Water Stress Enhances the Progression of Branch Dieback and Almond Decline under Field Conditions

Branch dieback and tree decline have been described as a common complex disease worldwide in woody crops, with Botryosphaeriaceae and Diaporthaceae being considered the most frequent fungi associated with the disease symptoms. Their behaviour is still uncertain, since they are considered endophytes becoming pathogenic in weakened hosts when stress conditions, such as water deficiency occur. Therefore, the main goal of this study was to determine if water stress enhances general decline on weakened almond trees subjected to different irrigation treatments under natural field conditions. In parallel, the occurrence of fungal species associated with almond decline was also determined in relation to disease progression by fungal isolation, and morphological and molecular based-methods. The symptoms of branch dieback and general decline were observed over time, mainly in the experimental plots subjected to high water deficiency. Botryosphaeriaceae were the most consistently isolated fungi, and Botryosphaeria dothidea was the most frequent. Collophorina hispanica was the second most frequent species and Diaporthe and Cytospora species were isolated in a low frequency. Most of them were recovered from both asymptomatic and symptomatic trees, with their consistency of isolation increasing with the disease severity. This work reveals the need to elucidate the role of biotic and abiotic factors which increase the rate of infection of fungal trunk pathogens, in order to generate important knowledge on their life cycle.


Introduction
Almond [Prunus dulcis (Mill.) D.A. Webb] represents the second highest woody crop by acreage in Spain after cultivated olive (Olea europaea subsp. europaea L.). To date, Spain currently leads the world in almond cultivation, with 657,768 hectares of cultivated almond (34% of the global surface) and 339.033 tonnes of production [1,2]. In this country, Andalusia region (southern Iberian Peninsula) represents the 30% of the Spanish cultivated almond surface [2].
Until recently, almond crop has been associated with traditional dry farming systems in marginal areas of southern Spain with unfavorable conditions to produce high yields. However, due to the global economic impact of almond kernels, as well as the necessity to find extensive alternative crops in Andalusia, almond plantings are increasingly being established in regions with better favorable conditions (i.e., moderate-warm temperatures, high humidity, irrigation-water resources, (1) the typical environmental conditions in southern Spain are characterized by scarce rains and warm temperatures during summer (from May to September), which predispose plants to water deficiency for a long time; and (2) the need to optimize water-irrigation treatments within the frame of eco-friendly agriculture towards a sustainable use of water resources. Therefore, the main goal of this study was to determine whether water stress enhances the general decline of weakened almond trees, subjected to different irrigation treatments, under natural field conditions. In parallel, the occurrence of fungal species, associated with branch dieback and almond decline, was also determined in relation to the disease progress by fungal isolation, and morphological and molecular based-methods for their identification.
Plants 2020, 9, x FOR PEER REVIEW 3 of 29 by scarce rains and warm temperatures during summer (from May to September), which predispose plants to water deficiency for a long time; and 2) the need to optimize water-irrigation treatments within the frame of eco-friendly agriculture towards a sustainable use of water resources. Therefore, the main goal of this study was to determine whether water stress enhances the general decline of weakened almond trees, subjected to different irrigation treatments, under natural field conditions. In parallel, the occurrence of fungal species, associated with branch dieback and almond decline, was also determined in relation to the disease progress by fungal isolation, and morphological and molecular based-methods for their identification.

