High Genetic Diversity and Species Complexity of Diaporthe Associated With Grapevine Dieback in China

Grapevine trunk diseases have become one of the main threats to grape production worldwide, with Diaporthe species as an emerging group of pathogens in China. At present, relatively little is known about the taxonomy and genetic diversity of Chinese Diaporthe populations, including their relationships to other populations worldwide. Here, we conducted an extensive field survey in six provinces in China to identify and characterize Diaporthe species in grape vineyards. Ninety-four isolates were identified and analyzed using multi-locus phylogeny. The isolates belonged to eight species, including three novel taxa, Diaporthe guangxiensis (D. guangxiensis), Diaporthe hubeiensis (D. hubeiensis), Diaporthe viniferae (D. viniferae), and three new host records, Diaporthe gulyae (D. gulyae), Diaporthe pescicola (D. pescicola), and Diaporthe unshiuensis (D. unshiuensis). The most commonly isolated species was Diaporthe eres (D. eres). In addition, high genetic diversity was observed for D. eres in Chinese vineyards. Haplotype network analysis of D. eres isolates from China and Europe showed a close relationship between samples from the two geographical locations and evidence for recombination. In comparative pathogenicity testing, D. gulyae was the most aggressive taxon, whereas D. hubeiensis was the least aggressive. This study provides new insights into the Diaporthe species associated with grapevines in China, and our results can be used to develop effective disease management strategies.


