Shift of Sensitivity in Botrytis cinerea to Benzimidazole Fungicides in Strawberry Greenhouse Ascribing to the Rising-lowering of E198A Subpopulation and its Visual, On-site Monitoring by Loop-mediated Isothermal Amplification

Grey mold disease results from Botrytis cinerea, a classical “high-risk” plant pathogenic fungus in meaning of resistance development to fungicides, and its management depends largely on the frequent applications of fungicides. The evolution of resistance to benzimidazole chemicals during 2008 and 2016 was monitored continuously in strawberry greenhouses located in Zhejiang province. Results showed that extensive applications of the mixture of carbendazim and diethofencarb caused the rapid spread of Ben MR subpopulation. The withdraw of this mixture lead to the sharply decrease of Ben MR and re-dominance of Ben HR isolates of B. cinerea with the E198A mutation in β-tubulin gene. The LAMP primers, based on the E198A point mutation, were designed to detect the E198A genotype specifically. HNB (Hydroxynaphthol blue), a metalion indicator, acted as a visual LAMP reaction indicator that turned the violet colored into a sky-blue color. The detection limit of concentration of DNA was 100 × 10−2 ng/μL and this LAMP assay could be applied to detect the E198A genotype with 100% accuracy in strawberry greenhouses of three Province and was more rapid and easier to operate. In summary, we establish a simple and sensitive on-field LAMP assay which can be adopted to determine within 1.5 h whether the benzimidazoles or the mixture of a benzimidazole fungicide and diethofencarb is suitable for management of B. cinerea.

Management of grey mold disease is achieved by frequent applications of fungicides. However, B. cinerea is a typical "high-risk" pathogenic fungus. Under the selection pressure of a fungicide, once the resistance occurs, the level of resistance will develop rapidly for its short life cycle and prolific reproduction rate 7 . The failure of control due to the rapid evolution of resistance has become a important concern. Benzimidazole fungicides, the first group of chemicals with therapeutic activity were one of the main agents used in crop plant productions [8][9][10] . The phenomena that fungi resistant to benzimidazoles has been demonstrated by a number of studies in different plant diseases [11][12][13] . B. cinerea resistant to benzimidazoles has been reported to carry point mutations at the codon 198 (E198A or E198K or E198V) or 200 (F200Y) in the β-tubulin gene [14][15][16] . In general, two types of benzimidazole-resistants, Ben HR (high resistance to benzimidazole fungicides) and Ben MR (moderately resistant to benzimidazoles), were widely detected in fields. Ben HR isolates were caused by the E198A mutation which simultaneously showed more sensitive to diethofencarb, a phenylcarbamate, than the wild benzimidazole-sensitive (Ben S) isolates. Ben MR isolates, simultaneously resistant to diethofencarb just like the Ben S isolates and caused by the F200Y mutation, were detected after the applications of the mixture "a benzimidazole fungicide + diethofencarb 14,15,17 ".
Resistance to benzimidazoles in plant-pathogenic fungi such as B. cinerea is traditionally detected by the tests of discriminatory doses, which takes 3 to 5 days even extended to a week, and is time-consuming and laborious 10,18,19 . Several detection methods based on PCR including AS-PCR and RAPD-PCR have been applied for the detection of mutant isolates responsible for fungicide resistance 9,20 . Previous studies have developed PCR-RFLP and real-time PCR to detect benzimidazole-resistance with mutations at codon 198 in the β-tubulin gene in B. cinerea and Monilinia laxa 9,21 . However, these techniques have inherent shortcomings, including the need for prolonged time and expensive equipment, which limit these methods to laboratories and are not suitable for field assays. Loop-mediated isothermal amplification (LAMP) 22 is a relatively new and rapid nucleic acids amplification technique which does not need a precision thermal cycler, thus making it appropriate for using in the fields. The LAMP assay has been used to detect several plant-pathogenic fungi, including Magnaporthe grisea 23 , Erysiphe necator 24 , Phytophthora sojae 25 , and others [26][27][28] . The current study is conducted to (i) monitor the evolution of resistance in B. cinerea from strawberry greenhouse following different applications of benzimidazoles, (ii) develop a simple, rapid and sensitive on-field LAMP assay for E198A subpopulation to guide the applications of benzimidazole fungicides for management of B. cinerea.

