A Survey of Trunk Disease Pathogens within Citrus Trees in Iran

Citrus trees with cankers and dieback symptoms were observed in Bushehr (Bushehr province, Iran). Isolations were made from diseased cankers and branches. Recovered fungal isolates were identified using cultural and morphological characteristics, as well as comparisons of DNA sequence data of the nuclear ribosomal DNA-internal transcribed spacer region, translation elongation factor 1α, β-tubulin, and actin gene regions. Dothiorella viticola, Lasiodiplodia theobromae, Neoscytalidium hyalinum, Phaeoacremonium (P.) parasiticum, P. italicum, P. iranianum, P. rubrigenum, P. minimum, P. croatiense, P. fraxinopensylvanicum, Phaeoacremonium sp., Cadophora luteo-olivacea, Biscogniauxia (B.) mediterranea, Colletotrichum gloeosporioides, C. boninense, Peyronellaea (Pa.) pinodella, Stilbocrea (S.) walteri, and several isolates of Phoma, Pestalotiopsis, and Fusarium species were obtained from diseased trees. The pathogenicity tests were conducted by artificial inoculation of excised shoots of healthy acid lime trees (Citrus aurantifolia) under controlled conditions. Lasiodiplodia theobromae was the most virulent and caused the longest lesions within 40 days of inoculation. According to literature reviews, this is the first report of L. theobromae and N. hyalinum on citrus in Iran. Additionally, we report several Phaeoacremonium species, S. walteri, Pa. pinodella and C. luteo-olivacea on citrus trees for the first time in the world.

Fungal trunk diseases have been studied in detail in grapevine, which are the main biotic factor limiting vineyard productivity and longevity [2]. However, recent findings of high incidence in stone and pome fruits, small fruits, nut crops, citrus, and olive worldwide highlight the need for a focus on this novel group of hosts [3]. Trunk diseases are caused by a broad range of taxonomically unrelated fungi that primarily infect wood hosts through winter pruning wounds, thus colonizing the vascular tissues. Members of the families Botryosphaeriaceae, Togninaceae, Diatrypaceae, Diaporthaceae, as well as several basidiomycetes are included in this group of fungi. Members of Diatrypaceae (G) Co-occurrence of brown wood streaking (black arrow), wedge-shaped necrosis (white arrow) and irregular wood necrosis (red arrow) on C. sinensis; (H) Arch-shaped necrosis on C. aurantifolia; (I) a young wedge shaped necrosis on Citrus reticulata; (J) Co-occurrence of wedge-shaped necrosis is indicated by the white arrow and black spots are indicated by the black arrow on the C. sinensis.