INTRODUCTION
In natural ecosystems, plant pathogens play important roles such as regulating host populations and host plant geographic and ecological distributions. Consequently, they can affect the availability of food sources to other living organisms (Lindahl and Grace, 2015). Most microbial pathogens have short generation times and large population sizes, which can result in high genetic variations and rapid adaptations to environmental stresses and to human-mediated factors such as fungicide resistance (Alberts et al., 2002;Lindahl and Grace, 2015). Hence, it is important to understand the genetic diversity and population variation of plant pathogens to develop sustainable control measures.
Grape is one of the most important fruit crops in China. China is the second largest grape-cultivating country and the top producer in the world (OIV, 2016). In 2016, the total grape cultivation area was estimated at 847 kha, and 14.5 million metric tons of fresh grapes were produced in China (OIV, 2016). Therefore, infectious diseases with significant risks to grape production have drawn broad attention from the grapevine industry. Grapevines are affected by several foliar diseases (Gadoury et al., 2012;Zhang et al., 2017), fruit diseases (Daykin and Milholland, 1984;Hong et al., 2008;Greer et al., 2011;Jayawardena et al., 2015), and trunk diseases (Yan et al., 2013;Dissanayake et al., 2015a,b). Grapevine trunk diseases have drawn considerable attention, as these diseases affect the perennial parts of the vine and can limit grape production for many years (Yan et al., 2013(Yan et al., , 2015. The genus Diaporthe Nitschke., belongs to the family Diaporthaceae, and is typified by Diaporthe eres (D. eres) Nitschke (Senanayake et al., 2017). Following the nomenclature rules Rossman et al. (2014) proposed that the genus name Diaporthe over Phomopsis as it was introduced first, represents the majority of species. In earlier species names were given to Diaporthe taxa based on their host specificity. This resulted in over 100 names listed under the genus Diaporthe (http://www.indexfungorum. org/Names/Names.asp and http://www.mycobank.org). With advances in molecular techniques, multi-locus DNA sequence data together with morphological characteristics have been extensively used for the delimitation of Diaporthe species (Udayanga et al., 2011;Gomes et al., 2013;Gao et al., 2017). The internal transcribed spacer (ITS), translation elongation factor-1a (EF-1α), β-tubulin, partial histone H3 (HIS), calmodulin (CAL), genes are the most commonly used gene regions for molecular characterization (Udayanga et al., 2011;Gao et al., 2017;Yang et al., 2018). Multiple studies have used different gene combinations to resolve the species boundaries in this genus (Udayanga et al., 2011(Udayanga et al., , 2014aGao et al., 2017;Marin-Felix et al., 2019). Species belonging to genus Diaporthe are endophytes, pathogenic, and saprobic on wide range of hosts worldwide Hyde et al., 2016;Marin-Felix et al., 2019). They are well-known pathogens on economically important crops (Udayanga et al., 2011). Several common disease among those are dieback on forest trees (Yang et al., 2018), leaf spots on tea (Guarnaccia and Crous, 2017), leaf and pod blights and seed decay on soybean , melanose, stem-end rot, and gummosis on Citrus spp. (Mondal et al., 2007;Udayanga et al., 2014a;Guarnaccia andCrous, 2017, 2018) and stem canker on sunflower (Muntañola-Cvetković et al., 1981;Thompson et al., 2011).
Phomopsis cane and leaf spot caused by Diaporthe species on grapevine is one of the most complex grapevine trunk diseases worldwide (Úrbez- Dissanayake et al., 2015a;. The disease symptoms of Diaporthe Dieback include shoots breaking off at the base, stunting, dieback, loss of vigor, reduced bunch set, and fruit rot (Pine, 1958(Pine, , 1959Pscheidt and Pearson, 1989;Pearson and Goheen, 1994;Wilcox et al., 2015). In woods brown to black necrotic irregularshaped lesions could be observed. Once clusters are infected rachis necrosis and brown, shriveled berries close to harvest could be observed (Pearson and Goheen, 1994). More than one Diaporthe species is frequently reported as causative agents from one country (Dissanayake et al., 2015a;. Currently, 27 species have been identified as causal organisms of Diaporthe dieback in grape-producing countries worldwide Udayanga et al., 2011Udayanga et al., , 2014aWhite et al., 2011;Baumgartner et al., 2013;Úrbez-Torres et al., 2013;Hyde et al., 2014;Dissanayake et al., 2015a;Lesuthu et al., 2019). Even though these species characterized under the one disease, disease symptoms, and aggressiveness are varying according to the species. Diaporthe ampelina (D. ampelina) has a long history as the most common and severe pathogenic species together with D. amygdali . Diaporthe ampelina and Diaporthe kyushuensis (D. kyushuensis) are the causal agent of grapevine swelling arm (Kajitani and Kanematsu, 2000;. Diaporthe perjuncta (D. perjuncta) and D. ampelina caused cane bleaching (Kuo and Leu, 1998;Kajitani and Kanematsu, 2000;Mostert et al., 2001;Rawnsley et al., 2006). Lesuthu et al. (2019) showed that D. ampelina, Diaporthe novem (D. novem), and Diaporthe nebulae (D. nebulae) as the most virulent species of Diaporthe associated with grapevines in South Africa. Diaporthe eres was found as a weak to moderate pathogen in several different studies Baumgartner et al., 2013). These results indicate the complexity and high species richness of Diaporthe associated with the grapevines. Up to now in China four Diaporthe species have been reported causing grapevine dieback (Dissanayake et al., 2015a). Those are D. eres, Diaporthe hongkongensis (D. hongkongensis), Diaporthe phaseolorum (D. phaseolorum), and Diaporthe sojae (D. sojae). Their taxonomic placements and pathogenicity under a controlled environment were also studied.
The study conducted by  showed that species of Diaporthe also associated as endophytes on grapes as well. In that study they observed that Diaporthe bohemiae (D. bohemiae), which was isolated from grape was unable to induce lesions. In addition to grapevines, Diaporthe have been reported on broad range of hosts (Udayanga et al., 2011). However, the most important charter is the ability of endophytic Diaporthe species to be opportunistic pathogens. Huang et al. (2015) observed that some Diaporthe species associated with citrus in China shown to act as opportunistic plant pathogens. Diaporthe foeniculina (D. foeniculina) has been found as both endophyte and opportunistic pathogen on various herbaceous weeds, ornamentals, and fruit trees (Udayanga et al., 2014a;Guarnaccia et al., 2016). So far it is not confirmed the factor that driven into pathogenicity from endophytes either due to environmental changes or the reduction of host's defense. Therefore, further studies are required to understand this in both field level and genomic level.
However, the genetic diversity of Diaporthe spp. associated with Vitis spp., relationships among isolates from different geographical regions, and relationships among isolates from China and those from other countries were not investigated. Therefore, to expand our knowledge on these issues, we performed an extensive field survey to isolate and identify Diaporthe species associated with grapevine dieback in China. We reconstructed a phylogenetic tree for the genus Diaporthe. The present study analyzed the genetic diversity of Diaporthe species associated with grapevines in China and constructed haplotype networks for Diaporthe species from different geographical origins for the first time. Finally, we analyzed the relationship between Diaporthe species from European and Chinese grape vineyards, as Diaporthe dieback is becoming an emerging trunk disease in both regions .

