Chestnut Brown Rot and Gnomoniopsis smithogilvyi: Characterization of the Causal Agent in Portugal

Sweet chestnut (Castanea sativa Miller) is a nutritious food with high social and economic impacts in Portugal. The fungus Gnomoniopsis smithogilvyi (syn. Gnomoniopsis castaneae) is the causal agent of chestnut brown rot, and is currently considered one of the major threats to the chestnut production chain worldwide. Considering the lack of knowledge on both the disease and the causal agent in Portugal, studies were conducted in an attempt to develop the necessary control strategies towards the mitigation of the disease in a timely way. Isolates of G. smithogilvyi were selected from three varieties of chestnut from the northeast of Portugal, and were characterized at the morphological, ecophysiological and molecular levels. Tests of pathogenicity and virulence were also developed. Gnomoniopsis smithogilvyi was confirmed as the causal agent of brown rot disease in Portuguese chestnut varieties, which showed high susceptibility. The fungus showed high adaptability to chestnut substrates. The Portuguese isolates of G. smithogilvyi are morphologically and genetically similar to those from other countries, even though some physiological variability was observed among them.


Introduction
The fruits of the European chestnut (Castanea sativa Miller) play an important role in human nutrition, especially for populations in the Mediterranean basin [1]. The chestnut also represents a relevant economic contribution to many European countries [2], such as Portugal, especially in the northeastern region of Trás-os-Montes, which is responsible for 88% of the Portuguese chestnut area and 82% of the production [3]. The region holds three chestnut protected denominations of origin (PDOs)-Terra Fria, Soutos da Lapa and Padrela. Among the 13 registered varieties in the region, three are of high productive and commercial significance-Longal, Judia and Martaínha [4,5]. Longal is highly distributed in the region and is considered the best variety for transformation, while Judia and Martaínha are preferably consumed fresh due to their higher caliber. Due to the nutritional characteristics of the sweet chestnut, combined with its high water content, the fruits are a substrate conducive to the development of fungi that cause post-harvest rot, affecting quality and, consequently, the commercial value of the chestnut.
Over the last decades, a sharp increase in the incidence of chestnuts presenting rot, with symptoms atypical to those of diseases that normally affect the fruits, has been de Montanha), Instituto Politécnico de Bragança, Portugal, and preserved at 4 • C until analysis (max. 48 h).
The isolates of G. smithogilvyi used in the present study were previously collected from local chestnuts, cultivar Longal (isolates G1 and G2) and Judia (isolate G3), from the production year 2018/2019, and identified molecularly as described in Rodrigues et al. [25]. Isolates G1, G2 and G3 were deposited in the culture collection "Micoteca of Universidade do Minho (MUM)", Braga, Portugal, with the codes MUM 20.139, MUM 20.140 and MUM 20.141, respectively. All isolates were grown on potato dextrose agar (PDA, BioLife, Milan, Italy) for 5 days at 25 • C before analysis.

Tests of Pathogenicity and Virulence of G. smithogilvyi Isolates 2.2.1. Preparation of Chestnuts
Chestnuts from the three varieties were carefully inspected for defects, such as discoloration, broken pericarp (shell), irregular tenderness, signs of pest infestation and fungal infection. Only apparently healthy nuts were selected. Nuts were washed with running tap water, surface-disinfected with 10% commercial bleach (0.5% sodium hypochloride) for 10 min, rinsed twice with sterilized water and blot-dried in sterile paper towels in a laminar flow hood. After that, the nuts were submitted to UV light for 20 min.

Preparation of Spore Suspensions
For this assay, the isolates G1 and G2 were used from 7-day-old PDA cultures grown at 25 • C. Conidiomata were collected from the plate using a sterile loop and homogenized by vortexing in sterile water with 0.1% Tween 80. Spore concentrations were determined using a Neubauer counting chamber, and were adjusted to 2 × 10 7 spores/mL in sterile water with 0.1% Tween 80.

