Differential carbohydrate-active enzymes and secondary metabolite production by the grapevine trunk pathogen Neofusicoccum parvum Bt-67 grown on host and non-host biomass

ABSTRACT

Neofusicoccum parvum is one of the most aggressive Botryosphaeriaceae species associated with grapevine trunk diseases. This species may secrete enzymes capable of overcoming the plant barriers, leading to wood colonization. In addition to their roles in pathogenicity, there is an interest in taking advantage of N. parvum carbohydrate-active enzymes (CAZymes), related to plant cell wall degradation, for lignocellulose biorefining. Furthermore, N. parvum produces toxic secondary metabolites that may contribute to its virulence. In order to increase knowledge on the mechanisms underlying pathogenicity and virulence, as well as the exploration of its metabolism and CAZymes for lignocellulose biorefining, we evaluated the N. parvum strain Bt-67 capacity in producing lignocellulolytic enzymes and secondary metabolites when grown in vitro with two lignocellulosic biomasses: grapevine canes (GP) and wheat straw (WS). For this purpose, a multiphasic study combining enzymology, transcriptomic, and metabolomic analyses was performed. Enzyme assays showed higher xylanase, xylosidase, arabinofuranosidase, and glucosidase activities when the fungus was grown with WS. Fourier transform infrared (FTIR) spectroscopy confirmed the lignocellulosic biomass degradation caused by the secreted enzymes. Transcriptomics indicated that the N. parvum Bt-67 gene expression profiles in the presence of both biomasses were similar. In total, 134 genes coding CAZymes were up-regulated, where 94 of them were expressed in both biomass growth conditions. Lytic polysaccharide monooxygenases (LPMOs), glucosidases, and endoglucanases were the most represented CAZymes and correlated with the enzymatic activities obtained. The secondary metabolite production, analyzed by high-performance liquid chromatography–ultraviolet/visible spectophotometry–mass spectrometry (HPLC-UV/Vis-MS), was variable depending on the carbon source. The diversity of differentially produced metabolites was higher when N. parvum Bt-67 was grown with GP. Overall, these results provide insight into the influence of lignocellulosic biomass on virulence factor expressions. Moreover, this study opens the possibility of optimizing the enzyme production from N. parvum with potential use for lignocellulose biorefining.

GRAPHICAL ABSTRACT

KEYWORDS:

• CAZymes
• enzymatic activities
• lignocellulosic biomass
• transcriptomic
• virulence factors

INTRODUCTION

Plant-pathogenic fungi deploy a wide array of cell wall–degrading enzymes (CWDEs) to overcome the host’s physical defenses and utilize their degradation products for growth and reproduction (Kubicek et al. 2014). The fungal degradation capacity relies on complex degradative machinery able to catalyze two processes: direct enzymatic depolymerization and generation of oxidative species that act on lignocellulosic biomass (Cragg et al. 2015). According to the cell wall polymer that they degrade, the CWDEs are (i) ligninolytic enzymes (e.g., laccases and peroxidases), (ii) cellulolytic enzymes (e.g., endoglucanases, exoglucanases, and β-glucosidases), and (iii) hemicellulolytic enzymes (e.g., endoxylanases, αarabinofuranosidase, β-xylosidase, and others) (Saldarriaga-Hernández et al. 2020). The enzymes using carbohydrates as substrates are grouped in carbohydrate-active enzyme (CAZymes) families within five classes: glycoside hydrolases (GHs), glycosyltransferases (GTs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), and auxiliary activities (AAs) (Lombard et al. 2014). These enzymes can be accompanied by carbohydrate-binding modules (CBMs), which promote the association of the enzyme with the substrate (Boraston et al. 2004). This wide variety of CAZymes in fungal plant pathogens makes them a promissory source for novel lignocellulolytic enzymes (Gibson et al. 2011).

Neofusicoccum parvum, included in the Botryosphaeriaceae family (phylum Ascomycota), is a ubiquitous fungus that causes disease in a broad range of plant species (Phillips et al. 2013). Its pathogenicity has been confirmed on 90 different hosts, mainly woody angiosperms (Sakalidis et al. 2013). Neofusicoccum parvum is one of the most aggressive Botryosphaeriaceae species in grapevine, commonly reported as a causal agent of Botryosphaeria dieback (Mondello et al. 2018; Travadon et al. 2013; Úrbez-Torres 2011; Úrbez-Torres and Gubler 2009). This disease, included among the grapevine trunk diseases, significantly affects the productivity of vineyards worldwide (Bertsch et al. 2013; Fontaine et al. 2016). Neofusicoccum parvum may enter the grapevine through fresh wounds (e.g., pruning wounds) or via natural openings (e.g., lenticels) and colonizes the woody tissues, causing discolored and decayed zones in the wood (Claverie et al. 2020; Reis et al. 2020). As a wood colonizer, this pathogen secretes enzymes capable of degrading plant cell polymers (Massonnet et al. 2018; Stempien et al. 2017). Genomic analyses of different N. parvum strains have pointed out that this fungal species possesses an important repertoire of lignocellulolytic enzymes (Belair et al. 2023; Blanco-Ulate et al. 2013; Massonnet et al. 2018; Morales-Cruz et al. 2015; Nagel et al. 2021; Yu et al. 2022). Interestingly, it has been found that N. parvum possesses one of the highest numbers of CAZymes within the Botryosphaeriaceae family (Belair et al. 2023; Garcia et al. 2021; Nagel et al. 2021; Yu et al. 2022). Although numerous genomic studies have been conducted for N. parvum, only one has attempted to investigate the CAZyme-coding gene expression profile (Massonnet et al. 2018). This study reported 73 putative CAZymes involved in host cell wall degradation when in vitro fungal growth occurred with grapevine wood. Despite these advances, there is a paucity of information on how N. parvum adapts its transcriptome under different conditions. Further, it has been suggested that N. parvum, along with other Botryosphaeriaceae species, is a good candidate for producing biotechnologically relevant CAZymes (Belair et al. 2023; Esteves et al. 2014). Whether the N. parvum CAZymes can be applied in a biotechnological context and which conditions are conducive to their production still require substantial support.

In addition to the necrosis in the woody tissue of grapevine, the infection by N. parvum also causes foliar symptoms that are characterized by wine-red or yellowish-orange spots on the leaf margins or leaf blade (Larignon et al. 2001; Mondello et al. 2018). As this fungus has never been isolated from leaves, it is thought that foliar symptoms may be remotely induced (Claverie et al. 2020). One hypothesis is that N. parvum may produce phytotoxic compounds that can translocate to leaves via transpiration stream (Bertsch et al. 2013). Multiple studies have described and evaluated the toxic secondary metabolites produced by N. parvum (Abou-Mansour et al. 2015; Andolfi et al. 2011; Evidente et al. 2010; Ramírez-Suero et al. 2014; Reveglia et al. 2019; Salvatore et al. 2021; Trotel-Aziz et al. 2022). These secondary metabolites are derived from either polyketide precursors or nonribosomal peptides and can interfere with enzymatic reactions and cellular transport, as well as cause damage to the cell membrane in the host (Massonnet et al. 2018). Among the different secondary metabolites produced by N. parvum, terremutin and mellein have been considered representative of their respective chemical families, inducing severe necrosis when applied to grapevine leaf disks and in planta (Abou-Mansour et al. 2015; Trotel-Aziz et al. 2022). Moreover, those two compounds and some of their derivatives were detected in grapevine wood from plants showing Botryosphaeria dieback (Abou-Mansour et al. 2015). Besides low-molecular-weight toxins, the virulence of N. parvum might also rely on exopolysaccharides (Martos et al. 2008) and necrosis- and ethylene-inducing proteins (Nazar Pour et al. 2020).

In order to develop effective strategies to control N. parvum, it is relevant to understand the factors facilitating host colonization, such as the CWDEs and phytotoxic compounds. Here, we evaluated the capacity of N. parvum to produce CAZymes related to plant cell wall degradation and secondary metabolites when growing in liquid culture in contact with two lignocellulosic biomasses. We focused on the strain N. parvum Bt-67 (Estremadura, Portugal; Reis et al. 2016) due to its high aggressiveness and ability to produce and secrete enzymes involved in pathogenicity and production of phytotoxic metabolites (Abou-Mansour et al. 2015; Belair et al. 2023; Ramírez-Suero et al. 2014; Stempien et al. 2018, 2017; Trotel-Aziz et al. 2022). As lignocellulosic biomasses, we selected one woody host biomass (grapevine canes) and one non-woody biomass from a non-host (wheat straw), which differ in chemical composition, carbon content, and structure. We hypothesized that the production of CWDEs, related to the expression and the secretion of CAZymes, and secondary metabolites in N. parvum varies according to the lignocellulose biomass composition. In this article, we tested this hypothesis using a multiphasic study combining enzymology, transcriptomic, and metabolomic analyses.

MATERIALS AND METHODS

Fungal strain and cultivation.

