Fungal canker agents in apple production hubs of Iran

To identify apple canker casual agents and evaluate their pathogenicity and virulence in apple production hubs including West Azarbaijan, Isfahan and Tehran provinces; samples were collected from symptomatic apple trees. Pathogenic isolates on the detached branches were identified as Cytospora cincta, Diplodia bulgarica, Neoscytalidium dimidiatum and Eutypa cf. lata. E. cf. lata was reported as a potential apple canker causal agent in Iran for the first time based on the pathogenicity test on the detached branches, whereas it caused no canker symptoms in apple trees until 6 months after inoculation. Currently, E. cf. lata seems to be adapted to a single city. C. cincta, D. bulgarica and N. dimidiatum caused canker symptoms in apple trees. “C. cincta” and also “C. cincta and N. dimidiatum” were the most widespread and aggressive apple canker species, respectively, associated with apple canker in Iran. Therefore, they are considered to be the main threat to apple production in Iran and should be carefully monitored. Disease progress curve, area under the disease progress curve and optimum temperatures were determined for mentioned species. It is concluded that the establishment of each species occurs in appropriate areas and times in terms of the optimum temperature for their growth.

The disease symptoms, pathogenicity and morphological features for each one of the species have been shown in Figs. 3, 4, 5 and 6.
Diplodia bulgarica A.J.L. Phillips, J. Lopes & S.G. Bobev. The colonies of OU12 and KH40 had aerial mycelium with rosette-shaped growth, colony color was initially white which turned to dirty green, in reverse side gray olive, colony diameter was about 70-80 mm after 5 days on PDA at 25 °C. Conidiomata were produced on poplar twigs on water agar (WA) under near-ultraviolet . Diplodia bulgarica KH40. Symptoms of canker on infected apple trees in orchard (a,b), pathogenicity on detached branch and comparing it with control (c), pathogenicity on 2-year-old apple tree and comparing it with control (d = KH40 isolate and e = control), culture growing on potato dextrose agar (f), conidiomata developing on poplar twigs on water agar (g), conidia exuded from conidiomata developed on poplar twigs (h), vertical section through pycnidia (i), hyaline immature conidia developing on conidiogenous cells (j), hyaline aseptate conidia (k) brown mature conidium on conidiogenous cells (l), brown aseptate conidia (m), brown one-septateconidia (n). . Neoscytalidium dimidiatum SK109. Symptoms of canker on infected apple trees in orchard (a-c), pathogenicity on detached branch and comparing it with control (d), pathogenicity on 2-year-old apple tree and comparing it with control (e-g = SK109 isolate and h = control), culture growing on potato dextrose agar (i), arthroconidia developing (j), mature 0-1-septate arthroconidia (k). The colonies of SK109 and KO120 displayed uniform radial growth, initially white and gradually olivaceous and finally dark, also the colony attained a diameter of about 90 mm after three days on PDA at 25 °C. The conidiomata were not produced on poplar twigs on water agar (WA) under NUV or PDA after 6-months. The aerial mycelium became arthrospore with age. The arthrospores were initially pale olive and eventually dark brown, Figure 5. Cytospora cincta D139. Symptoms of canker on infected apple trees in orchard (a-c), pathogenicity on detached branch and comparing it with control (d), pathogenicity on the 2-year-old apple tree and comparing it with control (e = KH40 isolate, f = control), culture growing on potato dextrose agar (g), stromata developing on PDA (h), stromata developing on poplar twigs on water agar (i), conidiophores (j), hyaline allantoid conidia (k,l).  25 and Phillips et al. 19 , these species (SK109 and KO120) were identified as Neoscytalidium dimidiatum. The molecular analysis confirmed the result of the morphological identification (Fig. 8). . Isolate "KH27" belonged to Eutypa genus according to the identification key for Diatrypaceae by Rappaz 26 . Since the identification of Eutypa spp. is based on sexual fruiting bodies on infected tissues and there were no fruiting bodies in sampling time, therefore, Eutypa sp. was completely identified based on molecular studies.

