Molecular characterization and overexpression of the difenoconazole resistance gene CYP51 in Lasiodiplodia theobromae field isolates

Stem-end rot (SER) caused by Lasiodiplodia theobromae is an important disease of mango in China. Demethylation inhibitor (DMI) fungicides are widely used for disease control in mango orchards. The baseline sensitivity to difenoconazole of 138 L. theobromae isolates collected from mango in the field in 2019 was established by the mycelial growth rate method. The cross-resistance to six site-specific fungicides with different modes of action were investigated using 20 isolates randomly selected. The possible mechanism for L. theobromae resistance to difenoconazole was preliminarily determined through gene sequence alignment and quantitative real-time PCR analysis. The results showed that the EC50 values of 138 L. theobromae isolates to difenoconazole ranged from 0.01 to 13.72 µg/mL. The frequency of difenoconazole sensitivity formed a normal distribution curve when the outliers were excluded. Difenoconazole showed positive cross-resistance only with the DMI tebuconazole but not with non-DMI fungicides carbendazim, pyraclostrobin, fludioxonil, bromothalonil, or iprodione. Some multifungicide-resistant isolates of L. theobromae were found. Two amino acid substitutions (E209k and G207A) were found in the CYP51 protein, but they were unlikely to be related to the resistance phenotype. There was no alteration in the promoter region of the CYP51 gene. However, difenoconazole significantly increased the expression of the CYP51 gene in the resistant isolates compared to the susceptible isolates. These results are vital to develop effective mango disease management strategies to avoid the development of further resistance.


Materials and methods
Isolates and culture conditions. In 2019, 138 single-spore field isolates of L. theobromae were obtained from diseased mango fruits in Hainan Province, China, previously identified and preserved in our laboratory. At 28 °C in the dark, the isolates were grown on potato dextrose agar (PDA) medium. All the procedures followed for using the mango fruit comply with relevant institutional, national, and international guidelines and legislation.
Determination of the baseline sensitivity of field isolates to difenoconazole. A mycelial growth inhibition assay was used to investigate the baseline sensitivity to difenoconazole of 138 L. theobromae isolates 13 . To prepare stock solutions, difenoconazole (97.2%; Zhengye Chemical Industrial Co., Hainan, China) was dissolved in 100% acetone to obtain 5 × 10 3 μg/mL solutions. A mycelial plug (5 mm in diameter) from the edge of the 3-day-old culture of each isolate was inoculated in 90 mm diameter Petri plates containing difenoconazole PDA media. The difenoconazole concentrations were 51.2, 12.8, 3.2, 0.8, and 0.2 μg/mL. The final concentration of acetone solvent in the medium was 0.1%. There was also a control medium with the same amount of acetone but no fungicide. Inoculated plates were cultured in the dark at 28 °C. The diameter of each colony was measured, and the inhibition rate of mycelial development calculated after cultured for 36 h. There were three replicate plates per treatment. The entire experiment was repeated twice independently. The frequency distribution of 138 EC 50 values of L. theobromae was plotted to represent the baseline sensitivity. The baseline sensitivity level of L. theobromae was used to develop classification criteria for difenoconazole-sensitive phenotypes 13,14 . The resistance factor (RF) of each isolate to fungicide was computed using the baseline sensitivity: sensitive isolates (S): RF < 5; resistant isolates (R): RF > 5 15 . Cross-resistance of difenoconazole with other fungicides. The cross-resistance of difenoconazole with other regularly used fungicides was investigated using 20 isolates. These comprised DMIs fungicides tebuconazole and five fungicides of other action modes, which were carbendazim (benzimidazole), iprodione (dicarboximides), bromothalonil (bromomethyl glutaronitrile), fludioxonil (phenylpyrrole) and pyraclostrobin (strobilurin). As previously stated, EC 50 values were calculated using a mycelial growth inhibition experiment. The EC 50 values of the fungicides tested were used to determine cross-resistance correlations. The experiment was conducted three times independently, using three replicate plates for each treatment.
