Myoinhibitory peptide regulates feeding in the marine annelid Platynereis

During larval settlement and metamorphosis, marine invertebrates undergo changes in habitat, morphology, behavior and physiology. This change between life-cycle stages is often associated with a change in diet or a transition between a non-feeding and a feeding form. How larvae regulate changes in feeding during this life-cycle transition is not well understood. Neuropeptides are known to regulate several aspects of feeding, such as food search, ingestion and digestion. The marine annelid Platynereis dumerilii has a complex life cycle with a pelagic non-feeding larval stage and a benthic feeding postlarval stage, linked by the process of settlement. The conserved neuropeptide myoinhibitory peptide (MIP) is a key regulator of larval settlement behavior in Platynereis. Whether MIP also regulates the initiation of feeding, another aspect of the pelagic-to-benthic transition in Platynereis, is currently unknown. Here, we explore the contribution of MIP to the regulation of feeding behavior in settled Platynereis postlarvae. We find that in addition to expression in the brain, MIP is expressed in the gut of developing larvae in sensory neurons that densely innervate the hindgut, the foregut, and the midgut. Activating MIP signaling by synthetic neuropeptide addition causes increased gut peristalsis and more frequent pharynx extensions leading to increased food intake. Conversely, morpholino-mediated knockdown of MIP expression inhibits feeding. In the long-term, treatment of Platynereis postlarvae with synthetic MIP increases growth rate and results in earlier cephalic metamorphosis. Our results show that MIP activates ingestion and gut peristalsis in Platynereis postlarvae. MIP is expressed in enteroendocrine cells of the digestive system suggesting that following larval settlement, feeding may be initiated by a direct sensory-neurosecretory mechanism. This is similar to the mechanism by which MIP induces larval settlement. The pleiotropic roles of MIP may thus have evolved by redeploying the same signaling mechanism in different aspects of a life-cycle transition.

Results: Here, we explore the contribution of MIP to the regulation of feeding behavior in settled Platynereis postlarvae. We find that in addition to expression in the brain, MIP is expressed in the gut of developing larvae in sensory neurons that densely innervate the hindgut, the foregut, and the midgut. Activating MIP signaling by synthetic neuropeptide addition causes increased gut peristalsis and more frequent pharynx extensions leading to increased food intake. Conversely, morpholino-mediated knockdown of MIP expression inhibits feeding. In the long-term, treatment of Platynereis postlarvae with synthetic MIP increases growth rate and results in earlier cephalic metamorphosis. Conclusions: Our results show that MIP activates ingestion and gut peristalsis in Platynereis postlarvae. MIP is expressed in enteroendocrine cells of the digestive system suggesting that following larval settlement, feeding may be initiated by a direct sensory-neurosecretory mechanism. This is similar to the mechanism by which MIP induces larval settlement. The pleiotropic roles of MIP may thus have evolved by redeploying the same signaling mechanism in different aspects of a life-cycle transition.

