RESEARCHFree Access icon

Susceptibility of Garden Plants to Phytophthora Root Rot

    Affiliations
    Authors and Affiliations
    • Elizabeth J. Beal1
    • Ian A. G. Waghorn1
    • Joe N. Perry2
    • Gerard R. G. Clover1
    • Matthew G. Cromey1
    1. 1Royal Horticultural Society (RHS), Department of Plant Health, RHS Wisley, Woking, Surrey, GU23 6QB, United Kingdom
    2. 2Oaklands Barn, Norfolk NR35 2HT, United Kingdom

    Published Online:https://doi.org/10.1094/PDIS-04-20-0765-RE

    Abstract

    Phytophthora root rot (PRR) is a serious disease of horticultural, forest, and ornamental plant species caused by species of the oomycete genus Phytophthora. Their wide host range makes the choice of resistant plants in the management of the disease difficult. We used the Royal Horticultural Society diagnostic dataset of PRR records from U.K. gardens to compare the susceptibility of different host genera to the disease. The dataset was compared with existing reports of plants recorded as notably resistant or notably susceptible to PRR. An index-based approach was used to separate 177 genera of woody plants into three categories: 85 were low index (<0.10: rarely affected), 34 were medium index (0.10 to 0.20: sometimes affected) and 58 were high index (>0.20: frequently affected). Similarly, genera of nonwoody plants were separated into: 45 low index (<0.22), 16 medium index (0.22 to 0.44), and 18 high index (>0.44). Taxus was the genus with the highest index, while most genera in the Malvales and Ericales were in the high-index group. Most genera in the Myrtales, Fabales, and Monocotyledons were low index. While 30 Phytophthora species were recorded in our study, the wide host range spp., P. plurivora, P. cryptogea, and P. cinnamomi, represented 63% of identifications. P. plurivora was the most common species on woody plants and P. cryptogea on nonwoody plants. These results provide confidence in the use of host resistance as part of the integrated management of PRR.

    Phytophthora is a genus of fungus-like organisms in the Oomycota (kingdom Straminipila, class Oomycetes, order Peronosporales, family Peronosporaceae). Species of Phytophthora are among the most significant pathogens affecting a broad range of horticultural, forest, and ornamental plant species, including trees, shrubs, annuals, and perennials (Henricot et al. 2014). Jung et al. (2016) reported that Phytophthora is responsible for >66% of all fine root diseases and 90% of all collar rots of woody plant species. They found that nursery stands across Europe were almost ubiquitously infected with a large array of Phytophthora species and that 83% of horticultural plantings were infected. At the Royal Horticultural Society (RHS) Advisory Service, Phytophthora diseases are among the most frequently diagnosed causes of plant death in gardens (Henricot et al. 2014).

    Symptoms of Phytophthora root rot (PRR) most often observed in trees and shrubs are the abnormal presence of small, yellow, or sparse leaves, reduced extension growth, crown die-back, and even death (Strouts 2012). The disease itself occurs on the roots or on the stem at ground level, where some or all of the root system becomes water-soaked, brown or black in color, and soft (O’Neill and Ann 2016). Attacks by Phytophthora may be short-lived and not progressive, although they may recur (Strouts 2012).

    The disease is frequently associated with the death of trees and shrubs in low-lying, poorly drained sites, and reinfection usually results where the site is replanted with the same or a closely related species (Robertson 1970). Phytophthora species can survive unsuitable conditions with resting structures in soil, infected roots, and organic debris (Jung et al. 2016). They can therefore be introduced to gardens in nursery stock or in soil (Prigigallo et al. 2015, Schwingle et al. 2007) and, once in a location, can survive in the absence of a susceptible host. Where a broad choice of plants to grow is available, management could consist of the use of relatively resistant species, attention to drainage, and other cultural measures. Evidence is lacking on the relative susceptibility of different garden plants to damage by Phytophthora species, but this information would assist gardeners in making planting choices in areas where the disease has been recorded.

    Most studies on the association of Phytophthora species with trees have focused on woodlands and forests within natural ecosystems, with very few studies within the urban environment (Barber et al. 2013). The urban environment (including gardens) is usually highly disturbed (Barber et al. 2013). Soil compaction, combined with irrigation, promotes PRR and increases its impact (Barber et al. 2013). However, gardens also provide an opportunity to manage both the environment and planting choices.