Inoculation and Incubation
For each binomium fungal isolate and chestnut variety, three different methods of inoculation were tested. In all methods, chestnuts were placed in 250 mL glass jars (three chestnuts each), and 20 µL of the spore suspension was used as inoculum. The inoculation methods were as follows: (i) intact chestnuts were superficially inoculated (on the pericarp) in three spots-near the hilum, near the embryo and in the center; (ii) chestnuts were perforated with a sterile needle in three spots, and inoculated on the whole; (iii) chestnuts were longitudinally cut in half with a sterile scalpel, and the two halves of the endosperm were inoculated. For each treatment and each variety, six non-inoculated chestnuts were used as the control. In total, 48 jars and 144 chestnuts were used. Chestnuts were incubated in the dark at 25 • C for 15 days.

Evaluation of Infection Rate in Chestnuts Inoculated with G. smithogilvyi
Chestnuts were longitudinally cut and the fungal infection rate was registered in terms of incidence, and the severity of infection rate was registered in terms of percentage of chestnut area with visible rot as proposed by Donis-González et al. [26]: Level 0 (L0): no visible rot; Level 1 (L1): 1-25% of the chestnut area with rot; Level 2 (L2): 26-50%, Level 3 (L3): 51-75% and Level 4 (L4): 76-100%. Incidence was registered as the percentage of infected chestnuts. The severity was then calculated from these infection levels using the McKinney index [27]. This index takes into account the intensity of the rot (severity), its frequency and the maximum possible value, according to the equation where d is the category of rot intensity scored for chestnuts, f is the rot frequency, N is the total number of examined chestnuts (healthy and rotted) and D is the highest category of rot intensity that occurred.

Verification of Koch's Postulates
The Koch's postulates were verified in the inoculated chestnuts that showed the development of brown rot. The rot-causing agent was re-isolated from each infected fruit in PDA Petri dishes and was incubated for 7 days at 25 • C in the dark, after which it was morphologically identified and compared with the original cultures.

Natural Incidence of G. smithogilvyi in Chestnuts
This assay was performed to evaluate the endophytic incidence of the fungus. For this, 36 glass flasks and 108 chestnuts (36 of each variety) were used. Chestnuts were selected and surface-disinfected as previously described, and were divided into four sets (for each variety). They were longitudinally cut and incubated for 7 days, and the intact chestnuts were incubated for 7, 14 and 21 days. All treatment groups were incubated at 25 • C, and carried out in triplicate (total of 9 chestnuts per treatment for each chestnut variety). After the incubation period, the intact chestnuts were longitudinally cut and the fungal infection rate was registered in terms of percentage of chestnut area with visible rot [26], as previously described.

Biological Material and Culture Media
In this assay, the isolates G1, G2 and G3 were used. Growth parameters were evaluated in three media: potato dextrose agar (PDA, Biolife, Italy), malt extract agar (MEA, Himedia, Mumbai, India) and chestnut medium (MC). PDA and MEA were used as standard media for comparison purposes with previously reported data [9,11]. The media were prepared following the manufacturers' instructions. MC was used as a model medium to mimic the chemical and nutritional conditions of chestnuts to better understand the isolates' ecophysiology, in particular, the adaptability to chestnut as a substrate. For the preparation of MC, fresh and healthy chestnuts (balanced mix of the three varieties Judia, Longal and Martaínha) were cooked for 15 min in a microwave, shelled and blended using a kitchen blender in the proportion of 200 g per 1 L of distilled water. Agar was added at 2%, and the medium was autoclaved for 121 • C for 15 min.

Inoculation and Growth Parameters
Spore suspensions of each isolate were prepared at 3 × 10 7 spores/mL as previously described. Petri dishes with a 9 cm diameter containing 20 mL of each medium were central point-inoculated with 20 µL of each spore suspension (in triplicate), and were incubated at 5, 12, 20, 25, 30 and 35 • C for 9 days in the dark.
Fungal growth was measured in two directions daily for nine days or until the colony achieved the maximum diameter (8.5 cm). From these data, the following parameters were calculated: (i) lag phase (λ, in days), corresponding to the number of days from inoculation until mycelial growth was visible and (ii) mean growth rate (κ, day −1 ), calculated from the slope of the regression line in the exponential phase from the colony diameter plotted against incubation time. Spores were counted for each culture medium 9 days after inoculation, but only for the 25 • C incubation. For this, spores were collected from the entire plate by flooding the medium with 10 mL of 0.9% NaCl with 0.1% Tween 80, and were counted with a Neubauer hematocytometer.