Background
Many organisms have a complex life cycle consisting of distinct stages that differ in form, physiology, behavior and habitat. Among benthic marine invertebrates, a common life cycle strategy, the biphasic life cycle, consists of a free-swimming larva that settles to the ocean floor and undergoes metamorphosis to a bottom-dwelling adult [1]. Marine invertebrate larval settlement is often coupled to the initiation of feeding or a change in diet [2]. These behavioral, physiological and morphological changes have to be tightly coordinated for the successful transition to a benthic life style. Knowledge of how this transition is regulated is important for understanding population structure in the ocean, life-history evolution, and how the environment influences animal life cycles [3][4][5][6][7][8].
The marine annelid Platynereis dumerilii has recently proven to be a useful marine invertebrate model for studying the molecular details of marine larval behavior, including settlement [9][10][11][12]. Platynereis has a biphasic life cycle with free-swimming, non-feeding larval (trochophore and nectochaete) stages and bottom-dwelling, feeding postlarval, juvenile and adult stages [13]. Larval settlement is followed by a period of growth and feeding, during which juvenile Platynereis add additional posterior segments. Cephalic metamorphosis, in which the first pair of parapodia are transformed into a second pair of tentacular cirri on the head, occurs after the juveniles have begun to add their sixth posterior segment [13][14][15][16].
Recently, we identified myoinhibitory peptide (MIP) as an inducer of rapid larval settlement behavior in Platynereis [11]. MIP is expressed in anterior chemosensoryneurosecretory neurons of the larva. Exogenous application of MIP inhibits the activity of the locomotor cilia, resulting in rapid sinking, and induces sustained contact with the substrate. Platynereis MIP belongs to an ancient neuropeptide family of Wamides, which are characterized by their amidated C-terminal tryptophan residue preceded by a small aliphatic residue [11,17]. Wamides are widespread among eumetazoans, except deuterostomes, and recently emerged as conserved regulators of life-cycle transitions [18]. For example, in some insects, MIP (also known as prothoracicostatic peptide (PTSP) or allatostatin-B (AST-B)) regulates ecdysone [19][20][21] and juvenile hormone levels [22], potentially influencing the timing of larval ecdysis and pupation. In cnidarians, including some corals and hydrozoans, GLWamide (also called metamorphosin) is known to induce larval settlement and metamorphosis [23][24][25].
How changes in feeding are regulated during marine life-cycle transitions is less well understood. Many neuropeptides are known to have roles in regulating different aspects of feeding [26][27][28]. MIPs/Wamides are also pleiotropic [29][30][31][32][33][34][35] and regulate aspects of feeding and gut muscle activity in some insects and cnidarians. The first MIP described had a myoinhibitory function on adult locust hindgut [36]. In several insects, MIP is expressed in the adult stage and can suppress muscle contractions of the hindgut [36][37][38][39][40]. MIP is also expressed in the stomatogastric nervous system of the adult crab, Cancer borealis, where it decreases the frequency of pyloric rhythm [41,42]. In addition, cnidarian GLWamide increases myoactivity in both hydra and sea anemone polyps, potentially influencing feeding [43,44]. Although none of these studies directly quantified feeding in the whole organism, MIP is a strong candidate for the regulation of feeding during marine life-cycle transitions.
Here, we study the expression and function of MIP in Platynereis during late larval (3-6 days) and early juvenile development (6-30 days). We found MIP expression in sensory neurons of the gut of 6 days and older Platynereis. We used both peptide-soaking and morpholino-mediated knockdown approaches to establish a role for MIP in the regulation of postlarval feeding and gut peristalsis. MIP treatment also resulted in faster juvenile growth, probably as a consequence of increased food ingestion and gut movement. Our results establish MIP as a pleiotropic neuropeptide in Platynereis that links behavioral and physiological components of a lifecycle transition.

