Discovery of novel representatives of bilaterian neuropeptide families and reconstruction of neuropeptide precursor evolution in ophiuroid echinoderms

Neuropeptides are a diverse class of intercellular signalling molecules that mediate neuronal regulation of many physiological and behavioural processes. Recent advances in genome/transcriptome sequencing are enabling identification of neuropeptide precursor proteins in species from a growing variety of animal taxa, providing new insights into the evolution of neuropeptide signalling. Here, detailed analysis of transcriptome sequence data from three brittle star species, Ophionotus victoriae, Amphiura filiformis and Ophiopsila aranea, has enabled the first comprehensive identification of neuropeptide precursors in the class Ophiuroidea of the phylum Echinodermata. Representatives of over 30 bilaterian neuropeptide precursor families were identified, some of which occur as paralogues. Furthermore, homologues of endothelin/CCHamide, eclosion hormone, neuropeptide-F/Y and nucleobinin/nesfatin were discovered here in a deuterostome/echinoderm for the first time. The majority of ophiuroid neuropeptide precursors contain a single copy of a neuropeptide, but several precursors comprise multiple copies of identical or non-identical, but structurally related, neuropeptides. Here, we performed an unprecedented investigation of the evolution of neuropeptide copy number over a period of approximately 270 Myr by analysing sequence data from over 50 ophiuroid species, with reference to a robust phylogeny. Our analysis indicates that the composition of neuropeptide ‘cocktails’ is functionally important, but with plasticity over long evolutionary time scales.

indicates that the composition of neuropeptide "cocktails" is functionally important, but with 1 0 5 plasticity over long evolutionary time scales. Here, we have utilized transcriptome sequence data for the first comprehensive identification 1 0 8 of neuropeptide precursors in ophiuroids (Figure 1). Representatives of over thirty bilaterian 1 0 9 neuropeptide precursor families were identified. Identification of ophiuroid representatives of 1 1 0 these neuropeptide precursor types has in some cases provided new insights into neuropeptide 1 1 1 precursor structure and evolution, as discussed in more detail below. First, however, we will 1 1 2 highlight representatives of bilaterian neuropeptide precursor families that have been  Comprehensive analysis of transcriptome sequence data from three ophiuroid species,  Eclosion hormone (EH) was first isolated and sequenced in the insects Manduca sexta 1 2 6 (tobacco hornworm) and Bombyx mori (silk moth) and shown to alter the timing of adult 1 2 7 emergence [33,34]. EH is one of the main peptide/protein hormones involved in control of 1 2 8 ecdysis (i.e. shedding of the cuticle) behavior in arthropods [35,36]. It binds to and activates 1 2 9 a receptor guanylyl cyclase that is expressed in epitracheal Inka cells and causes the 1 3 0 secondary release of ecdysis-triggering hormone (ETH) that is also expressed in Inka cells insect EHs, but aside from this there is little sequence conservation (Figure 2A). The 1 4 4 echinoderm EH-like precursor sequences were also analysed using a sequence-similarity- BLAST results with arthropod EHs and, to a lesser extent, with arthropod ion transport 1 4 9 17 of the long and short peptides occurred before the divergence of the Asterozoa and Further studies are now needed to identify the proteins that act as receptors for the KP-type 4 0 4 peptides identified here in ophiuroids and previously in other echinoderms [8]. [110]. In vertebrates, calcitonin is produced by the thyroid gland [111] and regulates calcium 4 1 0 (Ca 2+ ) levels in the blood, antagonizing the effects of parathyroid hormone [112,113]. The bridge that is characteristic of calcitonin-type peptides in vertebrates. More recently, it has 4 1 5 been discovered that both DH 31 -type and vertebrate calcitonin-type neuropeptides occur in 4 1 6 some protostomian invertebrates, including the annelid Platynereis dumerilii and the insect 4 1 7 Locusta migratoria [4,115]. Hence, it is proposed that an ancestral-type calcitonin precursor 4 1 8 gene duplicated in the common ancestor of protostomes to give rise to DH 31 -type and 4 1 9 calcitonin-type peptides, but with subsequent loss of calcitonin-type peptides in some 4 2 0 protostomes. Consistent with this hypothesis, calcitonin-type precursors but not DH 31 -type  18 related peptide [117]. Furthermore, a complex interplay of receptors and accessory proteins 4 2 7 determines the pharmacological profile of these peptides [118,119]. Alternative splicing of 4 2 8 DH 31 and calcitonin precursors in insects has also been previously reported [115,120,121].

