Titolo della tesi: Evolutionary biology of Raphitoma Bellardi, 1847 (Neogastropoda, Conoidea)
The biodiversity of Earth is not equally spread across the globe: marine areas host ~16% of the total known species (but represent also the majority of unexplored areas), and more than 98% of them are benthic organisms. The reason behind the abundance of species richness in the benthos is the high levels of ecosystem complexity that is found on the seafloor. The large number of combinations of environmental factors (heat, light, resources, salinity, hydrodynamics, grain size) acted as a powerful evolutionary driver, supporting speciation processes. The adaptations of benthic organisms can be related, among others, to their trophic and reproductive biology and ecology: the prey can be either elusive or protected by robust armours, the predators possess advanced detection senses and specialised offensive techniques; reproduction at sea can be sexual or asexual, involve brooding or spawning, species can be hermaphroditic, gonochoric, protandrous or protogynous, and larval development of benthic organisms can include a pelagic phase, harnessing the power of horizontal currents to achieve high dispersal.
Some taxa were particularly successful in adapting to the benthic conditions. Gastropoda is one of the most diverse marine taxon, and almost half of gastropod diversity is represented by the order Neogastropoda. Neogastropods are coiled, shelled, carnivorous snails, mostly predators. Predatory adaptations in this group are evident in the foregut anatomy, diet, and feeding behaviour. The most species rich taxon in Neogastropoda is the superfamily Conoidea, well known for the cone snails of family Conidae, which produce powerful venoms for securing their prey and escape their predators. Nevertheless, the most diverse family among conoideans is probably Raphitomidae, whose ecology and biology has not been completely assessed.
The focus of my PhD project is the evolutionary biology and ecology of gastropod species of the genus Raphitoma Bellardi, 1847. These raphitomids are well studied in the Mediterranean, but their trophic and reproductive ecology is still largely unknown. Recent assessments with integrative taxonomy revealed complex morphological characters, and therefore a probably vexed taxonomy, and an evolutionary history of intermittent loss of planktotrophy. Furthermore, two species, namely Raphitoma philberti (Michaud, 1829) and Raphitoma laviae (R. A. Philippi, 1844) were proposed as candidate poecilogonic species, a rare state in which dimorphic larval development occurs. During my PhD I used an integrative approach, combining traditional techniques with molecular data, including Next Sequencing Generation (NGS) data. NGS technology implements the parallel sequencing of millions of short DNA fragments, allowing to perform studies on a genomic scale. I carried out three studies which are reported as the three main chapters of this thesis: the first explores the trophic ecology of Raphitoma purpurea (Montagu, 1803) by analysing the transcriptome of the salivary glands and venom duct, and has been published in Toxins; the second describes the sexual dimorphism and geographic polymorphism retrieved in Raphitoma purpurea, and is currently under revision in the Journal of Molluscan Studies; the third addresses the populations genetic of the poecilogonic species Raphitoma philberti and Raphitoma laviae, plus a closely related undescribed species, and is currently being drafted for a submission in Molecular Ecology.
Chapter I - Potential ancestral conoidean toxins in the venom cocktail of the carnivorous snail Raphitoma purpurea (Montagu, 1803) (Neogastropoda: Raphitomidae)
In the animal kingdom, a wide range of different venoms has evolved for both defensive and offensive functions in predation and intraspecific competition. Marine gastropods include many venomous taxa, the most studied being the cone snails of the family Conidae Fleming, 1822, which employ powerful neurotoxins to incapacitate their prey or elude their predators. The conid venom cocktail is characterised by a mixture of short neuroactive peptides named conotoxins, which are highly diversified and able to affect a wide range of vertebrate and invertebrate taxa. Our knowledge of neogastropod venoms is mostly limited to cone snails, and taxonomic coverage of predatory venomous gastropods outside of Conidae is still very low. To gain a better understanding of the evolutionary history of the venom, a broader taxonomic scope is needed. Given that the evolution of a venom system might have constituted a key evolutionary innovation leading to adaptive radiation in Conoidea, one of the best targets for such studies is certainly the most diverse of Conoidea, Raphitomidae Bellardi, 1875. The trophic ecology of raphitomids has not yet been determined, but preliminary evidences point at a vermivorous diet.
