Cytotoxicity and mycotoxin production of shellfish-derived Penicillium spp., a risk for shellfish consumers



In order to assess the putative toxigenic risk associated with the presence of fungal strains in shellfish-farming areas, Penicillium strains were isolated from bivalve molluscs and from the surrounding environment, and the influence of the sample origin on the cytotoxicity of the extracts was evaluated. Extracts obtained from shellfish-derived Penicillia exhibited higher cytotoxicity than the others. Ten of these strains were grown on various media including a medium based on mussel extract (Mytilus edulis), mussel flesh-based medium (MES), to study the influence of the mussel flesh on the production of cytotoxic compounds. The MES host-derived medium was created substituting the yeast extract of YES medium by an aqueous extract of mussel tissues, with other constituent identical to YES medium. When shellfish-derived strains of fungi were grown on MES medium, extracts were found to be more cytotoxic than on the YES medium for some of the strains. HPLC-UV/DAD-MS/MS dereplication of extracts from Penicillium marinum and P. restrictum strains grown on MES medium showed the enhancement of the production of some cytotoxic compounds. The mycotoxin patulin was detected in some P. antarcticum extracts, and its presence seemed to be related to their cytotoxicity. Thus, the enhancement of the toxicity of extracts obtained from shellfish-derived Penicillium strains grown on a host-derived medium, and the production of metabolites such as patulin suggests that a survey of mycotoxins in edible shellfish should be considered.

Significance and Impact of the Study

Penicillium strains isolated from bivalve molluscs produce extracts exhibiting a higher cytotoxicity than extracts from Penicillium strains isolated from the surrounding marine environment. The use of a mussel-based medium for cultures of some shellfish-derived strains enhances the cytotoxicity of extracts when compared with classical media. The production of cytotoxic compounds and of the mycotoxin patulin on such a host-derived medium highlights a potential health risk for shellfish consumers.


In marine environments such as hypersaline waters, sediments, beaches sands or shellfish-farming areas, Penicillium is one of the predominant genera (Khudyakova et al. 2000; Sallenave-Namont et al. 2000; Mancini et al. 2005; Butinar et al. 2011). Among the wide diversity of Penicillium species, numerous marine-derived Penicillium strains have been described for the production of toxic metabolites (Bugni and Ireland 2004; Blunt et al. 2009, 2011; Ebel 2010), and some are known to produce mycotoxins such as ochratoxin A, penitrem A, roquefortine C or patulin (Frisvad et al. 2004; Paterson et al. 2004). Marine-derived fungi form a group of potential contaminants for marine bivalves such as mussels (Zvereva and Vysotskaya 2005; Zielinski et al. 2009). In previous studies, it has been shown that Penicillium strains sampled from shellfish-farming areas produced roquefortine C and patulin (Vansteelandt et al. 2012) and that some toxic fungal metabolites like gliotoxin or peptaibols produced by marine-derived fungi can be detected in marine sediments and can be accumulated in filter-feeding molluscs under laboratory conditions (Sallenave et al. 1999; Grovel et al. 2003). In situ contamination of bivalves by fungal compounds could be one of the origins of the unexplained toxicities of seafood observed along the French coast since the beginning of the 1990s (Belin et al. 2009).

In this study, the cytotoxicity of shellfish-derived Penicillium strains in comparison with environment-derived ones was investigated. Subsequently, a host-derived medium was used to evaluate the influence of mussel flesh on the production of secondary metabolites and mycotoxins by shellfish-derived Penicillia.

