Introduction

Seagrasses are a narrow ecological and taxonomical group of marine sessile vascular macrophytes well adapted to aquatic life that form extensive underwater meadows with significant ecosystem functions [1]. Seagrass meadows rank among the most productive aquatic ecosystems [2] and can store up to twice as much carbon as temperate and tropical forests [3]. Although the center of seagrass biodiversity lies in the Indomalaya ecozone [4], they occur in most coastal shallow areas, except for polar regions [5]. All seagrasses belong to the Alismatales and comprise some 12 genera with ca. 50–70 species [4, 6]. The Mediterranean Sea is home to four seagrass genera (Cymodocea, Halophila, Posidonia, and Zostera) encompassing four autochthonous (Cymodocea nodosa, Posidonia oceanica, Zostera marina, and Zostera noltii) and one alien invasive (Halophila stipulacea) species [7, 8]. Most of the Mediterranean sublittoral area occupied by seagrasses is dominated by the endemic P. oceanica (=Neptune seagrass). Its clonally reproducing meadows can spread up to 15 km while being hundreds to thousands years old [9]. However, many observations suggest that in certain areas, its populations are significantly declining due to a combination of mostly human-induced factors [7, 10, 11].

Seagrasses take up nutrients through their shoots and leaves, and hence, in comparison with terrestrial plants, significance of nutrient uptake through the roots is generally reduced [2, 12]. Congruently, seagrasses were reported to lack mycorrhizal symbioses [13], i.e., root-fungus symbioses utilized by most land plants to scavenge nutrients from recalcitrant substrates [14]. On the other hand, this lack might be as well due to comparably lower attention paid to seagrasses, especially in comparison with other plant guilds hosting mycorrhizal fungi, including freshwater aquatic plants and plants from saltmarshes and mangroves [15–17]. Nevertheless, similarly to terrestrial plants, seagrasses host endophytic fungi [18–22], although their ecophysiological function in the marine environment has not yet been understood. Fungal endophytes are commonly defined as mycobionts which live inside living plant tissues, lack localized interfaces or specialized hyphae for nutrient transfer, do not synchronize their development with plant development, and do not provide nutritional benefits to the plant [23]. The term “mycobiont” as used here is more general and may specifically refer to the situations where the ecophysiological function of the fungal symbiont remains unknown.

Despite its ubiquity and dominance in the Mediterranean Sea, the mycoflora of P. oceanica has been studied only by a few authors. The arguably oldest published information dates back to 1840s and reports two fungal species, Sphaeria biturbinata and Sphaeria posidoniae, in P. oceanica rhizomes [24]. More than a century later, Kohlmeyer [25] transferred these fungi in two new genera, Halotthia [with Halotthia posidoniae (Dur. et Mont.) Kohlm.] and Pontoporeia [with Pontoporeia biturbinata (Dur. et Mont.) Kohlm.]. Some 20 years later, Cuomo et al. [26] investigated the mycoflora of P. oceanica leaves, sheaths, rhizomes, and roots and found seven marine lignicolous fungi, namely H. posidoniae (occurred in 87 % of samples), Corollospora maritima Werdermann (70 %), Papulaspora halima Anastasiou (60 %), P. biturbinata (56 %), Lulworthia sp. (54 %), Phoma sp. (44 %), and Corollospora intermedia Schmidt (9 %). The authors pointed out that the detected diversity was comparably lower than in the intertidal grass Spartina alterniflora (29 species detected) and the mangrove Rhizopora mangle (31 species detected) as reported by Kohlmeyer and Kohlmeyer [27]. More recently, fungi associated with P. oceanica leaves, rhizomes, roots, and dead parts (matte) were studied by Panno et al. [28] who isolated 88 species in total, 14 of them being associated with the roots. Intriguingly, none of them belonged among those reported by Cuomo et al. [26], and most of them belonged to fast-growing sporulating species also known from terrestrial habitats, probably as a consequence of no surface sterilization of the seagrass organs prior to isolation. Another recent report is by Torta et al. [29] who however isolated only one fungus, contrary to the previous studies. Most recently, an anatomically and morphologically unique dark septate endophytic (DSE) association was discovered in the roots of P. oceanica at 11 localities in the northwest Mediterranean Sea [22], but the fungus or fungi forming this symbiosis have not been identified.