Symptoms of different species in orchards.
The symptoms caused by different species were determined in the orchards. It is noteworthy that the canker symptoms caused by C. cincta, Diplodia bulgarica, N. dimidiatum and E. cf. lata were different in appearance. www.nature.com/scientificreports/ Symptoms of C. cincta were twig and branch dieback and necrotic bark. Cytospora canker is characterized by diffuse resinous branch cankers, with fruiting bodies of the causal fungi usually forming on infected parts.
D. bulgarica caused stem and trunk cankers, bark discoloration and scaling-off of the bark. Symptoms of cankers caused by N. dimidiatum included bark lesions, discoloration of xylem tissues, longitudinal wood necrosis and extensive spore production as black powder under bark, spur and shoot blight.
Apple trees infected by E. cf. lata displayed branch and scaffold dieback, with dead leaves still attached indicating rapid death during the season.
Only N. dimidiatum was identified in Khomeini Shahr County (Isfahan Province). But both species "N. dimidiatum and C. cincta" were identified in Semirom County of Isfahan Province (Supplementary Table S1, Figs. 7 and 8). The severity of canker caused by C. cincta was similar to that with N. dimidiatum (Data not shown).
In Damavand, only C. cincta was identified (Supplementary Table S1 and Fig. 7). In Urmia County (West Azerbaijan Province), C. cincta and D. bulgarica were identified (Supplementary  Table S1 and Figs. 7 and 8). The severity of canker caused by C. cincta was higher than that with D. bulgarica (Data not shown).  Table S1 and Figs. 8 and 9). The severity of canker caused by D. bulgarica was higher than with E. cf. lata (Data not shown).
Pathogenicity test on 2-year-old apple trees. The symptoms were evaluated 6 months after inoculation (26 May to 24 November 2013). The result showed that N. dimidiatum, D. bulgarica and C. cincta caused canker symptoms, but E. cf. lata did not cause canker symptoms on 2-years-old apple trees (Fig. 10). Comparison of Means (LSD, 0.05 level) for both parameters "CL and CP/SP" indicated a significant difference between pathogenic isolates except Eutypa cf. lata (KH27) as compared to the control ( Fig. 10 and Supplementary Table S3). Also, the various species differently affected CL and CP/SP parameters. In this test, the highest and the lowest virulence were recorded for "N. dimidiatum SK109" and "C. cincta OU4, SS65, SH86, SS98 and SS100", respectively ( Fig. 10 and Supplementary Table S3).
Disease progress curve. Disease progress curves (DPCs) were plotted for the pathogenic isolates ( Fig. 11 and Supplementary Fig. S1, and Supplementary Table S4). DPC of N. dimidiatum SK109 can be divided into three sections. In the first and second sections, the slope of the curve is increasing while the increment of the curve slope in the second section is higher than the first one. Indeed, ascending temperature resulted in ascending disease progresses, and the curve slope increased, so that, it is the highest level from July 7 to Aug 4. The  www.nature.com/scientificreports/ temperature of this section was about 31.6 °C which the upper portion of the canker was dried. The curve slope in the third section is lower than the first and second ones. Descending temperatures led to descending the curve slope until October 13, which was stopped (≤ 19.9 °C) (Fig. 11). DPC of D. bulgarica KH40 can be divided into two sections. In the first section, the curve slope is higher than the second one (May 26 to Jun 9 with temperature about 27 °C). The curve slope (CL development) in the second section is lower as compared to the first ones, and disease progress was stopped on Nov 10 (≤ 14 °C) (Fig. 11).
It is noteworthy that canker symptoms in apple trees inoculated with E. cf. lata KH27 was not observed. Therefore, the progress curve of apple canker caused by E. cf. lata KH27 was linear, similar to control (Fig. 11).