Results and discussion

Influence of strain origin on cytotoxicity of culture extracts

Cytotoxicity of culture extracts from 24 Penicillium strains isolated from shellfish-farming areas and grown on malt extract agar medium (MEA) medium was evaluated on KB cells. Activities obtained for isolates from shellfish were compared with those from their immediate environment (Fig. 1). The number of inactive extracts from the two groups was similar. However, amongst toxic extracts, high cytotoxicities were found to be predominant in the extracts obtained from shellfish-derived isolates, with 45% of the extracts exhibiting an IC50 < 10 μg ml−1, whereas only 10% of extracts from environment-derived strains exhibited such high cytotoxicity. As samples of shellfish, sediment and seawater were gathered in the same place at the same moment, it seems that the origin of the samples has an influence on the cytotoxicity profiles of strains. Some Penicillium strains isolated from shellfish have been previously reported to be more toxic on an Artemia salina bioassay than others isolated from the marine environment (Sallenave-Namont et al. 2000; Matallah-Boutiba et al. 2012). This influence of the sample origin on the activity has also been reported for extracts of fungi isolated from different species of sponges (Holler et al. 2000; Thirunavukkarasu et al. 2011) or type of sediment (Khudyakova et al. 2000). Thus, the observation of a highest number of cytotoxic extracts from shellfish-derived isolates could suggest the operation of some mechanism of selection from either of the organisms, leading to a kind of association.

Figure 1.

Comparison of crude extracts cytotoxicity for (a) shellfish-derived (N = 14) and b) environment-derived (N = 10) Penicillium spp. strains cultured on MEA medium. Dark grey for IC50 < 10 μg ml−1; medium grey for 10 < IC50 < 30 μg ml−1; light grey for IC50 > 30 mg ml−1.

Associations between fungi and marine macro-organisms have been mainly studied through the investigation of the biodiversity or chemodiversity of sponge-derived fungi. Sponges are often associated with micro-organisms that have a profound impact on host biology (Wilkinson 1983). Some sponge-derived fungi may be opportunistic and contribute to localized lesions and diseases (Li and Wang 2009). These fungi may also contribute to host defence via the production of biologically active metabolites (Holler et al. 2000; Han et al. 2009; Thirunavukkarasu et al. 2011). However, limited data are available on the ecological function of fungal communities living within these invertebrates. Associations of fungi with bivalves have rarely been studied: some marine fungi have been reported to exhibit a pathogenic character towards different species of molluscs (mussels, oysters, abalones) by infecting the shell or the flesh (Alderman and Jones 1971; Grindley et al. 1998; Van Dover et al. 2007), but none of the other kind of associations such as commensalism or mutualism have been described. It could be envisaged that the different types of relationship described between sponges and fungi could be transposed to bivalve molluscs.

Influence of shellfish constituents on the cytotoxicity of shellfish-derived Penicillia

Ten fungal strains isolated from bivalves were grown on a mussel flesh-based medium (MES) obtained by substituting the yeast extract of the usual YES medium by a mussel flesh aqueous extract, in order to assess the influence of mussel constituents on the production of cytotoxic metabolites. Table 1 summarizes the cytotoxicity of culture extracts obtained for the MES medium as well as five other common media.

Table 1. Cytotoxicity of extracts of Penicillium spp. strains isolated from bivalves and grown on the MES medium and the five classical media (IC50, μg ml−1)
StrainSpeciesCulture medium
MMS50 P. venetum >30.0>30.012.5>30.0>30.025.0
MMS163 P. antarcticum 5.016.9>30.0>30.08.8>30.0
MMS231 P. polonicum 21.6>30.06.324.1>30.020.9
MMS266 P. marinum 15.7>30.0>30.020.515.026.6
MMS270 P. bialowiezense >30.0>30.025.0>30.0>30.0>30.0
MMS330 P. ubiquetum 26.911.119.116.813.918.7
MMS393P. sp.>30.0>30.0>30.0>30.0>30.0>30.0
MMS399 P. ligerum 14.425.716.521.4>30.0>30.0
MMS404 P. brevicompactum >30.0>30.0>30.0>30.0>30.0>30.0
MMS417 P. restrictum 4.9>30.06.3>30.012.5>30.0