Results
Platynereis MIP is expressed in the brain and gut of postlarvae, juveniles and adults Expression profiling of the MIP precursor gene by RNA in situ hybridization showed that MIP is expressed during both larval and postlarval development and continues to be expressed after cephalic metamorphosis, in the early adult stage (Figure 1A-C, G-H; Additional file 1). At 6 days and older, MIP is expressed in both the median brain and the trunk nervous system, in paired cells and also in single cells closer to the larval midline. We also found MIP expression in the digestive system, in the fore-, midand hindgut. The different regions of the gut are delineated by the differential expression patterns of Platynereis digestive system marker genes ( Figure 1D, I; discussed below). The MIP-expressing cells in the gut have sensory dendrites that project toward the lumen of the gut (Figure 1B-C; Additional files 2, 3). In some of these dendrites we could even detect the MIP RNA in situ signal, allowing the unambiguous assignment of these acetylated tubulin-positive cellular projections to the MIP-expressing cells (Additional file 3).
In addition to MIP, we also attempted to characterize the expression of the MIP receptor in 6 dpf and older larvae. We previously described the expression of the MIP receptor in the head of 2 dpf Platynereis larvae [11], however, in older larvae and postlarvae, the levels of MIP receptor expression proved too low to detect reliably with our RNA in situ hybridization method. The low expression of the Platynereis MIP receptor is typical of most G protein-coupled receptor expression levels [45].
Immunostaining with an antibody against Platynereis MIP showed that in addition to the neurosecretory plexus of the brain, MIP peptide is transported throughout the ventral nerve cord. At 6 days post fertilization (dpf), MIP-expressing neurons in the digestive system innervate both the foregut and hindgut ( Figure 1E, F). As larvae progress from 3 to 6 dpf, during which time the digestive system develops, MIP expression first emerges in the developing hindgut at 4 dpf, followed by the expression in the foregut at 6 dpf (Additional file 4A-L). By one month, MIP-expressing cells densely innervate the entire length of the gut, forming a nerve-net. Using an antibody against the conserved C-amidated dipeptide VWamide [11], we found similar immunolabeling in the brain, ventral nerve cord and gut of larvae of Capitella teleta, a distantly related annelid species [46] (Additional file 4M-P).
By combining phalloidin staining and MIP immunostaining in Platynereis 1 month post fertilization (1 mpf), we could assess the location of MIP-expressing neurons in the gut in relation to the digestive system musculature. In the foregut and in the sphincter muscle that separates foregut from hindgut, MIP-expressing neurons are intermingled with the muscle tissue of the pharynx and sphincter (Figure 2A-C). In the mid-and hindgut, MIPexpressing neurons sit in the inner epithelial cell layer underlying the smooth muscles of the gut ( Figure 2D-K). The axons of the MIP-expressing cells in the mid-and hind-gut of 1 month post fertilization (mpf) Platynereis run parallel to and just beneath the muscle fibers of both circular and longitudinal smooth muscles ( Figure 2I-K). The spatial expression patterns of Platynereis MIP and MIP peptide suggest a potential role for MIP signaling in feeding and digestion during larval and early juvenile stages of the life cycle.
Characterization of normal gut development and the initiation of feeding in postlarvae At 6 dpf Platynereis postlarvae have a through gut with clearly recognizable fore-, mid-and hindgut regions ( Figure 1). The foregut contains the muscular and extendable pharynx with the jaws and salivary glands (Additional file 2). Phyllodocid polychaetes, such as Platynereis, have an axial muscular pharynx consisting of circular, longitudinal and radial muscle fibers, which allow for complex sucking and swallowing movements [47,48]. The foregutmidgut boundary is marked by the presence of a sphincter  muscle. This muscle showed regular contractions in larvae expressing a genetically encoded calcium indicator GCaMP6 (Additional file 5). The broad midgut does not show regionalization and is followed by a short and narrow hindgut.
To gain insight into the morphology and maturation of the Platynereis digestive system, we carried out wholemount RNA in situ hybridization on 6 dpf, 14 dpf and 1 mpf Platynereis with marker genes selected from an ongoing broad RNA in situ hybridization screen, based on their expression in the digestive system at 6 dpf. The digestive system marker genes were identified through domain conservation, reciprocal BLAST and phylogenetic analyses as: extracellular digestive enzymes, peptidases subtilisin-1 and subtilisin-2 (Peptidase_S8; Pfam domain: PF00082), the protease enteropeptidase, responsible for the activation of proteolytic enzymes [49], the polysaccharidedigesting enzyme alpha-amylase, and the intracellular digestive enzyme legumain protease precursor ( Figure 1D, I; Additional files 6,7,8). Alpha-amylase and subtilisin-1 expression was restricted to the midgut at 6 dpf, but expanded to mid-and hindgut at 14 dpf and 1 mpf. Legumain protease precursor was constantly expressed in both mid-and hindgut from 6 dpf to 1 mpf, while subtilisin-2 expression was restricted to the midgut from 6 dpf to 1 mpf. Enteropeptidase was the only gene with expression in the foregut, including the salivary glands, at 6 and 14 dpf. At 1 mpf, enteropeptidase remained strongly expressed in the foregut, but expression also extended to the midand hindgut. Registration of these marker gene expression patterns [50] at 6 dpf to a common nuclear stain reference scaffold, along with the average 6 dpf MIP expression, highlighted the close association of MIP-expressing cells with the digestive system at this stage ( Figure 1D; Additional file 9).
We also looked at the change in expression of these digestive system marker genes and MIP across the Platynereis life cycle in stage-specific RNA-seq datasets [51].
With the exception of legumain protease precursor, the expression of all digestive system marker genes was undetectable in non-feeding larval stages but sharply increased between 4 and 10 dpf (Additional file 10). In accordance with a digestive function, these genes were strongly down-regulated in the adult non-feeding epitokes. MIP expression also increased sharply between 4 and 10 dpf, although it continued to be expressed in the non-feeding epitokes, suggesting further functional roles beyond feeding in Platynereis.
Following settlement, Platynereis larvae have been reported to begin feeding between 5 -8 dpf, with considerable variation between individuals [13,16]. Due to this variability, we decided to document feeding initiation in our own laboratory culture ( Figure 3A). We added Tetraselmis marina microalgae to the larval cultures and documented feeding based on chlorophyll fluorescence in the gut ( Figure 4C). Most larvae initiated feeding between 6 and 7 dpf; by 8 dpf, nearly all larvae had started feeding ( Figure 3A).