    Very few experimental studies have compared the susceptibility of hosts to PRR. Robertson (1970) tested 67 plant species and cultivars in an inoculated experiment with Phytophthora cinnamomi, while Shearer et al. (2013) tested the relative susceptibility of Australian flora to P. cinnamomi using a soil inoculation method of plants in pots. Experimental assessment of host resistance is not always straightforward. For instance, a stem inoculation method to assess resistance to P. cinnamomi in Leucospermum cultivars was found not to approximate a natural disease situation (Denman and Sadie 2001).

    Other than these few experimental studies, there are reports listing plants that are notably resistant or susceptible to PRR. These include two reports from the United Kingdom (O’Neill and Ann 2016; Strouts 2012) and seven from the United States (Abbey 2011; Anon 2018; Anon 2019; Creswell et al. 2011; Hagan and Mullen 2000; Meadows et al. 2017; Moorman 2014). These lists are derived from host reports, disease clinic records, and expert opinion (Creswell et al. 2011). These host lists provide limited information on the relative susceptibility of plants to PRR but some, such as Taxus, are notably susceptible (Strouts 2012). A survey of U.K. gardens between 2006 and 2009 found Taxus, Rhododendron, Prunus, and Buxus to be the most common hosts (Denton 2014). However, these figures do not take into account the popularity of these plants in gardens, and so may not indicate their degree of susceptibility to PRR.

    As of August 2007, ∼110 Phytophthora species had been described, with an estimated 200 to 500 unknown species not yet described (Brasier 2009). By 2019, the number of species named or provisionally named had reached 180 (Riddell et al. 2019). While many Phytophthora species exist, a much smaller subset of species with wide host ranges are recorded around the world. Phytophthora species with a wide host range found in many geographic locations around the world include P. plurivora, P. cinnamomi, P. nicotianae, P. cryptogea, and P. cactorum (Jung et al. 2016). These species predominated in a survey of U.K. gardens between 2006 and 2009 (Denton 2014), suggesting that wide host-range species predominate in U.K. gardens.

    The RHS receives samples from its members for disease diagnosis. Records held in the RHS database provide a potentially valuable source of information on resistance to PRR. It has consistently been in the top five diseases identified by plant pathologists at the RHS Gardening Advice Service (GAS) for >20 years. We used this database to group garden plants according to their susceptibility to Armillaria root rot (Cromey et al. 2020). That study showed that numbers of records of the disease in the database must be adjusted by a measure of the popularity of garden plants.

    Determination of the degree of resistance or tolerance of plants to PRR would provide a valuable tool in the integrated management of the disease. The objectives of this study were therefore to (i) develop and validate a list of garden plant genera, according to their relative likelihood of damage, (ii) determine any relationship between plant order or clade with susceptibility, and (iii) scrutinize the host preferences of Phytophthora spp. associated with the disease in U.K. gardens and compare this with other disease environments. This study only considered Phytophthora root rot, so Phytophthora species that primarily affect aerial plant parts such as P. ramorum, P. kernoviae, and P. ilicis were not included in the dataset.

    Materials and Methods

    RHS dataset.

    The RHS operates a diagnostic clinic as a part of an advisory service for its ∼500,000 members. The advisory service receives samples from members for diagnosis of possible plant diseases. Above-ground symptoms of root disease are often the rapid die-back or death of plants. The diagnosis of PRR requires appropriate symptomatic material (root and root collar tissue). Such symptoms include blackening and necrosis of fine roots, and a blackened lesion spreading from below the soil up into the root collar. Where symptoms were consistent with PRR, samples were tested with an immunological assay or by floating the sample in sterile pond water for 24 to 48 h and then looking for characteristic features such as sporangia. The immunological assay used was a lateral flow device for Phytophthora detection (Abingdon Health, Sand Hutton, United Kingdom). Symptomatic material was tested according to the product instructions.

    RHS records are from U.K. gardens and represent the only large database of PRR records in gardens that we are aware of. Data relevant to this study include the numbers of enquiries on each genus to the RHS GAS and the number of confirmed PRR diagnoses. As a popularity measure, we have used the total enquiries (TE) to the RHS GAS on each genus for the 12-year period (1 January 2007 to 31 December 2018) for which data are available. The TE measure includes enquiries on all aspects related to that genus in gardens, including identification, horticulture, and plant health. The number of Phytophthora root rot records on each host genus was recorded between 1 January 1997 and 1 September 2019. The TE for each genus was used to normalize the PRR record, and compared with the original PRR record to determine whether PRR records provide a useful measure of the likelihood of damage owing to Phytophthora infection (see below).