To address the trophic ecology of raphitomids, I performed the transcriptome analysis of the salivary glands and venom ducts of Raphitoma purpurea using NGS data. The secretome was characterised and main toxic compounds were identified on the basis of tissue-specific expression, the scoring system of DeTox, a comprehensive pipeline for toxin discovery, and a manually curated annotation. This large-scale approach aimed to identify both peptides with toxic properties (with a focus on conotoxin-like peptides) and proteins responsible for toxin maturation and regulation. Finally, I provided evidence of predatory behaviour in Raphitoma by identification of prey taxa from the gut contents of Raphitoma purpurea using the mitochondrial marker 16S rDNA.
I performed the dissection of the salivary glands and the venom duct of fifteen specimens of Raphitoma purpurea from Ploubazlanec (Brittany, France), which were processed for RNA sequencing. Two of the venom ducts were dissected into their most proximal and distal sections, removing the middle portion to be processed separately. The entire digestive tract was also dissected and preserved in ethanol for the gut contents analyses. After assessment of the concentration and quality of extracted RNA, three carcases, three salivary glands, three whole venom ducts, two distal fragments of venom duct, and two proximal fragments of venom ducts were sequenced. Raw reads were trimmed and assembled using Trinity and a differential expression analysis among different tissues was performed using DESeq2.
A reference dataset was built, including all toxins from UniprotKB and Conoserver, integrated with the published toxins of 33 gastropod species. Candidate venom components were identified using DeTox, a software for toxin detection that considers four features: (i) peptide sequence similarity with a reference database, (ii) detection of functional domains, (iii) detection of a signal sequence and no transmembrane domain — features typical of secretion components —and (iv) presence of a cysteine framework, which is often well conserved in conotoxins. The most relevant putative venom components (PVC) based on the DeTox score and expression levels were manually annotated. Putative raphitoxin families (PRF) were retrieved using ConoDictor and by scanning for transcript families with high expression and conserved cysteine frameworks. PVCs with high similarity scores with known toxins were selected to perform Maximum Likelihood analyses to infer their evolutionary history.
DeSeq2 identified 9990 differentially expressed transcripts, of which 2614 were overexpressed in the salivary glands and 7208 in the venom duct samples. Only 34 transcripts were differentially expressed between the two sections of the venom duct, not including any PVC but mostly related to cellular transport and muscle contraction. A total of 74,520 transcripts were selected by the DeTox pipeline as candidate venom components, resulting in 104 PVC which were processed for functional annotation. After annotation, they were classified as venom components with function related to post-translational modification (n = 12), regulation (n = 23), or biosynthesis (n = 69). Neuropeptides represented the highest number of PVC (n = 49), and the vast majority were predominantly expressed in the venom duct. Through the manual annotation of the venom of Raphitoma purpurea I performed a broad characterisation of the main venom components and reported similar compounds retrieved in other venomous taxa. ConoDictor identified 9530 conotoxin-like peptides in the DeTox output, which after filtering constituted 145 transcripts ascribed to 20 putative raphitoxin families (PRF), including conotoxins (n = 105), conodipines (n = 6), conoCAP (n = 1), conkunitzins (n = 9), contryphan (n = 1), and conatokins (n = 6), while 17 transcripts remained unclassified. From the gut content, two polychaete 16s rDNA fragments were successfully amplified, matching genera of the family Terebellidae and Spionidae.
