Introduction

The genus Botrytis, which belongs to the family Sclerotiniaceae, is comprised of ascomycetous plant pathogenic fungi. While this genus was originally considered to be composed of 22 species and one hybrid (Staats et al. 2005; Walker et al. 2011), recent improvements in robust molecular techniques have allowed for the discovery of previously undescribed Botrytis species. Accordingly, the current literature suggests that the genus Botrytis is defined by approximately 30–35 species (Garfinkel et al. 2017, 2019), most of which reportedly have a narrow host range and are thought to be highly specific (Staats et al. 2005; Garfinkel et al. 2019). These species are mostly referred to as specialists and are responsible for pre- and post-harvest blossom blights, fruit rots, leaf spots and bulb rots in various afflicted hosts (Staats et al. 2005). An exception to this specialist concept is B. cinerea, which exhibits a broad host range. This notorious generalist pathogen is responsible for large-scale agricultural and economic losses, a consequence that can be attributed to its ability to infect approximately 1400 hosts (Shaw et al. 2016). While this broad-spectrum pathogen is considered to be most ruinous on mature and senescent plant tissues, infections are generally established during early developmental stages of the hosts, where it remains quiescent until environmental conditions become favourable for large-scale degradation (Williamson et al. 2007). This period of dormancy poses a considerable problem for many crop production systems as seemingly healthy plant material may only begin to display signs of infection after harvest. At this stage, whole batches may be at risk and are often discarded, resulting in significant reductions in post-harvest yields despite promising pre-harvest forecasts. Although less prolific, B. cinerea is also known to cause disease symptoms in many crops during the growth and development stages, resulting in noteworthy pre-harvest losses (Williamson et al. 2007).

It is evident that B. cinerea causes extensive damage on a global scale. This translates into both direct and indirect monetary losses. Direct losses are a consequence of the in planta necrotrophic activity of the pathogen, while indirect losses result from input costs required for the prevention and management of the pathogen. Even though biocontrol agents are increasing in popularity (Moser et al. 2008), chemical fungicide application remains the primary method for Botrytis management (Dean et al. 2012).

As of 2018, the South African blueberry industry had ± 2000 hectares of established orchards across several provinces, generating an estimated yield of 11,300 tonnes per annum (Pienaar et al. 2019). Increasing international demand, high profitability and significant job creation have contributed to the rapid growth of this industry, which has doubled in value each consecutive season over the past 4 years (Pienaar et al. 2019). Despite increasing crop yields across the industry, individual blueberry orchards are known to be plagued by grey mould spoilage. While strains of B. cinerea that are resistant to several fungicides have been reported to occur on blueberries (Saito et al. 2016a; Weber 2011), it is not known if the same resistance profiles are found in South African orchards. Additionally, B. cinerea is known to establish infections in almost all plant organs, including flowers, fruits, leaves, shoots and soil storage structures (Williamson et al. 2007). This is particularly problematic for crops that have overlapping phenological stages throughout the growing season. For example, certain varieties of blueberries grown in South Africa flower and fruit on the same season’s growth, a consequence of limited chilling hours below 7 °C (Polashock et al. 2017). That is, at certain times, a single blueberry plant can simultaneously hold flower buds, flower blooms, immature fruit and mature fruit. Such an overlap makes it difficult to identify the critical control points when designing prevention strategies.

Therefore, this study was conducted with the primary aim of identifying and partially characterising putative B. cinerea strains found in a blueberry orchard located in the Western Cape province of South Africa. The objectives were to (i) isolate B. cinerea strains and other prominent fungal pathogens in an afflicted blueberry orchard from different types of tissues of the crop, (ii) identify the isolates using a combination of molecular and phylogenetic techniques and (iii) determine the explicit potential of the B. cinerea population to cause decay in healthy blueberries.

Materials and methods

Collection and isolation

Two varieties of southern highbush blueberries (Vaccinium corymbosum) were sampled 3 times throughout the 2020 growing season from a farm in Stellenbosch, located in the Western Cape province of South Africa. At each sampling event, a predetermined strategy was followed. Briefly, five rows were chosen for each of the two varieties of interest. The material was then collected from these identified rows, sampling randomly from the plants on both sides of the aisles. Depending on the availability, four different healthy tissue types were collected, namely fruit, flowers, leaves and undifferentiated buds. Representative samples were taken from the top, middle and bottom areas of the plants to maintain uniformity in the sampling. Additionally, visibly diseased plant material showing symptoms typical of Botrytis infections was collected from each row. While the type and incidence of symptomatic tissue varied per variety and time, the majority of samples taken throughout the study were diseased berries and flowers.

