Development of a TaqMan qPCR assay for the detection and quantification of Gnomoniopsis castaneae in chestnut tissues

Gnomoniopsis castaneae G. Tamietti (syn. Gnomoniopsis smithogilvyi L.A. Shuttleworth, E.C.Y. Liew & D.I. Guest) (Crous et al., 2012; Tamietti, 2016; Visentin et al., 2012) is an emerging fungal pathogen causing brown rot of Castanea sativa, C. mollissima and the hybrids C. sativa x C. crenata (Sakalidis et al., 2019). The disease severely impacts fruit production and marketing in Europe with very high incidence, as for Switzerland with 91% of infected fruits (Dennert et al., 2015) and Italy, up to 93.5% (Lione et al., 2015). Brown rot of kernels is also reported in Australia (Shuttleworth et al., 2013), North America (Sakalidis et al., 2019) and Chile (Vannini, unpublished). Gnomoniopsis castaneae can also incite shoot blight and leaf necroses on chestnut as well as bark cankers on both chestnut (Dar & Rai, 2015) and hazelnut (Corylus avellana), as exhaustively reviewed by Lione et al., (2019). The ecology, biology and epidemiology of this fungus are particularly complex. Gnomoniopsis castaneae is a cryptic species commonly found as an endophyte in all tissues of chestnut and additional hosts such as Quercus spp., Fraxinus ornus L. and Pinus pinaster Ait. (Lione et al., 2019). The severe impact of the pathogen in the last decade was associated with a massive presence of inoculum in the environment boosted by climate change Received: 15 February 2021 | Revised: 3 May 2021 | Accepted: 10 May 2021 DOI: 10.1111/efp.12701

tissues is urgently required to better understand G. castaneae ecology, biology and epidemiology. Primers and a species-specific probe for G. castaneae were designed based on the sequence of the single-copy elongation factor 1 alpha (EF1α) gene. The amplification efficiency of target DNA was 105.3% and the limit of detection of the assay was calculated at approximately 40 fg of pure fungal DNA. The pathogen was consistently detected in artificial mixtures of plant and pathogen DNAs with the same Limit of Detection (LOD) as pure fungal DNA. In naturally infected samples, the assay rapidly revealed the presence of the pathogen in all symptomatic specimens, as well as in asymptomatic tissues. Notably, a significant relationship between the results of a metagenomic HTS analysis and the qPCR assay on DNAs extracted from bulk fruit was found. This molecular tool will be of substantial aid in detecting and quantifying G. castaneae, even in the endophytic state, and in different host tissues. (Lione et al., 2015), in synergy with infestation by the Chinese Gall Wasp, Dryocosmus kuriphilus Yasumatsu (Fernández et al., 2018;Magro et al., 2010;Vannini et al., 2017). The colonization and necroses caused by Chinese Wasp galls are believed to start from the endophytic inoculum (Vannini et al., 2018), while indirect evidence supports floral infection by external inoculum as the main pathway of fruit colonization and rot (Shuttleworth & Guest, 2017). Although artificial inoculations reproduced the symptoms, it is not yet clear how the pathogen infection process leads to the development of bark cankers (Pasche et al., 2016). In this context, the endophytic behaviour of G. castaneae might play a key role in the biology of the fungus and its epidemiology. The possibility of monitoring the distribution of the inoculum in the different tissues and the patterns of endophytic inoculum accumulation as a function of host phenology and environmental parameters is a paramount requirement in clarifying the biology of the fungus and understanding the ability of this organism to shift from an endophytic to a pathogenic phase. At the moment, no protocols are available to monitor the presence of G. castaneae in chestnut tissues other than the classical biological detection through isolation in pure culture, highly specific but of low sensitivity (Manias et al., 2020). Quantitative PCR (qPCR) represents a powerful, accessible tool to address this issue. Thus, the aim of this work was to design and develop a highly sensitive and specific TaqMan assay for the detection of G. castaneae and submit it to a robust set of validation tests, including a comparison with parallel metagenomic data.

| Primer and probe design
A preliminary screening was carried out to select the most informative DNA regions for G. castaneae in fungal barcoding genes (i.e. ITS rDNA, partial β-tubulin, nuclear ribosomal RNA gene large subunit [LSU] and translation elongation factor 1 alpha EF1α). Available sequences of G. castaneae and additional Gnomoniopsis spp. were downloaded from the NCBI database and aligned with MUSCLE, implemented in Unipro UGENE v.37 (Okonechnikov et al., 2012).
The most favourable regions for the primers and probe design were manually selected and Primer3, implemented in the same UGENE software, was used with the default search criteria to precisely identify the regions. Primers and probe were synthesized by Eurofins Genomics. The TaqMan probe was labelled with the reporter dye FAM (6 -carboxyfluorescein) on the 5′ end, and the quencher BHQ1 (Black Hole Quencher 1) on the 3′ end.

