although EH involvement in processes other than ecdysis has not yet been described, these findings illustrate that ERNs generally serve multiple functions. Given that they are not only present in ecdysozoans, but also in the two remaining, non-molting bilaterian superphyla, Lophotrochozoa and Deuterostomia, it is likely that ecdysis regulation is a derived function of ERNs that emerged during early ecdysozoan evolution Zieger et al. 2020).
Based on their temporal expression profiles, it has been suggested that ERNs play a conserved role in nephrozoan (Protostomia + Deuterostomia) life cycle transitions such as hatching in direct-developing species or metamorphosis in indirect developers (Zieger et al. 2020). However, in situ evidence for ERN expression in a non-ecdysozoan is currently lacking. This study provides the first documentation, to our knowledge, of ERN-expressing cells in key developmental stages of the polyplacophoran mollusk Acanthochitona fascicularis. Polyplacophorans have a fossil record that dates back ∼ 500 mya and exhibit a number of mollusk-specific and putatively ancestral features (Vinther et al. 2012;Scherholz et al. 2013;Vinther 2015;Wanninger and Wollesen 2019), including a non-ganglionated tetraneurous nervous system composed of an anterior circumoesophageal nerve ring and four main longitudinal nerve cords that are interconnected via commissures (Wanninger 2009;Richter et al. 2010;Sigwart et al. 2014;Fritsch et al. 2016).
During polyplacophoran larval development, the scaffold of the adult nervous system is established through successive formation of cerebral, pedal, and lateral neurons Voronezhskaya et al. 2002). Interestingly, this does not occur in an anterior-to-posterior progression, in contrast to segmented lophotrochozoans such as Annelida . The non-feeding polyplacophoran trochophore further possesses sensory structures that are restricted to the larval stage, including an apical organ and a so-called anterolateral "ampullary system" Haszprunar et al. 2002;Voronezhskaya et al. 2002). Apical organ-like structures are found in many aquatic invertebrate larvae, including those of cnidarians, lophotrochozoans, and deuterostomes, where they are thought to regulate locomotion, settlement, and metamorphosis (Kempf et al. 1997;Hadfield et al. 2000;Voronezhskaya and Khabarova 2003;Voronezhskaya et al. 2004;Rentzsch et al. 2008;Conzelmann et al. 2013;Hou et al. 2020). The ampullary system, on the other hand, is probably a polyplacophoran-specific feature of yet unknown function Haszprunar et al. 2002;Voronezhskaya et al. 2002).
Here, we used immunofluorescence, RNA-seq differential expression analysis, and in situ hybridization to assess ERN expression in the developing nervous system of a non-ecdysozoan animal, the polyplacophoran mollusk Acanthochitona fascicularis. Our data shed new light on the dynamic roles of neuropeptides and larval sensory systems in animal life cycle transitions and contradict common assumptions concerning the involvement of the apical organ in settlement and metamorphosis.

ANIMAL COLLECTION, CULTURE, AND FIXATION
Adult specimens of Acanthochitona fascicularis were collected in the intertidal zone near the Station Biologique de Roscoff in France (48°43´44.70"N 3°59´13.53"W). The animals spawned spontaneously at room temperature (20-22°C) 2 to 4 days after collection. After adding several drops of concentrated sperm solution and incubation for 40 min, eggs were washed and reared in filtered seawater at room temperature. Late trochophore larvae (48-60 h) were provided with empty gastropod shells from the collection site to induce settlement and metamorphosis . For immunofluorescence, specimens were fixed in 4% PFA in PBS (0.01 M phosphate buffered saline, pH 7), washed three times in PBS, and stored at 4°C in PBS containing 0.1% sodium azide. For RNA extraction, water was removed by centrifugation and developmental stages were flash-frozen in liquid nitrogen and stored at -80°C. For in situ hybridization, larvae and juveniles were fixed in 4% PFA in MOPS-EGTA (0.1 M MOPS, 2 mM MgSO4, 1 mM EGTA, 0.5 M NaCl) for 1-2 h, washed three times in 100% methanol, and stored in methanol containing 50 mM EGTA at -20°C.