    Comparison between RHS data and existing reports on host resistance or susceptibility.

    A comparison was made between the RHS dataset and existing reports on host resistance or susceptibility to validate our approach and to set thresholds for “resistance” categories.

    We used all the reports we were able to find that included lists of resistant (R) or susceptible (S) genera for a comparison with the RHS dataset. Reports used were seven U.S. university extension reports (Abbey 2011; Anon. 2018, 2019; Creswell et al. 2011; Hagan and Mullen 2000; Meadows et al. 2017; Moorman 2014), two U.K. reports (O’Neill and Ann 2016; Strouts 2012), and the results of inoculated experiments in New Zealand (Robertson 1970) and Australia (Shearer et al. 2013). Each record was treated as equally valid. Where reports were inconsistent (genera referred to as “resistant” and “susceptible” in different reports), genera were not used for the comparison unless at least 75% of reports were in agreement.

    A graphical comparison of hosts appearing both in the RHS database and other reports was made (data not included). The evaluation was restricted to plants grown outdoors in the ground because plants primarily grown in pots or grown indoors are less likely to be exposed to Phytophthora. Initial examination of the plots indicated that trees and shrubs (woody plants) and other garden plants (mostly perennials) needed to be considered separately because the relationship between our data and the previous reports was clearly different for the two (Fig. 1).

    Fig. 1.

    Fig. 1. Fitted and observed relationships between PRR records on woody plants and nonwoody plants (log10[PRR records + 1]) versus total enquiries (log10[TE + 1]), with two parallel fitted lines (line S above, line R below: solid and dotted lines are woody and nonwoody plants, respectively) for genera reported to be susceptible or resistant to PRR.

    Download as PowerPoint

    The analysis used data for the 67 woody and 17 nonwoody genera in other reports that also had TE of at least 100. Genera with <TE 100 were not included because a small change in the number of PRR records will have a large effect on the relationship between TE and PRR records.

    Values of TE and PRR records were transformed logarithmically to new variables x and y, where x = log10(TE + 1) and y = log10(PRR records + 1). The transformed variables x and y were plotted (separately for woody and nonwoody plants) against one another, where the points represented genera categorized previously as R or S to PRR. Three linear regression models of transformed PRR records (y) on TE (x) were compared for the best fit: a single line combining both reported resistance categories; two parallel lines allowing the intercept to vary with resistance category; and two separate lines allowing both slope and intercept to vary with category.

    This analysis was performed using the software package Genstat, 16th edition (VSN International 2013). Discrimination between models was by standard partial F tests with two degrees of freedom (Hawes et al. 2003; Perry 1982).

    PRR index.

    A PRR index was developed from RHS data for each genus. PRR records were adjusted by TE to provide a weighted index for each genus. The index formula differed between woody and nonwoody genera (see the “Results” section).

    Comparison between RHS data and infection levels in ornamental plantings.

    PRR indices were compared with Phytophthora infection levels in ornamental plantings in Europe (Jung et al. 2016). Separate analyses were carried out for woody and nonwoody plants. Regressions were done on the logit transformed proportion of infected plants on the PRR index.

    Relationship between plant taxonomic group and PRR index or Phytophthora spp.

    Angiosperm genera were allocated to family, order, and clade according to the Angiosperm Phylogeny Group (2016). There were relatively few gymnosperms; thus, these were treated as a single group.

    The relationships between taxonomic position of the host and PRR index or associated Phytophthora spp. were examined. Median, minimum, maximum, and first and third quartile for PRR indices for genera in each taxonomic group were calculated. Woody and nonwoody plants were treated separately. The number of host genera in each PRR index group was recorded, along with the number of each Phytophthora species identifications, from each host taxonomic group.

    Finally, χ2 analyses were performed for comparison of the proportion of genera in taxonomic groups assigned to each resistance category, as well as for the proportion of Phytophthora spp. that had infected each taxonomic group. Differences worthy of note between observed and expected outcomes were defined as (o − e)2/e > 2, where o is the observed count and e is the expected.

    Isolation and identification of Phytophthora spp.

    Isolation and species identification was done from approximately half the samples received between 1998 and 2015. A total of 930 samples were identified to species. Of these, 460 were isolated as part of a Ph.D. study (Denton 2014). Phytophthora spp. were isolated from root and stem samples using baiting techniques (apples, hemp seeds, or rhododendron leaves) or by direct plating of necrotic roots or stem onto selective media as described by Henricot et al. (2014). Single hyphal tip cultures were obtained and then subcultured onto 3% oatmeal agar slopes in storage tubes under paraffin. Tubes were then stored at 10°C.

