Suspension Array for Multiplex Detection of Eight Fungicide-Resistance Related Alleles in Botrytis cinerea

A simple and high-throughput assay to detect fungicide resistance is required for large-scale monitoring of the emergence of resistant strains of Botrytis cinerea. Using suspension array technology performed on a Bio-Plex 200 System, we developed a single-tube allele-specific primer extension assay that can simultaneously detect eight alleles in one reaction. These eight alleles include E198 and 198A of the β-Tubulin gene (BenA), H272 and 272Y of the Succinate dehydrogenase iron–sulfur subunit gene (SdhB), I365 and 365S of the putative osmosensor histidine kinase gene (BcOS1), and F412 and 412S of the 3-ketoreductase gene (erg27). This assay was first established and optimized with eight plasmid templates containing the DNA sequence variants BenA-E198, BenA-198A, SdhB-H272, SdhB-272Y, BcOS1-I365, BcOS1-365S, erg27-F412, and erg27-412S. Results indicated that none of the probes showed cross-reactivity with one another. The minimum limit of detection for these genotypes was one copy per test. Four mutant plasmids were mixed with 10 ng/μL wild-type genomic DNA in different ratios. Detection sensitivity of mutant loci was 0.45% for BenA-E198A, BcOS1-I365S, and erg27-F412S, and was 4.5% for SdhB-H272Y. A minimum quantity of 0.1 ng of genomic DNA was necessary to obtain reliable results. This is the first reported assay that can simultaneously detect mutations in BenA, SdhB, BcOS1, and erg27.


INTRODUCTION
Botrytis cinerea is the causal agent of gray mold, which often causes heavy losses on many economically important crops, including vegetables, fruits, ornamentals, and bulbs. B. cinerea can attack leaves, stems, fruits, and even stored and transported agricultural products (Mbengue et al., 2016). It has been named one of the top 10 fungal pathogens in molecular plant pathology (Dean et al., 2012). Synthetic fungicides are widely used for controlling disease caused by B. cinerea. A few years after the introduction of fungicides, development of resistance in pathogen populations was observed, and B. cinerea was one of the first fungi for which resistance was described (Yourman et al., 2001;Hahn, 2014).
The most important resistance mechanism of B. cinerea is modification of the fungicide target caused by mutations in the encoding genes. Point mutations at codon 198 or 200 in the β-Tubulin gene (BenA) are related to benzimidazole resistance (Leroux et al., 2002). E198A mutant strains that possess amino acid replacement of glutamic acid by alanine have been detected worldwide as major benzimidazole-resistant strains in the field (Banno et al., 2008). A histidine to tyrosine replacement at codon 272 (H272Y) in Succinate dehydrogenase iron-sulfur subunit gene (SdhB) is responsible for the boscalid resistance phenotype (Yin et al., 2011). Resistance to dicarboximide can be associated with point mutations in the putative osmosensor histidine kinase gene (BcOS1), and strains possessing the I365S mutation are dominant (Cui et al., 2002;Oshima et al., 2002). A mutation at codon 412 (F412S) in the 3-ketoreductase gene (erg27) that causes amino acid replacement of a phenylalanine residue by serine is related to fenhexamid resistance (Debieu et al., 2013).
At present, many molecular methods have been used to detect fungicide resistance in B. cinerea, including sequencing (Leroux et al., 2010;Fillinger et al., 2012), polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP; Ishii et al., 2009;Van der Heyden et al., 2014), allele-specific PCR (AS PCR;De Miccolis et al., 2012, 2014Yin et al., 2012;Van der Heyden et al., 2014), tetra-primer amplification refractory mutation systems (ARMSs) PCR (Munoz et al., 2009), realtime PCR (Banno et al., 2008;Billard et al., 2012), and highresolution melting (HRM) analysis (Chatzidimopoulos et al., 2014). However each of these molecular assays can only detect a single allele or simultaneously detect different alleles in a single gene. Simultaneous detection of multiple alleles in a single reaction vessel is a technical difficulty in fungicide resistance identification. Increasing fungicide resistance has been a serious problem in disease control. Undoubtedly, investigating the large numbers of strains resistant to different kinds of fungicides requires massive effort. Therefore, a simple and high-throughput genotyping assay is required for large scale detection and monitoring of the emergence of resistant strains of B. cinerea.
The Bio-Plex R suspension array could be potential candidate for developing such a genotyping tool. The Bio-Plex R suspension array is based on the flexible multi-analyte profiling (xMAP) technology (Houser, 2012). It enables multiplex biological testing of up to 500 analytes within a single sample volume (Spierings and Dunbar, 2013). These 500 analytes are realized by 500 distinctly colored bead sets created by the use of three fluorescent dyes at different ratios. Each bead couples to sequence-specific capture oligos that recognize and bind target DNAs. Once bound, target DNA molecules, which are labeled with biotin, are tagged with Streptavidin-R-phycoerythrin (SAPE; Dunbar, 2006; Figure 1). The beads are individually analyzed by cytometry on the Bio-Plex 200 platform. A red laser distinguishes the internal bead dyes to identify each microsphere particle, and a green laser provides a quantitative result of the target molecules. Many applications for nucleic acid assay have FIGURE 1 | Schematic illustration of the suspension assay. The target regions BenA, SdhB, BcOS1, and erg27 containing SNPs are amplified by 4-plex PCR in a single tube. TAG-ASPE primers overlap the SNP site in the PCR-amplified target DNA, and only perfectly matching primers can be extended (eight TAG-ASPE primers were mixed in one reaction). Biotin-labeled extended products then hybridize to the microsphere coupled with anti-TAG sequences and are detected on the Bio-Plex platform.
This study established a novel and universal multiplex genotyping method. We describe a genotyping assay that combines an allele-specific primer extension (ASPE) with the Bio-Plex suspension array. Four mutations most commonly associated with fungicide resistance in B. cinerea are included in this assay, including BenA (E198A), SdhB (H272Y), BcOS1 (I365S), and erg27 (F412S). The development of this multiplex assay will greatly reduce the time, labor, and cost of screening compared to single-reaction-based detection methods. It represents important progress in the development of multiplex genotyping technologies for molecular detection of fungicide-resistance in B. cinerea.

