In autumn 2015, during a survey of hazelnut diseases, we noticed plants with leaf blight symptoms in a commercial plantation in the Province of Vojvodina, Serbia. To isolate the pathogen, hazelnut branches were surface-sterilized with 5% sodium hypochlorite for 2 min, followed by three washes with sterile distilled water. Surface-sterilized tissue was transferred to sterile filter paper, placed on potato dextrose agar (PDA) containing streptomycin, and incubated at 24B0C in the dark for 10 days. Individual germinating conidia were selected, transferred directly to PDA plates according to the procedures described by Choi et al. (1999) , and stored on PDA in tubes at 4B0C. Colony morphology (color, shape and growth rate) was determined after 7-10 days of incubation on PDA at 25B0C in darkness. Dimensions of microscopic structures were calculated based on 30 measurements for conidial morphology (shape, color and cell number), size (length and width), and the presence and size of apical and basal appendages where possible. Images were captured by an Olympus SC100 color camera mounted on a BX31 microscope (Olympus, Japan).

Pathogenicity tests were performed on detached leaves and hazelnut fruits. Unhardened developed leaves and fruits were collected and surface-disinfected with 70% ethanol. Disinfected leaves and fruits were placed in glass Petri dishes containing moistened sterile filter paper. Plant tissue was inoculated using a 5-mm mycelial cube from an actively growing edge of the 10-day-old fungal culture. Samples were incubated for 14 days at 20-25B0C with a 12 h photoperiod ( Sezer & Dolar, 2015 ). Isolates RS-Le-1 and RS-Le-5 were tested and PDA cubes were used as negative control.

For molecular analysis, fungal DNA was extracted from cultured mycelia with 2% CTAB buffer ( Day & Shattock, 1997 ). Three separate PCR reactions were performed using ITS1/ITS4, EF1-728F/EF-2, and T1/Bt2b primer pairs, amplifying the fragments encompassing ITS1, ITS2, and 5.8S rDNA gene (ITS), partial translation elongation factor 1-α region (EF1-α), and 5' end of the β-tubulin gene (TUB), respectively ( White et al. , 1990 ; Glass & Donaldson, 1995 ; O'Donnell & Cigelnik, 1997 ; O'Donnell et al. , 1998 ; Carbone & Kohn, 1999 ). The polymerase chain reactions (PCR) were carried out in a Tpersonal thermal cycler (Biometra, Germany). Amplified PCR products were analyzed by electrophoresis in 1.5% agarose gel, stained with ethidium bromide, and visualized under UV light with a Gel Doc EZ System (Biorad, USA). The PCR products of tested isolates obtained with all three primer pairs were purified and custom sequenced (Macrogen, the Netherlands). Sequences of the Serbian isolates were aligned and compared with closely related sequences retrievedfromtheGenBank.Multiplesequence alignments were carried out using the software package BioEdit v. 7.0.5.2 ( Hall, 1999 ); phylogenetic analysis was performed with MEGA 6.0 ( Tamura et al. , 2013 ).

Two isolates, RS-Le-1 and RS-Le-5, were isolated from the hazelnut branches. Isolates are maintained as culture collections in the Institute for Forage Crops. On PDA, fungal colonies grew up to 55 mm in diameter within seven days at room temperature (25 B1 2B0C). Colonies had a smooth, even to undulating, colorless margin. Aerial mycelium was cottony pure white (Fig. 1a). Acervuli formed on the aerial mycelium contained black, slimy conidial masses. Conidiophores were hyaline and branched. Conidiogenous cells were annellidic, hyaline and smooth. The conidia were fusiform, five-celled, straight or slightly curved (Fig. 1b, c). The cells comprised three colored median cells and apical and basal hyaline cells with appendages. Conidia measured 16.57-27.53 x 5.05-9.13 μm (mean 25.66 x 7.54), five-celled with three brown central cells, the first two darker than the third one. The basal cell had a single 2.95-12.38 μm (mean 7.49) long appendage. The apical cell had 2-3 appendages with the following dimensions: first 3.67-30.39 μm long (mean 17.14), median 5.73-27.88 μm long (mean 20.39), and third 10.12-33.49 μm long (mean 17.17). Based on these morphological characteristics, the isolates were initially identified as belonging to the genus Pestalotiopsis ( Sutton, 1980 ; Maharachchikumbura et al. , 2014 ).

In pathogenicity tests, both tested isolates caused blight and necrosis on leaves and fruits (Figs. 1d,e,f). The pathogen was re-isolated from diseased leaves and fruits, confirming Koch's postulates.

In PCR analysis, fragments of 548, 493, and 824 bp from ITS, EF1-α, and TUB regions were amplified in both isolates, respectively. The PCR products of the tested isolates were sequenced; the obtained nucleotide sequences were deposited in the GenBank under the accession numbers KY552665-552666 (ITS), KY568911-568912 (TUB), KY568913-568914 (EF1-α). Sequences of ITS, EF1-α, and TUB regions of the two Serbian isolates were identical. The ITS sequences of our isolates showed 100% nucleotide (nt) identity with the sequences of accessions KM510410 of P. mangiferae and AY924281 of P. virgatula . In the reconstructed neighbor-joining phylogenetic tree (Kimura 2-parameter model), Serbian isolates were grouped with P. mangiferae , P. virgatula , P. oxyanthi , P. ixorae , and P. asiatica isolates (Fig. 2). Based on the EF1-α sequences, Serbian isolates showed the highest nt identity (99.79%) with P. asiatica accession JX399049. The TUB sequences of Serbian isolates showed the highest nt identity (99.51%) with the accession JQ762258 of P. ixorae . Phylogenetic analysis based on EF1-α and TUB sequences did not yield more information and was sufficient for proper species identification of the examined isolates (Figs. S1 and S2 [suppl]). The ITS gene is often used for molecular identification of Pestalotiopsis isolates, but the analysis of at least two genes may be more informative ( Hu et al. , 2007 ). In our study, phylogenetic analysis generated from three genes was not sufficient for species identification. A limited number of deposited EF1-α and/or TUB sequences of Pestalotipsis species in the GenBank impedes species identification. Previously, identification of Pestalotiopsis species was dependent on host association and on morphological and cultural characteristics ( Wei et al. , 2007 ; Hu et al. , 2007 ). Phylogenetic analysis facilitated species identification and become a necessity in species description ( Maharachchikumbura et al. , 2014 ). The detection of Pestalotiopsis sp. on hazelnut in Serbia is one of the several reports of these fungal species on this nut tree in the world. Our paper will provide valuable information for further study of Pestalotiopsis sp., reported here as a new causal agent of leaf blight on hazelnut in Serbia.