Prevalence of Carbendazin Resistance in Field Populations of the Rice False Smut Pathogen Ustilaginoidea virens from Jiangsu, China, Molecular Mechanisms, and Fitness Stability

Rice false smut (RFS), caused by Ustilaginoidea virens, is an important fungal disease of rice. In China, Methyl Benzimidazole Carbamates (MBCs), including carbendazim, are common fungicides used to control RFS and other rice diseases. In this study, resistance of U. virens to carbendazim was monitored for three consecutive years during 2018 to 2020. A total of 321 U. virens isolates collected from Jiangsu Province of China were tested for their sensitivity to carbendazim on PSA. The concentration at which mycelial growth is inhibited by 50% (EC50) of the carbendazim-sensitive isolates was 0.11 to 1.38 µg/mL, with a mean EC50 value of 0.66 μg/mL. High level of resistance to carbendazim was detected in 14 out of 321 isolates. The resistance was stable but associated with a fitness penalty. There was a statistically significant and moderate negative correlation (r= −0.74, p < 0.001) in sensitivity between carbendazim and diethofencarb. Analysis of the U. virens genome revealed two potential MBC targets, Uvβ1Tub and Uvβ2Tub, that putatively encode β-tubulin gene. The two β-tubulin genes in U. virens share 78% amino acid sequence identity, but their function in MBC sensitivity has been unclear. Both genes were identified and sequenced from U. virens sensitive and resistant isolates. It is known that mutations in the β2-tubulin gene have been shown to confer resistance to carbendazim in other fungi. However, no mutation was found in the Uvβ2Tub gene in either resistant or sensitive isolates. Variations including point mutations, non-sense mutations, codon mutations, and frameshift mutations were found in the Uvβ1Tub gene from the 14 carbendazim-resistant isolates, which have not been reported in other fungi before. Thus, these results indicated that variations of Uvβ1Tub result in the resistance to carbendazim in field isolates of Ustilaginoidea virens.


Introduction
Rice false smut (RFS), caused by the filamentous fungus Ustilaginoidea virens (teleomorph: Villosiclava virens), is responsible for significant losses in the rice industry. RFS has long been a minor disease in rice production and occurs sporadically in some rice-growing regions, such as south and east Asia. However, it has recently become one of the most important diseases in most rice-growing regions of the world [1]. The false smut pathogen (U. virens) infects the plant during the flowering stage [2]. The infected grains convert first into whitish, yellowish-orange to green chlamydospores, which later turn greenish-black in colour [3]. Worse still, the U. virens can produce cyclopeptide mycotoxins, which are poisonous to humans and livestock and pose a serious problem for food security and rice production [4,5].
To date, although some RFS-resistant rice cultivars are commercially produced, disease control of RFS relies mainly on fungicide application [1]. An ideal prevention effect can be Isolates of U. virens used in this study were obtained during 2018 to 2020 from Jiangsu Province, which is the major rice production province in China. The rice fields are located in 10 counties and cities, namely Gaoyou, Jurong, Nanjing, Taizhou, Xinhua, Xuyi, Yizheng, Yangzhou Guanglin (YZ Guanglin), Yangzhou Hanjiang (YZ Hanjiang), and Zhenjiang ( Figure 1). Rice varieties grown in those areas include Wanxian No.98, Najing No. 9108, Nanjing No. 46, Nanjing No. 5718, Nanjing No. 5055, Yangchannuo No. 1, Huiliangyou No. 898, and Yongyou No. 2640. Three to five rice plants with symptoms of rice false smut were randomly selected from various blocks in each field and preserved at 4 • C. Individual false smut balls collected from different rice plants served as the source for all isolates. The false smut balls were divided in half and placed on PSA plates amended with streptomycin sulfate after being surface-sterilized in 0.1% sodium hypochlorite for 5 min. The plates were cultured in a growth chamber for 3-7 days at 27 • C (12 h photoperiod). Pure cultures were obtained by transfer of a single germinated chlamydospore and maintained on PSA slants at 4 • C for 2-4 weeks. A total of 321 isolates were collected throughout Jiangsu Province.
Yizheng, Yangzhou Guanglin (YZ Guanglin), Yangzhou Hanjiang (YZ Hanjiang), and Zhenjiang ( Figure 1). Rice varieties grown in those areas include Wanxian No.98, Najing No. 9108, Nanjing No. 46, Nanjing No. 5718, Nanjing No. 5055, Yangchannuo No. 1, Huiliangyou No. 898, and Yongyou No. 2640. Three to five rice plants with symptoms of rice false smut were randomly selected from various blocks in each field and preserved at 4 °C. Individual false smut balls collected from different rice plants served as the source for all isolates. The false smut balls were divided in half and placed on PSA plates amended with streptomycin sulfate after being surface-sterilized in 0.1% sodium hypochlorite for 5 min. The plates were cultured in a growth chamber for 3-7 days at 27 °C (12 h photoperiod). Pure cultures were obtained by transfer of a single germinated chlamydospore and maintained on PSA slants at 4 °C for 2-4 weeks. A total of 321 isolates were collected throughout Jiangsu Province.

