Samples of the jellyfish Pelagia noctiluca were collected from the “Cala Cupa” cove (4222008.0300 N, 1055004.0900E), Giglio Island (Tuscan Archipelago, North Tyrrhenian Sea) at 10 m depth near a previously studied meadow in May 2019. After capture by scuba divers, five entire living jellyfishes were placed separately in sterile containers and then stored in a refrigerator (icebox), and brought to the laboratory. The five animals were asymptomatic and well-developed, with equivalent size (umbrella range 9–11 cm in diameter). The samples were prepared for the isolation of fungi as follows. The samples were washed (five time) with sterile seawater to remove sediments, debris, and transient microorganisms (i.e., propagules not strictly associated with jellyfish) and gently blotted with sterile filter paper to remove water (Yue et al., 2015). To evaluate the effectiveness of this procedure in removing all the transient propagules, all washing solutions were plated on Potato Dextrose Agar Sea water (PDAs: PDA 39 g—Sigma-Aldrich, St. Louis, MO, United States—dissolved in 1 L of filtered seawater) and incubated at 20°C; to assess possible propagule growth (no growth was observed plating the last rinsing solution). Samples of the umbrella (U) and oral arms (OA) were excised manually from each animal and separately placed into sterile glass Petri dishes (80 mm × 15 mm). The inner tissues (IT) of the animals, including gonads, were also removed and placed in sterile Falcon tubes (50 mL).

In order to optimize the isolation procedures for the umbrellas and the oral arms, some preliminary tests were carried out to compare the homogenization and the direct plating methods. The results demonstrated that the isolation yield using the homogenization method was definitely lower than that achieved by the direct plating. Moreover, all the fungal species isolated by the homogenization method were also found by direct plating. Accordingly, the isolation was carried out using two different techniques for the different districts of the animals: “direct plating” for OA and U, and “homogenization and plating” for IT, within 24 h from sampling.

Oral arms and U samples were cut into small pieces (10 pieces of ca. 5 mm² for each substratum and sample) and directly placed on Petri dishes (five pieces for each plate 9 cm Ø) containing PDAs. Twenty Petri dishes were set up for OA and U.

For IT, 5 g (1 g from each animal) of tissues was added to 10 mL of sterile seawater and homogenized (ULTRA-TURRAX®-IKA, Staufen, Germany). The homogenate and 1:10 and 1:100 dilutions, in sterile sea water, were used to inoculate the plates: 0.5 mL of each dilution was spread in five Petri plates (9 cm Ø) containing PDAs. Five plates for each substratum and dilution were performed.

To prevent bacterial growth all media were supplemented with an antibiotic mix of 0.2 g/L streptomycin sulfate and 0.007 g/L penicillin G (Sigma-Aldrich, St. Louis, MO, United States). All procedures were carried out under sterile conditions.

A total of 50 plates were incubated at 20°C in the dark for up to 6 weeks and regularly monitored. All developed colonies were collected, transferred on appropriate media, and isolated in pure cultures. Rejection of duplicates (dereplication) was carried out by preliminary morphological, taxonomical and physiological analyses on different cultural media: PDAs, Corn Meal Agar Sea water (CMAs: 17 g CMA—Sigma-Aldrich dissolved in 1 L of filtered sea water), Malt Extract Agar Sea water (MEAs: 50 g MEA—Sigma-Aldrich dissolved in 1 L of filtered seawater), and Czapeck Sea water (CZs: 49 g CZ—Sigma-Aldrich, dissolved in 1 L of filtered seawater). However, every strain showing any morphological differences at macroscopic (colony) or microscopic (reproductive structures: asexual conidia, conidiphores, conidioma, or sexual structures) level were maintained as different morphotypes. The same dereplication approach was used considering physiological features revealed by the growth on different substrata (rate of growth, texture of the colonies, exudate, and pigment production).

