Genome Sequencing of Paecilomyces Penicillatus Provides Insights into Its Phylogenetic Placement and Mycoparasitism Mechanisms on Morel Mushrooms

Morels (Morchella spp.) are popular edible fungi with significant economic and scientific value. However, white mold disease, caused by Paecilomyces penicillatus, can reduce morel yield by up to 80% in the main cultivation area in China. Paecilomyces is a polyphyletic genus and the exact phylogenetic placement of P. penicillatus is currently still unclear. Here, we obtained the first high-quality genome sequence of P. penicillatus generated through the single-molecule real-time (SMRT) sequencing platform. The assembled draft genome of P. penicillatus was 40.2 Mb, had an N50 value of 2.6 Mb and encoded 9454 genes. Phylogenetic analysis of single-copy orthologous genes revealed that P. penicillatus is in Hypocreales and closely related to Hypocreaceae, which includes several genera exhibiting a mycoparasitic lifestyle. CAZymes analysis demonstrated that P. penicillatus encodes a large number of fungal cell wall degradation enzymes. We identified many gene clusters involved in the production of secondary metabolites known to exhibit antifungal, antibacterial, or insecticidal activities. We further demonstrated through dual culture assays that P. penicillatus secretes certain soluble compounds that are inhibitory to the mycelial growth of Morchella sextelata. This study provides insights into the correct phylogenetic placement of P. penicillatus and the molecular mechanisms that underlie P. penicillatus pathogenesis.


Introduction
Morels (Morchella spp.) are valuable edible fungi that belong within the order Pezizales. There is a long history of morel consumption in Asia, Europe, and North America [1]. The increasing demand for dried and fresh morels has driven the growth of the industry and the expansion of suitable morel cultivation areas [2,3]. In China, the morel cultivation area reached approximately 9000 ha in the production season of 2018-2019 and continues to expand [4]. With the expanding cultivation range and density of production, disease has become a major limiting factor for morel production. Typical diseases include stipe rot disease caused by the Fusarium incarnatum-F. equiseti species complex and pileus rot disease caused by Diploöspora longispora [5,6]. White mold disease, caused by Paecilomyces

Fungal Material
The P. penicillatus strain CCMJ2836 was provided by the Engineering Research Center of the Chinese Ministry of Education for Edible and Medicinal Fungi, Jilin Agricultural University (Changchun, China), which was isolated from white mold-diseased morel fruiting bodies in Sichuan Province, China. The mycelial plugs of strain CCMJ2836 were grown on potato glucose agar (PDA) medium overlaid with cellophane sheets for 10 d at 25 • C under a light/dark photoperiod (12/12 h).

Dual Culture Assay
Cultures of P. penicillatus strain CCMJ2836 and M. sextelata strain GB-Ch-102 were inoculated on HIMEDIA corn meal peptone yeast agar (CMA) and incubated at room temperature for 5 d. P. penicillatus agar discs were inoculated on one side of regular or two-compartment CMA dishes. After 6 d, M. sextelata agar discs were inoculated on the other side of the petri dish at about 4 cm from the P. penicillatus agar disc and then incubated for two more days. Each confrontation culture was replicated 10 times and incubated at room temperature. Cultures were photographed with an Epson Perfection V700 Photo scanner.

Genome Sequencing and Assembly of P. penicillatus
The genome of P. penicillatus CCMJ2836 strain was sequenced on the PacBio SMRT Sequel platform using two SMRT cells. A total of 4,266 Mb clean data (~105 × coverage) was generated (Table S1). The de novo assembly of the P. penicillatus genome was~40.20 Mb, consisting of 52 scaffolds with 2.60 Mb in N50 value and 44.70% in guanine-cytosine (GC) content ( Table 1). The completeness of the P. penicillatus genome was assessed through the CEGMA and BUSCOs analyses, with completeness scores of 95.56% and 97.20%, respectively. The whole genomic sequence of P. penicillatus CCMJ2836 has been deposited at GenBank and is available under accession number JACGSR000000000. The version described in this paper is version JACGSR010000000.

