While resembling Eumycotan fungi with the production of hyphae, the genus is placed in the kingdom Straminipila, class Oomycetes, order Peronosporales, and family Peronosporacae. The type species is P. infestans described by de Bary in 1876. Since this time over 128 species have been described, many of which are important plant pathogens capable of significantly impacting agricultural production and natural ecosystems. Some species have a rather narrow host range (P. infestans, P. lateralis, P. sojae) while others are capable of infecting a wide range of plant host species (P. cinnamomi, P. nicotianae, P. ramorum). From a historical perspective, most investigations on the genus have focused on the impact of the genus on agricultural production systems, however, more recently there has been an increased interest in investigating the role this genus plays in natural ecosystems as exemplified by the number of publications concerning species such as P. ramorum and P. alni, as well as the description of many new species recovered from environmental sampling (Martin et al. 2012).
Although Phytophthora species resemble Eumycotan fungi with the production of hyphae, evolutionarily they are more closely related to chromophyte algae and plats than to Eumycotan fungi (Wainright et al. 1993). They have cell walls that are primarily cellulose rather than chitin as observed in fungi and they are incapable of synthesizing β-hydroxysterols (which are required for synthesis of hormones regulating sexual reproduction). In addition, Oomycetes are diploid throughout their life cycle in contrast to most true fungi.
An excellent overview of the ecology, biology and taxonomy of the genus (although missing more recently described species) can be found in Erwin and Ribeiro (1996), a review of the recent taxonomic status in Kroon et al. (2012) and an overview of the genus, including molecular identification and diagnostics, in Martin et al. (2012). There are several publically available databases that provide a wealth of up to date information on the genus, along with sequences useful for species identification via BLAST analysis, including the Phytophthora Database (www.phytophthoradb.org), Phytophthora ID (www.phytophthora-id.org) and Q-Bank (www.q-bank.eu). Cline et al. (2008) have published an online list of Phytophthora spp. with a hyperlink for each species to the USDA SMML database that includes host range, distribution and supporting literature.
Species identification and numbers
A complicating factor when trying to identify Phytophthora species or investigate phylogenetic relationships is hybridization among distinct evolutionary lineages. While this does not appear to be a common occurrence, several stable hybrid species have been identified, e.g. P. andina (Goss et al. 2011; Blair et al. 2012); P. alni (Brasier et al. 1999); P. x pelgrandis (Nirenberg et al. 2009); P. x serindipita (Man in ’t Veld et al. 2012) as well as hybrid clade 1 species recovered from the field (Man in ’t Veld et al. 1998, 2007; Hurtado-Gonzales et al. 2009; Bonants et al. 2000). While conducting a detailed evaluation of clade 6 Phytophthora spp. from natural ecosystems in Australia, Burgess et al. (2010) observed ‘hybrid swarms’ that contained mixtures of parent, offspring, and intermediate isolates with high tendencies for back-crossing and out crossing. The authors’ concluded that the presence of such hybrid swarms was indicative of sexual and somatic hybridization events; the high proportion of these variant isolates within the population also suggested that these hybridization events were not uncommon. Recently four interspecific hybrid clade 6 species have been recovered from riparian ecosystems in Australia and South Africa that reflect outcrossing between P. amnicola, P. thermophila and P. taxon PgChlamydo (Nagel et al. 2013). Additional putative interspecific hybrids from riparian ecosystems in Australia were reported by Hüberli et al. (2013). Hybridization is a topic that requires a more detailed investigation as it could have a profound influence on gene flow among species and the evolution of new species with an expanded host range that could impact agricultural and natural ecosystems (as observed with P. alni).
Traditional classification to species level has been based on morphological characterization of reproductive structures (reviewed in Martin et al. 2012). This includes the sporangium (asexual) and oospore (sexual) as well as the production of chlamydospores (asexual structure not produced by all species). Important features of the sporangium include their dimensions (length and breadth), shape, thickening at the terminus (papilla), length of stalk (pedicle), whether or not the sporangium can be easily dislodged from the sporangiophore (caducity), and proliferation of sporangia (internal, external or nested).
