Fosp9, a Novel Secreted Protein, Is Essential for the Full Virulence of Fusarium oxysporum f. sp. cubense on Banana (Musa spp.)

ABSTRACT

The banana vascular wilt pathogen, Fusarium oxysporum f. sp. cubense, delivers a number of different secreted proteins into host plant tissues during infection. Until now, only a few of the secreted proteins from this fungus have been shown to be virulence effectors. Here, the product of fosp9, which is a gene in this pathogen, was found to be a novel virulence effector. The fosp9 gene encodes a hypothetical 185-amino-acid protein which has a functional signal peptide but contains no known motifs or domains. The fosp9 disruptants displayed a significant reduction in producing wilt symptoms on bananas, indicating that fosp9 is essential for the full virulence of this pathogen for banana. These disruptants did not exhibit a change in either saprophytic growth or conidiation on potato dextrose agar medium, but their invasive growth in the rhizomes of banana was markedly compromised, suggesting a pivotal role for fosp9 in the colonization of banana rhizome tissues by this fungus. Live-cell imaging revealed that the Fosp9-GFP fusion protein accumulated in the apoplast of the plant cells. Moreover, transcriptome profiling revealed that a number of virulence-associated genes were differentially expressed in the fosp9 disruptant relative to the wild type. Taken together, these findings suggest that Fosp9 is a genuine effector of F. oxysporum f. sp. cubense.

IMPORTANCE Fusarium wilt of bananas (also known as Panama disease), caused by the fungus F. oxysporum f. sp. cubense, is one of the most devastating banana diseases worldwide. The understanding of the molecular mechanism of its pathogenicity is very limited so far. We demonstrated that the secreted protein Fosp9 from this fungus contributes to its virulence against banana hosts and is essential for colonization of banana rhizome tissues by this fungus. In particular, Fosp9 contains no known domains or motifs and has no functionally characterized homologs, implying that it is a novel secreted effector involved in F. oxysporum f. sp. cubense-banana interactions. This work provides insight into molecular mechanisms of F. oxysporum f. sp. cubense pathogenicity, and the characterization of the fosp9 gene will facilitate development of transgenic banana and plantain strains resistant to this disease by silencing this effector gene through host-induced gene silencing or other strategies.

INTRODUCTION

Fusarium wilt of banana (commonly known as Panama disease) caused by the soilborne fungus Fusarium oxysporum f. sp. cubense is a destructive disease, which has been a major constraint to banana production worldwide (1, 2). In the 20th century, replacing Gros Michel bananas with the Cavendish varieties reduced the losses caused by this disease. However, in recent decades, Cavendish bananas have succumbed to this disease, and the discovery of the disease in more and more countries is attributed to a variant of the fungal pathogen called F. oxysporum f. sp. cubense tropical race 4 (2). This epidemic disease is particularly difficult to control, because the fungal pathogen can survive in the soil as chlamydospores for decades and cannot be totally eradicated. In particular, race 4 can infect most varieties of banana and plantain. The use of resistant varieties and preventing the fungal pathogen from entering plantations are the most effective means of control (1, 2). A proper understanding of the pathogenesis of F. oxysporum f. sp. cubense would allow the development of effective novel strategies to prevent the disease. Specifically, investigations into the secreted virulence proteins involved in the F. oxysporum f. sp. cubense-banana interaction would provide insight into the mechanisms underlying the resistance or susceptibility of Cavendish banana varieties to this fungus.

Many studies have revealed that secreted proteins play important roles in the F. oxysporum-plant interactions. For instance, secreted lipases from F. oxysporum f. sp. lycopersici have been implicated in pathogenicity against tomato, as the single or double deletion mutants of the regulator genes ctf1 and ctf2 which participate in lipase regulation caused severe reductions in both total lipase activity and virulence for tomato (3). The metalloprotease FoMep1 and serine protease FoSep1 secreted by this fungus are required for full virulence in tomato, because both proteases have plant chitinase-cleaving activity and are able to significantly reduce the antifungal activity of tomato chitinases (4). In addition to these secreted enzymes, F. oxysporum f. sp. lycopersici secretes several small proteins into the xylem sap, which are called Six (“secreted in xylem”) proteins. Of these proteins, SIX1 (Avr3), SIX3 (AvrII), SIX5, and SIX6 are effectors that can promote virulence for tomato, while SIX1 (Avr3), SIX3 (AvrII), and SIX4 (AvrI) are avirulence proteins that can trigger disease resistance in tomato (5). Intriguingly, homologues of some SIX proteins have been found in other formae speciales of F. oxysporum, including F. oxysporum f. sp. cubense (6, 7). Of particular interest, homologues of SIX1 and SIX8 in F. oxysporum f. sp. cubense were found to contribute to the full virulence for bananas, indicating that they are effectors (7, 8).

To identify secreted proteins (SPs) that might play crucial roles during F. oxysporum f. sp. cubense-banana host interactions, we analyzed transcriptome profiling of this fungus during early infection of Cavendish bananas and found that the expression of 43 secreted-protein (SP) genes significantly increased (9). Curiously, no counterparts of these SPs were found in other fungal pathogens when their sequences were used to query the pathogen-host interaction database PHI-base (www.phi-base.org/) using BLASTP, suggesting that the virulence functions of all these SP genes have not been elucidated. Therefore, we employed a large-scale disruption analysis of these SP genes to search for novel virulence SPs. We found that most of the candidate SP gene disruptants did not show significant changes in pathogenicity compared to the wild-type strain. However, disruption of one SP gene, fosp9, led to a severe reduction in pathogenicity. Further analysis demonstrated that the Fosp9 protein was required for normal colonization of F. oxysporum f. sp. cubense. The results will improve our understanding of molecular mechanisms underlying the pathogenesis of this fungus and facilitate development of novel strategies to combat Fusarium wilt disease of banana in future.

RESULTS

Characterization of the fosp9 gene in F. oxysporum f. sp. cubense.

The fosp9 gene sequence (GenBank accession no. EMT68949.1) in F. oxysporum f. sp. cubense race 4 is 558 bp in length and contains no introns. It was predicted to encode a protein of 185 amino acid residues with a putative signal peptide at the cleavage site between amino acids 18 and 19 (SignalP-5.0 server), suggesting that Fosp9 may be a secretory protein (Fig. 1a).

FIG 1.

Sequence analysis of Fosp9. (a) Deduced amino acid sequence and signal peptide (boxed) of Fosp9 protein. (b) Phylogenetic tree constructed based on amino acid sequences of Fosp9 and its homologs from other fungi using the neighbor-joining method with 1,000 bootstrap replicates. Numbers at the nodes denote the bootstrap confidence values, and numbers in parentheses are sequence identities between Fosp9 and other homologs.