3 2
Our analysis of the ophiuroid transcriptomes also identified two transcript variants for these three coding regions represent three or more exons due to the lack of genomic data, but 4 3 6 for the sake of simplicity, we refer to them here as 'exons'. Transcript variant 1 comprises 4 3 7 'exons' 1 and 3 but lacks 'exon' 2 whereas transcript variant 2 contains all 3 'exons'. Interestingly, 'exons' 2 and 3 both encode a calcitonin-type peptide. Hence, transcript variant 4 3 9 1 encodes a precursor that produces one calcitonin-type peptide and transcript variant 2 4 4 0 encodes two non-identical calcitonin-type peptides. These alternatively spliced transcripts 4 4 1 were found in several brittle star species (Figure 8)  two cysteine residues) ( Figure S5). It is difficult to predict the structural differences that may 4 6 4 arise in the heterodimer due to this variability in the number of cysteine residues. The  In addition to the neuropeptides discussed above, we have also identified three 4 7 2 neuropeptide precursors that could not be classified into any known neuropeptide families.   filiformis are predicted to generate eight copies of the mature peptide. Some of the mature 4 9 3 peptides have a C-terminal SGW motif, which is similar to the C-terminus of predicted 4 9 4 mature peptides derived from O. victoriae pedal peptide precursor 3 ( Figure S7). However, 4 9 5 the lack of sequence similarity in other parts of the peptide suggests that the C-terminal 4 9 6 similarity may reflect convergence rather than homology. Ovnp27 (GenBank; MF155251) was identified following a HMM-based search for 5 0 0 SIFamide-type peptides [127,128], albeit with a high E-value. This neuropeptide precursor 5 0 1 comprises two putative amidated mature peptides that are located immediately after the signal  victoriae precursor has a C-terminal IFamide motif just like in insect SIFamides ( Figure S9).

0 4
However, there is no sequence similarity with SIFamides in the rest of the peptide. This

21
[6] suggests that the sequence similarity that peptides derived from Ovnp27-type precursors 5 0 7 share with SIFamides may reflect convergence rather than homology.

Neuropeptide precursors not found in brittle stars
Our analysis of ophiuroid transcriptome sequence data did not reveal orthologs of the 5 1 1 Spnp9 precursor from S. purpuratus or the Arnp21, Arnp22, Arnp23 and Arnp24 precursors 5 1 2 from A. rubens [8,16]. An Spnp9 ortholog is found in A. japonicus but not in A. rubens [15] 5 1 3 and therefore this neuropeptide precursor type may be restricted to the Echinozoa. Orthologs suggests that these may be Asteroidea-specific precursors.

1 6
Previous studies have shown that receptors for leucokinin, ecdysis-triggering ligands should also be present in ambulacraria. However, our search approaches failed to 5 2 0 identify any proteins in ophiuroids that resemble precursors of these neuropeptides.

2 3
Many neuropeptide precursors comprise several structurally similar but non-identical 5 2 4 bioactive peptides -i.e. the precursor protein gives rise to a neuropeptide "cocktail". This These are important questions in neuroendocrinology for which answers remain elusive. Evidence that neuropeptide copy number may be functionally important has been peptide copies in each neuropeptide precursor was found to be identical (except for the 5 3 3

22
FMRFamide precursor) when compared between the twelve species, suggesting that 5 3 4 stabilising selection has acted to conserve neuropeptide "cocktails" in the Drosophila lineage. precursors and their putative mature peptides revealed that fourteen neuropeptide precursors 5 3 7 comprised multiple neuropeptide copies. In certain cases, the number of the mature peptides 5 3 8 derived from a particular precursor varied across species, whereas in other cases the numbers 5 3 9 remained constant (Figure 4). Interestingly, these three species belong to two of the three Gnathophiurina infraorder (clade C). Hence, this prompted us to examine the evolution of resolution. To do this, we utilized a unique dataset comprising 52 ophiuroid transcriptomes.

4 5
These transcriptomes were recently used as part of a phylotranscriptomic approach to 5 4 6 reconstruct the phylogeny of ophiuroids, generating a robust phylogenetic tree that comprises 5 4 7 three major clades [12]. Hence, this dataset allowed us to explore the evolution of over an unprecedented timescale of ~270 million years. We selected for analysis neuropeptide precursors comprising more than a one putative 5 5 1 mature neuropeptide, which include AN peptide, calcitonin, cholecystokinin 1, kisspeptin, 5 5 2 np18, np26, np27, NG peptide, PDF, SALMFamide (L-type and F-type), tachykinin and TRH (1 and 2). Pedal peptide precursors (1, 2 and 3) were excluded from the analysis because 5 5 4 orthology relationships between these precursors could not be established with confidence 23 constant across all the species examined, which share a common ancestor estimated to date 5 6 1 from ~270 million years ago [12].