SEQUENCING AND TRANSCRIPTOME ASSEMBLY
RNA was extracted from samples stored at -80°C in RNAlater (AMBION, Inc., Austin, Texas) using the RNeasy ® Plus Mini kit (74134, QIAGEN, Hilden, Germany). RNA-seq libraries were constructed for each sample using the NEBNext® Ultra TM II Directional RNA Library Prep Kit (#E7760, New England Biolabs, Frankfurt am Main, Germany) and the samples were multiplexed and sequenced on an Illumina NovaSeq 6000 using the SP protocol. Both library prep and sequencing were performed by the Next Generation Sequencing Facility at Vienna BioCenter Core Facilities (VBCF), member of the Vienna BioCenter (VBC), Austria.
For command line details of all bioinformatic steps, see Data S1, Supplementary Information). Libraries were preprocessed with bbduk (version 37.61, BBMap. http://www.sourceforge.net/ projects/bbmap/) to remove low-quality bases and adapter sequences (Bushnell et al. 2017). Library quality was assessed with fastqc (version 0.11.8, www.bioinformatics.babraham.ac. uk/projects/fastqc/). De novo transcriptomes were built for each developmental stage with transabyss (version 2.0.1) (Robertson et al. 2010) and these were then merged with transabyss-merge. Open reading frames were annotated with TransDecoder (version 5.02) (Haas et al. 2013) and redundancy of the output was reduced by using cd-hit (version 4.8.1) (Li and Godzik 2006) . Only those transcripts that encoded a peptide with a complete start and stop codon were retained for further analysis. It was noticed that no orthologue of eclosion hormone passed the stringent filtering performed during transcriptome construction, however a complete EH was found in the concatenated transcriptome file prior to ORF annotation. As such, EH was manually annotated and added to the reference transcriptome, bringing it to a total of 27,263 genes.

OF TRANSCRIPT ABUNDANCE
Using previously annotated neuropeptide sequences as queries Zieger et al. 2020), Acanthochitona fascicularis orthologues of eclosion hormone (EH), crustacean cardioactive peptide (CCAP), myoinhibitory peptide (MIP), bursicon alpha (Bursα) and bursicon beta (Bursβ) were retrieved through BLAST searches with default tBLASTn parameters and an e-value threshold of 1. Candidate sequences were checked by reciprocal BLASTs and aligned with their respective bilaterian orthologues using the EINSI algorithm of MAFFT (Katoh and Standley 2013). Sequence orthology was initially determined based on the presence of shared conserved amino acid regions and residues ( Figure S1, Supplementary Information). EH and Burs pro-neuropeptide sequences are characterized by a pattern of six and eleven conserved cysteine residues, respectively (except polychaete, scaphopod and cnidarian EH with only five cysteine residues), CCAP by a K-Rx-F-C-N-x(3)-F-T-G-C motif and MIP by a repeated W-x(5,6,7)-WGKR motif (Žitňan et al. 2007;Conzelmann et al. 2013;Zieger et al. 2020). Alignments and histograms were produced in Jalview (Waterhouse et al. 2009). Sequence annotation was confirmed through phylogenetic analyses ( Figure S1, Supplementary Information). For each alignment, best fitting models were determined using ModelFinder (Kalyaanamoorthy et al. 2017). Accordingly, Maximum Likelihood (ML) trees were inferred using IQTREE v1.6.2 (Nguyen et al. 2015) with the VT+R3, VT+I+G4, PMB+G4, and rtREV+F+R3 models for the Bursα/β, CCAP, EH, and MIP trees, respectively. Node support of the ML trees was estimated by 200 bootstrap pseudoreplicates and by Shimodaira-Hasegawa approximate Likelihood Ratio Test. Bayesian Inference trees were inferred using the LG+G4, WAG+G4, WAG +G4, and mtREV+G4 models in PhyloBayes 4.1c (Lartillot et al. 2009) for the Bursα/β, CCAP, EH, and MIP trees, respectively. Two independent chains were run simultaneously for 10000 cycles and convergence was checked using the maxdiff parameter (< 0.1). A majority-rule consensus tree was then built using a sample of 900 trees per chain, discarding the first 10% of the trees as burn-in. ML and BI phylograms were edited in FigTree (http://tree.bio.ed.ac.uk/software/figtree/).