Strains and Culture Conditions
The Bc05.10 strain of B. cinerea used as a sensitive strain was provided by Jingao Dong of the Agricultural University of Hebei (Hebei, China). Field strains resistant to different types of fungicides were provided by Chaoxi Luo of the Huazhong Agricultural University (Hubei, China). B. cinerea isolates were incubated on potato dextrose agar (PDA) medium at 22 • C for 5 days. To extract genomic DNA, mycelial disks of B. cinerea isolates were cultured in potato dextrose (PD) liquid medium at 22 • C for 7 days. After 7 days incubation, mycelia were harvested by filtration through sterile filter paper, and dried between two layers of filter paper. Fungal material was frozen at −80 • C for later use.

DNA Extraction and Plasmid Construction
Genomic DNA was extracted from each isolate using a previously described modification of the CTAB (Agbagwa et al., 2012). Template DNA was stored at −20 • C until use. Based on the B. cinerea sequence of BenA, SdhB, BcOS1, and erg27 in GenBank (accession numbers Z69263, AY726618, AF435964, and AY220532), eight nucleotide sequences containing the variants , and erg27-412S (TCC) were cloned into pUCm-T plasmids (Shanghai Sangon, CN). Gene splicing by overlap extension PCR was used for site-directed mutagenesis. All plasmid DNA constructs were confirmed by DNA sequencing. These plasmids were used to optimize reaction conditions and determine the analytical sensitivity and specificity of the bead-based suspension array. Genomic DNA and plasmid DNA were quantified using a NanoDrop 1000 Spectrophotometer (Thermo Scientific, USA).

Primers and TAG Sequences
All primers used in multiplex PCR and ASPE PCR were designed using the software primer premier 5. TAG sequences were designed by the Luminex Corporation. All oligonucleotides used in the assay were purified by HPLC and are listed in Tables 1-3.