Fosp9 orthologs have been found in various filamentous fungi, including F. oxysporum f. sp. cubense race 1 strain N2 (GenBank accession no. ENH71512), F. oxysporum f. sp. lycopersici 4287 (XP_018250655), F. oxysporum f. sp. vasinfectum (EXM19413), F. oxysporum f. sp. pisi (EXA35636), F. oxysporum f. sp. radicis-cucumerinum (PCD23483), F. fujikuroi (XP_023437779), F. proliferatum (CVL12864), F. acutatum (KAF4443784), F. agapanthi (KAF4478537), F. nygamai (PNP83985), F. verticillioides (XP_018758080), Colletotrichum orbiculare (TDZ14156), Verticillium dahlia (KAF3353668), and V. longisporum (CRK38439), with 47 to 99% amino acid sequence identity. Fosp9 exhibits a higher amino acid identity (95 to 99%) with orthologs in other Fusarium species than those in other filamentous fungi. Amino acid sequence identity between Fosp9 and its orthologs in F. oxysporum f. sp. race 1 strain N2 and F. oxysporum f. sp. lycopersici 4287 reached 99%. Notably, the ortholog of fosp9 in F. oxysporum f. sp. lycopersici is located on chromosome 13, an accessory chromosome, which suggests that this gene is not a core member of the F. oxysporum genome (10). In addition, a phylogenetic tree based on the amino acid sequences showed that Fosp9 is more closely related to its homologs in Fusarium species than those in other fungi (Fig. 1b). Among their common features, these amino acid sequences do not contain known domains/motifs, but they do contain two conserved cysteine residues in the mature protein region (see Fig. S1 in the supplemental material). These findings suggest that fosp9 encodes a novel secreted protein with unknown functions.

fosp9 is expressed during conidial germination and the infection progress.

Quantitative real-time-PCR (qRT-PCR) assays showed that expression of fosp9 was induced in F. oxysporum f. sp. cubense race 4 strain B2 during both conidial germination and the early infection stages of both susceptible and resistant banana roots. The expression level of fosp9 in conidia germinated for 24 h was over 4 times higher than that in the conidia germinated for 0 h. Moreover, the expression level of fosp9 was at least 400 times higher at 24 postinoculation (hpi) and 48 hpi than at 0 hpi either in the fungus inoculated into susceptible banana (Musa sp. AAA cv. Brazilian) or resistant banana (Musa sp. AAA cv. Nan tian huang) plants (11). Interestingly, the expression level of fosp9 was significantly higher in the susceptible banana (Musa sp. AAA cv. Brazilian) than in the resistant banana (Musa sp. AAA cv. Nan tian huang) plants at 24 h and 48 hpi, although the transcript level declined at 48 hpi compared to 24 hpi during infection of the susceptible plants.

When the fungus was subjected to oxidative stress (6 mM H₂O₂), the expression of the fosp9 gene was depressed, with an approximate 35% decrease in the transcript level relative to the mock treatment. Similarly, when this fungus was subjected to osmotic stress (1.2 M sorbitol), the expression of the fosp9 gene was also severely suppressed, with a nearly 80% decline in the transcript level compared to the mock treatment. These results indicate that oxidative and osmotic stresses have a negative effect on transcription of the fosp9 gene in F. oxysporum f. sp. cubense race 4 strain B2.

Functional validation of the signal peptide of the Fosp9 protein.

The yeast signal peptide trap assay showed that the constructs pSuc-Fosp9 and pSuc-SIX1a (which served as a positive control) enabled the invertase mutant yeast strain YTK12 to grow on CMD-W and YPRAA media (see Materials and Methods), and their culture filtrates could induce red color reactions after 2,3,5-triphenyltetrazolium chloride (TTC) treatment. In contrast, the negative-control construct pSuc2-Mg87 (12), in which the first 25 amino acids of the nonsecreted Mg87 protein from Magnaporthe oryzae were fused in frame to the mature sequence of yeast invertase in the vector pSuc2, did not enable the yeast strain YTK12 grow well on YPRAA, and the TTC-treated culture filtrate remained colorless (Fig. 2). Similar results were observed for the yeast strain YTK12. These results indicate that Fosp9 carries a functional secretory signal peptide.

FIG 2.

Functional verification of the signal peptide of Fosp9. The Fosp9 signal peptide fragment was fused in frame to the invertase gene in the pSUC2 vector, resulting in the vector pSUC2-Fosp9. pSUC2-Fosp9, pSUC2-Six1a (positive control), and pSUC2-Mg87 (negative control) were used for the yeast signal sequence trap assay. (a) Normal growth of each strain on yeast extract peptone dextrose agar (YPDA) plates. (b) CMD-W (minus Trp) plates were used to select yeast strain YTK12 carrying the pSUC2 vector. (c) YPRAA medium (containing raffinose as the only carbohydrate source) was used to indicate invertase secretion. (d) Enzymatic activity test. The dye 2,3,5-triphenyltetrazolium chloride (TTC) was reduced to the insoluble red compound triphenylformazan, indicating invertase secretion.

Fosp9 protein accumulated in the apoplast.

A vector expressing green fluorescent protein (GFP)-tagged Fosp9 protein was generated and used for transformation of Nicotiana benthamiana leaves by the agroinfiltration method (Fig. 3a). The subcellular localization of Fosp9 was determined by fluorescence microscopic imaging. The results showed that GFP alone accumulated in the cytoplasm and nucleus; in contrast, GFP-tagged Fosp9 accumulated in the cytoplasm but not in the nucleus. Meanwhile, GFP-tagged Fosp9 also appeared in the apoplasts of plasmolyzed plant cells (Fig. 3b). These results imply that Fosp9 might localize to the apoplast of host plant cells.

FIG 3.

Subcellular localization of Fosp9 after transient expression in N. benthamiana leaves. (a) Schematic diagram of the construct 35S:Fosp9-GFP. (b) Subcellular localization of GFP alone and Fosp9-GFP were observed by confocal microscopy at 36 h after Agrobacterium-mediated transformation of N. benthamiana, and merged GFP and bright-field images are shown. The N. benthamiana cells were plasmolyzed in 0.8 M mannitol solution. Cytoplasm (C) and nucleus (N) are indicated with white and red arrows, respectively, and the apoplasts (A) of plant cells are indicated with red arrowheads. Bars = 50 μm.

Phenotypic analysis of fosp9 disruptants and fosp9-complemented strains.

The fosp9 disruptants sp9-1 and sp9-4 were generated in the wild-type strain B2 of F. oxysporum f. sp. cubense race 4 by targeted replacement of the fosp9 gene with the hygromycin resistance gene (HPH) cassette (Fig. S2a). Multiple PCR assays with gene-specific primer sets (Table 1) revealed that the fosp9 gene was replaced by the HPH cassette in these two mutants (Fig. S2b), and this result was further confirmed by Southern blotting using an HPH-specific probe (Fig. 4a). Additionally, multiple PCR assays with gene-specific primer sets revealed that the complementation strain c-sp9 was successfully generated (Fig. 4a). As expected, the transcripts of fosp9 could be detected in the WT and complemented strain c-sp9 but could not be detected in strains sp9-1 and sp9-4 by quantitative PCR using cDNAs as templates (data not shown). These results indicated that the fosp9 gene was deleted in mutant strains sp9-1 and sp9-4 and that an intact copy of the fosp9 gene was successfully introduced into strain c-sp9.