IN SITU HYBRIDIZATION AND IMAGING
Gene-specific primers (Data S2, Supplementary Information) were designed using OligoCalc (Kibbe 2007). PCR amplification, cloning, and ligation were performed as previously described (Fritsch et al. 2015). Probes were generated using digoxigenin and fluorescein RNA Labelling kits (#11277073910 and #11685619910, Roche Diagnostics GmbH, Mannheim, Germany). Probe sequences are provided in (Data S2, Supplementary Information). Whole-mount in situ hybridization was performed as previously described (Fritsch et al. 2015(Fritsch et al. , 2016, but with a probe concentration of 1 ng/μL and a hybridization temperature of 58°C. Larvae were incubated in digoxigenin or fluorescein antibodies conjugated to alkaline phosphatase (#11093274910 and # 11426338910 Roche, 1:3000 dilution) for 24 h at 4°C.

SETTLEMENT AND METAMORPHOSIS
RNA-seq data from six embryonic, six larval, and two juvenile stages of the polyplacophoran mollusk Acanthochitona fascicularis was pseudo-aligned and quantified as TPM (transcripts per million) values in order to compare ERN transcript levels throughout development (Fig. 1a). ERN expression is first upregulated in late trochophore stages and strongly increases during settlement and metamorphosis. In juveniles, ERN transcript levels decrease only slightly. These results correspond well to our in situ hybridization data (Figs. 2-7). ERNs with high TPM values are expressed by more neurons than ERNs with lower TPM values (e.g., MIP by 24 neurons and Bursβ by only two neurons).

SYSTEM
Key steps of A. fascicularis development were documented with special focus on larval settlement, using serotonin (5-HT) and acetylated alpha-tubulin as markers for neurogenesis and for stage-dependent changes in ciliation.
Within 24 h after fertilization, A. fascicularis embryos reach the early trochophore stage (Fig. 1c). The trochophores possess a well-developed apical tuft and a ciliated prototroch, which persist throughout the free-swimming larval phase. Several 5-HTlike immunoreactive (5-HT-lir) cells are present, including four flask-shaped sensory neurons that belong to the apical organ (AO, Fig. 1c) and four ventral sensory neurons (VS, Fig. 1c). All of these neurons project to the cerebral commissure. During later trochophore stages (Fig. 1d,e,h), the number of flaskshaped sensory neurons in the apical organ increases to six and numerous 5-HT-lir somata are detected along the ventral nerve cords. In addition, a pair of dorsolateral 5-HT-lir sensory neurons (DLS, Fig. 1e) is present near the approximate location of the ampullary system, a polyplacophoran-specific larval sensory complex Haszprunar et al. 2002). 5-HTlir neurites can further be observed in the cerebral commissure, the circumesophageal nerve ring, the prototrochal nerve grid, the ventral and lateral nerve cords, and in the commissures that interconnect the main nerve cords.
The onset of settlement ( Fig. 1h-l) is characterized by the almost instantaneous loss of the apical tuft, the prototroch, the apical organ, the prototrochal nerve grid, as well as the dorsolateral and the ventral sensory neurons (Fig. 1h-l). Presence of these structures thus overlaps precisely with the duration of larval swimming in A. fascicularis. Formation of the ciliated foot sole also progresses quickly during early settlement stages (Fig. 1k). Newly settled juveniles (Fig. 1g) lack 5-HT-lir epidermal sensory neurons, but do possess additional somata that exhibit weak 5-HT-lir around the cerebral commissure and a denser network of 5-HT-lir neurites that innervates the foot.

INDIVIDUALLY IDENTIFIABLE NEURONS
In A. fascicularis, ERN expression starts around the midtrochophore stage and is particularly prominent in metamorphic competent trochophores (Figs. 2-3) and settlement stages . All ERNs are expressed in individually identifiable cells that can be traced throughout development. These cells are most likely neurons, since they are arranged in bilaterally symmetrical pairs that localize to major nervous system structures, such as the cerebral commissure, the ampullary system, and the ventral and lateral nerve cords. However, since in situ hybridization rarely labels neurites, it is possible that some cells correspond to endocrine cells instead. Most ERN-expressing neurons occupy basi-epidermal positions and their distribution patterns are highly consistent between specimens (Fig. 2-7). ERNexpressing neurons in the episphere (larval region anterior to the prototroch) are mostly flask-shaped, suggesting sensory roles. Notably, however, no ERN expression could be detected in the apical organ. After settlement and metamorphosis, several ERN neurons are lost together with parts of the episphere, whereas others are retained in early juveniles (Figs. 2-3).