Limits of Detection
To determine the limits of detection for each genotype, each plasmid was diluted to 10 4 , 10 3 , 10 2 , 10, and 1 copies/uL and was tested with the multiplex suspension array. To determine the lowest concentration of genomic DNA of field isolates needed in the suspension assay, genomic DNA of the IR-2 strain (benzimidazole-resistant strain) was diluted to 10, 1, 0.1, and 0.01 ng/µL. One microliter was then used as the template for multiplex PCR. The detection limit and the lowest concentration of genomic DNA were defined as the values where all replicates gave positive results for each allele. All assays were performed three times.

Determination of Sensitivity
To determine sensitivity for each Mut allele, we mixed 10, 10 2 , 10 3 , 10 4 , and 10 5 copies of Mut plasmid DNA with 10 ng genomic DNA (equivalent to 2.2 × 10 5 genomic DNA copies) as the template for 4-Plex PCR. These correspond to rates of 0.045, 0.45, 4.5, 45, and 450%. The lowest rate at which all three replicates were positive was defined as the sensitivity of the assay.

Data Analysis
For each sample tested with the Bio-Plex 200 system, MFI values were collected for each of the eight microsphere sets corresponding to each allele. Results can be exported from the Bio-Plex Manager 4.0 software into an Excel file. For each allele, NET MFI values were obtained by subtracting the respective notarget (multiplex PCR negative control) MFI values. A criterion was set as the mean MFI of 22 blank controls +5 standard deviation (SD). MFI values for each allele were required to meet this criterion, and alleles could then be determined based on allelic ratios (AR) as follows: Mutant allelic ratio = NET MFI Mut /(NET MFI WT + NET MFI Mut ) Based on preset criteria, AR was used to discriminate WT, mutant, and heterozygous samples. In this study, a mutant AR ≥ 0.03 was set to call the presence of BenA-198A, AR ≥ 0.2 was set to call the presence of SdhB-272Y, AR ≥ 0.02 was set to call the presence of BcOS1-365S, and AR ≥ 0.06 was set to call the presence of erg27-412S. When the AR was lower than these values, the WT was called. All data were analyzed and plotted using Excel 2013 or GraphPad prism 5.01.

Application to Field Isolates of B. cinerea
DNA samples from 10 field isolates of B. cinerea were analyzed using the Bio-Plex suspension array. One nanogram of genomic DNA was used as the template for multiplex PCR. To evaluate accuracy, genotyping results were compared with the results obtained by sequencing (Sangon, China).

Multiplex PCR (4-Plex)
Multiplexed PCR of genomic DNA was performed under optimized conditions. The PCR products contained mutations of interest. Four expected PCR products were amplified in a single FIGURE 2 | Specificity of bead-based assay. Individual plasmids (1 × 10 4 copies/test) representative of each of the eight alleles were used as the template in a 4-plex PCR to determine specificity. Results showed that median fluorescence intensity (MFI) values were all above 1200. The data shown are the means of three replicates.

Specificity of Bio-Plex Suspension Array
The specificity of the assay was verified with amplimers of each individual plasmid. Then, PCR products were hybridized to all eight sets of microbeads. As shown in Figure 2, no crossreactivity occurred between each probe. The mean background MFI for each allele ranged from 12 to 27, and the signal-tonoise ratio (positive MFI/ background) ranged from 81 to 137. Thus, the genotyping assay for these eight alleles was highly specific.

Detection Limit
A series of 10-fold dilutions of each plasmid (1-10 4 copies/µL) were used to determine the limits of detection for this method. The results indicated that MFI values reached threshold values when DNA concentrations of BenA-E198 (GAG), SdhB-H272 (CAC), SdhB-272Y (TAC), erg27-F412 (TTC), and erg27-412S (TCC) plasmids were as low as 10 FIGURE 4 | Detection limit of genomic DNA. Ten-fold serial dilutions of the field isolate of Botrytis cinerea genomic DNA template were tested with the suspension assay. The data shown are the means plus SD (n = 3). Genomic DNA of field isolate strains was serially diluted (0.01-10 ng/µL) to determine the lowest DNA concentration needed in the assay. Results showed that more than 0.1 ng of input DNA was required for discrimination of all eight alleles (Figure 4).

