A new Phytophthora species was detected (i) in the USA, infecting foliage of Kalmia latifolia, (ii) in substrate underneath Pieris, and (iii) in Germany in soil samples underneath Aesculus hippocastanum showing disease symptoms. The new species Phytophthora obscura sp. nov. is formally named based on phylogenetic analysis, host range, Koch’s postulates and morphology. Phytophthora obscura is homothallic with paragynous antheridia and semipapillate sporangia. It is genetically closely related to P. syringae and P. austrocedrae and together these three species define a new Phytophthora subclade 8d, with significant support for all genetic loci analysed including seven nuclear genes and the mitochondrial gene coxII. The morphological and ecological characteristics are very similar to P. syringae, and it is likely that P. obscura was not described earlier because it was identified as P. syringae. Artificial inoculations indicated that horse chestnut, kalmia, pieris and rhododendron might be hosts, and Koch’s postulates were confirmed for kalmia from which it was isolated. This pathogen was named after its elusive nature since it has to date rarely been detected in the US and Germany.
The genus Phytophthora is well known for repeated introductions of exotic species or strains that pose significant plant biosecurity threats. One notable introduction is that of the sudden oak death pathogen P. ramorum affecting oak trees and many woody ornamentals (Werres et al., 2001; Grünwald et al., 2008). It is now evident that this pathogen has undergone at least four global migrations and was introduced into North America at least three times (Prospero et al., 2007, 2009; Mascheretti et al., 2008, 2009; Goss et al., 2009a,b, 2011). Recently, P. ramorum has been found to cause unexpected, landscape-scale dieback of Japanese larch plantations in the UK (Brasier & Webber, 2010). With continued globalization and relaxation of trade barriers, the movement of exotic pathogens to new regions and hosts is likely to accelerate the emergence of another high impact pathogen such as P. ramorum.
This paper describes another newly emerging Phytophthora species recovered to date in Germany and the United States. This novel species was occasionally found in the Werres and Grünwald culture collections and was originally classified as P. syringae based on morphology (Werres et al., 1995) and ITS sequencing (Grünwald et al., 2011). However, preliminary amplified fragment length polymorphism (AFLP) analysis indicated that these isolates belonged to a unique cluster. In Germany the isolates originated from soil underneath a diseased horse chestnut (Aesculus hippocastanum) grown in an allée, displaying a bleeding canker (Werres et al., 1995). Isolation of a pathogen from these cankers was attempted but unsuccessful. In Oregon, USA the new species was detected in substrate underneath diseased Pieris japonica and was isolated from a leaf of a diseased Kalmia latifolia showing typical foliar blight symptoms similar to P. syringae.
The objective of this study is to formally describe the new species Phytophthora obscura sp. nov. based on extensive nuclear and mitochondrial DNA sequencing, AFLP, morphological characterization, and host range analysis including the closest relatives P. syringae and P. austrocedrae. Discovery of P. obscura has also led to the definition of a new subclade 8d in the genus Phytophthora with significant phylogenetic support based on DNA sequences from nuclear and mitochondrial loci.
Materials and methods
Cultures and morphological work
All isolates used in this work are listed in Table 1. Isolates of the new species from single zoospore culture were obtained from the Phytophthora culture collections in the Grünwald and Werres laboratories. In the Grünwald laboratory, cultures were maintained on V8 agar (100 mL V8 juice; 2 g CaCO3; 30 mg L−1β-sitosterol (EMD Chemicals, Inc.); 15 g agar; 900 mL deionized water). In the Werres laboratory, isolates were stored on oatmeal agar under paraffin oil at 15°C and refreshed on carrot piece agar (see below).
Table 1. Phytophthora isolates used in this study
Year of isolation
aOriginal isolate name in parentheses; isolate obtained from Paul Tooley, USDA Foreign Disease-Weed Science Research Unit, USA. This isolate was originally sampled by R. G. Linderman.
bOriginal isolate name in parentheses; isolate obtained from Everett Hansen, Oregon State University, USA.
cOriginal isolate name in parentheses; isolate obtained from Lani Yakabe, University of California, USA.
P.obscura sp. nov.
Soil (Aesculus hippocastanum)
P. obscura sp. nov.
BBA 2/94-IIB (Type strain)
Soil (Aesculus hippocastanum)
P.obscura sp. nov.
Leafy debris underneath Pieris sp.
P. obscura sp. nov.
Kalmia latifolia leaf
Soil (Quercus sp.)
