Molecular phylogenies confirm the presence of two cryptic Hemimycale species in the Mediterranean and reveal the polyphyly of the genera Crella and Hemimycale (Demospongiae: Poecilosclerida)

Sampling

Fragments of what a priori was thought to be Hemimycale columella were collected by SCUBA diving across the Northwestern, central and eastern Mediterranean, and Adriatic Sea, between 12 and 45 m of depth during several campaigns (Coconet, Benthomics, and MarSymbiOmics projects) (Table 1). Moreover, fragments of Hemimycale arabica and Crella cyathophora from the Red Sea (Dedalos and Ephistone) and Pacific (Bempton Islands) between 5 and 20 m depth were also collected (Table 1). Individuals were photographed underwater before sampling. Collected fragments were divided into two pieces, one of them was preserved in 100% ethanol, and after three alcohol changes, kept at −20 °C until DNA extraction; the other fragment was fixed in 5% formalin in seawater and preserved in 70% ethanol as a voucher for morphological and spicule studies. All vouchers have been deposited at the Sponge collection of the Centre d’Estudis Avançats de Blanes (numbers CEAB.POR.GEN.001 to CEAB.POR.GEN.029).

DNA extraction, amplification, and sequencing

DNA extractions were performed on two to three specimens per species and locality (totaling 18 individuals). Hemimycale spp. were extracted with QIAmp DNA stool kit (Qiagen), while Crella spp. were extracted with DNeasy Blood & Tissue kit (Qiagen) according to the manufacturer’s protocol. Standard primers were used for COI partitions M1–M6 (Folmer et al., 1994) and 18S rRNA (1F and 1795R, from Medlin et al., 1988), and Porifera primers for the D3–D5 partition of 28S rRNA (Por28S–830F and Por28S–1520R, from Morrow et al., 2012). Different amplification protocols were performed for each gene (Table 2). COI (M1–M6 partition) amplifications were performed in a 50 μL volume reaction, containing 37.6 μL H₂O, 5 μL buffer KCl (BIORON; F Holzinger Sales & Support, Germany), 2 μL BSA, 2 μL dNTP (Sigma; Sigma_Aldrich, Germany), 1 μL of primers, 0.4 μL Taq (BIORON; F Holzinger Sales & Support, Germany), and 1 μL of genomic DNA. 18S rRNA amplifications were performed in a 50 μL volume reaction, containing 36.85 μL H₂O, 5 μL buffer (Invitrogen, Carlsbad, CA, USA), 0.75 μL MgCl (Invitrogen, Carlsbad, CA, USA), 1.2 μL DMSO (dimethyl sulfoxide), 1 μL BSA, 1.5 μL dNTP (Sigma; Sigma_Aldrich, Germany), 1 μL of primers, 0.7 μL Taq (Invitrogen, Carlsbad, CA, USA), and 1 μL of genomic DNA. Finally, partition D3–D5 of 28S rRNA amplifications were performed in a 50 μL volume reaction, containing 36.85 μL H₂O, 5 μL buffer (Invitrogen, Carlsbad, CA, USA), 0.75 μL MgCl (Invitrogen, Carlsbad, CA, USA), 2 μL BSA, 2 μL dNTP (Sigma; Sigma_Aldrich, Germany), 1 μL of primers, 0.4 μL Taq (Invitrogen, Carlsbad, CA, USA), and 1 μL of genomic DNA. Polymerase chain reaction products were purified and sequenced in both directions using Applied Biosystems 3730xl DNA analyzers in Macrogen, Korea.

Sequence alignment and phylogenetic reconstructions

Sequences of COI, 28S, and 18S were aligned using Clustal W v.1.81, once their poriferan origin was verified using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), as implemented in Genieous 9.01 (Kearse et al., 2012). When sequences were identical, only one sequence per locality and species was included in the phylogenetic trees. After alignment, ambiguous regions were determined with Gblocks v.091 b software (Castresana, 2000), which removes from 1 to 4% of poorly aligned positions and divergent regions of an alignment of DNA. Representatives of family Hymedesmiidae (i.e., genera Phorbas and Hymedesmia) and Crambeidae (i.e., genera Crambe and Monanchora) were selected as outgroups. The inclusion of Crambeidae as an outgroup was decided because the species Hemimycale arabica had been reported to contain similar secondary metabolites (polycyclic guanidine alkaloids) to those of Crambe and Monanchora (Ilan et al. 2004).

