Convergent evolution of phytopathogenic pseudomonads onto hazelnut

Comparative Genomics of Prunus-Associated Members of the Pseudomonas syringae Species Complex Reveals Traits Supporting Co-evolution and Host Adaptation

Members of the Pseudomonas syringae species complex cause symptoms that are ranging from leaf spots to cankers on a multitude of plant species, including some of the genus Prunus. To date, a total of two species of the P. syringae species complex and six different pathovars have been associated with diseases on Prunus spp., which were shown to belong to different phylogenetic units (phylogroups, PG) based on sequence similarity of housekeeping genes or whole genomes, suggesting that virulence to Prunus spp. may be the result of convergent pathoadaptation. In this study, a comparative genomics approach was used to determine genes significantly associated with strains isolated from Prunus spp. across a phylogeny of 97 strains belonging to the P. syringae species complex. Our study revealed the presence of a set of orthologous proteins which were significantly associated with strains isolated from Prunus spp. than in strains isolated from other hosts or from non-agricultural environments. Among them, the type III effector HopAY predicted to encode for a C58 cysteine protease was found to be highly associated with strains isolated from Prunus spp. and revealed patterns supporting co-evolution and host adaptation.

Recombination of ecologically and evolutionarily significant loci maintains genetic cohesion in the Pseudomonas syringae species complex

Background

Pseudomonas syringae is a highly diverse bacterial species complex capable of causing a wide range of serious diseases on numerous agronomically important crops. We examine the evolutionary relationships of 391 agricultural and environmental strains using whole-genome sequencing and evolutionary genomic analyses.

Results

We describe the phylogenetic distribution of all 77,728 orthologous gene families in the pan-genome, reconstruct the core genome phylogeny using the 2410 core genes, hierarchically cluster the accessory genome, identify the diversity and distribution of type III secretion systems and their effectors, predict ecologically and evolutionary relevant loci, and establish the molecular evolutionary processes operating on gene families. Phylogenetic and recombination analyses reveals that the species complex is subdivided into primary and secondary phylogroups, with the former primarily comprised of agricultural isolates, including all of the well-studied P. syringae strains. In contrast, the secondary phylogroups include numerous environmental isolates. These phylogroups also have levels of genetic diversity typically found among distinct species. An analysis of rates of recombination within and between phylogroups revealed a higher rate of recombination within primary phylogroups than between primary and secondary phylogroups. We also find that “ecologically significant” virulence-associated loci and “evolutionarily significant” loci under positive selection are over-represented among loci that undergo inter-phylogroup genetic exchange.

Conclusions

While inter-phylogroup recombination occurs relatively rarely, it is an important force maintaining the genetic cohesion of the species complex, particularly among primary phylogroup strains. This level of genetic cohesion, and the shared plant-associated niche, argues for considering the primary phylogroups as a single biological species.

Electronic supplementary material

The online version of this article (10.1186/s13059-018-1606-y) contains supplementary material, which is available to authorized users.

Some strains that have converged to infect Prunus spp. trees are members of distinct Pseudomonas syringae genomospecies and ecotypes as revealed by in silico genomic comparison

A complementary taxonomic and population genetic study was performed to delineate genetically and ecologically distinct species within the Pseudomonas syringae complex by assessing 16 strains including pathovar strains that have converged to infect Prunus spp. trees, and two outgroups. Both average nucleotide identity and genome-to-genome distance comparison methods revealed the occurrence of distinct genomospecies, namely 1, 2, 3 and 8 (sensu Gardan et al.), with the latter two being closely related. Strains classified as P. s. pv. morsprunorum clustered into two distinct genomospecies, namely 2 and 8. Both the AdaptML and hierarchical Bayesian analysis of population structure methods highlighted the presence of three ecotypes, and the taxonomically related genomospecies 3 and 8 strains were members of the same ecotype. The distribution of pathogenic and virulence-associated genetic traits among Pseudomonas strains did not reveal any distinct type III secretion system effector or phytotoxin distribution pattern that characterized single genomospecies and strains that infect Prunus spp. The complete WHOP (Woody HOst and Pseudomonas spp.) genomic region and the entire β-ketoadipate gene cluster, including the catBCA operon, were found only in the members of genomospecies 2 and in the two P. s. pv. morsprunorum strains of genomospecies 8. A reduced gene flow between the three ecotypes suggested that point mutations played a larger role during the evolution of the strains than recombination. Our data support the idea that Prunus trees can be infected by different strains of distinct Pseudomonas genomospecies/ecotypes through diverse mechanisms of host colonization and infection. Such strains may represent particular lineages that emerged from environments other than that of the infected plant upon acquiring genetic traits that gave them the ability to cause plant diseases. The complementary assessment of bacterial strains using both taxonomic approaches and methods that reveal ecologically homogeneous populations has proven useful in confirming the cohesion of bacterial clusters.