In Vitro Sensitivity Determination of U. virens to Carbendazim
Sensitivity to carbendazim was assessed on fungicide-amended PSA at 0, 0.03, 0.1, 0.3, 1, 3, 10, 30, and 60 µg a.i./mL. Carbendazim-resistant isolates were subjected to a second series of sensitivity assays involving PSA amended with carbendazim at 0, 100, 250, and 500 µg a.i./mL. To inoculate test plates, mycelial plugs were removed with a 4 mm cork borer from the margins of 14-day-old colonies and placed upside down on the centers of 9 cm plastic Petri dishes containing the fungicide-amended or unamended media. Each isolate was tested three times, and the plates were incubated at 27 • C for 21 days in darkness. For each treatment, the mean colony diameter (minus the diameter of the inoculation plug) was measured and expressed as a percentage of growth inhibition. The EC 50 values were calculated by regression analysis of probit values of corresponding percent growth inhibitions against logarithms of fungicide concentrations. Baseline sensitivity of carbendazim was constructed on the basis of frequency distribution of logarithmically transformed EC 50 values of the 321 isolates.

Investigation of Mycelial Growth, Mycelial Dry Weight, Conidiation, and Conidial Germination Rate In Vitro
Mycelial growth diameters of 14 resistant isolates and 5 sensitive isolates were determined on fungicides-free PSA plates. Mycelium plugs (4 mm in diameter) taken from the periphery of 14-day-old cultures were transferred to new PSA plates and incubated at 27 • C in the dark. The colony diameters were measured after 14 days, with three replicates per isolates. Fourteen-day-old mean colony diameters of all resistant isolates or sensitive isolates were calculated as mycelial growth diameters. For conidiation test and measurement of conidia germination, after isolates were cultured on PSA at 27 • C for 14 days, two 4 mm diameter mycelia plugs were transferred into 50 mL of PSB, and the conidiation was counted with a hemocytometer after shaking at 27 • C, 150 rpm for 7 days. The mycelial were collected by filtration through two layers of gauze and measured for dry weight after drying at 50 • C in an oven for 3 days. To measure conidial germination rate, mycelial were removed by filtration and the conidia were collected from the filtrate by centrifugation at 7000× g for 5 min. The conidia were washed twice by resuspension in sterile distilled water, in which they were finally resuspended and adjusted to a concentration of 1.0 × 10 5 conidia/mL. Then, 100 µL conidia suspension was spread on the surface of PSA or WA (water agar: 15 g agar per liter water) plates, and the plates were incubated at 27 • C in the dark for 24 h. The conidial germination rate was recorded with a Nikon E200 microscope. Conidia was regarded as germinated when the germ tube length exceeded half the conidia length. Germination was quantified by counting at least 200 conidia per plate for each isolate. Each isolate was repeated three times and each experiment was conducted twice.

Stability of Resistant to Carbendazim
Mycelium plugs were taken from the periphery of the colonies and transferred to fresh fungicide-free PSA, then the plate was incubated at 27 • C for 10 days. The process was repeated for 5 generations, but new plates were always inoculated with the colony of the previous generation. EC 50 values for carbendazim were established before the first and after the last transfer, as described above.