The characterisation of the venom cocktail of Raphitoma purpurea was negatively impacted by the lack of data on the venom of closely related taxa, as the reference venoms were all from other families. Most of putative neuropeptides retrieved in the PVC and PRF did not have a positive match, and could be either false positives, or toxins new to science. Nevertheless, a fair number of venom compounds with high affinity with known toxins was retrieved, especially in the salivary glands, with function related to predatory behaviour. Two isomers of disulfide isomerase and elevenin were retrieved among the putative venom components. These peptides’ families expanded to perform the specific function of toxin folding and regulation, which was fundamental for the evolution of cone snails’ venom, and their retrieval in Raphitoma purpurea shows that this process may have originated before the radiation of Conidae. Furthermore, the neuropeptides retrieved in the salivary glands (turritoxins, Pmag02 conotoxins, and conkunitzins) in Raphitoma could also represent ancestral adaptations of the conoidean venom. If this hypothesis is correct, most conoideans families could share these traits. Alternatively, the presence of these venom components in Raphitoma would represent instances of convergent evolution. Overall, these results might indicate that ancestral venom elements were originally produced in the salivary glands, possibly before the development of the venom duct. Subsequently, these gene families expanded following the evolution of the venom duct, where a larger number of toxins could be secreted at a higher concentration. If this hypothesis is correct, evidence of venom evolution in Conoidea is more likely to be found in toxin maturation and regulation factors and venom components of the salivary glands rather than in venom duct conotoxin-like peptides, which probably had their evolutionary burst during the radiation of the family Conidae.
Chapter II - Sexual dimorphism in Raphitoma purpurea (Neogastropoda: Raphitomidae)
Dimorphism is a condition where two separate phenotypes occur within the same population or species. Sexual dimorphism concerns differences between males and females in morphology, anatomy, metabolism, and behaviour. In Gastropoda, contrary to most cases among animals, the females are dominant in size, an adaptation presumably related to the reproductive necessity of sustaining and accumulating large numbers of eggs. Although gastropods are the second most diverse taxon on Earth, sexual dimorphism has been addressed only in a limited number of species. Differing phenotypes can also be related to the ecological context, determining the presence of multiple phenotypes in a species, a condition known as polymorphism. Polymorphisms due to environmental factors in gastropods have been linked to the evolutionary mechanisms of adaptive radiation and divergent natural selection.
In this study I address the reproductive biology and ecology of Raphitoma purpurea (Montagu, 1803) and the morphological variation which occurs in this taxon. The collection of male-female pairs during mating in the Atlantic allowed me to perform a broad morphological and anatomical study. Specimens of the nominal species Raphitoma andrehoaraui Pelorce & Horst 2020 from the Mediterranean were examined in an integrative approach to test synonymy with Raphitoma purpurea. Furthermore, I gathered a large amount of DNA-barcoding sequences of Mediterranean raphitomids, and performed species delimitation and phylogenetic analysis to advance the taxonomy of the genus.
The morphometric analyses were based on the following characters: shell height, shell width, number of whorls, axial and spiral sculpture, last whorl height, outer lip width, outer lip colour, and number of internal indents. I performed the anatomical dissection of two male-female pairs and described the internal anatomy. Molecular barcoding included 252 newly produced sequences, which were combined with 134 sequences from Genbank. Species delimitation was performed with an automatic partitioning algorithm based on barcoding distances and phylogenetic analyses were run with maximum likelihood and Bayesian approaches.
Copulation of Raphitoma purpurea was observed in situ and in aquarium. The male, visibly smaller than the female, initiated reproduction, pursuing the female and positioning over or beside it on the right side. The female was standing on the foot leaving an opening in the posterior canal, through which the male introduced the penis into the mantle cavity. Average adult shell height was of 11 mm in males and 21 mm in females, and average shell width was of 5 mm in males and 8.5 mm in females. Measurements taken at the fifth whorl were also significantly different between the sexes. Mediterranean specimens of Raphitoma andrehoaraui were not sexed, and were overall shorter (height ~ 9 mm), of lighter colour, had fewer sculptured axes compared to the Atlantic Raphitoma purpurea (12–16 vs 17–22), and did not display the white thickened labrum typical of Raphitoma purpurea. One female specimen from Alboran had a remarkable morphology with intermediate traits of the two nominal taxa. The general anatomy of males and females of Raphitoma purpurea was described and graphically displayed, with focus on reproductive and feeding apparatuses. The molecular analyses retrieved 14 species, three more from the last molecular assessment. Furthermore, specimens of Raphitoma andrehoaraui were not distinguishable from specimens of Raphitoma purpurea. Additionally, 25% of samples were morphologically identified with further twelve nominal taxa, which did not make monophyletic clades.