JModelTest 0.1.1 (Posada, 2008) was used to determine the best-fitting evolutionary model for each dataset. The model GTR + I + G was used for both mitochondrial and nuclear genes. Phylogenetic trees were constructed under neighbor joining (NJ) (default parameters), Bayesian inference (BI), and maximum likelihood (ML) using Geneious software 9.01 (Kearse et al., 2012). NJ generates unrooted minimum evolution trees (Gascuel & Steel, 2006). BI analyses were performed with MrBayes 3.2.1 (Ronquist & Huelsenbeck, 2003). Four Markov Chains were run with one million generations sampled every 1,000 generations. The chains converged significantly and the average standard deviation of split frequencies was less than 0.01 at the end of the run. Early tree generations were discarded by default (25%) until the probabilities reached a stable plateau (burn-in) and the remaining trees were used to generate a 50% majority-rule consensus tree. ML analyses were executed with PhyML v3.0 program (Guindon & Gascuel, 2003; Guindon et al., 2005). The robustness of the tree clades was determined by a nonparametric bootstrap resampling with 1,000 replicates in PhyML. MrBayes and PhyML were downloaded by Genieous.

Incongruence length difference (ILD) test (PAUP 4.0b10) was run (Swofford, 2002) to verify sequence homogeneity or incongruence between the 18S rRNA and COI markers and the 18S and 28S rRNA markers. The ILD test indicated no significant conflict (p = 0.93 and p = 0.91, respectively) between the marker pairs to be concatenated. Thus, concatenated 18S COI and 18S–28S rRNA datasets were constructed for the species with sequences available for both markers. The phylogeny on the three genes concatenated was not performed due to the few species/individuals for which the three genes were available.

Phenotypic characters

To assess whether molecular differences among the target populations and species (Hemimycale columella, Senso latus, Hemimycale arabica, and C. cyathophora) were supported by morphological and spicule traits, the target species were observed both in situ and on recently collected samples. Moreover, spicules of all the species were observed through light and scanning electron microscopes (SEM) after removing the sponge organic matter from small (3 mm³) pieces of each individual by boiling them in 85% Nitric acid in a Pyrex tube and then washed three times with distilled water and dehydrated with ethanol 96% (three changes). A drop of a spicule suspension in ethanol per individual was placed on 5 mm diameter stuffs, air dry, and gold–palladium metalized (Uriz, Turon & Mariani, 2008) in a Sputtering Quorum Q150RS. Observation was performed through a Hitachi M-3000 Scanning Electron Microscope at the Centre d’Estudis Avançats de Blanes.

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: urn:lsid:zoobank.org:pub:48910653-0343-4A8D-911F-3498A755F305. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Code of Nomenclature for algae, fungi, and plants, and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. In addition, new names contained in this work have been submitted to MycoBank from where they will be made available to the Global Names Index. The unique MycoBank number can be resolved and the associated information viewed through any standard web browser by appending the MycoBank number contained in this publication to the prefix “http://www.mycobank.org/mb/283905”. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central, and CLOCKSS.

18S rRNA phylogeny

The resulting phylogeny using the 18S rRNA partition on 25 sequences (17 new) of 695 nt. (46 variable positions, from which 38 were parsimony informative) was congruent under BI, and ML and just differed in the position of Hemimycale arabica which appeared as a sister group of the remaining Crella spp. and Hemimycale spp. under NJ (Fig. S1). The representatives of the family Crambeidae (Monanchora) appeared as outgroups and the genus Phorbas was a sister group of the remaining species. In the BI, NJ, and ML trees, the genera Hemimycale and Crella appeared polyphyletic, with the Red Sea species Hemimycale arabica and C. cyathophora, far away from the Atlanto-Mediterranean Hemimycale and Crella species. The Atlanto-Mediterranean Crella formed a well-supported clade (1/81/98, posterior probability/bootstrapping values), which was the sister group of the Atlanto-Mediterranean Hemimycale (1/97/98). Moreover, the deep Hemimycale columella clustered with an Atlantic sequence downloaded from the GenBank (0.89/89/88) forming a separate clade from the also well-supported (1/97/98) group containing the shallow Mediterranean Hemimycale. No genetic differences for this partition were found among shallow individuals. In the BI and ML trees, the two individuals of Hemimycale arabica appeared in unresolved positions while they formed a poorly supported (75%) clade in the tree under the NJ criterion (not shown).