After collection, all plant material was kept at 4 °C until processing. To recover fungi from the symptomatic samples, diseased plant material was placed directly onto potato dextrose agar (Neogen, MI, USA), amended with streptomycin and chloramphenicol (PDA + s/c) at 0.1 ppm each and incubated at 26 °C in the dark for 7–10 days. Botrytis-like colonies were sub-cultured on PDA and maintained as pure cultures until further analysis. To recover endophytic fungi from the asymptomatic tissues, 10 g of each sample of healthy plant material was first washed in a 2% NaOCl solution for 1 min, then a 70% ethanol solution for 30 s and finally sterile distilled water (dH2O) 3 times for 1 min (Rajaguru and Shaw 2010). The surface sterilised material was then added to 50-ml sterile dH2O and macerated in a commercial blender for 20 s on the highest speed setting. Between each sample, the blender was rinsed with dH2O and sprayed with 70% ethanol. The resulting mixtures were diluted in sterile dH2O, whereafter 100 μl of the 10−1, 10−2 and 10−3 dilutions were spread in triplicate on PDA + s/c, and incubated at 26 °C in the dark for 10–14 days. The subsequent filamentous fungi were sub-cultured on unamended PDA and maintained in an axenic state until further processing.

DNA extraction, amplification and sequencing

Genomic DNA was extracted from 1- to 2-week-old pure cultures using Quick-DNA Fungal/Bacterial Miniprep Kits (Zymo Research, CA, USA) according to the manufacturer’s instructions. Polymerase chain reaction (PCR) was used to amplify three genes of interest from all isolates that morphologically resembled Botrytis spp., namely the glyceraldehyde-3-phosphate dehydrogenase (G3PDH), heat-shock protein 60 (HSP60) and DNA-dependent RNA polymerase subunit II (RPB2) genes. The primer pairs used for the amplification of these nuclear DNA genes were the same as those used by Staats et al. (2005). Therefore, the relevant primers were as follows: G3PDHfor + (5’-ATTGACATCGTCGCTGTCAACGA-3’) and G3PDHrev + (5’-ACCCCACTC GTTGTCGTACCA-3’); HSP60for + (5’-CAACAATTGAGATTTGCCCACAAG-3’) and HSP60rev + (5’-GATGGATCCAGTGGTACCGAGCAT-3’); RPB2for + (5’-GATGATCGTGATCATTTCGG-3’) and RPB2rev + (5’- CCCATAGCTTGCTTACCCAT-3’). The thermocycling conditions for these three genes were as described by Staats et al. (2005) where HSP60 and RPB2 gene fragments were amplified using the following settings: an initial denaturation of 94 °C for 5 min; 35 cycles of 94 °C for 30 s, 55 °C for 30 s and 72 °C for 90 s; a final extension of 72 °C for 10 min. The same parameters were used for the amplification of G3PDH gene fragments, except the annealing temperature was increased to 64 °C.

All PCR amplifications were performed in 10-µl reactions containing 4.6-µl nuclease-free water, 5-µl Taq ReadyMix (Kapa Biosystems, Cape Town, South Africa), 0.2 µl of both the forward and reverse primers (final concentration of 0.2 µM each) and 0.5-µl template DNA. Following confirmation of amplification by visualisation on a 1% agarose gel, PCR products were sequenced in the forward direction (CAF, Stellenbosch, South Africa) using the same forward primers as for the initial amplification. Sequences were obtained at the Central Analytical Facility, Stellenbosch University, using an ABI3730xl sequencer.