Phylogenomic Analysis
In view of the polyphyletic status of the Paecilomyces genus and the uncertain phylogenetic placement of P. penicillatus, a phylogenetic analysis was conducted using the single-copy orthologous genes ( Figure 1). As expected, species in Eurotiomycetes and Sordariomycetes were clearly separated into two phylogenetic clades. In agreement with the conclusion made from both morphological characteristics and ITS sequences [9], P. variotii was placed under Eurotiomycetes. Within Sordariomycetes, Hypocreales species grouped together and were separated from N. crassa in Sordariales. P. penicillatus was placed under Hypocreales and was closest to Hypocreaceae, which includes many mycoparasitic fungi that can infect edible mushrooms. The next closest families were Nectriaceae and Niessliaceae. P. penicillatus is distantly related to Stachybotryaceae, Clavicipitaceae and Ophiocordycipitaceae, and has the most distant relationship with Cordycipitaceae. Our results indicate that Paecilomyces is polyphyletic and that P. penicillatus and P. variotii are clearly distantly related. We conclude that P. penicillatus is in the order of Hypocreales but not Eurotiales, which was previously concluded based on morphological characters. Our findings support a previous study based on 18S rRNA and β-tubulin gene sequences [8,11]. Our phylogenetic analysis further demonstrated that P. penicillatus is closely related to Hypocreaceae. A previous phylogenetic analysis of the 5.8S rDNA and ITS sequences of entomopathogenic Paecilomyces spp. demonstrated that most Paecilomyces species are Hypocreales and a few Paecilomyces species are Eurotiales. Within Hypocreales, Paecilomyces spp. was further classified into three subgroups. One subgroup includes Paecilomyces viridis, P. penicillatus, and Paecilomyces carneus. Most strains of Paecilomyces lilacinus and Paecilomyces marquandii are closely related and are classified as a distinct subgroup. The other subgroup contains Paecilomyces farinosus, Paecilomyces fumosoroseus, and other Paecilomyces species [9]. Due to the lack of genomic resources of most Paecilomyces spp., a single-copy orthologous genes-based phylogenetic analysis is not technically possible yet. Our study, therefore, lays the foundation for further phylogenetic classification of Paecilomyces spp. by deploying the single-copy orthologous gene sequences.

Phylogenomic Analysis
In view of the polyphyletic status of the Paecilomyces genus and the uncertain phylogenetic placement of P. penicillatus, a phylogenetic analysis was conducted using the single-copy orthologous genes ( Figure 1). As expected, species in Eurotiomycetes and Sordariomycetes were clearly separated into two phylogenetic clades. In agreement with the conclusion made from both morphological characteristics and ITS sequences [9], P. variotii was placed under Eurotiomycetes. Within Sordariomycetes, Hypocreales species grouped together and were separated from N. crassa in Sordariales. P. penicillatus was placed under Hypocreales and was closest to Hypocreaceae, which includes many mycoparasitic fungi that can infect edible mushrooms. The next closest families were Nectriaceae and Niessliaceae. P. penicillatus is distantly related to Stachybotryaceae, Clavicipitaceae and Ophiocordycipitaceae, and has the most distant relationship with Cordycipitaceae. Our results indicate that Paecilomyces is polyphyletic and that P. penicillatus and P. variotii are clearly distantly related. We conclude that P. penicillatus is in the order of Hypocreales but not Eurotiales, which was previously concluded based on morphological characters. Our findings support a previous study based on 18S rRNA and β-tubulin gene sequences [8,11]. Our phylogenetic analysis further demonstrated that P. penicillatus is closely related to Hypocreaceae. A previous phylogenetic analysis of the 5.8S rDNA and ITS sequences of entomopathogenic Paecilomyces spp. demonstrated that most Paecilomyces species are Hypocreales and a few Paecilomyces species are Eurotiales. Within Hypocreales, Paecilomyces spp. was further classified into three subgroups. One subgroup includes Paecilomyces viridis, P. penicillatus, and Paecilomyces carneus. Most strains of Paecilomyces lilacinus and Paecilomyces marquandii are closely related and are classified as a distinct subgroup. The other subgroup contains Paecilomyces farinosus, Paecilomyces fumosoroseus, and other Paecilomyces species [9]. Due to the lack of genomic resources of most Paecilomyces spp., a single-copy orthologous genesbased phylogenetic analysis is not technically possible yet. Our study, therefore, lays the foundation for further phylogenetic classification of Paecilomyces spp. by deploying the single-copy orthologous gene sequences.