Prunus domestica cv. Top
Morphology was determined on four media: carrot piece agar (CPA) with 50 g carrot pieces (Daucus carota cv. Parano, HILD) and 15 g agar per L deionized water; potato dextrose agar (PDA) with 39 g PDA (BD-Diagnostics no. 214030) per L deionized water; oat meal agar (OMA) with 40 g oat meal and 15 g agar per L deionized water; and V8 agar (V8A) prepared with V8 juice (Campbells), 1 g CaCO3 and 15 g agar per L deionized water. To establish temperature response, isolates were grown in four replicates at 2, 5, 10, 15, 20, 25, 28, 30, 32, 35 and 37°C on CPA, PDA, OMA and V8A. For morphological measurements, such as spore size, isolates were grown on CPA as described in Werres et al. (2001). For each isolate, 100 sporangia, oogonia, oospores and antheridia were measured with the exception of P. syringae (only isolate BBA 12/99-5), where only 40 gametangia could be measured due to low abundance. For the calculation of the morphological data the isolate P-09-011 was excluded because it appeared damaged under the microscope, possibly due to lengthy storage in liquid nitrogen. A representative isolate, BBA 2/94-IIB, has been deposited as holotype with the Centraalbureau voor Schimmelcultures (CBS 129273) in Utrecht, The Netherlands. The new species was also submitted to Mycobank (MB 519834).
Refreshment of isolates
With the exception of P. austrocedrae where apples could not be infected, all isolates were refreshed for infection studies using apple cv. Boskoop. Two apples per isolate were inoculated via two holes each (diameter 0·8 cm, depth c. 1 cm). Each hole was filled with an agar plug from a growing colony on CPA. The incubation holes were closed with apple tissue and wrapped with Parafilm for 2 days to prevent drying. The apples were incubated in wet chambers at 15°C (12 h day) and 10°C (12 h night) until development of discoloration, but no longer than 6 weeks. For reisolation, tissue pieces from the edge of apple tissue showing symptoms were cut out and placed on CPA. If no symptoms developed, tissue pieces from the symptomless apple tissue were placed on CPA 6 weeks after inoculation.
Pathogenicity studies were carried out with A. hippocastanum, K. latifolia, Pi. japonica and Rhododendron catawbiense cv. Boursault. For the pathogenicity studies with horse chestnut, 5-year -old seedlings of A. hippocastanum were obtained. The seedlings originated from a nursery but were cultivated for 4 years without chemical control at JKI before inoculation. In October 2009, four containerized horse chestnuts were inoculated per isolate at four different points each, two inoculation points in older stem parts and two in younger ones. In November 2009 the trees were inoculated at two further points. Inoculation was similar to grafting. Instead of the scion wood, a plug from the edge of a growing Phytophthora colony was placed on the fresh wound. The inoculation points were sealed with Parafilm for 6 days. Due to safety concerns, the horse chestnuts were kept in the greenhouse. Incubation temperatures were similar to outdoor temperatures but remained above 0°C. Occurrence of disease symptoms were observed monthly during the season. In June 2010 reisolations were carried out onto CPA plates for confirmation of identity.
Whole plant, non-wounded dip inoculations were carried out on 10 plants of each of Pi. japonica, K. latifolia and R. catawbiense cv. Boursault. Five isolates were selected, four of P. obscura sp. nov. and one of P. syringae, for comparison with these plant species. Small plants in 4-inch containers were inoculated by dipping the above ground portion of the plant into a zoospore suspension (10 000 zoospores mL−1), misted and bagged for 48 h to encourage infection. Plants were incubated in a containment chamber at 20°C with 16 h day length. Following successful lesion development, re-isolations were carried out onto V8A and PARP medium plates for confirmation of identity.
Amplified fragment length polymorphism (AFLP) genotyping
AFLP was performed on genomic DNA using the AFLP Microbial Fingerprinting protocol (Applied Biosystems (ABI)) with slight modifications as described previously (Grünwald & Hoheisel, 2006). Total genomic DNA was obtained from mycelia using the FastDNA SPIN kit (MP Biomedicals LLC). DNA (20 ng) was digested, adaptors were ligated, and preselective amplifications with EcoRI core primer, MseI core primer and AFLP amplification core mix were performed on a GeneAmp 9700 thermal cycler (ABI) as described. Pre-amplified samples (10 μL) were diluted with 190 μL TE buffer and selective PCR was performed with MseI-AC/EcoRI-AC and MseI-CC/EcoRI-AA primer combinations. Products (0·5 μL) were run with 25 μL loading buffer (24 μL deionized formamide and 0·05 μL GeneScan-500 size standard, ABI) on a capillary sequencer (ABI Prism 3100). Electropherograms were analysed using GeneMapper software (v.3·7; ABI) to extract a matrix of presence and absence of alleles in a range of 50–500 bp after visual inspection. All AFLP analyses were replicated two or more times including independent DNA extractions and PCR reactions, until all polymorphic alleles could be unambiguously binned.
