Abstract
A rapid, high-resolution melting (HRM) analysis protocol was developed to detect sequence variations associated with resistance to the QoIs, benzimidazoles and dicarboximides in Botrytis cinerea airborne inoculum. HRM analysis was applied directly in fungal DNA collected from air samplers with selective medium. Three and five different genotypes were detected and classified according to their melting profiles in BenA and bos1 genes associated with resistance to benzimidazoles and dicarboximides, respectively. The sensitivity of the methodology was evident in the case of the QoIs, where genotypes varying either by a single nucleotide polymorphism or an additional 1205-bp intron were separated accurately with a single pair of primers. The developed two-step protocol was completed in 82 min and showed reduced variation in the melting curves' formation. HRM analysis rapidly detected the major mutations found in greenhouse strains providing accurate data for successfully controlling grey mould.
HRM analysis is a very sensitive method allowing to detect mutation-based fungicide resistance.
HRM analysis is a very sensitive method allowing to detect mutation-based fungicide resistance.
Introduction
Botrytis cinerea Pers.:Fr. is the causal agent of grey mould disease, which is responsible for serious losses in more than 200 crop species worldwide (Williamson et al., 2007). In breakouts, the disease is efficiently controlled with the application of botryticides. Compounds with site-specific mode of action, inhibiting either the mitotic division (benzimidazoles; MBCs) or involving osmoregulation of the fungal membranes (dicarboximides; DCs), were exclusively applied since the 1970s to control the disease. Evolving for over three decades, B. cinerea isolates exhibiting dual resistance against both MBCs and DCs were spread widely (Chatzidimopoulos et al., 2013). Resistance to the quinone outside inhibitors (QoIs) was developed 2 years after their introduction into the market (Bartlett et al., 2002). The application of the QoIs, such as pyraclostrobin, alone or in mixtures with boscalid (SDHI class) as an antiresistance strategy had only partially effect in delaying the resistance. Latest reports showed that dual resistance against both respirator inhibitors may arise from a population already resistant to certain fungicides including the QoIs (Amiri et al., 2013; Chatzidimopoulos et al., 2014).
Nucleotide substitutions in BenA gene, which encodes β-tubulin in the microtubule assembly, and bos1 gene, which encodes histidine kinase, serve as molecular markers to discriminate resistant field isolates of B. cinerea to MBCs and DCs, respectively (Yarden & Katan, 1993; Oshima et al., 2006). In addition, several molecular mechanisms associated with QoI resistance were reported, the most prominent being the substitution of glycine (G) to alanine (A) at position 143 of the cytb gene (Fernández-Ortuño et al., 2008). Molecular methods such as direct sequencing and RFLP analysis are regularly applied in fungicide resistance studies to detect the associated polymorphisms (Luck & Gillings, 1995). However, more recently, qPCR-based methods were developed for genotyping MBC-, DC-, and QoI-resistant isolates of B. cinerea (Banno et al., 2008, 2009).
Introduced in 2002, high-resolution melting (HRM) analysis reveals differences in melting curve shape that is highly correlated with the genotype. The method could be considered as a qPCR descendant that carries all the closed-tube method benefits such as reduced risk of contamination and DNA degradation (Madesis et al., 2014). HRM analysis has many applications in clinical studies associated with resistance of human pathogens in antibiotics (Ong et al., 2010; Gabriel et al., 2012). Hitherto, however, its applications in pesticide studies are quite limited (Pasay et al., 2008; Chatzidimopoulos et al., 2014). The developed two-step HRM analysis method presented in this study highlights the discrimination of multiple SNPs associated with resistance to three different classes of fungicides in B. cinerea isolates and establishes a reference pattern for tracking resistance.