All different morphotypes were cryo-conserved (Cryoinstant Mixed, VWR, Leuven, Belgium) and maintained in the culture collection of the “Laboratory of Ecology of Marine Fungi,” (CoNISMa) Department of Ecological and Biological Sciences (DEB), University of Tuscia, Viterbo, Italy. The strains of new taxa were also deposited at the international public institution, Mycotheca Universitatis Taurinensis (MUT).

A polyphasic approach that included morpho-physiological, molecular, and phylogenetic analyses was used to identify fungal strains.

The effect of salinity on fungal growth was investigated on PDA modified with different NaCl concentrations: 0, 35, 70, 105, and 120‰. The plates (90 mm Ø) were inoculated with a single agar disk (diameter 3 mm) cut from the actively growing margin of 7-day-old colonies grown on PDAs. Plates were incubated at 25°C and colony diameter was measured daily for 7–30 days (in relation to the different growth rates) to estimate fungal growth. Experiments were carried out in triplicate.

Seven PCR primers, designed for the highly conserved sequences of β-ketoacyl synthase (PKS) domains and the most conserved A domain of the NRPS gene, were used. PCR were performed as previously reported in section 2.2.2. PCR conditions and primers are reported in Table 1.

Principal Component Analysis (PCA) on salt growth preferences were made on the normalized colony diameters on the maximum growth (CANOCO Engine Version 5.2 2017; Šmilauer and Lepš, 2014).

One hundred and sixty-four fungal isolates were obtained from samples of P. noctiluca. The fungal assemblages associated to umbrella and oral arms of each jellyfish were quite homogeneous, therefore they were merged to describe the P. noctiluca mycobiota. Considering the animals’ district, the inner tissues hosted the highest number of fungal colonizers (89 isolates), followed by umbrella (48 isolates), and oral arms (27 isolates). According to the morpho-physiological dereplication, 40 morphotypes were obtained. The morphotypes were characterized by molecular analysis using the nrITS region, the universal DNA barcode for fungi (Bergerow et al., 2010; Schoch et al., 2012). The analysis of the sequences (BLASTn) and subsequent phylogenetic analysis (Figure 1) allowed to group the 40 morphotypes in 23 distinct taxa. Nine morphotypes were identified at the species level, 26 at the genus level, and for the remaining five strains, all belonging to the Pleosporales order, no further attribution was possible with nrITS region. Considering these results, additional genes were analyzed to attain the attribution at the species level.

All Cladosporium and Penicillium isolates were identified at the species level by the actin and β-tubulin genes, respectively (Supplementary Figures S1, S2; Bensch et al., 2012; Visagie et al., 2016), with the only exception of PN20. For this morphotype, the analysis of the CaM gene was also assessed for its species attribution to P. fundyense. The β-tubulin gene was also used to confirm the morphological identification of Chaetomium elatum PN36. The two Botrytis strains PN7 and PN40 were analyzed by a polyphasic approach including both molecular and morphological analyses. A multi-locus phylogenetic analysis was carried out using the rpb-2, HSP60, and G3PDH markers, as reported by Staats et al. (2005) to distinguish the most closely related neighbor species (Supplementary Figure S3). The strain PN7 is included in a well-supported group with three different species: B. cinerea, B. eucalypti, and B. pelargonii. Nevertheless, as already observed by Garfinkel (2021), these species must be considered conspecific, thus the strain PN7 was attributed to B. cinerea; this attribution was also confirmed by morphological analysis. Also, PN40 is grouped with three different species B. fabiopsis, B. galanthina, and B. caroliniana that have different morphological characters including presence/absence of sclerotia and conidial dimensions (Li et al., 2012). These features lead to the attribution of PN40 to B. caroliniana since it does not produce sclerotia, and conidia dimensions are 10–14 × 6–9 μm.

