The diversity and ecological roles of Penicillium in intertidal zones

Members of the genus Penicillium are commonly isolated from various terrestrial and marine environments, and play an important ecological role as a decomposer. To gain insight into the ecological role of Penicillium in intertidal zones, we investigated the Penicillium diversity and community structure using a culture-dependent technique and a culture independent metagenomic approach using ITS (ITS-NGS) and partial β-tubulin (BenA-NGS) as targets. The obtained isolates were tested for halotolerance, enzyme activity, and polycyclic aromatic hydrocarbons (PAHs) degradation. A total of 96 Penicillium species were identified from the investigated intertidal zones. Although the BenA-NGS method was efficient for detecting Penicillium, some species were only detected using conventional isolation and/or the ITS-NGS method. The Penicillium community displayed a significant degree of variation relative to season (summer and winter) and seaside (western and southern coast). Many Penicillium species isolated in this study exhibited cellulase and protease activity, and/or degradation of PAHs. These findings support the important role of Penicillium in the intertidal zone for nutrient recycling and pollutant degradation.


Results
Penicillium diversity from culture-dependent approach. A total of 818 Penicillium strains were isolated from mudflat (626 strains) and sand (192 strains), collected along the western (388 strains) and southern coast (430 strains) of Korea (Table S1). On the basis of morphological comparisons and ITS sequences, 818 strains were categorized in 65 groups. One to three representative strains of each group were identified to the species level based on a section-by-section phylogenetic analysis using BenA, with a total of 187 strains chosen. These strains were identified as 57 species, with eight strains that could not be identified that represent eight potential new species. The species numbers of Penicillium differed depending on season, seaside, and substrate, although it was not statistically significant. The differences of Penicillium according to season, seaside, and substrate are shown in Fig. 1 and Table S2.
Different numbers of Penicillium species were isolated depending on the isolation medium: 49 species were isolated on DRBC, 33 species on GYP, and 23 species on PDA (Table S1). Ten species were isolated on all media. A total of 21, 9, and 5 species were exclusively isolated on DRBC, GYP, and PDA, respectively ( Fig. 2A). Penicillium crustosum was the predominant species, followed by P. oxalicum and P. raperi.
Penicillium diversity from culture-independent approach. For ITS-NGS, a total of 2,363,934 fungal sequence reads were obtained, of which 1,813,137 reads were retained after filtering low quality sequences. Most Good's coverage estimates were 0.955-0.997, indicating a sufficient coverage for diversity analysis. One sample had a score of 0.890. Among the total number of sequences, Penicillium sequences accounted for 1,301 reads, and 21 mOTUs (99% threshold) were assigned to 21 species based on the phylogenetic analysis (Table S2 and Fig. S1). The mOTUs were identified to the species level based on a section-by-section phylogenetic analysis using ITS sequences. A total of 13 species were identified to known Penicillium species, and the remaining eight species could not be confidently identified because of unclear phylogenetic relationships. Penicillium crustosum (27.8%) was the predominant species, followed by P. raperi (18.7%), Penicillium sp. 36 (7.8%), and P. oxalicum (7.4%) (Fig. 2B).
comparison of culture-dependent and -independent approaches. In total, 96 Penicillium species were detected from mudflat and sand using the three different methods (conventional isolation, ITS-NGS, and BenA-NGS) ( Fig. 2B and Table S2). Each method identified a different number of species: 65 species from isolation, 21 species from ITS-NGS, and 76 species from BenA-NGS. A total of 13 species were shared across all three methods. Particularly, P. antarcticum, P. crustosum, and P. raperi were commonly detected from mudflat and sand in all methods. Several species were detected by only one method: isolation (12 species), ITS-NGS (8 species), and BenA-NGS (23 species) (Fig. 2B). Penicillium rubefaciens and P. concentricum were commonly detected in BenA-NGS, but not detected in isolation and ITS-NGS. community structure of Penicillium. Diversity and community analyses were conducted based on the BenA-NGS dataset because it represents Penicillium communities the best, with exact species-level identification and high coverage power for detecting Penicillium species compared to other methods. Alpha diversity indices for Penicillium communities were compared between season (summer and winter), seaside (western and Figure 2. Penicillium diversity detected from intertidal zone. (A) Total, unique, and shared Penicillium species and relative abundance of major species (3%) from three different media (DRBC, GYP, and PDA). (B) Total, unique, and shared Penicillium species and relative abundance of major species (3%) obtained from isolation, NGS using ITS, and NGS using benA. southern coast), and substrate (mudflat and sand) (Fig. 3). Species richness and diversity were significantly different based on season (Richness: P < 0.001; Diversity: P < 0.001), while evenness was comparable (P = 0.702). The winter season had a higher richness and diversity compared to the summer season. However, no index showed a significant difference for seaside (Richness: P = 0.189; Diversity: P = 0.164; Equitability: P = 0.180) and substrate (Richness: P = 0.634; Diversity: P = 0.634; Equitability: P = 0.634).