Correlation Analysis in Sensitivity of Carbendazim with Diethofencarb, Azoxystrobin, Pyraclostrobin, and Tebuconazole
For the correlation analysis assay, the sensitivity of nine carbendazim-resistant mutants and nine carbendazim-sensitive isolates to diethofencarb, azoxystrobin, pyraclostrobin, and tebuconazole were determined by measuring mycelial growth. Sensitivity to diethofencarb, azoxystrobin, pyraclostrobin, and tebuconazole was assessed on fungicide-amended PSA at 0, 0.03, 0.1, 0.3, 1, 3, 10, and 30 µg a.i./mL. For carbendazim-sensitive isolates, PSA media were amended with diethofencarb at final concentrations of 0, 10, 30, 100, 250, and 500 µg a.i./mL. In sensitivity determination to azoxystrobin and pyraclostrobin, salicylhydroxamic acid (SHAM) was not included in the medium, since all tested isolates were sensitive to these two QoI fungicides, and SHAM showed strong toxicity to U. virens, according to our previous studies [12,20]. The EC 50 value was estimated as previously described. The experiments were performed three times with four replicates plates per concentration. The cross-resistance relationships were analyzed by the correlation procedure in GraphPad Prism (version 5.01, GraphPad Software, San Diego, CA, USA).

Cloning and Sequence Analysis
Fourteen carbendazim resistant isolates and twenty-three sensitive isolates collected from different cities were selected for Uvβ1Tub and Uvβ2Tub analysis. All of them were grown on PSA at 27 • C for 14 days in the dark. Single agar plugs containing actively growing mycelium were transferred to 100 mL flasks containing 50 mL of PSB. Flasks were shaken at 150 rpm for 7 days at 27 • C. Mycelium was filtered from the broth, rinsed under sterile deionized water, and subjected to DNA extraction using the DNeasy Plant Mini Kit (Qiagen Inc., Valencia, CA, USA) according to the manufacturer's instructions.
On the basis of the whole nucleotide sequence of Uvβ1Tub (Gene ID: UV8b_05680) and Uvβ2Tub (Gene ID: UV8b_05383) from the sequenced isolate UV8b, the primer pairs tub1-5N/tub1-3N: ACAGTGATGCGTGATGCGAT/TGTTGGCTCAACGAGGTCAA were designed to amplify the Uvβ1Tub and its 711-bp upstream and 459-bp downstream fragment, and the primer pairs tub2-F/tub2-R: GGTACTCCGTAAACGTAATC/TCACCCTTCT GCTGGTTGCG were designed to amplify the Uvβ2Tub and its upstream fragment from the isolates analyzed. The polymerase chain reaction (PCR) mixtures contained 1 × PCR buffer, 20 ng of template DNA, 0.6 mM each primer, 200 mM each dNTP, and 1 U of Taq DNA polymerase (New England Biolabs, Ipswich, MA, USA). The following PCR parameters were used: an initial denaturation step of 5 min at 94 • C; followed by 35 cycles of 30 s at 94 • C, 30 s at 55 • C, and 3 min at 72 • C; and a final extension step of 7 min at 72 • C.

Phylogenetic Analysis
Amino acid sequences of β1Tub and β2Tub from U. virens and other plant pathogens were obtained from the public database GenBank. Multiple alignments were conducted using DNASTAR (DNASTAR Inc., Nevada City, CA, USA) and CLUSTAL X v. 2.1. For the ML method, the phylogram was inferred on the basis of the JTT matrix-based model. The bootstrap consensus tree, inferred from 1000 replicates, was constructed with the following parameters: the Poisson correction model, gamma distribution (five categories), and heuristic method using SPR-extensive. All positions containing gaps and missing data were eliminated.

Data Analysis
Significant differences of EC 50 values from different populations were evaluated by one-way analysis of variance with a least significant difference test in SPSS Software (version 22.0; IBM SPSS Inc. Chicago, IL, USA). To determine cross-resistance in isolates to fungicides, the EC 50 values were correlated, and the correlation coefficients (r) were calculated by SPSS. There were three replicates of each concentration for each isolate.