The CAZyme
To successfully infect fungal hosts, mycoparasitic fungi commonly produce chitinases and glucanases to aid in the digestion of the fungal host cell wall, which is composed of chitin and glucan [19,55,56]. In light of the mycoparasitic nature of P. penicillatus, we hypothesized that P. penicillatus may also produce enzymes that can degrade the cell wall of its host. To assess this possibility, the genome of P. penicillatus was mapped to the CAZymes database (http://www.cazy.org), which includes all the enzyme families known to function in carbohydrate metabolism [57]. A total of 299 CAZymes were identified including 170 GHs, 67 GTs, 28 AAs, 24 CBMs, seven CEs, and three PLs (Figure 2). The most abundant CAZyme family in P. penicillatus is the GH family 18. The second and the third most abundant CAZyme families are the GH family 16 and the AA family 3, respectively ( Figure 2). Pathogens 2020, 9, x FOR PEER REVIEW 6 of 13

The CAZyme
To successfully infect fungal hosts, mycoparasitic fungi commonly produce chitinases and glucanases to aid in the digestion of the fungal host cell wall, which is composed of chitin and glucan [19,55,56]. In light of the mycoparasitic nature of P. penicillatus, we hypothesized that P. penicillatus may also produce enzymes that can degrade the cell wall of its host. To assess this possibility, the genome of P. penicillatus was mapped to the CAZymes database (http://www.cazy.org), which includes all the enzyme families known to function in carbohydrate metabolism [57]. A total of 299 CAZymes were identified including 170 GHs, 67 GTs, 28 AAs, 24 CBMs, seven CEs, and three PLs (Figure 2). The most abundant CAZyme family in P. penicillatus is the GH family 18. The second and the third most abundant CAZyme families are the GH family 16 and the AA family 3, respectively ( Figure 2). The GH family 18 contains all the enzymes that have been previously shown to contribute to the degradation and replenishment of both intrinsic and extrinsic fungal chitin [58] (Figure 2). GH family 18 can be further divided into three subgroups, namely A, B, and C [59,60]. We were particularly interested in subgroup C (sgC) chitinases due to their potential involvement in mycoparasitism and their other interesting features [41,61]. Currently, known sgC chitinases contain CBM 18 (chitin binding) domains and/or CBM 50 (LysM) domains [18,62]. An analysis of sgC chitinases was performed. Based on the domain prediction, out of the 20 chitinases in GH family 18, five were predicted to be in sgC. Three of them have the CBM 18 domain and the other two have both the CBM 18 and the CBM 50 domains. In view of the close phylogenetic relationship between P. penicillatus and Trichoderma spp. and the well-studied production of fungal cell wall-degradation enzymes by Trichoderma spp., a phylogenetic analysis of sgC chitinases of P. penicillatus, Trichoderma atroviride and T. virens was conducted ( Figure 3). Interestingly, our results showed that out of five genes, only two orthologue pairs could be found in the sgC chitinases of P. penicillatus and T. virens (PP_10005833/TVC9, PP_10007271/TVC10). The drastic differences in the number and sequence divergence of sgC in these phylogenetically closely-related mycoparasitic organisms suggests a strong evolutionary pressure on this protein family.
Beside chitinases, <!--MathType@Translator@5@5@MathML2 (no namespace).tdl@MathML 2.0 (no namespace)@ --> The GH family 18 contains all the enzymes that have been previously shown to contribute to the degradation and replenishment of both intrinsic and extrinsic fungal chitin [58] (Figure 2). GH family 18 can be further divided into three subgroups, namely A, B, and C [59,60]. We were particularly interested in subgroup C (sgC) chitinases due to their potential involvement in mycoparasitism and their other interesting features [41,61]. Currently, known sgC chitinases contain CBM 18 (chitin binding) domains and/or CBM 50 (LysM) domains [18,62]. An analysis of sgC chitinases was performed. Based on the domain prediction, out of the 20 chitinases in GH family 18, five were predicted to be in sgC. Three of them have the CBM 18 domain and the other two have both the CBM 18 and the CBM 50 domains. In view of the close phylogenetic relationship between P. penicillatus and Trichoderma spp. and the well-studied production of fungal cell wall-degradation enzymes by Trichoderma spp., a phylogenetic analysis of sgC chitinases of P. penicillatus, Trichoderma atroviride and T. virens was conducted ( Figure 3). Interestingly, our results showed that out of five genes, only two orthologue pairs could be found in the sgC chitinases of P. penicillatus and T. virens (PP_10005833/TVC9, PP_10007271/TVC10). The drastic differences in the number and sequence divergence of sgC in these phylogenetically closely-related mycoparasitic organisms suggests a strong evolutionary pressure on this protein family.
Beside chitinases, β-(1,3)-glucanases are another enzyme family that decomposes the fungal cell wall and belongs to GH family17, 55, 64, and 81 [41]. Compared with Trichoderma spp. and other representative fungal pathogens of plants or animals, P. penicillatus and Trichoderma spp. encode the highest number of GH family 18 proteins among the fungal organisms included in this study [41]. For GH families 64 and 81, we found a higher number in P. penicillatus and Trichoderma spp. in comparison to the other filamentous fungi. Similarly, P. penicillatus contains the highest number of GH family 75, which could be involved in the decomposition of chitinous carbohydrates [41]. Our results indicated that P. penicillatus produces a plethora of chitinases and β-(1,3)-glucanases in resemblance to the well-characterized mycoparasitic Trichoderma spp.