The sexual reproductive structures consist of the antheridium and oogonium (paternal and maternal gametangia, respectively) and are produced when cultures are grown on the appropriate sterol-containing medium. Their fusion leads to the formation of an oogonium that matures into a thick-walled resting structure referred to as an oospore. While most species are homothallic and form oospores in single culture, there are heterothallic species where pairing with opposite mating types is essential to stimulate production of sexual reproductive structures. Since Phytophthora is sexually dimorphic (an isolate of a heterothallic species can function either as the maternal or paternal parent depending on the isolate it is paired with) it is advisable to pair self-sterile isolates with two tester isolates of opposite mating type. While the use of tester isolates of the same species is advisable, isolates of other heterothallic species (such as P. cryptogea or P. cambivora) may also be used. Characteristics such as the diameter of the oogonium and oospore, thickness of the oospore wall, whether or not the oospore fills the oogonium (plerotic), ornamentation on the oogonial wall, and mode of attachment of the antheridium are useful for species classification.
In an effort to simplify isolate identification and establish groupings of isolates for comparison of morphological features (but not phylogenetic relationships), Waterhouse (1963) introduced the concept of morphological groups I through VI based on a number of characteristics, and is still useful today. Unfortunately a dichotomous key that includes recently described species is not available for identification of isolates but there are several recent efforts to simplify morphological identification of species, including a manual for identification of 60 species of Phytophthora by integration of a dichotomous key with a DNA fingerprinting technique based on PCR-single strand conformational polymorphism (SSCP) (Gallegly and Hong 2008). A LUCID key for identification of 55 common Phytophthora spp. is available (Ristaino 2011) and an expanded LUCID key including most described species should be available on a dedicated website in the near future (G. Abad and Y. Balci, personal communication). A tabular presentation of morphological features enabling comparison among 117 species may be found in Martin et al. (2012; a downloadable file of the table alone is available on the journal website).
In 1999 the number of described species in the genus Phytophthora was approximately 55 (Brasier 2007) but since then there has been a significant increase., Brasier (2007) reported a doubling in number to 105 described species, with this number recently increasing to 117 (Martin et al. 2012). Additional species have recently been described; P. lacustris (Nechwatal et al. 2012) P. pluvialis (Reeser et al. 2013), P. mississippiae (Yang et al. 2013), P. cichorii, P. dauci and P. lactucae (Bertier et al. 2013), P. pisi (Heyman et al. 2013), P. stricta and P. macilentosa (Yang et al. 2014) and the hybrid species P. x serendipita and P. x pelgrandis (Man in ’t Veld et al. 2012), bringing the total to at least 128 described species. With the number of provisional species names used in the literature, and research efforts to evaluate the distribution of this genus in natural ecosystem, this number is likely to continue to increase in the future.
Historically the genus Phytophthora has been placed in the Pythialeswith Pythium and related genera but more recent phylogenetic analysis with the large (LSU) or small (SSU) rDNA sequences or cox2 gene has indicated a closer affiliation with downy mildews and white rusts (Albugo.) in the Peronosporales (Beakes and Sekimoto 2009; Thines et al. 2009). However, additional multigene analyses with a larger number of downy mildew species is needed to better characterize this relationship and the placement of Phytophthora spp. in clade 9 and 10 (Blair et al. 2008). The relationship between the Peronosporales and Pythium (Pythiales) needs clarification as well. A new genus, Phytopythium, was erected to accommodate an inconsistency between taxonomic and phylogenetic grouping for certain “intermediate” Pythium species (Bala et al. 2010), and it is likely that additional taxonomic revisions of the Peronosporomycetidae will be needed to fully resolve taxonomic conflicts.
Early efforts to understand phylogenetic relationships in Phytophthora focused on the use of the nuclear encoded rDNA, primarily the ITS region (Förster et al. 2000; Cooke and Duncan 1997; Crawford et al. 1996). Cooke et al. (2000) published the first comprehensive phylogenetic analysis of the genus using the ITS region to examine the phylogeny of 50 species. Most isolates grouped within eight primary clades (numbered 1 to 8) with several other species placed in two additional clades (clades 9 and 10). Kroon et al. (2004) expanded this analysis using two nuclear (translation elongation factor 1α, β-tubulin) and two mitochondrial (cox1 and nad1) genes. While in general the results were congruent with those reported by Cooke et al. (2000), there were some notable differences in the grouping of some species. Subsequent analysis by Blair et al. (2008) using seven nuclear genes (60S ribosomal protein L10, ß-tubulin, enolase, heat shock protein 90, large subunit rDNA, TigA gene fusion and translation elongation factor 1α) representing 8.1 kb of sequence data for 82 Phytophthora spp. clarified these differences. This larger, multi-marker analysis supported the observations of Cooke et al. (2000) with eight main clades plus two additional closely affiliated clades (clades 9 and 10) as the sister clades to the rest of the genus. More recently, Martin et al. (2014) expanded on this analysis by adding four mitochondrial genes (cox2, nad9, rps10 and secY) and additional species. The resulting phylogeny from this 11-marker analysis (10,828 bp per isolate) was similar to the prior observations of Blair et al. (2008) and subsequent analysis indicated that similar results could be obtained when using only five markers (LSU, β-tubulin, cox2, nad9 and rps10).