Results
Shift of sensitivity to benzimidazoles during 2008 and 2016. The kind of chemicals used showed great impact on the evolution of sensitivity in B. cinearea to benzimidazoles (Fig. 1). In these tested 27 greenhouses located in Zhejiang province, the mixture of carbendazim and diethofencarb was adopted to control the Ben HR isolates since 2006, the rapid increase of Ben MR isolates (22.2%, 47.0%) and simultaneous decrease of Ben HR isolates (53.3%, 27.3%) were observed respectively for 2008 and 2010. After this mixture was not used in these greenhouses since 2011, Ben MR isolates decreased rapidly according to the frequency of 41.0%, 20.3%, and 17.6%, respectively for 2012,2014, and 2016. Meanwhile, re-increase of Ben HR subpopulation was observed by the frequency of 33.8%, 43.8%, and 51.1%. Moreover, the analysis of the β-tubulin gene fragments of the twenty isolates of different phenotype of benzimidazole sensitivity chosen at random per sampled year indicated all the Ben HR isolates had the E198A mutation and all Ben MR had the F200Y mutation.
LAMP primers for detection of E198A B. cinerea. Template DNA from CL-3, a Ben HR isolate carrying the E198A mutation, and a Ben S isolate (TM-10) were used to assess the six sets of mismatched LAMP primers (FIP was mismatched) both visually based on color changes in Hydroxynaphthol blue (HNB) (Fig. 2b). The results indicated that the different FIP primer sets S2 (F3/B3/FIP2/BIP) and S5 (F3/B3/FIP5/BIP) could be used to distinguish the E198A genotype of B. cinerea ( Fig. 2 and Table 1). The primer set S2 showed the most intensive sky blue color which distinguish the E198A isolates from Ben S isolates with ease and therefore was adopted in subsequent tests in this study. Specificity and repeatability of the LAMP assay. The LAMP assay was positive only for all 6 isolates harboring the E198A mutation. No positive DNA products were observed when other type of mutants or Ben S isolates were used as templates (Fig. 4). These results suggested that the established LAMP assay had good specificity, accuracy and stability.   www.nature.com/scientificreports www.nature.com/scientificreports/

on-site LAMp detection and traditional tests in laboratory.
To evaluate the reliability and significance in fields, we adopted samples from different greenhouses in five different geographical regions located in three different Province by using the on-site LAMP assays. Results indicated that 153 (67.4%) samples were positive reactions. In laboratory, the traditional tests by the discriminatory dose of 5 μg mL −1 showed that 169 out of 227 isolates (74.4%) were resistant (Ben R) which could grow on PDA plates amended with 5 μg mL −1 carbendazim and the remaining 58 isolates were sensitive (Ben S). Further mutation analysis by PCR showed that a total of 169 isolates had mutant genotypes of β-tubulin gene (153 with E198A, 5 with E198V, 4 with E198K, and 7 with F200Y) and 58 were Ben S isolates. Thus, the on-site LAMP assay could specifically detect all isolates with the E198A mutation with 100% accuracy in fields (Table 2).

Discussion
As a high risk pathogen, resistance of B. cinerea to fungicides can emerge rapidly after continuous exposures in the fields 29 . Once resistance occurs, the control efficiency reduces sharply in addition to the increased fungicide residues threatening the health of humans as well as other non-target organisms. At present, there is a need for environment-friendly and sustainable control measures that requires smarter usages of fungicides in order to delay or manage resistance 1 . Therefore, monitoring the evolution of fungicide resistance is an important requisite for integrated control of grey mold.