Fungal Isolation and Morphological Identification
In this survey, 326 fungal isolates were collected from citrus trees (Table 1). According to colony appearance, culture characteristics, and microscopic structures, the main fungal isolates were classified as Phaeoacremonium spp., Botryosphaeriaceae spp. Cadophora sp., Colletotrichum spp., Peyronellaea sp., Phoma spp., and Biscogniauxia mediterranea. Thirty-nine isolates (11.96% of total isolates) were identified as Phaeoacremonium species and characterized by beige to medium brown flat slow-growing cultures on potato dextrose agar (PDA; Merck, Darmstadt, Germany) and on malt extract agar (MEA; 2% malt extract, Merck, Darmstadt, Germany). Septate hyphae were single or fasciculate, and three types of phialides, variable in shape and size (I, II, and III types), were recorded in these isolates [41]. Morphological features of 49 isolates (15.03%) were consistent with the description of species of Botryosphaeriaceae [16,17,42]. These isolates were characterized by dark green to gray or fast-growing gray mycelium on the PDA. All isolates produced fruit bodies, pycnidia, on pine needles within 15-35 days. Conidia were pigmented or hyaline. These isolates belonged to the genera Lasiodiplodia, Neoscytalidium, and Dothiorella. Twelve isolates of the phialides fungus were identified as Cadophora sp. These isolates formed flat, felty, and black-olivaceous and white to gray colonies on PDA, and their conidia were ellipsoid or elongate. Cultural and morphological characteristics observed were similar with the description of the Cadophora spp. [43,44]. Based on morphological characteristics, the remaining isolates were classified to Colletotrichum, Peyronellea, Pestalotiopsis, Fusarium, Microsphaeropsis, Alternaria, Trichoderma, Paecilomyces, Aspergillus, Penicillium, Phoma, Biscogniauxia, and Stilbocrea genera.  3  7  3  10  0  2  25  Total fungal isolates  80  94  40  66  19  27  326  Total number of trees  surveyed  39  57  19  43  18  12  188 Plants 2020, 9, 754 5 of 20 No association was found between wood symptoms and fungal species. Dual infections by trunk disease fungi in a single tree occurred. Phaeoacremonium parasiticum and P. italicum were isolated from one tree of C. aurantifolia; P. parasiticum, P. croatiense, and Do. viticola from one tree of C. sinensis, L. theobromae and Neoscytalidium hyalinum from one tree of C. aurantifolia, and C. luteo-olivacea and P. croatiense from one tree of C. limetta. In addition, some fungal species grew from an individual wood segment, such as Stilbocrea walteri and P. fraxinopennsylvanicum from C. limon.
Datasets of the BT and actin (ACT) alignments of Phaeoacremonium were congruent and could be combined (p = 0.225). The Hasegawa-Kishino-Yano model (HKY) with gamma distributed with invariant sites rates (G+I) was identified as the BIC best-fit nucleotide substitution model by the jModelTest for the Phaeoacremonium multi-locus analysis. Maximum likelihood (ML) of the combined ACT-BT regions provided a phylogeny with 98 to 100% ML bootstrap support for all species-level clades, with the exception of P. alvesii (paraphyletic, 87% bootstrap support), P. griseorubrum (paraphyletic, 66% bootstrap support), P. roseum (89% bootstrap support), and P. viticola (paraphyletic with regard to P. roseum and P. angustius) ( Figure 2). The 39 strains from Iran clustered in eight clades (P. italicum, Phaeoacremonium sp., P. rubrigenum, P. parasiticum, P. minimum, P. iranianum, P. fraxynopennsylvanicum, and P. croatiense). The isolates of the clade 2 grouped together in a polyphyletic clade with 100% bootstrap support with the P. italicum as a closely related species. The BT and ACT sequences of the second clade of Phaeoacremonium isolates were 98% (BT) and 98.77% (ACT) identical to those of P. italicum CBS 137763 (GenBank KJ534074, KJ534046). Three nucleotides varied in the ACT region and ten nucleotides in the BT region between the second clade of Phaeoacremonium isolates and the P. italicum CBS 137763 sequences.
Plants 2020, 9, 754 6 of 20 a polyphyletic clade with 100% bootstrap support with the P. italicum as a closely related species. The BT and ACT sequences of the second clade of Phaeoacremonium isolates were 98% (BT) and 98.77% (ACT) identical to those of P. italicum CBS 137763 (GenBank KJ534074, KJ534046). Three nucleotides varied in the ACT region and ten nucleotides in the BT region between the second clade of Phaeoacremonium isolates and the P. italicum CBS 137763 sequences.