Measurement of Conidia and Conidiomata
At least 100 conidia and 50 conidiomata of 7-day-old cultures of each culture medium were measured transversally and longitudinally using a microscope and a stereomicroscope, respectively, both coupled to Leica Application Suite V4 (LAS V4.12) software. Measures are expressed as (minimum)-mean-(maximum) (length × width).
Alignments were generated using ClustalX [32]. Phylogenetic trees were constructed using MEGA11 [33], using the neighbor-joining method [34] and the Jukes-Cantor parameters method [35] with 1000 bootstrap replicates. The evolutionary distances were computed using the maximum composite likelihood method [36]. Sequences of G. smithogilvyi isolates from different geographic regions obtained in previous studies were retrieved from GenBank and used for comparison.

Statistical Analysis
Since both the severity and incidence raw and transformed data failed the normality criteria, a non-parametric analysis was applied using the Kruskal-Wallis test followed by the post hoc Mann-Whitney test. The level of significance was established for p < 0.05. Tests were run with the IBM SPSS ® Statistics v. 22 software.
In the ecophysiology assays, the logarithm (Ln) of the diametral growth was used to determine the mean growth rate (κ) and the R 2 value was used to determine the experimental data fitting the model. All the growth rate values were compared to determine significant differences between conditions using GraphPad v. 8.0. Since the data failed the normality premise, a non-parametric analysis was applied using the Kruskal-Wallis test followed by multiple comparisons to evaluate the influence of culture media, isolates and temperature on the growth parameters lag phase, mean growth rate and sporulation.

Natural Incidence of G. smithogilvyi in Chestnuts
The results for the natural incidence and severity (as given by the McKinney index) of G. smithogilvyi in chestnuts are presented in Table 1. Given the high variability of incidence, no significant differences were observed between varieties (p = 0.721) or incubation conditions (p = 0.920).

Tests of Pathogenicity and Virulence of G. smithogilvyi Isolates
Overall, there was no significant difference in either incidence or severity between chestnut varieties (p = 0.596) and fungal strains (p = 0.774) in inoculated chestnuts (Table 2; Figure 1). Differences were found between treatments for both parameters (p < 0.0001). After the Mann-Whitney tests, no significant difference was detected in rot incidence and severity between cut and perforated fruits (p = 1.000), where the fungus had direct contact with the endosperm. In contrast, in intact fruits, the fungus caused significantly lower disease incidence and severity, resulting in significant differences between the treatments intact/cut and intact/perforated (p = 0.000). In the exposed cotyledons (cut chestnuts), the fruits were fully infected (100% of infected area) by the fungus, which was always re-isolated, following the Koch's postulates. Table 1. Natural incidence (%) and severity (McKinney index) of brown rot caused by G. smithogilvyi in chestnut varieties Longal, Judia and Martaínha (n = 9 ± SD).

Morphology and Ecophysiology of G. smithogilvyi
Under standard conditions, G. smithogilvyi produced a relatively dense and woolly mycelium, with a beige color, diffuse to regular margins, with concentric rings of yellowish conidiomata on the obverse and dark on the reverse of the colony ( Figure 2). In CM, the morphology of the fungus differed from that in PDA by being denser and more aerial, with fewer but bigger conidiomata showing an irregular distribution (no evident ring distribution). Conidiomata were abundant, globose to subglobose, viscous, with colors varying from cream to light orange, mostly yellow to light orange ( Figure 3).   Table 2. Incidence (%) and severity (McKinney index) of brown rot caused by G. smithogilvyi in chestnut varieties Longal, Judia and Martaínha artificially infected with two strains of the fungus, G1 and G2 (n = 6). Different letters mean significant differences between methods of inoculation), after Kruskal-Wallis non-parametric analysis followed by the post hoc Mann-Whitney test.