While the ITS region may be useful for species identification (see below), length variation among species makes it impossible to construct an unambiguous alignment across the entire genus, thus hampering the utility of this marker for phylogenetic analysis. Likewise, the translation elongation factor 1α has been used for phylogenetic analysis, but recent analysis of Phytophthora genomic data indicates that the gene is duplicated; divergence among duplicates may complicate phylogenetic interpretations of species evolution (J. E. Blair, unpublished).
While the above noted phylogenetic analyses have provided insight into the broader evolutionary relationships within the genus, there is still ambiguity when examining some closely related species and species complexes. Significant progress has been made with the clarification of the P. megasperma complex and other clade six species (Brasier et al. 2003; Durán et al. 2008; Hansen et al. 2009; Jung et al. 2011a, b) but there are still several provisional species awaiting more comprehensive analysis (for example, P. taxon PgChlamydo, P. taxon raspberry, P. taxon canalensis, P. taxon erwinii, P. taxon hungarica, P. taxon oregonsis and P. taxon paludosa). While there have been advances in understanding the relationships among some clade 2 species, there is need for additional analysis to clarify species complexes such as P. citricola and P. citrophthora. One clade 8 species complex where phylogenetic resolution has been elusive is P. cryptogea and the closely related species P. drechsleri. The multigene analysis of Mostowfizadeh-Ghalamfarsa et al. (2010) confirmed that while P. drechsleri was monophyletic, the P. cryptogea complex formed three well-defined phylogenetic groups with group I closely affiliated with P. erythroseptica and group II and III as a separate clade (group III isolates have been reported as the provisional species, P. sp. kelmania; Martin et al. 2014). Some isolates were placed intermediate between groups II and III and exhibited a greater amount of heterozygosity than the other isolates, suggesting possible outcrossing between these groups. Using a parsimony-based ancestral recombination graph and genealogies inferred from the β-tubulin and translation elongation factor 1-α genes from greenhouse recovered isolates, Olson et al. (2011) suggested that divergence between P. cryptogea and P. drechsleri was recent and that speciation is still in progress.
In addition to the choice of markers to use for phylogenetic analysis, another important consideration is the type of analysis used for estimating phylogenetic relationships or for the description of new species. While traditional methods of phylogenetic analysis (maximum likelihood, neighbour-joining, Bayesian) have adequately described relationships among most species, they have been unable to fully resolve the deeper relationships among the ten Phytophthora clades or among related genera. A recent study by Martin et al. (2014) used a novel variation of a multispecies coalescent approach to evaluate the ten clades; in general support was higher than that observed in the phylogenetic analysis for the recovered relationships, but the position of certain clades (Clade 3 and the unique grouping of P. sp. ohioensis and P. quercina) remained ambiguous. Here we present an analysis using a more powerful and complex Bayesian method (Drummond et al. 2012) with five genetic markers (Fig. 20), and recover strong support for basal relationships among the clades that are quite similar to the 11-marker study of Martin et al. (2014). Newer phylogenetic methods may allow for more complex modelling of the evolutionary process, however they are still sensitive to the accuracy of a priori information provided by the user. Additional studies will be needed to provide more basic information on the tempo of molecular evolution within this group.