28S rRNA (D3–D5) phylogeny

The 28S rRNA (D3–D5) dataset comprised 31 sequences (24 new) of 623 nt. (84 variable positions from which, 60 parsimony informative).

The resulting phylogenies were congruent with the three clustering criteria and matched in most cases the phylogeny based on the 18S rRNA partition, although the supporting values of some clades were in some cases slightly lower (Fig. S2).

The three phylogenies retrieved the representatives of Family Crambeidae (Monanchora and Crambe) as an outgroup. The monophyly of the in-group containing Crella spp. and Hemimycale spp. was strongly supported under the BI, NJ, and ML criteria (1/100/100). The genus Phorbas was a sister group of the remaining species considered. Crella was polyphyletic, with C. cyathophora separated from the well-supported clade (1/100/100) encompassing the Atlanto-Mediterranean Crella. The latter appeared as a sister clade of a poorly supported group (0.7/77/70) harboring C. cyathophora and Hemimycale spp. The Hemimycale spp. group, although monophyletic, was poorly supported under the NJ and ML criteria (77/70) while the Atlanto-Mediterranean Hemimycale clade was well supported under the three clustering criteria (1/92/95).

The deep and shallow Mediterranean populations of Hemimycale formed two well-supported monophyletic groups (0.96/87/83 and 0.96/ 100/98, for deep and shallow individuals, respectively), the former containing the Atlantic sequence of Hemimycale columella. No genetic differences for this partition were retrieved for shallow individuals despite their spread distribution across the Mediterranean. The individuals of C. cyathophora from the Red Sea clustered with those from the Pacific collected between Australia and Nouvelle Caledonie (1/89/76).

COI phylogeny

The COI dataset included 21 sequences (15 new) of 535 nt. (169 variable positions, from which 149 parsimony informative).

The COI phylogeny, which was congruent under BI, NJ, and ML, also retrieved the representatives of Crambeidae as outgroups of the group formed by Crella, Phorbas, and Hemimycale. The genus Phorbas clustered with the Atlanto-Mediterranean Crella spp. (0.98/100/86) likely because we only included one individual/species of Phorbas (Fig. S3).

A clade containing Hemimycale spp. and C. cyathophora was well supported (0.94/94/80). The Hemimycale clade was divided into two subclades corresponding to deep and shallow individuals. No genetic differences among shallow individuals were found. A sister, well supported group (1/100/94) contained C. cyathophora and Hemimycale arabica representatives with almost no genetic differences between them (Fig. S3).

Concatenated trees

The concatenated 18S + 28S rRNA (Fig. 1) confirmed the outgroup position for the Crambeidae representative (Monanchora), the polyphyly of Crella with the Red Sea and Pacific species forming a separate clade (1/100/100) from the Atlanto-Mediterran Crella, which appeared in a non-resolved position. Hemimycale also appeared polyphyletic, but the position of Hemimycale arabica was unresolved. The Atlanto-Mediterranean Hemimycale clade was confirmed as well as its division into two subclades: one containing the deep Mediterranean individuals together with two Atlantic sequences of the species and the other one harboring the shallow Mediterranean individuals, which did not show any genetic difference across the Mediterranean and Adriatic Sea.

The concatenated 18S rRNA + COI (Fig. 2) tree contained only 13 sequences and no representative of Crambeidae could be included. The representatives of the Atlanto-Mediterranean Crella appeared as outgroups of the remaining target species, which formed two well-supported clades: one containing C. cyathophora and Hemimycale arabica representatives (1/100/100) and the other with the Atlanto-Mediterranean Hemimycale (1/100/100) divided into two monophyletic well-supported groups (deep and shallow individuals).