Effect of Water Stress on Branch Dieback of Almond under Natural Field Conditions
At the beginning of the evaluation period (June 2018), the number of almond trees of each category among the 80 evaluated trees was as follow: Category 0 = 34 trees 0 (asymptomatic trees with 0% of affected surface by branch dieback), Category 1 = 30 trees (<25% of affected surface by branch dieback), Category 2 = 8 trees (25-50% of affected surface by branch dieback); Category 3 = 1 tree (51-75% of affected surface by branch dieback); Category 4 = 0 trees (76-90% of affected surface by branch dieback); and Category 5 = 7 trees (>90% affected surface by branch dieback or dead trees). In general, Disease Severity (DS) progresses in significantly higher values of relative area under the disease progress curve (RAUDPC; P = 0.0173) and final disease severity (P = 0.0012) when almond trees were subjected to T3 (Severe Regulated Deficit Irrigation; RAUDPC = 47.2 ± 8.7%; Final disease severity = 80.4 ± 3.6%), followed by T2 (Moderate Sustained Deficit Irrigation; RAUDPC = 31.3 ± 5.1%; Final disease severity = 71.3 ± 9.1%) and T1 (Moderate Regulated Deficit Irrigation; RAUDPC = 28.5 ± 4.5%; Final disease severity = 59.1 ± 6.5%) (Figures 1 and 2). Almond trees used as control (T0) showed the lowest values of RAUDPC (22.2 ± 1.5%), as well as the lowest values of final disease severity (42.2 ± 4.1%) (Figures 1 and 3). Control trees did not show internal wood discoloration.  conducted from June 2018 (next spring after first symptoms of branch dieback occur) to September 2019 (end of the experiment). For each disease parameter, columns represent the means of sixteen trees and vertical bars are the standard errors of the means. Columns with different capital or lowercase letters differ significantly for RAUDPC, or Final disease severity, respectively, according to Fisher's LSD test at P = 0.05.

Occurrence, Consistency and Frequency of Isolated Fungi
Fungal species, associated with branch dieback and almond decline, were isolated from all the categories of severity evaluated, with the exception from the trees belonging to the category 5 from which only saprophytes (i.e., Alternaria spp. Nees, Penicillium spp. Link, Sordaria spp. Ces. and De Not., etc.) were recovered. The consistency of isolation of each isolated fungi is shown in Table 1. In

Occurrence, Consistency and Frequency of Isolated Fungi
Fungal species, associated with branch dieback and almond decline, were isolated from all the categories of severity evaluated, with the exception from the trees belonging to the category 5 from which only saprophytes (i.e., Alternaria spp. Nees, Penicillium spp. Link, Sordaria spp. Ces. and De Not., etc.) were recovered. The consistency of isolation of each isolated fungi is shown in Table 1.  In general, the occurrence of fungal species and the consistency of their isolation increased with the  DS, with trees belonging to the categories 3 and 4 showing the highest number of fungal species, as well as the highest consistency of isolation. However, three fungal species were isolated from trees belonging to category 0, while only one fungal species was recovered from trees belonging to category 1. The consistency of isolation in these two categories was somewhat lower (≤5.4%) than those obtained from the remaining ones (up to 16.6%) (Figure 4). The total fungal biomass in the sampled trees per category was three fungal species in the trees belonging to Category 0, one fungal species in the trees belonging to Category 1, three fungal species in the trees belonging to the Category 2, and four fungal species in the trees belonging to the Category 3 and 4. Therefore, there was no linear correlation between the in planta abundance (biomass) of the studied fungal species and the severity of the category of severity (r = −0.6455; P = 0.2394).
with trees belonging to the categories 3 and 4 showing the highest number of fungal species, as well as the highest consistency of isolation. However, three fungal species were isolated from trees belonging to category 0, while only one fungal species was recovered from trees belonging to category 1. The consistency of isolation in these two categories was somewhat lower (≤5.4%) than those obtained from the remaining ones (up to 16.6%) (Figure 4). The total fungal biomass in the sampled trees per category was three fungal species in the trees belonging to Category 0, one fungal species in the trees belonging to Category 1, three fungal species in the trees belonging to the Category 2, and four fungal species in the trees belonging to the Category 3 and 4. Therefore, there was no linear correlation between the in planta abundance (biomass) of the studied fungal species and the severity of the category of severity (r = −0.6455; P = 0.2394).