—The strain N. parvum Bt-67 was obtained from vineyards of the Estremadura area (Portugal) and registered in the Instituto Superior de Agronomia (University of Lisbon, Portugal). The fungus was grown on potato dextrose broth (PDB; BD, Le Pont-de-Claix, France) with agar (BD) at 28 C in the dark for 7 days before cultivation.

A portion of N. parvum Bt-67 mycelium (0.5 × 1.0 cm) was cultured in 50 mL of liquid minimum medium EP (Eriksson 1975). Two lignocellulosic biomasses were used as a carbon source: grapevine canes (GP) (Vitis vinifera cv. Chardonnay, 2–3 mm; collected in Champagne vineyard after winter pruning) and wheat straw (WS) (Apache variety, <2 mm; obtained from ARD, Pomacle-Bazancourt, France), with a concentration of 5 g·L⁻¹. As control, a culture medium using only d-glucose (GLU; Carl Roth, Karlsruhe, Germany) at 10 mM as the carbon source was used. For each growth medium, control flasks were prepared without fungal inoculation. All flasks were incubated at 30 C with agitation at 100 rpm and in triplicates for 7–14 days. After cultivation, mycelia from each growth condition (i.e., EP medium with GP, WS, or GLU) were harvested and stored at −80 C before RNA and secondary metabolite extractions.

Chemical composition of the lignocellulosic biomasses.

—The chemical composition in hemicellulose, cellulose, lignin, and ash of both lignocellulosic biomasses was determined following the Van Soest method (Van Soest and Robertson 1980). This method consisted of sequential extractions using a neutral detergent (30 g sodium dodecyl sulfate, 20.61 g ethylenediaminetetraacetic disodium salt dihydrate, 6.81 g disodium tetraborate, 4.56 g disodium hydrogen phosphate anhydrous, and 10 mL ethylene glycol monoethyl ether per liter distilled water), an acid detergent (20 g cetyltrimethyl ammonium bromide and 26.63 mL sulfuric acid at 96% per liter distilled water), and sulfuric acid at 72%. The wheat straw chemical composition was previously determined by Cassarini et al. (2022).

Enzymatic activity assays.

—Extracellular enzymes (secretome) were obtained by collecting 5 mL of supernatant from each fungal culture at two time points, 7 (T7) and 14 (T14) days of incubation. Enzymatic activity assays were measured in triplicates. Xylanase activity was estimated by measuring reducing sugars produced from beechwood xylan (Carl Roth) at 0.5% w/v mixed with 0.1 mL of the secretome in Tris-HCl 25 mM, pH 7, at 30 C for 10 min (Kidby and Davidson 1973). β-d-Xylosidase, α-l-arabinofuranosidase, and β-d-glucosidase activities were evaluated by determining the hydrolysis rate of p-nitrophenyl-β-d-xylopyranoside (Carbosynth, Compton, UK), p-nitrophenyl-α-l-arabinofuranoside (Sigma, St. Louis, MI, USA), and p-nitrophenyl-β-d-glucopyranoside (Carbosynth), respectively, to p-nitrophenol (pNP). Each reaction was performed in 1 mL containing 0.1 mL of secretome, 0.45 mL of substrate (10 mM), and 0.45 mL of buffer (Tris-HCl 25 mM, pH 7) for 5 min at 30 C. The extinction coefficient of pNP was 10 625 L·mol⁻¹·cm⁻¹ (Cassarini et al. 2021). Total peroxidase activities were quantified as described by Bach et al. (2013). In brief, 0.1 mL of secretome was mixed with 0.1 mL of pyrogallol (PYGL; 160 mM) in ethylenediaminetetraacetic acid (EDTA; 20 mM), 0.1 mL of H₂O₂ (10 mM), and 1.7 mL of Tris-HCl 25 mM, pH 7, for a final reaction volume of 2 mL. For estimation of laccases activities, 0.1 mL of syringaldazine (SGZ; 0.5 mM in ethanol 96%) was mixed with 0.1 mL of secretome and buffer (Leonowicz and Grzywnowicz 1981). The reactions were at 30 C for 10 min. Endoglucanase activity was evaluated according to Rahman et al. (2018). A series of enzymatic dilutions were prepared to determine the enzyme concentration capable of releasing 0.5 mg of glucose. For this, 0.5 mL of carboxymethyl cellulose (CMC; Sigma, St. Louis, MI, USA) solution (2% w/v in Tris-HCl 25 mM, pH 7) was mixed with 0.5 mL of each enzymatic dilution and incubated at 30 C. After 30 min, 2 mL of 3,5-dinitrosalicylic acid (DNS) reagent was added and the reaction was stopped in boiling water for 5 min. The values from “enzyme blank” (without CMC) and “substrate blank” (without enzymatic dilutions) were subtracted from the absorbance of each sample. Absorbance for xylose and PYGL was measured at 420 nm. For pNP, SGZ, and glucose derived from CMC, the wavelengths used were 401, 530, and 540 nm, respectively. All the spectrophotometric measurements were performed using a SPECORD 200 PLUS spectrometer (Analytik Jena, Jena, Germany).

Enzymatic activities were expressed in units (IU) per milliliter, where 1 IU was defined as the quantity of enzyme required to release 1 µmol of equivalent xylose (for xylanase activity), 1 µmol of pNP (for β-d-xylosidase, α-l-arabinofuranosidase, and β-d-glucosidase activities), or 1 µmol of degraded PYGL or SGZ (for peroxidase/laccase activities) per minute. Endoglucanase activity was expressed as unit (IU) of CMCase per milliliter, where 1 IU corresponds to the amount of enzyme able to release 1 µmol of glucose equivalent per minute from the substrate.

Substrate relative degradation analyses by ATR-FTIR.

—After cultivation, the plant biomass (GP and WS) was separated from the culture medium by filtration and placed in 50-mL Falcon tubes. Two successive washes with 5 mL of Milli-Q water followed by centrifugation were performed. Agitation by either pumping or vortexing was necessary when the mycelium was attached to the biomass. The samples were dried in ovens at 50 C for at least 4 days. The Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy was used to analyze the chemical bonds in GP and WS before and after N. parvum Bt-67 cultivation. ATR-FTIR was performed using a Nicolet 6700 (Thermo Fisher Scientific, Madison, WI, USA) spectrometer and a UATR Two (PerkinElmer, Norwalk, CT, USA) measurement module. The measurements were performed in the wavenumber range of 4000–500 cm⁻¹ with a resolution of 4 cm⁻¹. Each sample was measured three times. A mean spectrum from the three readings was calculated, the baseline was corrected using OMNIC software, and the peak heights were analyzed using TQ Analyst EZ Edition. The area of interest (1800–800 cm⁻¹) was normalized. Ratios of absorbance of biomass in the absence (control) and presence of N. parvum Bt-67 were calculated. The measurements were performed from microbial growth triplicate assays.

Transcriptomic analysis.—

RNA extraction

Frozen mycelia from three replicates (7 days of growth on EP medium with GP, WS, or GLU) were used for RNA extraction. Total RNAs were extracted as described in Chang’s protocol, with minor modifications (Chang et al. 1993). For each sample, 200 mg of frozen tissue was ground to a fine powder in liquid nitrogen, lysed with 900 µL of cetyltrimethylammonium bromide (CTAB) extraction buffer, and then chloroform:isoamyl alcohol (24:1) was used to dissociate RNAs from proteins. After overnight precipitation with ¼ volume of 10 M LiCl, RNAs were collected and dissolved in RNase/DNase-free water. Then, total RNAs were purified with RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s recommendations. Treatment with RNase-free DNase I was applied during this purification step to ensure the elimination of genomic DNA. The quantity of RNA extracts was measured with a NanoDrop One spectrophotometer (Thermo Fisher, Madison, WI, USA), and quality was checked on Experion RNA StdSens Analysis Kit (Bio-Rad, Hercules, CA, USA), according to the manufacturer’s recommendations.

RNA-seq analysis

For the transcriptomic analysis, extracted total RNAs were sent to Novogene (Cambridge, UK). cDNA libraries were prepared and sequenced on an Illumina NovaSeq 6000 system (San Diego, CA, USA). The read quality was checked using FastQC (Andrews 2010). The high-quality reads were mapped to N. parvum UCRNP2 (Blanco-Ulate et al. 2013) using HISAT2 (Kim et al. 2015). The expression of transcripts was assessed by StringTie (Pertea et al. 2016). FPKM (fragments per kilobase of transcript sequence per million mapped reads) was used for estimating gene expression level. Gene Ontology (GO) analysis for significantly expressed transcripts (P-adjust < 0.05) was performed. First, the number of genes for each GO term was calculated by mapping all source genes in the GO database (http://www.geneontology.org/; Ashburner et al. 2000), then significantly enriched GO terms in source genes were estimated using Wallenius noncentral hypergeometric distribution. R package ballgown (Frazee et al. 2015) was used to determine differentially expressed genes. Genes with P < 0.05 (parametric F-test) and a log₂(fold change) > 2 or < −2 were considered to be differentially expressed. Annotation of differentially expressed genes (DEGs) coding for carbohydrate-active enzyme (CAZyme) families was performed according to the CAZy database (http://www.cazy.org/; Lombard et al. 2014) and by using the dbCAN2 server (Zhang et al. 2018). Venn diagrams were created with Venny 2.1 (https://bioinfogp.cnb.csic.es/tools/venny/index.html; Oliveros 2007).