Sampling and Pathogen Isolation
Field surveys were conducted during 2014 and 2015 in 20 vineyards in the six following provinces in China: Guangxi, Heilongjiang, Hubei, Jilin, Liaoning, and Sichuan (Figure 1). Samples were collected from symptomatic grapevine woody branches that exhibited bark discoloration, shoots breaking off at the base, stunting, wedge-shaped cankers, and light brown streaking of the wood from the following Vitis vinifera (V. vinifera) cultivars: Centennial Seedless, Red Globe, and Summer Black (Figure 2). Symptomatic tissue samples were collected into zip-lock plastic bags that contained wet sterilized tissue papers to maintain humidity. Once the samples were taken into the laboratory, infected trunks or shoots were photographed, and symptoms, location, and other relevant data were documented. The fungal pathogens were isolated using the following procedures. Infected shoots/trunks were cut into small pieces (1-3 mm thick). These pieces were then surface-sterilized by dipping into 70% ethanol for 30 s and then transferred into 1% NaOCl for 1 min. This step was followed by two washes with sterile distilled water. Once the wood pieces were dried, they were placed onto potato dextrose agar (PDA) plates supplemented with ampicillin (0.1 g L −1 ) and incubated at 25 • C. After 5-7 days of incubation, hyphal tips of fungi immerging from wood pieces were transferred onto new PDA plates and incubated until they produce conidia. Once the conidia were developed single spore isolation was done. For the strains do not developed conidia after 4 weeks two-three times hyphal tip isolation was done. All the pure cultures obtained in this study were deposited in the culture collection of Institute of Plant and Environment Protection of Beijing Academy of Agriculture and Forestry Sciences (JZB culture collection) at 4 • C.

DNA Extraction, PCR Amplification, and Sequence Assembly
Approximately 10 mg of aerial mycelium was scraped from 5-7 days old isolates grown on PDA (Potato Dextrose Agar) at 25 • C. Total genomic DNA was extracted using the DNeasy Plant Mini Kit (QIAGEN GmbH, QIAGEN Strasse 1, 40742 Hilden, Germany). For species confirmation, the internal transcribed spacer (ITS) regions were sequenced for all isolates. The obtained sequences were compared to those in GenBank using the MegaBLAST tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi). After isolates were confirmed as belonging to the genus Diaporthe, six additional gene regions, those encoding translation elongation factor-1α (EF-1α), β-tubulin, calmodulin (CAL), partial histone H3 (HIS), partial actin (ACT), and DNA-lyase (Apn2), were sequenced. Table 1 presents the primer pairs with their respective amplification conditions for each of the above gene regions. PCR mixtures of 25 µl total volume consisted of 0.3 µl of TaKaRa Ex-Taq DNA polymerase, 2.5 µl of 10 × Ex-Taq DNA polymerase buffer, 3.0 µl of dNTPs, 2 µl of genomic DNA, 1 µl of each primer, and 15.2 ddH2O. The PCRs were conducted in a Bio-Rad C1000 thermal cycler (Germany). The resulting products were visualized on a 1% agarose gel stained with ethidium bromide under UV light using a Gel DocTM XR Molecular Imager (Bio Rad, USA). All positive amplicons were sequenced by Beijing Biomed Gene Technology Co LTD. The sequence quality was confirmed by checking chromatograms using BioEdit v. 5 (Hall, 2006). Sequences were obtained using both forward and reverse primers, and consensus sequences were generated using DNAStar v. 5.1 (DNASTAR, Inc.). The sequence data generated in the present study have been deposited in GenBank ( Table 2).