Morphology and Ecophysiology of G. smithogilvyi
Under standard conditions, G. smithogilvyi produced a relatively dense and woolly mycelium, with a beige color, diffuse to regular margins, with concentric rings of yellowish conidiomata on the obverse and dark on the reverse of the colony (Figure 2). In CM, the morphology of the fungus differed from that in PDA by being denser and more aerial, with fewer but bigger conidiomata showing an irregular distribution (no evident ring distribution). Conidiomata were abundant, globose to subglobose, viscous, with colors varying from cream to light orange, mostly yellow to light orange (Figure 3).

Morphology and Ecophysiology of G. smithogilvyi
Under standard conditions, G. smithogilvyi produced a relatively dense and woolly mycelium, with a beige color, diffuse to regular margins, with concentric rings of yellowish conidiomata on the obverse and dark on the reverse of the colony (Figure 2). In CM, the morphology of the fungus differed from that in PDA by being denser and more aerial, with fewer but bigger conidiomata showing an irregular distribution (no evident ring distribution). Conidiomata were abundant, globose to subglobose, viscous, with colors varying from cream to light orange, mostly yellow to light orange (Figure 3).   The conidia from G. smithogilvyi showed variable shapes and sizes, mostly straight or slightly curved, cylindrical or ellipsoidal ( Figure 3C,F,I). In general, the average conidia size for the three isolates did not show major differences between culture media. In contrast, the size of the conidiomata showed significant differences between both culture media and isolates (p < 0.05), despite the high size diversity observed (Table 3). On the other hand, the substrate significantly influenced the size of the conidiomata, which were significantly bigger in CM (average of the three strains 435 µm) than in the other tested media (209 µm in PDA; and 67 µm in MEA). Table 3. Conidiomata size (n = 50) and conidia length and width (n = 100) of G. smithogilvyi isolates G1, G2 and G3 in potato dextrose agar (PDA), malt extract agar (MEA) and chestnut medium (CM). The conidia from G. smithogilvyi showed variable shapes and sizes, mostly straight or slightly curved, cylindrical or ellipsoidal ( Figure 3C,F,I). In general, the average conidia size for the three isolates did not show major differences between culture media. In contrast, the size of the conidiomata showed significant differences between both culture media and isolates (p < 0.05), despite the high size diversity observed (Table 3). On the other hand, the substrate significantly influenced the size of the conidiomata, which were significantly bigger in CM (average of the three strains 435 µm) than in the other tested media (209 µm in PDA; and 67 µm in MEA). Table 3. Conidiomata size (n = 50) and conidia length and width (n = 100) of G. smithogilvyi isolates G1, G2 and G3 in potato dextrose agar (PDA), malt extract agar (MEA) and chestnut medium (CM). Min-Mean-Max (SD: standard deviation). The three isolates of G. smithogilvyi presented intraspecific differences in colony morphology and growth (Figures 4 and 5) when grown in the media PDA and CM at different temperatures (12,20,25,30 Figure 6 shows the growth rates of the three isolates in each of the three culture media for the temperature range 5 to 35 °C (MEA was not tested at 35 °C). Even though no significant differences were observed for each isolate between the temperatures 20, 25 and 30 °C (p > 0.05), optimal growth occurred between 25 and 30 °C with growth rates of 3.5 to 6 mm/day. The growth rate was similar between PDA and MEA (p = 1.000), but significantly higher in CM (p < 0.05) for all isolates.  Figure 6 shows the growth rates of the three isolates in each of the three culture media for the temperature range 5 to 35 • C (MEA was not tested at 35 • C). Even though no significant differences were observed for each isolate between the temperatures 20, 25 and 30 • C (p > 0.05), optimal growth occurred between 25 and 30 • C with growth rates of 3.5 to 6 mm/day. The growth rate was similar between PDA and MEA (p = 1.000), but significantly higher in CM (p < 0.05) for all isolates. The results for lag phase and conidia production are presented in Figure 7. The lag phase was not significantly different between the isolates for the same temperature, considering each culture medium (p > 0.05). Within the 20 to 30 °C range, the fungus started to grow immediately after inoculation (lag phase = 0 days), while at 5, 12 and 35 °C, the lag phase was significantly longer (p = 0.000). It is noteworthy that for these suboptimal temperatures, the lag phase was significantly lower in CM when compared with PDA and MEA. The results for lag phase and conidia production are presented in Figure 7. The lag phase was not significantly different between the isolates for the same temperature, considering each culture medium (p > 0.05). Within the 20 to 30 • C range, the fungus started to grow immediately after inoculation (lag phase = 0 days), while at 5, 12 and 35 • C, the lag phase was significantly longer (p = 0.000). It is noteworthy that for these suboptimal temperatures, the lag phase was significantly lower in CM when compared with PDA and MEA.