All ten strains selected exhibited a similar growth on MES and YES media, showing that mussel constituents are also favourable to the development of fungi. In this way, fermentation studies have shown that mussel processing wastewaters can be depurated by fungi and yeasts, and such mussel-based media have proven to be suitable for the industrial production of amylase by several fungal species (Gonzalez et al. 1992; Torrado et al. 2013). Among the 60 extracts, 33 exhibited no activity, 22 a medium cytotoxicity and five were highly cytotoxic. IC50 obtained for MES and CYA media appeared to have the similar distribution, that is, four IC50 were higher than 30 μg ml−1, four were between 10 and 30 μg ml−1 and two were lower than 10 μg ml−1, implying that these two media seemed to be equally favourable to induce the production of active metabolites by fungi although their composition were highly different. The CYA medium has been reported to be one of the most strongly inducing production of the widest range of secondary metabolites (Frisvad et al. 2004). This study showed that the MES medium seems to induce in a similar manner the production of a large number of extrolites some of them being cytotoxic. On the contrary, the YES medium was the less favourable for the production of cytotoxic compounds, with only three extracts exhibiting an activity. For five strains (P. antarcticum, P. polonicum, P. marinum, P. restrictum and P. ligerum), the MES extracts were more cytotoxic than the corresponding YES extracts showing that shellfish-derived fungi grown on a host-derived medium could express a specific metabolome and could enhance the production of cytotoxic compounds. This influence of host-derived media on the specific expression of metabolites has been previously demonstrated with the stimulation of the production of corymbiferan lactones by necrotrophic fungi (Overy et al. 2006).

Cytotoxic metabolites responsible for cytotoxicities of MES extracts

In order to investigate the compounds responsible for the enhanced activities on MES media, metabolic profiles of strains were analysed using HPLC-UV/DAD-HRMS/MS, and the composition of extracts obtained on the MES and YES media were compared. The dereplication of extracts was performed on the basis of the accurate mass, UV spectrum and MS/MS spectrum of the main chromatographic peaks. The results obtained for three strains presenting either a high cytotoxicity on MES medium or a high differential activity between the MES and YES extracts are shown. Table 2 lists peaks that could be annotated.

Table 2. Annotated peaks observed on YES and MES chromatograms for MMS266 and MMS417
PeakExperimental m/z [M+H]+Molecular formulaTheoretical m/z [M+H]+ppm diffIdentification parametersAnnotation
  1. nf, data not fully congruent with literature; ni, no information on MS/MS or UV spectra in literature.

Strain MMS266
A, B345.2418C22H32O3345.24241.74XXXPenostatin derivatives
C327.23C22H30O2327.23195.81XXXPenostatin derivatives
E327.2283C22H30O2327.23198.86XXnfPenostatin derivatives
Strain MMS417
B231.1214C11H18O5231.12275.62XXniHydroxy pestalotin or Pestalrone A
C215.1262C11H18O4215.12787.44XXnfPestalotin or Pestalrone B

MES and YES extracts of MMS266 (P. marinum) exhibited different chromatographic profiles (Fig. 2a,b). Surprisingly, none of the cytotoxic indole alkaloids communesins or chaetoglobosins, some characteristic metabolites of this species, were found to be produced on these two media. The most abundant peaks observed in the YES extract were not observed in the MES extract, whereas several peaks were specific of the latter. Four of them (peaks A, B, C and E) were identified as pentostatin derivatives, a group of compounds initially described from marine Penicillium strains and known to exhibit significant cytotoxicities on various cell lines (Takahashi et al. 1996; Iwamoto et al. 1999). The fifth identified compound, also specifically produced on MES medium, was fusoxysporone, a viscidane-type diterpene first isolated from a Fusarium oxysporum (Abraham and Hanssen 1992). No biological activities have been described for this compound. Thus, further investigations on toxic effects of both pentostatins and fusoxysporone would be required to assess their potential involvement in the cytotoxicity of MES extracts and their putative human health effects if they were produced in situ in mussels.

Figure 2.

ESI positive mode base peak chromatograms (BPC) of crude extracts of (a) MMS266 on YES medium, (b) MMS266 on MES medium, (c) MMS417 on YES medium and (d) MMS417 on MES medium. Peaks annotated YES = YES medium component and MES = MES component.