Phylogenetic Analyses
For the phylogenetic analyses, reference sequences representing related taxa in Diaporthe were downloaded from GenBank Yang et al., 2018; Table 3) and aligned with the sequences obtained in this study ( Table 2). The sequences were aligned using MAFFT (Katoh and Toh, 2010) (http://www.ebi.ac.uk/Tools/msa/mafft/) and manually adjusted using BioEdit v. 5 (Hall, 2006) whenever necessary. Phylogenetic relationships were inferred using maximum parsimony (MP) implemented in PAUP (v4.0) (Swofford, 2003), maximum likelihood (ML) in RAxML (Silvestro and Michalak, 2010) and Bayesian analyses in MrBayes v. 3.0b4 (Ronquist and Huelsenbeck, 2003). In phylogenetic analysis, single-gene trees were constructed first using ML in RAxML. The phylogenetic tree topologies for different gene fragments were compared for evidence of incongruences with a focus on comparing branches with high bootstrap values. If no conflict was observed, a combined phylogenetic tree was generated.
In PAUP, ambiguous regions in the alignment were excluded for further analyses, and gaps were treated as missing data. The stability of the trees was evaluated by 1000 bootstrap replications. Branches of zero length were collapsed, and all multiple parsimonious trees were saved. Parameters, including tree length (TL), consistency index (CI), retention index (RI), relative consistency index (RC), and homoplasy index (HI) were calculated. Differences between the trees inferred under different optimality criteria were evaluated using Kishino-Hasegawa tests (KHT) (Kishino and Hasegawa, 1989). The evolutionary models for each locus used in Bayesian analysis and ML were selected using MrModeltest v. 2.3 (Nylander, 2004). ML analyses were accomplished using RAxML-HPC2 on XSEDE (8.2.8) (Stamatakis et al., 2008;Stamatakis, 2014) in the CIPRES Science Gateway platform (Miller et al., 2010) using the GTR + I + G model of evolution with 1000 non-parametric bootstrapping iterations. Bayesian analysis was performed in MrBayes v. 3.0b4   (Ronquist and Huelsenbeck, 2003), and posterior probabilities (PPs) were determined by Markov chain Monte Carlo sampling (MCMC). Six simultaneous Markov chains were run for 106 generations, sampling the trees at every 100th generation. From the 10,000 trees obtained, the first 2,000 representing the burn-in phase were discarded. The remaining 8,000 trees were used  (Jayasiri et al., 2015) and Index fungorum (http://www.indexfungorum.org). New species are described following Jeewon and Hyde (2016).

Morphology and Culture Characteristics
Colony morphology and conidial characteristics were examined for Diaporthe species identified by phylogenetic analysis. Colony colors were examined according to Rayner (1970) after 7 days of growth on PDA in the dark at 25 • C. Digital images of morphological structures mounted in water were taken using an Axio Imager Z2 photographic microscope (Carl Zeiss Microscopy, Oberkochen, Germany). Measurements were taken using ZEN PRO 2012 (Carl Zeiss Microscopy). Conidial length and width were measured for 40 conidia per isolate, and the mean values were calculated for all measurements. Conidial shape, color, and guttulation were recorded.

Genetic Diversity and Population Structure Analysis
Among the identified species, only one, Diaporthe eres, had a count of >20 individuals. As a result, only D. eres was selected for the analysis of genetic diversity and population relationships. For the D. eres population, diversity indices were calculated for each gene region and the combined sequence dataset. DnaSP v. 6.12 (Librado and Rozas, 2009) was employed to calculate haplotype richness (hR), the total number of haplotypes, Watterson's theta ( w), and pairwise nucleotide diversity (JI).
To overcome the population size effects, hR, w and JI were calculated after 1,000 repetitions, and the median estimate was recorded for each parameter. To understand the potential departure from an equilibrium model of evolution, Tajima's D was calculated using DnaSP v. 6.12 with a permutation test of 1,000 replicates. The minimum numbers of recombination events (ZnS) used by Kelly (1997) and the recombination parameters Za and ZZ used by Hudson (1983) were calculated for each gene region and the combined data set. Diaporthe eres haplotype networks were constructed using Network v. 5.0 (Bandelt et al., 1999).

Network Analysis
To understand the relationship among different geographical populations, recombination parameters were calculated, and haplotype networks were constructed. In this analysis, the combined dataset of Diaporthe eres isolates from China alone and Chinese isolates combined with European isolates  were used. ZnS, used by Kelly (1997), and the recombination parameters Za and ZZ (Hudson, 1983;Kelly, 1997) were calculated using DnaSP v. 6.12. The haplotype data generated using DnaSP v. 6 were used to construct a median-joining network in Network v. 5.0 (Bandelt et al., 1999).

Pathogenicity Assay
The pathogenicity and aggressiveness of the Diaporthe species were tested using detached green shoots of the V. vinifera cultivar Summer Black. Healthy, 30-50 cm long green shoots (including at least two nodes) were obtained from "Shunyi Xiangyi" vineyard in Beijing, China, where Diaporthe species were not recorded. The cuttings were surface-sterilized with 70% ethanol by wiping with cotton swabs. A shallow wound (5 mm length, 2 mm deep) was made in the center of each shoot using a sterilized scalpel. Mycelial plugs were taken from the growing margin of a 5-day-old culture grown in PDA and inoculated at the wound site. Non-colonized sterile PDA plugs were used for inoculation of shoots as a negative control. To prevent drying, all inoculated areas were covered with Para-film (Bemis, USA). Inoculated shoots were kept in a growth chamber for 21 days at 25 • C with a 12 h photoperiod. The experiment was organized with 10 replicates for each isolate. Pathogenicity test was repeated three times with same controlled environment. A total of 16        strains from eight species were tested. The presence of lesions advancing beyond the original 0.5 cm diameter inoculation point was considered indicative of pathogenicity. The experimental design was completely randomized. Data were analyzed with a one-way ANOVA (analysis of variance) using Minitab v. 16.0 (Minitab Inc., Boston, MA, USA), with statistical significance set at the 5% level. The pathogens were re-isolated to confirm their identity.

