Fusarium mindanaoense sp. nov., a New Fusarium Wilt Pathogen of Cavendish Banana from the Philippines Belonging to the F. fujikuroi Species Complex

The pathogen causing Fusarium wilt in banana is reported to be Fusarium oxysporum f. sp. cubense (FOC). In 2019, wilt symptoms in banana plants (cultivar: Cavendish) in the Philippines were detected, i.e., the yellowing of the leaves and discoloration of the pseudostem and vascular tissue. The fungus isolated from the vascular tissue was found to be pathogenic to Cavendish bananas and was identified as a new species, F. mindanaoense, belonging to the F. fujikuroi species complex (FFSC); species classification was assessed using molecular phylogenetic analyses based on the tef1, tub2, cmdA, rpb1, and rpb2 genes and morphological analyses. A reciprocal blast search using genomic data revealed that this fungus exclusively included the Secreted in Xylem 6 (SIX6) gene among the SIX homologs related to pathogenicity; it exhibited a highly conserved amino acid sequence compared with that of species in the FFSC, but not with that of FOC. This was the first report of Fusarium wilt in Cavendish bananas caused by a species of the genus Fusarium other than those in the F. oxysporum species complex.


Introduction
Bananas are one of the most common agricultural exports, though they are also widely consumed in the countries that produce them [1]. Musa sapientum cv. Cavendish (AAA group) is exclusively cultivated in many tropical countries as a commercial crop; 21 million tons of Cavendish bananas were exported in 2019 (FAO 2022). In the 1950s, the planting of Cavendish bananas instead of cv. Gros Michel began to increase worldwide because of an epidemic of Fusarium wilt disease (Panama disease) caused by Fusarium oxysporum f. sp. cubence (FOC) race 1. Thereafter, the causal pathogens of Fusarium wilt disease in bananas were found and characterized as race 1, race 2, race 4, subtropical race 4 (STR4), and tropical race 4 (TR4) based on their pathogenicity. In the 1990s, TR4 was identified as the causal agent of Fusarium wilt disease infecting Cavendish bananas in Taiwan [2]. The disease caused by TR4 has been reported in 23 countries (predominantly in Southeast Asia, South Asia, Africa, and Latin America) [3]. Moreover, TR4 affected the banana industry and reduced banana yield in the Philippines (FAO 2022). Mostert et al. [4] and Solpot et al. [5] investigated the Fusarium wilt pathogen in the Philippines; mostly TR4 (VCG01223/16), J. Fungi 2023, 9,443 2 of 15 and less commonly R4 (VCG0122), was detected in Mindanao. However, information on FOC other than TR4 is scarce and no reports of other Fusarium species are available.
In this study, during a survey of the Fusarium wilt disease in Mindanao conducted in September 2019, a new species was found that belonged to the F. fujikuroi species complex (FFSC); it caused symptoms of leaf yellowing ( Figure 1A) to emerge in older leaves and a reddish-brown discoloration of the pseudostem and vascular tissues of bananas (Cavendish) ( Figure 1B,C). This pathogen had not been previously reported to cause Fusarium wilt in banana. Therefore, we aimed to identify this causal agent, conducted molecular and morphological analyses, and proposed an isolate as a new pathogenic species of Fusarium wilt.
banana industry and reduced banana yield in the Philippines (FAO 2022). Mostert et al. [4] and Solpot et al. [5] investigated the Fusarium wilt pathogen in the Philippines; mostly TR4 (VCG01223/16), and less commonly R4 (VCG0122), was detected in Mindanao. However, information on FOC other than TR4 is scarce and no reports of other Fusarium species are available.
In this study, during a survey of the Fusarium wilt disease in Mindanao conducted in September 2019, a new species was found that belonged to the F. fujikuroi species complex (FFSC); it caused symptoms of leaf yellowing ( Figure 1A) to emerge in older leaves and a reddish-brown discoloration of the pseudostem and vascular tissues of bananas (Cavendish) ( Figure 1B,C). This pathogen had not been previously reported to cause Fusarium wilt in banana. Therefore, we aimed to identify this causal agent, conducted molecular and morphological analyses, and proposed an isolate as a new pathogenic species of Fusarium wilt.
Furthermore, to provide fundamental information relating to factors of its pathogenicity, we searched the Secreted in Xylem (SIX) genes from the whole-genome data of the new species. In addition, we predicted whether the genome of the new Fusarium species obtained SIX genes via horizontal transfer from FOC or by other means; it is reported that SIX genes are one of the most important factors for infecting banana plants [6][7][8].