A multi-locus phylogenetic analysis based on the nrLSU, nrSSU, and nrITS was used to analyze the 11 morphotypes included in the Pleosporales in order to determine their taxonomic position within this large order (Figure 2). Subsequently, different molecular markers (β-tub, tef1-α, and/or rpb-2) were selected and analyzed for each family to reach specific attribution of the strains (Table 2). The pleosporelean morphotypes were included in six families. The strains included in Lentitheciaceae (PN6), Lophiostomataceae (PN27), Neopyrenochaetaceae (PN31), and Roussellaceae (PN1) were attributed to Keissleriella sp. (PN6), Neopyrenochaeta acicola (PN31), Neovaginatispora fuckelii (PN27), and Roussoella intermedia (PN1). For PN6, the analyses carried out on nrITS, nrLSU, and nrSSU did not allow to distinguish the morphotype from the related species K. cladophila, K. camporesi, K. sparticola, and K. rosarum. The tef1-α sequences, efficiently used to attribute some species in the genus Keissleriella, are not available for the species strictly related to PN6; in addition PN6 is a mycelia sterilia and morphological analysis is useless for species attribution. It is worth to note that some authors reported the inconsistency of some species included in the genus Keissleriella and suggested the necessity of a large revision of this genus (Hyde et al., 2020).

The remaining Pleosporales strains included in Pleosporaceae (PN38, PN39, PN23, PN28, and PN32) and Phaeosphaeriaceae (PN9 and PN33) families can be all ascribed to new taxa. The pleosporeacean strains were all included in a strong supported clade with a strain of Tamaricicola sp. IG108, previously collected in the same area (Pasqualetti et al., 2020). The clade is strictly related with the monospecific genus Tamaricicola and seem to represent a new lineage inside this genus (Figure 2).

The strains PN9 and PN33 represent two new lineages (genera) inside the family Phaeosphaeriaceae (Figure 2). This is also strongly supported by the analyses of genetic distances among PN9 and PN33 and the closest taxa of the family for all tested markers (Supplementary Tables S6, S7).

According to Pasqualetti et al. (2020) analyses of salinity preferences were carried out to exclude, from the P. noctiluca fungal assemblage, the non-marine strains possibly present as propagules derived from terrestrial contamination and unable to growth at the sea salinity. All isolated strains are able to grow at sea salinity and included both halotolerant (36%) and facultative halophilic (64%) fungi (slight—and moderate–facultative halophiles, 56 and 8%, respectively; Gunde-Cimerman and Zalar, 2014; Raghukumar, 2017; Pasqualetti et al., 2020). However, the majority of strains showed their optimal growth at the sea salinity or higher (Table 2). Multivariate analysis carried out on growth data at different salinities, clearly identified four main groups (Figure 3). Strains characterized by optimal growth in the absence of salt or at salinities lower than that of seawater (halotolerant) were included in group I (PN7 and PN40) and group II (PN1, PN6, PN9, PN31, and PN33): Group I presented at sea salinity a growth reduction of 30% and Group II of 20% from the optimum. These strains showed a relatively low halotolerance with an 80–95% growth reduction (from the optimum) at salinities of 120‰. Isolates included in group III (PN16, PN36, PN29, PN43, PN44, PN3, PN35, PN15, PN20, and PN38) showed optimal growth at the sea salinity (35‰) and could be considered as facultative halophiles. Group IV included three strains, PN17, PN22, and PN25 presenting optimal growth at 70‰ (moderate-halophiles) and a remarkable euryhalinism, with strain PN25 showing an optimal growth in the range 35–105‰.

Overall, the mycobiota of P. noctiluca was mainly composed of Ascomycota (21 taxa; 91.3%) with a small contribution of Basidiomycota (two taxa, 8.7%). Considering the phylum Ascomycota, Dothideomycetes and Eurotiomycetes were the most representative classes with 13 and 5 taxa, respectively. This agrees with previous reports describing these classes as the most representative in marine environments, considering both the number of taxa and abundance (Gonçalves et al., 2022; Wijayawardene et al., 2022).

