Genome-wide prediction of secretory proteins in Ceratocystis paradoxa and expression analysis of effectors during early infection Authors: Xiuli Meng, Zhaowei Lin, Fengyu Yu, Xiaoqing Niu Published: 2026 ABSTRACT BackgroundCeratocystis paradoxa, a filamentous ascomycete and soil-borne wound pathogen, attacks various host plants such as Ananas comosus, Cocos nucifera and Saccharum officinarum. Secretory proteins are critical for phytopathogenic infection, yet related studies on C. paradoxa remain scarce.MethodsUsing WoLF PSORT, TMHMM, and big-PI Predictor, we predicted the secretory proteins in C. paradoxa. Carbohydrate-active enzymes and pathogenicity-related proteins were predicted by HMMER and PHI-base. Effectors were predicted by EffectorP v3.0. RT-qPCR was used to screen effectors highly expressed during early infection.ResultsIn this study, 253 secretory proteins were predicted from its 6,931 annotated genomic proteins. These proteins vary from 61 to 1894 residues in length, with signal peptides ranging from 13 to 35 residues in length. 58.89% of the signal peptide cleavage sites follow the canonical “Ala-X-Ala” motif, a substrate recognized by signal peptidase I. Using HMMER software, 78 carbohydrate-active enzymes (CAZymes) were predicted, including 45 glycoside hydrolases, 24 auxiliary activities, 4 polysaccharide lyases, 3 carbohydrate esterases, 1 glycosyl transferase and 1 carbohydrate binding module. Additionally, 144 secretory proteins matched known entries in the PHI-base, suggesting potential roles in pathogenicity. We further predicted 56 candidate effectors, classified into 9 cytoplasmic and 47 apoplastic types. Notably, six effector proteins were highly expressed during early infection, and three of them were predicted to be CAZymes.ConclusionsThis study systematically predicted 253 potential secretory proteins from C. paradoxa via multi-tool prediction, which may provide a comprehensive inventory of its pathogenicity-related secretome. These proteins potentially include 78 carbohydrate-active enzymes, 144 pathogenicity-related proteins, and 56 candidate effectors. Notably, six effectors highly expressed during early infection—three of which may have carbohydrate-active enzyme activity—could be key candidates involved in initial infection. Our findings may fill the gap in C. paradoxa secretory protein research, lay a potential foundation for elucidating its pathogenic mechanisms, and provide potential targets for controlling this soil-borne pathogen.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12866-026-05166-6. Background Plant pathogens secrete an arsenal of proteins and molecules into their hosts to modulate plant immune defense and facilitate colonization [15]. Fungal secretory proteins mainly consist of carbohydrate-active enzymes (CAZymes), effectors, toxins, peptidases, and other key protein types [15, 26, 35, 36, 41]. Studies of these secretory proteins are essential for unraveling the molecular mechanisms underlying pathogenicity. Protein secretion occurs via two primary routes: one is the classical endoplasmic reticulum-Golgi dependent pathway which is mediated through the presence of an N-terminal signal peptide [51],the other is the non-classical pathway, through which secretory proteins lacking a signal peptide are released via mechanisms such as membrane blebbing, the endolysosomal pathway, exosome-mediated secretion, and translocation across the plasma membrane [34]. CAZymes are capable of hydrolyzing carbohydrate in plant biomass for microbial uptake and consumption [46]. These enzymes primarily encompass six major classes: glycoside hydrolases (GHs), glycosyl transferases (GTs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), auxiliary activities (AAs) and carbohydrate binding modules (CBMs) [7]. One of the most well-characterized examples is MoChi1 from Magnaporthe oryzae – a GH -family secreted protein plays a critical role in the fungus's cell wall-degrading ability and pathogenicity [13]. Another important group of secretory proteins in plant pathogens is effectors which undermine plant immunity and promote colonization in the host [28]. Extensive research has shown that the majority of effector proteins are expressed during the early stages of infection [1, 25]. Thus, screening for effectors highly expressed in the early infection stage is valuable for identifying proteins related to pathogenicity. For instance, three effectors of Fusarium graminearum (FGSG_01831, FGSG_03599, and FGSG_12160) were highly expressed at 12 h post inoculation (hpi), and the deletion mutant Δ01831 displayed a significant reduction in initial infection [14]. Classical secretory proteins of fungi exhibit the following characteristics—(1) possession of an N-terminal signal peptide; (2) extracellular subcellular localization; (3) absence of transmembrane domains; and (4) lack of glycosyl phosphatidyl inositol (GPI) anchor site [58]. Whole-genome sequencing combined with bioinformatics analysis provides a powerful toolset for predicting the fungal secretome. For instance, whole-genome sequencing revealed that Ustilago maydis contains 6,902 predicted protein-encoding genes, among which 426 were secretory proteins [19, 27],Melampsora laricipopulina contained 6044 unique sequences, with 689 being secretory proteins [18],and in F. graminearum, the total number of secretory proteins reaches 574 [6]. Ceratocystis paradoxa (anamorph Thielaviopsis paradoxa), a pathogen with a broad host ranges, causes “pineapple disease” of sugarcane (Saccharum officinarum), stem-bleeding disease of coconut (Cocos nucifera) and postharvest disease of coconut and pineapple (Ananas comosus), posing a threat to crop yield and food security [17, 30, 37, 42]. As a necrotrophic fungus, C. paradoxa kills host cells first and then extracts nutrients from the necrotic tissue, leading to the yellowing