Terrestrial root endophytes may, among other possible functions, engage in nutrient uptake and transport to the host plant [30] and protect host roots against pathogen attacks [31]. On the other hand, nearly nothing is known about possible roles of endophytes in seagrass roots, and this is probably due to both the low number of seagrass mycobiont studies and their differing results. Therefore, we conducted a culture-dependent screening of P. oceanica endorhizal mycobionts at 11 localities in the NW Mediterranean Sea (the same as in [22]), using an isolation technique similar to those employed in the previous reports, but modified as follows: (1) In contrast to Panno et al. [28], we included a surface sterilization step to eliminate common saprobic contaminants, and (2) in contrast to Torta et al. [29], we prolonged the isolation period as some marine fungi are notoriously slow-growing [27]. We hypothesized that this approach coupled with the relatively large area under investigation would lead to detection of common P. oceanica endorhizal mycobionts, possibly comprising at least some of the previously reported fungi associated with P. oceanica roots. We also hoped that this approach could shed light on the fungus or fungi forming the recently discovered DSE association in P. oceanica roots [22].

Materials and Methods

Root Sampling

P. oceanica (L.) Delile roots were collected using free and scuba diving between June and September 2012 at 11 localities in the NW Mediterranean Sea in Croatia, Italy, France, and Spain (Fig. 1, Table 1; [22]). At each locality, roots of five different individuals (at least 3 m apart) were carefully excavated from the substrate, separated from the shoots, mixed to make a compound sample, and inserted into 50-ml plastic beakers filled with seawater. These were stored in the dark in a portable fridge and processed for isolation of mycobionts in the evening of the respective collection day.

Fig. 1
figure 1

Location of the sampling sites within the NW Mediterranean Sea. Posidonia oceanica roots were sampled at 11 localities in Croatia (HR), Italy (IT), France (FR), and Spain (ES). The numbering of the localities follows Table 1 and [22]

Full size image
Table 1 List of localities screened in this study
Full size table

Isolation of Posidonia oceanica Root Fungal Symbionts

Mycobionts were isolated from fresh P. oceanica roots using a portable flow box to prevent air contaminants. Healthy looking turgescent terminal roots (ca. 1–2 mm in diam.) were selected from bulk samples, surface-sterilized in 10 % SAVO (household bleach, Unilever ČR Ltd., Czech Republic; contains 4.5 % available chlorine) for 30 s, rinsed two times in sterile deionized water, cut into ca. 3–5-mm segments, and placed on agar media in four-compartment Petri dishes. We used potato dextrose agar (PDA; HiMedia Pvt. Ltd., India) and modified Melin-Norkrans agar (MMN) [32], both amended with Novobiocin sodium salt (50 mg/L; Sigma-Aldrich, Germany) to prevent growth of bacteria and NaCl (38 g/L) to adjust osmotic pressure [33]. Roots from two localities (#3 Cogoleto and #9 l′Escala) were additionally plated on PDA and MMN without NaCl. There were five dishes each with 16 root segments totaling 80 root segments per each medium/locality combination. The dishes were sealed with an air-permeable tape, kept at room temperature in the dark, and observed for hyphal growth each day during the first 2 weeks then once each week until 101 days. Obtained fungal colonies were conservatively grouped into morphotypes, and several representatives of each morphotype/medium/locality combination were transferred to new petri dishes with MMN + NaCl.

Microscopic Observations

Hyphal colonies emerging from the root segments were screened using an Olympus SZX12 stereomicroscope and a FEI Quanta 200 scanning electron microscope equipped with the Olympus ESEM mode. Hand semi-thin cross sections were made from several root segments with emerging dark mycelia to investigate their possible intraradical origin. The cross sections were screened with an Olympus BX60 microscope equipped with DIC. Photos were taken with an Olympus DP70 camera and adjusted for clarity in Adobe Photoshop (Adobe Systems, USA) and Paint.NET (Brewster, Jackson and contributors + Microsoft Corporation, USA) as needed. Figures were assembled in Adobe Photoshop and Paint.NET.