Initial Species Identification and Phylogenetic Analyses
During our field survey on six grape-growing provinces in China (Figure 1), we collected samples with typical symptoms associated with Diaporthe dieback, such as wedge-shaped cankers, and light brown streaking of the wood (Figure 2). However, these symptoms are sometimes confused with other grape trunk disease symptoms caused by Botryosphaeria dieback, Eupta, and Esca (Mondello et al., 2018). Hence, further confirmation is required by isolating and identifying causal organisms. One hundred and eleven Diaporthe isolates were initially identified by colony characteristics, such as abundant tufted white aerial mycelia on agar medium. The ITS gene regions were sequenced for all fungi isolated from diseased shoots and compared with those in GenBank using the MegaBLAST tool in GenBank. The isolates showed 95-99% similarity to known Diaporthe species in GenBank, and these closely related known species were included in the phylogenetic analysis.
To understand the taxonomic placements of our isolates, additional gene regions, including those encoding EF-1α, β-tubulin, and CAL, were sequenced. Then, phylogenetic trees were constructed for each individual gene region. The concatenated sequence data set consisted of 94 isolates (out of 111, due to sequencing errors) from the current study (Table 3) and 197 isolates originating from GenBank ( Table 2), with one outgroup taxon, Diaporthella corylina (CBS 121124). A comparison of maximum likelihood (ML) analysis results for each gene region is given in Table 4. In the ML analysis, the resulting tree of the combined data set of ITS, β-tubulin, CAL, and EF-1α genes had the best resolution of taxa (Figure 3). Therefore, in the present study, we used the combined sequence data to understand the taxonomic placements of the Diaporthe species isolated from grapevines in China. A Bayesian analysis resulted in 10,001 trees after 2,000,000 generations. The first 1,000 trees, representing the burn-in phase of the analyses, were discarded, while the remaining 9,001 trees were used for calculating posterior probabilities (PPs) in the majority-rule consensus tree. The dataset consisted of 1,494 characters with 727 constant characters and 1,006 parsimony-informative and 213 parsimony-uninformative characters. The maximum number of trees generated was 1,000, and the most parsimonious trees had a tree length of 9,862 (CI = 0.249, RI = 0.805, RC = 0.201, HI = 0.751).
Etymology-In reference to the Guangxi Province, from where the fungus was first isolated. Holotype-JZBH320094.

Culture Characteristics
Colonies on PDA reach 70 mm diam. after 7 days at 25 • C, producing abundant white aerial mycelia and reverse fuscous black.
Etymology-In reference to the Hubei province, from where the fungus was first isolated.

Culture Characteristics
Colonies on PDA reach 90 mm after 10 days at 25 • C (covers total surface), abundant tufted white aerial mycelia, buff, numerous black pycnidia 0.5 mm in diam. occur in the mycelium, typically in the direction of the edge of the colony; reverse buff with concentric lines.