Penicillium communities were compared between season, seaside, and substrate using CAP analysis (Fig. 3). The community structure of Penicillium was significantly different depending on season (P = 0.001, 16.5% explanatory power) (Fig. 3A). Penicillium brevicompactum, P. concentricum, P. expansum, P. koreense, P. mexicanum, and P. rubefaciens were more abundant in the summer than the winter (Fig. 4). In contrast, P. antarcticum and P. terrigenum were more abundant in the winter. Seaside had a significant effect on the clustering of communities (P = 0.001, 10.3% explanatory power) (Fig. 3B). Among major species, P. antarcticum, P. brevicompactum, P. koreense, P. rubefaciens, and P. terrigenum were significantly more abundant from the southern coast compared to the western coast (Fig. 4). In contrast, P. concentricum, P. expansum, and P. mexicanum were significantly more abundant on the western coast. Substrate did not significantly influence the Penicillium community structures (P = 0.194, 2.0% explanatory power) (Fig. 3C).

Discussion
Penicillium diversity using culture-dependent and -independent approaches. We conducted a comphrensive study to determine the species richness of Penicillium found in the intertidal zone along the Korean coast, testing for impact of season, seaside, and substrate. A total of 96 Pencillium species were identified using three different methods (isolation, ITS-NGS, BenA-NGS). We identified 54 more species than the number of Penicillium species previously recorded in Korea 27 and 58 more species than the number of Penicillium species reported from the marine environment globally 4-6,28-31 . Most of the previous studies of marine Penicillium from intertidal zones focused on screening for industrially useful enzymes as well as novel bioactive compounds. Therefore, these studies primarily identified strains to the genus level only and a handful to the species level (Penicillium chrysogenum, Penicillium citrinum) using ITS sequences 32,33 . Compared to our previous studies 5, 34 , we identified 41 species new to intertidal zones, as well as 17 new species candidates.
Most previous studies used either a culture-dependent method 35,36 or a culture-independent method using a single locus (e.g. ITS) 37,38 . We employed three methods to detect as many Penicillium species as possible and www.nature.com/scientificreports www.nature.com/scientificreports/ to compare the effectiveness of the three methods. Generally, a culture-dependent method isolates only a small subset of the microbial diversity from environments 39 . In our study, a relatively high proportion of Penicillium species (63.1%) were detected using a culture-dependent method. This is likely due to the fact that we focused on Penicillium, which are generally easy to culture; many Penicillium species are isolated from various substrates from terrestrial environments 1,2,40,41 and marine environments [4][5][6][7] . Another reason for the high detection is that we used three different media for isolation. More Penicillium species were isolated from the DRBC medium compared to the other two media. DRBC medium suppresses fast growing fungi, while the other isolation media do not 42 . Some Penicillium species grow slowly and may be covered by fast growing fungi 43 . Several species were only isolated on one agar medium ( Fig. 2A), but there was no apparent tendency for them to prefer certain media. Therefore, more different media have the opportunity to isolate the more diverse Penicillium species.