Parameter
In terms of conidia production, there was no significant difference between the culture media (p = 0.379), but temperature significantly influenced this parameter, ranging from 3.5 log10 spores/mL (MC, 12 • C) to 9 log10 spores/mL (MC, 30 • C). It is noted that the isolates produced the lowest amount of spores in CM for all temperatures except at 30 • C.
The results for lag phase and conidia production are presented in Figure 7. The lag phase was not significantly different between the isolates for the same temperature, considering each culture medium (p > 0.05). Within the 20 to 30 °C range, the fungus started to grow immediately after inoculation (lag phase = 0 days), while at 5, 12 and 35 °C, the lag phase was significantly longer (p = 0.000). It is noteworthy that for these suboptimal temperatures, the lag phase was significantly lower in CM when compared with PDA and MEA. In terms of conidia production, there was no significant difference between the culture media (p = 0.379), but temperature significantly influenced this parameter, ranging from 3.5 log10 spores/mL (MC, 12 °C) to 9 log10 spores/mL (MC, 30 °C). It is noted that the isolates produced the lowest amount of spores in CM for all temperatures except at 30 °C.

Molecular Characterization
The molecular analysis of the G. smithogilvyi isolates was performed on the basis of the ITS region and the TEF1-α gene (Figure 8). In this analysis, sequences of G. smithogilvyi isolates with different geographical origins were retrieved from GenBank (19 ITS and 13

Molecular Characterization
The molecular analysis of the G. smithogilvyi isolates was performed on the basis of the ITS region and the TEF1-α gene ( Figure 8). In this analysis, sequences of G. smithogilvyi isolates with different geographical origins were retrieved from GenBank (19 ITS and 13 TEF1-α sequences). Gnomoniopsis paraclavulata Sogonov was used as the tree root for both molecular markers. The phylogenetic analysis of the G. smithogilvyi isolates showed a very close relation between the isolates from this study and those retrieved from GenBank, including the type of strain. This is in agreement with other studies [15,16], where no differences were detected between isolates with worldwide geographic distribution. TEF1-α sequences). Gnomoniopsis paraclavulata Sogonov was used as the tree root for both molecular markers. The phylogenetic analysis of the G. smithogilvyi isolates showed a very close relation between the isolates from this study and those retrieved from GenBank, including the type of strain. This is in agreement with other studies [15,16], where no differences were detected between isolates with worldwide geographic distribution. The optimal tree is shown, as inferred using the neighbor-joining method [34]. The tree is drawn to scale, with branch lengths (above the branches) in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances, computed using the maximum composite likelihood method [36], are in the units of the number of base substitutions per site. All ambiguous positions were removed for each sequence pair (pairwise deletion option).

Discussion
In recent years, many studies have been conducted on the fungal pathogen G. smithogilvyi, mostly from countries where the chestnut is a significant commercial product [37][38][39][40]. In Portugal, Coelho and Gouveia [23] described G. smithogilvyi as the causal agent of brown chestnut rot for the first time, and it has been isolated from Portuguese chestnuts at frequencies ranging from 9.3% to 15.3% [23,25,41], but no studies have been previously reported on the morphology and ecophysiology of the isolates from Portugal and their The optimal tree is shown, as inferred using the neighbor-joining method [34]. The tree is drawn to scale, with branch lengths (above the branches) in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances, computed using the maximum composite likelihood method [36], are in the units of the number of base substitutions per site. All ambiguous positions were removed for each sequence pair (pairwise deletion option).