Molecular Identification of Posidonia oceanica Root Fungal Symbionts

DNA was extracted from representatives of all morphotypes (Table 2) using Extract-N-Amp Plant Kits (Sigma-Aldrich, Germany) following manufacturer’s instructions. SSU (18S) ribosomal DNA (rDNA) was amplified using the NS1 + NS24 primer pair [34], ITS1-5.8S-ITS2 rDNA using the ITS1F + ITS4 primer pair [34, 35] and LSU (28S) rDNA using the LR0R + LR5 primer pair [36]. For PCR parameters and gel electrophoresis, see [37]. PCR products were purified and sequenced by Macrogen Europe Laboratory (Macrogen Inc., The Netherlands/South Korea) using the NS1, NS4, NS5 and NS24, ITS1F, and LR0R primers.

Table 2 Identity of fungal isolates obtained from Posidonia oceanica roots based on ITS nrDNA similarity to sequences deposited in GenBank, according to BLAST
Full size table

The obtained ITS rDNA sequences were screened in Finch TV v1.4.0 (geospiza.com/finchtv) for possible machine errors and edited when needed. They were subsequently aligned in BioEdit v7.0.5.3 using ClustalW [38], and the alignment was used as a matrix for NJ analysis in TOPALi v2.5 (topali.org). The threshold limit for grouping of sequences was set at 97 %. Sequences within separate groups were further aligned to screen their heterogeneity, and the most divergent were subjected to BLAST searches (megablast/blastn algorithms) in GenBank [39], and their taxonomic position was further checked with Blast Tree View (NJ, max. seq. difference 0.75). This was only sufficient for identification of the least abundant sequence group at the species level. The same approach with LSU sequences enabled identification of the second most frequent group at the genus level. The remaining group of the most frequent dark septate mycelia was subjected to the following phylogenetic analysis.

Taxon selection was patterned on reduced LSU-SSU dataset of Liu et al. [40] (TreeBASE matrix no. 14874). Sequence alignments were obtained using MAFFT 6 (http://mafft.cbrc.jp/alignment/software/) [41]. The dataset consisted of 57 sequences, 1894 positions (463 variable and 353 parsimony-informative sites). Maximum likelihood analyses were performed using PhyML 3.0 [42], and bootstrap support was obtained using 500 replicates. Evolutionary model (TN93+G+I) was determined for all datasets using MEGA 5.2.1 [43].

Results

From the total 880 surface-sterilized root segments, we obtained 177 relatively slow-growing fungal colonies. Contrary to endophyte isolations from roots of terrestrial plants [44], there were no contaminations by bacteria or fast-growing sporulating fungi. First hyphal colonies started to emerge from the surface-sterilized roots after ca. 4 weeks, and several dark brown to black colonies emerged even after 12 weeks of incubation. The best medium for mycobiont isolation was PDA with NaCl (141 isolates) followed by MMN with NaCl (32 isolates).

The 177 fungal isolates were conservatively grouped into three morphotypes (black, yellow, and ochre) with 160, 16, and 1 isolate, respectively. These morphotypes differed not only in colony color and morphology but also in the isolation time and growth rate, the black morphotype emerging significantly later and growing slower than yellow and ochre. Isolates of the most frequent black morphotype produced thick slow-growing melanized hyphae on the root surface and formed either aerial colonies on the MMN medium (Fig. 2a) or compact colonies on the PDA medium (Fig. 2b–d). The surface of the compact colonies was covered by terminally swollen hyphae resembling capitate cystidia (Fig. 2e; cf. [37]). The microscopic screening of the hand sections showed that most of the colonies emerging from root segments actually developed from intracellular melanized microsclerotia occurring in the P. oceanica root hypodermis (Fig. 2f). Isolates of the most frequent black morphotype were obtained from all but two localities (#8 Anse de Paulilles and #10 Tamariu) and dominated P. oceanica root mycobiont communities at five localities (Table 2). The second most frequent yellow morphotype grew faster and produced non-melanized hyaline to yellow hyphae forming compact colonies which were partly submerged in the cultivation medium. The yellow isolates were obtained at seven localities and were the only isolates obtained at one locality (#8 Anse de Paulilles), although in a low number (Table 2). In total, only one ochre isolate was obtained at a single locality (#10 Tamariu; Table 2). The great majority of black isolates were obtained on PDA with NaCl (134 isolates); on the other hand, there seemed to be no medium preferences in the case of the yellow isolates.