Material Examined
CHINA Hubei Province, Wuhan, on diseased trunk of V. vinifera, 30 June 2015, X. H Li (JZBH320123, holotype); ex-type living cultures JZB320123. Notes: In phylogenetic analysis, D. hubeiensis was placed in a well-supported clade together with D. alangi (CFCC52556), D. tectonae (MFLUCC 12-0777) and D. tulliensis (BRIP62248b) with 100% ML, 100% MP bootstrap values and 0.99 posterior probabilities. Diaporthe hubeiensis developed sister clade with D. alangi (CFCC52556) with 99% ML, 83% MP bootstrap values and 0.99 posterior probabilities. Morphologically, Diaporthe hubeiensis has smaller conidiophores and smaller conidia (6.1× 1.8 µm) than D. alangi (7 × 2 µm), and it has no beta conidia in D. alangi (Yang et al., 2018). Diaporthe hubeiensis differs from D. tectonae by developing wider but shorter conidia (6.1× 1.8 µm vs 5.5 × 2.6 µm) (Doilom et al., 2017). Compared to D. tulliensis, D. hubeiensis has smaller conidia (6.1× 1.8 µm vs 5.5-6 µm) (Yang et al., 2018). In the ITS sequence comparison between D. hubeiensis and D. alangi, 44.6% of the 461 nucleotides across the ITS (+5.8S) were different. Of the three proteincoding genes, the two species showed 4.26% and 1.16% and 5.3% polymorphic nucleotide site differences for CAL, β-tubulin and EF-1α genes, respectively.  (Figure 3). The current species has a particular close relationship with D. pandanicola . In the original description of D. pandanicola, morphological characteristics were not given (Tibpromma et al., 2018). Therefore, these two species were compared based on only DNA sequence data. ITS sequence comparison between D. viniferae and D. pandanicola revealed that 2.9% of the 478 nucleotide sites across the ITS (+5.8S) regions were different. Similarly, 1.7% of the β-tubulin gene fragment was different. Table 5 summarized the genetic diversity data of D. eres associated with grapevines which were estimated using DnaSP V.6. In the analysis, the combined data set of ITS, β-tubulin, HIS, APN, and CAL gene sequences showed 0.16226 segregation sites per sequence and a haplotype diversity of 0.955. A haplotype network was developed for the D. eres species isolated from China using Network v. 5.0 (Figure 5). The resulting network combining ITS, β-tubulin, HIS, EF-1α, and CAL gene sequences gave two main clusters according to geographic origin. In the network, isolates from Hubei province were clustered into two main clades. A single haplotype (H-11) was clustered within the main Jilin clade. Haplotype 7 (from Hubei) and h-13 (from Sichuan Province) were connected with one intermediate haplotype to the two main clusters.

Genetic Diversity and Population Structure Analysis
To understand the relationship between Diaporthe isolates from Chinese vineyards and those from European vineyards, we calculated recombination parameters Z and ZnS. The combined data set consists of 135 sequences with 2203 sites. The estimate of R per gene was 6.6, and the minimum number of recombination events (Rm) was 15. Median-joining networks were constructed using both single-gene data files and a combined data set of ITS, β-tubulin, HIS, EF-1α, and CAL genes. The singlegene networks differed from each other, and the resulting patterns did not give a significant grouping. Therefore, in this study, only the combined network was considered (Figure 6). A total of 33 haplotypes were identified using DnaSP, and the haplotype data file was used to generate the haplotype network. In the resulting network, we found that Chinese haplotypes and Europe haplotypes were not shared and that there was no sharing of haplotypes among different provinces in China. However, the Chinese haplotypes were dispersed in the combined network, with the majority of isolates from Hubei located in two related clusters surrounded by European haplotypes. Similarly, the haplotypes from Sichuan and Jilin provinces were also dispersed in the network and close to both European and Chinese haplotypes.

Comparative Aggressiveness Among Diaporthe Species
Pathogenicity and aggressiveness among eight Diaporthe species isolated in our study were compared by inoculating them into the V. vinifera cultivar Summer Black. The inoculated shoots did not show significant lesion development within the first 2 weeks after inoculation. Brown necrotic lesions were detected both on the tissue surface and internally, advancing upwards, and downwards through the inoculation point. Twenty-one days after inoculation, D. gulyae developed the largest lesions (1.23 cm), followed by D. eres (0.94 cm). The remaining species, D. unshiuensis, D. viniferae, D. guangxiensis, D. pescicola, and D. sojae, exhibited similar levels of aggressiveness on grape shoots (Figure 7). Diaporthe hubeiensis was the least aggressive (0.5 cm) among the eight species.