Knockdown of MIP delays the initiation of feeding in Platynereis larvae
To explore the function of MIP in the Platynereis digestive system, we employed morpholino microinjection to knockdown MIP expression. We used two different translation blocking morpholinos and two mismatch control morpholinos (Additional file 11). To test the effectiveness of MIP-knockdown, we immunostained knockdown and control larvae with an antibody against Platynereis MIP. We observed a strong reduction in MIP immunostaining in Platynereis MIP-knockdown larvae, but not in controls, up to at least 6 dpf ( Figure 3D-G, Additional file 12). These experiments confirmed that the MIP morpholinos were capable of strongly reducing MIP expression. Next, we documented feeding in MIP-knockdown and control larvae. Similar to untreated larvae, most larvae injected with a control morpholino had initiated feeding between 7-9 dpf, whereas a significantly lower number of MIP-knockdown larvae had food in the gut at 7-9 dpf. This effect was still observed between 10-12 dpf ( Figure 3B-C).
To rule out that the reduced feeding in MIP-knockdown larvae is due to a developmental delay, we compared the morphology of the nervous system of MIP-knockdown and control larvae. There were no detectable differences in the nervous system of control larvae and MIPknockdown larvae based on acetylated tubulin immunostainings at 6 dpf (Figure 3D-G; Additional file 13). We also treated uninjected larvae at different ages between 24 hours post fertilization (hpf ) and 5 dpf with synthetic MIP peptide to see whether MIP-treated larvae initiate feeding sooner, indicating a potential developmental acceleration. MIP treatment did not significantly alter the timing of feeding initiation, even when food was available earlier than 5 dpf (Additional file 14). These experiments indicate a critical physiological role for MIP in the initiation of feeding behaviour in Platynereis postlarvae.
MIP treatment has a myostimulatory effect on the digestive system of Platynereis postlarvae In order to understand how the morpholino knockdown of MIP resulted in reduced larval feeding, we examined the effect of synthetic MIP treatment on postlarvae, focusing on the digestive system. Treatment of 6.5 dpf postlarvae with synthetic MIP caused a significant increase in gut To determine whether these effects on gut and pharynx movement resulted in increased ingestion of algal cells in MIP-treated postlarvae, we then scored the number of algal cells consumed by MIP-treated versus control 7 dpf postlarvae. Treatment with 5 μM and 20 μM MIP significantly increased postlarval algal cell consumption ( Figure 4C, F; Additional file 17). Additionally, MIPtreated postlarvae have decreased locomotion, indicating a switch in the nervous system from a locomotory to a feeding program (Additional file 18). These results show that MIP up-regulates feeding activity and gut peristalsis.

Long-term MIP treatment enhances growth in Platynereis postlarvae
Given the effect of MIP on the digestive system and feeding in Platynereis postlarvae, we next investigated the long-term effects of MIP treatment on postlarval growth. At approximately two weeks of age, feeding Platynereis begin to add new posterior segments [13]. After the development of the 5 th segment, juveniles undergo cephalic metamorphosis, a morphogenetic process in which the first chaetigerous segment loses its chaetae, develops a pair of tentacular cirri and fuses with the head (Figure 5A-D). The timing of cephalic metamorphosis and the addition of new segments vary between individuals. Even juveniles cultured individually showed variation in the timing of posterior segment addition, with segment number varying between 4 and 8 segments at 34 dpf (Additional file 19D). On a diet of Tetraselmis, the shortest interval for an individually-raised juvenile to develop an additional posterior segment was 4 days. The addition of new segments required that larvae begin to feed. Unfed larvae never develop beyond the 3-segmented stage (Additional file 19E). Growth in other nereid species depends on culture density [52][53][54][55]. We documented the growth of Platynereis juveniles cultured at different densities with excess food and determined the maximal density that still allowed optimal growth (3 larvae/ml) (Additional file 19 A-C). Under these conditions, juvenile Platynereis begin to develop the 5 th segment at 16 dpf, and start to undergo cephalic metamorphosis at 24 dpf. Morpholino knockdown methods are not applicable to such late stage animals, therefore we tested the effects of MIP treatment on errant juvenile growth. We found that the time to the addition of new posterior segments, and to cephalic metamorphosis, was reduced by sustained exposure to five different versions of mature MIP encoded by the Platynereis MIP preproneuropeptide gene ( Figure 5E, F). At 25 dpf, some MIP-treated individuals had completed cephalic metamorphosis, while control individuals were yet to undergo cephalic metamorphosis. Comparing the body length of MIP-treated and control Platynereis at 25 dpf revealed that MIP-treated individuals were on average approximately 100 μM longer than control individuals ( Figure 5G). The effect of MIP treatment on growth was only seen in the presence of food. In the absence of food, MIP treatment could not induce the addition of any new posterior segments, and postlarvae remained at the 3-segmented stage (Additional file 19E). Additionally, MIP-treated larvae, both fed and unfed, exhibited altered pigmentation of the gut and the body ( Figure 5H, Additional file 19F, G).

