Identification of differentially expressed genes involved in spore germination of Penicillium expansum by comparative transcriptome and proteome approaches

Abstract In this study, Penicillium expansum, a common destructive phytopathogen and patulin producer was isolated from naturally infected apple fruits and identified by morphological observation and rDNA‐internal transcribed spacer analysis. Subsequently, a global view of the transcriptome and proteome alteration of P. expansum spores during germination was evaluated by RNA‐seq (RNA sequencing) and iTRAQ (isobaric tags for relative and absolute quantitation) approaches. A total of 3,026 differentially expressed genes (DEGs), 77 differentially expressed predicted transcription factors and 489 differentially expressed proteins (DEPs) were identified. The next step involved screening out 130 overlapped candidates through correlation analysis between the RNA‐seq and iTRAQ datasets. Part of them showed a different expression trend in the mRNA and protein levels, and most of them were involved in metabolism and genetic information processing. These results not only highlighted a set of genes and proteins that were important in deciphering the molecular processes of P. expansum germination but also laid the foundation to develop effective control methods and adequate environmental conditions.

a maximum tolerance limit of 50 μg·L −1 in apple-derived food products. EC Regulation 1881/2006 set the maximum limit of patulin for certain contaminants in food equal to 10 μg·kg −1 (or 10 μg·L −1 ) in apple juice and solid apple products for infants and young children (Sarubbi, Formisano, Auriemma, Arrichiello, & Palomba, 2016;Snini et al., 2016). In addition, P. expansum is a psychrophilic mold with a minimum growth temperature near freezing and an optimum growth temperature of 25°C. It can be easily disseminated by different vectors in orchards and by equipment (Morales et al., 2010). Therefore, P. expansum may cause serious economic losses to the international horticultural industry every year.
In addition to proper sanitation and careful handling during storage and processing, cold temperature, fungicide, controlled atmosphere, natural antimicrobials from plants, generally recognized as safe substances, and antagonistic microorganisms are usually utilized to inhibit the growth and patulin production of P. expansum (Cheon et al., 2016;Daniel, Lennox, & Vries, 2015;De Paiva et al., 2017;Moake, Padilla-Zakour, & Worobo, 2005;Morales et al., 2010). However, the overuse of synthetic fungicides has resulted in extensive concern about its effects on the environment and human health. In addition, the overuse of certain control strategies has led to increasing resistance of pathogens (Palou, Ali, Fallik, & Romanazzi, 2016). Therefore, it is imperative to understand the genetic basis of the infection, pathogenicity, toxigenicity, and virulence before developing new ecologically friendly control approaches and risk assessment models.
Fungal asexual spores that involve reproduction without the fusing of gametes are the main form of contamination or inoculation, since they are easily dispersed through the air or water. These spores are considered resistant to a broad spectrum of unfavorable environmental conditions due to both their cell wall properties and the accumulation of protective compatible solutes in the cytoplasm (Latgé & Beauvais, 2014). Mature conidia possess a condensed cell wall and a dense cytoplasm that contains a high content of energystorage materials such as the major carbon storage compounds trehalose and mannitol and an exogenously imposed dormancy (Voegele et al., 2005). Germination of the fungal conidia refers to the emergence of cells from resting asexual spores and that form sporeling hyphae or thalli under optimal growth conditions (Van Leeuwen et al., 2013). During this process, the hydrophobicity of the conidia results a progressive reduction that leads to the swelling of the conidia knows as isotropic growth and subsequent polarized growth characterized by germ tube formation (Dague, Alsteen, Latgé, & Dufrêne, 2008). It greatly affects fungal aggressiveness and colonization and can be considered to be one of the main steps of the disease cycle. Various studies have been conducted on the physiological and biochemical changes of P. expansum spores under different environment conditions including pH, nutrients, antimicrobial substances, temperature, and atmosphere composition (Kalai, Anzala, Bensoussan, & Dantigny, 2017;Long et al., 2017;Nanguy, Perrier-Cornet, Bensoussan, & Dantigny, 2010). Ballester et al. (2015) and Li et al. (2015) have completed genomic sequencing of P. expansum and some other Penicillia strains that will aid further molecular genetic analysis. However, no information is available regarding the intrinsic flexibility and variation in P. expansum at the transcriptional or translational level at different developmental stages such as germination.
Therefore, this study aimed to assess the transcriptome and proteome alternation of P. expansum during its germination stage using high throughput sequencing techniques. The results should provide useful information to better understand the molecular basis of development and pathogenicity of P. expansum and help to develop efficient control strategies.