and subsequent browning of host tissues before they wither and die [39]. C. paradoxa primarily invades hosts through wounds, such as those caused by agricultural operations, or via infestations by Coleoptera insects like Rhynchophorus ferrugineus [4, 31]. The diseases caused by C. paradoxa are difficult to control due to the short interval between the appearance of symptoms and the death of the plant [54]. Studying the pathogenicity- related secretory proteins of C. paradoxa is conducive to developing new fungicides and screening resistant host plants, which holds great significance for future disease management [24]. Although whole-genome sequencing of C. paradoxa was completed in 2022 [32], its secretory proteins remain uncharacterized. In this study, a detailed account of the prediction and potential functions of the secretome and effectorome of C. paradoxa was presented. We also analyzed the CAZymes and pathogenicity-related, secretory proteins in this fungus. Notably, six effectors were found to be highly expressed during the early infection stage, and three of them may function as CAZymes. This research provides crucial insights into the systematic analysis of the pathogenic factors of C. paradoxa and the mechanism underlying pathogen-host interactions. Materials and methods Sequence data and preparation The genomic resource used in this study was derived from our previously published genome assembly [32], and the corresponding annotated protein sequences were provided as matched supporting data by the sequencing company. These protein sequences served as the fundamental dataset for our subsequent secretory protein prediction. Prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins The software tools used to predict secretory proteins and effectors are listed in Table 1. First, SignalP v5.0 was used to identify N-terminal signal peptides [2]. Subsequently, WoLF PSORT was applied to predict the subcellular localization of proteins with N-terminal signal peptides [16], and only extracellular secretory proteins were retained for further analysis. Next, TMHMM v2.0 was used to predict transmembrane domains [22]. The proteins containing transmembrane domains were excluded and the remaining proteins were used for further analysis. The big-PI Predictor was used to predict the glycosyl phosphatidyl inositol (GPI) anchor site [29], and the proteins without GPI anchors were selected as secretory proteins. EffectorP v3.0 was used to identify effectors from the secretory proteins [44]. HMMER [61] was used to predict CAZymes (E-value < 1.0e-15). The pathogen-host interaction (PHI) database [38] was used to predict the proteins involving the process of plant-host interaction (E-value: 1.0e-5).Table 1The bioinformatics tools adopted for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins from C. paradoxaSoftwareObjects predictedReferencesSignalP v5.0Signal peptideshttps://services.healthtech.dtu.dk/service.php?SignalP-5.0WoLF PSORTSubcellular localizationhttps://wolfpsort.hgc.jpTMHMM v2.0Transmembrane domainshttps://services.healthtech.dtu.dk/services/TMHMM-2.0/big-PI PredictorGPI-anchor sitehttps://mendel.imp.ac.at/gpi/fungi_server.htmlEffectorP v3.0Effectorshttp://effectorp.csiro.au/HMMERCAZymeshttps://bcb.unl.edu/dbCAN2/blast.phpPHI databasePathogen-host interaction proteinshttp://www.phi-base.org/CD-SearchDomainhttps://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgiInterProDomainhttps://www.ebi.ac.uk/interpro/SWISS-MODEL3D-structurehttps://swissmodel.expasy.org/ Table 1The bioinformatics tools adopted for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins from C. paradoxaSoftwareObjects predictedReferencesSignalP v5.0Signal peptideshttps://services.healthtech.dtu.dk/service.php?SignalP-5.0WoLF PSORTSubcellular localizationhttps://wolfpsort.hgc.jpTMHMM v2.0Transmembrane domainshttps://services.healthtech.dtu.dk/services/TMHMM-2.0/big-PI PredictorGPI-anchor sitehttps://mendel.imp.ac.at/gpi/fungi_server.htmlEffectorP v3.0Effectorshttp://effectorp.csiro.au/HMMERCAZymeshttps://bcb.unl.edu/dbCAN2/blast.phpPHI databasePathogen-host interaction proteinshttp://www.phi-base.org/CD-SearchDomainhttps://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgiInterProDomainhttps://www.ebi.ac.uk/interpro/SWISS-MODEL3D-structurehttps://swissmodel.expasy.org/ The bioinformatics tools adopted for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins from C. paradoxa Inoculation with C. paradoxa and sample collection during the early infection stage C. paradoxa was cultured on potato dextrose agar (PDA) medium for 3 days. To prepare the spore suspension, 10 mL of sterile water was added to the PDA plate, and spores of C. paradoxa were scraped off using a sterile glass slide. The spore suspension was filtered through sterile medical gauze, and the spore concentration was quantified using a hemocytometer, and adjusted to 1 × 10⁶ spores/mL for inoculation [11]. Healthy fresh coconut fruits were selected, and uniform wounds were made with a syringe. A 10 μL aliquot of the spore suspension was inoculated into each wound, which was then covered with moist sterile cotton to maintain humidity. Samples were collected at 8, 12, 24, 36, and 48 h post-inoculation (hpi), immediately snap-frozen in liquid nitrogen, and stored at −80 °C. RNA extraction, RT-PCR and RT-qPCR Total RNA was extracted using RNA prep Pure Plant Plus Kit (Catalog number:DP441, Tiangen Biotech, Beijing, China) according to the manufacturer’s protocols, and 2 μg RNA was reverse-transcribed to cDNA using an RNA reverse transcription kit (Catalog number:KR118, Tiangen Biotech, Beijing, China). The resulting cDNA was then diluted tenfold and used as the template of RT-PCR and RT-qPCR. RT-PCR was performed using cDNA of mycelium as template. Each 25 μL PCR reaction contained 1 μL of cDNA template, 2.5 μL of 10 × PCR buffer for KOD-Plus-Neo (Catalog number-KOD-401,Toyobo, Shanghai, China), 2.5 μL of dNTPs (2 mM), 1.5 μL of MgSO4 (25 mM), 1 μL of each primer (10 μM), 0.5 μL of KOD-Plus-Neo and nuclease-free water to a final volume of 25 μL. The PCR protocol involved a pre-denaturation at 94 °C for 2 min, followed by 30 cycles of denaturation at 98 °C for 10 s, annealing at 60 °C for 30 s, extension at 68 °C for 30 s, and a final extension at 68 °C for 10 min. The primers used in this article are listed in Table S5. Subsequently, RT- qPCR was performed using cDNA of mycelium and inoculation material as template. The inoculation materials were collected at 8, 12, 24, 36, and 48 h after the coconut fresh fruits were inoculate by C. paradoxa. Each 20 μL PCR reaction contained 0.01 μg cDNA template, 10 μL of 2 × FastReal qPCR PreMix (Catalog number: FP217-02, Tiangen Biotech, Beijing, China), 0.4 μL of 50 × ROX Reference Dye, 1 μL of each primer (10 μM), and nuclease-free water to a final volume of 25 μL. The PCR protocol involved a pre-denaturation at 95 °C for 2 min, followed by 40 cycles of denaturation at 95 °C for 5 s, extension at 60 °C for 30 s. The reference gene β-tublin was used for the normalization of RT-qPCR data. The 2-▵▵CT method was used to calculate the relative expression levels with three technical repeats [21]. Three biological replicates and three technical replicates were used for each sample. Results Bioinformatic workflow for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins in C. paradoxa In our previous work, a total of 6,931 protein-coding genes were annotated based on the whole-genome sequence of C. paradoxa [32]. Among them, 525 proteins (7.57% of the total) were predicted to contain signal peptides via SignalP 5.0 (Fig. 1). Subcellular localization analysis using WoLF PSORT further predicted 427 putative extracellular proteins from this set. Of these candidates, 329 were estimated to lack transmembrane domains according to TMHMM prediction. Further screening with big-PI Predictor identified 253 proteins predicted to have no GPI anchor domain. Finally, these 253 proteins (3.65% of the total annotated proteins) were regarded as putative candidate secretory proteins of C. paradoxa (Fig. 1 and Table S1). From the 253 identified secretory proteins of C. paradoxa, 56 were predicted as candidate effectors using EffectorP 3.0, and 78 were annotated as CAZymes via HMMER. In addition, a total of 144 proteins were assigned as pathogenicity-related proteins based on searches against the Pathogen–Host Interactions (PHI-base) database (Fig. 1).Fig. 1Bioinformatic workflow for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins in C. paradoxa Bioinformatic workflow for the prediction of secretory proteins, effectors, CAZymes and pathogenicity-related proteins in C. paradoxa Characteristics of secretory proteins and their signal peptides The 253 predicted secretory proteins varied in length from 61 to 1894 amino acids (aa), with the majority (217 proteins, 85.8%) ranging between 100 and 700 aa (Fig. 2 and Text S1). The signal peptides of the 253 secretory proteins had lengths ranging from 13 to 35aa, with an average of 19aa (Fig. 2 and Table S2). The prevalent signal peptide length was 18aa, accounting for 21% of the total (Fig. 2). Analysis of the abundance of 20 amino acids in the signal peptides revealed their frequencies in descending order as follows: A, L, S, V, M, T, F, G, I, P, K, R, Q, H, N, Y, C, W, E, and D (Fig. 2). Alanine (A) was the most abundant (26.52%), followed by Leucine (L) (17.76%), while aspartate (D) was the least common (0.08%) (Fig. 2). Classification statistics showed that nonpolar hydrophobic amino acid residues (A, V, L, P, I, F, M, G, and W) accounted for 72.33%; polar uncharged residues (S, T, Y, Q, C, and N) made up 21.78%; positively charged residues (K, R, and H) constituted 5.54%; and negatively charged residues (D and E) represented 0.35% (Fig. 2). The flanking amino acids of signal peptidase cleavage sites were analysed (Table 2 and Table S2). The three amino acids at the N-terminus relative to the cleavage sites were designated as positions −3, −2, and −1, while the two amino acids at the C-terminus were defined as positions + 1 and + 2. Results showed that alanine (A) was the most abundant at position −3 (58.89%); serine (S) was most prevalent at position −2 (22.92%); alanine (A) dominated at position −1 (87.75%); alanine (A) was the most common at position + 1 (28.06%); and proline (P) was the most abundant at position + 2 (32.41%) (Table 2 and Table S2). The recognition site of signal peptidase I (SPase I) is primarily determined by the "Ala-X-Ala" motif, where the residues at positions −3 and −1 are alanine (Ala), and any amino acid can occupy position −2 [3, 57]. In this study, 149 secretory proteins contained the "Ala-X-Ala" motif (Table 2), indicating they could be cleaved by SPase I. Conserved domain analysis of the 253 predicted secretory proteins was performed using CD-Search and InterPro. A total of 63 conserved domain types were identified via CD-Search, among which CAZymes were the most abundant, followed by peptidases, and subsequently WSC (Cell Wall Integrity and Stress Response Component) and Fn3-like (Fibronectin type III-like domain) (Fig. 2 and Table S6). InterPro analysis further annotated 162 conserved domain categories. Similarly, CAZymes dominated in quantity, with peptidases ranking second, followed by FAD (Flavin Adenine Dinucleotide) and RlpA-like domains (Rare lipoprotein A-like domains) (Fig. 2 and Table S7).Fig. 2Analysis of amino acid length and residue of secretory proteins and signal peptides. a Analysis length of the predicted secretory proteins amino acid residues in C. paradoxa; (b) Analysis of the candidate secretory proteins with different length of signal peptides in C. paradoxa; (c) Percentage of 20 amino acid residues in C. paradoxa