Extensive Field Survey, Laboratory and Greenhouse Studies Reveal Complex Nature of Pseudomonas syringae-Associated Hazelnut Decline in Central Italy

Pseudomonas avellanae (Pav) has been reported as the causal agent of bacterial decline and bacterial canker of hazelnut in Italy and Greece, respectively. Both hazelnut diseases were reported to be similar in terms of symptoms, severity and persistence. In this study, we found that both symptomatic and asymptomatic trees in the field were colonized by Pav. Multilocus Sequence Typing (MLST) analysis showed that Pav strains isolated during this study in Italy belong to the P. syringae phylogroup 1 and they are closely related to Pav strains previously isolated in Greece from hazelnut bacterial canker. On the other hand, strains isolated in earlier studies from hazelnut decline in Italy belong to both phylogroup 1 and 2 of P. syringae. Both phylogroup 1 strains of P. syringae from Greece and Italy are different than strains isolated in this study in terms of their capacity to excrete fluorescent pigments on different media. Despite the same plant genotype and cropping practices adopted, the incidence of hazelnut decline ranged from nearly 0 to 91% across our study sites. No disease developed on plants inoculated with Pav through wounding while leaf scar inoculations produced only mild disease symptoms. Based on our results and the previously reported correlation between pedo-climatic conditions and hazelnut decline, we conclude that hazelnut decline in central Italy could be incited by a combination of predisposing (adverse pedo-climatic conditions) and contributing factors (Pav). Because this is a true decline different from “bacterial canker” described in Greece, we refer to it as hazelnut decline (HD).

Comparative Genomic Analyses of Multiple Pseudomonas Strains Infecting Corylus avellana Trees Reveal the Occurrence of Two Genetic Clusters with Both Common and Distinctive Virulence and Fitness Traits

The European hazelnut (Corylus avellana) is threatened in Europe by several pseudomonads which cause symptoms ranging from twig dieback to tree death. A comparison of the draft genomes of nine Pseudomonas strains isolated from symptomatic C. avellana trees was performed to identify common and distinctive genomic traits. The thorough assessment of genetic relationships among the strains revealed two clearly distinct clusters: P. avellanae and P. syringae. The latter including the pathovars avellanae, coryli and syringae. Between these two clusters, no recombination event was found. A genomic island of approximately 20 kb, containing the hrp/hrc type III secretion system gene cluster, was found to be present without any genomic difference in all nine pseudomonads. The type III secretion system effector repertoires were remarkably different in the two groups, with P. avellanae showing a higher number of effectors. Homologue genes of the antimetabolite mangotoxin and ice nucleation activity clusters were found solely in all P. syringae pathovar strains, whereas the siderophore yersiniabactin was only present in P. avellanae. All nine strains have genes coding for pectic enzymes and sucrose metabolism. By contrast, they do not have genes coding for indolacetic acid and anti-insect toxin. Collectively, this study reveals that genomically different Pseudomonas can converge on the same host plant by suppressing the host defence mechanisms with the use of different virulence weapons. The integration into their genomes of a horizontally acquired genomic island could play a fundamental role in their evolution, perhaps giving them the ability to exploit new ecological niches.

