Can you find me? A new sponge-like nudibranch from the genus Jorunna Bergh, 1876 (Mollusca, Gastropoda, Discodorididae)

The nudibranch diversity of the western Indian Ocean is comparatively one of the least studied in the world. In this paper a sponge-like Discodoridae nudibranch Jorunna liviae sp. nov. is described. The description is based on integrative anatomy, including molecular analysis of two genes (the mitochondrial COI and the nuclear H3), dissections, electron microscopy (SEM) of buccal elements, micro tomography of the spicule’s arrangements and ecological observations. This study provides the first ever molecular data of Jorunna species from the western Indian Ocean, helping to fill the gap to further understand this apparent paraphyletic genus

Despite current research efforts to raise the biodiversity knowledge of nudibranchs from the western Indian Ocean (e.g.Manson-Parker 2015;Tibiriçá et al. 2017aTibiriçá et al. , b, 2018Tibiriçá et al. , 2020)), this region remains comparatively far less studied than other areas of the Indo-West Pacific.The high number of undescribed species often hampers comprehensive biogeographic studies.Thus, the discovery of new species is of primary importance to advance our global knowledge on how biodiversity is formed and how species diversity spreads across oceans.Moreover, the lack of molecular data from the western Indian Ocean (WIO) limits phylogenetic studies.Of all specimens sequenced from the genus Jorunna so far, none are from the WIO.The present study contributes to fill this gap by providing a description of a new Jorunna species from Mozambique, including molecular, morphological and ecological data.

Morphological study
Specimens were dissected by dorsal incision under a dissecting microscope Nikon SMZ18.Their reproductive system was separated, examined and drawn under a dissecting microscope Leica 80 with an attached camera lucida.Surrounding radula tissue was removed by immersing in 10% sodium hydroxide for about 8 hours or on a solution containing 180 mL of the tissue lysis buffer ATL with 20 mL of proteinase K-solution incubated in 56 °C for 48h (Holznagel 1998).Labial cuticle and radula were then mounted for electron microscopy (SEM) examination.Imagines were obtained under a FEI Nano-SEM 450 scanning microscope at the Servicios Centrales de la Ciencia y Tecnlogia de la UCA (MEB), Universidad de Cadiz.Microcomputed tomography (µCT) was carried out to inspect the spicules arrangement by the Servicio de Técnicas No Destructivas del Museo Nacional de Ciencias Naturales de Madrid (MNCN-CSIC).This technique uses x-ray attenuation of biological tissues in three different planes allowing for 2D and 3D image reconstructions (Ziegler et al. 2018).Images were reconstructed using VGSTUDIOMAX 2.2 and visualized in myVGL by Volume Graphics (https://www.volumegraphics.com).Spicules sizes were measured in the µCT images using the distance instrument tool.Measurements were taken from spicules that were clearly visible and from different parts of the body.

DNA extraction, amplification and sequence
DNA extraction and amplification were conducted by the peripheric services of the Instituto Universitario de Investigacion Marina (INMAR-UCA).Genomic DNA was extracted from a small sample of foot tissue using the Qiagen DNeasy Blood & Tissue extraction kit, following the manufacturer's instructions.One mitochondrial gene cytochrome c oxidase subunit (COI) and one nuclear gene histone H3 (h3) were amplified by polymerase chain reaction (PCR), using the universal primers LCO1490 and HCO2198 (Folmer et al. 1994) and H3AD-F and H3BD-R (Colgan et al. 2003), respectively.We tried to amplify the gene 16S using the 16S universal primers 16Sar-L and 16Sbr-H (Palumbi et al. 2002) but all attempts were unsuccessful.PCRs were performed in 25-μl reactions with 2 μl of DNA template.COI amplifications were performed with an initial denaturation for 3 min at 94 °C, followed by 40 cycles of 30 s at 94 °C, 30 s at 46 °C and 1 min at 72 °C with a final extension of 5 min at 72 °C.H3 amplifications were performed with an initial denaturation for 3 min at 95 °C, followed by 25 cycles of 45 s at 94 °C, 45 s at 50 °C (annealing temperature) and 2 min at 72 °C, with a final extension of 10 min at 72 °C.Once completed, successful PCR products were sent to Macrogen, Inc. (Madrid, ES) for purification and sequencing.
All sequences were revised and examined in Geneious v.10.2.4 (Kearse et al. 2012).Possible contamination was verified using the Basic Local Alignment Search tool (BLAST) web server (https://blast.ncbi.nlm.nih.gov/Blast.cgi,Altschul et al. 1990).New sequences were uploaded in Genbank, NCBI and ascension numbers are provided in Appendix 1. Outgroup sequences and other Jorunna spp.sequences were obtained from GenBank.The outgroup selection followed Neuhaus et al. (2021).Additionally, one species of each available genus of Discodorididae Bergh, 1891 from GenBank was included in the analysis with preference given for type species.When available, up to three sequences of each morpho-species of Jorunna from GenBank, NCBI were included in the phylogeny.Preference was given to specimens with COI and H3.Sequences were aligned in Geneious (https:// www.geneious.com)using Muscle and default settings.