Sensitivity of the Field Isolates to Carbendazim
In total, 321 isolates were collected from 10 cities in Jiangsu Province. The average frequency of resistance to carbendazim was 4.36% (Table 1). Except for Gaoyou, Nanjing, and Taizhou, carbendazim-resistant isolates were found in all of the other cities. The EC 50 values of the resistant isolates were more than 500 µg/mL for most of the resistant isolates. The resistance factors (defined as the ratio of EC 50 for a fungicide resistant isolate relative to the mean EC 50 of the sensitive isolates) were approximately 100 or even higher.
However, as for the sensitive isolates, carbendazim was quite effective, with EC 50 range from 0.108 to 1.378 µg/mL and mean EC 50 value of 0.663 µg/mL (Table 1). Among the sensitive isolates, there was no difference (p > 0.05) in mean EC 50 values of isolates from different cities.
The frequency distribution of the EC 50 values for carbendazim was unimodal ( Figure 2A). In terms of years, although no resistant isolates were detected in 2020, the proportion of isolates insensitive to carbendazim in 2020 was higher than those in 2018 and 2019 ( Figure 2B). range from 0.108 to 1.378 µg/mL and mean EC50 value of 0.663 µg/mL (Table 1). Among the sensitive isolates, there was no difference (p > 0.05) in mean EC50 values of isolates from different cities. The frequency distribution of the EC50 values for carbendazim was unimodal (Figure 2A). In terms of years, although no resistant isolates were detected in 2020, the proportion of isolates insensitive to carbendazim in 2020 was higher than those in 2018 and 2019 ( Figure 2B).

Fitness and Stability of Resistant Isolates
Four fitness components were tested for all 14 resistant isolates and 5 sensitive isolates. Overall, fitness penalties were recorded for the resistant isolates. The mycelial growth diameters of the resistant isolates were significantly lower than those of the sensitive isolates ( Figure 3A). In addition, mycelial dry weight was lower in the resistant isolates, but the difference was statistically non-significant ( Figure 3B, p > 0.05). The resistant isolates produced significantly fewer conidia in PSB compared with the sensitive isolates ( Figure 3C). Since most of resistant isolates produced no or little conidia, five resistant isolates (GL11, HA17, XH5a, XH7b, and XH43b) were tested for conidial germination rate on PSA and WA media. The conidial germination rate of the resistant isolates was also significantly lower than that of the sensitive isolates ( Figure 3D). sensitive isolates ( Figure 3A). In addition, mycelial dry weight was lower in the resistant isolates, but the difference was statistically non-significant ( Figure 3B, p > 0.05). The resistant isolates produced significantly fewer conidia in PSB compared with the sensitive isolates ( Figure 3C). Since most of resistant isolates produced no or little conidia, five resistant isolates (GL11, HA17, XH5a, XH7b, and XH43b) were tested for conidial germination rate on PSA and WA media. The conidial germination rate of the resistant isolates was also significantly lower than that of the sensitive isolates ( Figure 3D). The stability of resistance to carbendazim in 14 isolates was determined by culturing the isolates on fungicide-free PSA medium. Unchanged resistance factors before (T0) and after five generations (T5) indicated that resistance was stable ( Figure 4 and Table 2). The stability of resistance to carbendazim in 14 isolates was determined by culturing the isolates on fungicide-free PSA medium. Unchanged resistance factors before (T0) and after five generations (T5) indicated that resistance was stable ( Figure 4 and Table 2).

Correlation in Sensitivity of Carbendazim to Diethofencarb, Azoxystrobin, Pyraclostrobin, and Tebuconazole
As shown in Figure 5A, based on the EC 50 values, the correlation coefficient between carbendazim-diethofencarb was −0.74 (p < 0.001), indicating a moderate but significant negative cross-resistance between them. As for other fungicides used to control rice false smut, none of the carbendazim-resistant mutants exhibited resistance to azoxystrobin, pyraclostronbin, or tebuconazole. Spearman rank correlations in sensitivity between carbendazim and those three fungicides were analyzed, and no significant correlation was detected in sensitivity between carbendazim and azoxystrobin (r = 0.24, p = 0.33), pyraclostronbin (r = 0.29, p = 0.24), or tebuconazole (r = −0.32, p = 0.19) ( Figure 5B-D).