Definition of Plant-Pathogenic Pseudomonas Genomospecies of the Pseudomonas syringae Complex Through Multiple Comparative Approaches

A total of 34 phytopathogenic strain genomes belonging to the Pseudomonas syringae species complex and related species, including many pathotype strains, were assessed using average nucleotide identity (ANI) analysis. Their taxonomic relationships were consistently confirmed by the tetranucleotide frequency correlation coefficient (TETRA) values, multilocus sequence typing analysis (MLSA) performed with seven housekeeping genes, using both maximum likelihood and Bayesian methods, and split consensus network analyses. The ANI, MLSA, and split consensus analyses provided consistent and identical results. We confirmed the occurrence of the well-demarcated genomospecies inferred sensu Gardan et al. using DNA-DNA hybridization and ribotyping analyses. However, some P. syringae strains of the pathovars morsprunorum and lachrymans were placed in different genomospecies in our analyses. Genomospecies 1, 2, 4, 6, and 9 resulted well demarcated, whereas strains of genomospecies 3 and 8 had ANI values between 95 and 96% in some cases, confirming that this threshold reveals very closely related species that might represent cases of splitting entities or the convergence of different species to the same ecological niche. This study confirms the robustness of the combination of genomic and phylogenetic approaches in revealing taxonomic relationships among closely related bacterial strains and provides the basis for a further reliable demarcation of the phytopathogenic Pseudomonas species. Within each species, the pathovars might represent distinct ecological units. The possibility of performing extensive and standardized host range and phenotypic tests with many strains of different pathovars can assist phytobacteriologists for better determining the boundaries of these ecological units.

Harnessing population genomics to understand how bacterial pathogens emerge, adapt to crop hosts, and disseminate

Crop diseases emerge without warning. In many cases, diseases cross borders, or even oceans, before plant pathologists have time to identify and characterize the causative agents. Genome sequencing, in combination with intensive sampling of pathogen populations and application of population genetic tools, is now providing the means to unravel how bacterial crop pathogens emerge from environmental reservoirs, how they evolve and adapt to crops, and what international and intercontinental routes they follow during dissemination. Here, we introduce the field of population genomics and review the population genomics research of bacterial plant pathogens over the past 10 years. We highlight the potential of population genomics for investigating plant pathogens, using examples of population genomics studies of human pathogens. We also describe the complementary nature of the fields of population genomics and molecular plant-microbe interactions and propose how to translate new insights into improved disease prevention and control.

A Genomic redefinition of Pseudomonas avellanae species

The circumscription of bacterial species is a complex task. So far, DNA-DNA hybridization (DDH), 16S rRNA gene sequencing, and multiocus sequence typing analysis (MLSA) are currently the preferred techniques for their genetic determination. However, the average nucleotide identity (ANI) analysis of conserved and shared genes between two bacterial strains based on the pair-wise genome comparisons, with support of the tetranucleotide frequency correlation coefficients (TETRA) value, has recently been proposed as a reliable substitute for DDH. The species demarcation boundary has been set to a value of 95-96% of the ANI identity, with further confirmation through the assessment of the corresponding TETRA value. In this study, we performed a genome-wide MLSA of 14 phytopathogenic pseudomonads genomes, and assessed the ANI and TETRA values of 27 genomes, representing seven out of the nine genomospecies of Pseudomonas spp. sensu Gardan et alii, and their phylogenetic relationships using maximum likelihood and Bayesian approaches. The results demonstrate the existence of a well demarcated genomic cluster that includes strains classified as P. avellanae, P. syringae pv. theae, P. s. pv. actinidiae and one P. s. pv. morsprunorum strain all belonging to the single species P. avellanae. In addition, when compared with P. avellanae, five strains of P. s. pv. tomato, including the model strain DC3000, and one P. s. pv. lachrymans strain, appear as very closely related to P. avellanae, with ANI values of nearly 96% as confirmed by the TETRA analysis. Conversely, one representative strain, previously classified as P. avellanae and isolated in central Italy, is a genuine member of the P. syringae species complex and can be defined as P. s. pv. avellanae. Currently. The core and pan genomes of P. avellanae species consist of 3,995 and 5,410 putative protein-coding genes, respectively.

Pseudomonas syringae pv. actinidiae from recent outbreaks of kiwifruit bacterial canker belong to different clones that originated in China