Secondary Metabolites and Pathogenicity-Related Genes
We identified 74 putative secondary metabolite biosynthesis gene clusters in P. penicillatus. Among them, seven gene clusters had significant hits to known secondary metabolite biosynthesis gene clusters (Table 2), including two antifungal (squalestatin 1 and AbT1), one antibacterial (cephalosporin C), one insecticidal (leucinostatin), two mammal-toxic (monascorubrin and 6-methylsalicyclic acid), and one plant-toxic (dimethylcoproge) secondary metabolic compounds. Multiple P. penicillatus genes in these gene clusters showed significant sequence similarity to genes in the Pathogen-Host Interactions database (PHI-base) and/or the database of fungal virulence factor (DFVF), indicative of possible involvement of these gene clusters in the pathogenesis of P. penicillatus. Specifically, Squalestatin 1 (40% similarity) is a potent inhibitor of squalene synthase, an essential enzyme for sterol biosynthesis [63]. Squalestatin 1 exhibits broad-spectrum antifungal activity [64] and was identified in Phoma sp. and several ascomycetes [65,66]. AbT1 (100% similarity) is a precursor of the cyclic peptide antibiotic Aureobasidin A (AbA), known to be produced by Aureobasidium pullulans [67]. AbA has been shown to exhibit inhibitory activity to yeast inositol phosphorylceramide synthase and is toxic to several other fungal organisms, such as Aspergillus nidulans and A. niger [67,68]. The genome of P. penicillatus also contains a gene cluster that has a 40% similarity to that of a gene involved in the biosynthesis of leucinostatin, a compound with antitrypanosomal properties. Interestingly, most Paecilomyces spp. identified thus far are entomophagous [69], and leucinostatin was previously reported to be produced by a Paecilomyces sp. isolated from soil [70]. In addition, several causative gene clusters for mammal or plant mycotoxins or their precursors were also predicted in P. penicillatus with 100% similarity, including dimethylcoproge [71], monascorubrin [72], and 6-methylsalicyclic acid [73]. It remains to be investigated whether these mycotoxic secondary compounds facilitate the infection of P. penicillatus on morel mycelia and fruiting bodies, or whether they coordinate or contribute to interspecies competition against other microbes living in the same environmental niche.  [41]. Compared with Trichoderma spp. and other representative fungal pathogens of plants or animals, P. penicillatus and Trichoderma spp. encode the highest number of GH family 18 proteins among the fungal organisms included in this study [41]. For GH families 64 and 81, we found a higher number in P. penicillatus and Trichoderma spp. in comparison to the other filamentous fungi. Similarly, P. penicillatus contains the highest number of GH family 75, which could be involved in the decomposition of chitinous carbohydrates [41]. Our results indicated that P. penicillatus produces a plethora of chitinases and β-(1,3)-glucanases in resemblance to the wellcharacterized mycoparasitic Trichoderma spp.