Discussion
Our results established the MIP neuropeptide as a regulator of postlarval feeding and gut activity in Platynereis. At 6 dpf, MIP is expressed in both the pharynx and the hindgut in neurons with a sensory morphology with dendrites projecting to the lumen. The MIP-expressing neurons of the Platynereis gut possess several hallmark features of mammalian enteroendocrine cells, including a scattered distribution, dendrites extending towards the gut lumen and long branching axons in the gut epithelial layer underlying the gut musculature [56]. These results are consistent with a model where MIP cells receive sensory signals from inside the mouth and the gut and respond by releasing MIP in a neurosecretory manner in the vicinity of the pharynx and hindgut muscles. However, given the use of bath-application and whole-body morpholino knock-down, we could not analyze the function of individual MIP-expressing cells. In principle, MIP-expressing neurons in other parts of the body may also affect gut activity by hormonal action.
Contrary to its name, MIP plays a myostimulatory role in the Platynereis digestive system. This may be the result of a direct effect whereby MIP directly acts on the digestive system musculature to increase the rate of pharynx extensions and peristaltic movements. Alternatively, the myostimulatory action of MIP may be caused indirectly through the regulation by MIP of other neurons in the gut, for example, in an as yet unidentified central pattern generator circuit responsible for regular gut contractions, as seen in crustaceans [57]. Knowledge of the spatial expression pattern of the MIP receptor in 6 dpf and older Platynereis larvae could help to resolve this.
Our results show that the increase in pharynx extensions in Platynereis postlarvae has a direct effect on the amount of food ingested. Increased gut peristalsis could promote the passage of food within the gut or the mixing of food with digestive enzymes, speeding up digestion. The fact that the highest concentration of MIP treatment, 50 μM, did not increase the amount of food ingested compared to control postlarvae is likely a result of the simultaneous reduction in locomotor activity caused by MIP treatment. At the highest concentrations of MIP (20 -50 μM), increased gut peristalsis and pharynx extension activity may be offset by a decrease in locomotion, resulting in treated individuals encountering fewer algal cells.
The sustained expression of MIP in the gut and the long-term effects of MIP on juvenile growth indicate that MIP also has an important physiological role later in the life cycle. We interpret the enhancement of juvenile growth in long-term MIP treatment experiments to be a consequence of a sustained increase in feeding caused by MIP. Given its effect on both settlement and growth, MIP treatment may be a useful means of enhancing both larval settlement and juvenile growth in polychaete aquaculture [58].
Interestingly, MIP has a myostimulatory role in cnidarians and in the Platynereis digestive system, but a myoinhibitory role in the arthropod digestive system [36,41,44]. This could mean that either MIP was independently recruited to regulate gut activity in different phyletic lineages, or that the sign of the regulation switched during evolution. In the latter case, MIP would represent a conserved bilaterian gut peptide influencing feeding. Further comparative morphological and molecular studies of MIP cells and signaling pathways in a broader range of taxa will be needed to resolve this.
MIP regulates both settlement behavior and feeding, two aspects of the pelagic-to-benthic transition of the non-feeding Platynereis larvae. What could be the reason for the redeployment of the same peptidergic signal at different times during development and in different contexts? One possibility is that the anterior MIPexpressing sensory-neurosceretory cells of the larva and the MIP cells in the gut of the postlarva sense the same chemical cues released by potential food sources. Some marine larvae are induced to settle by their future juvenile food source [2]. Testing this hypothesis will require the identification of naturally occurring settlement cues and their corresponding receptors in Platynereis.
In Platynereis, juvenile feeding is an essential requirement for the completion of cephalic metamorphosis. In other polychaete species, where feeding often begins in the pelagic larval stage before settlement, feeding is also an essential component for settlement and metamorphosis. Starved larvae of Capitella sp., Polydora ligia, Hydroides elegans and Phragmatopoma lapidosa all lose or have decreased ability to complete settlement and metamorphosis [59][60][61][62]. Exploration of the roles of MIP in polychaete species with feeding larvae would increase our understanding of the links between MIP signaling, larval settlement and feeding.

Conclusions
We have described a role for MIP in Platynereis postlarval feeding and established methods for studying the neuroendocrine regulation of feeding, providing the basis for future studies in this area. The amenability of Platynereis larvae to peptide treatments by soaking, their transparent body wall, and a neuropeptide complement that overlaps with that of both vertebrates and arthropods, make Platynereis an ideal model with which to study the neuroendocrine regulation of feeding in an evolutionary context.

Platynereis culture
Platynereis larvae were obtained from an in-house culture as previously described [15]. After fertilization of eggs, developing embryos and larvae were kept in an incubator at a constant temperature of 18°C with a regular light-dark cycle.
Genes were named according to their common conserved domains, reciprocal BLAST to the Homo sapiens peptidome, and neighbor-joining and maximum likelihood phylogenetic analyses (Additional files 7,8). Genes were analyzed for the presence of a signal peptide assigned using the SignalP 4.1 server (http://www.cbs.dtu.dk/services/ SignalP/) and conserved domains assigned by searches in the Pfam database with an e-value cutoff 1e-06 (http:// pfam.xfam.org/search). To find additional sequences for use in phylogenetic analyses, the candidate gene sequences were used as queries in BLAST searches against the NCBI nr and Swissprot databases, taking the top 50 hits from each BLAST search. To diversify the range of phyla represented, BLAST searches were also performed with different restrictions, including 'non-mammal' , 'non-Drosophila' and 'Lophotrochozoans'. Sequence redundancy was reduced to 90% identity using CD-HIT [63]. Genes were aligned to candidate orthologues from other taxa with MUSCLE (http://www.ebi.ac.uk/Tools/msa/muscle/). Conserved regions of sequence alignment were select for phylogenetic analysis using Gblocks with minimal stringency settings [64]. Phylogenetic trees were constructed from trimmed sequence alignments using the neighbor-joining methods with 1000 bootstrap replicates in CLC Genomics Workbench 5.5.1 (CLC Bio, Qiagen), with a Gap Open Cost of 10 and a Gap Extension Cost of 1. Maximum likelihood trees with 100 bootstrap replicates were constructed using PhyML 3.0 using an LG substitution model and SPR and NNI tree searching methods [65]. Trees were inspected and taxa with long branches were removed to avoid long-branch attraction bias. The phylogenetic analyses were then re-run with the remaining taxa. We then went on to examine the expression of the digestive system marker genes in 6 dpf, 15 dpf and 1 mpf Platynereis by RNA in situ hybridization methods as described below.