A recently emerged plant disease, bacterial canker of kiwifruit (Actinidia deliciosa and A. chinensis), is caused by Pseudomonas syringae pv. actinidiae (PSA). The disease was first reported in China and Japan in the 1980s. A severe outbreak of PSA began in Italy in 2008 and has spread to other European countries. PSA was found in both New Zealand and Chile in 2010. To study the evolution of the pathogen and analyse the transmission of PSA between countries, genomes of strains from China and Japan (where the genus Actinidia is endemic), Italy, New Zealand and Chile were sequenced. The genomes of PSA strains are very similar. However, all strains from New Zealand share several single nucleotide polymorphisms (SNPs) that distinguish them from all other PSA strains. Similarly, all the PSA strains from the 2008 Italian outbreak form a distinct clonal group and those from Chile form a third group. In addition to the rare SNPs present in the core genomes, there is abundant genetic diversity in a genomic island that is part of the accessory genome. The island from several Chinese strains is almost identical to the island present in the New Zealand strains. The island from a different Chinese strain is identical to the island present in the strains from the recent Italian outbreak. The Chilean strains of PSA carry a third variant of this island. These genomic islands are integrative conjugative elements (ICEs). Sequencing of these ICEs provides evidence of three recent horizontal transmissions of ICE from other strains of Pseudomonas syringae to PSA. The analyses of the core genome SNPs and the ICEs, combined with disease history, all support the hypothesis of an independent Chinese origin for both the Italian and the New Zealand outbreaks and suggest the Chilean strains also originate from China.

The mangotoxin biosynthetic operon (mbo) is specifically distributed within Pseudomonas syringae genomospecies 1 and was acquired only once during evolution

Mangotoxin production was first described in Pseudomonas syringae pv. syringae strains. A phenotypic characterization of 94 P. syringae strains was carried out to determine the genetic evolution of the mangotoxin biosynthetic operon (mbo). We designed a PCR primer pair specific for the mbo operon to examine its distribution within the P. syringae complex. These primers amplified a 692-bp DNA fragment from 52 mangotoxin-producing strains and from 7 non-mangotoxin-producing strains that harbor the mbo operon, whereas 35 non-mangotoxin-producing strains did not yield any amplification. This, together with the analysis of draft genomes, allowed the identification of the mbo operon in five pathovars (pathovars aptata, avellanae, japonica, pisi, and syringae), all of which belong to genomospecies 1, suggesting a limited distribution of the mbo genes in the P. syringae complex. Phylogenetic analyses using partial sequences from housekeeping genes differentiated three groups within genomospecies 1. All of the strains containing the mbo operon clustered in groups I and II, whereas those lacking the operon clustered in group III; however, the relative branching order of these three groups is dependent on the genes used to construct the phylogeny. The mbo operon maintains synteny and is inserted in the same genomic location, with high sequence conservation around the insertion point, for all the strains in groups I and II. These data support the idea that the mbo operon was acquired horizontally and only once by the ancestor of groups I and II from genomospecies 1 within the P. syringae complex.

Extensive remodeling of the Pseudomonas syringae pv. avellanae type III secretome associated with two independent host shifts onto hazelnut

BACKGROUND

Hazelnut (Corylus avellana) decline disease in Greece and Italy is caused by the convergent evolution of two distantly related lineages of Pseudomonas syringae pv. avellanae (Pav). We sequenced the genomes of three Pav isolates to determine if their convergent virulence phenotype had a common genetic basis due to either genetic exchange between lineages or parallel evolution.

RESULTS

We found little evidence for horizontal transfer (recombination) of genes between Pav lineages, but two large genomic islands (GIs) have been recently acquired by one of the lineages. Evolutionary analyses of the genes encoding type III secreted effectors (T3SEs) that are translocated into host cells and are important for both suppressing and eliciting defense responses show that the two Pav lineages have dramatically different T3SE profiles, with only two shared putatively functional T3SEs. One Pav lineage has undergone unprecedented secretome remodeling, including the acquisition of eleven new T3SEs and the loss or pseudogenization of 15, including five of the six core T3SE families that are present in the other Pav lineage. Molecular dating indicates that divergence within both of the Pav lineages predates their observation in the field. This suggest that both Pav lineages have been cryptically infecting hazelnut trees or wild relatives for many years, and that the emergence of hazelnut decline in the 1970s may have been due to changes in agricultural practice.

CONCLUSIONS

These data show that divergent lineages of P. syringae can converge on identical disease etiology on the same host plant using different virulence mechanisms and that dramatic shifts in the arsenal of T3SEs can accompany disease emergence.