Cloning and sequencing of LtCYP51 gene. Mycelia of L. theobromae were snap-frozen in liquid nitrogen and processed with tungsten beads in a LUKYM-II Mixer-Mill to extract total genomic DNA from 30 isolates (Guangzhou Luka sequencing instrument Co., LTD, Guangzhou, China). Total genomic DNA was extracted according to the manufacturer's instructions using the E.Z.N.A. ® HP Plant DNA Mini kit (Omega Bio-Tek, Norcross, United States). Primer pairs, LtCYP51-F1/LtCYP51-R1 and Per-1F/Per-1F (Table 1), were designed to amplify the LtCYP51 coding sequence and LtCYP51 promoter of resistant and sensitive isolates. Amplications were performed in a My Cycler thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA). A 40 µL reaction volume was used for PCR amplification, with 20 µL of 2 × Phanta Max Master Mix, 0.8 µL of template DNA, 1.6 µL of (10 mM) each primer, and 16 µL of ddH2O. The PCR settings for the coding sequence were 95 °C for 3 min, followed by 35 cycles of 95 °C for 30 s, 58 °C for 50 s, and 72 °C for 90 s, followed by a final 5-min extension at 72 °C.
For the promoter, all conditions were the same as for the coding region, except that the extension time was 15 s. With a Tolo PCR Clean-Up Kit (Tolo Biotech Co., LTD, Shanghai, China), PCR products were purified. The PCR fragment was ligated into the p ESI-blunt vector (YEASEN Biotech Co. Ltd) and sequenced with vector primers M13F and M13R at Tianyihuiyuan Biotech Co., Ltd (Guangzhou, China).
The LtCYP51 DNA sequence was studied using the programs DNAMAN (version 6.0; LynnonBiosoft, U.S.A.) and InterPro Scan (http:// www. ebi. ac. uk/ inter pro/ search/ seque nce-search), and it was compared LtCYP51 genes of other fungi. The amino acid sequences of the LtCYP51 genes from difenoconazole-resistant and sensitive www.nature.com/scientificreports/ isolates were compared using the computer program EMBOSS Transeq (https:// www. ebi. ac. uk/ emboss/ trans eq/), which translated the DNA sequences into amino acid sequences using standard code.
Quantitative expression of LtCYP51 gene. Two sensitive isolates (YC70 and JS10) and six resistant isolates (LD10, SY31, YZ90, DZ11, SY05 and YC80) were randomly selected to study the expression of the LtCYP51 gene. The EC 50 values of 8 isolates see Table 2. Five mycelial plugs were transferred into a flask containing 100 mL of potato dextrose broth (PDB) and incubated at 28 °C for 24 h on a rotary shaker at 150 rpm. Three flasks were treated with difenoconazole to reach a final concentration of 150 µg/mL, with three replicates. Three flasks containing 100 mL of PDB were used as untreated controls. After being treated with difenoconazole for 12 h, the mycelia of each isolate were removed for RNA extraction 16 . Total RNA was extracted using the TIANGEN RNA simple Total RNA kit (Tiangen Biotech Co., Ltd, Beijing, China), and cDNA was synthesized using the HiScript III 1st Strand cDNA Synthesis Kit with g DNA Eraser (Vazyme Biotech Co., Ltd, Nanjing, China) following the manufacturer's instructions. Every reaction had three biological replicates and three technological replicates. Real-time PCR was carried out in a total volume of 20 µL using the qTOWER3 G REAL-TIME PCR thermocycler (Analytik Jena AG, Jena, Germany). To amplify the genes, 2 × ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd, Nanjing, China) was employed. The actin gene was amplified as a reference using the primer pair RT-Act LtF/RT-Act LtR to standardize the quantification of LtCYP51 expression 17 . Three repeats of the experiments were carried out. Statistical analysis. The inhibition rates were converted to the probability values, and difenoconazole concentrations were log 10-transformed before using a line regression model. The effective concentration to inhibit mycelial growth by 50% (EC 50 ) was calculated by the regression equation. The EC 50 values were checked for homogeneity of variances using Levene's test, then the EC 50 values were calculated for each isolate by combining the data from both replications. The Shapiro-Wilk test was used to determine the normality of the frequency distribution of difenoconazole sensitivity, and the outliers were detected using the boxplot in SPSS 21.0. The histograms were built utilizing log 10-transformed EC 50 values when the outliers were removed 13,14 . Spearman's rank correlation coefficient using log-transformed EC 50 values was used to examine cross-resistance among seven fungicides 18,19 . To assess the differences in the relative expression of genes, one-way ANOVA with the LSD test was used (P < 0.01). The differences in the mean expression levels were compared by the Mann-Whitney U test (P < 0.001). DNAMAN software was used to examine DNA sequences (version 6.0; LynnonBiosoft, U.S.A.).