TABLE 1.

Primer sets used in this study

Primer names	Oligonucleotide sequence (5′–3′)	Remark
FOC4_g10014531_F	CCGTCCACACCAAGAACCTCAAC	qRT-PCR
FOC4_g10014531_R	CCAGACTCCTTGTAGCCTCCGAAT
FOC4_g10012152_F	CTTATCATGGTTGGCGGTGCTATCA	qRT-PCR
FOC4_g10012152_R	AGGATGCGTTGAACAGAGAAGTGAG
FOC4_g10007426_F	CGTCCAAGATGGTCGTGTCAAGG	qRT-PCR
FOC4_g10007426_R	AGGCGTTCCAGGCAGAGCAA
FOC4_g10012762_F	ACCTCAACGCTGCCAACAAGAC	qRT-PCR
FOC4_g10012762_R	TTGCGGATAAGGACGAGGATGGT
FOC4_g10003282_F	CTCCAAGACTTCCGTCCTGTTCCT	qRT-PCR
FOC4_g10003282_R	AACCGAGAGCGATGCCAGTGT
FOC4_g10012064_F	CAGCGGATGTCTTGTCGTGTCTC	qRT-PCR
FOC4_g10012064_R	CCATTCTCAGGCTCAGGCTCTTG
FOC4_g10016768_F	TTCTCCACGGTCTCTGCTCCTTC	qRT-PCR
FOC4_g10016768_R	GTTGCCTTCTCGCCACATCTCTG
FOC4_g10010444_F	AGGCTGCTATGTTGGTCAGGAACT	qRT-PCR
FOC4_g10010444_R	CGTCTGCGTGATACTGAAGTGTTGT
FOC4_g10016782_F	CGCTGGAGTTGCCATGTGGATAC	qRT-PCR
FOC4_g10016782_R	ACGAACGATGCTACCAACGGAGA
FOC4_g10010870_F	GCTTCTAGCCGCACTGGTCTCT	qRT-PCR
FOC4_g10010870_R	CATTCCTCCATCTTCTCCGCATGAT
FOC4_g10006535_F	CGACGACATTACCAGCTACGATACG	qRT-PCR
FOC4_g10006535_R	AGGCACACCACTGACACCACT
FOC4_g10003693_F	GTCGCTCAAGTCGTTGCCTCATT	qRT-PCR
FOC4_g10003693_R	GATGCCTAAGCCGAAGCCGTATG
Foc-actin_F	CCAAGTCCAACCGTGAGAAGATGA	qRT-PCR
Foc-actin_R	CCAGAGTCCAGAACGATACCAGTG
Foc-ef1a_F	CCAGTGCGGTGGTATCGACAAG	qRT-PCR
Foc-ef1a_R	TGACGGTGACATAGTAGCGAGGAG
Fosp9-qrt-F	ACCATCTTCTCCATCCTCGCCATC	qRT-PC
Fosp9-qrt-R	GCCTTCGTCTGTGTCGTAGATTGC
MusaACT1-F	TGCTTGATTCTGGTGATGGTGTGAG	Quantitative PCR
MusaACT1-R	TCAGCAGTGGTCGTGAAGGAATAAC
sp9-psuc-F	GAATTTTAATTAAGAATTCATGCTTGCTCAAACCATCTTCTCC	Construction of the vector pSuc2-Fosp9
sp9-psuc-R	CTATAGGGAGAACCTCGAGAGGAGCAGCGAGGCCAGTAGT
SIX1a-psuc-F	GAATTTTAATTAAGAATTCATGGCGCCCTATAGCATGGTA	Construction of the vector pSuc2-SIX1a
SIX1a-psuc-R	CTATAGGGAGAACCTCGAGCTCTTGAGCATAAGCCCCAAACC
fosp9-pfgc-F	ttaccatggggcgcgccATGCTTGCTCAAACCATCTTCTCCATCC	Construction of the vector pFGC-Fosp9
fosp9-pfgc-R	GCTCACCATggcgcgccAGCTGATAACAATACGAGGACCAGAC
nsfosp9-pfgc-F	ttaccatggggcgcgccATGGCCATCGCCACTACT	Construction of the vector pFGC-nsFosp9
nsfosp10-pfgc-R	GCTCACCATggcgcgccAGCTGATAACAATACGAGGACCAGAC
sp9-1F	ggggtaccTGATAGGTGGGTGTATGATG	PCR amplification of the upstream fragment of fosp9
sp9-2R	gggagctcCCTCCAACTAAGAACTACTTG
sp9-3F	ggaattcGAATATGACCTGCCACAATGA	PCR amplification of the downstream fragment of fosp9
sp9-4R	cgagatctAAGGATCTCCAACGACTGAT
sp9-cF	ggttctcgaggtcgacTTGGCGGAATGTGAGCATGAATGT	PCR amplification of an intact copy of the fosp9 gene
sp9-cR	gggaacaaaagctggagctcATTACGAGTCAGGACGGTGTCAGAA
upF1	CAAGAGACAGAAAGTTGGTG	Identification of fosp9 mutants
upR2	GAATAGAGTAGATGCCGACC	Identification of fosp9 mutants
dpF3	GATCCGATGATAAGCTGTCA	Identification of fosp9 mutants
dpR4	CAACCGAACACAAGAACTCT	Identification of fosp9 mutants
9F5	ATGCTTGCTCAAACCATCTT	Identification of fosp9 mutants
9R6	GCTGATAACAATACGAGGAC	Identification of fosp9 mutants
sp9-C80A-F	GCAGCCAGGGGTAAGCAAGCCCAATTCGCTTTCTACCTCG	Site-directed mutagenesis of Fosp9
sp9-C80A-R	CGAGGTAGAAAGCGAATTGGGCTTGCTTACCCCTGGCTGC
sp9-C151A-F	GACCCAGAAGACGGATGCTAAGAAGCCTGGTACTGTTGAG	Site-directed mutagenesis of Fosp9
sp9-C151A-R	CTCAACAGTACCAGGCTTCTTAGCATCCGTCTTCTGGGTC

FIG 4.

Characterization and vegetative growth of fosp9 disruptant and fosp9-complemented strains. (a) Southern blot of fosp9 disruptant strains (sp9-1 and sp9-4). The hygromycin resistance cassette (HPH) fragments were used as the probes. (b) Colony, mycelial, and conidial morphology of mutant and wild-type strains. Bars = 20 μm.