Specificity of Pseudomonas isolates on healthy and Fusarium head blight-infected spikelets of wheat heads

The specificity of culturable bacteria on healthy and Fusarium head blight (FHB)-infected spikelets of wheat heads was investigated to find a candidate of biocontrol agents against FHB. The bacterial genus Pseudomonas was commonly isolated from the tissues, and phylogenetic analysis using 16S ribosomal RNA gene sequences of isolates of the genera revealed that particular phylogenetic groups in the genus specifically inhabited either healthy or infected spikelet tissues. The specificity of each group was suggested to be due to differences in the ability to form biofilms and colonize spikelet tissues; isolates originated from healthy spikelets formed biofilms on polyvinyl chloride microplate wells and highly colonized the spikelet tissues. Other bacterial groups obtained from FHB-infected spikelets less formed biofilms and attached with low densities on the spikelet tissues. Their colonization on the tissues, however, was promoted when co-inoculated with the causal pathogenic fungus, Fusarium graminearum, and several isolates were observed to smash the mycelia in vivo. Moreover, based on results of in vitro mycelial growth inhibition activity, the diseased tissue-originated isolates were verified to have a negative effect on the fungal growth. These results suggest that Pseudomonas isolates obtained from infected spikelet tissues were highly associated with the FHB pathogen and have potential as candidates for biological control against FHB.

High-resolution melting analysis as a powerful tool to discriminate and genotype Pseudomonas savastanoi pathovars and strains

Pseudomonas savastanoi is a serious pathogen of Olive, Oleander, Ash, and several other Oleaceae. Its epiphytic or endophytic presence in asymptomatic plants is crucial for the spread of Olive and Oleander knot disease, as already ascertained for P. savastanoi pv. savastanoi (Psv) on Olive and for pv. nerii (Psn) on Oleander, while no information is available for pv. fraxini (Psf) on Ash. Nothing is known yet about the distribution on the different host plants and the real host range of these pathovars in nature, although cross-infections were observed following artificial inoculations. A multiplex Real-Time PCR assay was recently developed to simultaneously and quantitatively discriminate in vitro and in planta these P. savastanoi pathovars, for routine culture confirmation and for epidemiological and diagnostical studies. Here an innovative High-Resolution Melting Analysis (HRMA)-based assay was set up to unequivocally discriminate Psv, Psn and Psf, according to several single nucleotide polymorphisms found in their Type Three Secretion System clusters. The genetic distances among 56 P. savastanoi strains belonging to these pathovars were also evaluated, confirming and refining data previously obtained by fAFLP. To our knowledge, this is the first time that HRMA is applied to a bacterial plant pathogen, and one of the few multiplex HRMA-based assays developed so far. This protocol provides a rapid, sensitive, specific tool to differentiate and detect Psv, Psn and Psf strains, also in vivo and against other related bacteria, with lower costs than conventional multiplex Real-Time PCR. Its application is particularly suitable for sanitary certification programs for P. savastanoi, aimed at avoiding the spreading of this phytopathogen through asymptomatic plants.

Multilocus sequence typing of Xylella fastidiosa causing Pierce's disease and oleander leaf scorch in the United States

Using a modified multilocus sequence typing (MLST) scheme for the bacterial plant pathogen Xylella fastidiosa based on the same seven housekeeping genes employed in a previously published MLST, we studied the genetic diversity of two subspecies, X. fastidiosa subsp. fastidiosa and X. fastidiosa subsp. sandyi, which cause Pierce's disease and oleander leaf scorch, respectively. Typing of 85 U.S. isolates (plus one from northern Mexico) of X. fastidiosa subsp. fastidiosa from 15 different plant hosts and 21 isolates of X. fastidiosa subsp. sandyi from 4 different hosts in California and Texas supported their subspecific status. Analysis using the MLST genes plus one cell-surface gene showed no significant genetic differentiation based on geography or host plant within either subspecies. Two cases of homologous recombination (with X. fastidiosa subsp. multiplex, the third U.S. subspecies) were detected in X. fastidiosa subsp. fastidiosa. Excluding recombination, MLST site polymorphism in X. fastidiosa subsp. fastidiosa (0.048%) and X. fastidiosa subsp. sandyi (0.000%) was substantially lower than in X. fastidiosa subsp. multiplex (0.240%), consistent with the hypothesis that X. fastidiosa subspp. fastidiosa and sandyi were introduced into the United States (probably just prior to 1880 and 1980, respectively). Using whole-genome analysis, we showed that MLST is more effective at genetic discrimination at the specific and subspecific level than other typing methods applied to X. fastidiosa. Moreover, MLST is the only technique effective in detecting recombination.