For the MMS417 strain (P. restrictum), YES and MES extracts exhibited chromatographic profiles of greater similarity (Fig. 2c,d). The main difference was the relative abundance of peak C that represented 4.8% of the total peak areas on YES and 14.1% on MES. This compound could be annotated as pestalotin, a metabolite previously isolated from a Penicillium sp. strain, or less likely as pestalrone B, a Pestalotiopsis karstenii molecule (Kimura et al. 1986; Luo et al. 2012). All the other annotated compounds were also pyran-2-one derivatives (Table 2). As for peak C, peak B was thought to be hydroxypestalotin rather than pestalrone B, as it has once been reported in a Penicillium sp. strain (Kirihata et al. 1992). Pestalotin derivatives have been described to exhibit only weak cytotoxicity against the human tumour cell line U-251 (Luo et al. 2012). Thus, the high activity of the MES extract could also be attributed to one of the minor peaks that remained unidentified.

Extracts of the strain MMS163 (P. antarcticum) were highly cytotoxic when it was grown on MES and MEA media, whereas no activity was observed for those obtained from cultures on CYA, potato dextrose agar medium (PDA) and dextrose casein agar medium (DCA) media. As marine isolates of this species are known to produce patulin (Vansteelandt et al. 2012), a mycotoxin cytotoxic on various cell lines (Iwamoto et al. 1999; Heussner et al. 2006), the six extracts were dereplicated with a focused attention to patulin detection (Table 3).

Table 3. Annotated peaks observed on extracts of MMS163
MediumIC50 (μg ml−1)Annotated compounds
PatulinChrysogineCladosporin5′-hydroxy-asperentinTerrestric acid

Cladosporin and 5′-hydroxy-asperentin were identified in six and five extracts, respectively. Cladosporin is a secondary metabolite initially isolated from Cladosporium cladosporioides (Scott and Van 1971). This compound has been shown to exhibit antifungal, antimicrobial and antimalarial activities (Scott and Van 1971; Anke et al. 1978; Hoepfner et al. 2012). To our knowledge, bioactivities of the cladosporin derivative 5′-hydroxy-asperentin have not yet been reported. Patulin was detected in the three cytotoxic extracts and not in the others, contrary to the alkaloid chrysogine that was only present in the nonactive ones. This indicates that the cytotoxicity of the extracts is strongly related to the presence of patulin and that there is a biosynthetical shift in the routes leading to these two compounds. Several investigations have shown that the biosynthesis of patulin is highly regulated and dependent of abiotic factors such as medium components. Glucose is a favourable carbon source, whereas nitrogenous compounds inhibit its production (Grootwassink and Gaucher 1980; Rollins and Gaucher 1994), and manganese is required for the expression of the isoepoxydon dehydrogenase gene implicated in the synthesis of patulin intermediates (Scott et al. 1986; Puel et al. 2010). In the present study, no correlation could be observed between any medium component and patulin or chrysogine production. Due to its cytotoxic, mutagenic, genotoxic, immunotoxic, neurotoxic and teratogenic effects (Andersen et al. 2004), patulin is considered as a major mycotoxin and its presence in food is regulated (FAO 2004). Thus, the stimulation of patulin production by mussel flesh constituents could lead to its accumulation in edible shellfish and then induce a putative health risk for their consumers.

In summary, this study shows that the most toxigenic Penicillium strains were found in association with natural samples corresponding to bivalve molluscs as opposed to sediment and seawater. As advised by Overy et al. (2006), who proposed to grow necrotrophic fungi on host-derived media to enhance the production of specific metabolites, the use of a mussel-based medium for the cultivation of shellfish-derived strains led to the observation of more cytotoxic compounds for some strains. Thus, the production of such toxic compounds may be generally enhanced by shellfish constituents. Further investigations must focus in priority on the isolation and identification of these compounds to assess the potential risk for human shellfish consumers. Moreover, the identification of the fungal toxin patulin in a Penicillium strain isolated from shellfish strengthens the interest of monitoring Penicillium mycotoxins in shellfish-farming areas.

Materials and methods

Fungal strains

Twenty-four Penicillium strains were randomly selected from the marine-derived fungal strain collection of the laboratory. These strains had been isolated from cultured shellfish (cockles and mussels) and their immediate environment (sediment and seawater) gathered along the western coast of France. Strains were identified using their phenotypic characteristics, by metabolite profiling or by sequencing the internal transcribed spacers (ITS-1 and -2) and beta-tubulin regions of their genomic DNA. Sequences were compared with CBS and GenBank databanks. Three strains remained unidentified as they did not grow during the identification procedure. The list of all strains is given in Supporting Information (Table S1). For the study of shellfish-derived Penicillia, 10 strains, isolated from shellfish and all belonging to different species, were kept from the previous selection.