After growth on potato dextrose agar (PDA) plates at 25°C for 6 days, the fosp9 mutants sp9-1 and sp9-4 had no changes in colony diameters, color, mycelial or conidial morphology, or conidial production (Fig. 4b) compared to the wild type (WT), indicating that disruption of the fosp9 gene might have no influence on the fungal growth and development. When the disruptants sp9-1 and sp9-4 were cultured on PDA plates supplied with 5 mM H₂O₂, the fungal growth was slightly inhibited, with about an 18% reduction in colony diameters relative to WT and fosp9-complemented strains (∼14% reduction). In contrast, growth on the PDA plates supplied with 1.2 M sorbitol showed no difference between the fosp9 mutant strains and the WT. As expected, both oxidative (5 mM H₂O₂) and osmotic (1.2 M sorbitol) stresses did not affect colonial morphology of these disruptants, and they were as round as those of WT and fosp9-complemented strains. Additionally, both mutant and wild-type strains could penetrate cellophane when they were grown on PDA plates covered with a cellophane membrane for 3 days (data not shown). These results suggest that deletion of the fosp9 gene might have a slight impact on the fungal tolerance to oxidative stress, but it has no effect on the tolerance to osmotic stress or the secretion of cell wall-degrading enzymes.

Fosp9 is required for full virulence of F. oxysporum f. sp. cubense.

The pathogenicity of the fosp9 mutants sp9-1 and sp9-4, the fosp9-complemented strain c-sp9, and the WT was tested on banana (Musa sp. AAA cv. Brazilian) plantlets at the three-leaf stage. At 45 days postinoculation, the symptoms in rhizome tissues were conspicuous enough for evaluating the disease severity (Fig. 5a). In contrast to the WT, the fosp9 mutants sp9-1 and sp9-4 exhibited significantly reduced virulence on banana plants, with diminished browning symptoms in the rhizome tissues (Fig. 5a) and a lower disease severity index (DSI) (Fig. 5b), while the fosp9-complemented strain c-sp9 showed virulence similar to that of the WT on banana plantlets, with severe symptoms in banana rhizome tissues and a DSI comparable to those infected with the WT (Fig. 5b), implying that the fosp9-complemented mutant recovered its virulence to a level as effective as that of the WT. Additionally, the mock‐treated plantlets displayed no symptoms in rhizome tissues. These results indicate that Fosp9 plays an essential role in the full virulence of the fungus.

FIG 5.

Disease symptoms and disease severity indices (DSI) of banana plantlets inoculated with wild-type and mutant strains. (a) Internal discoloration in rhizome tissues of banana plantlets inoculated with wild-type and mutant strains (red arrows). The wild-type strain B2, the fosp9 disruptants sp9-1 and sp9-4, the fosp9-complemented strain c-sp9, and the fosp9-complemented strain sp9(C80A, C151A) with mutations in two cysteine residues were inoculated into banana plantlets (Musa sp. AAA, cv. Brazilian) for 45 days. Inoculation with sterile water was used as a control (Mock). (b) DSI of banana plants inoculated with wild-type and mutant strains. The disease severity was recorded using a scale ranging from 0 (healthy plant) to 5 (dead plant). Values are means for 90 banana plants tested for each fungal strain in three independent experiments. *, P < 0.05 (Tukey test) relative to the WT.

Colonization of banana rhizome tissues by the pathogen.

Colonization in banana rhizome tissues is a pivotal step in the infection process of banana by F. oxysporum f. sp. cubense race 4 strain B2 after root penetration. Whether disruption of fosp9 gene influenced the fungal colonization in banana rhizome tissues was unknown. Therefore, the relative biomass of fosp9 mutant strains in banana rhizome tissues was analyzed by quantitative PCR. The results showed that fungal biomass in the rhizome tissues of plantlets inoculated with either disruptant sp9-1 or sp9-4 was approximately 30% of that in plants inoculated with the WT at 45 days postinoculation. The fungal biomass in banana plantlets inoculated with the fosp9-complemented strain c-sp9 was comparable to that of plantlets infected with the WT. As expected, no fungal DNA was detected in the rhizome tissues of mock-treated banana plantlets, while the positive-control pure culture of this fungus gave high biomass. These results imply that disruption of the fosp9 gene reduces the colonization of F. oxysporum f. sp. cubense in banana rhizome tissues.

fosp9 disruptants had a less negative impact on the growth of banana plantlets.

It has been reported that the pathogenic fungus infection could affect the growth of banana plants (13). Whether fosp9 mutants influenced the growth of banana plantlets was unknown. Therefore, we assessed plant height, stem weight, leaf area, and fresh weight of the banana plantlets inoculated with WT and fosp9 mutants. Compared to the mock-treated banana plantlets, those inoculated with the WT, sp9-1, sp9-4, and c-sp9 displayed marked declines in growth (Fig. S3). Specifically, plant height, stem diameter, fresh weight of the aboveground part, and leaf area of banana plantlets inoculated with WT were reduced by 22%, 6%, 32%, and 19% relative to those of the mock-treated plants (CK), respectively. In contrast, the growth of plantlets inoculated with sp9-1 was less inhibited, with 6%, 5%, 18%, and 7% declines in plant height, stem diameter, fresh weight of the aboveground part, and leaf area of banana plantlets relative to CK, respectively. Similar results were observed for sp9-4. As expected, the growth of banana plantlets inoculated with c-sp9 was also depressed by levels comparable to those inoculated with WT. These results indicated that disruption of fosp9 alleviated its negative impacts on the growth of banana plantlets relative to WT.

Effect of the fungal infection on the activities of some defense enzymes in banana plantlets.

The plant defense enzymes peroxidase (POD) and catalase (CAT), as well as phenylalanine ammonia-lyase (PAL), are related to the defense response of plants upon challenge with various pathogens (14–17). We measured the activities of these enzymes to determine the influence of the fungal infection on defense response of bananas (Musa sp. AAA cv. Brazilian) (Fig. S4). Compared to mock-treated plantlets, both POD and PAL enzyme activities displayed increases in plantlets inoculated with sp9-1, sp9-4, c-sp9, or the WT at 45 days postinoculation (dpi), and CAT activity showed an increase in plantlets inoculated with sp9-1 or sp9-4 but exhibited a decrease in those inoculated with the WT or the fosp9-complemented strain c-sp9. More specifically, POD activity was 1.4-fold higher in plantlets inoculated with sp9-1 or sp9-4, while it was about 0.64-fold higher in plantlets inoculated with the WT, compared to mock-treated plantlets. Higher POD and CAT activities might contribute to a lesser degree of disease symptoms in plantlets infected by the fosp9 mutants sp9-1 and sp9-4.

The two cysteine residues are not essential for Fosp9 function.