Although NGS methods help to detect more diverse fungi within the environment, molecular marker selection has a significant impact on NGS methods. The ITS region is less discriminative in identifying species of Penicillium than BenA. However, we use ITS, as well as BenA, since it is widely accepted fungal barcode 22 . In our study, certain species have been detected only in ITS-NGS. Among the 84 species detected in the NGS analysis, 15.5% of the Penicillium species were shared between ITS-NGS and BenA-NGS, while 9.5% were exclusive to ITS-NGS and 75.0% exclusive to BenA-NGS. A total of 20 Penicillium species were not detected in BenA-NGS (8 species from ITS-NGS and 12 species from isolation). The phenomenon of mismatched fungal diversity depending on the survey methods has been seen in previous studies [44][45][46][47] , which can result from the technical bias associated with DNA extraction, PCR amplification, and sequencing 48 . Low abundance species were generally prone to technical bias, which agrees with our results (Table S2). In the case of ITS-NGS, low sequence variation of the ITS2 region can be one of the sources of diversity difference to BenA-NGS; some Penicillium species cannot be assigned to species level based on the ITS2 region. If we consider the potential species based on ITS2 similarity (Table S2 and Fig. S1), five more species can be added to the number of shared mOTUs (Penicillium sp. 33,34,36,38,39). The unique species detected using the isolation method may correspond to the species occurring at low concentration; the culture-dependent method has been shown to detect bacteria at a broader range of concentrations that is 10 2 -10 4 times lower than that of the culture-independent method 49 . The higher number of mOTU detected by BenA-NGS compared to ITS-NGS suggests that it is useful to develop genus-specific primers for protein-coding genes (compared to ITS) for species identification and species richness studies. This is the first study using BenA-NGS for Penicillium diversity, and demonstrates its usefulness for detecting Penicillium in the environment. However, as with using multiple media for culture studies, combining multiple molecular markers will likely identify a greater species richness.
Penicillium community structure depends on environmental factors. We found that Penicillium diversity was significantly higher in winter than summer. Given that the diversity did not differ between substrate and seaside, seasonal variation seems to be an important factor influencing Penicillium diversity. High diversity in winter was also detected in studies of terrestrial fungi 50,51 , which may relate to severe environmental conditions. Various species can co-exist together in harsh environments, and, when conditions are relaxed, a single or small number of species are able to dominate and exclude other species 50 . In the winter season, Penicillium may www.nature.com/scientificreports www.nature.com/scientificreports/ be in spore state or grow slowly, which prevents the dominance of a single species. On the other hand, nutrient deficiency may occur in winter. Since coastal zones are influenced by river flows and oceanic currents, nutrient supply can fluctuate seasonally; nutrient supply from rivers is high in summer and low in winter 52,53 . Oligotrophic condition in winter may lead to co-existence of various Penicillium species.
There were different patterns of the Penicillium community between the western and southern coasts. The difference in Penicillium community between seasides is likely influenced by the difference in oceanic currents in each region. The western coast is influenced by the cold currents of West Korea Coastal Current during winter and summer, whereas the southern seaside is influenced by both the Jeju Warm Current and Taiwan Warm Current during winter and summer 54 . Ocean current patterns influence on the community structure of bacteria, ammonia-oxidizing archaea, and animal biodiversity [55][56][57] . Previous studies showed that temperature influences fungal community in marine environments 58,59 . The ocean currents likely provide temperature and physical barriers limiting Penicillium dispersal and affect Penicillium diversity. Some marine fungi have shown the differential preference on the plant species 60,61 , thus the vegetation composition in the coasts can be another environmental factor that influences on the fungal community structure. In South Korea, the flora of western and southern coasts is different 62,63 , which likely influences the Penicillium communities.