Secondary metabolite analysis.

—Two hundred milligrams of frozen mycelia (7 days of growth on EP medium with GP, WS, or GLU) was freeze-dried for 24 h. Freeze-dried samples were extracted with 1 mL methanol (MeOH). After centrifugation (2 min at 13 200 rpm), 20 µL of supernatant was analyzed by high-performance liquid chromatography–mass spectrometry (HPLC-MS; series 1200; Agilent, Waldbronn, Germany) equipped with a ultraviolet diode array detector (UV-DAD) and a coupled liquid chromatograph/mass selective detection (LC/MSD) trap atmospheric pressure chemical ionization mass spectrometer with positive and negative polarization. Methods were applied as described previously (Buckel et al. 2017).

Secondary metabolite analysis was performed using MetaboAnalyst 5.0 (Pang et al. 2021). First, raw MS data were aligned using built-in XCMS program following the automatic parameter adjustment algorithm. No correction for missing values and no filtering were necessary (all signatures were initially kept). The metabolic signatures detected within the biomass conditions and undetected in the GLU condition were removed, resulting in a list of 564 metabolic signatures (see SUPPLEMENTARY TABLE 3) in order to avoid segregation of the conditions via biomass compounds. The list of 564 metabolic signatures was later submitted to sum normalization and Pareto scaling before further statistical analysis.

Statistical analysis.

—For the enzymatic activities, significant differences among conditions were determined by one-way analysis of variance (ANOVA; P < 0.05) at each time point. Tukey’s honestly significant difference (HSD) tests (P < 0.05) were performed for multiple comparisons of means. One-way ANOVA was also used to determine significant differences (P < 0.05) in the substrate degradation analysis between the absorbance from the lignocellulosic biomass with N. parvum growth and the control group (without fungal growth). The secondary metabolite statistical analysis was performed using MetaboAnalyst 5.0. To inspect sample groupings, a principal component analysis (PCA) was performed. Subsequently, a partial least squares discriminant analysis (PLS-DA) was used to enhance discrimination between groups of samples. Q² was determined to assess the predictability of the model.

Data availability.

—Raw transcriptome data have been deposited to the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA; https://www.ncbi.nlm.nih.gov/sra) under accession number PRJNA852183.

RESULTS

Chemical composition of grapevine canes and wheat straw.

—The grapevine material used in this study contained (in percentage of dry matter) 18.2 ± 0.7% of solubles (including pectins, gums, galactans, soluble proteins, etc.), 20.1 ± 0.8% of hemicellulose, 39.2 ± 0.7% of cellulose, 21.6 ± 0.8% of lignin, and 0.2 ± 0.1% of ash. On the other hand, wheat straw contained 10.4 ± 1.5% of solubles, 31.1 ± 1.2% of hemicellulose, 47.4 ± 0.2% of cellulose, 8.5 ± 0.9% of lignin, and 2.2 ± 0.2% of ash (Cassarini et al. 2022).

Enzymatic activities.

—Activities of extracellular enzymes (secretome) acting on xylans (xylanases, xylosidase, and arabinosidase), glucans (endoglucanases and glucosidase), and lignin (peroxidases and laccases) were evaluated. Not all the enzyme activities were observed. In particular, no activity was detected for peroxidases and laccases, whereas endoglucanase activities were <0.02 IU·mL⁻¹ when N. parvum grew with the lignocellulosic biomasses (WS and GP) and null with GLU.

Xylanase activities were higher when N. parvum Bt-67 grew in the presence of WS than GP (FIG. 1A). In both biomass growth conditions, the enzymatic activity was higher at T14. For WS, xylanase activity levels were 1.30 ± 0.49 and 1.64 ± 0.23 IU·mL⁻¹ at T7 and T14, respectively. In contrast, N. parvum Bt-67 showed a low xylanase activity when cultured with GP, with values of 0.18 ± 0.04 and 0.24 ± 0.03 IU·mL⁻¹. Secretome from the GLU condition did not present any measurable activity. The β-d-xylosidase activity was significantly higher in the WS condition than for the GP secretome at the two time points (FIG. 1B). In both lignocellulosic biomass conditions, the β-d-xylosidase activity increased over time. For T7 and T14, the activity levels for WS were 4.08 ± 1.22 and 10.04 ± 1.34 IU·mL⁻¹, respectively, whereas GP showed activity values of 0.89 ± 0.27 and 1.80 ± 0.37 IU·mL⁻¹ at the respective time points. No β-d-xylosidase activity was detected for the GLU condition. Like xylanase and β-d-xylosidase activities, α-l-arabinofuranosidase activity was higher in the case of WS than for the GP secretome, particularly at T14 (FIG. 1C). At T7, α-l-arabinofuranosidase activity reached 0.51 ± 0.65, 0.19 ± 0.24, and 0.21 ± 0.18 IU·mL⁻¹ for WS, GP, and GLU, respectively. No significant differences were found at this time point. At T14, α-l-arabinofuranosidase activities increased in the lignocellulosic biomass conditions but decreased in the GLU condition, showing 0.99 ± 0.13, 0.19 ± 0.04, and 0.07 ± 0.02 IU·mL⁻¹ for WS, GP, and GLU, respectively. It should be noted that the variability in all the conditions was high at T7; however, it turned to low values at T14. β-d-Glucosidase was the highest enzymatic activity in N. parvum Bt-67 regardless of the type of lignocellulosic biomass. This enzymatic activity was significantly higher in WS, showing 6.53 ± 0.81 and 16.21 ± 0.98 IU·mL⁻¹, respectively, for T7 and T14, whereas GP presented 4.84 ± 0.70 and 8.11 ± 0.54 IU·mL⁻¹, respectively (FIG. 1D). No β-d-glucosidase activity for the GLU condition was measured at T7, although a weak activity was detected at T14 (0.39 ± 0.36 IU·mL⁻¹).

Figure 1. Enzymatic activities of Neofusicoccum parvum Bt-67 secretomes when grown on minimal media with grapevine canes (GP), wheat straw (WS), or glucose (GLU) at 7 and 14 days after inoculation. (A) Xylanase, (B) β-d-xylosidase, (C) α-l-arabinofuranosidase, and (D) β-d-glucosidase activities (IU·mL⁻¹). Blue, red, and gray lines represent secretome activities from the GP, WS, and GLU conditions, respectively. All the measurements were performed in biological triplicates. Error bars indicate standard deviation. Different letters indicate significant differences among groups at each time point (Tukey’s HSD test; P < 0.05).

Evaluation of biomass degradation by ATR-FTIR.

—Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy analysis was performed in order to evaluate the chemical degradation of the lignocellulosic biomasses by the action of N. parvum Bt-67. TABLE 1 shows the chemical bonds at the characteristic peaks (wavenumber) as well as the polymer potentially represented. The degradation ratios for each lignocellulosic biomass (GP and WS) in the absence (control) and presence of N. parvum Bt-67 are shown in FIG. 2. Ratios below 1 represent a relative degradation of the carbon bond. Contrarily, ratios above 1 represent a relative enrichment of the carbon bond due to the degradation of other carbon bonds.

Figure 2. Degradation ratios of chemical bonds representative of grapevine canes and wheat straw determined by attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy at 14 days after inoculation. Ratios correspond to the value obtained from the lignocellulose biomass with fungal growth over lignocellulose biomass without fungal growth. Bond vibration: ν (stretching), δ (bending), s (symmetric), a (asymmetric). All the measurements were performed in biological triplicates. Error bars indicate standard deviation. Asterisk stands for chemical bonds significantly degraded or enriched after fungal growth (one-way ANOVA; P < 0.05; see also SUPPLEMENTARY FIGS. 1 and 2).

Table 1. Remarkable absorption bands in the attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra of wheat straw (WS) and grapevine canes (GP) (Adapa et al. 2009; Álvarez et al. 2020; Levasseur et al. 2013; Viel et al. 2018; Xu et al. 2013).

Wavenumber (cm^−1)	Chemical bonds^a/functions	Biomass	Associated polymers
1732	ν(C=O) characteristic of xylan	GP, WS	Hemicellulose
1600	va(COO^−)	GP	Hemicellulose, lignin
	v(C=C) aromatics
1510	v(C=C) aromatics	GP, WS	Lignin
	v(C-O) aromatics, aromatic cycles
1370	v(C-O), δa(C-H), δs(C-H)	GP, WS	Cellulose, lignin
1230–1245	v(C-O), v(C-C), δ(O-H)	GP	Polysaccharides, hemicelluloses, lignin
1100	va(C-O-C)	GP, WS	Cellulose
1030	v(C-O), v(C=O), v(C-C-O), v(C-C)	GP, WS	Polysaccharides
896	Glycosidic bonds β1-4, δ(C-H)	GP, WS	Cellulose, hemicellulose

^aBond vibration: ν (stretching), δ (bending), s (symmetric), a (asymmetric).