candidate secretory protein signal peptides; (d) Conserved domain prediction by CD-Search, WSC: Cell Wall Integrity and Stress Response Component, Fn3-like:Fibronectin type III-like, Inhibitor_I9:Peptidase inhibitor family I9, Egh16-like: Erysiphe graminis f.sp. hordei16-like virulence factor, PA_PoS1_like: Protease-associated (PA) PoS1-like, NPP1: Nuclear Pyrophosphatase/Phosphatase 1, MPP_YHR202W_N:Mitochondrial Processing Peptidase YHR202W N-terminal, CuRO_3_Abr2_like:Copper Reductase/Oxidase 3 Abr2-like, FTP:Fungal Transmembrane Protein, FhaB:Filamentous hemagglutinin B, FAD_binding_4: Flavin Adenine Dinucleotide binding domain, type 4, BetA:Betaine aldehyde dehydrogenase A, G_glu_transpept: Gamma-glutamyl transpeptidase; (e) Conserved domain prediction by InterPro, FAD: Flavin Adenine Dinucleotide, RlpA-like domain: Rare lipoprotein A -like protein domain, CAP domain: Cysteine-rich secretory protein, DnaJ domain: DnaJ heat shock N-terminal, FMN: Flavin Mononucleotide binding, SUN: Sad1-UNC-84 homologous domain, Vid27: Vacuolar Import and Degradation protein 27Table 2The Frequency of 20 amino acid residues around the signal peptidase cleavage siteAmino acidThe signal peptidase cleavage sites position−3−2−112No%No%No%No%No%A14958.895421.3422287.757128.06114.35C62.3720.7951.9810.400D0010.400145.53103.95E0062.370041.58114.35F0083.160072.7751.98G83.1651.98145.5393.5693.56H00135.1400155.9341.58I51.9862.370062.3762.37K0020.790041.5851.98L10.43614.2300228.7103.95M0093.560062.3710.4N0062.370051.9893.56P000000008232.41Q00103.9510.4207.9172.77R10.451.980083.1672.77S145.535822.92103.95239.092811.07T114.35155.9310.4103.95145.53V5822.92103.9500197.51187.11W0031.190031.1962.37Y0041.580062.37103.95No. number of amino acid residues Analysis of amino acid length and residue of secretory proteins and signal peptides. a Analysis length of the predicted secretory proteins amino acid residues in C. paradoxa; (b) Analysis of the candidate secretory proteins with different length of signal peptides in C. paradoxa; (c) Percentage of 20 amino acid residues in C. paradoxa candidate secretory protein signal peptides; (d) Conserved domain prediction by CD-Search, WSC: Cell Wall Integrity and Stress Response Component, Fn3-like:Fibronectin type III-like, Inhibitor_I9:Peptidase inhibitor family I9, Egh16-like: Erysiphe graminis f.sp. hordei16-like virulence factor, PA_PoS1_like: Protease-associated (PA) PoS1-like, NPP1: Nuclear Pyrophosphatase/Phosphatase 1, MPP_YHR202W_N:Mitochondrial Processing Peptidase YHR202W N-terminal, CuRO_3_Abr2_like:Copper Reductase/Oxidase 3 Abr2-like, FTP:Fungal Transmembrane Protein, FhaB:Filamentous hemagglutinin B, FAD_binding_4: Flavin Adenine Dinucleotide binding domain, type 4, BetA:Betaine aldehyde dehydrogenase A, G_glu_transpept: Gamma-glutamyl transpeptidase; (e) Conserved domain prediction by InterPro, FAD: Flavin Adenine Dinucleotide, RlpA-like domain: Rare lipoprotein A -like protein domain, CAP domain: Cysteine-rich secretory protein, DnaJ domain: DnaJ heat shock N-terminal, FMN: Flavin Mononucleotide binding, SUN: Sad1-UNC-84 homologous domain, Vid27: Vacuolar Import and Degradation protein 27 Table 2The Frequency of 20 amino acid residues around the signal peptidase cleavage siteAmino acidThe signal peptidase cleavage sites position−3−2−112No%No%No%No%No%A14958.895421.3422287.757128.06114.35C62.3720.7951.9810.400D0010.400145.53103.95E0062.370041.58114.35F0083.160072.7751.98G83.1651.98145.5393.5693.56H00135.1400155.9341.58I51.9862.370062.3762.37K0020.790041.5851.98L10.43614.2300228.7103.95M0093.560062.3710.4N0062.370051.9893.56P000000008232.41Q00103.9510.4207.9172.77R10.451.980083.1672.77S145.535822.92103.95239.092811.07T114.35155.9310.4103.95145.53V5822.92103.9500197.51187.11W0031.190031.1962.37Y0041.580062.37103.95No. number of amino acid residues The Frequency of 20 amino acid residues around the signal peptidase cleavage site No. number of amino acid residues Pathogenicity-associated secretory proteins Among the 78 predicted CAZymes, GHs were the most abundant, comprising 45 proteins distributed across 29 subfamilies (GH43, GH5, GH10, GH11, GH16, GH18, GH3, GH30, GH131, GH132, GH20, GH76, GH55, GH93, GH53, GH63, GH51, GH37, GH13, GH32, GH31, GH28, GH17, GH115, GH7, GH6, GH45, GH74 and GH81) (Fig. 3 and Table S3). The second most abundant category was AAs, containing 24 proteins from 9 subfamilies (AA9, AA1, AA2, AA3, AA7, AA8, AA5, AA11, and AA16). PLs included 4 proteins spanning 4 subfamilies (PL1, PL3, PL4 and PL26). CEs contained 3 proteins from 3 subfamilies (CE4, CE5, and CE8). GT and CBM each contained a single protein, classified as GT61 and CBM63 respectively (Fig. 3 and Table S3).Fig. 3Predicted CAZymes and PHI proteins of candidate secretory proteins of C. paradoxa. a Predicted CAZymes; (b) Predicted PHI proteins Predicted CAZymes and PHI proteins of candidate secretory proteins of C. paradoxa. a Predicted CAZymes; (b) Predicted PHI proteins The Pathogen–Host Interactions database (PHI-base) is widely used to predict pathogenesis-related genes and effectors of pathogens [49]. Among the 253 secretory proteins identified, 144 (56.9%) had at least one hit in PHI-base annotations (Fig. 3 and Table S4). Of these 144 proteins, knockout of 90 secretory protein-encoding genes was predicted to reduce virulence,deletion of another 38 such genes had no impact on pathogenicity. Nine secretory proteins were identified as putative effectors. Disruption of four additional genes led to a complete loss of pathogenicity, while deletion of the remaining three was associated with enhanced virulence. (Fig. 3 and Table S4). Analysis of 56 predicted effectors The 56 predicted effectors included 47 apoplastic effectors (secreted into the extracellular space of the host plant) and 9 cytoplasmic effectors (secreted inside the host’s cytoplasm) (Table 3). Analysis of the amino acid lengths