| Fungal strains and DNA extraction
To validate the assay, 17 isolates of G. castaneae from different sampling sites and years, 3 strains of a different Gnomoniopsis species and 13 additional fungal taxa commonly isolated from chestnut tissues, were obtained from the DIBAF fungal collection of the Regional project Sancast (www.sanca st.it). Taxonomic details of isolates and GeneBank accession numbers of ITS rDNA sequences for representative strains are listed in Table 1.
Pure isolates were subcultured to PDA and incubated at 27℃ in the dark for 7 days before scraping the mycelium from the agar surface and extracting the DNA using the NucleoSpin Plant II mini kit (Macherey Nagel), following the manufacturer's instructions.
DNA concentration was measured with Qubit (Thermo Fisher) using the High Sensitivity dsDNA Assay kit. Extracted DNA was stored at −20℃ until further analysis.

| Field samples and DNA extraction
A series of 26 samples including 17 individual fruits (10 symptomatic and seven asymptomatic), five leaves and four twigs were collected from chestnut trees in the Monti Cimini area in September/October 2020 (Table 2). Tissues from these samples were divided into two parts after surface sterilization: one half was used in a standard isolation procedure on PDA, whereas the remaining part was ground in a TissueLyser II (Qiagen, Hilden) and 200 mg of the powder used for total DNA extraction with the same kit described above. To obtain pure and putative endophytes-free chestnut DNA, the same method was applied to in vitro plantlets of C. sativa, kindly provided by Dr. Beatriz Cuenca Valera (Grupo TRAGSA-Sepi). Finally, for comparative metagenomic analysis DNA was extracted from 15 bulk samples, obtained by separating the endocarp and pericarp from 500 g fresh chestnut fruit, and grinding them independently, as above. All extracted DNA was stored at −20℃ until further analysis.

| Tuning qPCR assay
After a series of optimization experiments aimed to determine the best performing concentration for primers and probe (data not shown), the qPCR reaction mix comprised 10 μl 2× GoTaq Probe

| qPCR performance testing
Primers and probe were designed to provide the best discrimination at species level, that is maximizing the number of mismatches in primers and, more importantly, in the probe. To test these factors, a preliminary in silico check of the specificity of the whole assay was carried out by NCBI BLAST analysis on the NCBI nucleotide collection (nr/nt) database. Subsequently, this analytical feature of the assay was wet-lab tested using the panels of DNA extracted from the fungal taxa listed in Table 1.
To assess the efficacy of the assay for detecting the pathogen, the efficiency was evaluated through the proportionality of Cq values in respect to the amount of target template DNA and sensitivity measured according to the limit of detection (LOD) and limit of quantification (LOQ) of target DNA, respectively defined as the lowest concentration of target DNA at which 95% of the positive samples can be detected or quantified. Analytical parameters were tested on a set of 10-fold serially diluted DNA from G. castaneae strain GN01, ranging in concentration from 10 ng μl −1 to 10 fg μl −1 . Five replications were amplified for each dilution to prepare the standard curve. LOD and LOQ were estimated using a curve-fitting modelling approach (Merkes et al., 2019) with the R script code available at https://github.com/cmerk es/qPCR_LOD_Calc. Repeatability was determined on 10 replicates of 4 standard DNA concentrations (10 ng, 1 ng, 100 pg and 10 pg per PCR), while reproducibility was assessed on 3 replicates of standard DNA at the TA B L E 1 List of fungal taxa used in the present study. Their source tissue, sampling site and accession numbers (when available) of ITS rDNA sequences in NCBI database are indicated The effect of inhibitors potentially contained in chestnut tissues was tested using DNA extracted from endophyte-free in vitro plantlets of Castanea sativa. Fifty nanograms of plant DNA was spiked with increasing DNA concentrations of the pathogen (10 ng, 1 ng, 100 pg, 10 pg and 1 pg per reaction, 3 replicates each).