For WS, a significant degradation value was found for β1-4, bending C–H glycosidic bonds, representative of cellulose and hemicellulose (0.85 ± 0.03) (SUPPLEMENTARY FIG. 1B). In contrast, significant enrichment was found for stretching C=C and C–O bonds, representatives of aromatics (2.49 ± 0.18) (SUPPLEMENTARY FIG. 1F). The other ratios related to WS did not vary significantly. Regarding GP, degradation values were found for stretching C=O (hemicellulose), stretching C=C and C–O (lignin), stretching C–O, bending asymmetric C–H, and bending symmetric C–H (cellulose and lignin), stretching asymmetric C–O–C (cellulose), and stretching C–O, stretching C–C, and bending O–H (polysaccharides, hemicellulose, and lignin) bonds, with ratios of 0.79 ± 0.12, 0.91 ± 0.09, 0.84 ± 0.10, 0.90 ± 0.10, and 0.93 ± 0.05, respectively. Conversely, enrichment was observed for β1-4, bending C–H glycosidic bonds (cellulose and hemicellulose), showing a ratio of 1.10 ± 0.05. However, these values of degradation or enrichment between the groups with the presence of N. parvum Bt-67 and the absence of fungal growth were not significantly different (SUPPLEMENTARY FIG. 2).

Gene expression of N. parvum Bt-67 grown in presence of grapevine canes, wheat straw, and glucose.

—In total, 4156 genes were significantly expressed (P < 0.05). Based on these expressed genes, a heatmap together with hierarchical clusters was created (FIG. 3). This revealed that gene expression of N. parvum Bt-67 was segregated according to the carbon source (WS, GP, or GLU). Moreover, replicates of the same carbon source condition were clustered together, showing robustness in the experiment.

Figure 3. Heatmap and hierarchical clusters of significantly expressed genes of Neofusicoccum parvum Bt-67 grown with different carbon sources. GP = grapevine canes; WS = wheat straw; GLU = glucose. Numbers after carbon source treatments indicate biological repetitions. The scale from red to blue represents the log₁₀(FPKM+1) value, where red indicates high expression levels and blue low expression levels. FPKM = fragments per kilobase of transcript sequence per millions mapped reads.

Gene Ontology (GO) of significantly expressed transcripts

In order to correlate biological functions and metabolic pathways (GO categories) with the significantly expressed genes, an enrichment analysis was performed with the data sets from N. parvum Bt-67 grown with GP and WS, comparing each with the control (GLU condition) data sets. For induced genes in the GP condition, the most relevant enrichments were hydrolase activity acting on glycosyl bonds (GO:0016798) and carbohydrate metabolic process (GO:0005975) (FIG. 4A). Similarly, the most significant enrichment for the WS condition was hydrolase activity acting on glycosyl bonds (GO:0016798) (FIG. 4B). In addition, the WS condition also showed significant enrichment for ribosome (GO:0005840), structural constituent of ribosome (GO:0003735), and structural molecule activity (GO:0005198). In contrast, induced genes in the GLU condition showed enrichment for isomerase activity (GO:0016853), generation of precursor metabolites and energy (GO:0006091), kinase activity (GO:0016301), and nucleobase-containing compound catabolic process (GO:0034655), when compared with GP, and enrichment for nucleobase-containing compound catabolic process (GO:0034655) and generation of precursor metabolites and energy (GO:0006091), when compared with WS.

Figure 4. Gene ontology (GO) classifications for up-regulated (Up) and down-regulated (Down) genes and metabolic pathways in Neofusicoccum parvum Bt-67 when grown in presence of (A) grapevine canes (GP) or (B) wheat straw (WS). Comparisons were performed against gene expression of growth with glucose (GLU). Asterisk denotes statistical significance (P-adjust < 0.05). Significance was calculated using Wallenius noncentral hypergeometric distribution.

Differentially expressed genes (DEGs)

Following the identification of the biological function for significantly expressed genes, a differentially expressed gene (DEG) analysis was performed, where the growth in presence of glucose was used as the reference. As result, 1302 genes were identified as DEGs (P < 0.05; log₂(fold change) > 2 or < −2) for GP (493 up-regulated and 809 down-regulated) and 1282 for WS (455 up-regulated and 827 down-regulated) with respect to reference (SUPPLEMENTARY FIG. 3). Out of the up-regulated genes, 222 and 184 were specific for GP and WS, respectively, whereas 356 and 374 genes were specifically down-regulated for GP and WS, respectively (FIG. 5).

Figure 5. Venn diagrams of the differentially expressed genes (DEGs) in Neofusicoccum parvum Bt-67 grown in presence of different lignocellulosic biomasses. GP = grapevine canes; WS = wheat straw. Growth with glucose was used as control. The number in each circle indicates the number of DEGs for each condition. The overlapping number indicates the number of DEGs in both conditions.

Among the up-regulated genes in the GP condition, the top five encoded for three unknown proteins (fold change = 17.09, 10.67, and 9.80 for ranks 1, 3, and 5, respectively), an endo-xylogalacturonan hydrolase A (Q9UUZ2; fold change = 11.75), and a putative glycoside hydrolase active on cellulose (Q00023; fold change = 10.14) (SUPPLEMENTARY TABLE 1). For up-regulated genes in the WS condition, the top five was represented by two unknown proteins (fold change = 11.18 and 9.82 for ranks 2 and 5, respectively), two previously highly expressed proteins in GP, which are an endo-xylogalacturonan hydrolase A (Q9UUZ2; fold change = 12.52) and a putative glycoside hydrolase active on cellulose (Q00023; fold change = 10.74), and a rhamnogalacturonan acetylesterase (Q00017; fold change = 10.07). Notably, these genes were related to carbohydrate-active enzymes (see Differentially expressed CAZymes). In contrast, out of the top five down-regulated genes in GP, it was only possible to annotate one gene, which encodes for a zinc-regulated transporter (O94639; fold change = −11.67). For WS, the top five down-regulated genes encoded for two unknown proteins (fold change = −13.85 and −12.59 for ranks 2 and 5, respectively), an alpha/beta (AB) hydrolase superfamily protein C1039.03 (Q9US38; fold change = −16.80), a cell wall protein ECM33 (fold change = −11.81), and an E3 ubiquitin protein ligase ZNRF3 (fold change = −11.33). None of these down-regulated genes were related to the CAZy classification.

Differentially expressed CAZymes

Out of 493 up-regulated genes in the GP condition, 115 genes encoded for CAZymes, including 20 auxiliary activities (AAs), 12 carbohydrate esterases (CEs), 70 glycoside hydrolases (GHs), one glycosyltransferase (GT), and 12 polysaccharide lyases (PLs) (SUPPLEMENTARY TABLE 2). Among these 115 CAZyme genes, 79 presented signal peptides, suggesting that these enzymes may be secreted. The most represented families were GH5 and AA9 with 9 genes each, followed by GH43 and GH3 with 8 and 6 genes, respectively. Some of these genes were also associated with carbohydrate-binding modules (CBMs), for example, one GH5_4 was linked to a CBM46 (gene ID = UCRNP2_7122). For the WS condition, 113 up-regulated genes encoded for CAZymes, including 20 AAs, 10 CEs, 69 GHs, one GT, and 13 PLs. Among them, 74 had signal peptides. AA9 and GH3 were the most represented CAZymes families with 9 genes each, followed by GH43 and GH5 with 7 genes each. In total, 94 up-regulated CAZymes were expressed under both conditions; moreover, 21 and 19 up-regulated CAZyme genes were specifically expressed in GP and WS conditions, respectively (FIG. 6; TABLE 2).

Figure 6. Venn diagram of up-regulated genes encoding carbohydrate-active enzymes (CAZymes) in Neofusicoccum parvum Bt-67 grown in presence of different lignocellulosic biomasses. GP = grapevine canes; WS = wheat straw. Growth with glucose was used as control. The number in each circle indicates the number of up-regulated CAZyme genes for each condition. The overlapping number indicates the number of up-regulated CAZyme genes in both conditions.

Table 2. Differentially expressed CAZymes (P < 0.05 and log₂(fold change) > 2) of Neofusicoccum parvum Bt-67 during growth on minimal media containing grapevine canes (GP) or wheat straw (WS) relative to glucose (GLU).