Fungal species, belonging to Botryosphaeriaceae, were the most frequent and they also showed the highest consistency of isolation of the whole of the experiment (Table 1). Botryosphaeria dothidea was the most frequent species, since it was isolated from the 50.0% of the sampled trees, from trees belonging to the categories 2, 3 and 4. This was followed by Collophorina hispanica (Gramaje, Armengol and Damm) Damm and Crous, which was isolated from the 41.6% of the sampled trees, from trees belonging to the categories 0, 2 and 3. The species belonging to Cytospora Ehrenb. were also isolated from trees belonging to the categories 0 (Cytospora cedri Syd., P. Syd. & E.J. Butler), 3 (Cytospora sp. 2) and 4 (Cytospora sp. 1) with a low frequency (8.3%). The frequency of the remaining species was also low [Dia.     Fungal species, belonging to Botryosphaeriaceae, were the most frequent and they also showed the highest consistency of isolation of the whole of the experiment (Table 1). Botryosphaeria dothidea was the most frequent species, since it was isolated from the 50.0% of the sampled trees, from trees belonging to the categories 2, 3 and 4. This was followed by Collophorina hispanica (Gramaje, Armengol and Damm) Damm and Crous, which was isolated from the 41.6% of the sampled trees, from trees belonging to the categories 0, 2 and 3. The species belonging to Cytospora Ehrenb. were also isolated from trees belonging to the categories 0 (Cytospora cedri Syd., P. Syd. & E.J. Butler), 3 (Cytospora sp. 2) and 4 (Cytospora sp. 1) with a low frequency (8.3%). The frequency of the remaining species was also low [Dia.

Molecular Identification of Isolated Fungi
For all Datasets, the topology obtained by Maximum Parsimony (MP) was confirmed with those obtained by BI analysis. The model used in BI analysis, and the gene boundaries, the number of total characters (T), parsimony-informative characters (PI), parsimony-uninformative characters (PNI) and conserved sites (C) processed in each maximum parsimony analysis, as well as TL, consistency index (CI), retention index (RI), homoplasy index (HI) and rescaled consistency index (RC) values obtained from the one most parsimonies trees in each Dataset are shown in Table 2.

Discussion
Studying whether the effect of abiotic factors, such as water stress enhances the incidence and development of branch dieback and decline syndromes on weakened trees, is essential in improving our understanding of the endophytic behaviour of fungi associated with this complex disease. In fact, to date, the role of the fungal trunk pathogens, causing tree decline, is still uncertain, since their aggressiveness could vary markedly depending on abiotic (i.e., ecological, environmental and

Discussion
Studying whether the effect of abiotic factors, such as water stress enhances the incidence and development of branch dieback and decline syndromes on weakened trees, is essential in improving our understanding of the endophytic behaviour of fungi associated with this complex disease. In fact, to date, the role of the fungal trunk pathogens, causing tree decline, is still uncertain, since their aggressiveness could vary markedly depending on abiotic (i.e., ecological, environmental and agronomical aspects) and/or biotic (i.e., plant-pathogen interactions) factors. Consequently, several authors consider that most of the fungal trunk pathogens are secondary or opportunistic, causing damage when biotic or abiotic circumstances occur [11,19,24,25].
The environmental and agronomic conditions regarding the availability of irrigation-water resources in southern Spain could be a limiting factor, enhancing the development of branch dieback and decline syndromes on fruit and nut crops. Therefore, we have evaluated the effect of water stress enhancing the disease development on weakened almond trees under natural field conditions. The first symptoms of branch dieback occurred in late-summer autumn 2017 in an eight-year old experimental field, subjected to four different irrigation treatments, since 2013. The symptoms included branch dieback, canker formation, internal wood discoloration and general decline (Figure 2d-g) were observed mainly in the experimental plots subjected to high water-stressed conditions (T2, T3). In fact, the DS progress was significantly higher in almond trees subjected to T3 than in those subjected to T0 after two consecutive years of periodic evaluations. Our results are in accordance with those previously obtained by several authors under the control conditions, which showed that stem cankers, developed by B. dothidea or S. musiva on water-stressed plants of peach, or Populus, respectively, were higher than those developed on non-water-stressed plants [22,23]. On the other hand, almond trees under full irrigation (T0) also showed minimum levels of dieback symptoms. Although, no internal wood discoloration was observed, several fungal species, such as C. hispanica, Cy. cedri and N. mediterraneum were isolated from those trees. These results reinforce the hypothesis that these fungi could cause latent infections in asymptomatic or lesser-symptomatic trees. In parallel, the conclusions obtained in this study should be considered to discard the high levels of irrigation water as potential abiotic factor associated with the prevalence of the disease in the newly established almond growing regions in southern Spain, as we initially hypothesized in the introduction. To the best of our knowledge, this is the first approach, which has demonstrated the endophytic behaviour of fungal trunk pathogens on weakened trees, subjected to water stress under natural field conditions.