of the 56 effectors revealed that 4 effectors contained fewer than 100 amino acids, while one exceeded 600 amino acids, and the majority were concentrated in the 100–300 amino acid range (Table 3 and Text S2). A characteristic feature of fungal effectors is a high content of cysteine residues [52]. Amino acid composition analysis showed that 46 out of 56 effectors contained more than 4 cysteine residues, accounting for 82.14% of the predicted effectors. Conserved domain analysis of the 56 effectors was performed using CD-Search, and conserved domains were identified in 33 of these effectors (Table 3). These domains could be divided into eight categories, including cell adhesion and cell wall regulation (WSC、Fasciclin、CVNH、Egh16-like), host interaction and immune modulation (LysM、AltA1), carbohydrate-active enzymes (GH11、GH7、GH43、GH28、GH45、Pectatelyase、AA9), hydrolase and protease activity (Abhydrolase、ZnMc、LapB、YoaJ), redox and metabolic homeostasis (cupin_OxOx、NPP1), endoplasmic reticulum protein quality control(EMP24_GP25L、ERp29、PRKCSH), nuclear homeostasis and stress response (SUN、BadF), as well as transcriptional and metabolic modification (Zn-ribbon_RPC11、Ribosomal_P2、PelB) (Table 3 and Table S8). Of the 23 effectors with no conserved domains identified by CD-Search, 3D structure prediction was performed using SWISS-MODEL. A total of 8 effectors were found to possess conserved domains, including EVM0002838 (Ubiquitin 3 binding), EVM0001001 (Root induced signalling), EVM0000590 (Fucose-specific lectin), EVM0005477 (DnaJ-related), EVM0004508 (Chitin-binding), EVM0003086 (Allergen Asp f4), EVM0006017 (DNase1 protein), and EVM0004591 (NTF2-like domain) (Fig. 4 and Table S9). CAZyme prediction by HMMER indicated that 15 of the 56 effectors may function as CAZymes, including 9 GHs, 2 CBMs, 2 PLs, 1 CE, and 1 AA (Table 3). Among the 56 effectors, 31 had at least one hit in PHI-base annotations (Table 3). Of these 31 effectors, deletion of 18 effector-encoding genes was predicted to reduce virulence, deletion of three such genes showed no influence on pathogenicity, seven proteins were defined as putative effectors, and deletion of three additional protein-encoding genes was predicted to cause pathogenicity loss.Table 3Analysis of 56 predicted effectors of C. paradoxaNoIDTypeLengthCysteine residuesCD-SearchCAZymesPHI prediction1EVM0006118.1Apoplastic3077AbhydrolaseCE5loss_of_pathogenicity2EVM0001457.1Apoplastic2009---3EVM0002838.1Apoplastic2717--reduced_virulence4EVM0002988.1Apoplastic1587BadF--5EVM0005248.1Apoplastic45110SUNGH132reduced_virulence6EVM0002749.1Cytoplasmic2193EMP24_GP25L-reduced_virulence7EVM0001001.1Apoplastic1584--unaffected_pathogenicity8EVM0001662.1Apoplastic2550GH11GH11effector9EVM0001415.1Apoplastic48324ZnMc-loss_of_pathogenicity10EVM0006308.1Apoplastic2374cupin_OxOx-effector11EVM0001672.1Apoplastic4099--reduced_virulence12EVM0003219.1Cytoplasmic5266LapB-reduced_virulence13EVM0003726.1Apoplastic1488CVNH--14EVM0001537.1Apoplastic1108---15EVM0000590.1Apoplastic1585---16EVM0005477.1Cytoplasmic4209--loss_of_pathogenicity17EVM0000959.1Cytoplasmic1840---18EVM0000448.1Apoplastic2533NPP1-effector19EVM0005292.1Apoplastic1588WSC-reduced_virulence20EVM0006727.1Apoplastic2018---21EVM0003091.1Cytoplasmic4552Fasciclin-reduced_virulence22EVM0005405.1Apoplastic17012Zn-ribbon_RPC11-unaffected_pathogenicity23EVM0001236.1Apoplastic1485--reduced_virulence24EVM0005140.1Apoplastic2784Egh16-like-reduced_virulence25EVM0006149.1Apoplastic3075AA9AA9effector_26EVM0004926.1Apoplastic1739---27EVM0002163.1Apoplastic23214PectatelyasePL3reduced_virulence28EVM0000791.1Apoplastic3456YoaJCBM63unaffected_pathogenicity29EVM0000664.1Cytoplasmic1010---30EVM0004508.1Apoplastic807---31EVM0002175.1Apoplastic1858LysMCBM50effector32EVM0003086.1Apoplastic3618---33EVM0006860.1Apoplastic836---34EVM0006847.1Apoplastic3619GH28GH28reduced_virulence35EVM0003861.1Apoplastic612---36EVM0006017.1Apoplastic1784---37EVM0005609.1Apoplastic705---38EVM0000708.1Apoplastic2028Egh16-like-reduced_virulence39EVM0001919.1Cytoplasmic56116PRKCSH-reduced_virulence40EVM0004929.1Apoplastic44633WSC-reduced_virulence41EVM0003315.1Apoplastic1468---42EVM0001470.1Apoplastic4484GH43GH43-43EVM0002373.1Apoplastic2330GH11GH11effector44EVM0002448.1Cytoplasmic4255ERp29-reduced_virulence45EVM0000175.1Apoplastic1666---46EVM0001048.1Apoplastic2290GH11GH11effector_47EVM0003467.1Apoplastic3246GH43GH43-48EVM0001673.1Apoplastic53924GH7GH7reduced_virulence49EVM0001978.1Apoplastic62643WSC-reduced_virulence50EVM0005163.1Apoplastic25012---51EVM0004591.1Apoplastic1866---52EVM0003780.1Apoplastic28418GH45GH45-53EVM0005548.1Cytoplasmic1090Ribosomal_P2--54EVM0003352.1Apoplastic43611PelBPL1reduced_virulence55EVM0004553.1Apoplastic1325---56EVM0005025.1Apoplastic1776AltA1--No data; Abhydrolase: Alpha/beta hydrolase superfamily, BadF Benzoyl-CoA reductase subunit F, SUN Sad1/UNC-84 homology, EMP24_GP25L Endoplasmic reticulum Membrane Protein of 24 kDa_ GlycoProtein 25L, ZnMc Zinc metallopeptidase, cupin_OxOx Cupin superfamily (Oxalate oxidase related), LapB Lipoprotein assembly factor LapB, CVNH Cyanovirin-N Homology, NPP1 Nucleotide pyrophosphatase/phosphodiesterase 1, Zn-ribbon_RPC11 RNA polymerase III subunit C11-like zinc ribbon, YoaJ Expansin-YoaJ, PRKCSH Protein kinase C substrate 80 K-H, ERp29 Endoplasmic reticulum protein 29, PelB Pectate lyase B, AltA1 Alternaria allergenFig. 