6 2
Some studies that have investigated the physiological significance of neuropeptide 5 6 3 "cocktails" indicate that neuropeptides derived from the same precursor protein are 5 6 4 functionally redundant. For example, this was found for myomodulin neuropeptides in the 5 6 5 mollusk Aplysia californica using the accessory radula closer muscle preparation as a  However, the authors of the latter study cautiously highlighted the need to "search for additional functions or processes in which these peptides may act differentially". Importantly, Inhibitory Peptide-related peptides in Aplysia were tested on three organ preparations 5 7 4 (oesophagus, penis retractor, body wall) it was found that the rank order of potency for the or structurally related neuropeptides is functionally important. For those neuropeptide precursors that did exhibit variation in neuropeptide copy to 20 copies (Figure 9). F-type SALMFamide precusors also showed variation in copy precursors than in TRH-type precursors. Furthermore, detailed analysis of sequence 5 8 5 alignments for these precursors revealed that loss of neuropeptide copies is usually a 5 8 6 consequence of non-synonymous mutations in codons for residues that form dibasic cleavage 5 8 7 24 sites or for glycine residues that are substrates for the C-terminal amidation. This is not 5 8 8 surprising since the C-terminal amide in smaller-sized peptides is usually important for 5 8 9 receptor binding and activation. What is unclear at the moment is how the peptide copy 5 9 0 number increases within a given precursor. Perhaps the increase in peptide copy number 5 9 1 occurs as a result of unequal crossing-over during recombination [130].

9 2
The number of peptides within the F-type SALMFamide precursors appear to be clade 5 9 3 specific. Thus, the average/median number of F-type SALMFamides in precursors from clade 5 9 4 A is 13, clade B is 12 and clade C is 11, with a few exceptions (Figure 8). Similarly, the 5 9 5 number of peptides within NP26-type precursors also appears to be clade specific. Hence the 5 9 6 number of peptides is highly stable at 7 peptides within clades A and B but a high variation in 5 9 7 peptide copy number is observed in clade C. When examining peptide copy number within certain species appears to be an exception/outlier. For instance, 16 copies of the mature 6 0 0 peptide in Ophioplax lamellosa TRH-1 precursor is distinctly different to the 19 copies found It could be argued that misalignments during transcriptome assembly may have 6 0 4 influenced the number of predicted peptides found in a given precursor. However, it is 6 0 5 unlikely that misalignments have affected the predicted sequences of neuropeptide precursors 6 0 6 comprising multiple copies of peptides that are similar but non-identical, which applies to the 6 0 7 majority of the precursor proteins analysed here in ophiuroids. The only exception to this are 6 0 8 the TRH-type precursors, where the encoded peptide sequences are short and often identical, 6 0 9 even at the nucleotide level (data not shown), Another limitation of using transcriptome data 6 1 0 is that the sequences of neuropeptide precursors may be partial or unknown for some species 25 precursor types, sequence data was obtained from a variety of species from each of the three 6 1 5 clades of ophiuroids. For example, complete F-type SALMFamide precursor sequences were 6 1 6 found in most of the investigated species (39 species + 3 reference species). Here we report the first detailed analysis of the neuropeptide precursor complement of "cocktails" is functionally important as evident from the conservation of neuropeptide copy 6 2 7 number for multiple precursors. Ophiuroid transcriptomes used in this study were sequenced and assembled as

TN -HC KG R L--PKF CFLH Pa O v i c 2 R G -I C SD----PLAC GAA Fa A r ub SR -RC S -----VKGCM VH F a A j a p K S -A CS N R H --PK LCIL H Pa D m el _ C C H 1 ---SC LE Y ---GHS CWGA Ha D m e l _CC H 2 ---G C QAY ---GHVC YGG Ha A m el _ C C H 1 ---SC L SY ---GHSCW GA H a A m e l _CC H 2 ---G CS A F ---GH SCFG G Ha L g ig _ G G N ---K CSG R WA -IHA CFGG Na A c a l _L1 1 P R ID C TRF VF-APAC RGV SA A m el _ L 1 1 ES V NC E LY P F-HHTCR GT M S H s a p _ED N 3
----CT C F T Y K DK ECVY Y CH L DI I W * . . . * . .      Ophiuroidea Am_cipu