Wide distribution of resistance to the fungicides fludioxonil and iprodione in Penicillium species

Fludioxonil and iprodione are effective fungicides widely used for crop protection and are essential for controlling plant pathogenic fungi. The emergence of fungicide-resistant strains of targeted pathogens is regularly monitored, and several cases have been reported. Non-targeted fungi may also be exposed to the fungicide residues in agricultural fields. However, there are no comprehensive reports on fungicide-resistant strains of non-targeted fungi. Here, we surveyed 99 strains, representing 12 Penicillium species, that were isolated from a variety of environments, including foods, dead bodies, and clinical samples. Among the Penicillium strains, including non-pathogenic P. chrysogenum and P. camembertii, as well as postharvest pathogens P. expansum and P. digitatum, 14 and 20 showed resistance to fludioxonil and iprodione, respectively, and 6 showed multi-drug resistance to the fungicides. Sequence analyses revealed that some strains of P. chrysogenum and Penicillium oxalicum had mutations in NikA, a group III histidine kinase of the high-osmolarity glycerol pathway, which is the mode of action for fludioxonil and iprodione. The single nucleotide polymorphisms of G693D and T1318P in P. chrysogenum and T960S in P. oxalicum were only present in the fludioxonil- or iprodione-resistant strains. These strains also exhibited resistance to pyrrolnitrin, which is the lead compound in fludioxonil and is naturally produced by some Pseudomonas species. This study demonstrated that non-targeted Penicillium strains distributed throughout the environment possess fungicide resistance.


Introduction
Fludioxonil is a member of the phenylpyrrole class of fungicides that acts on a broad spectrum of plant pathogenic fungi [1]. It is a derivative of pyrrolnitrin, a secondary metabolite produced by certain bacteria, including Pseudomonas species [2] ここをクリックまたはタップ してテキストを入力してください。. In many countries, fludioxonil is widely used for crop, as well as post-harvest, protection of pom fruits from fungal pathogens. Fludioxonil's mode of action consists of a fungal two-component system in the high-osmolarity glycerol iprodione were obtained from the Abe laboratory at Tohoku University, Sendai, Miyagi, Japan. Oligonucleotides were synthesized by Eurofins Genomics (Tokyo, Japan).

Antifungal susceptibility assay
Sensitivity to fungicides was determined by measuring colony growth on PDA plates in the presence of fungicides. Approximately 10,000 spores of each Penicillium strain were inoculated onto both PDA and PDA supplemented with 1 μg/mL fludioxonil, 10 μg/mL iprodione, or 0.05 μg/mL pyrrolnitrin and incubated at 25˚C for 5 days. The diameter of each fungal colony on PDA amended with fungicides was measured and compared with that on PDA alone. Strains having a growth rate of 50% or more were defined as "fungicide-resistant", whereas those having a growth rate of less than 50% were defined as "fungicide-sensitive".

Extraction of genomic DNA
Mycelia cultured in potato dextrose broth were frozen in liquid nitrogen and ground to a fine powder using a mortar and pestle. Total genomic DNA was extracted using a NucleoSpin Plant II Kit (Takara Bio, Ohtsu, Japan).

DNA sequencing
The genes encoding NikA (500 bp-upstream of the open reading frame to 500 bp-downstream of the open reading frame) in Penicillium chrysogenum and Penicillium oxalicum were amplified by PCR using genomic DNA as the template and specific primers (S1 Table). The PCR conditions were as follows: 30 cycles of 98˚C for 10 s, 53˚C for 5 s, and 68˚C for 1 min with KOD One PCR Master Mix (Toyobo, Osaka, Japan). The PCR product was subject to agarose gel electrophoresis and purified using a Gel/PCR Extraction Kit (NIPPON Genetics, Tokyo, Japan). The purified PCR products were subjected to DNA sequencing (Eurofins Genomics). The sequences were compared with those of the nikA genes from the reference genomes of P. chrysogenum P2niaD18 (GCA_000710275) and P. oxalicum 114-2 (GCA_000346795), which were retrieved from the National Center for Biotechnology Information database (https://www.ncbi.nlm.nih.gov/).