| qPCR validation on naturally infected samples
The qPCR assay was finally tested on the DNA extracted from 26 samples of fruit, leaves and twigs collected from chestnut trees

| Primer and probe design
Among the four barcoding genes initially trialled, the elongation factor 1 alpha (EF1α)  The in silico blast of the primer pair to the NCBI GenBank database returned no sequences without mismatches on the primers other than the target EF1α-gene fragment of G. castaneae. This finding, together with the highly specific probe sequence, suggested that the assay had the desired specificity.

| Specificity
The specificity of the assay was wet-lab validated by amplification of DNAs obtained from 17 isolates of G. castaneae from different sampling sites and years, three strains of a different Gnomoniopsis species, and 13 fungal species commonly isolated from chestnut tissues or organs. Gnomoniopsis castaneae DNAs were consistently amplified, whilst all the other species, including Gnomoniopsis sp., gave no amplification signal (Table 1).

| Sensitivity
Amplification of pure DNA from strain GN01 was always proportional to its concentration. The standard curve generated by plotting five repetitions of each log DNA concentration against the Cq value as determined by qPCR, resulted in a linear response over six logs, from 10 ng to 100 fg, with a high correlation coefficient (r 2 = 0.991;  (Table 3).  Table 4).

| qPCR validation on naturally infected samples
The connection between data obtained by qPCR and HTS metagenomics was tested by linear regression analysis. The equation of the best fit line, together with its graphical representation, and the experimental data are reported in Figure 3. The R 2 is 0.62 and the F-test reported significance p < 0.001 (F = 115.4, NDF = 73).

| DISCUSS ION
In this paper, we report the design, development and validation of a novel G. castaneae real-time PCR detection assay, to our knowledge the first tool of this type available for this threatening pathogen.
The assay reliably distinguished G. castaneae from other taxonomically related fungi and from other fungi usually recognized as endophytes or saprophytes of chestnut tissues that can putatively contaminate field samples. The specificity of the assay was strictly related to the choice of the EF1α-gene as the target for amplification. This gene is known to possess the necessary level of polymorphism to discriminate taxonomically related species (Roger et al., 1999;O'Donnell et al., 1998), is less sensitive to base composition artefacts (Hashimoto & Hasegawa, 1996) and provides robust phylogenies among major eukaryotic groups (Baldauf & Palmer, 1993), therefore representing an effective alternative to ITS ribosomal DNA markers.
Moreover, EF1α appears to be consistently present in genomes as a single-copy gene, a key feature that improves the likelihood of the assay giving precise quantification of the pathogen. Although the unique nature of the gene might slightly reduce the sensitivity of the assay in comparison with assays targeting multicopy genes (as with the ITS ribosomal DNA), it adds great accuracy and affordability to quantifying fungal infection in plant tissue because of the direct proportionality between the fluorescence measured and the number of fungal cells present (Guinet et al., 2016;Klymus et al., 2020;LaSalle et al., 2011). However, the assay proved to be excep-  This proficiency in detecting the pathogen in very low quantities, irrespective of the presence of other fungal or plant DNA, demonstrates the significant advantages and great diagnostic value of this assay. When applied in the field, this assay will enable the movement of infected plant material, regardless of the infected part of the plant, fruit or plantlets, and its level of infection, to be curtailed, conceivably helping to reduce the spread of the pathogen on regional, national and international scales. It can also have a crucial impact in postharvest fruit storage, where control strategies aimed at reducing the impact of rot-causing fungi are crucial in reducing significant economic losses over time (Maresi et al., 2013;Ruocco et al., 2016); the opportunity to detect a latent infection could greatly help in optimizing the nature and the timing of treatments.
Another further goal for researchers is the development of strategies for the control of G. castaneae in the field. Even in this application, this qPCR assay will be a significant tool in tracking the presence and the amount of pathogen infection in order to evaluate the effectiveness of control measures.
In addition, this qPCR assay will help shed light on the numerous still unknown biological facets of this emerging pathogen, such as its detection in Dryocosmus kuriphilus and in the galls induced by this wasp (Morales-Rodriguez et al., 2019;Seddaiu et al., 2017;Vannini et al., 2017) to better understand the interactions between these organisms. (Grupo TRAGSA -Sepi, Spain) for having provided 'in vitro' plantlets of C. sativa for the present study.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflict of interest.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/efp.12701.

DATA AVA I L A B I L I T Y S TAT E M E N T
Sequence data created and analysed in this research are openly available from Genbank® (https://www.ncbi.nlm.nih.gov/genba nk/) and the accession numbers for each data are available in the paper. Other data supporting the findings of this study are provided in full in the results section of this paper and available from the corresponding author upon request.