Concerning the occurrence of fungal species associated with branch dieback and almond decline, the following seven species belonging to four different families were identified: Botryosphaeriaceae: B. dothidea, N. mediterraneum, N. parvum and Neoscytalidum sp.; Diaporthaceae: Dia. neotheicola and Dia. rhusicola; Tympanidaceae: C. hispanica; and, Valsaceae: Cytospora cedri and Cytospora spp. Among them, B. dothidea, C. hispanica, Dia. neotheicola, N. mediterraneum and N. parvum have been previously described associated with branch dieback and decline on weakened almond trees in Spain [7,8,26]. Moreover, the pathogenicity of most of these species has been previously demonstrated in almond trees in Spain [5,8,9,26]. On the other hand, Dia. rhusicola and species belonging to Cytospora and Neoscytalidium genera, are associated with branch dieback and tree decline in other nut crops, such as English walnut or pistachio [6,[15][16][17], but to our knowledge, these species have not been previously reported in association with branch dieback and almond decline in Spain. However, their pathogenicity to almond should be demonstrated in the future to confirm they are canker pathogens of almond.
Botryosphaeriaceae were the most frequent isolated fungi and they also showed the highest consistency of isolation in the whole of the experiment, with B. dothidea being the species most frequently isolated. The differences in consistency of isolation of Botryosphaeriaceae fungi from weakened almonds can occur, depending on the scenario where the surveys are conducted, but in general, B. dothidea and Neofusicoccum species are usually the most frequent [9,25]. Likewise, according to the literature, our results also suggest that Botryosphaeriaceae spp. found on weakened almond trees are able to endanger the productivity and longevity of orchards in Spain, as well as in other countries [9].
Among Botryosphaeriaceae fungi, notice that B. dothidea has been reported worldwide causing canker diseases in a broad range of woody crops, including different Prunus spp. [6,7,27]. However, the role of this fungus as a trunk pathogen is still uncertain, given it has been reported as a latent pathogen of global importance for its endophytic behaviour in woody plant health [20]. In fact, studies conducted recently in southern Spain, which compared the pathogenicity of B. dothidea on inoculated detached and attached shoots of almond, the English walnut and pistachio, demonstrated that, in every case, the fungus is significantly higher aggressive on detached shoots than on attached [5,16,17]. It was confirmed that B. dothidea could remain latent in woody plants until trees become weakened as a consequence of different biotic and/or abiotic factors.
With respect to Diaporthaceae, Dia. neotheicola and Dia. rhusicola showed a low consistency of isolation and their occurrence was also low. These two species have been previously reported, associated with branch dieback and shoot blight of English walnut in California [28] and southern Spain [16], and the first one was also isolated from pistachio in southern Spain [17] and recently reported associated with twig cankers and shoot blight of almond in Spain [26]. Usually, Diaporthaceae species occur simultaneously with Botryosphaeriaceae in the same orchards, with Botryosphaeriaceae being always the most frequent [16,17,28]. In addition, studies conducted in California by Agustí-Brisach et al. [29] suggest that coinfections between Botryosphaeriaceae and Diaporthaceae species result in antagonistic interactions on infection and disease development on English walnut. But, Dia. amygdali, which is a common species associated with branch dieback and cankers of almond [5,7], was not found in this experiment.