    DNA isolation, amplification, and sequencing.

    Methods of DNA isolation, amplification, and sequencing on the above set of 1,033 isolates were as described by Henricot et al. (2014). Exceptions since 2013 include the replacement of polymerase chain reaction beads with Bio-Rad Master Mix and the replacement of ethidium bromide with SYBR Safe. For all isolates, DNA was extracted from either the roots and the stem bases of symptomatic plants and/or the Phytophthora cultures. Identifications were made using semi-nested polymerase chain reaction to amplify ITS-rDNA, followed by sequencing and comparison against sequences in databases such as GenBank and PhytophthoraDB (Jung et al. 2016).

    Results

    Phytophthora spp. associated with PRR in U.K. gardens.

    While other Phytophthora species were occasionally recorded as associated with root rot in U.K. gardens, the 17 most common species represent 99% of the total (Table 1). Three Phytophthora spp. represented 60% of the population in our study of U.K. gardens. Phytophthora plurivora and P. cryptogea were each responsible for >20% of the total, while P. cinnamomi made up 14%. The most common 10 species were responsible for >90% of the total.

    Table 1. Phytophthora species composition associated with root rot (number and percentage of identifications) in U.K. gardens in 1998–2015

    Three species, P. cryptogea, P. cactorum, and P. pachypleura, were relatively more common on nonwoody hosts (where they represented 56% of the total) than woody hosts (30%), while P. plurivora, P. cinnamomi, and P. citrophthora were particularly common on woody hosts (representing 48% of the total, compared with 23% on nonwoody hosts).

    Comparing the RHS dataset with previous reports.

    Plant genera in the RHS dataset were compared with the disease susceptibility ratings in previous reports (Fig. 1). Comparing RHS data with the disease susceptibility ratings of previous reports enables the establishment of a set of thresholds for the RHS data. Because there is a continuum between highly resistant and highly susceptible plants, the aim was to develop three categories: low index (rarely affected), high index (frequently affected), and an intermediate- or medium index for those that are neither notably resistant nor notably susceptible.

    For woody plants (Fig. 1), the fitted relationship for the susceptible category is y = −0.41 + 0.50x. For the resistant category, the fitted relationship is y = −0.99 + 0.50x.

    The regressions for woody plants showed that there was a linear relationship between y and x (F1,65 = 45.5, P < 0.001), that there was no need for two separate lines (F1,63 = 0.09, P > 0.5), and that the best-fitting model was two parallel lines with separate intercepts (F1,64 = 33.7, P < 0.001) and the same slope (Fig. 1). The parallel nature of the best-fitting model supports an index-based approach for RHS data.

    The slope of the fitted line for woody plants is not equal to unity, so there is density dependence. However, the slope is very close to one-half, implying that there is a proportional relationship between PRR records and the square root of TE. For susceptible genera, PRR records = 0.39√TE, while for resistant genera, PRR records = 0.10√TE. The threshold between “susceptible” and “not susceptible” was therefore computed as the line midway between the two fitted lines, i.e., y = –0.70 + 0.50x. This is equivalent to the relationship: PRR records = 0.20√TE. Each genus was assigned a PRR index, calculated as PR/√TE. Indices > 0.20 were therefore classified as susceptible. Examination of data and plant pathological experience indicate that resistance follows a graduated scale. Therefore, values between 0.00 and 0.20 were further divided. Indices between 0.00 and 0.10 were defined as “seldom affected” (equivalent to resistant), while those between 0.10 and 0.20 were defined as “sometimes affected” (equivalent to moderately resistant).

    For nonwoody plants (Fig. 1), the fitted relationship for the susceptible category is y = –0.22 + 0.27x. For the resistant category, the fitted relationship is y = –0.50 + 0.27x.

    The regressions for nonwoody genera also showed that there was a reasonably linear relationship between y and x (F1,28 = 3.1, P = 0.087), that there was no need for two separate lines (F1,26 = 0.33, P > 0.5) and that the best-fitting model was two parallel lines with separate intercepts (F1,27 = 7.9, P = 0.009) and the same slope (Fig. 1). The parallel nature of the best-fitting model again supports an index-based approach for RHS data.