Phylogenetic analysis
Maximum likelihood (ML) and Bayesian inference (BI) were used to infer evolutionary relationships.Analyses were conducted for individual genes as well as for the concatenated COI+16S.JModeltest was used to estimate the best fit-evolutionary model by applying the Akaike information criterion (AIC) for each gene.The model chosen was the GTR+I+G for COI and H3.Bayesian inference was performed via MrBayes v.3.2.6 (Ronquist and Huelsenbeck 2003) and run for 5,000,000 generations and four chains, with unlinked parameters, partitioned by genes and a burnin of 25%.Node support was assessed based on the posterior probability (PP) and considered strongly supported when PP ≥ 0.95 (Alfaro et al. 2003).Maximum likelihood analyses were performed in RAxML v8.2.4 implemented in the Cypres portal, applying 5,000 bootstrap (https://www.phylo.org, Miller et al. 2010).Support for nodes in the ML analysis was assessed with non-parametric bootstrapping (BP) using RAxML v.7.06 (Zhang et al. 2013).Maximum likelihood values of 70 or higher were considered statistically significant (Huelsenbeck and Rannala 2004).The trees obtained were visualized and collapsed (PP ≥ 0.5) in TreeGraph2 (http://treegraph.bioinfweb.info,Stöver and Müller 2010) and edited in Adobe Illustration 2021 v.25.2 (https://www.adobe.com/products/illustrator.html).

Species delimitation
Three molecular species delimitation analyses were conducted to aid the species hypothesis.Firstly, Species by Automatic Partitioning (ASAP) was performed on the in-group COI dataset applying the Kimura two Parameter (K2P) and the default setting parameters (Puillandre et al. 2021).Secondly, the Poisson Tree Processes model (bPTP) was implemented in the bPTP web server (https://species.h-its.org)applying default settings in the COI and concatenated tree resulted from the BI phylogeny (Zhang et al. 2013); and, third, the minimum COI p-distance was calculated applying default settings on Mega X version 10.2.4 (Kumar et al. 2018).Type locality.Ponta do Ouro, Mozambique (26°51'26"S, 32°53'4"E).