It is interesting to note that, in this study, C. hispanica was the second most frequent species isolated from weakened almond trees after B. dothidea. Our results are in concordance with those found by Olmo et al. [8], who indicated that this slow-growing species is common in declined almond trees. However, it is usually excluded in the diagnosis process probably because its presence goes unnoticed, due to its slow growth [8].
Finally, Cytospora species were also isolated in low consistency from the trees of three different categories of severity. According to our results, Cytospora spp. have also been reported associated with canker diseases in weakened tree nuts (English walnut and pistachio) in southern Spain showing less frequency and aggressiveness, and often simultaneously with Botryosphaeriaceae and/or Diaporthaceae fungi [16,17]. However, several Cytospora spp. have been already reported in California as canker pathogens of several fruit and nut crops including Prunus spp. such as almond, apricot or peach [30].
The isolations made during this study suggest that most of these fungi can occur in both asymptomatic and symptomatic trees, but their frequency of isolation increases with increasing DS. However, the consistency of isolation was low for all the species in the whole of the experiment, showing the highest values in weakened almond trees, belonging to categories 3 and 4 (51 to 90% of final disease severity). This information reinforces the hypothesis that the fungal species associated with tree decline could have a major endophytic behaviour, and its aggressiveness is probably enhanced by abiotic factors, such as water stress on previously infected and weakened trees. In general, studies on the etiology of fungal trunk diseases describe a broad list of fungi associated with the disease, but do not usually considering the strict pathogenic behaviour of each. Nevertheless, this work reveals the need to go on elucidating the role of biotic and abiotic factors, enhancing the infection of fungal trunk pathogens and disease development on woody crops towards generating important knowledge on their life cycle. Therefore, focus the research on such relevant challenge will provide a better understanding of the biology of fungi associated with tree decline syndrome. It will build a strong foundation for developing effective management approaches against the disease, by taking into consideration the optimum water management.

Experimental Field, Irrigation Treatments and Experimental Design
The present study was conducted in a nine-to 10-years-old experimental field of almond cv. Guara grafted onto GF-677 rootstock (5.5 ha; 7 × 6 tree spacing; 238 trees/ha) belonging to the Andalusian Institute of Agricultural and Fisheries Research and Training (IFAPA in Spanish) Centre 'Alameda del Obispo' located in Córdoba (Andalusia region, Spain; 37.8ºN, 4.8ºW), whose soil was classified as a Typic Xerofluvent of sandy loam texture and exceeds 1. The experimental field used in this study was established in February 2009. Pruning for tree formation was done along the two first years, and then there never were pruning interventions. Control management strategies to prevent pest (Acetamiprid 20%; Deltametrine 2.5%) and diseases (Boscalid 26.7% + Pyraclostrobin 6.7%; Thiram 50%; Tebuconazole 50% + Tryfloxistrobin 25%) were done according to a treatment-calendar based on the weather conditions which could favour the typical almond pest and diseases of this area. Weeds were controlled by mowing and herbicide applications (Glyphosate 36%; Oxifluorphen 24%). Mineral fertilization was calculated and applied following the recommendations of the California Fertilization Guidelines for Almonds (https://apps1.cdfa.ca.gov/ FertilizerResearch/docs/Almonds.html). Therefore, this experimental field is representative of the edapho-climatic characteristics and the standard crop management of the new almond plantings in Andalusia [31].