RNA In situ hybridization
Different developmental stages of Platynereis were collected for fixation for use in wholemount RNA in situ hybridization and immunostaining techniques. Individuals 6 days and older were relaxed using 1 M MgCl 2 [66] prior to fixation. Postlarvae and juveniles that had begun feeding were starved for a few days prior to fixation to avoid the presence of autofluorescent algae cells in the gut, which interfere with fluorescent signals from immunostaining. All animals were fixed in 4% paraformaldehyde (PFA) in 0.1 M MOPS (pH 7.5), 2 mM MgSO 4 , 1 mM EGTA, 0.5 M NaCl for 1 h at room temperature. Fixed larvae were dehydrated through a MeOH series and stored in 100% MeOH at -20°C.
DIG-labelled antisense RNA probes for the Platynereis MIP precursor (JX513877), MIP receptor (JX513876), alpha-amylase, subtilisin-1, subtilisin-2, legumain-protease precursor, and enteropeptidase were synthesized from purified PCR products of clones sourced from a Platynereis cDNA library [51]. RNA in situ hybridization using nitroblue tetrazolium (NBT)/5-bromo-4-chloro-3-indolyl phosphate (BCIP) staining combined with mouse antiacetylated-tubulin staining to highlight cilia and nervous system, followed by imaging with a Zeiss LSM 780 NLO confocal system and Zeiss ZEN2011 Grey software on an AxioObserver inverted microscope, was performed as previously described [50], with the following modification: fluorescence (instead of reflection) from the RNA in situ hybridization signal was detected using excitation at 633 nm in combination with a Long Pass 757 filter. Animals were imaged with a 40X oil objective.

Image registration of RNA in situ hybridization patterns
We projected average MIP expression pattern of four 6 dpf individuals onto a common 6 dpf whole-body nuclear reference template generated from DAPI signal of 40 individuals as described previously for 72 hpf larvae [50]. Acetylated tubulin and expression patterns of digestive system marker genes of select individuals were also projected onto the reference template. Snapshots and video of the projected genes were generated in Blender 2.7.1 (http://www.blender.org/).
For phalloidin stainings, we used freshly-PFA-fixed larvae dehydrated in 100% acetone for 5 minutes. Staining with rhodamine phalloidin (Molecular Probes) 1:100 in combination with the rabbit anti-MIP antibody was performed with the standard protocol adapted from [67]. After the staining procedure, samples were transferred to the mounting medium 87% glycerol containing 2.5 mg/ mL of anti-photobleaching reagent DABCO (Sigma, St. Louis, MO, USA), as phalloidin-conjugated rhodamine destabilizes in TDE [68].
Confocal images were processed with Imaris 6.4 (Bitplane Inc., Saint Paul, USA) software. Raw image stacks and the Platynereis 6 dpf nuclear stained reference are available at the Dryad data repository.

Calcium imaging
Fertilized eggs were injected with 500 ng/μl capped and polyA-tailed GCaMP6 [69] RNA generated from a vector (pUC57-T7-RPP2-GCaMP6) containing the GCaMP6 ORF fused to a 169 base pair 5′ UTR from the Platynereis 60S acidic ribosomal protein P2, as in [70]. Injection protocol is described in more detail in the 'Morpholino Knockdown of Platynereis MIP' section. Larvae were imaged with a 488 nm laser and transmission imaging with DIC optics on a Zeiss LSM 780 NLO confocal system on an AxioObserver inverted microscope (Additional file 5), or using a Zeiss AxioZoom V16 microscope with Hamamatsu Orca-Flash 4.0 digital camera (Additional file 13).

RNA-Seq
RNA-Seq analysis of digestive enzyme and MIP precursor gene expression was performed on an existing dataset of 13 different stages spanning the Platynereis life cycle, from egg to mature adults. Methods used were described in [51].