As for the phylum Basidiomycota, only two taxa (one strain of Tremellomycetes and one of Microbotryomycetes) were identified. Although the result is in line with those previously reported in marine habitats/substrata, it must be also considered that the Basidiomycota could be underestimated due to the use of not selective techniques (Ding et al., 2011, Gonçalves et al., 2022). With respect to the isolation districts, the highest number of species (18) and colonies (89) were associated with the inner tissues, while seven and four species were recovered from the umbrella and oral arms, respectively. Oral arms showed a scarce colonization (27 colonies) and none of the four species found were exclusively associated with this district: B. caroliniana and C. allicinum were isolated from OA and U, P. brevicompactum from OA and IT and C. delicatulum was detected in all districts. On the contrary, the 43 and 83% of colonizers recovered in the umbrella and the inner tissues, respectively, were isolated only from one district (Table 2).

The most represented genera, in terms of species, were Cladosporium (six species) and Penicillium (five species). This result is not surprising since these genera are well adapted to marine environments and appear to be generally ubiquitous in the oceans (Raghukumar, 2017). Some of these species were frequently recorded in several marine habitats and substrata. For example, P. antarcticum and P. brevicompactum are well-known as saprotroph in seawater and as biont/saprotroph of algae, phanerogams, and animals (Paz et al., 2010; López-Legentil et al., 2015; Bovio et al., 2017, 2018, 2019; Marchese et al., 2020, 2021). Other species (i.e., C. aggregatocicatricatum, C. westerdijkiae, and P. bialowienzense) have been recorded on fewer marine substrata, even if they are widespread in terrestrial habitats (Bovio et al., 2017, 2018; Maamar et al., 2020; Ben-Dor Cohen et al., 2021).

Finally, the 22% of taxa (including the new genera and species) were detected for the first time in the marine environment. The high percentage of new strains detected emphasize the gap in our knowledge about real diversity of fungi in the sea. In the meantime, it underlines the great importance to focus on conservation strategies of biodiversity, considering also present and future prospectives linked to dramatic global changes. In this context, the maintaining of marine species in culture collections of microorganisms represents an important tool for their preservation, possible ecosystem restoring, and other applications. Under the ecological point of view, these species might be exclusively associated with P. noctiluca and considered as species-specific (“specialized”) organisms (Rambelli et al., 2004). However, the available information regarding the associations of these fungi with jellyfishes and other marine substrata is too scant to allow more than a hypothetical consideration. To the best of our knowledge, P. noctiluca was never studied before. Moreover, it is worth noting that most mycological studies on jellyfishes focused only on single strains and their biotechnological potential, without providing any ecological information on the single species or on the whole mycobiota (Wright et al., 2003; Liu et al., 2011, 2012; La Kim et al., 2012a,b; Liu et al., 2013; Yue et al., 2015; Tao et al., 2016; Liu et al., 2018; Regalado and Ramirez, 2019; Tan et al., 2019; Yue et al., 2022; Li et al., 2023). Globally only 14 species belonging to Cladosporium, Purpureocillium, Tilletiopsis, Aspergillus, Epicoccum, Paecilomyces, Penicillium, and Phoma were collected from the different jellyfish species studied (A. aurita, N. nomurai, and Catostylus sp.) (Supplementary Table S8). However, the only study focusing on a jellyfish mycobiota regarded N. nomurai, which showed a relatively low diversity in the fungal assemblage represented by just seven isolates and five species (Yue et al., 2015).

In the present study, 23 different fungal species were found on P. noctiluca and none of them was detected in the mycobiota of the other studied jellyfishes. On the contrary, some common genera such as Cladosporium and Penicillium were observed. Nevertheless, the significance of this genera in the mycobiota jellyfish characterization appears of low relevance considering, as already observed, that species of these genera are generally ubiquitous in the seas. Lastly, it is worth noting that the mycobiota of P. notiluca showed a relatively high diversity, if compared to other marine epizoic fungal communities (López-Legentil et al., 2015; Bovio et al., 2019).