Systematics
Habitat.Specimens were collected on submerged subtropical compressed sandstone reefs in Ponta do Ouro, Mozambique.
Diagnosis.Body elongate-ovulated.Dorsum pale gray to pink, covered on highly dense caryphyllidia; rhinophores short, with up to nine lamellae, ending in a knob apex; six to nine bipinnate branchial leaves encircling the anal pore.Radula with five to seven very thin pectinated outermost teeth bearing long bundled fibrous denticles.Labial cuticle smooth.Copulatory spine with bifid apex.
Etymology.This species is dedicated to Livia Renée Cornelius, daughter of the second author of this paper.
Description.External morphology (Figs 1,2).Length varied from 11 to 30mm.Body elongate-ovulated, with gritty texture (Fig. 1A).Mantle covered on highly dense caryophyllid, evenly distributed on the dorsum (Fig. 2A).Caryophyllidia elongated, formed by five to eight spicules, projecting over tip, forming a crown of approximately 140 µm on the dorsum, taller on the margin of gill sheath (≈ 280 µm).Rhinophoral and branchial sheaths low, margin covered by caryophyllidia (Fig. 1D).Rhinophores short, retractable, with six to eight diagonal lamellae with a knob protruding apex (Fig. 1E).Gill with six to nine retractile, bipinnate branchial leaves, held vertically and forming a closed circle around the anal pore (Fig. 1F).Foot narrower than mantle, bilabiate anteriorly, upper lip bifurcate at center (Fig. 1B).Side of the foot covered by spicules (≈ 60 µm), spicules absent on foot sole (Figs 1B, C, 2B, C).Feet do not project beyond mantle in natural crawling position.Oral tentacles small and conical.Dorsum color pale pink to gray.Some specimens covered by pinkish-brown minute dots forming spots distributed on the notum.Gill and rhinophores translucent pinkish-white.Oral tentacle white.Upper lip translucent white with brownish dots.Foot pinkish-white.
Internal morphology.(Figs 3,4) The visceral mass is enveloped by a translucent-white tissue covered by brownish dots.Eye spots are visible by transparency.
Digestive system.Smooth labial cuticle (Fig. 3A).Oral tube long, about twice the size of oral bulb, with a pair of retractor muscles (Fig. 4A).Buccal bulb ovate, short, radular sac small and ovate, protruding ventrally, with a pair of strong retractor muscles (Fig. 4A).
Oesophagus passing through nerve ring, where it folds.Pair of salivary glands, relatively short, uniform, near the base of oesophagus (4A).Oesophagus connects to oval stomach.Intestine about half of oesophagus diameter.Caecum locate ventrally to stomach.Digestive gland cone-shaped, occupying approximately 30% of visceral mass.Anus opening at the center of gill circle.
Central nervous system.Central nervous system partially covered by blood gland.This is divided into two parts, anterior part about half the size of posterior part.Cerebral ganglia about half the size of pleural ganglia.Cerebral ganglia and pleural ganglia fused.Pedal ganglia ventrally located connected by a simple pedal commissure.Buccal ganglia short, ventrally located.Rhinophoral ganglia bulb-shaped, about 30% the size of cerebral ganglia.Eyes connected to cerebral gland by short rhinophoral nerve (Fig. 4B).
Reproductive system.Hermaphroditic duct leading to an ampulla long and convoluted, located between female gland and accessory gland.Ampulla branching into short oviduct and prostate.Flattened and ovulated prostate narrowing into a thin deferent duct, expanding into ejaculatory portion.Penis unarmed.Accessory gland size and shape varied according to the specimen, from pear-shaped and similar size to the female gland MNCN15.05/200187 to elongated and half of the size MNCN15.05/200189; in all specimens it narrows into a very thin, highly convoluted tube.Copulatory spine in accessory gland of approximately 1.25 mm (Fig. 3F).Vagina with similar length and width than deferent duct, leading to an oval bursa copulatrix.Thin duct near the vagina leads to oval seminal receptacle, about 2/3 of the size of the bursa copulatrix, which connects to a large female gland by a short uterine duct (Fig. 4C).
Natural history.This species has only been seen associated with the sponge A. brevispiculifera, on which the species is very cryptic (Fig. 5A, B).They are usually found at the base of the sponge branches but they have also been seen on other parts.When removed from the host sponge, the Jorunna liviae sp.nov.stretches the body curling the mantle toward the middle of the foot, similar to what Miller (1996) observed for J. ramicola.Perhaps this behavior aims to protect the sole of the foot which lacks caryophylliid.The white egg mass is also found on the same sponge and forms a close spiral ribbon of approximately five coils (Fig. 5F).A likely undescribed species of nudibranch egg-eater Favorinus sp. has been seen feeding on the J. liviae sp.nov.egg mass (Fig. 5C, D).Curiously, most of the time the egg ribbons are found on the tip of the sponge.Perhaps this strategy provides some protection against encrusting organisms due the higher water flux in this part of the sponge.Mating has been observed through July between specimens of different sizes and tonalities (Fig. 5E).Jorunna liviae sp.nov.seems to prefer sandy reefs with predominantly hydroids, soft coral and sponges.In Southern Mozambique, the flatter sand reefs have a higher density of sponges than the reefs with predominantly hard coral.
Molecular study and phylogeny.We successfully amplified the gene COI and H3 of four Jorunna liviae sp.nov.specimens.The phylogenetic trees constructed by BI and ML analyses of single gene datasets (Suppl.material 1) were not conflictive but differed in the ability to resolve phylogenetic relationships.The single gene H3 analysis retrieved the lowest resolution and the concatenate dataset the highest.Nevertheless, all Jorunna species were recovered with more than 50% support in all analysis.In general, the BI analysis better solved the relationship between species, while the ML analysis appears to reflect populational structure.Therefore, the results discussed below are based on the concatenated analysis (Fig. 6), except when stated otherwise.
The family Discodorididae formed a large polytomy.The genus Jorunna was divided in two paraphyletic clades, one containing all specimens of J. funebris (PP = 1; BS = 94) and another clade with the remaining Jorunna species (PP = 0.99; BS = 74).
The COI inter-specific variation (uncorrected p-distance) within the genus varied from 9.08% between J. tomentosa lineage B (LB) and J. artsdatabankia to up 16.92% between J. funebris and J. tomentosa lineage A (LA) (Table 1).The COI intra-specific variation of Jorunna liviae sp.nov.ranged from 0.16% to 1.08%.The closest species to Jorunna liviae sp.nov.was J. tomentosa lineage B with a minimum p-distance of 13.06%.ASAP retrieved 10 partitions, in both analysis (COI and concatenate) the partitions with higher score (asap-score 1.50-3) Jorunna liviae sp.nov.was retrieved as a distinct taxonomic unit.Curiously, J. funebris were retrieved as a species complex in all possible partitions.