Discussion
In recent years, many studies have been conducted on the fungal pathogen G. smithogilvyi, mostly from countries where the chestnut is a significant commercial product [37][38][39][40]. In Portugal, Coelho and Gouveia [23] described G. smithogilvyi as the causal agent of brown chestnut rot for the first time, and it has been isolated from Portuguese chestnuts at frequencies ranging from 9.3% to 15.3% [23,25,41], but no studies have been previously reported on the morphology and ecophysiology of the isolates from Portugal and their comparison to the ones isolated from other geographical areas.
In the present study, G. smithogilvyi was isolated from asymptomatic chestnuts from three C. sativa Portuguese varieties-Judia, Longal and Martaínha confirming its endophytic behavior, as also reported by others from various geographic origins [9,12,15,16]. No significant differences were observed for incidence between the three varieties ( Table 1). The Koch postulates confirmed G. smithogilvyi to be the causal agent of chestnut brown rot. Pathogenicity and virulence assays revealed that all of the tested G. smithogilvyi isolates were capable of colonizing the chestnut tissue, showing high virulence and causing intense symptoms of brown rot in all chestnut varieties ( Figure 1). These assays also revealed the role of the outer shell in controlling the development of the disease, mostly in the case where infection occurs from external propagules. Even though the fungus shows endophytic behavior, infections may also occur after harvest if the shell is ruptured. These assays demonstrate the importance of chestnut management practices at both pre-and post-harvest stages of chestnut production, as suggested by Silva-Campos et al. [39].
Our study provided detailed data on the main morphological and ecophysiological characteristics of three isolates of G. smithogilvyi. The colony showed a yellowish tone and the presence of numerous yellow to orange conidiomata (Figures 2 and 3), typical of the genus Gnomoniopsis [42]. The Portuguese isolates were similar to those from Australia, Italy, New Zealand and India [9,11,13], in terms of colony aspect as well as conidia and conidiomata size and morphology when grown in PDA. In contrast, in MEA, the conidiomata of the isolates were significantly smaller than those from Australia [11]. Additionally, the average growth rate at 25 • C in MEA was higher (approximately 4 mm/day) than that reported by Visentin et al. [9] (0.8 to 1.0 mm/day). This difference may be related to the method of calculation, since in our study, the growth rate considered only the logarithmic phase.
From a molecular perspective, our isolates showed a very high degree of genetic similarity (>99%) for the ITS region to the isolates from Australia, Chile, Slovenia, Spain, the USA, France, Italy, New Zealand, the United Kingdom and Switzerland [9,[17][18][19]43,44]; (Figure 8). The TEF1-α sequence showed a higher diversity among isolates, with the Portuguese ones closer to those from Chile and Australia. Silva-Campos et al. [40] developed and validated a multiplex PCR assay for G. smithogilvyi detection, and observed a very low level of genetic divergence between Australian isolates and those from other geographic origins.
In a study by Lione et al. [14], the authors reported that the production of conidia and other growth parameters by G. smithogilvyi are dependent on various factors such as the biological features of the isolate and the climatic and environmental conditions. In fact, in the present study, the ecophysiological studies were conducted in three different culture media, PDA, MEA and CM. While PDA and MEA were the media used for fungal characterization by others [9,11,42], a chestnut-based medium mimicking the substrate of interest had not been tested previously. By comparing growth parameters in CM with those in PDA and MEA, it was clear that G. smithogilvyi shows an optimal adaptation to chestnut, namely lower lag phase, higher growth rate and higher mycelial growth (lower conidia production). It is also interesting to note that growing in MC at 30 • C stimulates the production of conidia, which might reflect an adaptation to stressful temperatures. This effect may have a negative impact on the virulence of the fungus, which could potentially increase under the conditions of climate change already in place.