Genetic distance among AFLP genotypes was calculated using Nei’s unbiased genetic distance (Nei, 1978) as implemented in Tools for Population Genetic Analyses (tfpga) v.1·3 (Miller, 1997). The genetic distance matrix was imported into mega v.4 (Tamura et al., 2007) to construct a UPGMA dendrogram. Support for isolate groupings was assessed using 1000 bootstrap samples in tfpga.
Isolates for inclusion in this study were identified to species by sequencing of the internal transcribed spacer of the rDNA (ITS) in Phytophthora-ID as described previously (Grünwald et al., 2011). Eight nuclear loci and one mitochondrial DNA locus were sequenced in eight isolates for phylogenetic analysis (Table 2): the ITS, the seven nuclear genes sequenced in Blair et al. (2008), and coxI, coxII and the cox spacer region in the mitochondrial genome. Primers and PCR conditions were as previously reported for ITS (Cooke et al., 2000; Grünwald et al., 2011), the other seven nuclear genes (Blair et al., 2008), coxI and coxII (Martin & Tooley, 2003a), and the cox spacer region (Martin et al., 2004; Grünwald et al., 2011). Samples were sequenced by the Oregon State University CGRB on an ABI Prism 3730 Genetic Analyzer (Applied Biosytems). Nucleotide sequences from other Phytophthora clade 8 species and outgroups P. infestans and P. sojae were obtained from GenBank. Accession numbers for sequences included are given in Table 2. All indels were removed from sequence alignments for phylogenetic analysis.
Table 2. Phytophthora species studied and the GenBank sequence accession numbers for DNA sequences characterized
GenBank accession number
n.a. not available.
*Downloaded from GenBank as P. syringae, but shown here to be P. obscura sp. nov. in Figure 9.
ITS and coxII were analysed independently using maximum likelihood, maximum parsimony and Bayesian criteria. A multilocus phylogeny of clade 8 was also conducted by concatenating the seven nuclear loci from Blair et al. (2008). Phylogenetic inference by maximum likelihood was conducted using PhyML 3·0 (Guindon & Gascuel, 2003). Nucleotide substitution models were selected using modeltest 3·7 by AIC (Posada & Crandall, 1998). Parameter values were estimated in PhyML. The starting tree for the analysis was generated using bionj and nearest-neighbour interchange (NNI) was used for tree searching. Branch support was assessed using 1000 bootstrap samples for ITS and coxII, and 500 bootstrap samples for the multilocus data set. Inference by maximum parsimony was conducted in paup 4·0b10 (Swofford, 2002). Heuristic search was used with 20 replicates of random stepwise addition of sequences. Bootstrap analysis was conducted using heuristic search with 1000 samples. Bayesian phylogenetic inference was conducted in MrBayes 3·1·2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). Substitution models were specified based on modeltest 3·7 results (Posada & Crandall, 1998). For the seven concatenated loci, one substitution model was used across all seven loci. A second analysis in which sequences were partitioned by gene and evolutionary rates allowed to vary among genes, produced nearly identical results but required many more MCMC steps to obtain adequate sampling of parameters. Two simultaneous runs were conducted with four chains each, default priors, and tree sampling every 100 generations. Run lengths were 1 × 106 steps for ITS and coxII, and 5 × 106 steps for the unpartitioned, multilocus analysis. The first 2000 trees were discarded as burn-in for calculation of clade posterior probabilities. MCMC output was visualized in tracer 1·3 (Rambaut & Drummond, 2007) to assess convergence. Phylogenetic trees were visualized in mega v.4 (Tamura et al., 2007), TreeView 1·6·6 (Page, 1996) and geneious 4·8·2 (Biomatters Ltd).