Pathogenicity Test
Mean lengths of wood discolorations caused by inoculated isolates obtained from Citrus species on the detached shoots of C. aurantifolia are shown in Figure 3 and   shoots followed by Do. viticola (38.17 mm) and P. parasiticum (34.33 mm) (Figure 4a). In contrast, two species of S. walteri (6.33 mm) and P. rubrigenum (6.00 mm) produced the smallest wood lesions on the inoculated shoots, and no significant differences were observed between these species and the control treatments (3.67 mm).
Plants 2020, 9, x FOR PEER REVIEW the point of inoculation and re-isolation frequencies of inoculated isolates on lime shoots. L. theobromae was the most aggressive fungal species and produced the longest necrotic lesions (57.67 mm) on the inoculated shoots followed by Do. viticola (38.17 mm) and P. parasiticum (34.33 mm) ( Figure 4a). In contrast, two species of S. walteri (6.33 mm) and P. rubrigenum (6.00 mm) produced the smallest wood lesions on the inoculated shoots, and no significant differences were observed between these species and the control treatments (3.67 mm). All inoculated fungi caused longer basipetal than acropetal lesions on the lime shoots ( Figure  4b). Of the isolates inoculated, 10 species caused downward and upward wood lesions that were significantly different to those in the control (p < 0.05). L. theobromae also produced the longest wood lesion lengths both in upward (22.34 mm) and in downward (35.33) directions, while S. walteri (upward = 2.5, downward = 3.83 mm) and P. rubrigenum (upward = 2.5, downward = 3.50 mm) did not cause any significant necrotic lesion lengths both in the downward and in the upward directions compared to the control treatments (upward = 1.34, downward = 2.33 mm) on the inoculated shoots. Re-isolation percentages were between 40.0% (C. luteo-olivacea) and 100% (L. theobromae and N. hyalinum) on the inoculated lime shoots, and no fungal isolates were recovered from control treatments. All inoculated fungi caused longer basipetal than acropetal lesions on the lime shoots ( Figure 4b).
Of the isolates inoculated, 10 species caused downward and upward wood lesions that were significantly different to those in the control (p < 0.05). L. theobromae also produced the longest wood lesion lengths both in upward (22.34 mm) and in downward (35.33) directions, while S. walteri (upward = 2.5, downward = 3.83 mm) and P. rubrigenum (upward = 2.5, downward = 3.50 mm) did not cause any significant necrotic lesion lengths both in the downward and in the upward directions compared to the control treatments (upward = 1.34, downward = 2.33 mm) on the inoculated shoots. Re-isolation percentages were between 40.0% (C. luteo-olivacea) and 100% (L. theobromae and N. hyalinum) on the inoculated lime shoots, and no fungal isolates were recovered from control treatments. Plants 2020, 9, x FOR PEER REVIEW

Discussion
This study shows the high incidence and severity of fungal trunk pathogens associated with wood decay symptoms of six Citrus species (C. sinensis, C. aurantifolia, C. reticulate, C. limetta, C. aurntium, and C. limon) in Iran. During the last decade, extensive studies have been done on fungal trunk pathogens of fruit trees, including grapevine [36,37], stone [38], and pome fruit trees [28,30], pistachio, [45] almond [46][47][48], walnut [49,50], pomegranate, and fig trees [51] in Iran. The current study shows that Citrus also represents a rich catch host for fungi associated with trunk diseases in this country. Different trunk disease fungi often co-occurred in the same tree and even in the same type of symptom, thus showing the complexity of the etiology of wood symptoms observed. The coinfection of several trunk disease fungi on woody crops could lead to an increase in disease severity