Sample Collection and Fungal Isolation
Symptomatic banana plants were collected from a farm in Mindanao in 2019. The discolored vascular tissues ( Figure 1C) of the pseudostem were cut into pieces of approximately 3 mm 2 , which were then sterilized with 0.6% (v/v) sodium hypochlorite for 1 min, washed with sterilized water, dried with sterilized paper, and placed on a water agar (WA) plate. The hyphae that emerged on WA were transferred onto a potato dextrose agar (PDA) plate to produce conidia for monoculture. The PD20-05 isolate was maintained on a PDA plate.

Genomic DNA Extraction
DNA was extracted from the mycelia of each isolate, which were grown for 7-10 days in yeast glucose medium using the modified CTAB method [9]. After treatment with Furthermore, to provide fundamental information relating to factors of its pathogenicity, we searched the Secreted in Xylem (SIX) genes from the whole-genome data of the new species. In addition, we predicted whether the genome of the new Fusarium species obtained SIX genes via horizontal transfer from FOC or by other means; it is reported that SIX genes are one of the most important factors for infecting banana plants [6][7][8].

Sample Collection and Fungal Isolation
Symptomatic banana plants were collected from a farm in Mindanao in 2019. The discolored vascular tissues ( Figure 1C) of the pseudostem were cut into pieces of approximately 3 mm 2 , which were then sterilized with 0.6% (v/v) sodium hypochlorite for 1 min, washed with sterilized water, dried with sterilized paper, and placed on a water agar (WA) plate. The hyphae that emerged on WA were transferred onto a potato dextrose agar (PDA) plate to produce conidia for monoculture. The PD20-05 isolate was maintained on a PDA plate.

Genomic DNA Extraction
DNA was extracted from the mycelia of each isolate, which were grown for 7-10 days in yeast glucose medium using the modified CTAB method [9]. After treatment with chloroform-isoamyl alcohol (24:1), 2-Mercaptoethanol and 10% CTAB at 0.2% and 2%, respectively, were added to the supernatant and incubated for 40 min at 60 • C. After incubation, an equal volume of chloroform-isoamyl alcohol (24:1) was added, mixed gently for 10 min, and centrifuged for 10 min at 12,000 rpm for purification. The aqueous phase was carried out, and the above-mentioned purification was again conducted. Precipitation was achieved by adding 2.5 and 0.1 times the volume of ethanol and 3 M sodium acetate, respectively, which was then mixed for a short period of time and centrifuged for 10 min at 12,000 rpm. After removing the liquid, a pellet of DNA at the bottom was dried and dissolved with 30 µL of TE buffer.

Gene Prediction
Genome DNA was sequenced using the Illumina HiSeq genome analyzer platform and DNA libraries and paired-end (PE) genomic libraries were generated. The libraries were sequenced in PE mode with 150 bp reads on the Illumina HiSeq X instrument. Adaptors were eliminated from reads using the Trimmomatic read trimming tool for Illumina NGS data, with a quality cut-off of 30. The raw mate-pair read sequence quality was checked using FastQC vers. 0.11.8 [10] (http://www.bioinformatics.babraham.ac.uk/projects/fastqc accessed on 12 December 2022). Platanus allee vers. 2.0.2 [11] was used to assemble the reads and obtain contig data. The N50 values were calculated to measure the quality of the assemblies. Augustus 3.3.3 [12] was used to perform gene predictions using F. graminearum data as a reference.

Phylogenetic Analyses
Molecular analyses were conducted to identify the pathogen. To select the DNA sequences, the translation elongation factor 1-alpha (tef1), beta-tubulin (tub2), calmodulin (cmdA), RNA polymerase large subunit (rpb1), and RNA polymerase second-largest subunit (rpb2) genes were amplified according to the method reported by Yilmaz et al. [13]; the genes were then sequenced using the following primer pairs: EF1 and EF2 [14], T1 and T2 [15], CL1 and CL2A [16], Fa [17] and R8 [18], and 5F2 [19] and 7cr [20], respectively. The sequence data were deposited in the DNA Data Bank of Japan (Table 1). One hundred and three sequences (Table 1) of each DNA region were aligned using Clustal W in MEGA 7 [21], concatenated, and subjected to phylogenetic analyses using the maximum-likelihood (ML), maximum-parsimony (MP), and neighbor-joining (NJ) methods. The reliability of the branches on the phylogenetic tree was evaluated using the bootstrap (BS) [22] test with 1000 replicates.