The P. obscura sp. nov. isolate from Kalmia (P-09-011) appeared to be damaged in microscopic observations, although it was able to cause disease on Pieris, Kalmia and Rhododendron, but was not inoculated onto Aesculus, indicating that its pathogenicity was not compromised. This isolate was originally obtained by plating on selective media, and recently from liquid nitrogen storage, and either process might have damaged the isolate. It showed differing growth rates and had a high rate of distorted oospores in comparison to the other three isolates of this new species. Therefore the morphological characterization of the new species is based only on the three isolates from soil associated with A. hippocastanum and from substrate underneath Pieris.
Colonies grew smooth to slightly stellate on CPA, OMA and V8A, but cottony and slightly petaloid to stellate on PDA (Fig. 1). The vegetative growth rate varied with the medium. On CPA, P. obscura sp. nov. isolates grew nearly as fast as P. syringae but much faster than the P. austrocedrae isolates (Fig. 2, Table 3).
Table 3. Morphological observations for Phytophthoraobscura sp. nov. type strain and averages observed for P.obscura sp. nov., P. syringae and P. austrocedrae averaged across strains included in this study
P.obscura sp. nov. BBA 2/94-IIB (type strain)
P.obscura sp. nov.a
aOn CPA, summarized data from the isolates BBA 2/94-IIA, BBA 2/94-IIB, JP-09-059 (n per propagule and isolate = 100); the isolate P-09-011 looked damaged, the morphological data were not used for the official description.
bOn CPA, summarized data from the isolates BBA 1/99-28, BBA 12/99-5, PSY-06-099 (n per propagule and isolate = 100; n =40 for gametangia of P. syringae isolate BBA12/99-5, the only P. syringae isolate that produced gametangia).
cOn CPA, summarized data from the isolates PAU-09-001, PAU-09-002 (n per propagule and isolate = 100).
Hyphal swellings were rarely observed. Mycelium showed a relatively low rate of branching. Phytophthora obscura sp. nov. produced semipapillate, non-caducous sporangia on single, sympodial sporangiophores. Sporangia were generally ovoid but could also vary from globose to limoniform (Fig. 3). Spore measurements are presented in Table 3.
The species is homothallic and readily produced oogonia with paragynous antheridia in culture with one antheridium per oogonium (Fig. 3). Oogonia were usually round, and sometimes a little irregular. Antheridia were mainly globose. Oospores were mainly plerotic, but mature oospores were sometimes slightly aplerotic. Dimensions for the gametangia are presented in Table 3. Chlamydospores were not observed.
Amplified fragment length polymorphism (AFLP) analysis
Two selective primer combinations were evaluated. Both yielded qualitatively similar results supporting P. obscura sp. nov. as a new taxonomic species. Electropherograms among P. obscura sp. nov., P. austrocedrae and P. syringae were distinctly different, yet close to identical within species (Fig. 4). The genetic distance between isolates from Germany and Oregon reveals polymorphism within the species and is less than that observed between the three P. ramorum clonal lineages (Fig. 5).
All of the P. obscura isolates had nearly identical ITS sequences. The two Oregon isolates were heterozygous (A/G) at base 696, while the German isolates were homozygous (G). Across the seven other nuclear loci, there were six fixed nucleotide differences between the German and United States isolates. In the cox region, there were four single nucleotide differences between the isolates from Germany and the US: one in coxII, three in coxI, and two in the spacer (Fig. 6).
The best model for sequence evolution for ITS and the other nuclear genes was GTR+I+G. However, the estimates of the gamma shape parameter for ITS in MrBayes were large and highly variable, therefore this analysis was rerun with the GTR+I model, which gave very similar results. The GTR+I substitution model was selected for coxII. ITS sequences produced trees showing P. obscura sp. nov. isolates as a well-supported phylogenetic species (Fig. 7). Furthermore, P. obscura sp. nov., P. austrocedrae and P. syringae formed a well-supported clade distinct from previously defined clades (Cooke et al., 2000; Kroon et al., 2004; Blair et al., 2008). This clade has been labelled 8d (Figs 7–9). Maximum parsimony (MP) produced two equally parsimonious trees for ITS, which differed from each other and from the maximum likelihood tree and Bayesian consensus tree in the relative positions of species in clade 8a but were otherwise equivalent. Similar results were obtained for the seven concatenated nuclear loci (Fig. 8). Phytophthora obscura sp. nov., P. austrocedrae and P. syringae again formed a well-supported clade. The new clade 8d forms a larger highly supported clade with clade 8b. Phytophthora syringae was previously observed to group with clade 8b in a phylogeny based on these same genes (Blair et al., 2008). MP produced two tied trees, which differed in the relative positions of clades 8a and 8c. For the mitochondrial gene coxII, P. obscura isolates were also distinct from P. austrocedrae and P. syringae, and these three species clustered into a clade with 100% support across phylogenetic methods. However, clade 8b sequences were not available for this gene. One of the coxII sequences identified as P. syringae that was downloaded from GenBank (AY369367) was identical to coxII from the Oregon isolate from Pieris substrate JP-09-059. The AY369367 sequence was published by Martin & Tooley (2003b) and was obtained from an isolate collected from K. latifolia in a home garden in Oregon. This isolate (originally 469, here P-09-011) was kindly provided by P. Tooley and included in this study, thus the sequences labelled AY369367 and P-09-011 in Fig. 9 are from the same isolate.