Fig. 2
figure 2

Morphology of the pleosporalean mycobiont (Pleosporales sp. MV-2012, the black morphotype) isolated from Posidonia oceanica roots. a Aerial hyphae emerging from a surface-sterilized root plated on modified Melin-Norkrans agar (MMN) with sodium chloride (NaCl, 38 g/L). Stereomicroscopy, bar = 500 μm. b, c Typical structures formed by the pleosporalean mycobiont when growing out from segments of Posidonia oceanica roots incubated in the dark on potato dextrose agar (PDA) with NaCl. Scanning electron microscopy (SEM), bars = 1 mm. d Compact black colonies formed on a surface-sterilized root segment plated on PDA with NaCl, identical in nature with those depicted at b and c using SEM. Stereomicroscopy, bar = 500 μm. e A detail of some superficial hyphae with terminal swellings developed on the surface of the compact black colonies (as in d). Light microscopy with DIC, bar = 20 μm. f A cross section of a root segment with aerial hyphae originating from intracellular microsclerotia (arrows). MMN with NaCl, light microscopy with DIC, bar = 20 μm

Full size image

DNA was successfully isolated and amplified from all examined isolates, and gel electrophoresis yielded a single band in all tested samples. All sequences were of sufficient quality and length. Three separate groups were delimited within our sequence dataset. Using BLAST searches, two sequence groups could be identified at the species/genus level: The ochre isolate P07 was identified as Fuscoporia torulosa (Basidiomycetes, Hymenochaetaceae) and the yellow isolates P01–P04, P12, P13, P21, and P32 as Lulworthiales sp. MV-2012 (Ascomycetes) (Tables 2 and 3). The dominant conspecific black isolates (Pleosporales sp. MV-2012) showed the closest similarity to members of the Aigialaceae family (Ascomycetes, Pleosporales) such us Aigialus grandis (SSU: 98 % to AF441172), Rimora mangrovei (LSU: 93 % to GU479798), or Fissuroma aggregata (ITS: 87 % to JN846718) (Tables 2 and 3). In the phylogenetic analysis based on SSU-LSU rDNA, they formed a sister lineage to the genus Aigialus and together with Ascocratera manglicola and Rimora mangrovei formed a strongly supported clade within Aigialaceae (Fig. 3) [40]. The respective sequences were deposited in GenBank under accession numbers KC145421-4, KC145426-32, and KC412711-21 (ITS rDNA), KC736937-46 (LSU rDNA), and KJ210571 (SSU rDNA).

Table 3 Identity of fungal isolates obtained from Posidonia oceanica roots based on LSU nrDNA similarity to sequences deposited in GenBank, according to BLAST
Full size table
Fig. 3
figure 3

A PhyML tree based on a combined dataset of SSU (894 bp) and LSU (919 bp) rDNA sequences of the pleosporalean mycobiont (Pleosporales sp. MV-2014, the black morphotype) from Posidonia oceanica roots. Bootstrap support values greater than 50 % are given above the nodes. Thickened branches represent 100 % bootstrap values. The clade of the pleosporalean mycobiont dominating the cultivable fungal communities in P. oceanica roots is highlighted in green color. The tree was rooted with Myriangium hispanicum (Myriangiales) and Dothidea insculpta (Dothideales). For details on the sequences used, see “Materials and Methods”