    As with woody plants, the slope of the line for nonwoody plants is not equal to unity, so there is density dependence. The threshold between “susceptible” and “not susceptible” was computed as y = –0.36 + 0.27x, which is equivalent to the relationship: PRR records = 0.44(TE)0.27. Each genus was assigned a PRR index, calculated as PRR records/(TE0.27). Indices > 0.44 were therefore classified as “susceptible.” As above, index values between 0.00 and 0.44 were further divided. Indices between 0.00 and 0.22 were defined as “seldom affected” (equivalent to resistant), while those between 0.22 and 0.44 were defined as “sometimes affected” (equivalent to moderately resistant).

    Comparing RHS indices with infection levels.

    PRR indices were compared with Phytophthora infection levels in ornamentals plantings in Europe (Jung et al. 2016). For both woody and nonwoody host genera there was a positive relationship between the proportion of infected plants and the PRR index.

    A regression of logit transformed proportion of infected plants on the PRR index was just significant (P = 0.052) for woody plants (Fig. 2). The fitted line is logit(proportion infected) = –0.218 + 0.922 index. The standard error of the slope = 0.462.

    Fig. 2.

    Fig. 2. Fitted and observed relationships between logit transformed proportion of infected plants (Jung et al. 2016) versus PRR index for woody or nonwoody plants.

    Download as PowerPoint

    A regression of logit transformed proportion of infected plants on the index was significant (P = 0.012) for nonwoody plants (Fig. 2). The fitted line is logit(proportion infected) = –0.780 + 1.178 index. The standard error of the slope = 0.411.

    Notably susceptible hosts.

    Twelve genera of trees and shrubs had a PRR index ≥ 0.80 (Table 2). Of these, two are gymnosperms, including Taxus, which had a substantially higher index (6.03) than any other genus. The remainder are scattered through the Eudicot clades (rosids, asterids, and eudicot subclade). Two genera (Rhododendron and Calluna), in the Ericaceae, were in this group. No monocots, magnolids, superasterids, or superrosids were in this high-index group. Eight of the 12 genera are on lists of susceptible plants, while none have been recorded as resistant.

    Table 2. Host genera with PRR index > 0.80 (trees and shrubs) or 0.70 (nonwoody ornamentals)

    Ten genera of nonwoody ornamentals had a PRR index ≥ 0.70 (Table 2). They were spread through the Eudicot clades (rosids, superrosids, asterids, and eudicot subclade). Four genera had a PRR index > 1.0. The most common order in the set is the Ranunculales, which make up 30% of the group. Five of the 10 are on lists of susceptible plants, while one (Anemone) has been recorded as resistant.

    Susceptibility of host genera to PRR.

    The 177 genera of woody plants in the three PRR index categories are summarized in Table 3. They include all woody plant genera in the RHS database with at least 100 TE. Forty-eight percent (85) of genera had a low index, 19% (34) had a medium index, and 33% (58) had a high index.

    Table 3. Summary of PRR records in the RHS GAS database (1997–2019) on woody plant genera.

    The 79 genera of nonwoody plants in the three PRR index categories are summarized in Table 4. They include all nonwoody plant genera in the RHS database with at least 100 TE. Fifty-seven percent (45) of genera had a low index, 20% (16) had a medium index, and 23% (18) had a high index.

    Table 4. Summary of PRR records on nonwoody plants.

    Susceptibility of host order or clade to PRR.

    Relationships between the PRR index and host plant taxonomic group were examined (Tables 5 and 6).

    Table 5. PRR index for host plant taxonomic groups above genus levelz

    Table 6. Distribution of index categories in different taxonomic groups

    The median index for gymnosperms was higher than that for angiosperms (Table 5). More than half of angiosperms were in the low-index category, while almost half of gymnosperms were in the high-index category (Table 6).

    The distributions of genera across the index categories differed among the three angiosperm clades. The median index for monocots was 0.00 (Table 5) and most were in the low-index group (Table 6). The four magnolid genera were in the low- and medium-index categories, the maximum index being 0.16. Most angiosperms grown in U.K. gardens are Eudicots and these therefore fitted the overall angiosperm distribution. There were no major differences between Eudicot subclades in proportions of genera in the three index categories, although the five superasterids were in the low or medium groups (Table 6).