The irrigation system was formed by two pressure compensating drip irrigation laterals, spaced 1 m from the tree rows, and all the trees were fully irrigated until the irrigation treatments began, as described below. From April 2013 to October 2019 (before and along this present study), the experimental field was subjected to four irrigation treatments: (i) Control (T0): The trees were irrigated to cover their full water requirements (ET), which was calculated using the relationship between ground cover (GC) and a transpiration coefficient, proposed by Espadafor et al. [32]. An additional 15% of that quantity was supplied to account for the evaporation from emitter wet surfaces under the trees using Bonachela et al. [33] model; (ii) Moderate Regulated Deficit Irrigation (T1): seasonal irrigation was 65% of T0, but the deficit was mainly concentrated during kernel filling stage, where almonds are less affected by water stress. Specifically, the irrigation supplied was: 70% of T0 in spring; 40% in kernel filling stage (from middle-July to harvest period in middle-August) and 100% in the postharvest period; (iii) Moderate Sustained Deficit Irrigation (T2): This treatment consisted of 65% of T0 steadily throughout the irrigation season. In total, a similar amount of irrigation water to that of the T1 was supplied; and (iv) Severe Regulated Deficit Irrigation (T3): This treatment received 30% of seasonal irrigation in relation to T0, following a similar water allocation strategy as in T1. Irrigation was 40% of T0 in spring and after harvest, and only 15% during the kernel-filling stage.
A randomized complete block design, with four replicated blocks, each consisting of four irrigation treatments, was used in this experiment. There were 16 trees per treatment plot, from which the four central trees were used for experimental measurements and the remaining 12 ones served as guard line. Therefore, the experiment included a total of 256 trees (4 blocks × 4 irrigation treatments × 16 trees per elementary plot), from which 80 trees were evaluated. Weather data were collected from an automated weather station located at 300 m apart from the orchard.

Disease Severity Assessment and Data Analysis
In the summer-autumn of 2017, the experimental trees subjected to the different irrigation treatments started to weaken, showing the first symptoms of branch dieback and general decline. Since then, the progress of this syndrome was monitored over time by periodic assessments from June 2018 to September 2019. Two assessments per year were conducted, with a total of four assessments. DS was assessed based on the estimation of the percentage of the affected surface of the tree canopy using a 0-5 rating scale. Each scale value was referred as 'category' of severity (six categories in total) for further purposes of this study (see Section 4.3. Sampling and fungal isolation). The values of this scale have a linear relationship with the percentage of affected tissues (leaves and shoots) in order to satisfy the homogeneity of variances and normality for suitable statistical analysis [34]. The equivalences between the values of the scale and the percentage of affected surface of the tree canopy are: 0 = 0%, 1 = < 25%, 2 = 25-50%, 3 = 51-75%, 4 = 75-90%, 5 = ≥ 90%. The DS was assessed in June and September of each year (four evaluations in total), before, and after, harvest, respectively, and all the blocks of the whole of the experiment were evaluated each time. The relative area under the disease progress curve (RAUDPC) was calculated by the trapezoidal integration method from the disease severity values over time [35].
The dependent variables 'final disease severity (%) and RAUDPC (%) were subjected to ANOVA to determine the differences in DS between irrigation treatments. Data were tested for normality and homogeneity of variances, and logarithmically transformed where necessary. Treatment means for the global analyses were compared using Fisher's protected LSD test [36]. All the data were analysed using Statistix 10 [37].

Sampling and Fungal Isolation
Two almond trees per each category of disease severity (12 trees in total) were selected to temporarily monitor the fungi isolated from affected tissues. From each tree, branches and shoots showing dieback and cankers were collected in each disease assessment time. Samples were kept at 4 • C until being processed in the laboratory.