Species description

• Genus Hemimycale Burton, 1934

• Sequence accession Numbers GenBank (Table 1)

• Type species Hemimycale columella (Bowerbank, 1874)

Hemimycale is the only genus of Hymedesmiidae that has smooth diactines and monactines exclusively (Van Soest, 2002). The genus was described by Burton (1934) as “reduced Mycaleae with skeleton of loose fibers of styli, sometimes modified into anisostrongyles, running vertically to the surface; fibers tending to branch and anastomose; no special dermal skeleton, no microscleres.”

The spicule complement described by Burton; however, seems different from that reported in the several modern redescriptions of Hemimycale columella (Vacelet, Donadey & Froget, 1987), which report predominant straight anisostrongyles with rare or absent styles. Indeed, Burton stated that the Bowerbank representation of Hemimycale columella spicules was wrong because it figured anisostrongyles instead of styles, and was precisely the dominance of styles what induced Burton to place the species among the Mycaleae. The termination of the diactines either round or pointed ends may be the result of different silica concentration in the water masses, as reported for other siliceous sponge skeletons (Uriz, 2006), but it cannot be totally discarded that the Burton Hemimycale columella belonged in another Hemimycale species.

• Species: Hemimycale columella (Bowerbank, 1874)

• Sequence accession numbers GenBank (Table 1)

Description (Figs. 3A–3D): Encrusting to massive sponges. Surface smooth, covered with circular inhaling, areas up to 6 mm in diameter with an up to 3 mm high rim. Morbid and fleshy consistence. Translucent to whitish ectosome, difficult to separate from the choanosome. Thousands of calcareous spherules, 1 μm in diameter formed by intracellular calcifying bacteria (Uriz et al., 2012) are spread through the sponge mesohyl and specially accumulated at the sponge periphery of whitish individuals (L. Garate et al., 2013–2014, unpublished data).

• Color from pinkish-orange to whitish outside, dark orange inside.

Spicules (Table 3; Fig. 4F): Asymmetric strongyles (anysotrongyles), straight, 302–435 × 3–4 μm in size. Styles rare or completely absent from the Mediterranean specimens (this study) and Canary Islands (Cruz, 2002).

Skeletal arrangement: Plumose branching bundles of anysostrongyles together with spread spicules. A palisade of vertical anysotrongyles forms the rim around the inhaling areas.

Distribution: Northeastern Atlantic (United Kingdom and Ireland coasts) Canarias Islands (Cruz, 2002), western Mediterranean: Tossa de Mar, Arenys de Mar, from 28 to 60 m depth (this study). It is not possible to confirm whether previous Mediterranean records of the species (see Vacelet & Donadey, 1977) belonged to Hemimycale columella or to Hemimycale mediterranea.

Biology: Multiannual life span, ca. 60% survival after two monitoring years; maximum growth in autumn–winter (L. Garate et al., 2013–2014, unpublished data). Larval release occurs at the beginning of November in Mediterranean populations (M. J. Uriz, L. Garate & G. Agell, 2013–2014, unpublished data).

• Species: Hemimycale mediterranea sp. nov. (Figs. 3E–3H)

• Sequence accession numbers GenBank (Table 1)

Description: Thick encrusting sponges with aerolate inhaling areas up to 3 mm in diameter, surrounded by an up to 1.5–2 mm high rim, which in some cases barely surpasses the sponge surface. Thousands of calcareous spherules, 1 μm in diameter formed by intracellular calcifying bacteria are spread through the sponge mesohyl and specially accumulated at the sponge periphery (Garate et al., in press).

Ectosome: Firmly attached to the choanosome.

Color: Flesh to clear brownish externally, more or less whitish depending on calcibacteria accumulation at the surface, sometimes partially covered by an epibiotic (reddish or pinkish) cyanobacteria.

Spicules (Table 3; Figs. 4A–4E): Smooth, uniform in size, straight, anysostrongyles, 200–296 × 3–4 μm in size. Styles completely absent.

Skeletal arrangement: Plumose undulating bundles of anysostrongyles together with spread spicules. A palisade of vertical anysotrongyles forms the rim around the inhaling areas.