    There were significant differences between Eudicot orders in distribution of genera across the three index categories (Table 6). Two orders in particular, the Myrtales and the Ericales, diverged from other Eudicot orders. Seven of eight Myrtales genera were in the low-index category, while one was in the medium category. They also had the lowest median index. Conversely, the Ericales genera tended to have higher indices, with only 23% in the low-index category. Their median index was one of the highest (Table 5). Most genera in the Sapindales, Malvales, and Buxales were in the high-index group (Table 6). The Sapindales and Malvales had the highest median index (Table 5).

    Although, in general, there was a consistency in distribution of index categories between woody and nonwoody genera, there was one exception (Table 5). The Ranunculales had the highest median index among nonwoody plants, but a low median among the woody plants. Three of five woody Ranunculales were low index, while four of eight nonwoody genera were in the high-index category.

    Association between Phytophthora species and host plant groups.

    The eight most common Phytophthora species were recorded on both Gymnosperms and Angiosperms (Table 7). While only one-quarter of Phytophthora records were on Gymnosperms, there were more records of P. cinnamomi on Gymnosperms than on Angiosperms. Four species—P. cactorum, P. pachypleura, P. cambivora, and P. syringae—were relatively uncommon on Gymnosperms, each representing ≤2% of the total.

    Table 7. Distribution of Phytophthora species in different host taxonomic groups

    Ninety-eight percent of Phytophthora identifications on Angiosperms were from the Eudicot clade. There is some evidence that P. cinnamomi was relatively more common on Magnolids (50% of records) than on other clades, and that P. pachypleura was more common on Monocotyledons (20%) than on other clades (<5%).

    Phytophthora species distributions differed significantly (P < 0.001) between Eudicot subclades. The best represented subclades were the Rosids and Asterids. P. cactorum and P. cambivora were more common on Rosids than on Asterids (10% vs. 4% and 7% vs. 1%, respectively), while P. pachypleura was more common on Asterids (8%) than on Rosids (3%). P. citrophthora was particularly common on the Eudicot subclade (24%) and P. cactorum (21%) on the Superrosids.

    There was a significant effect of Eudicot order on Phytophthora species distributions (P < 0.001). Of particular note, P. cryptogea was especially common on the Asterales (57%), P. cinnamomi on the Ericales (24%), P. citrophthora on the Buxales (44%), P. pachypleura on the Garryales (54%: mostly on the genus Aucuba), and P. cambivora on the Fagales (25%) and Sapindales (18%). The most common species, P. plurivora, was spread across all the Eudicot orders, but was relatively uncommon on the Asterales (7%). Likewise, the second most common species, P. cryptogea, was recorded on all Eudicot orders, but was relatively uncommon on the Fabales (4%), Garryales (7%), and Dipsacales (3%).

    Discussion

    The PRR index provides a means to group plants according to their susceptibility to this disease. It calculates the relative frequency of occurrence of the disease on different hosts in U.K. gardens by adjusting numbers of records in the RHS diagnostic database with the popularity of the host in U.K. gardens. We have previously used the RHS diagnostic database for a quantitative assessment of the relative susceptibility of host plants to Armillaria Root Rot (Cromey et al. 2020). As far as we are aware, these are the only times that a diagnostic database has been used for a quantitative assessment of the relative susceptibility of host plants to a disease.

    The large RHS diagnostic database provides the opportunity both to evaluate existing reports on susceptibility to PRR and to add information on a much wider range of taxa. The association among plant taxa between the PRR index and existing reports of plants considered to be resistant or susceptible to PRR (Abbey 2011; Anon 2018, 2019; Creswell et al. 2011; Hagan and Mullen 2000; Meadows et al. 2017; Moorman 2014; O’Neill and Ann 2016; Strouts 2012) suggests that the categories are relevant across regions and ecosystems. It also appears to be valid across different assemblages of Phytophthora species. Jung et al. (2016) reported on the incidence of Phytophthora species in surveys of 27 mixed amenity/ornamental plantings in 20 European countries. The relationship between infection levels reported therein and our PRR indices provides additional support that high-index genera are notably susceptible to PRR, while low-index genera are less susceptible.

    Taxus, the highest index genus, is known to be particularly susceptible to PRR (Strouts 2012). For instance, Phytophthora is recorded as the cause of decline and mortality of Taxus in Greece (Tziros and Diamandis 2013) and in the United States (Murray and Hansen 1997). Castanea, another high-index genus, has long ago been wiped out in conducive microsites in the United States as the result of Phytophthora infection (Hansen et al. 2012).