Full size image

Discussion

Only a few studies focused on root mycobionts of seagrasses in general and P. oceanica in particular. Here, we detected only three cultivable fungal species associated with P. oceanica surface-sterilized roots. Moreover, one of them, F. torulosa, was represented by only a single isolate. This number is therefore comparable to Cuomo et al. [26] and Torta et al. [29] who detected only one species and comparably lower than Panno et al. [28] who detected 14 fungal species associated with the seagrass roots. In contrast to Cuomo et al. [26], we did not detect C. maritima, and in agreement with Torta et al. [29], we detected Lulworthiales sp. MV-2012 with nearly identical ITS rDNA sequences to their “Lulwoana sp.”. In contrast to our hypothesis, we did not detect any of the species reported by Panno et al. [28]. On the other hand, Panno et al. [28] did not detect any species reported by Cuomo et al. [26] and concluded that this might be due to seasonality of sampling as they took their samples in spring while Cuomo and colleagues in autumn. We argue with this conclusion and point out that a more likely explanation is the method used by Panno and colleagues for isolation of P. oceanica mycobionts: they used no surface sterilization but only serial washing and blended the samples with sterile seawater. Such an approach leads to preferential isolation of fast-growing surface-dwelling saprobes in contrast to true endorhizal mycobionts [45]. This is especially true in the aquatic environment where submerged solid substrates entrap substantial numbers of fungal propagules which may be present in dormant or minimally active state [46]. Congruently, the spectra of root-associated fungi reported by Panno and colleagues were dominated by well-known saprobes, including species of Acremonium, Penicillium, etc., which cannot be interpreted as true P. oceanica endophytes. The low effectiveness of serial washing as a surface sterilization technique in seagrass research is further discussed in Newell and Fell [47]. In agreement with our hypothesis, it thus seems that the sterilization technique used in this study eliminated all surface-dwelling fungi and revealed true endorhizal mycobionts of P. oceanica. On the other hand, Shoemaker and Wyllie-Echeverria [48] obtained an assemblage of 36 fungi, mostly dominated by saprobes/parasites probably well-adapted to the marine environment, such as Aspergillus, Cylindrocarpon, Penicilium, and Trichoderma, from rhizomes of three temperate Pacific seagrasses even after surface sterilization with sodium hypochlorite.

At most of the localities, the P. oceanica root cultivable mycobiont communities were dominated by a pleosporalean fungal species hitherto not reported in seagrasses, the Pleosporales sp. MV-2012. This was surprising especially when taking into account the results of Torta et al. [29] who used a very similar isolation protocol and indeed detected the “Lulwoana sp.”, probably conspecific with our Lulworthiales sp. MV-2012, i.e., the second most frequent mycobiont isolated in this study. There are two probable explanations for the lack of the pleosporalean mycobiont within the fungal spectra reported by Torta and colleagues. First, the fungus was simply absent or present in very low frequencies which prevented its detection by the isolation method used; this may be true also for the two of our localities not yielding the pleosporalean mycobiont. Second, the cultivation period utilized by Torta and colleagues, i.e., 4 weeks, was too short for successful isolation of the slow-growing pleosporalean mycobiont which new colonies appeared even after 12 weeks of incubation. This could be resolved by re-visiting the investigated localities and prolonging the cultivation period. Obviously, the pleosporalean mycobiont could be targeted also by culture-independent molecular methods, e.g., amplification of its DNA using specific primers.

The phylogenetic analysis placed the Pleosporales sp. MV-2012 in the vicinity of Aigialus ssp., Ascocratera manglicola, and Rimora mangrovei within Aigialaceae. The recently established Aigialaceae family currently encompasses five genera: Aigialus, Ascocratera, and Rimora represent marine intertidal fungi colonizing mangrove bark and wood in various tropical geographic locations while Fissuroma and Neoastrosphaeriella colonize dead bamboo and palm tissues and are known from Australia, Japan, Philippines, and Thailand [40, 49]. This study thus extends the know range of Aigialaceae habitats for the roots of the Mediterranean endemic seagrass P. oceanica. Whether the second most frequent mycobiont of P. oceanica roots isolated in this study, the Lulworthiales sp. MV-2012, has any closer relationship with the Lulworthia sp. detected in fibrous remains of old P. oceanica leaf sheaths by Cuomo et al. [26] will probably remain unresolved because at present, direct comparison of the respective sequences is impossible. The least frequent species in this study, F. torulosa, is mainly known as a basidiomycetous wood decay saprobe causing white pocket rot in dead and living hardwood trees in Europe [50]. This seems to be the first report on the occurrence of this species in the marine environment. However, due its low isolation frequency, significance of this finding remains unknown.