Fosp9 contains two cysteine residues (C80 and C151) in the mature protein, and a disulfide bond was predicted to be formed by these cysteine residues using the DISULFIND program (Fig. S5a). To test whether these cysteines are essential for Fosp9 function, a mutant allele of fosp9 was generated in which both C80 and C151 were replaced with alanine (Fig. S5b). The mutant allele with the two replacements [fosp9(C80A, C151A)] was expressed in the fosp9 mutant sp9-4. The virulence of sp9(C80A, C151A) for banana plantlets was restored to the level of the WT (Fig. 5b), with symptoms in rhizome tissues and disease severity (DSI = 54.0) similar to those of the WT (DSI = 56.0). Meanwhile, the fungal biomass in rhizome tissues of banana plantlets inoculated with sp9(C80A, C151A) was comparable to that in plantlets inoculated with the WT. These results indicate that C80 and C151, which are presumably involved in disulfide bond formation, are not necessary for Fosp9 to exert its function in F. oxysporum f. sp. cubense during infection and colonization of banana plantlets.

Transcriptome analysis of the fosp9 disruptant and wild-type strain.

To gain insight into differences in the molecular and cellular processes between the fosp9 mutant sp9-4 and the WT during the infection of banana (Musa sp. AAA cv. Brazilian) roots, these strains were used to inoculate banana root for 48 h followed by transcriptome profiling using transcriptome sequencing (RNA-seq). After filtering out the low‐quality reads, about 22.77 million raw sequencing reads and about 20.32 million clean reads from each sample were generated. The average genome mapping ratio was appropriately 84%, using the F. oxysporum f. sp. lycopersici 4287 strain genome (Broad Institute) as the reference genome. A total of 19,839 genes were found to be expressed in this fungus. Compared with the wild type, 2,139 genes were found to be changed dramatically (≥1.5-fold change) in the fosp9 disruptant sp9-4. Among these differentially expressed genes (DEGs), 1,183 DEGs had decreased expression levels, while 956 DEGs had increased expression levels (Table S1). To validate the RNA-seq data, the expression levels of 12 randomly selected DEGs were analyzed using qRT-PCR with primer sets designed for the targeted genes. The results showed that the expression levels of these genes analyzed using qRT-PCR were generally consistent with the corresponding variations in transcript levels determined by RNA-seq (Table S2), suggesting the reliability of the RNA-seq data.

All DEGs were then used to perform a Gene Ontology (GO) enrichment analysis. Only 262 of these DGEs were assigned a biological function, and they were categorized accordingly (Fig. 6a). Their molecular functions were primarily related to transporter activity (GO:0005215) and transmembrane transporter activity (GO:0022857). The cellular components involved included intrinsic component of membrane (GO:0031224), integral component of membrane (GO:0016021), membrane part (GO:0044425), and membrane (GO:0016020). These imply that proteins encoded by enriched DGEs are involved in transport of substances and the majority of them are components of a membrane. Additionally, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that seven pathways, including “Alanine, aspartate and glutamate metabolism (fox00250),” “Metabolic pathways (fox01100),” “Ribosome (fox03010),” “Starch and sucrose metabolism (fox00500),” “Glyoxylate and dicarboxylate metabolism (fox00630),” “Cyanoamino acid metabolism (fox00460),” and “Ascorbate and aldarate metabolism (fox00053),” were significantly enriched in a total of 201 DEGs (Fig. 6b). Thus, these DEGs were significantly enriched in the pathways related to amino acid and carbohydrate metabolism, as well as translation.

FIG 6.

GO and KEGG pathway enrichment analyses of differentially expressed genes (DEGs) in the fosp9 disruptant relative to wild-type strain. (a) GO functional annotation of DEGs. The black and gray histograms represent the numbers of DEGs enriched in cellular components and molecular function, respectively. The y axis shows the GO terms, and the x axis shows the numbers of DEGs enriched in the corresponding terms. (b) KEGG pathway enrichment analysis of DEGs. The x axis shows the number of DEGs enriched in the corresponding pathways, and the y axis shows the KEGG pathways.

To find potential virulence-associated genes among the DEGs, a BLAST analysis was conducted against PHI-base (version 4.9; http://www.phi-base.org/), which contains a collection of genes from fungi, oomycetes, and bacteria which have been proven to affect the outcome of pathogen-host interactions. A total of 304 PHI-base genes were identified as orthologous to the DEGs (Table S3). Approximately 83% of these matching PHI-base genes are from pathogenic fungi, and 17% are from pathogenic bacteria. Moreover, genes from the plant-pathogenic fungi Magnaporthe oryzae and Fusarium graminearum account for 19% and 16% of all matching PHI-base genes, respectively, and those from the animal-pathogenic fungi Candida albicans, Aspergillus fumigatus and Cryptococcus neoformans constituted around 8%, 7%, and 5% of all matching PHI-base genes, respectively. Notably, a number of genes whose expression was decreased in the fosp9 disruptant sp9-4 are orthologous to PHI-base genes which encode transporters implicated in toxin export and required for essential nutrient acquisition (5 genes), transcription factors regulating fungal virulence (9 genes), and key enzymes involved in fungal toxins, amino acid and siderophore biosynthesis (16 genes) and lipid metabolism (2 genes) in other fungi, respectively (Fig. 7). These results imply that Fosp9 might influence many facets of the fungus, including amino acid synthesis, carbohydrate metabolism, fungal toxin biosynthesis, and essential nutrient acquisition, by which it contributes to the pathogen virulence.

FIG 7.

Heat map showing the differences in gene expression levels (log₂ FPKM) of some genes in the fosp9 disruptant strain sp9-4 relative to the wild-type strain B2. Color from red to blue indicates high to low expression and log₂ FPKM value.

DISCUSSION

The F. oxysporum f. sp. cubense genome encodes more than 1,300 secreted proteins, including carbohydrate-active enzymes and peptidases, but only a few of the secreted proteins have been experimentally identified and characterized thus far. In this study, we identified and characterized Fosp9, a novel secreted protein in F. oxysporum f. sp. cubense race 4. Our findings showed that Fosp9 is required for full virulence of this fungus on banana plants.

The fosp9 gene (GenBank accession no. EMT68949.1) seems to be a unique gene without any paralogues in the pathogenic-fungus genome. The sequence of fosp9 could be mapped to chromosome 13 of F. oxysporum f. sp. lycopersici strain 4287, an accessory chromosome, which suggests that this gene is not a core member of the F. oxysporum genome. Also, it is different from genes encoding effector proteins (such as proteins secreted in xylem [SIX]) and secreted enzymes predicted to degrade or modify plant or fungal cell walls, which reside on other accessory chromosomes (10). Meanwhile, Fosp9 protein lacks known sequence motifs or conserved domains, and functions of other Fosp9 orthologs in other filamentous fungi (Fig. 1) are also unknown. Thus, Fosp9 is a novel secreted protein in F. oxysporum f. sp. cubense race 4.

Numerous studies have revealed that a single amino acid change could lead to alteration of protein function (18–20). In the present study, although amino acid sequence identity between Fosp9 (race 4 strain B2) and its ortholog in the fungus F. oxysporum f. sp. cubense race 1 strain N2 is up to 99%, it remains possible that they have distinctive roles in pathogenicity. Thus, it would be interesting to investigate roles of the ortholog in race 1 strain in pathogenicity and to determine whether function of a fosp9 disruptant in the race 4 strain can be restored with its ortholog in a race 1 strain. This will help determine whether Fosp9 is implicated in general pathogenicity or race-specific pathogenicity. Also, investigation of the roles of Fosp9 orthologs in pathogenicity in other filamented fungi would help determine whether it is involved in general pathogenicity or species-specific pathogenicity.