The observations of mating in Raphitoma purpurea are coherent with the reproductive ecology of many marine caenogastropods. The penis had a terminal papilla shaped like a spatula, which could play a role in postcopulatory sexual competition, for example removing sperms previously introduced by another male. No bursa copulatrix or spermatheca were found: storing could be accomplished in proximity of the large vestibule or in the dorsal fold of the capsule gland. Although no clearly defined albumen gland was found, the last lobe of the capsule gland looked like a separate compartment, with a presumably different function. No vestigial male reproductive organs were detected during dissections of the females, suggesting Raphitoma purpurea is a gonochoric species. The egg capsules were consistent with previous observation in this species, and other raphitomids. A secondary sexual dimorphism — i.e. not strictly related to reproductive organs — was also evident in the Raphitoma purpurea population of Ploubazlanec (Brittany, France). The smallest adult male was less the size of the largest female. Significant differences in height, width and sculpture were also confirmed when comparing the same whorl. This result provides further evidence of this species’ gonochorism and suggests that specimens may be sexed using shell characters, although further data is necessary. The extreme size ratio observed compares to few documented sexual dimorphisms among gastropods. This extreme size difference has also interesting ecological implications on trophic ecology. The foregut anatomy of Raphitoma purpurea was consistent with previous descriptions, but rostrum, proboscis, and venom duct were substantially shorter in males, hinting at a possible adaptive modification in the feeding apparatus of females to target larger preys.
The Mediterranean specimens morphologically ascribed to the recently described R. andrehoaraui were retrieved as conspecific with R. purpurea. The former nominal taxon was based on individuals from southern France, with colour pattern, number of axial ribs and spiral cords different from R. purpurea, especially the lack of secondary spiral cordlets in the latter. Interestingly, populations from the Alboran Sea may have an intermediate morphology: one female specimen in our dataset was consistent with the Atlantic morphotype in shell shape and height and white lip, but with overall paler colouration, fewer spiral cords and no secondary cordlets. In contrast, a male from the same area was more compatible with the morphology occurring in the Mediterranean. A denser geographical sampling is needed to confirm the polymorphism of this species, and to check for the presence of sexual dimorphism also in the Mediterranean. Determining which environmental factors affect shell development in Raphitoma would help to understand phenotypic plasticity and adaptations in Mediterranean vs Atlantic gastropods. Furthermore, retrieving a correlation between shell traits and environmental conditions could make raphitomids biological indicators of climate changes and pollution (Abdelhady, 2016). Similarly to Raphitoma andrehoaraui, several other nominal species were not retrieved as separate species, and could represent local ecomorphs.
CHAPTER III - Population genetics of poecilogonic species in the marine gastropod genus Raphitoma Bellardi, 1847 (Neogastropoda: Conoidea) using RAD-seq data
A species dispersal can be the source of many evolutionary advantages, such as the broadening of a species range and the maintaining of gene flow among populations. For marine benthic invertebrates, which have limited (if any) dispersal capabilities, larvae are crucial in performing this important function. Planktotrophic larvae can spend a long pelagic phase and travel far distances while feeding autonomously, whereas lecithotrophic larvae are usually produced in lower amounts and get their sustainment from the egg. Poecilogony is a state in which different larval developmental modes are present in a species or population Regrettably, only very few species are known to exhibit poecilogony, almost exclusively among polychaete annelids and sacoglossan gastropods. Shelled gastropods retain their embryonic/larval shell (protoconch) as adults, and the modality of the pelagic phase can be presumed directly from protoconch length and sculpture. Candidate cases of poecilogony were highlighted recently in the neogastropod family Raphitomidae, the most diverse taxon of Conoidea. Two species of genus Raphitoma — preliminarily identified as R. philberti (Michaud, 1829) and R. laviae (R. A. Philippi, 1844) — included specimens exhibiting both multispiral and paucispiral protoconch, indicating planktotrophic and lecithotrophic development, respectively.