Discussion
Four species, isolated from the apple trees displaying canker symptoms in Iran, were identified based on morphological characters and phylogenetic analysis. They belonged to four different genera in three different families, D. bulgarica and N. dimidiatum (Botryosphaeriaceae), C. cincta (Valcaceae) and E. cf. lata (Diatrypaceae). All identified species caused canker symptoms on the detached branches in the pathogenicity tests. D. bulgarica, N. dimidiatum and C. cincta caused canker symptoms on 2-year-old apple trees in the pathogenicity test. All the isolates of C. cincta (OU4, SS65, SH86, SS98, SS100, D131, D134 and D139) displayed a good growth rate at 15-25 °C (Fig. 11 and Supplementary Fig. S2). It should be noted that the temperature in the detached branch test was 25 °C, but the field temperature was more than 25 °C from May 26 to September 15 (Fig. 11). This could explain why all the isolates of C. cincta displayed more virulence on apple detached branches than that on 2-yearold trees. Also, E. cf. lata KH27 caused no canker symptoms until 6 months after inoculation. Accordingly, the optimum temperature for E. cf. lata KH27 was recorded 20-25 °C (Fig. 11), which was provided in the detached branch test, but the temperature in the pathogenicity on 2-year-old tree in the field was not optimal during 6 months of the experiment period except September 15-29 (Fig. 11). It can be the reason why E. cf. lata KH27 was not pathogenic on 2-year-old apple trees in the field. www.nature.com/scientificreports/ Eutypa lata was reported as an apple canker causal agent 20,21,27 , and also its pathogenicity test was performed on the detached branches 22 , but this is the first attempt of the pathogenicity test of E. cf. lata on the apple tree. Pathogenicity test for E. lata was conducted on 2-year-old Ribes rubrum and the symptoms were observed 21 months after inoculation 28 . Also, the pathogenicity test of this species was performed on rooted cuttings of grapevine under greenhouse conditions (with a temperature about 22 °C) and the pathogenicity was confirmed 54 months after inoculation 29 . According to the results of the pathogenicity test of E. lata on grapevine 28 and Ribes rubrum 29 , the pathogenicity test for this species should be done in controlled conditions (greenhouse and growth chamber) because disease symptoms appeared at a very long time after inoculation at the optimum temperature.
In this study, the disease symptoms did not appear in the apple trees inoculated with E. cf. lata until 6 months after inoculation under field conditions, whereas its pathogenicity and virulence were confirmed on the detached branch. E. cf. lata was reported as an apple potential canker causal agent in Iran based on pathogenicity test on the detached branches for the first time. Nevertheless, the pathogenicity test on the detached branch presents useful evidence, the field tests on apple trees should be conducted to confirm the pathogenicity and also determine disease severity in orchards.
Cytospora cincta (anamorph of L. cinctum) has been previously reported as canker causal agent in different regions of Iran 10-12 . It is noticed that only the pathogenicity of isolates from Karaj-Iran had been demonstrated 10 . Besides, the pathogenicity of other species of Cytospora isolated from the apple trees displaying canker symptoms has not been confirmed. In the world, only the pathogenicity of L. cinctum in Michigan has been demonstrated 9 . In this study, C. cincta was reported for the first time from West Azerbaijan and Damavand-Tehran provinces of Iran.
Diplodia bulgarica had been reported on apple trees displaying canker symptoms in West Azarbaijan and Kermanshah provinces of Iran, and also its pathogenicity had been confirmed 30,31 . Also, D. seriata and D. malorum were reported as apple canker causal agents in West Azarbaijan and Kermanshah provinces of Iran 32,33 . In this study, D. bulgarica was the highest incidence in West Azarbaijan province (Data not shown).