Dual Culture Assay
Based on the above analyses, we hypothesized that P. penicillatus may produce certain inhibitory compounds that can affect the hyphal growth of morels. To test this possibility, we conducted dual culture assays where agar plugs of both P. penicillatus and M. sextelata were inoculated onto the same petri dish. Indeed, M. sextelata had minimum growth in the direction towards the colony of P. penicillatus, and extensive growth of M. sextelata was observed in the direction away from P. penicillatus ( Figure 4A), suggesting that P. penicillatus does produce host-inhibitory substances in a contact-independent manner. In an attempt to further investigate whether the substances are volatile or soluble, we also conducted confrontation assays on enclosed two-compartment petri dishes, where P. penicillatus and M. sextelata were inoculated on the center of each compartments. We expected to observe a similar inhibitory effect of P. penicillatus on M. sextelata in the case of volatile compounds being produced by P. penicillatus. However, M. sextelata grew rapidly in all directions at nearly equal speed in these assays ( Figure 4B). This indicated that the inhibitory compounds were unlikely to be volatile, and rather, that they must be secreted and diffused in the medium. We are currently investigating whether P. penicillatus exhibits the antagonistic activity against M. sextelata through extracellular fungal cell wall-degradation enzymes or mycotoxins, as predicted in this study.
Pathogens 2020, 9, x FOR PEER REVIEW 9 of 13 speed in these assays ( Figure 4B). This indicated that the inhibitory compounds were unlikely to be volatile, and rather, that they must be secreted and diffused in the medium. We are currently investigating whether P. penicillatus exhibits the antagonistic activity against M. sextelata through extracellular fungal cell wall-degradation enzymes or mycotoxins, as predicted in this study. Normal petri dish (B). Two-compartment petri dish.

Conclusions
In this study, we present the first whole-genome sequence of P. penicillatus, the causative agent of the morel white mold disease. The genome of P. penicillatus is ~40.20 Mb in size with an N50 value of 2.59 Mbp and 9454 coding genes. The phylogenetic analysis based on the single-copy orthologous genes demonstrated that P. penicillatus is within Hypocreales and is closely related to Hypocreaceae, a family that is known to include many mycoparasitic species. CAZymes analysis further illustrated that the genome of P. penicillatus has a large number of fungal cell wall degradation enzymes. We identified a plethora of gene clusters that are known to be involved in the biosynthesis of cytotoxic secondary compounds that may function coordinatively during the pathogenesis of P. penicillatus. We also showed that P. penicillatus can produce contact-independent soluble host-inhibitory compounds. Taken together, we provide strong evidence in support of the phylogenetic placement

Conclusions
In this study, we present the first whole-genome sequence of P. penicillatus, the causative agent of the morel white mold disease. The genome of P. penicillatus is~40.20 Mb in size with an N50 value of 2.59 Mbp and 9454 coding genes. The phylogenetic analysis based on the single-copy orthologous genes demonstrated that P. penicillatus is within Hypocreales and is closely related to Hypocreaceae, a family that is known to include many mycoparasitic species. CAZymes analysis further illustrated that the genome of P. penicillatus has a large number of fungal cell wall degradation enzymes. We identified a plethora of gene clusters that are known to be involved in the biosynthesis of cytotoxic secondary compounds that may function coordinatively during the pathogenesis of P. penicillatus. We also showed that P. penicillatus can produce contact-independent soluble host-inhibitory compounds. Taken together, we provide strong evidence in support of the phylogenetic placement of P. penicillatus within Hypocreales and provide a basis for the future functional characterization of the fungal cell wall-degradation enzymes and secondary metabolites produced by this mycoparasitic fungus.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-0817/9/10/834/s1, Text S1: Annotation of the genome of P. penicillatus, Table S1: CEGMA analysis on the completion of the P. penicillatus genome, Table S2: Functional annotation of the genes in the P. penicillatus genome against several databases, Table S3: The statistics of the repeat sequences in the P. penicillatus genome, Table S4: Annotation of tRNA, rRNA, miRNA, and snRNA in the P. penicillatus genome.
Author Contributions: Conceptualization and supervision, Y.L. and Y.F.; Formal analysis and writing, X.W.; Methodology, J.P.; Investigation, X.W. and Y.G.; software, J.P., L.S., and X.W.; Writing-review and editing, J.P., Y.F., and G.B. All authors have read and agreed to the published version of the manuscript.