The potential biotechnological interest of the isolated fungi, within a blue growth strategy perspective, has been tested by both plate screening and molecular survey of the target genes PKS and NRPS. These genes are commonly used to find species of interest for production of bioactive metabolites (Bentley et al., 2002; Molnár et al., 2010) since they are involved in the biosynthesis of a broad range of compounds such as antibiotics, antifungal, antiviral, anticancer, mycotoxins, antifouling, and pigments (Raimundo et al., 2018; Stroe et al., 2024). Moreover, the presence of these genes in fungi associated to marine macro-organisms (i.e., sponges, corals, algae, and mangrove plants) could suggest their potential roles in the host chemical defense process (Zhou et al., 2011; Hafez Ghoran et al., 2023).

A preliminary screening (agar plug diffusion method) was carried out to evaluate the potential production of antimicrobial molecules by the isolated species cultivated on different substrata (Table 3). Different media were utilized to promote different metabolic pathways and metabolites (Pinedo-Rivilla et al., 2022). Globally, the 70% of the strains exhibited inhibitory activity against one or two of the tested microorganisms (B. pumilus; P. fluorescens; P. griseofulvum). In general, no activity was observed against P. fluorescens, 11 strains showed antibacterial activity against B. pumilus and 11 strains antifungal activity. Five strains (PN16, PN31, PN33, PN38, and PN40), were active against both P. griseofulvum and B. pumilus gram-positive bacteria. Strains included in the order Pleosporales (PN9, PN33, PN27, and PN38) showed the strongest antibacterial activity, while the two Penicillium strains PN22 (P. bialowienzense) and PN25 (P. brevicompactum) showed the strongest activity against P. griseofulvum. The antifungal activity showed by various fungal species could be due to the production of cell-wall degrading enzymes (i.e., chitinases); however, PN22 and PN25 did not show any chitinolytic activity (data not shown). Considering the different substrata, in general, the most efficient to induce the production of bioactive molecules was PDAs (81% of active strain), only B. caroliniana showed activity when growing on all saline media, while only P. brevicompactum showed antifungal activity in a salt-free medium (CYA). Even if the antimicrobial activity tested by the plug diffusion method is not exhaustive, this preliminary result confirms that jellyfish-associated fungi can be a good source of active metabolites (Liu et al., 2013; Yue et al., 2015).

The results of the molecular survey are generally in accordance with the antimicrobial activity observed with cultural methods. The 70% of active strains presented PKS and/or NRPS genes: NRPS was present in nine strains, while PKS only in four strains and only P. brevicompactum and P. anctarticum presented both genes. Considering that Cladosporium and Penicillium were the most representative genera in the P. noctiluca fungal assemblage and that the primers used were designed on these taxa, it was quite surprising that the number of positive strains was relatively low. This is particularly true considering that most Cladosporium and Penicillium strains are known to produce various bioactive compounds related to these genes (El Hajj Assaf et al., 2020; Li et al., 2024). In particular, in this study for the Cladosporium genus only C. delicatulum was positive to NRPS and none of them to PKS. Finally, it is worth noting that P. bialowienzense, showing the strongest antifungal activity, was negative to both genes and, as already observed, to chitinolytic activity; indicating that its biological activity is probably related to others metabolic pathways.

Some new (Phaeosphaeriaceae sp. PN33) or neglected species (i.e., N. fuckelii and R. intermedia) that have never been studied for the presence of these genes, positive in this study, deserve further investigation.

In addition, the four strain (PN36, PN3, PN44, and PN43) that did not show activity were also negative to the gene amplification. It is possible that these results underestimated the real presence of these target genes, since the primers used for their amplification were designed for other specific genera, i.e., Penicillium, Aspergillus, and Cladosporium (Bingle et al., 1999). Hence, it is possible that they were unable to amplify homologous genes in other genera and species, in particular considering new taxa (Lee et al., 2001).