Discussion
This study shows the high incidence and severity of fungal trunk pathogens associated with wood decay symptoms of six Citrus species (C. sinensis, C. aurantifolia, C. reticulate, C. limetta, C. aurntium, and C. limon) in Iran. During the last decade, extensive studies have been done on fungal trunk pathogens of fruit trees, including grapevine [36,37], stone [38], and pome fruit trees [28,30], pistachio [45], almond [46][47][48], walnut [49,50], pomegranate, and fig trees [51] in Iran. The current study shows that Citrus also represents a rich catch host for fungi associated with trunk diseases in this country. Different trunk disease fungi often co-occurred in the same tree and even in the same type of symptom, thus showing the complexity of the etiology of wood symptoms observed. The co-infection of several trunk disease fungi on woody crops could lead to an increase in disease severity compared to the single occurrence of a fungal species, as it has been previously demonstrated on grapevine with Botryosphaeriaceae and Ilyonectria spp. [52].
Three species of Botryosphaeriaceae, namely N. hyalinum, Do. viticola and L. theobromae were obtained from citrus trees in this study. Neoscytalidium hyalinum was isolated from C. aurantifolia and C. limetta, Do. viticola was recovered from C. sinensis, C. aurantifolia, and C. aurantium, and L. theobromae was associated with C. sinensis and C. aurantifolia. Several species of Botryosphaeriaceae are known to dieback and branch cankers in Citrus spp. worldwide [14,20,22,23,[66][67][68][69][70]. Dothiorella viticola has been previously reported to cause gummosis in citrus in California [20] and Tunisia [71]. This fungus has also been reported from cultivar Parent Washington on sour orange rootstock [68], C. sinensis and C. latifolia Tan. in California [20], and C. sinensis in New Zealand [72]. Abdollahzadeh et al. reported this species from Citrus sp. in Guilan province of Iran [73]. Our study provides the first report of this fungus from C. aurantifolia and C. aurantium.
Neoscytalidium hyalinum has been reported as the most prevalent Botryosphaeriaceae species associated with citrus branch cankers in the desert regions of southern California [14]. This fungus has been recovered from C. paradise showing gummosis in California [20] and also from C. sinensis in Italy [74]. Therefore, our work is the first report of N. hyalinum from two Citrus species, C. aurantifolia and C. limetta. L. theobromae has been previously reported from some Citrus species, including C. limon in Chile [23] and Persian lime (Citrus latifolia) trees in Mexico [70]. Our study represents the first report of this species on C. sinensis and C. aurantifolia.
In the current study, 12 isolates of Cadophora luteo-olivacea were obtained from C. reticulata and C. limetta. C. luteo-olivacea has previously been reported with black vascular streaking and a decline in the symptoms characteristic of Petri disease on grapevine [44,54,75,76], bark cracks of kiwifruit [62], and from pear fruits showing dark-brown and slightly sunken spots [77]. Aside from these reports, little is known regarding the role of Cadophora species involved in trunk diseases of trees. This is the first time that C. luteo-olivacea has been found on Citrus spp.
In the current study, eight isolates of Stilbocrea walteri were isolated from C. aurantifolia, C. aurntium, and C. limon. This species was originally reported from dead corticated branches of Quercus ilex in Portugal [91], and to our knowledge, it has not been reported from necrotic wood tissues of trees. Therefore, this study is the first report of this species in Iran and on Citrus species worldwide.
Peyronellea pinodella (Didymellaceae) is a destructive necrotrophic pathogen on some plant families, including Fabaceae, Amaranthaceae, Asteraceae, Amaryllidaceae, Appiaceae Rubiaceae, Malvaceae, Poaceae, and Polemoniaceae [92]. To date, there is no report on the occurrence of P. pinodella on Citrus species and this is the first data on the occurrence of this species on C. sinensis and C. aurantifolia.
Two species of Colletotrichum were found to be associated with trunk diseases of citrus trees in this work, C. gleoesporioides on C. sinensis and C. limetta and C. boninense on C. limetta. Several species of Colletotrichum are associated with fruit and leaf anthracnose diseases of Citrus species; however, other diseases such as twig and shoot dieback caused by Colletotrichum spp. have been documented on citrus trees [14,93]. Colletotrichum gloeosporioides has been reported from a wide range of fruit trees such as strawberry, olive, almond, mango, apple, avocado, and citrus [94]. This fungus was found to be associated with twig dieback of lemon trees in Portugal [93]. C. boninense has been associated with fruit and leaf anthracnose on citrus trees [95,96]. According to a recent study, some Colletotrichum species have been isolated and reported from stems of citrus trees in Iran. These included C. karstii from C. aurantifolia and C. sinensis and four species, C. gloeosporioides, C. novae-zelandiae, C. siamense, and C. fructicola from C. sinensis [97]. Therefore, our study represents the first report of C. gleoesporioides and C. boninense from branches of C. limetta.
In our work, six isolates of M. olivacea were obtained from sweet orange. This fungus has been reported from various plant species worldwide. This taxon has previously been isolated as an endophytic species from P. persica [98], from xylem and stems of Pinus sylvestris [99] and Chilean gymnosperms [100]. Carlucci et al. isolated this species from internal wood discoloration of olive trees in Italy [101]. Microsphaeropsis olivacea has also been isolated and reported from some woody plants, such as Prunus cerasus, P. avium [102], and Persian oak (Quercus brantii) [103] in Iran. To our knowledge, this is the first report of M. olivacea on citrus trees. Several isolates of Fusarium, Pestalotiopsis, Phoma, Penicillium, Aspergillus, Trichoderma, and Alternaria species were also obtained from Citrus species in this study. Therefore, more studies are needed on these taxa in order to elucidate their potential impact on citrus trunk diseases.
Pathogenicity of selected fungal species in detached shoots of lime tree were confirmed in the current study. Results revealed that L. theobromae was more virulent on lime shoots than other species. In contrast to our results, Bautista-Cruz et al. reported that L. theobromae was the least virulent species when inoculated in Persian lime branches [70]. Several factors differed from the study carried out by Bautista-Cruz et al. and might have contributed to the discrepancy between the experiments, including the type of planting material inoculated, the environmental conditions for disease development, the time for virulence assessment, and the fungal strain used in the pathogenicity test. L. theobromae has been considered the most aggressive species on Eucalyptus [104,105], grapevine [42,106], and pistachio trees [107]. Lasiodiplodia theobromae was considered an important pathogen on greengage, sour cherry, peach, apricot, cherry [38], and willow trees [29] in Iran. Our study improved the knowledge on the occurrence of fungal trunk pathogens on Citrus species showing a decline in symptoms. Further investigations are needed throughout the citrus orchards to determine the potential impact of these fungi on citrus decline.