DISCUSSION
Grapevine trunk disease has become one of the most devastating grapevine diseases in recent decades. According to data collected worldwide, ∼1.5 billion US dollars per year is spent to replace dead grapevines due to these trunk diseases (Hofstetter et al., 2012;Fontaine et al., 2016). This is a great concern among grapeproducing countries, as the disease infects perennial parts of the vine and reduces the productive lifespan of vines by several years (Gramaje and Armengol, 2011). The disease ultimately affects the sustainability of the wine industry and table grape production (Fontaine et al., 2016). As the world's top grapeproducing country, China has strived to improve the quality and quantity of grapes. Though they are the most important grapevine trunk diseases worldwide, there is no evidence of either the esca complex or Eutypa dieback in China (Fontaine et al., 2016). However, the third most common grapevine trunk disease, caused by the species in Botryosphaeriaceae (Yan et al., 2013(Yan et al., , 2018, has been identified as the leading grapevine trunk pathogen complex in China. Unfortunately, over the last few years, diseases caused by Diaporthe species (Dissanayake et al., 2015a have become the emerging trunk diseases in China. Understanding the diversity of the causative species and the genetic variation within pathogen populations could help in developing sustainable disease management strategies. In addition, understanding the relationships between European and Chinese isolates can help track disease spread, as both regions share similar disease severity and Diaporthe species that differ from those in North America (Fontaine et al., 2016;Úrbez Torres and O'Gorman, 2019). To achieve these objectives, disease surveys were conducted in six provinces. We isolated and identified 111 Diaporthe strains and showed that they belong to eight species. In 1958, D. ampelina (= Phomopsis viticola) was identified infecting green shoots of grapevines (Pscheidt and Pearson, 1989). The disease was named "Phomopsis cane and trunk disease." According to the USDA Fungal-host interaction database, there are 166 records of Diaporthe species associated with grapevines worldwide (https://nt.ars-grin. gov/fungaldatabases/fungushost/fungushost.cfm) (Farr and Rossman, 2019). These records are related to the following 27 Diaporthe species: Diaporthe ambigua (D. ambigua)   Farr and Rossman, 2019), Diaporthe hispaniae (D. hispaniae), D. hongkongensis , Diaporthe hungariae (D. hungariae) , D. kyushuensis (Kajitani and Kanematsu, 2000), D. nebulae  Diaporthe neotheicola (D. neotheicola) (Úrbez-Torres et al., distribution. The second most abundant taxon, D. sojae, has a wide range of hosts as well, including Camptotheca acuminata (C. acuminata) (Chang et al., 2005), Glycine max, Cucumis melo (Lehman, 1923;Santos et al., 2011), Capsicum annuum (C. annuum) (Pennycook, 1989), Stokesia laevis (S. laevis) (Sogonov et al., 2008), and Helianthus annuus (H. annuus) (Thompson et al., 2011). These two Diaporthe species were previously identified and characterized from grapevines in China by Dissanayake et al. (2015a).
The present study recorded three Diaporthe species, D. gulyae, D. pescicola, and D. unshiuensis, associated with Vitis dieback for the first time. Diaporthe gulyae was previously reported on H. annuus in Australia (Thompson et al., 2011), Canada, and the United States (Mathew et al., 2015a,b) and on Carthamus lanatus (C. lanatus) in Italy (Andolfi et al., 2015). Diaporthe pescicola was previously described in association with peach shoot dieback in China . Diaporthe unshiuensis was first described in China in 2015 as an endophyte of a Citrus sp. .
The identification and characterization of novel taxa and new host records is an indication of the high potential of Diaporthe to evolve rapidly. Host switching is often related to fungal adaptive ability (Bleuven and Landry, 2016). The changing environments and human interference present both challenges and opportunities for fungi, with some capable of switching from endophytic or saprobic lifestyles to pathogenic styles or becoming more aggressive and colonizing new hosts (Manawasinghe et al., 2018). The novel taxa and the new records reported here for grapevine trunk diseases in China might be due to these factors. During the past decade, northern China has become significantly warmer (Piao et al., 2010). The increased temperature could attract new pests and disease agents to the region. On the other hand, human-mediated factors can also influence the development of a new disease (McDonals, 2004). For example, in commercial grape vineyards, significant amounts of chemicals are applied annually in the form of pesticides and fungicides (Úrbez-Torres, 2011). Such applications could lead to the development of resistant strains of the target organism and non-target micro-fungi (Manawasinghe et al., 2018). Over time, strains and species that are more resistant and/or more aggressive could emerge. The recent identification of new species and new host records of Diaporthe in China and in Europe are consistent with the hypothesis. Studying the genetic diversity of pathogens provides clues to how host switches might have occurred and the genetic basis for new pathogen emergence.
The knowledge of the genetic diversity of a particular phytopathogen can be used to develop sustainable management strategies such as resistance breeding and fungicide screening. In this study, D. eres was analyzed, as it had a relatively large number of isolates from which to obtain reasonable estimates of various intraspecific diversity indices. In this study, multi-locus sequences were used as the marker of choice. The use of sequence data as genetic markers facilitated the analysis of genetic variations among isolates within a population. We selected ITS, β-tubulin, HIS, EF-1α, and CAL gene regions, as they were extensively used in phylogenetic analysis of the genus Diaporthe. In addition, ACT and Apn2 genes were selected since those regions provide a large number of polymorphic sites for the Diaporthe eres species complex (Udayanga et al., 2014b). Genetic polymorphisms are required for both phylogenetic and population genetic studies (Xu, 2006). Using these gene regions, we calculated haplotype richness (h R ), the total number of haplotypes, Watterson's theta ( w ), and pairwise nucleotide diversity (JI) for Diaporthe eres obtained from Chinese vineyards.
The combined effect of the mutation, recombination, marker ascertainment, and demography of a particular species can be revealed by analyzing and comparing gene genealogies and haplotype diversities within and between genes (Stumpf, 2004;Xu, 2006). The calculated haplotype diversities of Diaporthe eres were higher than 0.5 for Apn2, CAL, HIS, β-tubulin and the combined data, reflecting high genetic diversity. Tajima's D indicates how much population variation can be sustained over time (Tajima, 1989). In the present study, positive D values were observed for coding gene regions (Apn2, CAL, and HIS). This might be due to selective pressure causing a recent population contraction. The selection pressure could have come from the continuous application of fungicides, leading to the loss of certain genotypes. In contrast, Tajima's D for the combined sequences was negative (−0.20416), which indicates a possible recent population expansion of certain multi-locus genotypes (Tajima, 1989). In Hubei, several multi-locus genotypes were over-represented, consistent with this hypothesis.
The Hudson and Kaplan (1985) index for the recombination between Chinese and European isolates was calculated for this study. In our analysis, we calculated the number of recombination events in the history of a sample of sequences (R) and the number of recombination events that can be parsimoniously inferred from a sample of sequences (Rm) (Hudson, 1983;Kelly, 1997). When the rate of recombination equals zero, R gives zero (Hudson, 1983;Hudson and Kaplan, 1985). Since the R is given a value based on the history of the sample, Rm denotes the minimum number of recombination events implied by the data using the four-gamete test. A positive ZZ value, which reflects intragenic recombination, has played an important role in nucleotide variation and a high number of recombination events (Hudson, 1983). Therefore, we can conclude that recent recombination events might have occurred between the Chinese and European isolates. Haplotype networks provide a better understanding of the coexistence of ancestral and derived haplotypes by providing an account for recombination (Huson and Bryant, 2006). Therefore, haplotype networks are intensively used in intraspecific analyses. We used a median-joining network in which the number of mutations separate haplotypes (Castelloe and Templeton, 1994). In each network, the ancestral haplotype was predicted based on rooting probability (Posada and Crandall, 2001). The analyses suggested that the most recent ancestry of the Chinese haplotypes was shared with the Spanish and Hungarian haplotypes. In addition, haplotypes from the UK and Czech Republic shared ancestry with Chinese haplotypes. Overall, the Diaporthe population in China is genetically diverse and might have an admixture population. The current population is likely derived from a combination of endemic D. eres strains and introduced strains from other regions.