Pseudomonas cannabina pv. cannabina pv. nov., and Pseudomonas cannabina pv. alisalensis (Cintas Koike and Bull, 2000) comb. nov., are members of the emended species Pseudomonas cannabina (ex Sutic & Dowson 1959) Gardan, Shafik, Belouin, Brosch, Grimont & Grimont 1999

Sequence similarity in the 16S rDNA gene confirmed that crucifer pathogen Pseudomonas syringae pv. alisalensis belongs to P. syringae sensu lato. In reciprocal DNA/DNA hybridization experiments, DNA relatedness was high (69-100%) between P. syringae pv. alisalensis strains and the type strain of P. cannabina (genomospecies 9). In contrast, DNA relatedness was low (below 48%) between P. syringae pv. alisalensis and reference strains from the remaining genomospecies of P. syringae including the type strain of P. syringae and reference strain of genomospecies 3 (P. syringae pv. tomato) although the well-known crucifer pathogen, P. syringae pv. maculicola, also belongs to genomospecies 3. Additional evidence that P. syringae pv. alisalensis belongs to P. cannabina was sequence similarity in five gene fragments used in multilocus sequence typing, as well as similar rep-PCR patterns when using the BOX-A1R primers. The description of P. cannabina has been emended to include P. syringae pv. alisalensis. Host range testing demonstrated that P. syringae pv. alisalensis strains, originally isolated from broccoli, broccoli raab or arugula, were not pathogenic on Cannabis sativa (family Cannabinaceae). Additionally, P. cannabina strains, originally isolated from the C. sativa were not pathogenic on broccoli raab or oat while P. syringae pv. alisalensis strains were pathogenic on these hosts. Distinct host ranges for these two groups indicate that P. cannabina emend. consists of at least two distinct pathovars, P. cannabina pv. cannabina pv. nov., and P. cannabina pv. alisalensis comb. nov. Pseudomonas syringae pv. maculicola strain CFBP 1637 is a member of P. cannabina.

PAMDB, a multilocus sequence typing and analysis database and website for plant-associated microbes

Although there are adequate DNA sequence differences among plant-associated and plant-pathogenic bacteria to facilitate molecular approaches for their identification, identification at a taxonomic level that is predictive of their phenotype is a challenge. The problem is the absence of a taxonomy that describes genetic variation at a biologically relevant resolution and of a database containing reference strains for comparison. Moreover, molecular evolution, population genetics, ecology, and epidemiology of many plant-pathogenic and plant-associated bacteria are still poorly understood. To address these challenges, a database with web interface was specifically designed for plant-associated and plant-pathogenic microorganisms. The Plant-Associated Microbes Database (PAMDB) comprises, thus far, data from multilocus sequence typing and analysis (MLST/MLSA) studies of Acidovorax citrulli, Pseudomonas syringae, Ralstonia solanacearum, and Xanthomonas spp. Using data deposited in PAMDB, a robust phylogeny of Xanthomonas axonopodis and related bacteria has been inferred, and the diversity existing in the Xanthomonas genus and in described Xanthomonas spp. has been compared with the diversity in P. syringae and R. solanacearum. Moreover, we show how PAMDB makes it easy to distinguish between different pathogens that cause almost identical diseases. The scalable design of PAMDB will make it easy to add more plant pathogens in the future.

Variability of the 16S-23S rRNA gene internal transcribed spacer in Pseudomonas avellanae strains

The 16S-23S rRNA gene internal transcribed spacer region (ITS1) from 34 strains of Pseudomonas avellanae and some strains of Pseudomonas syringae pathovars was amplified and assessed by restriction fragment length polymorphism (RFLP) using 10 restriction enzymes. In addition, the ITS1 region of four representative P. avellanae strains was sequenced and compared by the neighbour-joining algorithm with that of P. syringae pathovars. Two main groups of P. avellanae strains were observed that did not correlate with their origin. The ITS1 region sequencing revealed a high similarity with the P. syringae complex. One group of P. avellanae strains showed high similarity to P. s. pv. actinidiae and P. s. pv. tomato; another group showed similarity with P. s. pv. tabaci and P. s. pv. glycinea. Two strains clustered with P. s. pv. pisi. The difficulties to unambiguously classify the strains associated with hazelnut decline in Greece and Italy are discussed.