Materials and methods
In December 2012, during the second growing season of a 3-year field experiment, portable air samplers (Burkard Manufacturing Co. Ltd, Hertfordshire, UK) containing plates (ø9 cm) with selective medium (Edwards & Seddon, 2001) slightly modified (Supporting Information, Fig. S1) were used to entrap B. cinerea airborne propagules. Traps were operated for 1 h near midday every 10-14 days in an experimental lettuce glasshouse located at Magnesia, Greece (39°13′10″Ν, 22°45′24″Ε). Fungicides were applied in 20-day intervals to control the disease (Fig. S2). To monitor the resistance to the QoIs, MBCs and DCs population of the pathogen, discriminatory concentrations of 10 mg L−1 pyraclostrobin (F-500 25% a.i. EC, BASF, Ludwigshafen, Germany) + 100 mg L−1 salicylhydroxamic acid (SHAM; Sigma-Aldrich, Steinheim, Germany), 1 mg L−1 carbendazim (Carbendazim 50% a.i. WP, Cequisa SA, Barcelona, Spain) or 3 mg L−1 iprodione (Rovral 75% a.i. WG, BASF AG) had previously added to the medium (Chatzidimopoulos et al., 2013). Unamended with fungicides plates were used as controls. Colony-forming units (CFUs) emerged after 6 days of incubation at 20 °C in the dark (Fig. S1). To determine the phenotype, CFUs were screened for resistance to several classes of fungicides by aseptically touch with a needle the sporulating CFU and then deposit on premarked points on the periphery of Petri dishes with agar media amended with fungicides (Fig. S3). To perform molecular analyses, fungal DNA was extracted as proposed by Chatzidimopoulos et al. (2014) with a slight modification: a loop of the forming colony was placed directly in a 2-ml tube and ground to a fine powder with the aid of liquid nitrogen using a micropestle. The final concentration was adjusted to 20 ng μL−1.
For HRM, novel primers CG143-F (5′-CCCTACGGGCAAATGTCACT-3′) and CG143-R (5′-GTCCAATTCATGGTACAGCACTCA-3′) targeting the cytb gene were designed to amplify a 73-bp product containing the target polymorphic region. In addition, primers BN198.200-F/R (5′-AGCCATATAACGCAACTC-3′/5′-TATCGTAAAGAGCCTCGT-3′) and OS365.73-F/R (5′- GTAGGAACTGAAGGTATTCT-3′/5′-TGACGTTCACTATCAATGT-3′) were used to generate 87- and 82-bp amplicons of BenA and bos1 genes. All primers were optimized to anneal at 60 °C.
Nucleic acid extracts (1 μL of each) were used directly in 20-μL reaction volumes containing 1× Kapa HRM Fast PCR kit (Kapa Biosystems, Cape Town, South Africa), 2.5 mM MgCl2 and 0.2 μM of each primer. PCR-HRM was performed with the Corbett Rotor-Gene 6000 instrument (Corbett Research, Mortlake, Vic., Australia) using a fast, two-step cycling protocol (45 cycles) with 5-s denaturation (95 °C) and 30-s annealing/extension (60 °C). Up to 36 samples were used for each assay. High-resolution melting analyses were performed at the temperature ramping and fluorescence acquisition setting recommended by the manufacturer, that is temperature ramping from 70 to 95 °C, rising by 0.1 °C/2 s. The melting curves were normalized by calculation of the ‘line of best fit’ inbetween two normalization regions before and after the major fluorescence decrease representing the melting of the PCR product using the accompanying software (rotor gene q series software v.2.0.2). This algorithm allows the direct comparison of the samples that have different starting fluorescence levels. All samples were plotted according to their melting profiles. Under the difference graph, melting profiles of the samples were compared to that of the wild type, which was converted to a horizontal line. Significant deviations from the horizontal line indicated the presence of mutation and were recorded as HRM positive. To examine the reproducibility of the method, that is consistency of each melting profile, the assays were conducted in triplicate, and the tests were repeated over several days, maintaining the same test conditions, reagent quantities and DNA concentrations.
Two different PCR assays were carried out to verify the results in HRM's clusters formation regarding CG143-F/R amplicon: an allele-specific PCR assay produced by the primer pair BcAR-F/R for the detection of the G143A substitution and a simple PCR produced by the primer pair cytb-Bc-F/R for the detection of the extra intron (Jiang et al., 2009). To verify the genotypes in partial BenA and bos1 genes, representative samples of each HRM's cluster were subjected to PCR with primers TUB-HP-F/R and BcOS1-HP-F/R (Banno et al., 2008) prior to sequencing. Primers were used in a rapid PCR protocol (Chatzidimopoulos et al., 2014) with annealing temperature at 55 and 58 °C, respectively. Purified PCR products (Nucleospin gel and PCR clean-up, Macherey-Nagel GmbH & Co., Düren, Germany) were sequenced according to dideoxy chain termination method (Sanger et al., 1977).