Documentation of normal feeding in Platynereis
To document variation in the commencement of feeding in Platynereis larvae from our laboratory culture, larvae were kept in Nunclon 6-well tissue culture dishes, with 10 ml sterile filtered seawater (FSW) per well. Each well contained 30 larvae. Larvae from 6 different batches, with different parents, were used in our analysis. Larvae were fed 5 μl Tetraselmis marina algae culture at 5 dpf. Larvae were then tested for feeding by checking for the presence of fluoresent Tetraselmis algae in the gut using a Zeiss Axioimager Z1 microscope with an AF488 fluorescent filter and a 20X objective. Larvae were checked for signs of feeding at 5.5, 6, 7, 8, 10, 12 and 14 dpf. After ingestion, algal cells can remain in the gut for up to 48 h before digestion causes a loss of fluorescence. Although larvae with a full gut can also be identified with normal light microscopy due to the transparent body wall, fluorescent microscopy enables the detection of even a single alga cell in the gut, due to the strong chlorophyll fluorescence of the Tetraselmis cells.

Morpholino knockdown of Platynereis MIP
Two translation blocking morpholinos (MOs) and two corresponding 5 base pair mismatch control morpholinos were designed to target the Platynereis-MIP-precursor (GeneTools, LLC): Nucleotides altered in mismatch control morpholinos are in italics. Information on the position of the morpholinos in relation to the MIP start codon can be found in Additional file 11.
MOs were diluted in water with 12 μg/μl fluorescein dextran (Mr 10,000, Invitrogen) as a fluorescent tracer. 0.6 mM MOs were injected with an injection pressure of 600 hPa for 0.1 s and a compensation pressure of 35 hPa using Eppendorf Femtotip II needles with a Femtojet microinjector (Eppendorf ) on a Zeiss Axiovert 40 CL inverted microscope equipped with a Luigs and Neumann micromanipulator. The temperature of developing zygotes was maintained at 16°C throughout injection using a Luigs and Neumann Badcontroller V cooling system and a Roth Cyclo 2 water pump.
For microinjection, fertilized Platynereis eggs developing at 16°C were rinsed 1 h after fertilization with sterile 0.2 μm filtered seawater (FSW) in a 100 μM sieve to remove the egg jelly, followed by a treatment with 70 μg/ml proteinase K for 1 min to soften the vitellin envelope. Following injection, embryos were raised in Nunclon 6-well plates in 10 ml FSW and their development was monitored daily.
Larvae were fed 5 μl Tetraselmis marina algal culture at 6 dpf. Feeding in 7 -14 dpf injected larvae was assessed by checking for the presence of fluoresent Tetraselmis marina algae in the gut using a Zeiss Axioimager Z1 microscope with an AF488 fluorescent filter and a 20X objective. Larvae were checked for signs of feeding as described above every 24 h from 7 dpf on. We scored a minimum of 62 larvae (maximum 424 larvae) from a minimum of 3 separate microinjection sessions (with 3 different batches of larvae) for each translation-blocking and control morpholino. Photomicrographs of morpholino-injected larvae were also taken and larval body length was measured from these pictures using Image J 64 software. Some morpholino-injected larvae were also fixed at 6 days for immunostaining with the anti-MIP antibody (as described above) in order to assess morpholino specificity and effectiveness.
Effect of synthetic MIP on Platynereis feeding behaviour Peptide functions can be investigated in Platynereis larvae by bath application of synthetic neuropeptides [10,11]. To test whether synthetic MIP treatment increased developmental speed, leading to early initiation of feeding in Platynereis larvae, experiments were performed in Nunclon 6-well plates, with 10 ml FSW per well. Each control and peptide treatment was replicated across three wells, with 30 larvae per well. Larvae were treated with 5 μM MIP7 or controls at 24 hpf, 60 hpf, 4 dpf or 5 dpf, then fed at 4 or 5 dpf (depending on the age at which MIP treatment occurred) with 5 μl Tetraselmis marina algal culture. Larvae were fed at an earlier age due to the possibility of MIP treatment causing an earlier initiation of feeding. Larvae were checked for feeding by monitoring algal cell fluorescence in the gut as described above. Larvae were monitored from 5 or 5.5 dpf (depending on age at which larvae were first fed) until 7 or 8 dpf. A control non-functional MIP peptide (MIPW2A, AANKNSMRVAamide), in which the two conserved tryptophan sites were replaced with alanines (this prevents MIP from activating its receptor, see [11]) was also tested. A further control of larvae treated with DMSO alone was also included, as MIP peptides require DMSO to be dissolved in solution.
To test the effects of synthetic MIP peptide treatment on the digestive system of Platynereis larvae, we recorded videos of groups of 60 larvae at 6.5 dpf in a square glass cuvette 1.5 cm x 1.5 cm x 0.3 cm in 500 μl of FSW using a Zeiss AxioZoom .V16 microscope with Hamamatsu Orca-Flash 4.0 digital camera. For each treatment and control, 3 biological replicates (larval batches with different parentage, fertilized on different days) were carried out. We tested three concentrations of synthetic MIP: 5, 20 and 50 μM, plus 50 μM control non-functional MIP peptide MIPW2A and a 0.1 % DMSO control. A 2.5 min video at 10 frames per second was recorded 10 min after peptide or DMSO addition. Videos were analyzed manually in Fiji (Image J 1.48s, Wayne Rasband, http://imagej.nih.gov/ij). For each video, 20 larvae that remained within the frame of the video for the entire 2.5 min were scored for gut peristalsis and pharynx extension activity. Distance traveled and speed of the larvae was also measured using the MTrack2 plugin [71]. Significant differences in gut peristalsis, pharynx extension activity and locomotion in MIP-treated versus control larvae were tested in an unpaired t test.
To test the effects of synthetic MIP treatment on short-term ingestion of algal cells in Platynereis larvae, experiments were performed in Nunclon 24-well plates, with 2 ml FSW per well. Each control and peptide treatment was replicated across three wells, with 20 larvae per well. 7 dpf postlarvae were treated with 5 μM MIPW2A control peptide, 5 μM MIP, 20 μM MIP or 50 μM MIP for 10 min. Following this, 20 μl Tetraselmis marina algal culture was added to each well and larvae were left to feed for 30 min. All larvae were then immediately fixed in 0.5 mL 4% paraformaldehyde in 1X PBS with 0.01% Tween (PTw) for 1 hour. Following 4 washes in 1 ml PTw, larvae were mounted on glass slides and the number of algal cells in the digestive system of each larva was counted using a Zeiss Axioimager Z1 microscope with an AF488 fluorescent filter and a 20X objective. Significant differences in MIP-treated versus control larvae were tested in an unpaired t test.