Eclosion hormone
Eclosion hormone (EH) expression begins in the midtrochophore stage in two pairs of bilaterally symmetrical neurons (e4/5) that are situated in the lateral epidermis of the episphere (Fig. 2a). In late trochophores, three more pairs of EH neurons are present in the vicinity of e4/5 (Fig. 2b-c,e), one in a basi-epidermal position (e1) and two in proximity to the brain (e2/3). The basi-epidermal neurons are flask-shaped and belong to the so-called ampullary system (Fig. 2), a larval sensory complex that is autapomorphic for Polyplacophora Haszprunar et al. 2002). Another three to four pairs of EH-expressing neurons are distributed in the hyposphere along the ventral nerve cords, with the anterior-most pair being located just posterior to the prototroch (e6-9, Fig. 2b-c,e). In early  juveniles, EH expression in basi-epidermal neurons is no longer visible (Fig. 2d, e1/4/5). Some specimens show additional cells with weak EH expression that are usually located in close proximity to one (or several) of the above-described neurons (i.e. Fig. 2b, arrowhead).

Crustacean cardioactive peptide (CCAP)
CCAP expression commences in late trochophore stages in five bilaterally symmetrical pairs of flask-shaped sensory neurons ( Fig. 2f-h,k). Two of these CCAP neuron pairs are located in the lateral episphere (c1/2) in the area of the ampullary system Haszprunar et al. 2002), while the other three are located in the ventral episphere (c3-5). In juveniles, CCAP expression is no longer detected laterally in c1/2 but is maintained ventrally in c3-c5 (Fig. 2i-j).

Myoinhibitory peptide (MIP)
MIP expression is first detected in late trochophore stages in 13 bilateral symmetrically arranged pairs of basi-epidermal neurons (m1-m13, Fig. 2l-m,p). While m1-4 are situated in the anterior ventral epidermis, m5-7 are located in slightly more posterior lateral positions and m8-13 are distributed along the ventral nerve cords. After settlement, MIP expression persists in the above-described cells (Fig. 2n-o). Additional neurons are present in the ventral and lateral episphere near m1-6 and in the lateral hyposphere dorsal to m8-10 ( Fig. 2no, arrows). The latter probably correspond to lateral nerve cord neurons. In some specimens, individual MIP neuron pairs were missing (semi-transparent neurons, Fig. 5-7), or additional MIP neurons with weak and possibly transient MIP expression were detected (e.g., Fig. 6e-f). These variable neurons were only observed in the lateral episphere and are likely transient.

Bursicon alpha (Bursα) and bursicon beta (Bursβ)
The expression of Bursα and Bursβ starts in the late trochophore stages. Bursα is expressed in two pairs of bilaterally symmetrical neurons (ba1 and ba2, Fig. 3a-b,d). The ba1 neurons are located in the anterolateral epidermis and are part of the ampullary system (Fig. 3a-d), while the ba2 neurons are located in a more postero-ventral position near the brain. Bursβ is expressed in only one pair of neurons occupying a similar position as the ba2 neurons (bb1, Fig. 3e-h). Expression of both genes persists in juveniles (Fig. 3c,g), although only a very weak signal was obtained.