Culture and extraction

Culture media: Strains were grown on six different culture media all prepared with 15 g l−1 of agar in sterilized natural seawater (salinity of 32·8 psu) containing additionally: 5 g Czapek extract, 5 g yeast extract and 30 g saccharose for Czapek yeast extract agar medium (CYA); 40 g glucose and 10 g casein enzymatic digest for DCA; 5 mg CuSO4, 10 mg ZnSO4, 1 g peptone, 20 g malt extract and 20 g glucose for MEA (medium used for the cytotoxicity screening of the 24 strains); 5 mg CuSO4, 10 mg ZnSO4, 4 g potato extract and 20 g glucose for PDA; 5 mg CuSO4, 10 mg ZnSO4, 0.5 g MgSO4, 20 g yeast extract and 150 g saccharose for yeast extract sucrose medium (YES). Mussel extract saccharose (MES) medium was prepared by incorporating 20 g of a freshly prepared mussel extract instead of the 20 g of yeast extract of the YES medium. Mussel extract was obtained by collecting total flesh of blue mussels (Mytilus edulis) from Normandy, France. The flesh was crushed and centrifuged at 6000 g, 5°C during 30 min. Supernatant was successively passed through filters of decreasing porosity (50, 5 and 1 μm; Sartorius, Göttigen, Germany) and finally lyophilized. The extract obtained was ground in order to obtain a homogeneous powder. Bioassay performed on noninoculated MES medium extracts showed no cytotoxicity of the mussel flesh constituents.

Fungal cultures and extract preparation

Each culture was carried out in triplicates in Erlenmeyer flasks containing 50 ml of solid medium; cultures were incubated at 27°C for 12 days under natural light. Fungal mycelium was extracted together with the agar layer, in order to obtain both internal and excreted compounds. The cultures were extracted twice with 100 ml of dichloromethane/ethyl-acetate 1 : 1 (v/v). Mixtures were then sonicated during 30 min and macerated overnight. After filtration on Büchner funnel, the organic phase was dehydrated on Na2SO4. Spores were removed by filtration on regenerated cellulose 0·45-μm filters (Sartorius). Organic phases were evaporated to dryness leading to crude extracts.

Cytotoxicity assay

Cytotoxicity was evaluated on KB cells, a human epidermoid carcinoma cell line, with a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) coloration after 72 h of incubation according to the protocol of Ruiz et al. (2007). Results were expressed as the concentration that inhibited 50% of cell growth (IC50). Experiments were carried out in duplicate. Culture medium containing MeOH 5% was used as negative control and penicillic acid was used as positive control (IC50 = 18.8 μg ml−1). Cytotoxicity was considered as high if IC50 was under 10 μg ml−1 and low if IC50 was between 10 and 30 μg ml−1. Extracts with IC50 higher than 30 μg ml−1 were considered as inactive.


HPLC-DAD/ESI-IT-TOF-MSn analyses were performed with a Shimadzu LCMS-IT-TOF instrument composed of two LC-20ADxr pumps, a SIL-20ACxr autosampler, a CTO-20AC column oven, an SPD-M20A PDA detector and a CBM-20A system controller, coupled to a mass spectrometer with an ESI ion source, and an hybrid Ion Trap-Time-Of-Flight mass analyser (Shimadzu, Kyoto, Japan). HPLC and EI-IT-MS analysis conditions are given as supporting informations. Dereplication of extracts was performed according to the method described by Vansteelandt et al. (2012). The data obtained were analysed using the fungal metabolite database of Nielsen and Smedsgaard (2003) and literature.


We acknowledge the Region des Pays de la Loire for funding of this study through the COLNACOQ and CHIMIMAR programs, and Yolaine Joubert and Solène Brochard for her help in identification of fungal strains, cultures and extractions.

Conflict of Interest

Authors have no conflict of interests to declare.