Results
Baseline sensitivity of L. Theobromae to difenoconazole. The EC 50 values of difenoconazole to inhibit mycelial growth of 138 L. theobromae field isolates ranged from 0.01 to 13.72 µg/mL. After the outliers were excluded by boxplot, a continuous unimodal log-normal distribution of sensitivity of 121 isolates to difeno- www.nature.com/scientificreports/ conazole was observed (W = 0.981, P = 0.087) (Fig. 1). The mean EC 50 value of 121 isolates was 1.12 ± 1.09 μg/mL, adopted as the resistance threshold concentration. Twenty-one of the 138 isolates were categorized as resistant to difenoconazole based on baseline sensitivity. The EC 50 values of resistant isolates ranging from 5.60 to 13.72 μg/ mL, and the resistance factors ranged from 5 to 12.25. The resistance frequency of L. theobromae isolates against difenoconazole was 15.22%. The resistant isolates could grow in the medium containing 150 μg/mL of difenoconazole (Fig. 2).
Cloning and characterization of the LtCYP51. The nucleotide sequences of the 1797 bp fragment of the LtCYP51 gene from the isolates were found to be 99% identical to that of L. theobromae (GenBank accession number MK107983.1). The LtCYP51 gene fragment encodes 523 amino acids and has two introns of 49 bp each at nucleotide positions 247 and 494, respectively. The BLAST search amino acid sequence of the LtCYP51 protein also showed 100%, 94.5% and 93.1% identity with that of the CYP51 protein in L. theobromae from cacao (XP_035367211.1), Diplodia seriata from grape (OMP84122.1) and Botryosphaeria dothidea from apple (KAF4310083.1), respectively.
Comparison of the LtCYP51 gene and its upstream region in sensitive and resistant isolates. The 30 isolates were analyzed for the sequence of LtCYP51 genes and their upstream regions. Based on the alignment, two mutant phenotypes were found. Compared with other isolates, the sensitive isolate YC70 has  www.nature.com/scientificreports/ two amino acid substitutions at positions 207 (from glycine to alanine, G207A) and 209 (from glutamic acid to lysine, E209k) on the LtCYP51 protein. Furthermore, the resistant isolate YC80 has one amino acid substitution at position 209 (from glutamic acid to lysine, E209k) on the LtCYP51 protein, and this substitution was consistent with the sensitive isolate YC70 (Fig. 4). In the other resistant isolates, no mutation was found. Fragments approximately 500 bp upstream of the LtCYP51 gene were obtained using the primer pair Per-1F/ Per-1R. The upstream regions were identical in all tested isolates. In any of the isolates tested, no mutations or insertions were identified in the promoter of the LtCYP51 gene.

Relative expression of LtCYP51 in sensitive and resistant isolates.
To explore the mechanism of resistance, the expression levels of the LtCYP51 gene in resistant and sensitive isolates were tested. Our results showed that difenoconazole significantly induced LtCYP51 expression in the resistant isolates (P < 0.01) (Fig. 5A). The mean constitutive relative expression levels of LtCYP51 without fungicide in the sensitive and resistant isolates were 1.05 and 1.7 times higher, respectively. Difenoconazole increased the relative expression of LtCYP51 by 1.87-2.06 times in two sensitive isolates with an average of 1.97, but 6.71-12.41 times in six resistant isolates with an average of 10.05 times. In the resistant isolates, the mean relative expression of LtCYP51 induced by difenoconazole was fivefold higher than that of sensitive isolates, and this difference was significant (P < 0.001) (Fig. 5B).