NEURONS EXPRESSING DIFFERENT ERNS
Double labeling was performed on settlement stages when ERN expression levels are particularly high (Fig. 1a). However, only a weak signal was obtained for Bursα and Bursβ, so that spatial co-expression could not be assessed for these neuropeptides. EH, CCAP and MIP are not spatially co-expressed ( Fig. 4-6), but double in situ hybridization showed that neurons expressing these ERNs tend to cluster together. Several EH (e1-5, Fig. 4), CCAP (c1/2, Figs. 4 and 6), and MIP neurons (asterisks, Fig. 6) are located in close proximity to each other in the lateral episphere. These neurons correspond to the ampullary system ( Fig. 4-6), an autapomorpic sensory structure of chiton larvae Haszprunar et al. 2002). Along the ventral nerve cords, EH neurons (e7-9) are likewise positioned just dorsal to MIP neurons (m9/11/12, Fig. 5, H-K). Most notably, each CCAP neuron (c1-5) is in direct contact with at least one MIP neuron (m1-4 and asterisks, Fig. 6).  Scopelliti et al. 2016;Žitňan and Daubnerová 2016;Zhang et al. 2017;Li et al. 2019;Williams 2020). Furthermore, a recent survey of temporal ERN expression profiles supports their ancestrally conserved involvement in key life cycle transitions, such as hatching and metamorphosis (Zieger et al. 2020). Consistent with this hypothesis, ERN transcript levels peak in settlement and metamorphosis stages of the polyplacophoran mollusk A. fascicularis (Fig. 1a).

Discussion
Although major neural structures are already established in early and mid-trochophore stages (Fig. 1b-e) Haszprunar et al. 2002;Voronezhskaya et al. 2002), our in situ hybridization experiments confirm that most ERNexpressing neurons form only in late trochophores, just before settlement (Figs. 2-3). Therefore, their appearance coincides with the acquisition of metamorphic competence. Some ERNexpressing neurons are lost towards the end of metamorphosis (Fig. 7), which strongly suggests that they serve a temporary purpose during this process. These data corroborate our previous findings on ERN upregulation during key life cycle events in Nephrozoa (Zieger et al. 2020).
To exert different effects, neuropeptides are released either synaptically as neurotransmitters/neuromodulators or as hormones into the circulatory system. Accordingly, arthropod ERNs are not only expressed by neurons and neurosecretory cells of the central and peripheral nervous system, but also by non-neural tissue types, including the midgut, the tracheal system, and specific endocrine glands (Davis 2003;Žitňan and Daubnerová 2016;Scott et al. 2020). In contrast, A. fascicularis ERNs were only detected in bilateral symmetrically arranged pairs of cells, which we consider to be neurons based on their distribution and cell morphology (Fig. 7a). Here, ERNs most likely act as neurotransmitters and/or neuromodulators rather than as neurohormones. Many of these ERN-expressing neurons are flask-shaped and occupy a basi-epidermal position, which further implies sensory functions.
The larvae of A. fascicularis are non-feeding and most likely initiate settlement primarily based on chemical cues, as has been shown for several polyplacophoran species (Barnes and Gonor 1973;Lord 2011). Thus, the temporary EH-and CCAPexpressing neurons of the ampullary system (Fig. 7a) are prime candidates for sensory cells that promote settlement and metamorphosis in response to suitable conditions. A similar task has been assigned to MIP-expressing neurons in the larval apical organ of annelids (Conzelmann et al. 2013;Hou et al. 2020). MIP-related neuropeptides of the Wamide superfamily are among the most extensively studied across metazoans and their ancestral role in metamorphic hormone signalling is well established (Schoofs and Beets 2013;Williams 2020). Yet, our findings suggest that such complex developmental transitions depend not only on MIP, but on the combined actions of several highly conserved neuropeptides, including different ERNs.
To our surprise, none of the investigated ERNs are expressed in the larval apical organ of A. fascicularis. The apical organ is considered responsible for controlling settlement and metamorphosis in various taxa (Kempf et al. 1997;Hadfield et al. 2000;Voronezhskaya and Khabarova 2003;Voronezhskaya et al. 2004;Rentzsch et al. 2008;Conzelmann et al. 2013;Hou et al. 2020), but its relatively early and rapid disintegration (Fig. 1,  b-l, Fig. 7b) argues against a major involvement in these processes in A. fascicularis. A similar situation has been reported in the scaphopod Antalis entalis, where the larval apical organ is lost even before metamorphic competence is achieved (Wanninger and Haszprunar 2003). Moreover, a recent study on the polychaete Hydroides elegans showed that laser ablation of the larval apical organ does not prevent settlement and metamorpho-sis (Nedved et al. 2021). Accordingly, involvement in larva-tojuvenile transitions might not be as important or as widespread as is often hypothesized. Our data instead suggest that these complex developmental processes are regulated by taxon-specific sets of neurons, such as the transient ERN-expressing neurons of the polyplacophoran-specific ampullary system.
In addition to these transient neurons, other ERN-expressing cells persist in A. fascicularis juveniles (Fig. 7). Our RNA-seq data corroborate this observation, since juveniles maintain a relatively high expression level for most ERNs (Fig. 1). This points towards a continued requirement of ERNs, possibly for the adoption of adult-like behaviours such as feeding and crawling. Contrary to the situation in Drosophila (Nässel and Zandawala 2019), spatial co-expression of ERNs was not detected in A. fascicularis. Rather, subsets of ERN-expressing neurons are located in close proximity to one another, which suggests that they contribute to the same processes. CCAP and MIP exert myoexcitatory and myoinhibitory effects, respectively, and both have been shown to affect feeding in various animals (Žitňan and Daubnerová 2016;Williams 2020). Thus, it is likely that the abutting CCAP-and MIP-expressing neurons anterior to the mouth opening of A. fascicularis larvae and juveniles (Fig. 6d,h) constitute neural microcircuits that influence food-intake. Likewise, the regular arrangement of EH-and MIP-expressing neuron pairs along the ventral nerve cords corresponds to that of the seven pairs of Importantly, a seven-fold seriality is not only typical for (larval and early juvenile) polyplacophoran muscular and skeletal elements, but is also considered ancestral for Aculifera (the monophyletic mollusk lineage that unites the polyplacophorans with the worm-like, shell-less aplacophorans) (Scherholz et al. 2013(Scherholz et al. , 2015. Interspaced pairs of MIP neurons are also present along the ventral side of annelids (Conzelmann et al. 2013;Williams et al. 2015) and the ventral nerve cord of arthropods (Santos et al. 2007;Nässel and Zandawala 2019). Given the pre-bilaterian origin of MIP-like neuropeptides and their putative loss in deuterostomes (Williams 2020), such a pattern could be a plesiomorphic feature of protostomes that was later adapted to match different body plans and muscular arrangements. Detailed neuroanatomical comparisons between additional species will be necessary to address these deep evolutionary questions. Our findings highlight that ERNs are promising targets for this purpose due to their expression in individually identifiable neurons across protostomes.