I investigated the population genetics of Raphitoma philberti, Raphitoma laviae, and the recently documented species Raphioma sp. C using a reduced-representation genome-wide SNP dataset. This study aimed at confirming the taxonomic and phylogenetic status of these species, possibly confirming or denying the presence of poecilogony, and testing the hypothesis that the occurrence of two differing developmental strategies affects the population structure of poecilogonic species, compared to purely lecithotrophic or planktotrophic ones.
A total of 133 specimens from the Mediterranean of the three species were genetically assessed and larval development was inferred by protoconch observation. Among them, 80 samples were processed following the specific locus amplified fragment sequencing (SLAF-Seq), including three localities, corresponding to three a priori populations: Croatia, Greece, and Corsica. After sequencing, polymorphic loci were filtered allowing a minimum allele frequency of 0.05 and including biallelic alleles present in at least 60% of samples with a sequencing depth of at least 6x was retained. A species delimitation analysis was performed using all COI sequences available for the genus Raphitoma. Maximum likelihood phylogenetic analyses were carried out with the mitochondrial marker COI dataset and the genome-wide SNP dataset. A random SNP per locus was selected to infer the population structure for each species and to perform cluster analyses and compute F statistics. Finally, a Mantel test was performed to test Isolation by Distance (Wright, 1943) was tested using samples of R. laviae from Croatia, which included a sufficient amount of planktotrophic (n = 10) and lecithotrophic (n = 11) samples from four localities.
The final dataset for all species included 9970, and those with each species separate had 4189 to 5822 loci. The species delimitation suggested hypotheses with more than three species, but they were not supported by monophyly in the phylogenetic reconstruction. SLAF-Seq data confirmed the monophyly of the species Raphitoma philberti, Raphitoma sp. C, and Raphitoma laviae instead. The clade of Raphitoma philberti was split in two groups, one without geographic structure, and the other geographically structured and including both specimens with both planktotrophic and lecithotrophic development. Raphitoma sp. C comprised two clades without a geographic structure, all with planktotrophic development. Samples of Raphitoma laviae could be split in at least four clades including samples from: all localities, only Croatia, or only Corsica. Specimens with paucispiral protoconch from Croatia were separated from the ones with a multispiral protoconch from the same area, and displayed a fine geographical structure. The genetic structure did not correspond with the sampling localities in any of the species. Similarly, the clusters analyses and shared ancestry matrix did not retrieve groups corresponding to the geographic distribution. Fst had high values (>0.05) between Croatia and Corsica for Raphitoma laviae and between Croatia and Greece in Raphitoma philberti and Raphitoma sp. C. High Fst was also retrieved between the lecithotrophic and planktotrophic Raphitoma laviae from Croatia, where Isolation by Distance was retrieved only for lecithotrophic specimens with 0.8 correlation.
The phylogenetic and population structure analyses including genome wide data consistently retrieved specimens with planktotrophic and lecithotrophic development within the species Raphitoma philberti and Raphitoma laviae, thus confirming poecilogony in both species. A complex pattern was retrieved including widespread haplotypes, retrieved in all the areas of study, and some more isolated, with a degree of geographic structure. In Raphitoma laviae, the ancestral clade following the phylogenetic reconstructions included specimens from distant localities in the central Mediterranean, suggesting high connectivity for this species within the central Mediterranean. Nevertheless, additional clades with a limited distribution were retrieved, an element associated with low levels of connectivity between areas. Few samples of Raphitoma philberti and Raphitoma sp. C were available in the SLAF-Seq dataset, but preliminary results based on mitochondrial markers shows that a similar pattern could be true for this species, with coexistence of widely distributed vs localised genetic clusters.