Results and discussion
HRM analysis discriminated the polymorphisms in the targeted loci. The amplicons with fluorescent dye were melted to generate the melting curves. Wild-type isolates were selected as reference genotypes to generate the difference plots. Three different melting profiles were formed in the CG143-F/R amplicon (Fig. 1a). The melting curves of three samples were different by 4 °C from the other two closely formed clusters (Fig. 1a). PCR revealed that these genotypes corresponded to a 1200- to 1300-bp product most probably carrying the Bcbi-143/144 intron. As expected (Venter et al., 2001), a G to C transversion (G143A substitution) would have a 0.3 °C in the curve shift (Fig. 1b). The presence of this mutation, which is responsible for the resistance to the QoI fungicides, was confirmed by allele-specific PCR.
Sensitivity of HRM analysis in detecting variants within CG143-F/R amplicon after normalization of the melting curves. (a) Detection of different genotypes ranging from a single nucleotide polymorphism conferring resistance to the QoIs (G143A) to the presence of an extra 1205-bp intron. (b) A G→T transversion (G143A change in amino acid composition) caused a 0.3 °C shift of the melting curves 445 × 543 mm (300 × 300 DPI).
In the BenA locus, isolates carrying either the E198A or the F200Y substitution formed two clusters clearly distinguishable from that of the wild type. Sequencing the amplicon TUB-HP-F/R (compared to the wild-type strain SAS56, Genbank accession number Z69263) showed that genotypes of the top major cluster had an A to C transversion, which resulted in amino acid substitution glutamic acid for alanine (E198A) to the protein product in 18 isolates. These isolates corresponded to the highly resistant to MBCs (BenHR) phenotype. Likewise, a T to A transversion changed amino acid composition of phenylalanine for tyrosine (F200Y) in moderately resistant to MBCs (BenMR) isolates. All these variants had difference plots distinguishable from the other genotypes (Fig. 2b). Thus, the resistant strains could be clearly distinguished from the wild type through the difference plots of these substitutions.
High-resolution melting analysis for the detection of multiple polymorphisms at positions 198 and 200 in the BenA protein, associating with high and moderate levels of resistance to benzimidazoles. (a) Normalized data. (b) In the difference graph the resistant genotypes were easily distinguished as the clusters were formed in both sides of the wildtype genotype. In this case, all the samples were included to demonstrate that the clusters were formed distinctly despite the experimental error (variation).
In partial bos1 gene, five distinct clusters were formed, including the wild type. As depicted in Fig. 3a and b resistant to DCs, isolates generated melting curves according to their polymorphisms in the amplicon. Compared with the wild-type genotype (Genbank accession number AF435964): nucleotide substitutions of T for A or G resulted in changes in amino acid synthesis isoleucine for (1) asparagine (I365N) or (2) serine (I365S); (3) two substitutions of A for C and A for G altered the amino acid compositions glutamine for proline (Q369P) and asparagine for serine (N373S); (4) nucleotide alterations G for T and G for C resulted changes in amino acid compositions valine for phenylalanine (V368F) and glutamine for histidine (Q369H), respectively. All these variants had difference plots distinguishable from the other genotypes (Fig. 3b). Overall, all substitutions revealed by sequencing were correctly identified by HRM assays giving a correlation of 100%.
Results of high-resolution melting analysis from iprodione-sensitive and iprodione-resistant B. cinerea isolates demonstrating five different genotypes. (a) Normalized melting curves. (b) Difference graph. Representative melting curves demonstrate five distinct clusters including the wild type. The I365S, I365N, V368F and Q369H,Q369P and N373S groups of amino acid substitutions are associating with moderate levels of resistance to dicarboximides 444 × 540 mm (300 × 300 DPI).
Spore traps with selective medium are an easy way for estimating the fungicide resistant proportion in B. cinerea populations. Although the addition of fungicides to the medium was not necessary for the HRM analysis it confirmed the association of the phenotype with the genotype. To extract an adequate amount of DNA from each CFU, a relatively extended cultivation period required. However, it may be possible to reduce this timely consuming process in only 24–36 h by modifying the phase of the selective medium and the overall approach of the molecular procedure. It has been demonstrated that the HRM analysis can detect fungal DNA in host tissue at concentrations higher than 3 pg μL−1 (Luchi et al., 2011). In this case, HRM analysis may deploy as a semi-quantification assay for the detection of the resistant alleles. Further research is currently under progress towards this direction.