Assay Sensitivity
Serial dilutions of Mut plasmid templates (10-10 4 copies) were mixed with 10 ng WT genomic DNA (corresponding to 2.2 × 10 5 genomic DNA copies per reaction), as shown in Figure 5, and a significant concentration-dependent relationship was observed for each Mut allele. The suspension assay could detect 0.45% of BenA-198A, BcOS1-365S and erg27-412S, and detected 4.5% of SdhB-272Y when these mutations were mixed within the genomic DNA background. These results indicate that the beadsbased array had a high sensitivity.

Assay Reproducibility
To assess reproducibility of the assay, three independent assays were performed on different days. In each reaction, individual plasmid DNA aliquots (1 ng/µL) were used as templates for The data shown are the means of three replicates.
4-plex PCR. The measurement of each sample was performed three times, and the mean MFI values were calculated. Results of reproducibility assays are shown in Table 4. The coefficient of variation (CV) for each probe was below 10%, indicating high reproducibility.

Sample Analysis
Ten field strains of B. cinerea were genotyped using the Bio-Plex suspension array. The measurement of each sample was performed twice, and results are shown in Table 5. Some alleles at one locus were found to be neither WT nor Mut using this method, illustrating that other mutations exist. To identify the accuracy of the assay, sequencing of each sample was performed. Sequencing results corresponded to the suspension array ( Table 6). Alleles that could not be detected using the beadsbased assay correspond to other mutations. For instance, the BenA gene in sample 1 was neither WT nor E198A, but rather contained the E198V mutation. These results indicate that this high-throughput platform can simultaneously detect eight alleles within a single sample volume. The results from this method are clear, intuitive, and highly accurate.