Discussion
The phylogenetic relationships within the family Discodoridae are poorly solved.Most of the type species of Discodoridae genera are not sequenced which hinders our capacity to further understand the family.Jorunna is one of the few genera of the Discodoridae family which has its type species (J.tomentosa) sequenced.However, a recent study based on three genes (COI+16S+H3) reveals that it is uncertain if J. tomentosa represents two distinct lineages (Neuhaus et al. 2021).In our phylogenetic analysis, J. tomentosa is divided into two sub-clades (lineage A and B), which form a clade which is sister of J. artsdatabankia and related to J. onubensis and J. liviae sp.nov.Additionally, the genus Jorunna appears paraphyletic as J. funebris did not nest within the large Jorunna clade.In Camacho-García and Gosliner's (2008) morphological study, J. funebris nested on a clade together with J. rubescens, J. parva and J. pardus, which is sister of the clade containing the remaining Jorunna species studied by the authors.Unfortunately, to date no other species from the J. funebris clade has been sequenced.Consequently, the lack of molecular data from several Jorunna species hampers any further conclusion about the phylogeny of the genus.Nevertheless, it is clear that the species here described belongs to the genus Jorunna, as it forms a clade with the type species.In addition, the new species fits all the morphological diagnosis characters of the genus (see Camacho-García and Gosliner 2008).Interesting, all the species delimitation analyses suggest that J. funebris is a species complex, or alternatively, as proposed by Ip et al. (2019), there is an identification error in their sequences.Jorunna liviae sp.nov. is similar in appearance to the Atlantic species J. spongiosa and J. tomentosa.This latter is typically found in European waters (Atlantic and Mediterranean), but few records exist from South Africa and none of them from the Indian Ocean side (Camacho-García and Gosliner 2008;Neuhaus et al. 2021).Apart from the geography and genetic distance, these three species can be clearly distinguished by their radulae, in particular by the shape of the outermost teeth.These are very thin and pectinate in J. liviae sp.nov., hooked with small branches on J. spongiosa and slender hamate with up to 8 short denticles in J. tomentosa.In fact, the outermost pectinate teeth of Jorunna liviae sp.nov.are quite unique, and only similar to J. parva, a species also found in the WIO but easily distinguishable by the yellow background and dark caryophyllidia.Camacho-García and Gosliner (2008) provided detailed anatomical descriptions and comparative tables of Jorunna species by region.To better illustrate the differences between the species described in this study, we adapted and updated Camacho-García and Gosliner's (2008) comparative table of the Indo-Pacific Jorunna species, including recent distribution and morphological data observed by us, as well as the species J. liviae sp.nov.and J. labialia (Table 2).This latter species is found in the western Indian Ocean and Red Sea but was under 'Mediterranean and Western Atlantic' species in Camacho-García and Gosliner's (2008) comparative tables.
The Indo-Pacific species that most resembles J. liviae sp.nov. is J. ramicola; a species first described from New Zealand and likely occurring in Mozambique (Tibiriçá et al. 2017a).Jorunna ramicola is dull in color and bears long, slender outermost teeth which may be described as pectinate.However, these teeth bear much shorter and less numerous denticles, which do not form the distinct bunch as it does in J. liviae sp.nov.Moreover, the innermost   Present, curved spine ≈ 1.25mm long teeth in J. ramicola are denticulated, while in J. liviae sp.nov.they are simple hamate.In addition, the labial cuticle in J. ramicola bears jaw elements (Camacho-García and Gosliner 2008), while in J. liviae sp.nov. it is smooth.
Externally, they can be easily separated by the color of the rhinophores, which in J. ramicola is dark pigmented and in J. liviae sp.nov.whitish pink.Additional differences are provided in the comparative Table 2.

Conclusions
Based on morphological and genetic data there is no doubt that J. liviae sp.nov. is a newly discovered species.Here we provide the first sequence of Jorunna species to the WIO.We recommend further efforts to sequence other Jorunna species in order to clarify the monophyly of the genus and phylogenetic relationships.In addition, J. funebris specimens from different geographic regions should be morphologically and genetically examined as they may represent a species complex.

Figure 6 .
Figure 6.Bayesian inference tree based on the concatenate sequence dataset (COI+H3) collapsed (PP< 0.5).Numbers at the top of nodes indicate Bayesian Posterior probability (PP) and on the bottom bootstrap support from the maximum likelihood analysis (BS).Colored bars on the right represent the results of the species delimitation analyses on the Jorunna spp., from left to right: ASAP on COI dataset, PTP on COI dataset, bPTP on COI+H3 dataset.

Table 2 .
Comparative morphology of valid Jorunna species from the Indo-Pacific Ocean.