Genome sequencing
Whole-genome sequencing using next-generation methods was performed as described previously [26]. Briefly, we prepared a fragmented DNA library from the genomic DNA of P. roqueforti using NEBNext Ultra II FS DNA Library Prep Kit for Illumina (New England BioLabs) and NEBNext Multiplex Oligos for Illumina (New England BioLabs). Paired-end sequencing was carried out by Novogene.

Single nucleotide variant detection
To search for single nucleotide polymorphisms in nikA of P. roqueforti, we performed read mapping using CLC Genomics Workbench (CLC bio, Aarhus, Denmark). The reads from each isolate were trimmed and mapped to the nikA (PROQFM164_S03g000214) of P. roqueforti FM164 (GCA_000513255).

Mutations in the group III HHK, NikA
As demonstrated in several fungi, including plant pathogens, fludioxonil/iprodione-resistance might be attributed to mutations in the group III HHK, NikA, in Penicillium strains. In further analyses, we focused on P. chrysogenum, P. oxalicum, and P. roqueforti because they contained multiple multi-drug resistant strains. The nikA genes in a set of P. chrysogenum, P. oxalicum strains were sequenced and compared with those of P. chrysogenum strain P2niaD18 (GCA_000710275), P. oxalicum 114-2 (GCA_000346795), respectively. The nikA genes in a set of P. roqueforti strains were obtained by genome sequencing. There were several amino acid alterations in NikA (Fig 2), such as the combination of A404S and S1332L in P. chrysogenum IFM 56829. The S1332L mutation was also present in P. chrysogenum IFM 61615. Both strains were sensitive to fludioxonil and iprodione, suggesting that the mutations are not involved in the resistance. Conversely, in P. chrysogenum IFM 57243, which showed resistance to fludioxonil and iprodione, glycine was changed to aspartic acid at position 693 (G693D) and threonine was changed to proline at position 1318 (T1318P). The mutation at position 693 is located in the HAMP domain region, suggesting that this mutation affects the sensing of, and interactions with, the fungicides (Fig 3). For P. oxalicum, 10 of 15 strains harbored both S94F and Q151R mutations regardless of their fungicide-resistance level, indicating that these mutations are not associated with fungicide resistance. In P. oxalicum IFM 54751, which is resistant to fludioxonil but not to iprodione, threonine was changed to serine at position 960 (T960S) in addition to the abovementioned two mutations. This mutation is located in the kinase domain and potentially affects the histidine kinase function and fludioxonil resistance. In P. roqueforti, no amino acid alterations were found in nikA compared with the reference sequence.

Resistance of Penicillium strains to high osmotic stress
As demonstrated in other species, the HOG pathway is involved in responses to fungicides and osmotic conditions. To test the link between resistance to fungicides and osmotic stress, we investigated the colony growth of P. chrysogenum and P. oxalicum strains on PDA containing high concentrations of KCl or sorbitol (1.5 M) (Fig 2). Among the 18 tested strains of P. chrysogenum, IFM 52203 showed a sensitivity to KCl (growth was less than 60% compared with under stress-free conditions). The growth rates of P. chrysogenum IFM 57243, which had G693D and T1318P mutations in NikA, in the presence of high concentrations of KCl and sorbitol were 92% and 124%, respectively. These values were comparable to those of other strains, indicating that the G693D and T1318P mutations had no effect on the strains' sensitivity to osmotic stress. Compared with P. chrysogenum, P. oxalicum strains were relatively sensitive to KCl. Growth was less than 30% in the presence of KCl, compared with no KCl stress, in 3 of 15 strains, whereas only 1 strain showed >20% reduction in the colony growth in the presence of sorbitol. The growth rates of P. oxalicum IFM 54751, with of the T960S mutation in NikA, in high concentrations of KCl and sorbitol were 32% and 86%, respectively. The moderate sensitivity to high osmotic stresses suggests the involvement of the T960S mutation in osmotic stress adaptation.