    The 13 genera of trees and shrubs with a PRR index > 2.00 should be regarded as highly susceptible to PRR and avoided where inoculum is known to be present, especially where waterlogging is likely. Eleven of the 13 genera were recorded as being susceptible to PRR by O’Neill and Ann (2016). Two genera (Aucuba and Sarcococca) were not included in their list, which focused on container-grown hardy ornamentals. However, Buxus (Sarcococca is in the family Buxaceae) is listed as susceptible. Aucuba is particularly susceptible to P. pachypleura, a species described a few years ago by Henricot et al. (2014) and may not be as susceptible to the more widespread species.

    Forty-three genera of woody plants with TE = 100+ had no PRR records (index = 0.00). These genera are likely to have adequate resistance to PRR in almost all situations. A further 13 genera had PRR indices > 0.00, but <0.05. Of these, we found only one (Callistemon) reported to be susceptible to PRR elsewhere (Robertson 1970).

    Identifying trends in PRR resistance at higher taxonomic levels would help to predict the status of genera where data are lacking. Shearer et al. (2013) evaluated the variation in susceptibility of 501 Western Australian taxa to P. cinnamomi infection after soil inoculation. They found trends within families that were reasonably consistent predictors of susceptibility. In most cases, these findings were consistent with the PRR indices presented here. They found that most taxa in the Fabaceae (order Fabales), Malvaceae (order Malvales), Myrtaceae (order Myrtales), and Poaceae (in Monocots) were resistant to P. cinnamomi. Comparing PRR indices, all Myrtales and Monocot genera had low indices and the median index (0.20) of the Fabales was the fourth lowest of the 14 Eudicot orders. Conversely, most of the Malvales were in the higher index bracket. However, P. cinnamomi was relatively rare on the Malvales in our study (P. plurivora was the most common species recorded) and it is possible that the Malvales are particularly resistant to P. cinnamomi.

    Wide-host-range Phytophthora species found in many geographic locations around the world, include P. plurivora, P. cinnamomi, P. nicotianae, P. cryptogea, and P. cactorum (Jung et al. 2016). These species represented >70% of identifications in this study, P. plurivora, P. cryptogea, and P. cinnamomi being responsible for 63%. Other species were less common in our study; for example, P. citrophthora and P. nicotianae represented 7.5 and 1.4% of identifications, respectively.

    P. plurivora, the most common (24% of identifications) species on trees and shrubs in U.K. gardens, was described in 2009, before which it was identified as P. citricola (Jung and Burgess 2009). P. plurivora was shown to be widespread and aggressive in forests, seminatural ecosystems, and nurseries on a wide range of hosts across Europe (Jung and Burgess 2009), and our results show this is also true in garden environments.

    A degree of specialization toward woody or nonwoody hosts is apparent among Phytophthora spp. The species responsible for PRR on herbaceous plants in North Carolina are primarily P. nicotianae, P. cactorum, P. cryptogea, P. drechsleri, P. palmivora, and P. tropicalis (Meadows et al. 2017). Similarly, in our study, P.cryptogea, P. nicotianae, and P. cactorum were more common on nonwoody than on woody hosts, P. cryptogea being the most common species on nonwoody hosts. The wide host-range species P. plurivora was also common (the second most common species) on nonwoody hosts, but it had a higher frequency on woody than on nonwoody hosts. Our study also supports the woody-host preference of P. cinnamomi (Erwin and Ribeiro 1996; Podger 1972) where it was the third most common species overall, but only the fifth most common species on nonwoody plants.

    P. cinnamomi, the most common species on Taxus baccata in our study, is reported to be a particular problem on the gymnosperms T. baccata, Chamaecyparis lawsoniana, and Araucaria araucana (Tziros and Diamandis 2013; Vegh and Bourgeois 1975). P. citrophthora has been reported to particularly cause a root and crown rot of Taxus (Erwin and Ribeiro 1996), which our data supports, as Taxus was the most common host of this species in our study (13 records). Conversely, P. cambivora is seldom recorded on gymnosperms (Erwin and Ribeiro 1996). Likewise, in our study only one of 35 records of P. cambivora was from a gymnosperm (Taxus).

    Shearer et al. (2013) found that taxa in the Ericaceae (order Ericales) were mainly susceptible to PRR. In our study, the Ericaceae were all high index, while only a single member of the Ericales (in family Theaceae) had a lower index (0.10). The Proteaceae (Wills 1992) and the Ericales (Shearer et al. 2013) are reported to be particularly susceptible to P. cinnamomi (Wills 1992). Similarly, in our study, P. cinnamomi represented 28.6% of identifications from the Proteales, compared with 7.5% across all Eudicot orders. It was also particularly common on the Ericales in our study, representing 23% of records on this order.