Tree Sampling and Fungal Isolation
During 2014 and 2015, several field surveys were performed in important citrus-producing regions of Bushehr province, Tallhe and Tang Eram. This province is located in the south of Iran, within 28.7621 • N latitude and 51.5150 • E longitude. Symptomatic wood samples were collected from various species of citrus trees including, acid lime, sweet orange, mandarin, sour orange, sweet lemon (C. limetta), and lemon showing yellowing, defoliation, canker, dieback, and gummosis. In total, 325 wood samples were collected from branches of 106 symptomatic trees (15-to 35-year-old) in 27 orchards. A map with the point locations of the sampled orchards is shown in Figure 5. Collected samples were brought to the laboratory and inspected for internal wood lesions and fungal isolation. Small fragments (4 × 4 mm) of symptomatic wood tissues were cut from the edges of wood lesions, surface-sterilized in sodium hypochlorite solution (1.5%) for 60 s, and rinsed three times in sterilized water. Wood chips were dried in sterilized filter paper and placed on PDA amended with 90 to 100 mg/L streptomycin sulfate (PDAS). For each branch sampled, three to five Petri dishes were obtained. All Petri dishes were incubated at 25 • C until fungal colonies were observed. Pure cultures of the fungal isolates were obtained by hyphal-tipping or transferring single conidia to fresh PDA.

Morphological Identification
All fungal isolates were identified initially to the genus level based on colony morphology and main microscopic structures using published articles and descriptions. Botryosphaeriaceae isolates were identified based on colony appearance and conidial morphology [16,108]. To induce sporulation, three to five mycelial plugs from each isolate were placed on 2% water agar (WA; Biokar-Diagnostics) plates amended with sterilized pine needles and incubated at 25 • C under near-ultraviolet light for 15-45 days [42]. Conidial characteristics (size, shape, color, and presence or absence of septa) were recorded for all isolates. Phaeoacremonium isolates were grouped based on colony appearance, pigment production on MEA, PDA and oatmeal agar (OA; 60 g oatmeal; 12.5 g agar; Difco, France) and the main microscopic structures (phialide shape and type, conidiophore morphology, size of hyphal warts, and conidial shape and size) [41,57,109]. Identification of Cadophora isolates was based on the colony and micro-morphological structures, such as conidiogenous cell size and shape, and conidia. The remaining fungal isolates were identified based on available identification keys and published papers [91,[110][111][112][113][114].

DNA Extraction, Amplification, and Sequencing
Identities of representative isolates were confirmed using molecular data. Fungi selected for molecular studies were grown on PDA for 10 to 15 days at 25 • C in the dark. DNA was extracted using an AccuPrep ® Genomic DNA Extraction Kit (Bioneer, South Korea) following the instructions of the manufacturer. Four primer sets, ITS1/ITS4 [115], EF1-728F/EF1-986R [116], T1/Bt2b [117,118], and ACT-512F/ACT-783R [116] were used to amplify the ITS region ITS1-5.8S-ITS2, portions of the tef1-α, BT and ACT genes, respectively. The identification of Botryosphaeriaceae isolates was confirmed by the sequencing of ITS and a partial sequence of tef-1a. For Phaeoacremonium isolates, a partial sequence of BT and ACT genes were amplified and sequenced. Molecular identifications of other isolates were confirmed by sequence analysis of ITS (Cadophora, Colletotrichum, Peyronellea, Stilbocrea, and Biscogniauxia isolates), BT (Colletotrichum and Peyronellea isolates), or tef1-α (Stilbocrea isolates). The polymerase chain reaction (PCR) was performed in a Techne TC-312 Thermal Cycler (Techne, Cambridge, UK), as described by Hashemi and Mohammadi [29]. For each isolate, 3-4 µL of PCR product was separated by electrophoresis on a 1% agarose gel (UltraPureTM Agarose, Invitrogen) containing ethidium bromide and visualized under UV illumination. The size of the products was evaluated using a 100 bp ladder (Gene Ruler, TMDNA Ladder Mix, Fermentas). PCR products were submitted to Bioneer Corporation (Daejeon, South Korea) for sequencing. MegaBLAST approach of the NCBI database (https://www.ncbi.nlm.nih.gov/) was initially used to identify fungal species.