Pyrrolnitrin-resistant Penicillium
Fludioxonil is an analog of the natural antifungal compound pyrrolnitrin, which is produced by some Pseudomonas species [27]. The mode of action for pyrrolnitrin is believed to be related to the fungal group III HKKs of the HOG pathway [28]. Therefore, we performed     fludioxonil and/or iprodione (Fig 4A). For P. chrysogenum, five and two of seven pyrrolnitrinresistant strains showed multi-drug resistance to fludioxonil and iprodione, respectively ( Fig  4B). For P. oxalicum, six and five of nine pyrrolnitrin-resistant strains showed multi-drug resistance to fludioxonil and iprodione, respectively (Fig 4C). Penicillium chrysogenum IFM 57243 with G693D and T1318P mutations in NikA showed resistance to pyrrolnitrin, whereas P. chrysogenum IFM 56829 and IFM 61615 were sensitive to pyrrolnitrin. Thus, the G693D and T1318P mutations may contribute to resistance against the three fungicides. Penicillium oxalicum IFM 54751 with the T960S mutation exhibited resistance to fludioxonil and pyrrolnitrin. The results of the antifungal susceptibility testing revealed that within a set of Penicillium strains isolated from various environments that are non-targeted fungi some members show resistance to an agricultural fungicide, as well as its lead compound, which is naturally produced by bacteria in the environment.