Although only isolated directly from kalmia, host range was evaluated for horse chestnut, kalmia, rhododendron and pieris. On horse chestnuts, first disease symptoms developed within 4 weeks after inoculation, but only with P. obscura sp. nov. and not on all seedlings or inoculation points. New symptoms at other inoculation points and further development of existing symptoms appeared mostly after the winter. Ten months after inoculation, in June 2010, the horse chestnuts inoculated with P. obscura sp. nov. showed severe lesions including some dead branches and bleeding stem cankers (Fig. 10). One seedling was completely killed. Phytophthora obscura caused foliar blight on K. latifolia, Pi. japonica and R. catawbiense cv. Boursault (Fig. 11). Koch’s postulates could be confirmed for K. latifolia isolate P-09-011 obtained from plant tissue. Reisolations and successful identification using ITS sequencing confirmed pathogenicity for four isolates of P. obscura sp. nov. on K. latifolia, A. hippocastanum, Rhododendron and Pieris. The damage with the P. syringae isolates was moderate to minor on all hosts and mainly restricted to the surrounding tissue of the infection points (Fig. 11). The two P. austrocedrae isolates did not cause any disease symptoms on any of the hosts (pictures not shown).
Phytophthora obscura Grünwald & Werres, sp. nov.
Etym.: obscura, obscure: referring to the cryptic nature and the fact that it has (to date) not often been found in nature.
Coloniae in agaris fragmentorum dauci (CPA), farinae avenae (OMA), et “V8” (V8A) laeves vel parum stellatae, in agaro dextrosi Solani tuberosi (PDA) gossypinae et petaloideae vel stellatae visae. Optima temperatura in CPA 25°C, medio incremento diurno radiali 4·3 mm. Tumores hypharum nulli vel rari. Mycelium in CPA tarde ramificans. Sporangia semipapillata, non caduca, terminalia in sporangiophoro, media longitudine latitudineque 36·2 × 26·4 μm, media proportione longitudinis:latitudinis 1·4:1, copiose facta in CPA post inundationem extracto humi sphagnosi. Species homothallica. Antheridia paragyna, fere globosa, media longitudine latitudineque 10·0 × 10·0 μm. Oogonia sphaerica diametro medio 33·4 μm, non ornata. Oosporae in cultura copiose prodientes, plerumque sphaericae et pleroticae vel fere apleroticae, diametro medio 29·8 μm. Nullae chlamydosporae facta.
Typus: Germany: isol. ex soil from Aesculus hippocastanum in an allée, S. Werres (holotypus: BBA 2/94-IIB; cultura ex tip CBS 129273; MB 519834).
Colonies grew smooth to slightly stellate on CPA, OMA and V8A and appeared cottony and petaloid to stellate on PDA. Temperature optimum on CPA at 25°C. Daily incremental radial growth average of 4·3 mm on CPA at 25°C. Hyphal swellings absent to rare. Low branching rate of mycelium on CPA. Sporangia semipapillate, non-caducous, born terminally on the sporangiophore. Sporangia average 36·2 × 26·4 μm with an average L:B ratio 1·4:1. Sporangia mainly ovoid. Sporangia abundantly formed on CPA after flooding with peat extract. Species homothallic. Abundant oospore production in culture. Oogonia not ornamented. Antheridia paragynous, nearly globose 10·0 × 10·0 μm in average. Oogonia spherical with 33·4 μm average diameter. Oospores mainly spherical and plerotic to nearly aplerotic, with 29·8 μm average diameter. No chlamydospores produced.