Known distribution: Northwestern Mediterranean, central Mediterranean, Adriatic, eastern Mediterranean (Spain: Cap De Creus, Tossa, Blanes, Arenys, South Italy: Croatia, Tremiti, Turkey, Greece) between 3 and 17 m deep.

Biology: Annual life span, maximum growth rates in summer (M. J. Uriz, L. Garate & G. Agell, 2013–2014, unpublished data). Larval release at the end of September beginning of October (M. J. Uriz, L. Garate & G. Agell, 2013–2014, unpublished data).

In most cases, it is difficult to ascertain whether individuals of Hemimycale columella recorded by other authors belong to Hemimycale columella or Hemimycale mediterranea. The redescription of Hemimycale columella by Van Soest (2002) based on the holotype (from the Atlantic) reported large aerolate porefields with elevated rims, which are shared with the deep Mediterranean specimens of Hemimycale columella (Figs. 3A–3D) in contrast to the small, short-rimmed porefields showed by Hemimycale mediterranea sp. nov. Both species have mainly straight slightly asymmetric strongyles but the spicule sizes are systematically larger in Hemimycale columella (Table 3). However, while styles were rarely present in Hemimycale columella individuals, they have not been found in specimens of Hemimycale mediterranea sp. nov. The external color also differs between the two species, being orange to pinkish in Hemimycale columella and flesh color to brownish Hemimycale mediterranea sp. nov. (Figs. 3E–3H). Vacelet & Donadey (1977) reported two different color forms occurring side by side on the littoral of Provence (France), one pink cream and the other one brownish. Likely the second color morph, which besides had smaller strongyles, corresponded to the Hemimycale mediterranea sp. nov.

Color has not received much attention as a diagnostic character in sponges because it has been generally considered to be a response to higher or lower light irradiance at the sponge habitat, or to the presence of epibiotic or symbiotic cyanobacteria. However, color has proven to be taxonomically relevant to distinguish other invertebrates such as shrimp species (Knowlton & Mills, 1992) and also sponge species of the genus Scopalina (Blanquer & Uriz, 2008), and thus it seems worthy to be taken into account in sponge taxonomy.

The slight phenotypic differences found between the two species appear; however, consistent across individuals and localities within the Atlanto-Mediterranean basin. Moreover, their ecological distribution and bacterial symbionts, strongly differentiate these two cryptic species. For instance, although calcareous spherules produced by intracellular bacteria are present in the two species, the producer bacteria belong in different species (Garate et al., in press), and the respective microbial communities totally differ (Garate et al., in press). Symbionts, as predators do (e.g., Wulff, 2006), often distinguish their target sponge preys or hosts while the species remain morphologically cryptic to taxonomists. Moreover, Hemimycale mediterranea sp. nov. shows an annual life span, with individuals disappearing after larval release, while Hemimycale columella has a multiannual life span (M. J. Uriz, L. Garate & G. Agell, 2013–2014, unpublished data) and growth dynamics also differs between the two species, as Hemimycale mediterranea sp. nov. grows more in summer, while Hemimycale columella grows preferentially in autumn–winter (M. J. Uriz, L. Garate & G. Agell, 2013–2014, unpublished data).

The contrasting ecological distribution of these two cryptic species in the Mediterranean helps to predict their identity in the field. Hemimycale mediterranean sp. nov. inhabits shallower zones than Hemimycale columella. However, it is likely that both species may share occasionally habitats, as the record of the two color morphs side by side (Vacelet & Donadey, 1977) indicate. Hemimycale mediterranea sp. nov. seems to be more abundant and widespread in the Mediterranean than Hemimycale columella. Molecular differences between groups of individuals of Hemimycale columella suggest the possible presence of additional cryptic species among the deep Mediterranean Hemimycale.

The presence of two morphologically cryptic Hemimycale species in the Mediterranean, which show contrasting biological traits, reinforces the idea that cryptic species represent something more than wrong taxonomic identifications or biodiversity underestimates. They may feature contrasting biological cycles and life spans, and puzzle biological studies, which may invalidate conservation policies based on those studies.