Also, N. dimidiatum was reported as apple dieback agent in tropical and subtropical regions of Iran, Khuzestan, Kerman and Fars Provinces 15 . However, its pathogenicity has not been reported on apple trees. Also, this species has not been reported from the main apple-growing areas in Iran (cold regions). In this study, the pathogenicity of N. dimidiatum was confirmed on the apple detached branches and trees. Also, N. dimidiatum was reported from Isfahan province of Iran for the first time.
Despite the other studies, AUDPC was measured and compared for these species. Canker's progress of N. dimidiatum (SK109) was more than other species (Fig. 11 and Supplementary Fig. S1, and Supplementary  Table S3) because the growth rate of N. dimidiatum SK109 was higher than other species. The other reason is the growth optimum temperature for this species (about 30 °C, Fig. 11) provided during the pathogenesis. The progress of the canker caused by C. cincta D139 was stopped 2 weeks after inoculation (Jun 9) because the average temperature of the assessment region was 27.8 °C after this period which D139 grew slightly at temperatures above 25 °C (Fig. 11).
Two pathogenic species (N. dimidiatum and C. cincta) were identified in Isfahan province (Semirom and Khomeini Shahr counties) (Supplementary Table S1, Figs. 7 and 8). Only N. dimidiatum was identified in Khomeini Shahr County (Kooshk city). But both species "N. dimidiatum and C. cincta" were identified in Semirom County (Supplementary Table S1, Figs. 7 and 8). In Semirom county, C. cincta was more prevalent than N. dimidiatum (Data not shown). Indeed, N. dimidiatum species was identified only in one city (Komeh) which is southern part of Semirom County and has warmer climate. It is interesting that C. cincta was not found in Komeh city (Supplementary Table S1 and Fig. 7). The progress peak of apple canker caused by N. dimidiatum was recorded on July and August (Data not shown). Additionally, May, June and September were the progress peak of apple canker caused by C. cincta (Data not shown).
In Damavand, only C. cincta was identified (Supplementary Table S1 and Fig. 7). There was a difference between the isolates of C. cincta in terms of virulence and the most virulent isolates on 2-year-old apple trees were the isolates (D13, D134 and D139) of Damavand ( Fig. 10 and Supplementary Table S1). May and September were apple canker progress peak in Damavand (Data not shown).
In Urmia County (West Azerbaijan Province), C. cincta and D. bulgarica were identified (Supplementary Table S1 and Figs. 7 and 8). The progress peak of canker caused by C. cincta occur on May, June and September (Data not shown). Also, the progress peak of canker caused by D. bulgarica was observed on July and August (Data not shown).
In Khoy County (West Azerbaijan Province), D. bulgarica and E. lata species were identified (Supplementary Table S1 and Figs. 7 and 8), which E. cf. lata is not considered as an important pathogen due to low disease progress. The progress peak of apple canker caused by D. bulgarica was observed from June to August (Data not shown).
It is concluded that the establishment of each species occurs in areas compatible with their growth, and also the highet progress in canker symptoms for each species occurs in appropriate times in regard to the optimum temperature for their growth (Supplementary Table S5).
Based on this study and the results of Proffer and Jones 9 , it is suggested that only C. cincta (L. cinctum) is an important pathogen among Cytospora spp. isolated from apple tree displaying canker symptoms. In this study, isolates of C. cincta were isolated from thick branches with high infection. Proffer and Jones 9 isolated different fungal species from apple trees displaying canker symptoms in Michigan. They conducted the pathogenicity test for V. malicola (teleomorph C. schulzeri), L. cinctum, B. stevensii and B. obtusa and evaluated their symptoms 6 weeks after inoculation. The results showed that isolates of L. cinctum, B. stevensii and B. obtusa were able to make canker symptoms, while isolates of V. malicola caused no canker symptoms.