Scanning electron microscopy (SEM)
Platynereis larvae and juveniles of different developmental stages were fixed with 3% glutaraldehyde in 0.1 M phosphate buffer pH 7.2, rinsed in phosphate buffer, further fixed with 1 % osmium tetroxide in water and dehydrated in an ascending EtOH series over several days. Critical point drying with carbon dioxide was performed in a Polaron E 3000. The samples were coated with goldpalladium in a Balzers MED 010. Images were taken on a Hitachi S-800 Scanning electron microscope.

Calculation of optimal culture density
The assessment of growth in larvae cultured individually was performed in a Nunclon 24-well tissue culture dish with 1 larva per well in 2 ml FSW. Larvae were fed from 6 dpf with 3 μl Tetraselmis marina algae culture. Larvae were scored under a dissection microscope for number of segments and cephalic metamorphosis every 48 h from 14 dpf to 34 dpf.
Documentation of growth in larvae cultured at different densities was carried out in Nunclon 6-well plates with 10 mL FSW/well and 30, 50 or 100 larvae per well. Three replicate wells were included for each culture density. Larvae were fed with surplus Tetraselmis marina algae throughout the experiment. Larvae were scored for segment number and cephalic metamorphosis every 4 days from 16 to 32 dpf.

Long term treatment of Platynereis with synthetic MIP
To test the effect of synthetic MIP treatment on growth in Platynereis larvae, we again carried out experiments in Nunclon 6-well plates as described above, with 30 larvae per well and 3 replicate wells per treatment and control. 5 μM synthetic peptides were added at 4 dpf. Different versions of mature MIP peptide tested were: MIP1 -AWNKNNIAWamide, MIP6 -AWGDNNMRV Wamide, MIP7 -AWNKNSMRVWamide, MIP8 -AW KGQSARVWamide, and MIP9 -GWNGNSMRVWamide. Larvae were also fed at 4 dpf with 5 μl Tetraselmis sp. algal culture. At 25 dpf (21 days after peptide addition), errant juveniles were scored for number of segments and cephalic metamorphosis (as above). Juveniles were also photodocumented using a Zeiss Axioimager Z1 microscope with differential interference contrast (DIC) and size of control and treated larvae (end of head to end of pygidium, excluding cirri) was measured in Fiji (Image J 1.48s, Wayne Rasband, http:// imagej.nih.gov/ij).
Additional file 2: Movie of MIP expression in foregut sensory cell of 6 dpf Platynereis. Whole-mount RNA in situ hybridization (WMISH) for the Platynereis MIP precursor (red) counterstained for acetylated tubulin (white) and DAPI nuclear stain (blue). Ventral view of foregut. Yellow arrowhead marks sensory dendrite of MIP-expressing cell. Green arrows indicate salivary glands.
Additional file 3: Movie of MIP expression in mid-and hindgut sensory cells of 6 dpf Platynereis. Whole-mount RNA in situ hybridization (WMISH) for the Platynereis MIP precursor (red) counterstained for acetylated tubulin (white) and DAPI nuclear stain (blue). Ventral view of mid-and hindgut. Yellow arrowheads mark sensory dendrites of MIP-expressing cells.