In China, the selection of resistance in B. cinerea from crops such as vegetables to benzimidazoles and double-resistance to benzimidazoles and diethofencarb has been reported in regions such as Zhejiang 16,17,30 . In this study, the development of resistance to benzimidazoles was monitored during 2008 and 2016 in strawberry greenhouses of Zhejiang Province. Results showed that the application strategy of fungicides had significant impact on the evolution of sensitivity. The extensive usages of the mixture of carbendazim and diethofencarb resulted in quick increase of Ben MR subpopulation. Interestingly, when its applications were withdrawn for serious double-resistance to benzimidazoles and diethofencarb, Ben HR isolates re-dominated in B. cinerea population although no selection pressure of benzimidazole fungicides existed in the latter process. And, all the Ben HR isolates detected after that re-dominance had the E198A (GAG → GCG) mutation in β-tubulin gene. This evolution phenomenon of sensitivity might be mainly attributed to their strong competitive ability of Ben HR isolates 1 . Many studies found that E198A was the dominant sub-population with resistance to benzimidazole fungicides 16,29 . In 2017, we tested samples from three Province in China, and the results showed that 81.9% resistant isolates had the single point mutation, E198A, and only a frequency of 3.1% was observed for F200Y. A total of frequency of 4.0% was detected for other two genotypes (E198V and E198K) of benzimidazole resistance. However, E198V, for example, was reported as lowly resistant (LR) which can grow on 5 mg L −1 but cannot on 10 mg L −1 carbendazim or thiophanate-methyl and were more sensitive to low temperature according to our previous study 16 . Therefore, monitor the E198A sub-population can give enough information for the shift of sensitivity and management of grey mold.
Previous studies also indicated that the frequency of sub-populations harboring the F200Y mutation in B. cinerea decreased rapidly if application of carbendazim and diethofencarb mixture was stopped 29 . We speculated that this evolution pattern of resistance would be utilized to control B. cinerea through a well-directed application   www.nature.com/scientificreports www.nature.com/scientificreports/ of benzimidazoles if we could determine the instantaneous status of population structure of B. cinerea in a greenhouse. When E198A subpopulation dominates, the mixture of carbendazim and diethofencarb can be used. The decrease of E198A isolates and rapid development of Ben MR will be expected due to extensive applications of the mixture. Then, the mixture and the benzimidazole fungicides should be suspended for some times until the first situation appeared which would happen within a shorter period. This suggests that we can get enough information for the management of grey mold through monitoring the situation of E198A sub-population. However, the prerequisite is that we can rapidly determine the status of E198A subpopulation.
The discriminatory doses assay is the commonly used method to detect resistance to benzimidazole fungicides 10,16,31 . However, it is not suitable for field applications where quick detection is preferred. Molecular-based methods such as Real-time PCRs 9,20,21 not only require 4 to 5 h but also expensive equipment and well-trained technicians that limit their applications in fields. In this study, the E198A mutants of B. cinerea were detected by a novel LAMP assay which could successfully differentiate E198A mutants from sensitive isolates and other mutants (E198V, E198K, and F200Y) with 100% accuracy. For sample pre-treatments, using All-DNA-Fast-Out to extract DNA within about 10 min, eliminated the process of centrifugation, extraction and other operational steps in the traditional procedure of extracting DNA, thereby saving time of the samples-pretreatment, and reducing the contamination between the samples, to achieve the purpose of rapid detection. This assay could be on-field finished within 1.5 h without expensive devices thus provided an alternative approach to rapidly detect fungicide resistance in the plant pathogens in fields. LAMP is recently