When compared to DNA derived from pure single-spore cultures (CDNA: 400–900 ng μL−1, A260/280: 2), the quantity and quality of DNA extracted from CFUs were significantly lower (CDNA: 40–50 ng μL−1, A260/280: 1.7). However, to smoothly operate the HRM analysis, all DNA samples should be prepared by a single extraction protocol. If differences exist in the concentrations of the template DNA, PCR amplification will result in different yields of PCR product, which will further influence the melting analysis (Chatzidimopoulos et al., 2014). In the present assay, it was observed that the use of the HRM kit reduced the variation in the melting curves' formation. Homogeneity of the PCR mixtures is the first step and one of the main determinants for a successful HRM analysis.
The data presented provide useful information for the dynamics of the population and the strategy that would take place. Three different cases of multiple SNPs were examined to test the behaviour of the HRM analysis. The resistance to the QoIs was quite common among different phenotypes. One of this study's objectives was to monitor the heteroplasmic state of the cytb gene via the HRM analysis. This would become visible by small shifts on the melting curves between the two intron-less clusters formed (Luchi et al., 2011). However, most resistant isolates showed a dominant homoplasmic population of A143 alleles. In fact, the presented analysis protocol worked exemplary in distinguishing the 143 alleles. The screening for resistant A143 alleles directly from CFUs was found suitable for detecting resistance. In vitro a loss of the resistance might be expected even from the first transfer of the culture in fungicide-free media (De Miccolis Angelini et al., 2012b). High resistance to anilinopyrimidine fungicides (AP) has also detected in B. cinerea isolates. The primary mode of action of these fungicides has not yet been clarified, but there are evidence involving methionine biosynthesis. Even though the role of several genes (cgs, cgl and cbs) coding for enzymes in methionine biosynthesis has been investigated, none was associated with AP resistance (De Miccolis Angelini et al., 2012a). It is conceivable that HRM analysis could apply in small fragments (up to 150–200 bp) of these genes, to screen differences in the genotype of numerous isolates (up to 96 samples can be screened simultaneously). The association of the genotypes with the phenotypes will provide useful evidence regarding the part of the genes involved in resistance.
In the present study, an HRM analysis method was developed for scanning sequence variants associated with resistance to fungicides in B. cinerea isolates. This new technology represents a tool to rapidly detect the major mutations found in field or greenhouse strains resistant to these fungicides. Remarkably, with a single pair of primers, the protocol was sufficient to discriminate up to five different sequence variants within a target area. The data presented provide further evidence for the integrity of this novel methodology in plant pathogens' mutation scanning (Chatzidimopoulos et al., 2014). The discriminatory capability and sensitivity of HRM analysis was demonstrated in a single plot, where genotypes carrying either a SNP or an extra 1205-bp intron were separated accurately. Natural polymorphisms may generate a different melting profile. In this case, sequencing is necessary to properly identify the new melting profile. This extra group of melting curves can then either be highlighted or excluded from the presentation graph. The most essential parameter for generating the melting curves is the melting temperature (Tm), which is closely related with the GC content and the length of the amplicon; when implemented correctly, the need of sequencing disappears (Reed et al., 2007).
In general, where multiple samples contain the same sequence variant, the melting profiles group together tightly. Although amplicon HRM has been used as a genotyping technique, it is as a scanning technique that it finds wider application (Taylor, 2009). Of all available scanning techniques, high-resolution melting is a closed-tube method that can be performed in the same container that was used for PCR amplification. Such a method requires no processing and is immediately available after scanning for genotyping or sequencing if necessary. Most real-time PCR instruments generally lack the necessary resolution for high-sensitivity scanning. A combination of qPCR-HRM could extend the possibilities in molecular research as detection, quantification and genotyping can be performed as an all-in-one assay, rendering HRM a method of choice for pesticide resistance studies (Reed & Wittwer, 2004; Rouleau et al., 2009).
References
Author notes
Editor: Holger Deising