DISCUSSION
Botrytis cinerea is a widespread fungus that causes gray mold rot in many plants. Due to the different fungicide-resistant molecular mechanisms, investigating the large numbers of resistant strains in the field will require a massive amount of work. However, none of the preexisting molecular methods meet the high-throughput demands for detection presently required. Here, we successfully developed and validated a multiplex microbead-based suspension array for simultaneous detection of eight alleles using the Bio-Plex 200 platform. This method will be broadly applicable in the field of plant pathology. The microbead-based suspension array combines advanced fluidics, optics, and digital signal processing. The platform provides a sensitive and ideal method for genotyping large numbers of samples at a relatively low cost. We successfully developed and validated a consolidated microbead-based genotyping platform for simultaneous detection of eight alleles in a single reaction. Four basic steps were included in this assay: (i) multiplex PCR, (ii) single-tube ASPE PCR with tagged primers, (iii) hybridization of single-stranded PCR products to microbeads carrying anti-TAG sequence, and (iv) detection on the Bio-Plex 200 platform.
With respect to the first step, we designed several primers for each gene. Different primer mixtures for the four genes were used to screen the optimized combination to ensure specificity of amplification. We also found that the 4-plex primers influenced the MFI of negative control samples. Some unsuitable primers resulted in non-specific hybridization with microbeads. After PCR primers were determined, a multiplex PCR assay was optimized for several factors, including annealing temperature, and concentrations of four primer pairs. Gel imaging illustrates Positive MFI values are the means of three replicates and are highlighted in bold. the high specificity of the PCR primers used in the multiplex reaction under optimized conditions (data not shown). Because multiplex PCR was the first step of the high throughput platform, it is essential that this reaction be as specific as possible to minimize potential problems. Multiplex ASPE PCR incorporated eight allele-specific primers to specifically target WT or mutant sites. Each 5 end of the ASPE primer attached to a unique TAG sequence. The 24-base TAG oligonucleotides only contain three kinds of bases and exclude guanine (G). xTAG technology ensures the same annealing temperature and hybridization efficiency, so that cross hybridization between targets and microbeads can be effectively avoided. Artificial changes of nucleotides were generally introduced into the 3 ends of ASPE primers to improve specificity. A series of allele-specific primers have been screened by real-time PCR to identify the eight alleles in our laboratory, and each primer introduced additional mutations. However, signal values that were too low were observed in suspension assays when using the same primers as used for real-time PCR. These results indicate that ASPE primers that introduced additional mutations were not suitable for this experiment. Later PCR cycles generated only single-stranded biotinylated products. This greatly improved the sensitivity of the assay and allowed for simple optimization.
Biotinylated ASPE PCR amplicons with TAG sequences were hybridized to complementary probes (anti-TAG). The eight complementary probes were each coupled to a unique colorcoded microsphere. Hybridization temperature affects signal output, so 54 and 37 • C were selected to identify the optimal temperature in this assay. When the temperature was 54 • C, the MFI of the blank was not affected. However, the MFI of targets was lower than when the hybridization temperature was 37 • C. We also found that using 1× Tm buffer containing 0.1% BSA to dilute SAPE can effectively reduce the background.
There were nine no calls obtained with the beads-based assay from 10 samples because the minimum signal requirements were not met ( Table 5). Sequencing results indicated that these nine calls corresponded to other mutation types ( Table 6) not within the scope of detection in this suspension assay. Except for the nine non-called alleles, the concordance between the beads-based assay and DNA sequencing was 100%. An advantage of the suspension assay is that the resulting data are easier to interpret. All results of the 10 samples were clear at a glance and can be obtained in a single day, whereas analysis of sequencing results requires additional time.
Although this assay requires four basic steps, it still saves time and effort compared to other traditional methods once reaction conditions are determined. The Bio-Plex platform uses 96-well plates, so 96 samples can be detected in one round of reactions. This is equivalent to an output of 96 × 8 = 768 genotype designations. If using traditional PCR to distinguish eight alleles, 768 reactions in 768 wells would be needed, which would be time-consuming and introduce more possibilities to make mistakes.
The Bio-Plex 200 platform allows for up to 100 different analytes to be measured simultaneously in a single reaction, so researchers can personalize their experiments according to different requirements. For this microbead-based method, more mutations in BenA, SdhB, BcOS1, and erg27 could be detected: for instance, F200Y and E198V/K in the BenA gene (Yarden and Katan, 1993), P225T/F/L and H272Y/R in the SdhB gene (De Miccolis et al., 2014), I365R/N, Q369H/P and N373S in the BcOS1 gene (Cui et al., 2004;Grabke et al., 2014), and F412I/V in the erg27 gene (Fillinger et al., 2008). All of these mutations have been reported to be associated with fungicide resistance. Reaction conditions for multiplex PCR would not need to be further optimized; screening for these mutants would only require developing TAG-ASPE primers and microbeads corresponding to these other alleles. The microbead-based assay is flexible, and researchers can use it to design highly personalized experiments.
As with any technology, there are also some limitations of suspension arrays. First, the platform requires a relatively high start-up cost due to the commercial beads and instrument. Second, as with most traditional methods, it can only analyze known fungicide-resistance mutations. Third, this microbeadbased assay requires two PCR amplifications, so it is a qualitative method and cannot be used for quantification. This is the first report that demonstrates simultaneous detection of resistance to benzimidazoles, dicarboximides, SDHIs, and SBIs. The high-throughput capacity of the beadbased methods in fungicide resistance detection is unmatched.
All results show that this multiplex, accurate, flexible, and sensitive assay represents a new efficient genotyping tool for fungicide resistance monitoring and administering of field strains of B. cinerea. The protocol presented in this study will be helpful to researchers for characterizing field strains of B. cinerea.

AUTHOR CONTRIBUTIONS
XZ wrote the paper and performed most of this work. BL participated in designing and screening of ASPE primers. FX and PZ interpreted the results and revised the paper. XM supervised the entire study. All authors reviewed the manuscript.