Upon infection, the pathogenic fungus F. oxysporum switches from a saprophytic to an infectious lifestyle, and expression of many secreted enzymes predicted to degrade or modify plant cell wall and secreted proteins is induced (10). The effector gene SIX1 of F. oxysporum f. sp. lycopersici could be induced during penetration and colonization of tomato roots (21). Similarly, the cerato-platanin family protein-encoding gene FocCP1 from the fungus F. oxysporum f. sp. cubense was highly expressed during spore germination and the infection process (22). Intriguingly, a similar expression pattern was observed in the secreted protein gene SIX8, which was induced during the early stage of F. oxysporum f. sp. cubense infection (8). In this study, we found that the expression of fosp9 gene was induced during infecting banana roots, which was consistent with the aforementioned secreted protein genes (SIX1, FocCP1, and SIX8). Thus, the induction of expression of these genes appeared to be related to host penetration and colonization early during infection, and F. oxysporum f. sp. cubense appears to release a set of secreted proteins, including Fosp9, FocCP1, and SIX8, to promote pathogenesis on banana during the early stage of infection.

It has been reported that the growth of tomato plants was suppressed by infection with the soilborne fungus F. oxysporum, and growth inhibition could be alleviated in plants infected by the effector gene SIX6 deletion mutant compared to those infected by the WT (23). Although inhibition of growth of banana plants infected with the fungus F. oxysporum f. sp. cubense has been reported (13, 24), enhanced growth of banana plants infected by strains of the fungus with reduced virulence relative to those infected with the WT has not been reported so far. In this study, we found that disruption of fosp9 could mitigate inhibition of growth of banana plants by F. oxysporum f. sp. cubense infection (Fig. S3). Thus, our result is consistent with that has been reported in the SIX6 deletion mutant of F. oxysporum f. sp. lycopersici. These results imply that disruption of fosp9 impairs pathogenicity, which significantly reduced the inhibitory effect on growth of the banana plantlet. Thus, the enhanced plant growth suggests that the infection process of fosp9 disruptants is slower, allowing the plant to continue to grow during the early stages of infection.

The cysteine residues in many secreted proteins are essential for virulence function in some plant-pathogenic fungi. For instance, two of the four cysteine residues in the secreted effector Pep1 were shown to be indispensable for the virulence function in Ustilago maydis (25). Similarly, the cysteine residues in the secreted protein MC69 are essential for the pathogenicity function in the rice blast fungus M. oryzae (26). Additionally, two cysteine residues were confirmed to play key roles in maintaining the structure and virulence function of the small secreted protein SsSSVP1 in the necrotrophic phytopathogen Sclerotinia sclerotiorum (27). Interestingly, cysteine residues that form disulfide bonds are required for the functioning of the secreted small cysteine‐rich protein VdSCP126, but not VdSCP27 or VdSCP113, in Verticillium dahliae (28), which play a pivotal role in the V. dahliae-plant interaction. Also, cysteine residues in the secreted protein FoEG1 are not required for induction of plant cell death (29). These differences show that cysteine residues are important for virulence function in some secreted proteins, but not in others. In the current study, two conserved cysteine residues (C81 and C150) in the mature protein of Fosp9 were predicted to be involved in disulfide bridge formation (Fig. S1), but they were not required for Fosp9 to exert its virulence function during infection of banana plantlets in the complementation assay (Fig. 5b). Therefore, our result is similar to that found for VdSCP27 and VdSCP113 of V. dahliae, and these findings indicate that some secreted virulence proteins play important roles in host-pathogen interactions via distinct cysteine residues that form disulfide bonds.

Secreted proteins that are delivered into the host plant tissues by pathogenic fungi can either display their activity in the apoplast or be translocated into host cells and function therein. For instance, the fungal effector Pit2 is secreted into the apoplasts of maize cells, where it inhibits host cysteine proteases, resulting in compatibility in the U. maydis-maize interaction (30). The secreted protein SsCP1, which is a cerato-platanin family protein, was found to interact with PR1 in the apoplast to facilitate infection of the necrotrophic phytopathogen S. sclerotiorum (31). Likewise, another cerato-platanin family protein, FocCP1, which is essential for the penetration and virulence of F. oxysporum f. sp. cubense, functions in the apoplast (22). Additionally, the U. maydis effector Pep1 is localized to the apoplast of plant cells and inhibits plant peroxidases to suppress the early immune responses of maize (25). In contrast, the M. oryzae effector Avr-Piz-t was translocated into the rice cytoplasm, where it targeted the ring E3 ubiquitin ligase apip6 to inhibit pathogen-associated molecular pattern-triggered immunity in rice (32). Here, we found that the Fosp9 protein was localized to the apoplasts of plant cells (Fig. 3) upon transient expression in N. benthamiana leaves. Thus, it can be inferred that the host target could be localized in the apoplast, where Fosp9 functions, or in the plasma membrane during the fungus-banana interaction. Since the host target of Fosp9 is as yet unknown, it would be interesting to uncover how Fosp9 contributes to the virulence of F. oxysporum f. sp. cubense and the plant processes in which it interferes.

Defense-related enzymes, such as PAL, CAT, and POD, play a role in plant defense upon pathogen infection (14–17). Several effector proteins secreted by plant-pathogenic fungi function as inhibitors of specific defense-related enzymes. For example, the effector protein Pep1 suppresses the early immune responses of maize by inhibiting plant peroxidases to promote infection by U. maydis (33). Similarly, the serine proteases FoMep1 and FoSep1, secreted by F. oxysporum f. sp. lycopersici, have chitinase-cleaving activity and are able to significantly reduce the antifungal activity of tomato chitinases (4). Additionally, the secreted effector protein Pit2 from U. maydis inhibits the maize cysteine proteases, which are involved in salicylic acid (SA)-associated defense (30). In this study, it was notable that an increase in CAT activity was shown only in the fosp9 disruptant-treated plantlets. Meanwhile, POD activity was higher in the fosp9 disruptant-treated plantlets than in the wild-type-treated plantlets (Fig. S4). Thus, we speculate that Fosp9 might inhibit the CAT and POD activities of susceptible banana plants to promote fungal virulence. However, to test this hypothesis, the interactions between Fosp9 and defense enzymes need to be experimentally verified.