4SWISS-MODEL-based 3D structural modeling of 8 effectors. a EVM0002838 (Ubiquitin 3 binding); b EVM0001001 (Root induced signalling); c EVM0000590 (Fucose-specific lectin); d EVM0005477 (DnaJ-related); e EVM0004508 (Chitin-binding); f EVM0003086 (Allergen Asp f4); g EVM0006017 (DNase1 protein); h EVM0004591 (NTF2-like domain) Table 3Analysis of 56 predicted effectors of C. paradoxaNoIDTypeLengthCysteine residuesCD-SearchCAZymesPHI prediction1EVM0006118.1Apoplastic3077AbhydrolaseCE5loss_of_pathogenicity2EVM0001457.1Apoplastic2009---3EVM0002838.1Apoplastic2717--reduced_virulence4EVM0002988.1Apoplastic1587BadF--5EVM0005248.1Apoplastic45110SUNGH132reduced_virulence6EVM0002749.1Cytoplasmic2193EMP24_GP25L-reduced_virulence7EVM0001001.1Apoplastic1584--unaffected_pathogenicity8EVM0001662.1Apoplastic2550GH11GH11effector9EVM0001415.1Apoplastic48324ZnMc-loss_of_pathogenicity10EVM0006308.1Apoplastic2374cupin_OxOx-effector11EVM0001672.1Apoplastic4099--reduced_virulence12EVM0003219.1Cytoplasmic5266LapB-reduced_virulence13EVM0003726.1Apoplastic1488CVNH--14EVM0001537.1Apoplastic1108---15EVM0000590.1Apoplastic1585---16EVM0005477.1Cytoplasmic4209--loss_of_pathogenicity17EVM0000959.1Cytoplasmic1840---18EVM0000448.1Apoplastic2533NPP1-effector19EVM0005292.1Apoplastic1588WSC-reduced_virulence20EVM0006727.1Apoplastic2018---21EVM0003091.1Cytoplasmic4552Fasciclin-reduced_virulence22EVM0005405.1Apoplastic17012Zn-ribbon_RPC11-unaffected_pathogenicity23EVM0001236.1Apoplastic1485--reduced_virulence24EVM0005140.1Apoplastic2784Egh16-like-reduced_virulence25EVM0006149.1Apoplastic3075AA9AA9effector_26EVM0004926.1Apoplastic1739---27EVM0002163.1Apoplastic23214PectatelyasePL3reduced_virulence28EVM0000791.1Apoplastic3456YoaJCBM63unaffected_pathogenicity29EVM0000664.1Cytoplasmic1010---30EVM0004508.1Apoplastic807---31EVM0002175.1Apoplastic1858LysMCBM50effector32EVM0003086.1Apoplastic3618---33EVM0006860.1Apoplastic836---34EVM0006847.1Apoplastic3619GH28GH28reduced_virulence35EVM0003861.1Apoplastic612---36EVM0006017.1Apoplastic1784---37EVM0005609.1Apoplastic705---38EVM0000708.1Apoplastic2028Egh16-like-reduced_virulence39EVM0001919.1Cytoplasmic56116PRKCSH-reduced_virulence40EVM0004929.1Apoplastic44633WSC-reduced_virulence41EVM0003315.1Apoplastic1468---42EVM0001470.1Apoplastic4484GH43GH43-43EVM0002373.1Apoplastic2330GH11GH11effector44EVM0002448.1Cytoplasmic4255ERp29-reduced_virulence45EVM0000175.1Apoplastic1666---46EVM0001048.1Apoplastic2290GH11GH11effector_47EVM0003467.1Apoplastic3246GH43GH43-48EVM0001673.1Apoplastic53924GH7GH7reduced_virulence49EVM0001978.1Apoplastic62643WSC-reduced_virulence50EVM0005163.1Apoplastic25012---51EVM0004591.1Apoplastic1866---52EVM0003780.1Apoplastic28418GH45GH45-53EVM0005548.1Cytoplasmic1090Ribosomal_P2--54EVM0003352.1Apoplastic43611PelBPL1reduced_virulence55EVM0004553.1Apoplastic1325---56EVM0005025.1Apoplastic1776AltA1--No data; Abhydrolase: Alpha/beta hydrolase superfamily, BadF Benzoyl-CoA reductase subunit F, SUN Sad1/UNC-84 homology, EMP24_GP25L Endoplasmic reticulum Membrane Protein of 24 kDa_ GlycoProtein 25L, ZnMc Zinc metallopeptidase, cupin_OxOx Cupin superfamily (Oxalate oxidase related), LapB Lipoprotein assembly factor LapB, CVNH Cyanovirin-N Homology, NPP1 Nucleotide pyrophosphatase/phosphodiesterase 1, Zn-ribbon_RPC11 RNA polymerase III subunit C11-like zinc ribbon, YoaJ Expansin-YoaJ, PRKCSH Protein kinase C substrate 80 K-H, ERp29 Endoplasmic reticulum protein 29, PelB Pectate lyase B, AltA1 Alternaria allergen Analysis of 56 predicted effectors of C. paradoxa No data; Abhydrolase: Alpha/beta hydrolase superfamily, BadF Benzoyl-CoA reductase subunit F, SUN Sad1/UNC-84 homology, EMP24_GP25L Endoplasmic reticulum Membrane Protein of 24 kDa_ GlycoProtein 25L, ZnMc Zinc metallopeptidase, cupin_OxOx Cupin superfamily (Oxalate oxidase related), LapB Lipoprotein assembly factor LapB, CVNH Cyanovirin-N Homology, NPP1 Nucleotide pyrophosphatase/phosphodiesterase 1, Zn-ribbon_RPC11 RNA polymerase III subunit C11-like zinc ribbon, YoaJ Expansin-YoaJ, PRKCSH Protein kinase C substrate 80 K-H, ERp29 Endoplasmic reticulum protein 29, PelB Pectate lyase B, AltA1 Alternaria allergen SWISS-MODEL-based 3D structural modeling of 8 effectors. a EVM0002838 (Ubiquitin 3 binding); b EVM0001001 (Root induced signalling); c EVM0000590 (Fucose-specific lectin); d EVM0005477 (DnaJ-related); e EVM0004508 (Chitin-binding); f EVM0003086 (Allergen Asp f4); g EVM0006017 (DNase1 protein); h EVM0004591 (NTF2-like domain) Expression profile analysis of effectors during early infection To screen the effectors whose expression was significantly upregulated during the early infection stage, the expression profiles of 56 effectors were analysed. Firstly, effectors highly expressed during the vegetative growth stage were excluded via RT-PCR. Twenty-seven out of the 56 effector proteins (No. 1, 6, 8, 12, 15, 16, 17, 18, 19, 20, 21, 26, 29, 31, 33, 34, 36, 39, 40, 42, 44, 46, 47, 48, 51, 52, 53) exhibited relatively high expression levels during vegetative growth stage (Fig. 5). The expression profile of the remaining 29 effectors were analysed using RT-qPCR. Each of the 5 time points (8 h, 12 h, 24 h, 36 h and 48 hpi) were selected to analyse expression profile of the 29 effectors and the vegetative growth stage was used as control. Compared with the vegetative growth stage, six effectors (No.5, 11,14, 25,35 and 43) were highly expressed in the early infection stage (Fig. 6). Their corresponding gene IDs were EVM0005248, EVM0001672, EVM0001537, EVM0006149, EVM0003861 and EVM0002373 respectively (Table 3). EVM0005248, EVM0001672, EVM0001537, EVM0006149 and EVM0003861showed a “low–high-low” expression pattern during the infection stage (Fig. 6). EVM0005248 reached its highest relative expression at 8 hpi, while EVM0001672 peaked at 36 hpi. The relative expressions of EVM0001537 peaked at 12 hpi and EVM0006149 maintained high relative expression levels at both 8 h and 12 hpi (Fig. 6). EVM0003861 exhibited a relatively complex expression pattern: it reached a peak at 12 hpi, decreased slightly at 24 hpi, increased again at 36 hpi, and decreased once more at 48 hpi (Fig. 6). EVM0002373 displayed a “low–high” pattern, where its relative expression level continuously increased and peaked at 48 hpi. Notably, among these four infection-stage highly expressed effectors, EVM0005248, EVM0006149 and EVM0002373 were CAZyme (Table 3).Fig. 5RT-PCR result of 56 candidate effectors during vegetative growth stage. M: Marker D2000, T: Beta-tubulin was used as a reference gene. The fragment sizes of 56 effectors are as follows: 1: 184 bp; 2: 153 bp; 3: 74 bp; 4: 173 bp; 5: 81 bp; 6: 142 bp; 7: 182 bp; 8: 162 bp; 9: 200 bp; 10: 245 bp; 11: 90 bp; 12: 171 bp; 13: 170 bp; 14: 152 bp; 15: 225 bp; 16: 237 bp; 17: 240 bp; 18: 236 bp; 19: 211 bp; 20: 226 bp; 21: 238 bp; 22: 238 bp; 23: 219 bp; 24: 236 bp; 25: 188 bp; 26: 241 bp; 27: 168 bp; 28: 153 bp; 29: 209 bp; 30: 230 bp; 31: 207 bp; 32: 247 bp; 33: 170 bp; 34: 248 bp; 35: 166 bp; 36: 167 bp; 37: 203 bp; 38: 164 bp; 39: 153 bp; 40: 224 bp; 41: 194 bp; 42: 196 bp; 43: 183 bp; 44: 220 bp; 45: 177 bp; 46: 176 bp; 47: 164 bp; 48: 164 bp; 49: 247 bp; 50: 179 bp; 51: 176 bp; 52: 178 bp; 53: 246 bp; 54: 208 bp; 55: 171 bp; 56: 209 bpFig. 