Phylogenetic analysis

The G3PDH, HSP60 and RPB2 partial gene sequences from 85 Botrytis spp. isolates collected in this study were viewed and trimmed using Geneious Prime (v. 2020.2.4). These sequences were aligned with reference sequences of Botrytis spp. and outgroups Sclerotinia sclerotiorum and Monilinia fructigena obtained from GenBank (Staats et al. 2005; Garfinkel et al. 2017, 2019). Multiple sequence alignments were carried out using ClustalW in MEGA (v. 6.06) (Tamura et al. 2013). The default parameters were used, where the pairwise and multiple alignments had a gap-opening penalty of 15, a gap extension penalty of 6.66, IUB DNA weight matrix, a transition weight of 0.5 and a delay divergent cut-off of 30% (Saito et al. 2016b). The alignments for each locus were trimmed to ensure that the sequences were of equal length, and all ambiguous sites were removed. The G3PDH, HSP60 and RPB2 gene alignments were combined to form a concatenated Botrytis dataset. A total of 2444 positions were used in the final dataset, and the evolutionary relationships were estimated using the Tamura–Nei model with gamma-distributed and invariant sites. Maximum likelihood (ML) analysis was used to construct a phylogenetic tree from this concatenated dataset, with the substitution models determined through the Bayesian information criterion (BIC). The ML tree for the G3PDH + HSP60 + RPB2 dataset was inferred using the Tamura–Nei model with gamma-distributed and invariant sites, which was executed in MEGA with 1000 bootstrap replicates.

Pathogenicity screening of Botrytis isolates

A total of 85 Botrytis isolates collected throughout the three sampling events were assessed for their rot-causing ability through a detached berry assay (Saito et al. 2016a). Conidial suspensions served as the inocula, which were prepared from cultures sporulating on 90-mm PDA plates that had been incubated at 26 °C for 7–10 days in the dark. After adding 2 ml of a spore suspension buffer containing 0.2% agar, 0.05% Tween 80 and 30% glycerol (Visagie et al. 2014), conidia were harvested by gently rubbing the flooded culture with a sterile glass rod and then collecting 1 ml of the resulting suspension. The spore suspensions were quantified using a combination of direct counting with a haemocytometer and spectrophotometry, similar to a technique described by Caligiore-Gei and Valdez (2015). To ensure an even amount of inoculum for each isolate, spore suspensions were diluted to approximately 1.0 × 105 conidia ml−1.

Mature but firm blueberries were picked from an orchard in Stellenbosch and processed the following day. Each berry was surface sterilised using the same protocol as described above and placed in a sterilised container. Each plastic container (330 × 265 × 75 mm) had 20 evenly spaced upside-down McCartney bottle lids fixed to the bottom with double-sided tape, serving as placeholders for single berries. Before the assay, all containers were sprayed with 70% ethanol and allowed to dry overnight under UV exposure in a laminar flow cabinet. Using a sterile hypodermic needle, a ± 3-mm deep wound was created near the equator of each berry. An aliquot of 1 µl per standardised isolate (approximately 100 conidia) was introduced directly into the blueberry wound using a micropipette. Each isolate was inoculated in triplicate (three berries per isolate). Most of the conidial suspensions formed a droplet on top of the area of inoculation, an observation noted previously by Fernández-Ortuño et al. (2014). In each container, two berries in diagonally opposite corners that were wounded but not inoculated served as negative controls. Once all the berries in a single container had been inoculated accordingly, 100–150-ml sterile dH2O was poured into the bottom of the container to maintain high relative humidity (Panebianco et al. 2015), and the lids were sealed. All berries were incubated at 22 °C for 5 days, whereafter the resulting rot was scored visually using a scale presented by Saito et al. (2016a). A score of 0 was equivalent to no visible symptoms, while a score of 1 = < 25, 2 = 25–50, 3 = 50–75 and 4 = > 75% of the blueberry surface covered with Botrytis mycelia. A rot severity score was calculated for each isolate using the average of the three replicates, which served as a proxy for pathogenicity.

Results

Isolate identities

A total of 85 Botrytis isolates were collected and analysed during this study. Of the collected isolates, all were obtained from visibly symptomatic material, with the majority recovered from diseased berries and flowers (Fig. 1). No Botrytis isolates were recovered from the internal tissues of surface sterilised asymptomatic blueberry plant material. Therefore, none of the Botrytis isolates explored in this study was considered to be endophytic.

Fig. 1
figure 1

Summary of the sources from which Botrytis isolates (n = 85) were collected during the survey of blueberry orchards in the Western Cape province of South Africa, including a diseased flowers; b diseased berries; c diseased leaves and d diseased buds

Full size image

Of the Botrytis isolates that were assessed phylogenetically, all 85 formed a monophyletic clade with the reference sequences of B. cinerea in the G3PDH + HSP60 + RPB2 tree (Fig. 2). When combined with the typical morphological features exhibited (Fig. 3), these isolates were identified as members of the species B. cinerea. Isolate C3, one of the two isolates that clustered differently, fell into a monophyletic clade with two B. pelargonii reference strains and one B. cinerea reference strain (Fig. 2).