Phytophthora obscura sp. nov. causes foliar blight on K. latifolia, Rhododendron and Pieris, and bleeding cankers on A. hippocastanum. To date, attempts to isolate P. obscura from bleeding cankers of the naturally infected A. hippocastanum have failed and inoculation studies conducted in the 1990s also failed (Werres et al., 1995). The latter may be due to the very short incubation period after inoculation and the high incubation temperatures used at the time (Werres et al., 1995). Recent experience indicates that the new Phytophthora species might require a long incubation period and low temperatures to infect trees and to establish in the tissue. Thus, it is not yet clear whether P. obscura was responsible for the bleeding cankers observed on horse chestnuts in Germany in the 1990s. One isolate was originally obtained from Kalmia and Koch’s postulates were confirmed for this species. Phytophthora obscura causes foliar blight reminiscent of P. syringae on Kalmia, Rhododendron and Pieris (Fig. 11). Aesculus, Kalmia, Rhododendron and Pieris are highly susceptible to two other clade 8 species including the sudden oak death pathogen P. ramorum and P. syringae (Werres et al., 2001; Parke et al., 2004; Grünwald et al., 2008; Grünwald & Parke, unpublished). All three pathogens are foliar pathogens on Kalmia. In a recent intensive sampling effort in Oregon, P. obscura was only sampled once in Oregon nurseries among c. 700 isolates sampled from Rhododendron, Kalmia, Pieris, substrate and water (Grünwald et al., 2011). To the best of the authors’ knowledge, this pathogen has not been observed anywhere else and is not found in GenBank or any other database. Thus, although found in Oregon once on Kalmia and in substrate in the US or soil in Germany, this species at this point appears to be rare and/or difficult to isolate. Given the similarity to P. syringae, some P. obscura isolates in culture might also be misidentified as being P. syringae.
Morphologically, P. obscura is more similar to P. syringae than to P. austrocedrae. Phytophthora obscura can easily be distinguished from P. austrocedrae mainly because of the slower growth, the bigger sporangia, and the different size and shape of the gametangia of the latter. Phytophthora obscura and P. syringae cannot be distinguished based on morphology alone. Ecologically, P. obscura might be comparable to P. syringae and considered to cause mostly foliar blight, twig blight and stem cankers on ornamentals or cankers on trees. All clade 8d species known to date are characterized as having semi-papillate, non-caducous sporangia and being homothallic. Homothallism indicates that these taxa might be inbreeding species.
This work also provides support for a novel clade 8d in the genus Phytophthora. Detection of novel taxa, and hence deeper taxon sampling, allowed definition of this new subclade 8d based upon the clade structure established by Blair et al. (2008). Blair et al. grouped P. syringae (and P. austrocedrae by association) with clade 8b. The analysis here provides phylogenetic support for P. syringae, P. austrocedrae and P. obscura as being distinct phylogenetic species. Support was also observed for a larger clade encompassing 8b and 8d. However, 8d itself also exhibited strong bootstrap support and the genetic distance between the 8d and 8b clades clearly indicates two separate clades. AFLP analysis, eight nuclear loci and the coxI-cox spacer-coxII mitochondrial locus support this revised structure in clade 8.
The origin of P. obscura is not known. While P. ramorum is now clearly recognized as an exotic pathogen (Mascheretti et al., 2008, 2009; Grünwald & Goss, 2009; Grünwald et al., 2009; Goss et al., 2011), it is not clear at this point whether P. syringae and/or P. obscura are introduced or exotic pathogens. Based on the AFLP data shown here, P. obscura shows some genetic diversity and two apparent clades for the US and German populations. To date, P. obscura has only been found rarely in nurseries (USA) or in association with horse chestnuts in urban plantations (Germany) most likely because it was mistaken for P. syringae, well known as having the same four hosts studied here. At this point the threat posed by P. obscura to woody plants and tree species is unclear and further research is necessary to assess if this pathogen is introduced and what the demographic history of it is (Grünwald & Goss, 2011). The full host range of this taxon also remains to be established. Phytophthora obscura might be involved in causing some of the horse chestnut cankers in Europe, but further work is needed to establish support for this fact.
We would like to thank Paul Tooley, Lani Yakabe and Everett Hansen for providing isolates. We would also like to thank Meg Larsen, Karan Fairchild, Caroline Press, Henrike Gottfried and Julia Hauffe for excellent technical assistance. The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an official endorsement or approval by the United States Department of Agriculture or the Agricultural Research Service of any product or service to the exclusion of others that may be suitable. Financial support was provided by United States Department of Agriculture–Agricultural Research Service (USDA-ARS) CRIS 5358-22000-034-00, the USDA ARS North-West Nursery Research Program and the USDA Floriculture Nursery Initiative.