Gene ID	CAZyme family^a	SP^b	Fold rate^c		UniProt ID	Predicted activity
			GP	WS
In GP condition
UCRNP2_10191	AA1_3	Y	2.6526	ND	Q96WM9	Laccase-2
UCRNP2_867	AA2	Y	3.0890	ND	PF00141	Peroxidase
UCRNP2_4400	AA3_2	Y	2.6204	ND	PF00732	Dehydrogenase
UCRNP2_4996	AA3_2	Y	5.8405	ND	PF05199	Dehydrogenase
UCRNP2_8476	CE2	Y	4.0472	ND	R1EAH5	Putative GDSL-like lipase acylhydrolase domain protein
UCRNP2_6889	CE12	Y	8.5533	ND	PF13472	GDSL-like lipase acylhydrolase protein
UCRNP2_1987	GH5_15	Y	6.0452	ND	Q0C8Z0	Probable glucan endo-1,6-β-glucosidase B
UCRNP2_5425	GH13_1	Y	3.2991	ND	Q02906	α-Amylase B
UCRNP2_691	GH15 + CBM20	Y	4.1470	ND	P69327	Glucoamylase
UCRNP2_9853	GH43_36	Y	2.7106	ND	—	α-l-Arabinofuranosidase
UCRNP2_8221	GH72 + CBM43	Y	2.5559	ND	D4ATC3	1,3-β-Glucanosyltransferase ARB_07487
UCRNP2_8718	GH106	Y	2.1084	ND	—	Unknown
UCRNP2_2665	GH131	Y	2.4279	ND	—	Broad-specificity exo-β-1,3/1,6-glucanase with endo-β-1,4-glucanase activity
UCRNP2_5778	CE9	N	2.2966	ND	Q8JZV7	N-acetylglucosamine-6-phosphate deacetylase
UCRNP2_8897	GH5_5	N	2.1609	ND	P07982	Endoglucanase EG-II
UCRNP2_7598	GH13_40	N	3.1361	ND	Q02751	α-Glucosidase
UCRNP2_4873	GH16_23	N	2.3045	ND	PF00722	Unknown activity
UCRNP2_5421	GH29	N	5.1080	ND	PF01120	α-l-Fucosidase
UCRNP2_2255	GH92	N	2.4344	ND	D4ATR3	Uncharacterized secreted glycosidase ARB_07629
UCRNP2_1979	GH125	N	3.0299	ND	—	Exo-α-1,6-mannosidase
UCRNP2_637	PL4_5	N	3.3754	ND	R1GLK8	Putative rhamnogalacturonate lyase protein
In WS condition
UCRNP2_6999	AA3_2	Y	ND	2.4160	PF05199	Dehydrogenase
UCRNP2_3571	GH3	Y	ND	7.0093	B0Y7Q8	Probable β-glucosidase F
UCRNP2_6948	GH3	Y	ND	3.9793	Q0CB82	Probable exo-1,4-beta-xylosidase bxlB
UCRNP2_8351	GH13_1	Y	ND	2.5415	P19269	α-Amylase 1
UCRNP2_6952	GH18	Y	ND	6.2977	D4B4X4	Class III chitinase ARB_03514
UCRNP2_8371	GH28	Y	ND	5.5763	A2RAY7	Putative galacturan 1,4-α-galacturonidase C
UCRNP2_156	GH31	Y	ND	4.7045	—	Unknown
UCRNP2_9371	PL4_5	Y	ND	2.7530	Q8RJP2	Rhamnogalacturonate lyase
UCRNP2_731	AA3_2	N	ND	2.3830	Q92452	Glucose oxidase
UCRNP2_8536	AA3_2	N	ND	4.8413	PF00732	Dehydrogenase
UCRNP2_4898	AA7	N	ND	3.8692	R1GQT1	Putative FAD-binding domain-containing protein
UCRNP2_6659	CE15	N	ND	8.3551	G0RV93	4-O-methyl-glucuronoyl methylesterase
UCRNP2_4352	GH3	N	ND	2.8462	Q5BFG8	β-Glucosidase B
UCRNP2_7888	GH16_4	N	ND	2.1641	Q8N0N3	β-1,3-Glucan-binding protein
UCRNP2_3072	GH31	N	ND	2.9419	P31434	α-Xylosidase
UCRNP2_51	GH31	N	ND	2.8994	A2QTU5	α-Xylosidase A
UCRNP2_5589	GH31	N	ND	2.2300	Q01336	Uncharacterized family 31 glucosidase
UCRNP2_348	GH154	N	ND	2.0518	R1GWQ6	Uncharacterized protein
UCRNP2_1907	PL1_4	N	ND	3.9401	Q5BA61	Pectin lyase B

^aAA = auxiliary activity; CBM = carbohydrate-binding module; CE = carbohydrate esterase; GH = glycoside hydrolase; PL = polysaccharide lyase.

^bSP = signal peptide (Y = yes; N = no).

^cND indicates genes that were not differentially expressed in a given condition.

Interestingly, under the two growth conditions (GP and WS), the four CAZymes families with the highest representation were the same, i.e., AA9, GH3, GH5, and GH43. These groups comprise cellulolytic (AA9s, GH3s, and GH5s) and hemicellulolytic (GH43s, but also some GH3s and GH5s) enzymes. The AA9 proteins are copper-dependent lytic polysaccharide monooxygenases (LPMOs). Almost all of the GH3s were annotated as β-glucosidase; in addition, a GH3 exo-1,4-β-xylosidase was only up-regulated in WS. GH5s comprise mainly endo-β-1,4-glucanases. In the GP condition, an additional putative glucan endo-1,6-β-glucosidase (GH5 subfamily 15) was identified. For GH43 enzymes, we identified a xylosidase/arabinosidase, an α-l-arabinofuranosidase, two exo-β-1,3-galactanases, one of which having a CBM family 32, a putative arabinan endo-1,5-α-l-arabinosidase, a putative exo-α-l-arabinofuranosidase, and an unspecific GH43 subfamily 29 enzyme. Moreover, an α-l-arabinofuranosidase (GH43 subfamily 36) was only expressed in the GP condition. Additional cellulolytic enzymes, such as cellobiohydrolases (GH6 and 7), and hemicellulolytic enzymes, such as xylanases (GH10, 11, 30), xyloglucan-specific endo-β-1,4-glucanase (GH12), β-mannosidase (GH2), acetylxylan esterases (CE1 and 5), and others, were found.

Regarding lignin-modifying enzymes, a laccase (AA1), two peroxidases (AA2), and two dehydrogenases (AA3) were differentially up-regulated for the GP condition (TABLE 2). In presence of WS, only three AA3 enzymes were differentially up-regulated. In addition, one peroxidase (AA2) and two dehydrogenases (AA3) were up-regulated in both biomass growth conditions. Furthermore, we also found up-regulation of pectinases, such as endo- and exo-polygalacturonases (GH28s), and different PLs (PL1, 3, 4, and 9). Interestingly, one endo-xylogalacturonan hydrolase (UCRNP2_438; GH28) was the most expressed CAZyme in both GP and WS, with fold rates of 11.7521 and 12.5253, respectively.

Secondary metabolite production.

—In total, 669 metabolites were detected among the three growth conditions. The metabolites associated with the lignocellulosic biomasses (e.g., ε-viniferin and vitisin C for GP) and undetected in the GLU condition were removed, resulting in a list of 564 compounds (SUPPLEMENTARY TABLE 3). Thirty-five out of 564 metabolites were only present when the fungus grew with the lignocellulosic biomasses, four were unique for the GLU condition, and 510 were produced in the three conditions (FIG. 7A). Based on the metabolite peak areas, a principal component analysis (PCA) and a partial least squares discriminant analysis (PLS-DA) were performed. In the PCA, the first two components explained 65.4% of the variance, where PC1 accounted for 44.6% and PC2 for 20.8% (FIG. 8A). In the PLS-DA, the first two components explained 56.8% of the variance, where component 1 accounted for 44.2% and the component 2 for 12.6% (FIG. 8B). The Q² was 0.75, indicating a good prediction ability. These analyses showed that the samples segregated according to the carbon source (GP, WS, or GLU).

Figure 7. Venn diagrams of metabolites produced by Neofusicoccum parvum Bt-67 grown in minimal media with grapevine canes (GP), wheat straw (WS), or glucose (GLU) as carbon sources. A. Total number of metabolites in the three conditions. B–D. The number of significantly different metabolites between each pair of conditions: GP vs. WS, GP vs. GLU, and GLU vs. WS. Significantly different metabolites were determined by one-way ANOVA (FDR < 0.05). Overlapping numbers indicate metabolites commonly produced in the compared conditions.

Figure 8. (A) Principal component analysis (PCA) and (B) partial least squares discriminant analysis (PLS-DA) based on the peak area of 564 metabolites produced by Neofusicoccum parvum Bt-67 grown in minimal media with different carbon sources: grapevine canes (blue), wheat straw (red), or glucose (gray). The percentage of variation of each component is indicated at each axis. Ellipses represent the 95% confidence interval. Q² for the PLS-DA was 0.75.