Fig. 2
figure 2

Maximum likelihood tree of the combined G3PDH + HSP60 + RPB2 partial gene sequence dataset. The tree describes the phylogenetic relationship between 85 Botrytis isolates collected in this study and reference sequences of previously described Botrytis spp. obtained from GenBank, with Sclerotinia sclerotiorum and Monilinia fructigena as outgroups. The numerical branch labels represent the bootstrap percentages (n = 1000). Branches with less than 50% bootstrap support are not shown. The tree is scaled such that the branch lengths are proportional to the number of nucleotide substitutions per site. The red box shows an exploded portion of the tree to emphasise the clustering pattern of isolates C3 and C177 in context of the reference sequences. Isolates obtained in this study are presented in boldface. Isolates that were assigned to B. cinerea are highlighted in blue

Full size image
Fig. 3
figure 3

General morphological characters of Botrytis cinerea. A–C. Fruiting structures in a diseased berry A and flower B, C, D. Colony morphology on PDA. E Conidiophore structure under a stereo microscope (scale bar = 500 µm). F Conidiophore with conidia (scale bar = 50 µm)

Full size image

Pathogenicity screening

Of the 85 Botrytis isolates that were screened for their level of pathogenicity on detached berries, 48 isolates (56.5%) resulted in an average score of 2 or more, while 37 isolates (43.5%) resulted in an average score of less than 2 out of a possible 4. Accordingly, the overall mean score for all 255 infected berries (85 isolates in triplicate) was 2.11, and the modal class was a score of greater than or equal to 1 but less than 2 (Fig. 4). For the two extremes, eight isolates achieved a maximum average score of 4 (all three replicate berries had more than 75% of their surface covered with grey mould), while six isolates achieved a minimum average score of 0 (all three replicate berries exhibited no visible signs of decay after the allotted time). To confirm the viability of the conidia from the group of isolates that resulted in no visible rot, 1 µl of the same spore suspensions used to inoculate the berries was spotted on PDA in triplicate and kept at 26 °C for 7 days. After incubation, conidia in all spots had germinated, therefore confirming that the low severity score was due to an inherently lower pathogenicity of those isolates. All the control berries that were wounded but left uninoculated were asymptomatic after the incubation period, suggesting that the applied procedure was sufficiently sterile and that no conidia were transferred between neighbouring berries within each container.

Fig. 4
figure 4

Proportions of the average pathogenicity scores (i) obtained by 85 Botrytis isolates. After inoculated blueberries had been incubated for 5 days, the severity of the rot was scored using a scale of 0–4 (Saito et al. 2016a)

Full size image

Discussion

As with the previous studies, B. cinerea isolates were recovered from blueberries experiencing pre-harvest grey mould decay (Rivera et al. 2013). Additionally, isolates were obtained from other sources that are less frequently explored in fruit crops, such as diseased leaves and diseased undifferentiated buds (Fig. 1). Despite the origin of the tissue, all source material was visibly diseased, and the isolates were in an active necrotrophic or saprotrophic lifestyle. In contrast, no Botrytis isolates were recovered from the internal tissues of healthy material, suggesting that this pathogen did not persist as an endophyte in the southern highbush blueberries that were sampled. As of yet, no literature directly contradicts this finding, but it does contrast results obtained in work on other hosts. For example, Botrytis strains exhibiting an endophytic lifestyle have previously been isolated from Centaurea stoebe (Shipunov et al. 2008), Primula (Barnes and Shaw 2003) and lettuce (Sowley et al. 2010). While these examples propose that certain Botrytis spp. are capable of persisting as endophytes in a cryptic state, descriptions of this lifestyle are still largely incomplete (van Kan et al. 2014).