Discussion
In this study, we explored Penicillium strains resistant to the widely used fungicides fludioxonil and iprodione, as well as pyrrolnitrin. Strains of Penicillium species that cause postharvest decay of citrus and fruits, and exhibiting resistance to these fungicides, have been investigated [24,25]. Interestingly, in the present study there were no resistant strains found in P. expansum and P. italicum, where one P. digitatum strain isolated from lemon showed resistance to iprodione. To the best of our knowledge, except for these pathogens, we are the first to demonstrate that several strains of non-targeted fungi, such as penicillin-producing P. chrysogenum, environmentally ubiquitous P. oxalicum, and cheese-producing P. roqueforti, show resistance to fludioxonil, iprodione, and pyrrolnitrin. A multi-drug resistance to these fungicides was detected in some strains. These results raised questions regarding the mechanisms and occurrence of resistance in these species.
The mechanisms underlying resistance to fludioxonil have been studied in several fungi, including plant pathogens such as B. cinerea, Cochliobolus heterostrophus, and M. oryzae [29][30][31]. The fungal HOG pathway involved in osmotic stress adaptation is a target of these fungicides [10,23,32]. Laboratory-generated mutants resistant to these fungicides provide clear perspectives on the mechanisms. Fillinger et al. showed that exposure to pyrrolnitrin or iprodione results in resistant B. cinerea mutants, most of which harbor de novo mutations in the Bos1 protein, a group III HHK of the fungus [33]. The mutations occurred in the protein's six repeated HAMP domains. The regeneration of site-directed clones clarified that the amino acid alterations in the HAMP domains are responsible for the fungicide resistance. Some mutations resulted in fungicide resistance and hypersensitivity to osmotic stress, whereas other mutations resulted in resistance to iprodione, but not to phenylpyrroles, and sensitivity to hyperosmolarity. Here, some Penicillium strains possessed mutations in NikA that were associated with fungicide resistance. Penicillium chrysogenum IFM 57243 showed multi-drug resistance to fludioxonil and iprodione, as well as pyrrolnitrin, and had G693D and T1318P mutations in the NikA protein. Penicillium oxalicum IFM 54751 showed resistance to fludioxonil and pyrrolnitrin, but not to iprodione, and had a T960S mutation in the NikA protein.
These mutations are located in the amino acid residues highly conserved among the 12 species (Fig 5). These mutations may contribute to antifungal compound resistance, while sitedirected clones of the mutations need to be created in the future.
A fitness penalty has been reported in fludioxonil-resistant isolates of plant pathogens, as indicated by their relatively slower mycelial growth rates or decreased pathogenicity levels [15,34,35] might explain the decreased fitness levels of fludioxonil-resistant strains in the field. Interestingly, the fungicide-resistant Penicillium strains found in the present work showed no apparent growth defects on PDA compared with sensitive isolates of the same species (S1-S3 Figs). This suggested that Penicillium species pay almost no fitness costs for phenylpyrrole and dicarboximide resistance, which should be investigated further.
Here, several strains of P. chrysogenum and P. oxalicum without mutations in the NikA showed resistance to one of the antifungal compounds. In N. crassa, the components of the HOG pathway, which function downstream of the group III HKK Os1, are responsible for fungicide resistance [21,36]. Indeed, a strain with a mutation in the os2 gene, which encodes a mitogen-activated protein (MAP) kinase in the HOG pathway, shows resistance to the fungicides. However, a mutant of the SakA MAP kinase, which is an ortholog of Os2, in A. nidulans shows only slight resistance to fludioxonil and iprodione [23]. The HOG pathway contributes to fungicide responses in different ways among fungal species. To date, only one study has investigated the HOG pathway's role in the fungicide responses of Penicillium. Wang et al. demonstrated that the gene deletion mutant of Pdos2, which encodes an Os2 MAP kinase, constructed in P. digitatum shows only slight resistance to fludioxonil and iprodione, suggesting that its HOG pathway has a limited impact on fungicide sensitivity [37]. Thus, it is unlikely that fungicide sensitivity can be attributed to mutations in HOG pathway components, because the fungicide resistance levels identified here were relatively high. The resistance mechanisms of fludioxonil and iprodione are poorly understood, and thus uncharacterized mutations may be present in the resistant strains. More comprehensive investigations are needed to fully understand how the non-targeted fungi possess resistance to synthetic fungicides.
The Penicillium strains used in this study were collected from diverse environmental and clinical sources. According to their records, they have no history of phenylpyrrole or dicarboximide fungicide exposure. The growth test indicated that each species was naturally susceptible, and some strains acquired resistance, to the fungicides. However, it is not known where and how these Penicillium strains have become resistant. One plausible cause for resistance is exposure to fungicide residues in environmental organic matters, such as plant litter or compost. Fludioxonil and iprodione are registered as fungicides for use on a wide variety of crops, and thus, huge amounts of plant debris contaminated with residual fungicides are generated in agricultural settings. Ubiquitously present non-targeted Penicillium strains may encounter such environments, resulting in their being placed under fungicide pressure. This might lead to the natural occurrence of resistance. Another possibility is exposure to environmental pyrrolnitrin produced by certain bacteria in the environment. Pyrrolnitrin is a lead compound of phenylpyrroles, and multi-drug resistance between pyrrolnitrin and phenylpyrroles has been reported [38]. Many bacterial species that belong to the genera Burkholderia and Pseudomonas produce pyrrolnitrin [39]. Some strains have been isolated from the rhizosphere and used as biological control agents against plant pathogenic fungi in agriculture. Therefore, there might The gene IDs are as follows: P. brasilanum PMG11_02111, P. camemberti PCAMFM013_S001g000092, P. chrysogenum EN45_023640, P. decumbens PENDEC_c002G07053, P. digitatum Pdw03_4331, P. expansum PEX2_037120, P. flavigenum PENFLA_c003G00500, P. griseofulvum PGRI_040000, P. italicum PITC_092520, P. oxalicum PDE_05313, P. roqueforti PROQFM164_S03g000214, and P. steckii PENSTE_c014G10375. The amino acid residues framed by orange and blue show the mutations detected in P. chrysogenum IFM 57243 and P. oxalicum IFM 54751, respectively. https://doi.org/10.1371/journal.pone.0262521.g005 be several environmental niches having high concentrations of microbial pyrrolnitrin. Thus, non-targeted Penicillium strains may have acquired resistance to pyrrolnitrin and multi-drug resistance to fludioxonil/iprodione owing to exposure to pyrrolnitrin produced by indigenous species. This study warns of the potential risk of non-targeted fungi around the world acquiring resistance to the fungicides. This issue requires further clarification.  PRN). Colony growth rate in the presence of pyrrolnitrin � 50%; "fungicide-resistant", < 50%; "fungicide-sensitive". (TIFF) S1 Table. PCR primers used in this study. (DOCX)