Morphological Analyses
Mycelial plugs (ϕ7 mm) of the isolate were placed in the center of the potato dextrose agar (PDA), synthetic nutrient-poor agar (SNA) [23], and oatmeal agar (OA) [24] plates and incubated for 6 days at 25 • C in the dark. The colony character on the surface and reverse sides was observed. The isolate (PD20-05) was cultured on carnation leaf agar (CLA) [25] and SNA, inducing sporodochial conidia and microconidia to observe its asexual morphological characteristics. Thirty conidia and conidiophores were observed under a light microscope (BX51, Olympus, Tokyo, Japan) to record their shape and size. For the mycelial growth test, 6 d cultures of the isolate grown at 25 • C on PDA plates were used. Mycelial plugs (ϕ7 mm) were then placed on the center of the PDA plates. These plates were incubated at 4 • C, 10 • C, 15 • C, 20 • C, 25 • C, 30 • C, and 35 • C in the dark. After incubation for 6 days, mycelial growth per day was calculated. The average growth rate per day for each temperature was determined from five replicates.

Pathogenicity Test
A pathogenicity test was conducted using a conidial suspension in sterilized water adjusted to 1 × 10 7 conidia/mL. Six Cavendish seedlings were used in this experiment, in which the roots of three seedlings were soaked in 500 mL of the conidial suspension for 3 h, before being planted in pots with a 1:1 mixture of red ball earth and humus. A 5 g/L solution of NPK 8-8-8 was added as a chemical fertilizer. The remaining three seedlings were treated with sterilized water as a control. The treated and control plants were inoculated for 34 days at 25 • C with an 8 h light/16 h dark cycle.

Detection of Secreted in Xylem Genes among the Whole-Genome Data
A tBLASTn analysis was conducted to identify the SIX genes using BLAST 2.11.0+ software [26,27]. In this analysis, previously reported SIX gene protein sequences were obtained from the NCBI protein database and were used as queries against the assembled whole-genome sequence (e-value = 0.001). The identified new SIX gene sequence was reciprocally BLAST searched (BLASTP) against the NCBI-deposited protein sequences of SIX genes to estimate their sequence similarity. The SignalP program (v. 5.0) [28] was used to identify the new SIX protein code signal peptides.

Identification of the Homologous SIX Genes of PD20-05 in the Foc TR4 Genome
To search for the homologous SIX genes of PD20-05 in Fusarium oxysporum f. sp. cubense (FOC) TR4, we initially constructed the genome sequence of FOC TR4 using the NGS data (SRR10054447) [29] and Platanus-allee. Subsequently, the SIX gene of PD20-05 was used as a query against the assembled FOC TR4 genome sequence (e-value ≤ 0.001).
To clarify whether the SIX6 gene identified in our isolate was from FOC, the putative SIX6 gene protein sequences of our isolate and those of FFSC and FOC were aligned using clustalW in MEGA7 [21], followed by the construction of an NJ tree with the option to completely delete the gap. The reliability of the branches of the phylogenetic tree was evaluated via the BS test [22] with 1000 replicates.