Nucleotide Sequence Analysis of the β-Tubulin Genes
Both Uvβ1Tub and Uvβ2Tub genes and their upstream fragments in the resistant and sensitive isolates were amplified and sequenced. Twenty-three sensitive isolates collected from different cities have the same amino acid sequences in both genes. The nucleotide sequences of Uvβ1Tub amplified from the twenty-three sensitive isolates, including exons and introns, were 1825 bp in length. The full-length cDNA of Uvβ1Tub amplified from RNA of the sensitive isolates was 1347 bp in length, encoding a putative polypeptide of 448 amino acids. Sequence comparison of the genomic DNA and cDNA revealed four introns located at nucleotide position 13-241, 266-333, 457-571, and 1363-1428. The Uvβ2Tub from carbendazim-sensitive isolates nucleotide sequences (including exon and intron) was 2103 bp. The full-length cDNA of Uvβ2Tub was 1359 bp in length, encoding a putative polypeptide of 452 amino acids. Sequence comparison of the genomic DNA and cDNA revealed six introns located at nucleotide position 13-156, 181-259, 328-492, 548-646, 960-1079, and 1379-1515 ( Figure 6A).
There was no amino acid difference between the carbendazim-sensitive isolates and carbendazim-resistant in the Uvβ2Tub. As for the Uvβ1Tub, compared with the sensitive isolates, several variations were found in the resistant isolates. The variations included point mutations, non-sense mutations, codon mutations, and frameshift mutations ( Table 3). As for point mutations, isolate GL11 had a variation at codon 91, resulting in an amino acid alteration from V to G, and isolate ZJ7 had a variation from A to D at codon 411. Two isolates, GL12b and JR12, had non-sense mutations. For GL12b, the mutation is a C-to-T transversion at nucleotide 871, transforming the codon 291 from Glutamine to a stop codon. For JR12, the mutation at nucleotide 63 (G to A) resulted in a stop codon at amino acid 21. Two isolates had codon mutations, JR11 had 15-bp deletion at codon 408-412, and YD8 had 3-bp deletion at codon 221. Five isolates had frameshift mutations, resulting in a change to a gene's reading frame-one to seven deletions were found in those isolates. The locations where mutations were found from different resistant isolates on the UvTub1 are marked in Figure 4A,B. The codon positions where mutations have been reported from other species in β1Tub and β2Tub are marked in Figure 4B. All of the mutations found in our study have never before been reported from other pathogens.

Nucleotide Sequence Analysis of the β-Tubulin Genes
Both Uvβ1Tub and Uvβ2Tub genes and their upstream fragments in the resistant and sensitive isolates were amplified and sequenced. Twenty-three sensitive isolates collected from different cities have the same amino acid sequences in both genes. The nucleotide sequences of Uvβ1Tub amplified from the twenty-three sensitive isolates, including exons and introns, were 1825 bp in length.   The two β-tubulin genes in U. virens exhibit high sequence identity (78.12%), especially at the MBC-binding sites (amino acid residues 6, 165 to 167, 198 to 200, 240, and 241) ( Figure 6B). The sequences of deduced amino acids of β-tubulin from A. nidulans from the NCBI GenBank database AAA3328.1. The shaded letters indicate the conserved residues. Letters above the sequences: substitution amino acids in β2Tub from other resistant field isolates [21]. Letters under sequences: substitution amino acids in β1Tub from resistant laboratory mutants [22,23]. The codon positions where mutations occurred in Uvβ1Tub are marked with arrows. The sequences of deduced amino acids of β-tubulin from A. nidulans from the NCBI GenBank database AAA3328.1. The shaded letters indicate the conserved residues. Letters above the sequences: substitution amino acids in β2Tub from other resistant field isolates [21]. Letters under sequences: substitution amino acids in β1Tub from resistant laboratory mutants [22,23]. The codon positions where mutations occurred in Uvβ1Tub are marked with arrows.
In addition, for isolates HA26, XH7b, and ZJ24, no mutations were found in Uvβ1Tub and Uvβ2Tub, nor in their upstream fragments. All of the sequences amplified in this study have been deposited in Genbank (Supplementary Table S1).

Phylogenetic Analyses of Predicted Amino Acid Sequences
The deduced amino acid sequence of Uvβ1Tub was 99.07% identical to that of Metarhizium humberi (KAH0601712), 99.07% identical to that of Epichloe typhina (P17938), 98.61% to that of Fusarium albosuccineum (KAF4455974), and 97.91% to that of Clonostachys rosea (VUC22392). Comparative analysis of the sequences yielded an E-value of 0.0, which confirmed Uvβ1Tub to be a member of the fungal β1Tub family.
A phylogenetic tree was constructed on the basis of the concatenated alignment of β-tubulin homologs of U. virens and other ascomycete fungi that are close to U. virens, including the outgroup fungus Candida glabrata (Figure 7). There were a total of 443 positions in the final data set. Maximum likelihood phylogenetic analyses of the predicted amino acid sequences for these two proteins (β1Tub and β2Tub) with sequences for 36 Tub genes from ascomycete species resolved two distinct clades, with β1Tub and β2Tub forming two monophyletic clades. Results confirmed that Uvβ1Tub and Uvβ2Tub were homologous to the β1Tub and β2Tub proteins from multiple other fungi, respectively.