Our microscopic observations suggested that the pleosporalean mycobiont developed its in vitro colonies from melanized intracellular microsclerotia occurring in the seagrass root hypodermis which were morphologically identical to those reported by Vohník and colleagues [22]. Since the pleosporalean mycobiont also produced thick, septate, a melanized hyphae, it can be ranked among DSE, a miscellaneous group of ubiquitous endophytes colonizing roots of most terrestrial plants [51–53]. Pleosporalean fungi are often reported as DSE in roots of plants from arid and semi-arid regions [54, 55]. On the other hand, although the Pleosporales comprise many freshwater and marine, mostly saprobic, genera [56], to our knowledge, they have not yet been reported as DSE in roots of seagrasses. Intriguingly, Torta et al. [29] claimed that the “Lulwoana sp.” (Lulworthiales) detected by them in P. oceanica roots belonged to DSE despite that no evidence was given that it produced dark septate hyphae or melanized intracellular microsclerotia. Under our cultivation conditions, isolates of the Lulworthiales sp. MV-2012 did not produce menalized hyphae typical for DSE. Torta et al. [29] also provided no evidence that their “Lulwoana sp.” had any connection to the reported intracellular microsclerotia, except for being isolated from some of the collected P. oceanica roots. The identity of microsclerotia reported by Torta and colleagues could be resolved by techniques targeting single cells, i.e., laser capture microdissection, followed by PCR, possibly employing specific primers targeted at the “Lulwoana sp.” DNA.

Based on the data available at present, relatively little can be speculated about the biological significance of the dominant pleosporalean mycobiont. Some DSE can form structures morphologically resembling hyphal interfaces produced by mycorrhizal fungi for communication and nutrient exchange with their host plant [52, 57, 58], and the DSE colonization pattern in P. oceanica roots reported by Vohník et al. [22] comprises dense parenchymatous nets/hyphal sheaths on the root surface which may morphologically resemble hyphal mantles produced by some ectomycorrhizal fungi. On one hand, based on the microscopic observations, it seems plausible that this DSE association is formed by the pleosporalean mycobiont discovered in this study. On the other hand, it remains to be investigated whether this mycobiont engages in nutrient uptake and transport to the host plant in exchange for carbohydrates as reported for other DSE [30]. Other possible functions may include protection of host roots against pathogen attacks [31] or modifications of the distribution of rhizosphere-associated organisms [59, 60] as suggested for other fungi possessing melanized hyphae. Moreover, P. oceanica forms characteristic peat-like sediment (matte) which can be several meters thick and which is exceptionally resistant to decay [1]. Such stocks of organic nutrients are usually inaccessible for plants without the aid of symbiotic bacteria or fungi, and it may be that the pleosporalean mycobiont plays a role during the decomposition of matte.

To conclude, we found that cultivable fungal assemblages colonizing healthy P. oceanica roots across the NW Mediterranean Sea were composed of only three species and were dominated by a single pleosporalean mycobiont. Such a narrow spectrum of root mycobionts across a wide distribution area markedly contrasts with available data on root mycobiont communities from terrestrial ecosystems. The pleosporalean mycobiont has not been previously reported in seagrasses or any other host/ecosystem and probably represents a member of a hitherto undescribed genus in the Aigialaceae family. It apparently forms melanized intracellular microsclerotia in P. oceanica roots and dark septate hyphae in culture and therefore possesses the main characteristics of the ubiquitous terrestrial DSE. This work thus extends the known range of pleosporalean DSE for the marine environment. Our microscopic observations support the view that this fungus forms the recently described DSE association in P. oceanica roots [22]. Regular occurrence of the pleosporalean mycobiont in P. oceanica roots across the NW Mediterranean Sea and its apparent intracellular lifestyle indicate its non-random symbiotic relationship with the dominant Mediterranean seagrass which certainly deserves further investigation.