In poecilogonic species, the genetic structure should be comparable to that of an high-dispersal species, since the planktotrophic component would determine high gene flow and hamper local adaptation and segregation of lecithotrophic specimens. But the results of this study showed that both Raphitoma laviae and Raphitoma philberti display a certain degree of structuring, mostly concerning the area of Croatia and in one case involving specimens which underwent a lecithotrophic larval development. In order to explain the pattern retrieved, I hypothesise an external effect on dispersal in combination with poecilogony. If a region within the poecilogonic species range became completely or partially isolated, then the local lecithotrophic population would be separated from the panmictic metapopulation, and a population genetic structure would emerge within the region. In fact, the internal genetic structure retrieved in Raphitoma laviae does not delimit geographic populations, but a grouping with a large distribution range (Corsica, Croatia, and Greece) and several with a limited one, including one with exclusively lecithotrophic specimens from Croatia, with a fine geographic structure and verified IBD. My hypothesis excludes an epigenetic control of the offspring development and also assumes that loss of planktotrophy is an irreversible process, but additional studies are needed to confirm the unknown biology of poecilogonic gastropods. Future studies should aim at understanding the biological mechanisms that guide the development switch of poecilogonic species, to define the effect of environmental variables and the possibility of reversal after loss of planktotrophy.
This collection of studies approached the evolutionary biology and ecology of species of the genus Raphitoma Bellardi, 1847 using an integrative biology approach, benefitting from traditional and Next Generation molecular data, successfully filling part of the knowledge gap that affects this and other neglected marine taxa. The transcriptome sequencing of Raphitoma purpurea (Montagu, 1803) and the molecular barcoding of gut content confirmed a vermivorous diet for the species and extended to this taxon the ability to produce venom components. Furthermore, the duplication of chaperone families which favoured the diversification of conid venom cocktails is also detected in raphitomids salivary glands, and may be a synapomorphy shared with other conoidean taxa. Anatomical and morphometric analysis of Raphitoma purpurea confirmed gonochorism and revealed a sexual dimorphism in a population in the Atlantic, where the size of females almost doubles that of adult males. Furthermore, the species range was updated after molecular confirmation that the Mediterranean nominal taxon Raphitoma andrehoaraui Pelorce & Horst, 2020 is a synonym of Raphitoma purpurea. The geographic polymorphisms retrieved suggests caution in phenotypes interpretation for the taxonomy of this genus in the Atlantic and Mediterranean. Inference of larval development based on protoconch length and sculpture integrated with reduced representation genomic data confirmed the poecilogony of Raphitoma philberti and Raphitoma laviae. The population structure retrieved was ambivalent, sharing features of both high- and low-connectivity.
By exploring the biology and ecology of raphitomids, which could represent one of the most diverse marine taxon, I have found multiple evidences of the same adaptations that may have boosted the radiation of Neogastropoda and, in general, of marine life. These outcomes from the genus Raphitoma may represent fundamental advancements in the study of the evolution of marine invertebrates, but also in the specific fields of venomics and population genetics. The venom cocktail of R. purpurea is the first characterised of the family, and will serve as a reference database for new venomic studies of the family Raphitomidae. The sexual dimorphism retrieved in Raphitoma purpurea can be extended to other species of the genus, and be implemented for sexing samples using shell characters. The study on the population genetics of Raphitoma philberti and Raphitoma laviae is the first assessment of poecilogony in species of Neogastropoda documented by NGS data. Finally, the molecular framework produced for Raphitoma will be indispensable to correctly reassess the taxonomy of the genus in an integrative approach. In order to fully understand the radiation of Neogastropoda and Conoidea, and the evolutionary processes that produced most of the biodiversity that we see today, similar studies should advance the understanding of biology and ecology of neglected taxa, focussing on additional raphitomids and eventually more families within Conoidea, to understand which adaptations were more important in promoting radiation for Gastropoda and marine species.