The current research presents the first in-depth study regarding the isolation, pathogenicity, virulence and phylogenetic analysis of fungal pathogens associated with apple canker in Iran. Given the spread of canker disease www.nature.com/scientificreports/ on apple as well as its economic importance in Iran, the identification of apple canker agents would pave the way for its integrated management including biological control with focusing on aggressive species and/or isolates. Besides, since different causal agents including D. bulgarica, N. dimidiatum, C. cincta, and E. cf. lata contribute to apple tree cankers, it is essential to take measures to manage their spread to new orchards. According to our results, C. cincta is the most widespread canker pathogens of apple in Iran. Additionally, C. cincta and N. dimidiatum is the most aggressive apple canker pathogens in Iran. Also, N. dimidiatum SK109 displayed the most virulence on Golden delicious cultivar, one of the most widely apple cultivar. Therefore, C. cincta and N. dimidiatum are considered to be a main threat to apple production in Iran and should be carefully monitored. Currently, E. cf. lata seems to be adapted to a single city (West Azerbaijan province-Khoy county-Firuraq city); but additional sampling will likely reveal its further geographical and host range. It is noteworthy that 77 of the tested isolates were not able to cause canker symptoms. A number of these isolates may be associated with apple canker and contribute to disease. Since apple trees are perennial plants, they may be infected with different fungal species inciting a complex disease in orchestrate with two or several fungal species. Meanwhile, additional pathogenic fungi may cause apple canker in Iran, remain to be investigated. Therefore, the further studies are still needed to identify apple canker causal agents and also discover their roles in disease establishment and severity in Iran.

Materials and methods
Sampling and fungal isolation. The survey was conducted in three major apple production provinces  Table S1), and 10 trees were randomly selected from four orchards in each site From May to September 2012. It is noteworthy that sampling was performed from Golden Delicious and Red Delicious, the main planted cultivars in Iran. Symptomatic branches with 2 to 5 cm in diameter were cut and transported to the lab in paper bags. The samples were firstly disinfected by 70% ethanol, then bark was removed, and small pieces (5 × 5 mm) of wood tissue were dissected from the margin of the canker. The pieces were placed on Petri dishes containing PDA medium amended with chloramphenicol (200 mg/L). Emerging colonies were purified using hyphal tip culture on 2% water agar (WA).

Plant materials.
It is declared that 2-year-old apple trees (cv. Golden Delicious) were legally bought from local commercial fruit trees provider company named ITA-Sadra (http:// itasa dra. ir/? page_ id= 2196& lang= en) and all methods involving plant studies were performed in accordance with the relevant guidelines and regulations.
Pathogenicity tests on detached branches. Firstly, 90 isolates were selected according to the geographic region and morphotype (Supplementary Tables S1 and S2). Then pathogenic isolates were detected based on the pathogenicity test on the detached branches 34 . For this purpose, the detached branches (20 cm long and 1.5 to 2 cm diameter) from 2-year-old apple trees (cv. Golden Delicious) were sealed by Parafilm at two both ends to avoid dehydration and surface-sterilized using 70% ethanol and inoculated in the middle after removing the cortex with a cork borer (5-mm diameter). Instantly a mycelium plug (5-mm diameter from 4-day-old cultures) of each isolate was placed on the wound in direct contact with vascular tissue, then wrapped using Parafilm. The control was inoculated with PDA plug. To keep the relative humidity high, the control and inoculated detached branches were placed on a plastic mesh plate in a plastic container covered with a moistened paper towel at the bottom and maintained at 25 °C. CL/SL and also CP/SP ratio were measured 4 weeks after inoculation. The experiment was conducted in a randomized complete block design (RCBD) with five replications.

Identification of pathogenic isolates. Morphological identification.