Studies on dual species transcript profiling during the interaction between F. oxysporum f. sp. cubense and banana (Musa spp.) revealed that a great number of genes which are related to pathogenicity, pectin and chitin metabolism, and reactive oxygen scavenging play pivotal roles during infection (34). Moreover, CCP1 (cytochrome c peroxidase 1) was shown to participate in cellulose utilization, the oxidative stress response, and pathogenicity of the fungus. Similarly, in this study, we found that a series of DEGs were homologous to virulence-associated genes in t PHI-base (Table S4). Of these DEGs, many encode transporters involved in toxin export and nutrient acquisition. and a substantial portion encode key enzymes required for fungal toxin, amino acid, and siderophore synthesis, while nine encode transcription factors implicated in regulation of fungal virulence (Fig. 7). Of note, these genes were decreased in the fosp9 disruptant compared to the WT. Furthermore, KEGG pathway enrichment analysis revealed that a number of the DEGs found here are genes involved in carbohydrate and amino acid metabolism (Fig. 6). Therefore, Fosp9 might play roles in modulating the fungal toxin synthesis and export and in essential nutrient synthesis, as well as roles by which it contributes to fungal virulence during F. oxysporum f. sp. cubense-banana interactions.

In summary, we demonstrated that the secreted protein Fosp9 contributes to the virulence of F. oxysporum f. sp. cubense and its colonization on banana plantlets, so it is therefore a genuine effector. Future work should include experiments aimed at identifying the Fosp9 host targets and plant processes perturbed by Fosp9, to uncover how this effector contributes to the virulence of the fungus for bananas.

MATERIALS AND METHODS

Strains and growth conditions.

The wild-type strain B2 (race 4) and mutant strains of F. oxysporum f. sp. cubense were cultured on potato dextrose agar (PDA; Difco Laboratories, Detroit, MI, USA) medium at 25°C in the dark. Preparation of protoplasts and transformation of this fungus were carried out according to the polyethylene glycol-mediated transformation method described previously (35). Hygromycin- and Geneticin-resistant transformants were selected on plates with either 100 μg mL⁻¹ of hygromycin B or 100 μg mL⁻¹ of G418, respectively.

Sequence analyses.

The signal peptide of Fosp9 was predicted using SignalP-5.0 (https://services.healthtech.dtu.dk/service.php?SignalP-5.0). The protein domain or motif was predicted by performing Pfam analysis (http://pfam.xfam.org/search/). Homologous proteins of Fosp9 were identified by searching the database of nonredundant protein sequences (nr), using the BLASTP program, at the National Center for Biotechnology Information (NCBI) (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Homologous proteins of Fosp9 (identity, >40%; query cover, >85%) from other fungi were selected and further compared using ClustalW. The phylogenetic tree was generated by the neighbor-joining method using the MEGA program (version X), and bootstrapping was performed with 1,000 replicates. The software DISULFIND (http://disulfind.disi.unitn.it/) was used for predicting whether disulfide bonds could be formed by the cysteine residues in the mature Fosp9.

RNA extraction and transcription level analysis.

For gene expression analysis, mycelia and conidia of F. oxysporum f. sp. cubense wild-type strain B2 were inoculated into banana plantlet (Musa spp. AAA cv. Brazilian or AAA cv. Nan tian huang) roots, and fungal samples were collected at 24 and 48 hpi, respectively. For analyzing gene expression during conidium germination, mycelia were collected after conidia were cultured in potato dextrose broth for 24 h. All these experiments included three repeats, and samples were collected from each repeat. For RNA isolation, conidia or mycelia of fungal strains were extracted using an RNAprep Pure Plant kit (Tiangen, Beijing, China) according to the recommended protocol, and genomic DNA was removed using DNase I.

First-strand cDNA was synthesized using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China) according to the manufacturer’s instructions. Transcription levels were analyzed by qRT-PCR assays. qRT-PCR was performed on a QuantStudio 6 Flex real-time PCR system (Applied Biosystems) using TB Green Premix Ex Taq (Tli RNaseH Plus) (TaKaRa, Dalian, China). The actin gene of F. oxysporum f. sp. cubense was used as an internal control gene. Relative expression levels of each gene were calculated by the 2^−ΔΔCT method (36). The experiment was performed in triplicate and repeated three times. The resulting data were used to calculate the means and standard deviations. The primers used for qRT-PCR are listed in Table 1.

Function validation of the N-terminal signal peptide of Fosp9.

To examine the secretion function of the putative Fosp9 signal peptide, the yeast invertase secretion system was utilized. The predicted signal peptide sequences and the subsequent two amino acids of Fosp9 and SIX1a proteins from F. oxysporum f. sp. cubense were amplified by PCR using the primers listed in Table 1 and then fused in frame to the mature invertase gene SUC2 in the yeast signal sequence trap vector pSuc2t7M13ori (pSuc2) (37), resulting in the vectors pSuc2-Fosp9 and pSuc2-SIX1a, respectively. These vectors were then transformed into the auxotrophic and invertase mutant strain YTK12 using the reagent Quick & Easy yeast transformation mix (TaKaRa, Dalian, China). All transformants were cultured on the tryptophan-deficient CMD-W medium (0.67% yeast N base without amino acids, 0.075% dropout (DO) supplement −Trp, 0.1% glucose, 2% sucrose, 2% agar) for selecting positive colonies. To test for invertase secretion, the successfully transformed strains were cultured on YPRAA medium (2% peptone, 1% yeast extract, 2% raffinose, 2 mg mL⁻¹ antimycin A). The yeast strain YTK12 and the transformed strain carrying pSuc2-Mg87 were the negative controls (12), while the strain carrying pSuc2-SIX1a was the positive control. The yeast strain YTK12 carrying either the recombinant or empty pSUC2 vector can grow on CMD-W medium, while only the yeast strains expressing the signal peptide fragment fused in frame to the mature invertase gene Suc2 are able to secrete invertase and grow well on both CMD-W and YPRAA media (12). In addition, the culture filtrates of the yeast strains were used to confirm the secretion of invertase by an enzymatic activity test based on reduction of the dye 2,3,5-triphenyltetrazolium chloride (TTC) to the insoluble red compound triphenylformazan (12).

Subcellular localization assays and fluorescence microscopy.

To determine the subcellular localization of the Fosp9 protein, the whole open reading frame of the fosp9 gene was fused with the GFP gene and cloned into the NcoI-XbaI sites of vector pFGC5941 using an In-Fusion HD cloning kit (Clontech, USA) to generate vectors pFGC-Fosp9. This vector was transformed into Agrobacterium tumefaciens strain GV3101, and Agrobacterium-mediated transient gene expression in N. benthamiana leaves was carried out according to the method reported by Ma et al. previously (38). The treated leaves were collected, and green fluorescence was observed 2 to 3 days postinfiltration using an Olympus FluoView FV1000 confocal laser-scanning microscope at 559-nm excitation and 56- nm emission wavelengths.

Generation of fosp9 disruptants.