For fungal isolation, the outer bark of affected wood samples was removed, and were subsequently washed under running tap water. Little wood pieces were collected from the margin of the affected area of symptomatic samples or randomly selected across the wood section, in the case of asymptomatic samples (Category 0). All the wood pieces were surface disinfected by dipping into a 10% (vol/vol) solution of commercial bleach (Cl at 50 g l −1 ) for 2 min. Subsequently, they were air dried on sterile filter paper and plated onto malt extract agar (MEA) [20 g of MEA (Merck KGaA, Darrmstadt, Germany), 20 g of agar (Rokoagar AF LAB, ROKO Industries, Llanera, Asturias, Spain; 1 l of sterile distilled water (SDW)] supplemented with 0.5 g l −1 of streptomycin sulphate (Sigma-Aldrich, St. Louis, MO, USA) (MEAS). From each category of severity, a total of 168 wood pieces, obtained from the margin of the affected tissues, were plated on Petri dishes for fungal isolation [attempts of isolation (wood pieces) onto Petri dishes: 2 trees per disease category × 3 Petri dishes per tree × 7 wood pieces per Petri dish × 4 sampling moments]. Petri dishes were incubated for 5 to 14 days at 25 • C in darkness and they were examined daily removing the isolation attempts (wood pieces) contaminated by saprophytes (i.e., Alternaria spp. Penicillium spp., etc.) by cutting the agar up to 1-cm-radio beyond the colony margin using a sterile scalpel. It was necessary to prevent the contamination of the whole agar surface of the Petri dishes before our target fungi developed onto MEAS, due to the low mycelial growth rate of some expected fungi for isolation.
When the colonies were large enough to be examined, hyphal tips from the margin of the fungal colonies were transferred to the potato dextrose agar (PDA; Difco Laboratories ® , Detroit) in order to obtain pure cultures. They were all incubated as previously described, and were grouped into four fungal groups (families) according to colony colour and mycelial growth development of each: Botryosphaeriaceae (light to dark grey, fast growing mycelium), Collophorina spp. (reddish to beige, very slow growing mycelia), Cytospora (beige to olive grey, middle-slow growing mycelium), and Diaporthe (beige-white, middle-fast growing mycelium). These preliminary morphological observations were helpful in selecting 30 representative isolates that were subsequently identified by molecular tools (Table 1). All the isolates were single-spored by a serial dilution method and they were registered and maintained at 4 • C in darkness (Fungal collection of the Department of Agronomy, University of Cordoba, Spain).

Assessment of Consistency and Frequency of Isolated Fungi
The consistency of isolation (%) of each isolate was calculated as the number of positive attempts of isolation (wood pieces) of a given fungus divided by the total attempts of isolation in the whole of the experiment [Consistency of isolation = (Nº. of positive wood pieces /168) × 100; where 168 is the total attempts of isolation (wood pieces) per category of disease severity in the whole of the experiment obtained as follow; 21 wood pieces per tree × 2 trees of each disease category × 2 years of evaluation × 2 sampling times per year]. The frequency of isolation (%) of each fungal species was estimated as the ratio between the number of trees from which each species was isolated and the total of sampled trees (12 trees). Additionally, the in planta abundance (biomass) of the studied fungal species and the severity of the dieback symptoms (category) was also compared by Pearson's linear correlation (n = 5; data from Category 5 was excluded since only saprophytes were isolates) using Statistix 10 [37].  (Table 1) previously grown on PDA were ground by means the FastPrep ® -24 grinder machine (MP Biomedicals, Santa Ana, CA, USA). Subsequently, genomic DNA was extracted using the E.Z.N.A ® Fungal DNA Kit (OMEGA BioTek, Norcross, GA, USA). A MaestroNano ® spectrophotometer (MaestroGen, Taiwan) was used to determine the concentration and purity of the extracted DNA.

PCR Analysis and Sequencing
The 5.8S nuclear ribosomal gene with two flanking internal transcribed spacers (ITS) was amplified for all the 30 isolated fungi. Subsequently, part of the beta-tubulin (TUB) gene, part of the translation elongation factor 1-alpha (EF) and/or a 200-bp intron of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were amplified for the different isolates according to the necessities to complete the further phylogenetic analysis. To this end, the protocols described in the literature for each family and genus were followed to identify our fugal isolates (Tables 1 and 3). The PCRs were performed in a total volume of 25 µl [20 ng of genomic DNA, 5 µl of 5× My Taq Reaction Buffer and 0.13 µl of My Taq DNA Polymerase (Bioline)]. Additionally, 0.4 or 0.2 µM (each) primer was added for the ITS; or for the TUB, EF, and GAPDH PCRs, respectively. A negative control was included in all PCRs using ultrapure water instead of DNA. Primer pairs and PCR cycling programs used to amplify each locus are shown in Table 3. Ultrapure water was used instead of DNA as negative control. A MyCycler™ Thermal Cycler (BIO-RAD) was used to conduct the PCRs.