Considering physiochemical differences between mudflat and sand (e.g. texture and nutrient composition), different Penicillium communities were expected. However, the substrate did not significantly influence the diversity and community structure of Penicillium (Fig. 3C). Though environmental filtering by substrates is important factor to differentiate fungal communities [64][65][66][67] , marine fungal community structures usually depend on the biological substrate such as animal, algae, or plant 15,35,68 . Macroalgae, sponge, and animal debris have been known The neighbor-joining phylogenetic tree was based on partial BenA gene sequences. The relative abundance of Penicillium isolated from intertidal zone is color coded on the outer ring (substrate -red; season -blue; seaside -purple). Halotolerance, enzyme activity, and PAH degradation of the Penicillium is color coded on the inner ring (Halotolerance -light blue; gallic acid reaction -orange; protease -brown; endoglucanase -green; β-glucosidase -yellow). (2019) 9:13540 | https://doi.org/10.1038/s41598-019-49966-5 www.nature.com/scientificreports www.nature.com/scientificreports/ to harbor high Penicillium diversity 5,7,69,70 . High enzyme activity associated with alginate, cellulose, and protein detected in Penicillium isolates supports this speculation. ecological roles of Penicillium associated with enzyme activities. Penicillium species from various environments produce a variety of enzymes such as alginase, cellulase, chitinase, and proteasee 5,7,71,72 . Protease activity has been reported in fungi isolated from various marine environments 73,74 . Particularly, members in Penicillium are known for their biotechnological potential in the production of proteases 72 . In our study, a relatively high proportion of Penicillium from the intertidal zone exhibited endoglucanase, β-glucosidase, and protease activity. A high proportion of species having enzyme activities and halotolerance implies that these species play an important role as decomposers of cellulose and protein in the intertidal zone. Macroalgae (over 400 species reported to date) 75 and animal debris such as crab, fish, shells, and sponges are frequently found in the intertidal zone. Penicillium species that have a high enzyme activity can use this abundant algae and animal debris on the intertidal zone as a their favorite nutrient source. Although the species in the Penicillium sections Brevicompacta, Citrina, Canescentia, Fasciculata, and Sclerotiora showed relatively higher enzyme activity than the species in the other sections, there was no distinct pattern of enzyme activity based on a phylogenetic group.
Microbes are known to degrade various pollutants that result from human activities in intertidal zones 76 . Members in the genus Penicillium have been shown to degrade environmental pollutants such as synthetic dyes and polycyclic aromatic hydrocarbons (PAHs) 72 . PAHs are produced by the incomplete combustion of organic matter such as oil and wood 77 . Penicillium canescens, P. chrysogenum, P. frequentans, P. italicum, P. janthinellum, P. montanense, P. simplicissimum, and P. restrictum are known to degrade PAHs 78 , but these species did not show PAHs degrading ability in our study. The PAHs-degrading ability of fungi varies by strain and species. The capability of PAHs degradation was observed in only a few Penicillium species in this study. Particularly, the species (P. decaturense and P. hetheringtonii) in section Citrina showed a relatively higher ability to degrade the PAHs compared to species in other sections.

conclusions
The number of studies investigating marine fungi has increased in recent years and Penicillium are one of the many fungi reported from this environment. In this study, we detected 96 Penicillium species from mudflat and sand from Korea using three different analysis techniques (conventional isolation, ITS-NGS, and BenA-NGS). Although BenA-NGS detected the highest number of Penicillium species, some Penicillium species were only detected by isolation and/or ITS-NGS method. If the goal is to identify the total species diversity, we suggest combining various approaches to detect more species. Many of the isolated Penicillium exhibited cellulase activity, protease activity, and/or degradation of PAHs. Penicillium is a decomposer of a variety of marine organisms and it is thought to play an important role in nutrient recycling and pollutant degradation in intertidal environments. However, it is unclear whether the Penicillia identified in this study are in an active or inactive form in the marine environment. An additional study will be needed to further identify what is the actual function of the Penicillia in the marine environment. We expect that this study will provide basic information regarding the Penicillium diversity and improve interest in the functional role of Penicillia in marine environments.

Materials and Methods
Sample collections. During summer (July and August in 2014) and winter (February 2015), mudflat and sand samples were collected from 30 coastal sites around Korea (10 western coast, 20 southern coast) for each season (Fig. 1). Samples were collected during low tide. To fully characterize each site, each sample contained five subsamples collected approximately 20 m from each other. Each subsample was collected at 2-3 cm depth after removing the top soil to avoid surface contamination. Samples were transported on ice and stored at 4 °C for isolation and at −80 °C for DNA extraction.