To induce sporulation, the isolates were cultured on WA medium having pieces of double-autoclaved and halved poplar twigs, then incubated under NUV light with a 12-h photoperiod at 23-27 °C for a suitable period ranging from 2 to 10 weeks 19 . The cultures were microscopically inspected at regular intervals during the incubation period. After emerging the conidiomata on poplar twigs, they were vertically cut using a sharp scalpel, and the sporogenous tissues were mounted on a microscope slide containing a drop of 100% lactic acid, then the characteristics of spores and other organs were recorded.  (KH40 and OU12), Neoscytalidium sp. (SK109 and KO120), and Cytospora sp. (OU4, SS65, SH86, SS98, SS100, D131, D134 and D139) together with the other sequences of related genus were already used by other studies 11,18,19,[40][41][42][43][44][45] were selected for phylogenetic analyses. The dataset was updated by investigations in the database for acquiring accurate sequences. The outgroup taxa for present dataset were taken based on previous studies 41,42,45 . All sequences were aligned using Q-INS-i algorithm of MAFFT version 7 (http:// mafft. cbrc. jp/ align ment/ server/) 46 and the online version of Gblocks 0.91b 47 was applied to remove ambiguous parts of the alignment, with all three options (including allow smaller final blocks, allow gap positions within the final blocks, and allow less strict flanking position) for a fewer stringent selection (http:// molev ol. cmima. csic. es/ castr esana/ Gbloc ks_ server. html). The most suitable substitution model for dataset was chosen using Akaike information criterion (AIC) by means of PAUP * /MrModeltest v2.2. A symmetrical model including a gamma distribution were picked out for rates with SYM + I + G for Cytospora with ITS analysis, and GTR + G for Eutypa with ITS and BT analysis, Diplodia and Neoscytalidium with ITS and TEF-1-α analysis. Bayesian inference (BI) was carried out by means of MrBayes v3.1.2 48 with choosing a random beginning tree and running the chains for 4 million for ITS and also combined sequences of ITS and TEF-1-α or ITS and BT. After casting off burn-in samples, the residual samples were reserved for additional analyses. The Markov Chain Monte Carlo (MCMC) method within a Bayesian framework was utilized to assess the posterior probabilities of the phylogenetic trees 49  Pathogenicity test on apple trees. Pathogenicity test was also carried out on 2-year-old apple trees under field conditions (research farm of agriculture faculty, Tarbiat Modares University) on 26 May 2013. The stems of 2-year-old apple trees (cv. Golden Delicious) were surface-sterilized using 70% ethanol. For inoculation, the stems were wounded in the middle part by removing the cortex (5 mm diameter) with a sterile cork borer. Instantly a mycelial plug (5 mm diameter) of 5-day-old culture was placed on the wound, then wrapped using Parafilm. The control was inoculated with a PDA plug. CL and also CP/SP ratios were measured 6 months after inoculation. The experiment was conducted in RCBD with three replications.
Disease progress curve. To measure AUDPC, CL was measured at 13-time points with 2 weeks intervals during the pathogenicity period on the apple trees. Local daily temperatures from 26 May to 24 November 2013 were provided by the synoptic weather station of Chitgar-Tehran-Iran. AUDPC was calculated for CL every 2 weeks until 6 months after inoculation according to Eq. (1) 51 .
where n denotes the number of measurements, x is CL at each measurement, t signifies the number of days between measurements.
where R 1 is the radius of the pathogen colony after 72 h, and R 2 signifies the radius of the pathogen colony after 24 h. Statistical analysis. The hypothesis of normality and equal variance were tested, and data transformation was performed using square root and log base 10 for the detached branch and 2-year-old apple tree experiments, respectively. Conventional parametric statistics were applied for the analysis. The data was statistically analyzed by using SAS (SAS 9.1) and SPSS (SPSS 15.0). ANOVA was conducted by GLM statistical method and means comparison was done by least significant difference (LSD) test. GraphPad Prism (GraphPad Prism 5) software was used for making graphs.
Ethical approval. All authors approve Ethics and consent for participation and publication. All authors of the manuscript have read and agreed to its content and are accountable for all aspects of the accuracy and integrity of the manuscript in accordance with ICMJE criteria. That the article is original, has not already been published in a journal, and is not currently under consideration by another journal.

Data availability
The dataset supporting the conclusions of this article is included in the article and Supplementary information.