To determine which metabolites were the most discriminant for the growth conditions, one-way ANOVA was performed. As result, 166 out of 564 compounds were significantly different (FDR < 0.05; SUPPLEMENTARY TABLE 3). In addition, a heatmap was created with the top 50 most significant metabolites (SUPPLEMENTARY FIG. 4). This heatmap confirmed the results obtained by PCA and PLS-DA, showing that the biological replicates were grouped. The two most discriminant compounds (FDR < 0.05) were identified as bisnorcholic acid (m/z = 381; C₂₂H₃₆O₅), followed by 10′-apo-β-caroten-10′-al (m/z = 415; C₂₇H₃₆O), an unknown compound (m/z = 521), and 2-hydroxy-4,7-dimethoxy-2H-1,4-benzoxazin-3(4H)-one (HDMBOA) (m/z = 226; C₁₀H₁₁NO₅) (TABLE 3).

Table 3. The five most discriminant metabolites (FDR < 0.05) produced by Neofusicoccum parvum Bt-67 during growth on minimal media containing grapevine canes, wheat straw, or glucose.

m/z	rt^a	Formula	Compound	FDR^b
381.2610	773.41	C22H36O5	Bisnorcholic acid	8.05 × 10^−9
381.2606	680.01	C22H36O5	Bisnorcholic acid	1.83 × 10^−4
415.2459	642.30	C27H36O	10′-Apo-β-caroten-10′-al	4.86 × 10^−4
521.4390	1411.98	—	Unknown	4.86 × 10^−4
226.0680	106.69	C10H11NO5	2-Hydroxy-4,7-dimethoxy-2H-1,4-benzoxazin-3(4H)-one (HDMBOA)	9.87 × 10^−4

^art = retention time (s).

^bOne-way ANOVA was used to determine discriminant metabolic compounds with a false discovery rate (FDR) < 0.05.

One-way ANOVA was performed for each pair of conditions (i.e., GP-WS, GP-GLU, and WS-GLU). Comparison between GP and WS showed 58 metabolites significantly different (FDR < 0.05), of which two were specifically produced in the GP condition and five in the WS condition (FIG. 7B). Between GP and GLU, 58 compounds presented significant differences, in which 14 were specific for GP and two for GLU (FIG. 7C). Between GLU and WS, 47 compounds presented significant differences, of which one was specific for GLU and three for WS (FIG. 7D). Subsequently, differentially produced metabolites were determined by considering a log₂(fold change) > 2 or < −2 and an FDR < 0.05 (SUPPLEMENTARY FIG. 5). In the comparison between GP and WS, 30 metabolites were more abundant in the GP condition, whereas 17 were highly produced in the WS condition. For GP vs. GLU, 40 metabolites were more abundant in the GP, whereas 16 were highly produced in the GLU condition. For WS vs. GLU, 27 metabolites were more abundant in the WS condition, whereas 19 were highly produced in the GLU condition.

TABLE 4 displays the top 10 differentially produced compounds in each growth condition. The complete list of differentially produced metabolites is shown in SUPPLEMENTARY TABLE 4.

Table 4. The top 10 differentially metabolites produced by Neofusicoccum parvum Bt-67 in comparison of each pair of carbon source conditions: grapevine canes (GP), wheat straw (WS), and glucose (GLU).

m/z	rt^a	Fold change^b		Formula	Compound
(1) In GP condition		WS	GLU
454.1362	512.09	9.4921	10.966		Unknown
718.4423	1388.49	8.6027	6.9422		Unknown
820.5264	1313.31	5.0373	8.5575	C44H80NO8P	PC(14:0/22:4(7Z,10Z,13Z,16Z))
757.559	1388.49	3.8596	6.7214	C43H81O8P	PA(16:1(9Z)/24:1(15Z))
138.0526	132.17	3.8138	2.9678	C7H7NO2	Trigonelline
682.2159	612.42	3.4865	6.7989	C30H39N3O13S	2-Amino-4-({1-[(carboxymethyl)-C-hydroxycarbonimidoyl]-2-[(4-{3-[2-(2,4-dihydroxyphenyl)-2-oxoethyl]-4,6-dihydroxy-2-methoxyphenyl}-3-hydroxy-2-methylbutan-2-yl)sulfanyl]ethyl}-C-hydroxycarbonimidoyl)butanoic acid
470.3698	1260.55	3.4531	5.1569		Unknown
298.1395	230.83	3.2927	4.8571	C18H19NO3	Oripavine
435.3588	1444.93	2.9832	7.0895		Unknown
861.2926	674.00	2.9827	6.0082		Unknown
(2) In WS condition		GP	GLU
381.2606	680.01	11.919	9.9015	C22H36O5	Bisnorcholic acid
399.2714	642.3	11.534	10.266	C27H36O	10’-Apo-beta-caroten-10’-al
400.2746	642.3	10.8900	9.3974		Unknown
297.1667	439.93	9.6958	7.6784	C16H24O5	Ovalicin
393.2243	584.5	9.6944	6.0614	C22H32O6	Dihydrofukinolide
381.261	773.41	8.2699	6.8772	C22H36O5	Bisnorcholic acid
393.2236	534.09	7.7364	6.2441	C22H32O6	Dihydrofukinolide
415.2459	642.3	7.6843	7.8798	C27H36O	10’-Apo-beta-caroten-10’-al
308.1548	530.05	5.9883	6.3070	C16H21NO5	5-Amino-2,3-dihydro-6-(3-hydroxy-4-methoxy-1-oxobutyl)-2,2-dimethyl-4H-1-benzopyran-4-one
337.1985	556.46	4.9602	6.2514		Unknown
(3) In GLU condition		GP	WS
671.4858	1473.09	3.9730	4.9128		Unknown
267.1336	328.05	3.9660	7.2704	C13H18N2O4	Phenylalanylthreonine
237.1229	285.37	3.9647	4.4048	C12H16N2O3	Hexobarbital
203.1386	126.45	3.7900	4.4604	C9H18N2O3	Alanylisoleucine
231.1695	301.65	3.2066	3.3383	C11H22N2O3	Isoleucylvaline
322.1865	280.67	3.1197	2.7628	C15H23N5O3	Arginylphenylalanine
219.1332	194.68	2.9623	3.1345	C9H18N2O4	Meprobamate
351.2503	1024.82	2.6804	3.5381	C21H34O4	Tetrahydrocorticosterone
203.1385	242.54	2.2995	3.0112	C9H18N2O3	Alanylisoleucine
277.2159	1333.35	2.2722	5.4675	C18H28O2	19-Norandrosterone

^art = retention time (s).

^bFold change of the most differentially produced metabolites in (1) GP condition compared with WS and GLU, (2) WS condition compared with GP and GLU, and (3) GLU condition compared with GP and WS.

DISCUSSION

Neofusicoccum parvum Bt-67 produces a broad set of CAZymes when grown in presence of lignocellulosic biomass.

—In this work, we aimed to evaluate the capacity of N. parvum Bt-67 in producing CAZymes during growth on woody host (grapevine canes [GP]) and non-woody non-host substrates (wheat straw [WS]). Our results showed that this pathogenic fungus could express a wide variety of CAZymes. These CAZymes mainly targeted cellulose and hemicellulose; however, a significant number of pectinases were also identified. This is in line with Massonnet et al. (2018), who observed that the presence of ground wood from grapevine (cv. Cabernet Sauvignon or cv. Merlot) in agar culture medium induced the expression of CAZymes in N. parvum isolate UCD646So. These authors identified 73 putative CAZymes involved in host cell wall degradation, mainly hemicellulases and pectinases. Here, we present 115 highly expressed CAZymes when growth occurred in presence of grapevine, in which CAZymes targeting hemicelluloses were the most abundant, followed by cellulolytic enzymes (TABLE 2; SUPPLEMENTARY TABLE 2). Besides lignocellulolytic enzymes expressed with the host material (GP), we also showed that N. parvum has the potential to produce CAZymes in presence of non-host material, such as wheat straw.

Our transcriptomic analysis revealed that similar sets of CAZymes were expressed when N. parvum Bt-67 grew in presence of GP and WS. A similar phenomenon has been observed in other wood decay fungi during growth on different substrates (Gaskell et al. 2016; MacDonald et al. 2011; Van den Wymelenberg et al. 2011; Wu et al. 2021). Notably, the most abundant up-regulated CAZymes families in N. parvum Bt-67 when grown on either GP or WS included LPMOs (AA9s), β-glucosidases (GH3s), endoglucanases (GH5s), and different GH43 enzymes, such as xylosidase and arabinosidase. These results match with genomic analyses, which indicated that those CAZymes families are the most abundant in N. parvum (Belair et al. 2023; Massonnet et al. 2018; Morales-Cruz et al. 2015). Significant expression of GH3, GH5, and AA9 enzymes has also been observed in wood decay Basidiomycetes (Eastwood et al. 2011; Umezawa et al. 2020). The predominance of GH5 endoglucanases, GH3 β-glucosidases, and AA9 LPMOs may lie in their ability to degrade cellulose, the main polymer in the plant cell wall. These enzymes, together with cellobiohydrolases (GH6s and GH7s), act synergistically on cellulose to raise d-glucose (Kubicek and Kubicek 2016; Saldarriaga-Hernández et al. 2020). It is noteworthy that three cellobiohydrolases (one GH6 and two GH7s) were also highly expressed in N. parvum Bt-67 in both lignocellulosic biomass growth conditions.