| Fungal physiological measurements
Conidial suspensions were obtained by flooding the sporulating cultures of P. expansum with sterile distilled water containing 0.05% (v/v) Tween-20. A suitable aliquot of fresh spore suspension was added to a 250-ml conical flask containing 100 ml potato dextrose broth (PDB) to obtain a final concentration of 1.0 × 10 6 spores·ml −1 with the aid of a hemocytometer. The germination rate and germ tube length of the spores were microscopically determined during an incubation of period of 3-16 hr at 25°C on a rotary shaker at 200 rpm. 4′, 6′-diamidino-2-phenylindole (DAPI) was utilized at a concentration of 50 mg/L to facilitate observation of the morphology of fungal spores as described by Lai et al. (2015). The dry weight of the hyphae was measured after centrifugation, and then they were repeatedly washed with distilled water and dried at 60°C in an oven to a constant weight at the indicated time. The mycelial growth in vitro and in vivo was measured daily on PDA plates and apple fruits separately as described by Lai et al. (2017).

| Determination of patulin production
Fungal spores with a final concentration of 1.0 × 10 7 spores/ml were cultured in PDB medium under static conditions at 25°C for 2, 3 and 4 days. After centrifugation, the supernatant was filtered through a 0.22 μm syringe filter (Albet, Spain), and then 10 μl of the filtrates was injected into a reversed-phase HPLC (high performance liquid chromatography) system with a Waters XTerra RP18 column (Waters, USA) for the quantitative determination of the patulin concentration.
The mobile phase consisted of 95% water and 5% acetonitrile. The operational parameters were as follows: a flow rate of 0.8 ml/min, a column oven temperature of 18°C, and a detector wave length of 276 nm.

| RNA-Seq analysis
Fungal spores were cultured in PDB medium under shaking conditions for 6 and 12 hr as described previously. The harvested spores and mycelia were washed twice by sterile distilled water and quickly frozen with liquid nitrogen for subsequent RNA-Seq and iTRAQ analysis.
Total RNA preparation and sequencing were performed using a service from Beijing Genomics Institute (BGI) Co., Ltd. The experiment was repeated once. Briefly, high-quality total RNA with a 28S:18S ratio >1.5 and an absorbance 260/280 ratio between 1.7 and 2.0 was extracted using TRIzol reagent (Invitrogen, USA) following the manufacturer's instruction. The cDNA library for each sample was constructed utilizing TruSeq RNA Sample Preparation Kits v2 (Illumina, USA) according to the manufacturer's instructions. After quantification and qualification during the quality control step, the sample library was sequenced on the BGISEQ-500 platform.
The raw data were filtered to remove adaptors and reads in which unknown bases were more than 10% or the percentage of low quality bases was over 50%. The remaining clean reads were stored as FASTQ format and mapped to reference (https://www.ncbi.

| iTRAQ-based quantitative proteomics analysis
Isobaric tags for relative and absolute quantitation (iTRAQ) analysis was performed utilizing a service from LC-Bio of LC Science (USA) as described by Lai et al. (2016). Protein preparation utilized the standard extraction procedure of LC-Bio of LC Science. One  Table S3.
Comparison of the RNA-seq and iTRAQ datasets was based on the semantic similarity of their GO terms. Detail information of the common DEGs acquired from the two approaches is presented in Figure 6.

| Statistical analysis
Except for the experiments where specified, the data were pooled across three independent repeat experiments, and statistical analysis were performed with SPSS software (SPSS Inc., Chicago, IL, USA).
Analysis of variance (ANOVA) was used to compare more than two means, and Duncan's multiple range test was used for mean separations. Differences at p < .05 were considered to be significant.

| Identification of a strain of P expansum from decaying market fruit in the Chinese province of Zhejiang
The fungus, isolated from decaying market fruit, proved to be a major causative agent by artificial infection tests, and it was highly pathogenic. Therefore, we decided to identify and characterize this pathogen by morphological observation and rDNA-internal transcribed spacer analysis. After an artificial re-inoculation in vivo and in vitro, the cultural and morphological characteristics of the purified pathogen were consistent with those of blue mold. In apple fruits, the color of the lesions was light brown. The infected tissue with an earthymusty odor and blue pustules became soft and mushy, and had a watery consistency. On the older lesions, the pathogen formed typically brush-like sporulating structures that produced single celled spores ( Figure 1a). The spore masses initially appeared as white mycelia that then turned bluish-green in color from the edge to the center of the colonies on PDA plates (Figure 1b). The conidia were spherical or elliptical and contained one or multiple nuclei (Figure 1c and d).
The pathogen was also highly pathogenic to orange, jujube, tomato, and pear fruits (Figure 1e-h). After amplification using ITS universal primers, an approximately 800-bp fragment was identified ( Figure S1). Through sequencing and MegaBLAST analysis, the submitted sequence was found to match perfectly with the ITS of P. expansum strains ( Figure 1i). Therefore, the isolated pathogen was confirmed as P. expansum.