CONCLUSION
Present study provides an account of Diaporthe species associated with Chinese vineyards by their phylogenetic placements. Collectively, in the present study, 111 Diaporthe strains were isolated and characterized into eight species using both morphological and molecular phylogenetic approaches. To identify those taxa, four gene regions were examined. The combination of ITS, CAL, β-tubulin, and EF-1α genes gave the best species delimitation in the genus Diaporthe. The present study introduced three novel taxa and three host records of Diaporthe associated with Chinese grapevines. The most abundant Diaporthe species was D. eres, which was moderately aggressive. D. gulyae was the most aggressive among the eight species on detached green shoots. The Chinese D. eres population was high in nucleotide diversity and haplotype diversity. In haplotype network analysis, the Chinese population was dispersed in the network but showed a certain degree of clustering according to their geographical origins. This result suggests that there is likely geographic structuring of D. eres in China. However, more in-depth analysis is required using more isolates from different provinces. Haplotype networks including Chinese and European isolates suggest a close relationship between the two populations. This is confirmed by the recombination among isolates from these two regions. Our results suggest that the D. eres population in China might be a result of an admixture. The results presented here provide opportunities for several fields, including grapevine breeding for disease-resistant cultivars, screening for new fungicides, and developing appropriate quarantine and management strategies to prevent and control grapevine dieback diseases.

DATA AVAILABILITY
The sequence data generated in this study is deposited in NCBI GenBank (https://www.ncbi.nlm.nih.gov/genbank) and the respective accession numbers are given in Table 2. The Alignment generated in the present study available in TreeBASE (https://treebase.org/treebase-web/home.html) under the 24324.

AUTHOR CONTRIBUTIONS
JY and XL conceived the research. JY, IM, AD, XL, and WZ planned the basic research. ML, YZ, and WSZ provided materials. IM and AD conducted the experiments and prepared manuscript. IM, AD, DW, and JX analyzed data. KH, SB, and JY revised the manuscript. All authors read and approved the final manuscript.