Phylogenetic Analysis
For the phylogenetic analysis using the five loci, the final dataset included 2606 positions (excluding gaps and including sites), comprising 462, 280, 521, 594, and 749 positions from tef1, tub2, cmdA, rpb1, and rpb2 gene sequences, respectively. The PD20-05 isolate was independent of known species and a sister lineage of the F. sacchari clade (BS value = 76; Figure 2).
Colonies on PDA reached 50-54 mm diam. after 6 d at 25 • C in the dark; the colonies were raised, aerial mycelia dense, covered the entire margin and surface in the center, and were white at the edge. Colonies were also reverse pink in the center and white at the margin. Hyphae grew at 10-35 • C, with an optimum temperature of 25 • C (avg. ± sd. 4.4 ± 0.16 mm/day; Figure 4). Colonies on OA reached 66-68 mm diam. after 6 d at 25 • C in the dark were raised, aerial mycelia dense, and covered colony margin entire; they were also surface white and reverse white. Colonies on SNA reached 59-61 mm diam. after 6 days at 25 • C in the dark; these colonies were raised, aerial mycelia sparse, covered the entire colony margin entire, and were surface white and reverse white.
Note: F. mindanaoense resembled F. concentricum regarding the size of sporodochial conidia (Table 2). However, F. mindanaoense could be distinguished by the characteristics of colonies. F. mindanaoense did not produce concentric aerial hyphae in its mycelium (Figure 3A), while F. concentricum did produce this symptom of fungal infection. A holotype and ex-holotype strain were deposited at Flora and Fauna Analytical and Diagnostic Center at Central Luzon State University. Note: F. mindanaoense resembled F. concentricum regarding the size of sporodochial conidia (Table 2). However, F. mindanaoense could be distinguished by the characteristics of colonies. F. mindanaoense did not produce concentric aerial hyphae in its mycelium ( Figure 3A), while F. concentricum did produce this symptom of fungal infection. A holotype and ex-holotype strain were deposited at Flora and Fauna Analytical and Diagnostic Center at Central Luzon State University.

Pathogenicity Test
Yellow leaves appeared on inoculated plants after 20-34 days; one dried-up seedling and leaves of other seedlings closed around the main veins ( Figure 5A,B). Part of the internal tissues of the corms turned black, while the tissues of pseudostem just above the corn were reddish-brown ( Figure 5C,D). Additionally, the roots turned black all around ( Figure 5E,F). The inoculated strain was re-isolated from the discolored roots and vascular lesions, whereas the control plants treated with water exhibited no symptoms.

Pathogenicity Test
Yellow leaves appeared on inoculated plants after 20-34 days; one dried-up seedling and leaves of other seedlings closed around the main veins ( Figure 5A,B). Part of the internal tissues of the corms turned black, while the tissues of pseudostem just above the corn were reddish-brown ( Figure 5C,D). Additionally, the roots turned black all around ( Figure 5E,F). The inoculated strain was re-isolated from the discolored roots and vascular lesions, whereas the control plants treated with water exhibited no symptoms.

Detection of Secreted in Xylem Genes in Whole-Genome Data
In this study, the genomic DNA of PD20-05 was sequenced and assembled into 3377 contigs with an N50 of 53.7 kb and a maximum length of 164.3 kb. In the tBLASTn analysis, SIX gene sequences were searched in the PD20-05 genome using the 1186 NCBIdeposited SIX protein sequences. The analysis showed that the SIX6 gene was the only SIX gene found in the PD20-05 genome sequence. Moreover, the SIX6 protein of PD20-05 was predicted to contain signal peptides ( Figure S1). Furthermore, a reciprocal BLASTp analysis showed that the SIX6 protein sequence of PD20-05 was identical to that of Fusarium sp. NRRL 25303, F. proliferatum, F. globosum, F. agapanthi, F. denticulatum, F. tjaetaba, F. napiforme, F. pseudocircinatum, F. circinatum, F. phyllophilum, F. mundagurra, and F. pseudoanthophilum (Table 3), which all belong to the FFSC species and made one clade in the phylogenetic tree ( Figure 6). These SIX6 gene sequences were greatly different from those of two FOC strains (accession nos. KX435007 and KX435008) [8], as assessed based on the alignment ( Figure S1). Two types of SIX6 genes in the F. hostae (HY9) genome were obtained by conducting a BLASTp using the SIX6 gene sequences of FOC (BRIP628956) and F. mindanaoense (PD20-05) as query sequences with low e-values (3 × 10 −82 and 2 × 10 −120 , respectively). Van Dam and Rep [33] reported that the strain acquired one SIX6 gene via horizontal transfer from the FOC. The two types of SIX6 genes fell into different clades in the phylogenetic tree ( Figure 6). ysis showed that the SIX6 protein sequence of PD20-05 was identical to that of Fusarium sp. NRRL 25303, F. proliferatum, F. globosum, F. agapanthi, F. denticulatum, F. tjaetaba, F. napiforme, F. pseudocircinatum, F. circinatum, F. phyllophilum, F. mundagurra, and F. pseudoanthophilum (Table 3), which all belong to the FFSC species and made one clade in the phylogenetic tree ( Figure 6). These SIX6 gene sequences were greatly different from those of two FOC strains (accession nos. KX435007 and KX435008) [8], as assessed based on the alignment ( Figure S1). Figure 6. NJ phylogenetic tree based on SIX6 gene sequences. The genes highlighted in red are the SIX6 genes of FFSC, whereas the genes highlighted in green are the SIX6 genes of FOC and F. hostae. F. hostae has two types of SIX6 genes. One belongs to FFSC type ( a ), and another one belongs to FOC type ( b ).
Two types of SIX6 genes in the F. hostae (HY9) genome were obtained by conducting a BLASTp using the SIX6 gene sequences of FOC (BRIP628956) and F. mindanaoense (PD20-05) as query sequences with low e-values (3e-82 and 2e-120, respectively). Van Dam and Rep [33] reported that the strain acquired one SIX6 gene via horizontal transfer from the Figure 6. NJ phylogenetic tree based on SIX6 gene sequences. The genes highlighted in red are the SIX6 genes of FFSC, whereas the genes highlighted in green are the SIX6 genes of FOC and F. hostae. F. hostae has two types of SIX6 genes. One belongs to FFSC type ( a ), and another one belongs to FOC type ( b ).