Discussion
Mango diseases are widely controlled using site-specific systemic fungicides in almost all mango-growing regions in the world. The detection of fungicide resistance is a crucial step in monitoring and regulating the spread of resistance in the field 20 . DMI fungicides are classified as a medium risk for resistance development by the Fungicide Resistance Action Committee 21 . DMIs fungicides were more favoured by orchardist due to their specific mode of action and broad anti-fungi spectrum at present. However, DMI resistance has been found in a variety of phytopathogenic fungi 16,17,20 . The resistance mechanisms of DMIs have been reported to be diverse: (I) point mutations in the target gene 14α-demethylase (CYP51) [22][23][24] ; (II) CYP51 gene overexpression 16,[25][26][27][28][29] ; and (III) overexpression of efflux proteins 30,31 . In this study, we established the baseline sensitivity of L. theobromae to difenoconazole using 121 isolates from five major mango-producing regions in Hainan, China. The results showed that the EC 50 values ranged from 0.01 to 13.72 µg/mL, with a mean EC 50 value of 1.1 µg/mL, suggesting that this method could be used as a criterion to judge difenoconazole resistance in further studies. Twenty-one difenoconazole-resistant isolates were found in this study; their EC 50 values ranged from 5.61 to 13.72 μg/mL. Among systemic fungicides, MBC fungicides are inhibitors of tubulin biosynthesis, which impedes cell division and inhibits mycelial growth 32 . MBC-resistant populations of L. theobromae have been confirmed from papaya, citrus and mango 11,33-36 . Our lab discovered that the resistance frequencies of L. theobromae isolates www.nature.com/scientificreports/ to carbendazimmany were more than 70%, and the highly resistant isolates grew normally with 1000 μg/mL carbendazim. Point mutations in the target gene β-tubulin were identified in resistant isolates 34 . Compared with carbendazim resistance, DMI resistance of L. theobromae from mango is not severe in Hainan. However, www.nature.com/scientificreports/ the EC 50 values of 79 isolates to difenoconazole were above 1 µg/mL, which accounted for more than half of the isolates. Thus, a large number of isolates had reduced sensitivity to difenoconazole. Meanwhile, resistant isolates showed positive cross-resistance to difenoconazole and tebuconazole. In addition, no cross-resistance was discovered in this investigation between DMI and non-DMI fungicides. This is consistent with the reported results of Botrytis cinerea and Colletotrichum gloeosporioides resistance to DMIs 16,37 . In the experiment of cross resistance, we found that there were a small number of isolates with multifungicide resistance (MFR). The result shows that there may be more MFR isolates of L. theobromae in fields. So it is necessary to continuously detect multifungicide resistance in the future. The commonly used site-specific fungicides (such as carbendazim and azoxystrobin) gave bad control effect against L. theobromae due to the development of resistant isolates. With frequent applications of DMI fungicides in fields, the development of DMI fungicide resistance is a major challenge for effective disease control of fruit in China. A appropriate management strategies for fungicide resistance and better management of SER have been suggested, such as reducing the usage of DMI fungicides by combining with alternative fungicides with distinct modes of action that have not been found to cause cross-resistance. For the control of L. theobromae, mixtures of difenoconazole and other chemical fungicides, as well as the botanical fungicide Thymol, have been reported to be particularly effective 38 . This management can reduce pathogen population selection pressure, slowing the development of DMI resistance. The target site of action of DMIs is the enzyme CYP51. The function of CYP51 is to remove the 14-methyl group of the sterol precursor. DMI fungicides interact with CYP51 to inhibit the demethylation of lanosterol and influence the production of ergosterol, destroying the integrity and fluidity of the fungal cell membrane. The use of DMI fungicides interfered with ergosterol synthesis activating a CYP51 response. The function of CYP51 has been verified in many pathogens [39][40][41][42] . DMI fungicides significantly induce CYP51 expression. Fan et al. proved that CYP51 gene deletion mutants of Fusarium graminearum increased the sensitivity