| Growth dynamic analysis of P. expansum in vitro and in vivo
As one of the most prevalent postharvest diseases, the growth dynamics of P. expansum that partially represent fungal competitiveness and dominance are important to determine. In this study, several growth indicators were determined at different incubation times. After 6 hr of cultivation in PDB, the average volume of the spores was visually enlarged. After 12 hr, the germination rate had reached more than 90% (Figure 2a). After 16 hr, the average length of the germ tube was approximately 60 μm ( Figure 2b). After 48 hr, the mycelial dry weight derived from 1.0 × 10 8 spores had reached about 1.2 g (Figure 2c).
After 6 days of cultivation on PDA, colonies nearly extended to the entire petri dish (Figure 2d). In addition, P. expansum was strongly pathogenic to apple fruits and was highly destructive (Figure 2e).
After 6 days of infection, the lesion diameter on apple fruits was almost 4 cm. Patulin is an important indicator of pathogenicity of P. expansum. Patulin production significantly increased with the time of culture. The concentration of patulin was greater than 130 μg·ml −1 after 4 days of incubation (Figure 2f).  (Table S2). However, the functions of most of them are unknown.

| Identification and functional classification of DEPs during spore germination of P. expansum
Although transcriptome profiling provides substantial biological information, its ability to provide a comprehensive perspective for this

| Comparative analysis of RNA-seq and iTRAQ dataset
Integrative analysis of transcriptomic and proteomic data may present complementary information that enables more informative conclusions to be drawn. After combination and simplification based on the homologies and biological function annotations and deleting the results with unknown functions, correlation analysis between iTRAQ and RNA-seq datasets was performed. A total of 130 genes in the overlap were identified that were involved in amino acid metabolism (11), carbohydrate metabolism (23), energy metabolism (6), lipid metabolism (9), metabolism of cofactors and vitamins (9), nucleotide metabolism (16), translation (25), folding, sorting and degradation (9), transcription (6), transport and catabolism (6), signal transduction (6) or other pathways. Generally, the alterations of the expression of these genes showed a larger variation range at the level of transcription, and the expression trend of about half of the candidates was different between the transcriptional and translational levels ( Figure 5 and Figure 6c and e). A poor correlation between the quantities of these two macromolecules indicated a complex regulatory mechanism controlling expression both at the RNA and the protein levels during spore germination of P. expansum. In fact, similar findings have previously been found in many other biological processes such as in humans, yeast, and plants (Hua et al., 2016).