Phylogenetic Analysis
Due to the broad range of Phaeoacremonium spp. obtained in this study, a phylogenetic analysis was carried out for the Phaeoacremonium spp. isolates. Sequences from citrus in Iran were aligned with sequences available in GenBank/NCBI. These were compared using MAFFT sequence alignment program v. 6 [119] with ex-type specimens from different hosts. Alignments were inspected in Sequence Alignment Editor v. 2.0a11 [120]. PAUP version 4.0 b 10 [121] was used to perform a partition homogeneity test. The congruence between the ACT and BT datasets was tested at 1000 replicates, and the maximum likelihood (ML) was carried out on the concatenated alignment. The MEGA version 7 software [122] was used for ML analysis. Bayesian information criterion in jModelTest 2.1.10 [123] was used to estimate the best fit model. Single and concatenated datasets were tested for branch support (1000 bootstrap replicates). We included sequences published by Spies et al. as reference sequences [58]. Pleurostoma richardsiae CBS 270.33 was included as an outgroup. Phaeoacremonium sequences obtained in this study were submitted to GenBank/NCBI (Table 2) and the sequence alignments were deposited in TreeBASE under study number 26006 (http://treebase.org). Table 2. Host, origin, and GenBank accession numbers of Phaeoacremonium isolates obtained from Citrus spp. in Iran (used in phylogenetic studies).

GenBank Accession Number
Phaeoacremonium Species Code b-Tubulin Actin Isolates used for pathogenicity tests on detached shoots of C. aurantifolia.

Pathogenicity Tests
Pathogenicity tests were carried out with 12 species on detached shoots of C. aurantifolia under controlled conditions. These include Do. viticola, P. italicum, P. minimum, P. rubrigenum, and P. parasiticum isolated from C. aurantifolia, L. theobromae, and Col. gloeosporioides obtained from C. sinensis, C. luteo-olivacea, and N. hyalinum recovered from C. limetta, P. fraxinopensylvanicum from C. limon, P. iranianum from C. reticulata and S. walteri isolated from C. aurantium. The shoots (38-40 cm in length and 2-2.5 cm in diameter) were surface-disinfected with alcohol (96%) and then were wounded at the uppermost internode with a 4-mm cork borer. To assess pathogenicity, wounds were inoculated with a 4-mm colonized PDA agar from 14-days-old cultures. All inoculated sites first were covered by moist cotton and then were wrapped with a strip of Parafilm (Pechiney Plastic Packaging, Menasha, USA). Six shoots per fungal isolate were used, and an equal number of shoots were also inoculated with 4-mm non-colonized PDA agar plugs for negative controls. Inoculated shoots were arranged at random, including the six inoculated shoots per isolate. Inoculated shoots were placed in moist chambers and incubated at 25 ºC. The total, upward, and downward lesion length data were evaluated individually, 40 days after inoculation. Recorded data were checked for normality of distribution by means of the Shapiro-Wilk and Kolmogorov-Smirnov tests. The data were subjected to analysis of variance (one-way ANOVA) using SAS v 9.1 (SAS Institute, Cary, NC, USA) (Dataset S1; Dataset S2). The least significant difference (LSD) test was used for comparison of treatment means at p < 0.05. Fungal re-isolations were made from the edges of the lesions on the test and control shoots and placed on PDA. The identity of the re-isolated fungi was confirmed based on morphological characteristics and molecular analysis in order to complete Koch's postulates. The pathogenicity of other species was not tested in this work because they were identified after the pathogenicity trials had begun on the detached shoots of C. aurantifolia.