Discussion
Fungicide application has been an effective option for controlling rice diseases in China. MBC fungicides, such as carbendazim and thiophanate-methyl, which were extensively used for the control of rice blast, sheath blight of rice, and bakanae disease of rice, were recently adapted for controlling RFS. However, few studies have reported the sensitivity of field isolates of U. virens to MBC fungicides. In this study, for the first time, the resistance of U. virens towards MBCs was monitored, and the resistance isolates within field population of U. virens were characterized.
A total of 321 isolates were collected from the main rice production area of China, and their sensitivities to the MBC fungicides were determined. The results showed the

Discussion
Fungicide application has been an effective option for controlling rice diseases in China. MBC fungicides, such as carbendazim and thiophanate-methyl, which were extensively used for the control of rice blast, sheath blight of rice, and bakanae disease of rice, were recently adapted for controlling RFS. However, few studies have reported the sensitivity of field isolates of U. virens to MBC fungicides. In this study, for the first time, the resistance of U. virens towards MBCs was monitored, and the resistance isolates within field population of U. virens were characterized.
A total of 321 isolates were collected from the main rice production area of China, and their sensitivities to the MBC fungicides were determined. The results showed the overall frequency of MBC resistant isolates was 4.36%, and all of the resistant isolates were of high resistance level. Except for the resistant isolates, other isolates were very sensitive to carbendazim, with a mean EC 50 value of 0.66 µg/mL. The results were similar to those for the Gibberella zeae and Fusarium species complexes causing pokkah boeng disease, with mean EC 50 values of 0.59 µg/mL and 0.60 µg/mL, respectively [25,26]. Although the sensitivity tested in this study could not be treated as the baseline, as the isolates have been previously exposed to MBC fungicides, the information provides a frame of reference for future issues with MBC insensitivity in U. virens.
The correlation analysis assay showed that the nine resistant isolates were sensitive to azoxystrobin, pyraclostrobin, and tebuconazole, indicating that there was no multiple fungicide resistance among MBC, QoI, and DMI fungicides. Thus, QoI and DMI fungicides can be used for the management of carbendazim resistance in U. virens. Moreover, the mixture of carbendazim with DMIs, such as tebuconazole (registration number: PD20110332), triadimefon (registration number: PD20060057), and hexaconazole (registration number: PD20181518), which has been registered to control rice diseases in China, could still be used to manage the resistance of U. virens to MBC at present. In addition, an important anti-resistance strategy would be to limit MBC fungicide sprays to a minimum and only in mixtures with multi-site, broad spectrum, protectant, low-risk fungicides [27].
A moderate negative cross-resistance between carbendazim and diethofencarb was observed in carbendazim-resistant isolates. Strong negative cross relationships could be observed between diethofencarb and benzimidazole-resistant isolates with E198A and E198V mutations [28,29]. In several pathogens, such as in Botrytis spp. and Monilinia spp., diethofencarb inhibits benzimidazole-resistant but not benzimidazole-sensitive isolates [30][31][32]. However, in U. virens, benzimidazole-sensitive isolates were also largely sensitive to diethofencarb (with EC 50 values of approximately 10 µg/mL). The binding mode between UvTub and diethofencarb needs to be further investigated.
Our results also indicate that the acquisition of MBC resistance is accompanied by a reduction in fitness, in that the resistance isolates grew more slowly, produced fewer conidia, and the conidia germinated less than the sensitive isolates. Fitness plays a significant role in the evolution of fungicide resistance in the fungal population. Thus, the investigation of fitness is important for the establishment of effective strategies for resistance management. In several pathogens, MBC-resistant isolates shared similar fitness with wild-type isolates and suffered little fitness penalty [32][33][34][35][36]. Similar to our results, fitness reduction was also observed in F. fujikuroi MBC-resistant isolates in terms of fewer conidia and less virulence [23]. In addition, MBC resistant isolates bearing different mutations might have different fitness, since different point mutation isolates possessed different predominance. Take B. cinerea for example; MBC resistant isolates collected from China usually harbored the E198A/V/K mutation. Among them, E198V was quite predominant compared with E198A/K [33,34,37]. In Japan, E198V and E198A were widely determined, while E198K and F200Y were barely found [29]. However, whether the reduction in fitness can be offset by the ability to withstand MBC need to be measured.