Electrophoresis of the amplification products from PCR was conducted on a 1.5% (w/v) agarose gel stained with RedSafe TM (Intron Biotechnology). A 100-bp DNA molecular weight marker (Ladder-GTP, gTPbio) was used, and the agarose gel was visualized under UV. Finally, the PCR products were purified by means the MEGAquick-spin TM Total Fragment DNA Purification kit (INTRON Biotechnology). The resulting amplicons were sequenced in both directions [Central Service Support Research (SCAI) of the UCO (Spain)].

Phylogenetic Analysis
Consensus sequences from DNA sequences generated with forward and reverse primers were obtained with the SeqMan software (DNASTART Lasergen SeqMan ® v. 7.0.0, Madison, WI, USA). They were compiled into a single FASTA file format. Subsequently, they were BLAST searched in GenBank (http://www.ncbi.nlm.nih.gov/genbank/) to determine the close related species for each sequence.
Firstly, a neighbor-joining (NJ) analysis was performed individually for each locus. It was useful to determine whether the sequence datasets were congruent and combinable (data not shown). To this end, the maximum composite likelihood method with 2000 bootstrap replications was used. Genetic distances were calculated using the Kimura 2-parameter mode and tree topologies of 70% reciprocal bootstrap generated individually for each locus were compared visually. The data of different loci were combined into single concatenated datasets when no supported nodes were in conflict.
Independent phylogenetic analyses were conducted for the isolates of each fungal group (family), previously established according to their main morphological characteristics and Blast analysis. The combined alignment of the ITS and TUB loci was analysed in order to infer the phylogeny of isolates belonging to Botryosphaeriaceae (Dataset I-A). Additionally, a little phylogeny combining ITS and EF loci was also conducted into Botryosphaeriaceae group to confirm the identification of Neoscytalidium sp. Isolate (Dataset I-B). In the case of Diaporthaceae, the combined alignment of the EF, TUB and ITS loci was conducted (Dataset II). Isolates belonging to Tympanidaceae were identified by means the combined alignment of the ITS, EF and GAPDH loci (Dataset III). Finally, the combined alignment of the ITS and EF loci was performed to infer the phylogeny of the isolates belonging to Valsaceae (Dataset IV). For each multilocus alignment, data of the reference taxa (including outgroup) downloaded from GenBank and the number of the taxa included in this study are shown in Table 1, and Table 2, respectively.
The reference Genbank taxa were selected based on their high similarity with our query sequences using MegaBLAST [38] and they were added and aligned with our sequences by Clustal W. Maximum parsimony (MP) analyses were conducted using MEGA version 7.0 software [38], and they were performed by means the Tree-Bisection-Regrafting (TBR) algorithm with search level one. The initial trees were obtained by the random addition of sequences (10 replicates). The gaps and missing data were treated as complete deletions. A total of 1,000 bootstrap replications were done to ensure the robustness of the topology [39]. Tree length (TL), consistency index (CI), retention index (RI), homoplasy index (HI) and rescaled consistency index (RC) were calculated for each resulting MP tree.
Additionally, Markov chain Monte Carlo (MCMC) methods were used to perform Bayesian inference (BI) analyses by means the software MrBayes v.3.2.6 [40]. They were useful for estimating the posterior probability of trees. The best fit models of the evolution used for each gene partition were also determined by MEGA v. 7.0 [38]. Two analyses with four MCMC chains each were run simultaneously for 1 × 10 7 generations, starting from a random tree topology. The trees were sampled every 100 generations, and the "temperature" parameter was set to 0.2. The first 25% of the saved trees was discarded as the burn-in phase of the analysis. The sequences derived in this study were uploaded at GenBank (Table 1). Table 3. Primer pairs and PCR conditions used for the amplification of the genes included in this study.