Conclusion
Our study shows that different subsets of ERN-expressing neurons serve distinct functions during polyplacophoran development. Some are only present during settlement and metamorphosis, while others persist in juveniles and probably contribute to the onset of adult behaviours such as feeding and muscular locomotion. Consistent with data from other nephrozoan clades (Zieger et al. 2020), our findings suggest that orchestrating the larvato-juvenile transition involves multiple ERNs in Acanthochitona fascicularis. Especially the transient ERN-expressing neurons of the polyplacophoran-specific and strictly larval ampullary system are likely important for this process. Absence of ERN expression in the apical organ of A. fascicularis as well as the relatively early and rapid loss of this structure in both polyplacophorans and scaphopods challenges the common notion that settlement and metamorphosis are chiefly regulated by the larval apical organ across protostomes. Instead, we propose that this complex task is carried out through interaction of several distinct subsets of neuropeptidergic sensory neurons, whose distributions and precise roles are likely taxon-specific, warranting further investigation in comparative studies. EMBRC France for facilitating our access to important marine biological resources and research infrastructures. We would further like to thank Alexander Vogt from Vienna BioCenter Core Facilities (VBCF) for technical help with RNA-seq. The research leading to the presented results has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No 730984 to EZ. Parts of this research were supported by the Austrian Research Fund (FWF) grant P29455-B29 to AW and N.S.M.R. was supported by the Austrian Science Fund grant P3219.

DATA ARCHIVING
The Acanthochitona fascicularis sequence data generated and analysed during the current study are available from NCBI (BioProject ID: PR-JNA706890). All other data generated or analysed during this study are included in this published article and its supporting information files. Command line details of all bioinformatic steps are provided in Data S1, Supplementary Information.

CODE AVAILABILITY
Command line details of all bioinformatic steps are provided in Data S1, Supplementary Information.