The upstream and downstream flanking sequences of the fosp9 gene were amplified using the primer sets sp9-1F/2R and sp9-3F/4R (Table 1), respectively. Then, the upstream flanking sequence was ligated to KpnI/XhoI-digested pCT74 plasmid to generate pCT74-fosp9L, and the downstream flanking sequence was ligated to EcoRI/XbaI-digested pCT74-fosp9L to produce pCT74-fosp9. The plasmids were linearized using KpnI/XbaI to generate the gene replacement construct. The gene replacement construct was then transformed into the prepared protoplasts of F. oxysporum f. sp. cubense by the polyethylene glycol-mediated transformation method described previously (39, 40). Each transformant was verified by PCR using three primer sets (upF1/upR2, dpF3/dpR4, and 9F5/9R6 [Table 1]) to detect the upstream-HPH segment, the GFP-downstream segment, and the target gene, respectively. Southern blot assays were further performed to confirmed the disruption of the fosp9 gene in those transformants.

Complementation of the fosp9 disruptant.

To test whether the observed phenotypes of fosp9 mutants were caused solely by the disruption of fosp9, a fragment containing the entire fosp9 gene and its 1,540 bp of upstream sequence was amplified with primers sp9-cF and sp9-cR, and cloned into the SalI/SacI-digested plasmid pSilent-Dual1 (pSD1) using the In-Fusion HD cloning kit (Clontech, USA), to generate pSD1-fosp9. pSD1-fosp9 was then transformed into protoplasts of the fosp9 disruptant sp9-4. The candidate complemented mutants were selected with G418 and confirmed by PCR using the primer set 9F5/9R6.

Mutations in cysteine residues of Fosp9.

To replace the two cysteine residues at amino acids 80 and 151 in Fosp9 with alanines, double point mutations were introduced in plasmid pSD1-fosp9 using the Fast site-directed mutagenesis kit (Tiangen, China) according to the manufacturer’s instructions with the two primer sets listed in Table 1. Mutations in pSD1-fosp9 were confirmed by DNA sequencing and ultimately used to create the pSD1-fosp9 (C80A, C151A) vector. This vector was transformed into protoplasts of the fosp9 disruptant sp9-4, and the candidate complemented mutants were selected by G418 and confirmed by PCR using the primer set 9F5/9R6.

Phenotypic characterization of fosp9 disruptants.

To analyze the influence of the fosp9 disruptants on the growth rate and conidiation, a 1-μl conidial suspension (2 × 10⁶ spores mL⁻¹) of each mutant or wild-type strain was placed on the center of PDA medium and cultured in the dark at 25°C. The diameters of the resulting colonies were measured, and conidia were counted 6 days postinoculation according to the method described previously (39). For H₂O₂ stress response analysis, the same volume of a conidial suspension of each strain containing 2 × 10⁶ spores mL⁻¹ was inoculated on the center of PDA supplemented with 6 mM H₂O₂ and cultured in the dark at 25°C. The diameters of the colonies were examined 6 days postinoculation. All of the experiments were repeated three times with three petri dishes in each replication.

Pathogenicity test, measurement of plant growth, and fungal biomass estimation.

Banana plantlets (Musa spp. AAA cv. Brazilian) at the three-leaf stage, which are susceptible to F. oxysporum f. sp. cubense race 4, were selected for pathogenicity testing using the root-dip method described previously (35). For each strain, 10 plantlets were used as the replicates, and pathogenicity tests were repeated three times, while the controls were inoculated with distilled water. After the treatments, banana plantlets were placed in a greenhouse with a 14-h photoperiod at 25°C. Disease severity in banana plantlets was evaluated, and data were collected at 45 days postinoculation according to the method reported previously (35, 39, 40). Meanwhile, plant height (in centimeters), fresh weight of above-ground part (in grams), stem diameter (in millimeters), and leaf area (in square centimeters) were measured according the methods described previously (41).

To analyze the fungal biomass inside plants, DNA was extracted from rhizome tissues of banana plants using an EZ-10 spin column plant genomic DNA purification kit (Sangon, Shanghai, China) following the protocol provided by the manufacturer. The levels of F. oxysporum f. sp. cubense and banana DNA were determined by quantitative PCR experiments using the primer pairs from the elongation factor 1-alpha gene of this fungus (FocEF1α; GenBank no. EXL94185) and the banana actin 1 gene (MusaActin1; GenBank no. HQ858237), respectively (Table 1). The relative fungal biomass was calculated by normalizing FocEF1α to MusaActin1 (42).

Determination of plant defense enzyme activities.

Banana leaves were collected at 45 days postinoculation from plantlets inoculated with fosp9 mutants or the wild-type strain. The activities of the plant defense enzymes peroxidase (POD), catalase (CAT), and phenylalanine ammonia-lyase (PAL) in the banana leaves were measured using POD, CAT, and PAL activity detection kits (Solarbio, Beijing, China), respectively, according the manufacturer’s instructions. Each experiment included triplicate samples and was repeated three times. One unit of POD activity was defined as the amount causing an increase of 0.01 in A₄₇₀ per min, and 1 U of catalase activity was defined as the amount of enzyme causing 1 nmol of H₂O₂ decomposition per min, while 1 U of PAL activity was defined as the amount of enzyme causing an increase of 0.1 in A₂₉₀ per min. The activities of all these enzymes were expressed as units per gram (fresh weight).

RNA-seq and validation of data by qRT-PCR.

The mycelial and conidial mixtures of either the fosp9-disruptant sp9-4 or the wild-type strain were inoculated into banana (Musa sp. AAA cv. Brazilian) roots in a hydroponics system for 48 h. Each fungal sample was then harvested, immediately frozen in liquid nitrogen, and maintained below −70°C until RNA isolation. Total RNA (5 μg) from each sample was prepared for Illumina RNA-seq. Purification of poly(A) mRNA, synthesis of cDNA, and the sequencing library construction were carried out according to methods reported previously (43). The cDNA library was sequenced using the Illumina HiSeq 2000 platform with a paired-end (single reads of 150 bases) sequencing strategy at Guangdong Longsee Biomedical Corporation, Guangzhou, China. To obtain clean reads, raw sequencing reads were subjected to a trimming process with Trimmomatic software (version 0.32) (44). Subsequently, clean reads were aligned to the F. oxysporum f. sp. lycopersici reference genome (GenBank assembly accession no. GCA_000149955) with HISAT2 software (version 2.0.5) (45). For quantification, gene expression levels were calculated in fragments per kilobase of exon model per million mapped reads (FPKM) using Cuffdiff software in the Cufflink package (46). Differentially expressed genes (DEGs) between control and disruptant samples were identified using a false discovery rate (FDR) of ≤0.001 and an absolute value of the log₂ ratio of ≥1 as the threshold. To characterize the biological functions of the DEGs and the metabolic pathways in which they are involved, the DEGs were subjected to a Gene Ontology (GO) functional analysis using topGO and a KEGG pathway enrichment analysis with clusterProfiler R package (47). The significant KEGG pathways for DEGs were selected using a P value cutoff of ≤0.05. To validate the RNA-seq results, 12 DEGs were randomly selected and qRT-PCR assays were performed as described above using the primers listed in Table 1.

Data availability.

Raw data were deposited in the National Center for Biotechnology Information (NCBI) database under the SRA study accession number PRJNA654435.