    Vettraino et al. (2010) reported P. citrophthora to be the cause of root rot of Buxus in Italy. Buxus (11) and Sarcococca (4), both in the Buxales, were common hosts of P. citrophthora in our study, confirming the susceptibility of these hosts of this species.

    P. pachypleura was described by Henricot et al. (2014), although it had previously been recorded as an unidentified Phytophthora species in the P. citricola complex since 2001. It is the main cause of mortality of Aucuba japonica in U.K. gardens, as well as being sporadically associated with other hosts (Henricot et al. 2014). While one-third of P. pachypleura records in our study were from Aucuba, the 25 hosts recorded in our study suggests it can be hosted in a wide range of garden plants. P. pachypleura represented 79% of records on Aucuba. The PRR index for Aucuba is high (1.25—the third highest index), probably because of the high susceptibility to P. pachypleura.

    In our study, P. plurivora was uncommon on the order Asterales, which is mostly comprised of herbaceous plants. This agrees with Jung and Burgess (2009), who list woody species as the primary hosts of this species and note that it causes fine root destructions, collar rots, and aerial bark cankers. It was, however, the second most common species on nonwoody plants, suggesting that it occurs on most garden host plants.

    P. cryptogea infects a wide range of herbaceous plants, especially in the Asteraceae, as well as woody plants (Erwin and Ribeiro 1996). In our study, P. cryptogea was by far the most common species of plant pathogen found on annuals and herbaceous perennials, especially those in the orders Asterales and Lamiales. It was also the second most common species of plant pathogen found on woody plants, demonstrating its wide host range across ornamental garden plants.

    Most attention is paid to PRR in trees and shrubs, where sudden decline or death can have a big impact on the appearance of a garden. However, PRR also occurs on nonwoody plants. Many ornamental annuals and herbaceous perennials in greenhouses, nurseries, and landscapes in North Carolina, United States, are susceptible to Phytophthora (Meadows et al. 2017). Symptoms depend on the host, but may be seen as a crown or stem rot resulting in rapid wilt or plant collapse or a root rot causing stunted or wilted plants. Affected roots appear brown to black, or roots may be mostly decayed (Meadows et al. 2017).

    Nonwoody genera tended to have a lower incidence of PRR than woody hosts in our study. Crone et al. (2013) found in Western Australia that 15 of 19 common herbaceous plant species tested were hosts of P. cinnamomi, but 10 of these were symptomless hosts. The study of Crone et al. (2013) included seven genera in the Asteraceae, of which P. cinnamomi was recovered from five, but root rot symptoms were not recorded in any. Similarly, we had only a single record of P. cinnamomi on samples of the Asterales (which are mostly herbaceous annuals or perennials) sent for diagnosis of dieback or root rot. Our study did not aim to detect Phytophthora growing asymptomatically within herbaceous plants. However, results of these other studies suggest that they may enable the maintenance of pathogen inoculum in soil. It is therefore important to deal with waterlogging issues that can facilitate pathogen spread in soils before replanting, irrespective of the apparent susceptibility or resistance of planting choice after diagnosis of PRR.

    The development of three categories of plants according to the likelihood that they will be adversely affected by PRR provides alternative planting choices in gardens. Choosing the right plant in the right place may reduce the risk that a plant that has been exposed to Phytophthora spp. will exhibit aerial symptoms. For instance, the most susceptible plants should be avoided in situations where PRR has been diagnosed, especially where conditions conducive to it, such as waterlogging, cannot easily be rectified. Planting plants in the least susceptible group will be advisable where the risk is particularly high. This approach is already being used by gardeners to manage the risk of Armillaria root rot, where results of a study on this disease (Cromey et al. 2020) has been interpreted for gardeners in a readily accessible web format (see “Author-Recommended Internet Resources”).

    Acknowledgments

    We thank RHS members for the provision of sample material, Jenny Denton and Geoff Denton for isolating and sequencing Phytophthora, and Beatrice Henricot for validating many of the Phytophthora identifications. The RHS is a registered charity under No. 222879/SC038262.

    Author-Recommended Internet Resources

    Royal Horticultural Society: https://www.rhs.org.uk/advice/pdfs/phytophthora-host-list.pdf

    The authors declare no conflict of interest.

    Literature Cited

    The authors declare no conflict of interest.