In order to successfully colonize plant tissue, pathogens must be able to degrade hemicellulose, which contributes to cell wall rigidity and, therefore, plant resistance (Miedes et al. 2014). Here, most of the CAZymes in N. parvum Bt-67 transcriptome were hemicellulolytic enzymes, including xylanases (GH10 and 11), xylosidases (GH3 and 43), arabinosidases (GH43, 62, and 146), acetylxylan esterases (CE1 and 5), and others. These enzymes act cooperatively in hemicellulose degradation. Interestingly, some hemicellulases have been reported as virulence factors in fungal pathogens (Brito et al. 2006; Nafisi et al. 2014; Nguyen et al. 2011; Yajima et al. 2009). However, different studies failed to demonstrate the role of these enzymes as virulence factors (Apel-Birkhold and Walton 1996; Gómez-Gómez et al. 2002; Sella et al. 2013; Wegener et al. 1999; Wu et al. 2006). Thus, the role of hemicellulases in pathogenicity, as well as other CWDEs in most plant pathogens, still needs to be demonstrated.

Herein, we found several pectin-degrading enzymes in the N. parvum Bt-67 transcriptome, including polygalacturonases (GH28) and different PLs (PL1, 3, 4, and 9). Remarkably, the highest expressed CAZymes in N. parvum Bt-67 when grown on both biomass growth conditions was a GH28 endo-xylogalacturonan hydrolase (UCRNP2_438). This is in line with Nazar Pour et al. (2022), who found that the most produced enzymes were pectinases when N. parvum grew in presence of Eucalyptus stem, compared with the control condition without plant biomass. Interestingly, GH28 polygalacturonases have been described as virulence factors, being the first enzymes to be secreted when phytopathogens encounter plant cell walls (Kubicek et al. 2014). Although the lignocellulosic biomass may present poor content of pectins, these latter have a significant presence in the primary cell wall, contributing to plant resistance against pathogens (Miedes et al. 2014).

Production of CAZymes in N. parvum Bt-67 is substrate dependent.

—Although most of the DEG-encoding CAZymes were expressed in both lignocellulosic biomass conditions (94 out of 134), there were differences in CAZyme production, where 21 CAZymes were differentially up-regulated in the GP condition and 19 in WS. Moreover, most of the expression levels of up-regulated CAZymes in both lignocellulosic biomass conditions were higher in the WS condition, especially for cellulases and hemicellulases. These differences in abundance and expression levels of CAZymes produced by N. parvum Bt-67 were correlated with the enzymatic activity assays (FIG. 1). Here, all the evaluated hemicellulases (i.e., xylanase, β-d-xylosidase, and α-l-arabinofuranosidase), as well as β-d-glucosidase, showed higher activity levels in the WS condition. Interestingly, expression levels of xylanases (GH10 and GH11), β-xylosidase (GH3 and GH43), α-l-arabinofuranosidase (GH43 and GH62), and β-glucosidase (GH3) were generally greater when N. parvum Bt-67 grew on WS (SUPPLEMENTARY TABLE 2).

These differences in CAZyme production in our growing conditions, particularly for the (hemi-)cellulolytic enzymes, may be explained by the fact that wheat straw contains a higher amount of cellulose and hemicellulose than grapevine canes (47.4 ± 0.2% vs. 40.4 ± 1.48% and 31.1 ± 1.2% vs. 19.2 ± 1.10%, respectively). Similar explanations have been indicated for the main fungal producer of cellulase and hemicellulase cocktails, Trichoderma reesei, when comparing the CAZyme production during growth on contrasting substrates, including wheat straw (Adav et al. 2012; Häkkinen et al. 2012). Besides the polysaccharide composition in the different substrates, the structure of hemicellulose may directly influence the fungal CAZyme production. For example, Häkkinen et al. (2012) provided evidence that the induction of CAZyme genes was dependent on the xylan type. These authors also pointed out that the side chains of hemicellulose may play a role in the induction process.

Apart from the cellulose and hemicellulose contents, lignin may influence the CAZyme production since this polymer acts as a barrier in the hydrolysis process. The lignin content in grapevine canes was much higher than in wheat straw (21.4 ± 0.64% vs. 8.5 ± 0.9%). This is in line with previous studies in other wood decay fungi, which reported lower expression and enzymatic activity levels of (hemi-)cellulolytic enzymes when the cultures occurred in presence of substrates with high lignin content (Metreveli et al. 2021; Mukherjee and Khowala 2016; Rajesh Banu et al. 2021; Rytioja et al. 2017). Interestingly, on the one hand, we found that when N. parvum Bt-67 grew on the biomass with high lignin content (GP), one AA1 laccase and one AA2 peroxidase, which might be involved in lignin modification, were up-regulated. On the other hand, the fungus in the WS condition only up-regulated one AA2 peroxidase, which was also up-regulated in presence of GP. It should be noted that the expression level of this peroxidase was higher in the GP condition. This is consistent with the higher production of lignin-modifying enzymes of woody-decay fungi in cultures with different wood substrates, compared with growths in non-woody substrates (Fernández-Fueyo et al. 2016; Rytioja et al. 2017).

Secreted CAZymes by N. parvum Bt-67 degrade the lignocellulosic polymers.

—The degradation of the lignocellulosic components in GP and WS caused by the N. parvum Bt-67 secreted CAZymes was confirmed by the FTIR analysis (FIG. 2). These results are consistent with the transcriptomic analysis and enzymatic activity assays, which demonstrated the cellulolytic and hemicellulolytic enzyme production in both lignocellulosic biomass conditions. In addition, degradation values for lignin were only observed in GP, which correlates with the higher abundance and expression levels of lignin-modifying enzymes (i.e., laccases and peroxidases) when the pathogen grew in presence of this biomass. Nevertheless, degradation values of lignin were not significant (FIG. 2; SUPPLEMENTARY FIG. 2), which accords with the non-detection of laccase/peroxidase activities.

Secondary metabolite production is substrate dependent.

—We detected 564 metabolites produced by the fungus in the three growth conditions (GP, WS, and GLU). We found a higher number of metabolites when the fungus grew with plant biomass (GP or WS) than GLU as the sole carbon source. These findings are not surprising since glucose, which is rapidly metabolized, interferes with the biosynthesis of many secondary metabolites (Demain 1986). This interference may be partially due to the carbon catabolite repression (CCR) mechanism in fungi (Brakhage 2013). Contrarily to glucose, polysaccharides, oligosaccharides, and oils may favor metabolite production (Demain 1986).

PCA and PLS-DA of the metabolites produced by N. parvum Bt-67 allowed the distinction of the growth conditions (FIG. 8). The most discriminant metabolite (FDR < 0.05) was bisnorcholic acid, which belongs to the class of trihydroxy bile acids (Kuramoto et al. 1990). Bile acids are steroid compounds with relevant biological functions in mammals, such as solubilization and adsorption of fats, cholesterol, and lipid-soluble vitamins (Kollerov et al. 2016). In fungi, bile acids have only been detected in Penicillium sp. (Ohashi et al. 2008). In addition, filamentous fungi can transform bile acids (Kollerov et al. 2016). Although bile acids might have similar functions in fungi as those described in animals (e.g., solubilization of ergosterol, the main constituent of the fungal plasma membrane), their exact role is unknown. Following bisnorcholic acid, 10′-apo-β-caroten-10′-al was one of the most discriminant compounds. This metabolite is an apo carotenoid produced by the cleavage of β-carotene at the 10′ position (Koschmieder et al. 2021). Besides being essential pigments for plants (Koschmieder et al. 2021), carotenoids can be produced by fungi (Avalos and Carmen Limón 2015). In the latter, the main function of carotenoids is associated with the protection of cells against adverse conditions, such as oxidative stress caused by reactive oxygen species (ROS) and radiation (Avalos and Carmen Limón 2015).