| DISCUSSION
Penicillium expansum is a destructive phytopathogen, capable of causing decay in deciduous fruits and vegetables during postharvest handling and storage. It also produces large amounts of secondary metabolites, especially patulin (Morales et al., 2010;Tannous et al., 2014). In this study, P. expansum was purified from naturally infected The X axis indicates the number of DEGs (the number is presented by its square root value). The Y axis represents GO terms. All of the GO terms are grouped into three ontologies: blue is for biological process, brown is for cellular component and orange is for molecular function; (b) Statistics of the pathway enrichment of DEGs in each pairwise. The rich factor is the ratio of DEGs numbers to all gene numbers annotated in this pathway term. A greater rich factor indicates greater intensity. The Q value is the corrected p value ranging from 0 to 1, and a lower Q value indicates greater intensity. The top 20 of the enriched pathway terms are displayed. (c) KEGG classification of DEGs. The X axis indicates the number of DEGs. The Y axis represents the second KEGG pathway terms. All of the second pathway terms are grouped in the top pathway terms indicated in different colors. The detailed information of these DEGs is listed in Table S1 of spore germination at an unprecedented depth and coverage. For example, using the Affymetrix GeneChip, Seong, Zhao, Xu, Güldener, and Kistler (2008) found that 185 genes were specifically expressed and 2,566 genes were significantly changed at different stages during spore germination of Fusarium graminearum. Most of these genes were However, these researchers focused on the molecular processes underlying spore germination at the single genome, transcriptional or protein level. In fact, molecular differences may be exhibited across multiple layers of gene regulation such as genomic variations, gene expression, protein translation and post-translational modifications.
Analyzing them separately may not be very informative. This necessitates the integrative analysis of such multiple layers of information to understand the interplay of the individual components.
When it comes to P. expansum, there are few reports about the changes in its gene expression during spore germination, since its F I G U R E 4 Scatter plots of the top 20 Gene Ontology (a) and KEGG (b) enrichment of DEPs involved in spore germination of P. expansum. A deeper color in the color code represents the higher confidence for the biological process. Rich factor: the number of DEPs in one GO or KEGG/ the number of total identified proteins in the same GO or KEGG. A higher value indicated higher enrichment level. The detailed information of these DEPs is listed in Table S3 F I G U R E 5 Expression changes of overlapping candidates involved in metabolism (a), genetic information processing (b) and others (c) from the iTRAQ and RNA-seq datasets. Each row in the color heat map indicates a single protein. The annotation and involved metabolic pathway of each protein are listed. Blue to red in the color code indicates down-to up-regulated expression of the candidates genome was just completely sequenced in 2015 (Ballester et al., 2015;Li et al., 2015). In addition, integrating the information from transcriptomic and proteomic data to gain meaningful insights of fungal spore germination has rarely been reported. Therefore, two high-throughput technologies of RNA-seq and iTRAQ were used to explore global gene regulation during spore germination of P. expansum in this study. A total of 3,026 genes was found to be differently regulated, and 489 proteins showed variation in their abundance. Pathway classification was compared between DEGs from RNA-seq and DEPs from iTRAQ analysis. Similar results were observed that the majority of DEGs or DEPs were involved in translation, transcription, and metabolism of carbohydrates, amino acids, lipids, energy, and nucleotides. They were absolutely required for spore metabolism and growth during germination. A total of 130 DEPs corresponded to gene sequences F I G U R E 6 Global overview of the changes in phenotype, transcriptome and proteome of P. expansum spores during germination. (a) Phenotypic changes of P. expansum spores during development. The red arrow corresponds to the focused stage in this study; (b) Scatter plots of all expressed genes acquired from the RNA-seq method during the spore germination stage. The X-axis and Y-axis show the log 2 value of gene expression. Blue, orange, and brown dots correspond to down-regulated, up-regulated and nonregulated genes, respectively; (c) Venn diagram of the overlap of DEGs from RNA-seq and iTRAQ datasets. (d) Protein ratio distribution acquired from the iTRAQ method during spore germination. The X-axis presents the −log 2 value of the protein ratio (6 hr/12 hr). Deep pink, azure, and gray dots correspond to up-regulated, down-regulated and nonregulated proteins, respectively. (e) Pathway classification of overlapped DEGs acquired from RNA-seq and iTRAQ methods that were obtained by RNA-seq, and this enabled a comparison of germination-related differences in specific transcripts or cognate proteins. Therefore, by narrowing of the scope and reducing the number, the 130 overlapped DEPs could probably yield more precise and valuable information to explore the molecular mechanism of P. expansum germination. A component protein of the proteasome whose function was to degrade unneeded or damaged proteins by proteolysis was decreased as well (Enenkel, 2014). Therefore, protein synthesis capacity during spore germination had increased and led to the subsequent changes in the biophysical or biochemical profiles of P. expansum.
Transcription is the first step of gene expression, in which a particular segment of DNA is copied into RNA, especially mRNA, by the enzyme RNA polymerase. In this study, almost all DEPs related to nucleotide metabolism and transcription were down-regulated at the mRNA level and up-regulated at the protein level. Among them, multiple types of RNA polymerases were notable, as they are necessary for constructing RNA chains using DNA genes as templates and are responsible for the synthesis of distinct subsets of RNA. These results indicated that genetic information processing had been activated and accelerated from dormancy to vegetative growth of P. expansum spores.
Energy metabolism is a set of metabolic reactions and processes to convert biochemical energy from nutrients into adenosine triphosphate (ATP) and release waste products. In this study, the expression levels of DEPs related to energy metabolism were all down-regulated at both mRNA and protein levels. The possible reasons for this are delineated. Before germinating, the dormant spores need to shed their cell walls, degrade compatible sugars accompanied by a decrease in microviscosity of the cytoplasm, and reorganize the transcriptome including mRNA breakdown and the selected up-regulation of different gene categories (Lamarre et al., 2008). Therefore, dormant spores need more energy than ones that have germinated. In addition, the DEPs related to transport, signal transduction, motility, and other molecular types of metabolism were recorded. It is reasonable that the conversion of dormant cells into vegetative cells requires active metabolism and cell division.
Overall, a large number of genes showed consistency between the transcript and protein levels, while some genes that exhibited inconsistency between these levels suggested that posttranscriptional regulation and modification serve important roles in the regulation of fungal germination. Additional research is needed to better understand the fact that the turnover and stability of mRNA levels are important for the translation of mRNA into proteins. For example, the downregulation of expression at the transcriptional level combined with the up-regulation of expression at the protein level of some DEGs implied that mRNA had already begun to be degraded. Conversely, there may be regulation of translational pathways or increased protein degradation leading to apparently enhanced transcription. In summary, the transcriptomic and proteomic data in this study not only highlighted a set of genes and proteins that were important in deciphering the molecular processes of the germination of P. expansum but also provided the basic knowledge for the development of effective control methods and adequate environmental conditions.

ACKNOWLEDGMENTS
This work was supported by the National Natural Science

CONFLICT OF INTEREST
None declared.