The mechanism of resistance to carbendazim was often associated with point mutations in the Tub genes that change the structure of the fungicide binding site to, in turn, decrease sensitivity [18,38]. Analysis of the U. virens genome revealed two β-tubulin genes, β1Tub and β2Tub. Both of them were identified and investigated in this study. Most of the ascomycetes contain only one β-tubulin gene [39], while two genes have been found in some fungi, including Aspergillus nidulans, Trichoderma spp., Colletotrichum spp., and Fusarium spp. [40][41][42][43]. In our phylogenetic analyses, the ascomycetes that have two β-tubulin genes were also partially listed.
The two tubulin genes in U. virens exhibited high sequence similarity. The deduced amino acid sequences of Uvβ1Tub and Uvβ2Tub were 78.12%. A similar high identity was also reported in G. zeae, with the identity of 76% [44]. However, in U. virens, whether their products have different functions need to be further investigated. Gene replacement demonstrated that the two genes are functionally interchangeable in A. nidulans [45] and C. graminicola [46]. In G. zeae, both tubulin isotypes function well by being assembled into cytoplasmic microtubules. In addition, the effects of β1Tub on mycelial growth, conidial germination, and pathogenicity have been verified by gene knockout in G. zeae [44]. MBC fungicides can interact with several regions of the tubulin molecule, including amino acid residues 6, 165 to 167, 198 to 200, 240, and 241 [44,47]. All these regions were conserved in Uvβ1Tub and Uvβ2Tub from sensitive isolates.
In our study, as for the carbendazim-resistant isolates, point mutations were found in the Uvβ1Tub gene, not in the Uvβ2Tub gene. As for MBC resistance, substitutions at the β2Tub gene have frequently been reported to cause resistance in the field or laboratory isolates of several pathogenic fungi, including mutations at codons 6, 50, 134, 165, 167, 198, 200, 235, 240, 241, and 257 [15,18,21,23,48]. Variations in β1Tub have also been reported to be involved in the MBC resistance. In the case of UV-mutants of F. moniliforme (synonym, F. verticillioides), resistance resulted from a Tyr50Asp mutation in β1Tub [22]. The point mutations V91G and A411D found in Uvβ1Tub in our study have never before been reported in other pathogens.
Other than point mutation, the majority of variations found in our study were nonsense mutations (isolates GL12b and ZJ7), codon mutations (isolates JR11 and YD8), and frameshift mutations (isolates GL23, HA17, XH5a, XH43b, and YZ11), which may cause the premature termination of Uvβ1Tub gene translation, opposing the formation of full-length transcripts, thereby impacting the gene expression. The increased resistance to benzimidazole fungicides in the β1Tub unexpressed mutants suggests β1Tub rather than β2Tub is the preferred binding target for MBC fungicides in U. virens. The binding preference of MBC with β1Tub rather than β2Tub has been verified in F. graminearum [43]. By gene knockout, MBC sensitivity was found to be significantly reduced in a β1Tub deletion isolate but increased in a β2Tub deletion isolate compared with that of a parental isolate, suggesting that β1Tub was involved in the MBC sensitivity of F. graminearum [43]. Homology modeling further verifies the possible MBC binding sites to be β1Tub rather than β2Tub. In addition, MBC fungicides are more likely to disrupt β1Tub microtubules rather than β2Tub microtubules in GFP-tub fusion mutants in vivo [43]. Thus, we assumed that the resistant mechanisms in those U. virense isolates seemed to be different from other pathogens. Unexpressed β1Tub gene may result in the lack of MBC target in U. virens, hence causing the resistance to the MBCs. For U. virens, the function and their involvement in MBC resistance of Uvβ1Tub and Uvβ2Tub need to be further studied. In addition, the reason why Uvβ1Tub is so prone to mutate remains to be studied.
We also confirmed the absence of mutations in the β1Tub and β2Tub sequence in the resistant isolates HA26, XH7b, and ZJ24, suggesting that there may be other pathways involved in the molecular mechanism of MBC resistance.
In conclusion, our study has revealed the emergence of the resistance of U. virens to MBC fungicides. We suggest that it is necessary to persistently monitor the resistance to MBCs. MBCs should be valuable when used in combination or alternation with other fungicides for the control of RFS in rice fields.