All but two of the Botrytis isolates collected in this study were identified as B. cinerea, a result inferred using phylogenetic analysis of the G3PDH + HSP60 + RPB2 dataset (Fig. 2). Consequently, the species diversity of the Botrytis population in this blueberry orchard seems to be low. This could be a direct result of the prominence of B. cinerea as a pathogen, which suggests that it readily dominates fungal populations. Nevertheless, the inherent inefficiencies of culture-dependent techniques (Rantsiou et al. 2005) could prevent accurate detection of true diversity. A recent study by Garfinkel et al. (2019) supports this possibility, where 16 different Botrytis spp. were isolated from a single crop (peonies). The results of the present study diverge, with B. perlagonii and an unidentified Botrytis sp. representing the only two non-B. cinerea isolates. The isolate that was identified as B. pelargonii, namely isolates C3, was assigned to this species based solely on its phylogenetic position (Fig. 2). However, the previous studies have shown that the interspecific variation between B. pelargonii and B. cinerea can be as little as only one unique polymorphism out of 2955 aligned base pairs, a feature that can result in very short internodes separating the two in phylogenetic trees (Staats et al. 2005). This makes it particularly difficult to differentiate between the two species using only the phylogenetic species concept. While some work has indicated that B. pelargonii persists as a cryptic member of a Botrytis species complex, others are doubtful of the existence of this species (Walker 2016). Irrespective of its classification, B. pelargonii has never been reported to occur on blueberries but instead has only been isolated from Pelargonium and Chinese ginseng (Lu et al. 2018). Interestingly, the B. cinerea strain that clustered with isolate C3 and the two B. pelargonii representatives (Fig. 2) was originally isolated from blueberries in California (Saito et al. 2016b). Given the phylogenetic position of isolate C3 and strain X793, both of these fungi may belong to B. pelargonii, making this study the first report of this species on blueberries. Taking this into account, isolate C3 was tentatively identified as B. pelargonii with the understanding that further morphological characterisation is required to confidently differentiate it from B. cinerea (Mirzaei et al. 2008).

A similar difficulty was observed for isolate C177, which was the only unidentified Botrytis isolate in this study. Even though it did not form a well-supported group with any of the reference sequences (Fig. 2), it fell into “Clade 1”, which was originally defined by Staats et al. (2005). This clade is currently considered to be composed of B. calthae, B. cinerea, B. fabae, B. pelargonii, B. pseudocinerea, B. sinoviticola and B. californica (Saito et al. 2016b). While novel species like B. eucalypti (Liu et al. 2016) have recently been included as members of Clade 1, it is difficult to infer putative new species based only on molecular information. Even though three housekeeping genes were used in combination (G3PDH + HSP60 + RPB2) in this study, these can still provide a limited resolution within this group of Botrytis spp. (Staats et al. 2005). Therefore, while the branching pattern of the Botrytis phylogenetic tree (Fig. 2) suggests that isolate C177 could potentially be novel, this is doubtful in the given context and still requires further characterisation.

The pathogenicity assay was used primarily to screen the Botrytis isolates for the level of severity at which they cause rot, and it provided sufficient resolution to identify the isolates that are potentially more problematic. Of the 85 isolates that were tested, the majority (56.5%) resulted in a severity score of 2 or higher out of a possible 4. This suggests that the population of B. cinerea in this blueberry orchard possesses the mechanisms to cause severe decay on asymptomatic, wounded berries. This infection assay (Saito et al. 2016a) is advantageous because it also offers the opportunity to confirm whether the isolates are causal agents of decay, thereby fulfilling Koch’s postulates. However, the interpretation of the results can present problems. The visual scoring system of 0–4 can be easily applied to distinguish decay on opposite ends of the spectrum (the difference between 0 and 4, for example), but becomes challenging when assessing closely related classes (the difference between 2 and 3, for example). Unfortunately, this introduces some degree of subjectivity, which can contribute to inconsistency. Ideally, a quantitative method should be employed as the primary metric, which can then be supplemented with qualitative scores based on visual and physical properties. For example, lesion diameter offers an absolute method of quantifying pathogenicity. This has been successfully implemented in other fruits, such as strawberries (Fernández-Ortuño et al. 2014) and table grapes (Panebianco et al. 2015), as well as other types of tissues, such as onion bulbs (Garfinkel et al. 2017) and peony leaves (Garfinkel et al. 2019). This was considered for the current study, but difficulties persisted when trying to implement this on blueberries. Due to the small size of the fruit, the inoculated Botrytis isolates quickly overgrew the berries, and their hyphae often covered the majority of the fruit surface before any measurable lesions developed. Furthermore, the lesions caused on blueberries did not have definite boundaries, which further added to the difficulty of accurately quantifying the severity of the decay. Therefore, the screening method employed in this study was preferable under the given circumstances.

The results of this study provide a glimpse into the fungal pathogen populations proliferating blueberry orchards in Stellenbosch in the Western Cape. Given the relatively incomplete body of the literature regarding the fungal microbiome of blueberries, particularly the endophytes, continued exploration holds great potential to generate novel information in this field. Overall, the work presented here serves as a foundation for the partial characterisation of Botrytis spp. occurring in the blueberry orchards studied and may prove useful in the management of these pathogens in this system.