Discussion
Fusarium sacchari (leaf blight on AAA genome group and fruit rot on AAA), F. proliferatum (fruit rot on AAB and sheath rot on ABB), F. fujikuroi (fruit rot on AA), F. concentricum (fruit rot on AAA), F. verticillioides (fruit rot on Musa sp.), and F. musae (fruit rot on Musa sp.) belonging to FFSC were reported as banana pathogens [34][35][36][37][38][39][40][41]. These species do not cause Fusarium wilt of bananas. Maryani et al. [39] also isolated F. proliferatum from a symptomatic tissue of Fusarium wilt of banana (AA) in 2019, concluding that the fungus was not a pathogen of Fusarium wilt of banana (Cavendish: AAA); rather, it was an endophyte because it caused only a slight discoloration in the corm without any further disease development. Moreover, in 2022, Thi et al. [42] also isolated FFSC species (F. fujikuroi) from symptomatic tissues of Fusarium wilt of banana (ABB). However, a pathogenicity assay was not carried out. To the best of our knowledge, no FFSC species have been reported as the pathogen underlying Fusarium wilt in banana. In this study, we identified a new causal agent, F. mindanaoense (which belongs to FFSC), of Fusarium wilt in banana in the Philippines. This is the first report of Fusarium wilt in banana caused by a fungus belonging to the FFSC.
As FOC affects Cavendish bananas, research on FOC has focused on managing Fusarium wilt disease. Therefore, rapid detection methods for FOC, such as loop-mediated isothermal amplification and PCR detection, have been developed for diagnosis and occurrence monitoring [43][44][45]. Our study reveals that a pathogen belonging to the FFSC also caused Fusarium wilt in the Cavendish banana. Focusing on FOC and other pathogenetic fungi to acquire basic knowledge that may contribute to controlling Fusarium wilt is necessary.
The SIX genes play a role in the pathogenicity of Fusarium wilt; SIX1, SIX2, SIX6, SIX7, SIX9G1, SIX11, and SIX13 were detected in the FFSC species [34,37]. The present study showed that the F. mindanaoense genome possessed the SIX6 gene exclusively, which matched with those of the FFSC with low e-values (Table 2; 0-3.92 × 10 −125 ). Van Dam and Rep [33] reported that the SIX6 gene from F. hostae (HY9), which belongs to the FFSC species, was horizontally transferred from FOC. We found that F. hostae (HY9) has two types of SIX6 genes: the FOC and FFSC groups ( Figure 6). Because the gene sequence of F. mindanaoense that was identified as the SIX6 gene did not belong to a clade of FOC, and one of the SIX6 genes obtained from F. hostae genome belonged to the FFSC in the phylogenetic tree, F. mindanaoense was thought not to have acquired its pathogenicity through horizontal gene transfer from FOC ( Figure 6). However, a functional analysis of the SIX6 gene of the FFSC is warranted to clarify whether the SIX6 gene acts as a functional gene in the pathogenicity of Fusarium wilt.