6Gene expression profiles of 6 effectors upregulated during early infection stage. The material collected at 8, 12, 24, 36 and 48 h after inoculation were used for RNA isolation and cDNA synthesis. Fungal β-tubulin was used as an internal control for transcript normalization. Fold changes of gene expression were relative to mycelium, which was grown on PDA plates for 3 days. Three biological replicates and three technical replicates were used for each sample. a EVM0005248 expression pattern; b EVM0001672 expression pattern; c EVM0001537 expression pattern; d EVM0006149 expression pattern; e EVM0003861 expression pattern; f EVM0002373 expression pattern. The error bars in this experiment were plotted based on the standard deviation RT-PCR result of 56 candidate effectors during vegetative growth stage. M: Marker D2000, T: Beta-tubulin was used as a reference gene. The fragment sizes of 56 effectors are as follows: 1: 184 bp; 2: 153 bp; 3: 74 bp; 4: 173 bp; 5: 81 bp; 6: 142 bp; 7: 182 bp; 8: 162 bp; 9: 200 bp; 10: 245 bp; 11: 90 bp; 12: 171 bp; 13: 170 bp; 14: 152 bp; 15: 225 bp; 16: 237 bp; 17: 240 bp; 18: 236 bp; 19: 211 bp; 20: 226 bp; 21: 238 bp; 22: 238 bp; 23: 219 bp; 24: 236 bp; 25: 188 bp; 26: 241 bp; 27: 168 bp; 28: 153 bp; 29: 209 bp; 30: 230 bp; 31: 207 bp; 32: 247 bp; 33: 170 bp; 34: 248 bp; 35: 166 bp; 36: 167 bp; 37: 203 bp; 38: 164 bp; 39: 153 bp; 40: 224 bp; 41: 194 bp; 42: 196 bp; 43: 183 bp; 44: 220 bp; 45: 177 bp; 46: 176 bp; 47: 164 bp; 48: 164 bp; 49: 247 bp; 50: 179 bp; 51: 176 bp; 52: 178 bp; 53: 246 bp; 54: 208 bp; 55: 171 bp; 56: 209 bp Gene expression profiles of 6 effectors upregulated during early infection stage. The material collected at 8, 12, 24, 36 and 48 h after inoculation were used for RNA isolation and cDNA synthesis. Fungal β-tubulin was used as an internal control for transcript normalization. Fold changes of gene expression were relative to mycelium, which was grown on PDA plates for 3 days. Three biological replicates and three technical replicates were used for each sample. a EVM0005248 expression pattern; b EVM0001672 expression pattern; c EVM0001537 expression pattern; d EVM0006149 expression pattern; e EVM0003861 expression pattern; f EVM0002373 expression pattern. The error bars in this experiment were plotted based on the standard deviation Discussion The secreted proteome is a crucial component of the ability of fungi to perceive and respond to the environment [9]. Genome-wide analysis of the secreted proteome in pathogens facilitates a comprehensive understanding of their pathogenic factors. Classical secretory proteins in fungi are generally based on the following four characteristics: the presence of an N-terminal signal peptide, extracellular type subcellular localization, the absence of transmembrane domains, and the absence of GPI anchor sites [58]. Based on the four characteristics mentioned above, we predicted 253 secretory proteins in C. paradoxa by SignalP, WoLF PSORT, TMHMM, and big-PI Predictor. The proportion of secretory proteins of C. paradoxa to total proteins is 3.65%, which is lower than that of most plant-pathogenic fungi such as M. oryzae (14.3%), F. graminearum (9.9%), Verticillium dahlia (10.5%) [45] and other Ceratocystis species, such as C. albifundus (4.61%), C. cacaofunesta (5.72%), C. fimbriata (5.49%), C. manginecans (7.26%) and C. platani (8.04%) [39]. The different proportions of secretory proteins in fungi may be related to their pathogenic strategies,however, further experiments are required to verify this hypothesis. The 253 predicted secretory proteins will serve as a precise blueprint, providing targeted insights for future research into pathogenic genes. The lengths of the secretory proteins of C. paradoxa were concentrated in the range of 100–700 amino acids, which is consistent with that of Botryosphaeria dothidea [56]. The lengths of the signal peptides were concentrated between 17–20 amino acids, with an average of 19 aa, a feature consistent with Colletotrichum graminicola [12], V. dahlia [48], and Stemphylium lycopersici [58]. However, the abundance of the 20 amino acids in the signal peptides of C. paradoxa differed from that reported in other pathogenic fungi. For C. paradoxa, the amino acids ranked in descending order of abundance were A, L, S, V, M, T, F, G, I, P, K, R, Q, H, N, Y, C, W, E, and D. In contrast, the order for S. lycopersici was A, L, S, T, V, M, F, I, G, P, R, K, Q, C, Y, H, N, W, E, and D [58], and for V. dahliae it was A, L, S, V, T, M, G, F, I, P, R, Q, K, W, C, Y, H, N, E, and D [48]. A common feature among these was that amino acids A, L, and S were highly represented in the signal peptides. In this study, 58.89% of the signal peptide cleavage sites conformed to the canonical “Ala-X-Ala” motif. This motif is a substrate recognized and processed by SPase I, which is consistent with findings in many pathogenic fungi, such as S. lycopersici [58], B. dothidea [56], and C. graminicola [12]. Fungi produce a variety of CAZymes for degrading plant