to DMI fungicides prochloraz and difenoconazole 42 . Point mutations of the CYP51 gene change the conformation of the target protein, resulting in a decrease in the binding ability of fungicides to the target protein. In this study, two amino acid substitutions, E209K and G207A, were found in LtCYP51 in the sensitive isolate YC70, and one of those amino acid substitutions, E209K, was also found in the resistant isolate YC80. We infer that point mutation of the CYP51 gene of L. theobromae may not be the cause of low-level resistance to difenoconazole. Our studies have reported overexpression of LtCYP51 in resistant isolates after treatment with difenoconazole. Although most research studies claim that target site changes cause resistance in the majority of DMI-resistant isolates, additional resistance mechanisms independent of CYP51 mutations cannot be ruled out. Some studies found that the CYP51 genes of DMI-resistant isolates of plant pathogenic fungi have various point mutations [42][43][44] . Studies have also shown that an increase in DMI fungicide application dosages would not improve their efficacy in the case of gene mutations. Some studies reported mutations and overexpression of the CYP51 gene simultaneously in some DMI-resistant isolates of plant pathogens 17,20,45 . Furthermore, in L. theobromae of papaya, Mycosphaerella graminicola of wheat, Blumeriella jaapii of cherry and Neophysopella meliosmae-myrianthae of grapevine no CYP51 gene point mutation linked to DMI resistance has been identified, but CYP51 overexpression has been observed 18,30,46,47 . A similar pattern of results was obtained in this study. The results of the present study, do not clearly explain the mechanisms leading to CYP51-independent resistance. Regarding the mechanism of CYP51 overexpression, overexpression of CYP51 caused by promoter insertions or retrotransposons has only been confirmed in a few phytopathogenic fungi thus far; for example, in Penicillium digitatum, increased expression was due to a 199-bp sequence duplication at the promoter of CYP51 48 . The mobile genetic element 'Mona' is believed to facilitate overexpression of CYP51 in Monilinia fructicola 49,50 . According to the report of Rallos, overexpression of CYP51 was associated with the presence of the Y136F mutant genotype 45 . However, the underlying mechanisms of CYP51 overexpression are not known for in the field DMI-resistant subpopulations of   37 . We tried to clone and sequence analysis the promoter of LtCYP51 in this study. However, the promoter of LtCYP51 from difenoconazole-resistant L. theobromae isolates did not show any mutations or insertions. The molecular mechanism of LtCYP51 overexpression needs further investigation by obtaining the complete sequence of the promoter region.
In brief, DMIs have diminished sensitivity in field populations due to their long-term and intensive use. Our results suggest a potential risk for DMI resistance development in L. theobromae. In the mango fields of China, Hainan Province, L. theobromae has acquired a low to moderate difenoconazole resistance. Although there existed obvious positive cross resistance between difenoconazole and tebuconazole, no cross-resistance was found between difenoconazole and non-DMI fungicides. Control measures such as rotation and mixture treatments with different modes of action fungicides can reduce the emergence of resistant isolates in the field. This means that isolates with multifungicide resistance will become more frequent over time. Compared with the difenoconazole-sensitive isolates, there were no mutations in the CYP51 gene of resistant isolates at positions 132 or 137 or at any other positions (markers for resistance in DMI fungicides). However, induced expression of CYP51 in resistant isolates is involved in resistance to difenoconazole. In the future, more research should focus on exploring the mechanisms that induce CYP51 expression, investigating nontarget site mechanisms of fungicide resistance and the mechanisms of multifungicide resistance in L. theobromae. Improved knowledge of fungicide resistance evolution and of the molecular mechanisms by which this occurs will be necessary to implement suitable control strategies that will reduce the likelihood of fungicide resistance outbreaks. Our findings are critical for controlling the high-risk pathogen L. theobromae and can help to slow down or even prevent the emergence of DMI fungicide resistance.