DNA extraction, PCR amplification, and sequencing from isolates. Genomic DNA was extracted from isolated Penicillium strains using the modified cetyltrimethylammonium bromide (CTAB) extraction protocol 80 . The PCR amplifications of the ITS and BenA region were performed using primers ITS1F/ITS4 81 and Bt2a/Bt2b 82 , respectively. Each PCR reaction was performed in a C1000 thermal cycler (Bio-Rad, Richmond, CA, USA) following previously described methods 5 . The PCR products were purified using the Expin TM PCR Purification Kit www.nature.com/scientificreports www.nature.com/scientificreports/ Identification of Penicillium isolates. Sequences were assembled, proofread, edited, and aligned using MEGA5 83 . Consensus sequences were deposited in GenBank (Table S1). Molecular identification of species was performed in two steps. First, the sectional position of the strains was determined by comparison of the ITS sequences in a dataset containing sequences of type strains. Next, the BenA sequence data were compared with type strain sequences to identify strains at the species level. Multiple alignments were performed using MAFFT v. 7 84  Bioinformatics process and statistical analysis. The Illumina MiSeq generated ITS (ITS-NGS) and BenA (BenA-NGS) sequence data sets were processed using Quantitative Insights Into Microbial Ecology (QIIME) v. 1.8.0 86 . Sequence pairs were assembled using fastq-join and low quality (QV < 20), short (<200 bp) sequences filtered during the de-multiplexing process. Molecular operational taxonomic units (mOTUs) were clustered with 99% similarity threshold and chimera sequences were filtered using de novo clustering in USEARCH 5.2.236 87 . For each mOTU, the most abundant sequence was chosen as the representative sequence and used for pre-identification using BLAST against type sequences of Penicillium for ITS and BenA. Final identification was conducted based on phylogenetic analysis as described above. Additional chimeric sequences were filtered based on each database using by UCHIME v4.2 88 . Non-fungal, non-Penicillium sequences, and mOTU with less than 10 reads were removed from the analysis. Before the next step, sequences were rarified based on the minimum number of sequences for normalizing dataset. Indices of alpha diversity were calculated for richness (Chao1), diversity and evenness (Shannon's index) in QIIME.
Statistical tests and graphical plotting were conducted using ggplot2 89 , phyloseq 90 , and basic packages in R 91 . Alpha diversities were compared between categories (season, seaside, and substrate) using Wilcox rank sum test adjusted by the false discovery rate (FDR) 92 . Community structures were compared by Constrained Analysis of Principal coordinates (CAP) based on Bray-Curtis dissimilarities. CAP models were constrained by each category with conditioning by the other factors (e.g. ~Seaside + Condition [Season + Source]), and significance was tested by ANOVA-like tests with 999 permutations.
Halotolerance, enzyme activity, and pAH degradation. Halotolerance was determined by measuring the colony diameter of a representative strain of each species on MEA supplemented with ASW instead of distilled water. Each species was inoculated in three-point fashion on MEA with and without ASW using spore suspensions. The plates were incubated at 25 °C and colony diameter was measured after 5 days. Growth difference between MEA with and without ASW was compared.
Endoglucanase, β-glucosidase, and protease activity were assessed for a representative strain of each species using a modified plate screening method 69 . Endoglucanase and β-glucosidase activity were assayed on Mandels' medium 71 , with 1% carboxymethylcellulose (Sigma-Aldrich, MO, USA) and 0.5% D-cellobiose (Sigma-Aldrich, MO, USA) as the primary carbon source, respectively. Protease activity was assayed by growing the fungi for five days on yeast extract agar (Oxoid, MD, USA) supplemented with 1.5% skim milk (Difco-Becton, MD, USA) as the primary carbon source 93 . To indicate the enzyme activity, the diameters (mm) of clear zones surrounding colonies were measured.
A gallic acid reaction was used as a screening method to determine the capability of degradation of PAHs 94 . The gallic acid reaction was assayed by growing the fungi on 1.5% MEA (Difco-Becton, MD, USA) supplemented with 5 g L −1 of gallic acid. Each plate was incubated at 25 °C for 14 days. The capability of degradation of PAHs was identified by a color change surrounding the colony. Each value for a strain is the average result from three experiments.