polysaccharide materials to facilitate infection and gain nutrition. Among them, CAZymes of classes GHs, PLs, AAs and CEs are known as cell wall degrading enzymes [33]. A total of 78 CAZymes were predicted in C. paradoxa, including 45 GHs, 24 AAs, 4 PLs, 3 CEs, 1 CBM and 1 GT. The 78 CAZymes may be involve in the cell wall degrading progress, however, their bioinformatics characteristics and specific functions need to be verified by gene knockout technology. Previous research has indicated that necrotrophic fungi produce CAZymes with higher diversity than biotrophic and hemibiotrophic fungi, in addition, plenty of CAZymes of necrotrophic fungi are up-regulated during the plant infection process [5, 59]. It is particularly noteworthy that three of predicted CAZymes (EVM0005248, EVM0006149 and EVM0002373) were proved to be highly up-regulated during the plant infection process, which is consistent with existing research findings [5, 59]. Further research is required to confirm the function of EVM0005248, EVM0006149 and EVM0002373 during the process of C. paradoxa invasion into the host. The plant's primary immune defense, PAMP-triggered immunity (PTI), is activated by pathogen-associated molecular patterns (PAMPs). Pathogens counteract PTI by secreting effectors, which play a critical role in early infection [8]. Criteria to fit the definition of candidate secreted fungal effector proteins include: proteins with a signal peptide for secretion, no trans-membrane domains, no similarity with other obvious protein domains, fairly small size and rich in cysteine residues [43]. For instance, known effectors of Cladosporium fulvum such as Avr2, Avr9, Avr4, and ECP2, are small cysteine-rich proteins [47]. In addition, the effectors can be secreted inside the host cytoplasm as well as in the extracellular space, and are subsequently classified as cytoplasmic and apoplastic effectors [43]. In this study, we predicted 56 effectors in C. paradoxa and 47 ones were apoplastic effectors. The 76.8% of predicted effectors are shorter than 400 amino acids in length and 83.9% of the effector proteins contain more than 4 cysteine residues, which align with established fungal effector characteristics so far. Except for CAZymes, numerous effector domains have been documented to be closely associated with fungal pathogenicity, including WSC, Egh16-like, LysM, AltA1, EMP24_GP25L, ERp29, and SUN [10, 20, 23, 35, 36, 53, 55, 60]. In the present study, conserved domain analysis via CD-Search revealed that EVM0005292 harbors a WSC domain, EVM0005140 contains an Egh16-like domain, EVM0002175 possesses a LysM domain, EVM0005025 has an AltA1 domain, EVM0002749 carries an EMP24_GP25L domain, EVM0002448 contains an ERp29 domain, and EVM0005248 harbors a SUN domain. The potential pathogenic roles of these seven candidate effectors remain to be further experimentally validated in future research. Notably, six out of the 56 predicted effectors were highly expressed during the early infection stage of C. paradoxa. These six effectors (EVM0005248, EVM0001672, EVM0001537, EVM0006149, EVM0003861, and EVM0002373) were selected as priority candidates for further functional characterization of their roles in host immune suppression and pathogen invasion. It should be emphasized that although transcriptome expression profiling can screen highly expressed effectors at different infection stages, it alone is insufficient to fully clarify their biological functions, particularly during the transition from the early asymptomatic phase to the necrotrophic phase. The Agrobacterium-mediated heterologous transient expression assay in Nicotiana benthamiana is a well-established approach to characterize candidate effectors as either cell-death inducers or immune suppressors [50]. Therefore, these six significantly upregulated effectors will be subjected to agroinfiltration assays in the subsequent work. In addition, subcellular localization analysis is required to determine the precise distribution of these effectors once they enter host cells [40]. In future studies, the subcellular localization assays on these key effectors should be also conducted to identify their exact targeting sites, providing further experimental evidence for their biological roles during pathogen–host interaction. Conclusion This article employed online software to predict the secretory proteins and effectors in C. paradoxa. Additionally, CAZymes and pathogenicity-related proteins were identified by HMMER software and PHI-base. Furthermore, among the predicted effector proteins, six candidates—including three putative CAZymes—exhibited high expression levels during the early infection stage of C. paradoxa. These findings provide novel insights into the pathogenic genes of C. paradoxa, though the functions of the target genes require further experimental validation. Supplementary Information Supplementary Material 1: Table S1: Secretory proteins prediction result. Table S2: Signal peptides sequence and flanking amino acid of cleavage site. Table S3: CAZymes prediction result. Table S4: PHI prediction result of 144 secretory proteins. Table S5: Primer sequences of 56 effectors. Text S1: Amino acid sequence of 253 secretory proteins. Text S2: Amino acid sequence of 56 effectors. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4. Supplementary Material 5. Supplementary Material 6. Supplementary Material 1: Table S1: Secretory proteins prediction result. Table S2: Signal peptides sequence and flanking amino acid of cleavage site. Table S3: CAZymes prediction result. Table S4: PHI prediction result of 144 secretory proteins. Table S5: Primer sequences of 56 effectors. Text S1: Amino acid sequence of 253 secretory proteins. Text S2: Amino acid sequence of 56 effectors. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4. Supplementary Material 5. Supplementary Material 6.