Another relevant compound was 2-hydroxy-4,7-dimethoxy-2H-1,4-benzoxazin-3(4H)-one (HDMBOA) (TABLE 3), which is a benzoxazine found in plants from different families, including Poaceae, Acanthaceae, Ranunculaceae, and Scrophulariaceae (Escobar et al. 1997). Benzoxazinones have inhibitory effects on plant growth, as well as antimicrobial and insecticide activities (Macías et al. 2005). Benzoxazinoids are considered traditional phytoanticipins (Pedras and Yaya 2015). Here, HDMBOA was detected in all the replicates of both biomass growth conditions and two out of three replicates in GLU (SUPPLEMENTARY TABLE 3). A possible explanation for this might be that N. parvum Bt-67 has obtained the capacity to produce this plant-derived compound through the so-called horizontal gene transfer (HGT). It has been hypothesized that HGT can take place during interactions between plants and their endophytic counterpart or vice versa (Nicoletti and Fiorentino 2015; Salvatore et al. 2021). Since N. parvum can have an endophytic lifestyle, HGT might occur. Indeed, the production of plant metabolites by Botryosphaeriaceae has been reported previously. For example, Shi et al. (2012) found that grapevine endophytes, including one Botryosphaeria strain, were able to produce resveratrol, a stilbene phytoalexin produced by plants in response to stress factors. Terpenoid compounds from plants have been extracted from Lasiodiplodia theobromae (Zaher et al. 2015) and Botryosphaeria sp. (Yuan et al. 2009). Interestingly, we observed that N. parvum Bt-67 was able to differentially produce a plant terpenoid compound recognized as dihydrofukinolide (TABLE 4), a sesquiterpene lactone metabolite identified in Petasites japonicus (Naya et al. 1972) and Glycine max (Gupta et al. 2018).

The discriminant compounds described above (TABLE 3) were highly produced in the WS condition. The presence of WS also influenced the production of ovalicin (TABLE 4), which is a sesquiterpene methyl ether with antibiotic properties (Cane and Levin 1976). Structurally, ovalicin is closely related to fumagillin, a well-known antibiotic metabolite produced by Aspergillus fumigatus (Cane and Levin 1976). Ovalicin, as well as fumagillin, targets the enzyme methionine aminopeptidase (type 2), which plays an essential role in posttranscriptional modifications of proteins (Griffith et al. 1997; Ross et al. 2005). Another metabolite positively affected by WS was 5-amino-2,3-dihydro-6-(3-hydroxy-4-methoxy-1-oxobutyl)-2,2-dimethyl-4H-1-benzopyran-4-one, a benzopyran analogous to the mycotoxin fusarochromanone found in Fusarium equiseti (Xie et al. 1990). This latter compound inhibits cell proliferation and reduces cell viability by arresting cells at the G₀/G₁ phase of the cell cycle (Gu et al. 2014).

It should be noted that the diversity of differentially produced metabolites was higher in presence of GP (SUPPLEMENTARY FIG. 5; SUPPLEMENTARY TABLE 4). In this growth condition, the most known accumulated metabolites were PC(14:0/22:4(7Z,10Z,13Z,16Z)) and PA(16:1(9Z)/24:1(15Z)). The first compound is a phosphatidylcholine (PC), whereas the second is a phosphatidic acid (PA). Phospholipids, such as PC and PA, are essential for maintaining the structure and functionality of membranes (Lagace and Ridgway 2013). They are found in large amounts in fungal membranes (Siebers et al. 2016). Phospholipids appear to be relevant for the interaction of symbiotic and pathogenic microorganisms with host cells (Wang et al. 2019). For example, PC is pivotal for the functional symbiosis between the bacterium Bradyrhizobium japonicum and soybean (Minder et al. 2001), as well as for the obligate biotrophic fungus Rhizophagus irregularis colonizing Lotus japonicus (Wewer et al. 2014). Phospholipids have also been described as virulence determinants. Wessel et al. (2006) observed that Agrobacterium tumefaciens mutants unable to synthesize PC exhibited a drastic virulence defect. Wang et al. (2019) showed that phospholipid biosynthetic genes are required for full virulence, mycotoxin production, fungal development, and abiotic stress adaptation of Fusarium graminearum. Hence, it is likely that the higher production of PC and PA by N. parvum Bt-67 in contact with GP corresponds to the fungus’ effort to establish interaction with the host tissue.

The GP growth condition also influenced the production of trigonelline, an alkaloid extracted from various plants, including Fabaceae species and coffee (Ashihara et al. 2015). In plants, trigonelline may have roles as a nutrient source, osmoregulator during drought or salt stress, or cell cycle regulator (Ashihara et al. 2015). Trigonelline seems to be another example of plant metabolites produced by endophytic fungi. Teimoori-Boghsani et al. (2020) found that isolates of Talaromyces spp. from Salvia abrotanoides produced trigonelline. This metabolite has also been detected in fermentation cultures of Trichoderma asperellum (Ni et al. 2019). Another plant metabolite differentially produced in the GP condition was oripavine, a morphinane alkaloid found in Papaver spp. (Hosztafi 2014).

In the GLU condition, dipeptides, such as phenylalanylthreonine, alanylisoleucine, isoleucylvaline, and arginylphenylalanine, were among the most differentially produced metabolites. Most of the dipeptides seem to be short-lived intermediates on their way to specific amino acid degradation pathways following further proteolysis (Thirumalaikumar et al. 2021). However, some dipeptides may have physiological roles. In the yeast Saccharomyces cerevisiae, dipeptides activate the ubiquitin ligase E3 component N-recognin 1 (UBR1), which mediates the degradation of Cup9, a transcriptional repressor of peptide import (Turner et al. 2000). Dipeptides can also have antioxidative activities, protecting cells from oxidative stress (Ozawa et al. 2022).

In addition to dipeptides, compounds belonging to the steroid class, such as tetrahydrocorticosterone, 19-norandrosteron, cortexolone, and androsterone, were differentially produced in the GLU condition (TABLE 4; SUPPLEMENTARY TABLE 4). In animals, these compounds are considered steroid hormones, i.e., signaling molecules that mediate vital physiological functions (Dai et al. 2012). Interestingly, steroid signal molecules are not exclusive to animals. Steroid hormones, such as testosterone and androstenedione, have been found in fungi, including the plant-pathogenic fungus Cochliobolus lunatus (Kastelic-Suhadolc et al. 1994) and the edible mushroom Pleurotus ostreatus (Plemenitaš et al. 1999). In addition, the existence of steroid-binding proteins with a high affinity for steroid hormones, together with the steroid biotransformation capacity, has been demonstrated in distinct fungal species (Črešnar and Žakelj-Mavrič 2009; Kollerov et al. 2020). That implies that fungi possess a steroid signaling system (Žakelj-Mavrič et al. 1995); however, the role of steroid hormones in fungi, as well as the other components associated with this signaling system, has not been well understood. Previous studies have found that steroid hormones induce stress responses in fungi. For example, progesterone applications on the yeasts S. cerevisiae and Candida albicans influenced the up-regulation of stress-responsive genes, such as those involved in multidrug resistance, heat shock, osmotic stress, nitrogen depletion, amino acid starvation, and oxidative stress (Banerjee et al. 2007, 2004). Steroid hormones may also affect fungal growth. Cvelbar et al. (2013) observed that steroid compounds reduced the amount of ergosterol in S. cerevisiae and Hortaea werneckii, but not in Aspergillus oryzae. Considering those previous studies, we could speculate that N. parvum Bt-67 differentially produced steroid hormones to regulate its hyphal growth (e.g., by interfering with the ergosterol biosynthesis) or to enhance its stress responses to starvation during the GLU condition. On the other hand, it might be possible that the differential production of these steroids is associated with other physiological processes in the fungus, such as the regulation of the reproductive phase. Lin et al. (2016) found that pregnenolone produced and secreted by Trichoderma citrinoviride inhibited its conidiation. The Trichoderma pregnenolone also inhibited the conidiation of Fusarium graminearum (Lin et al. 2016). Remarkably, the conidiation inhibition was specific for pregnenolone over other steroids (Lin et al. 2016). The role of steroid hormones in plant-pathogenic fungi remains to be elucidated.

CONCLUSION

In this study, we found that N. parvum Bt-67 has a differential production of pathogenicity- and virulence-related factors depending on the lignocellulosic biomass. Interestingly, CAZyme expression levels, enzymatic activities, and lignocellulose biomass degradation were higher when the fungus grew with the non-host biomass (WS). These results suggest that wheat straw might be a promising carbon source to produce CAZymes of interest from N. parvum.

In addition, the distinct carbon sources differently enhanced the production of secondary metabolites. Remarkably, the diversity of differentially produced metabolites was higher when N. parvum Bt-67 grew with the host biomass (GP). Some differentially produced metabolites, such as phospholipids and ovalicin, seem to be relevant to the plant-microbe interaction. Others may be involved in the fungal physiology and response to stress conditions, such as diverse steroid compounds and carotenoids. Nevertheless, the exact role of most of the secondary metabolites, either those associated with the plant-pathogen interaction or those with the physiology of the fungus itself, requires further investigation.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

SUPPLEMENTARY MATERIAL

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00275514.2023.2216122

Additional information

Funding

The authors are grateful to the Grand Reims for the financial support of the AFERE and MALDIVE chairs. Julián D. Restrepo-Leal PhD funding was provided by the University of Reims Champagne-Ardenne and Grand Reims. Additional funding for the project was provided by AFERE and MALDIVE chairs.