developed for detection of plant pathogenic fungi in infected plant materials 32,33 . Addition of the HNB dye prior to amplification, which is a metal ion that is widely used in LAMP assays, reduces the risk of contamination 34,35 . In this study, the negative and positive responses were successfully distinguished by HNB color changes 30,36,37 . The LAMP detection for B. cinerea based on Bcos5 and DNA was extracted using the Plant Genomic DNA Kit was 10-fold more sensitive than conventional PCR 32 . They also reported a specific, repeatable and sensitive LAMP assay for detection of B. cinerea with F200Y mutation (Ben MR) from different plants artificially inoculated with conidia 30 . Our study described a LAMP assay targeting E198A mutation of B. cinerea (Ben HR) according the evolution pattern of different sensitivity to fungicides in strawberry greenhouses to guide the application of benzimidazole fungicides. Moreover, as Botrytis cinerea is a ubiquitous and "high-risk"of fungicide resistance development plant pathogenic fungus worldwide, the smarter and precise usages of fungicides on plants through on-site monitoring techniques will significantly decrease the input of fungicides and therefore provide benefits for food security and human healthy.  38 . In these sampled greenhouses, benzimidazole fungicides, including carbendazim and thiophanate-methyl, were frequently used before 2004 and their application decreased rapidly, which were replaced by the mixture of carbendazim and diethofencarb (an N-phenylcarbamate) since 2006. From 2011, this mixture was not used in these greenhouses due to poor efficacy. Resistance to benzimidazoles of the collected isolates was determined through the tests of discriminatory doses according to previous described 16 . In brief, isolates which could not grow on potato dextrose agar (PDA) (200 g potato, 20 g dextrose, 35 g agar and 1 L H 2 O) plates amended with 5 mg/L carbendazim were considered as sensitive (Ben S); those could grow on 10 mg/L but not on 100 mg/L were defined as moderately resistance (Ben MR); and those that could grow on 100 mg/L were determined as highly resistant (Ben HR).
isolation the β-tubulin gene fragments of B. cinerea. Twenty isolates of different phenotype of benzimidazole sensitivity for each sampled year were chosen at random. The DNA fragment, which included the 198 th and 200 th codon of the β-tubulin gene, was amplified by conventional PCR using BCtubF and BCtubR ( Table 1). The volume of the reaction was 50 μL with the reagents as following: 25 μL 2 × PCR Master, 0.4 μM primers, 1 μL DNA template, ddH 2 O was supplemented to 50 μL and the thermal cycling of conventional PCR program was 95 °C for 5 min; 30 reaction cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 90 s with an extension at 72 °C for 5 min. The PCR products were analyzed through gel electrophoresis and purified using UNIQ-10 Coloum DNA Purification Kit (Sangon, Shanghai), following the manufacturer's protocol. All the PCR products were sequenced by Invitrogen Company, Shanghai, China. The sequences were aligned using Crustal W software (http://www.bi.ac.uk).
Primer design and evaluation for LAMP detection of E198A B. cinerea. The Primer explorer V5 software tools (http://primerexplorer.jp/e/) were used to design LAMP primers which targeted a fragment containing the codon 198 th and 200 th in the β-tubulin gene of B. cinerea (FQ790278.1). The detailed structure of LAMP primers used is shown in Fig. S1. Mismatching bases were introduced at the 3′end of the FIP, and six groups of specific LAMP primers (Table 1) were designed to differentiate E198A mutants from the wild-type. Two B. cinerea isolates, CL-3 (E198A, Ben HR) and TM-10 (Ben S) were used to develop the LAMP assay. The six groups of mismatched LAMP primes were screened to determine their specificity to distinguish the E198A genotype from the sensitive type. The reaction volume was a 25 μL mixture: 8 U Bst DNA polymerase (New England Biolabs, Beijing), 1 mM dNTPs (Sangon, shanghai), 4 mM Mg 2+ , 0.8 M betaine (Sigma), 1.2 μM FIP and BIP, 0.2 μM F3 and B3, 150 μM HNB (metalion indicator, Sigma), 2.5 μL 10 × Thermo Pol buffer, 1 μL double-stranded