The description of new species is also an area were traditional phylogenetic methods may not accurately describe species relatedness. Aside from morphological characterization, recent species descriptions typically contain molecular evidence from one or a few genetic markers (primarily ITS and perhaps cox1 or 2). However, as described above, alignment ambiguity and the presence of intraspecific polymorphisms can seriously impact the recovered phylogeny; recent hybridization events and incomplete lineage sorting of ancestral polymorphisms also violate the assumptions made by traditional phylogenetic methods. The use of coalescent-based approaches to estimate species trees from a collection of gene trees has been gaining popularity among many other taxonomic groups, but has seen little attention in Phytophthora or oomycete research in general. The recent description of P. pisi (Heyman et al. 2013) employed a multispecies coalescent approach, which confirmed the individual analyses of ITS and cox2 data. In addition, a recent study of the hybrid species P. andina (Blair et al. 2012) used several coalescent methods to determine the likely parental lineages of this species, one of which was clearly P. infestans. In the future, the use of more complex phylogenetic methods as well as coalescent-based approaches will be needed to clarify relationships at both ends of the spectrum, from deep basal nodes to recently evolved and potentially interbreeding species complexes.
A common observation among all phylogenetic studies is there is no consistent correlation between phylogenetic grouping and morphological features (Cooke et al. 2000; Kroon et al. 2004, 2012; Blair et al. 2008; Martin et al. 2014). While there is some correlation with sporangial type (clade 4, 5, and 10 have primarily papillate sporangia while clade 3 has primarily semipapillate sporangia and clades 6, 7, and 9 primarily nonpapillate sporangia), other clades show combinations of these features (clade 1, 2 and 8). Characteristics such as oogonial ornamentation, heterothallism, and mode of antheridial attachment are all polyphyletic.
Because of the large number of species, intraspecific variation of some morphological features, and overlapping morphology among closely related species, traditional methods of species identification can be challenging and require some level of expertise to be effective. The use of molecular criteria has simplified this task and provides a tool for delineating distinct taxa within morphologically similar species complexes. The most accurate molecular method for species identification is sequence analysis of specific markers. The internal transcribed spacer (ITS) region of the nuclear ribosomal DNA (rDNA) has been widely used and a large number of sequences are currently available in public databases. However, this marker may not be ideal for the identification of all species, especially those that are closely related. For example, many clade 1C species (P. infestans, P. mirabilis) cannot be distinguished using this marker alone, nor can P. fragariae and P. rubi. More recently a portion of the cox1 gene, along with the ITS region, have been proposed as the markers to use in the Barcode of Life Database (www.boldsystems.org) and representative sequences for all described and some provisional species have been deposited (Robideau et al. 2011).
Several nuclear (60S ribosomal protein L10, β-tubulin, enolase, heat shock protein 90, large subunit rRNA, TigA gene fusion, translation elongation factor 1α; (Blair et al. 2008; Kroon et al. 2004; Villa et al. 2006)) and mitochondrial (cox1, nad1, cox2, nad9, rps10 and secY; (Kroon et al. 2004; Martin 2008; Martin and Tooley 2003a, b; Martin et al. 2014) markers have been sequenced for phylogenetic analysis of Phytophthora and can also be used for species identification. Background information for amplification and sequencing of many of these markers, as well as the capability for BLAST searches against a curated database for isolate identification, may be found at the Phytophthora Database (www.phytophthoradb.org). A dataset for ITS and cox1 and 2 spacer sequences is also available at Phytophthora ID ((Grünwald et al. 2011), www.phytophthora-id.org) and sequence data for several markers (ITS, β-tubulin, elongation factor 1 alpha, and cox1), along with pictures of morphological features, may be found at Q-Bank (www.q-bank.eu).
There are several caveats to consider when using BLAST analysis to identify isolates to species level to prevent misidentification (Kang et al. 2010; Nilsson et al. 2012). BLAST scores are dependent on the length of the aligned sequences as well as the level of sequence identity; instances where high levels of sequence identity occur for only a portion of the target sequence may result in incorrect species identification. Also, it is common to encounter situations where scores are similar among multiple species, making it difficult to draw conclusions about an isolate’s identity (this can be especially problematic for isolates within or related to species complexes). In addition, the use of markers known to contain intraspecific polymorphisms may lead to inaccurate species identifications due to potentially lower similarities among closely related sequences. Heterozygosity in nuclear markers may also complicate identification efforts; while the presence of distinct alleles may indicate outcrossing (as Phytophthora is a known diploid), heterozygosity may also result from hybridization events between distinct lineages (as described above). Phylogenetic analysis of several markers is therefore suggested to confirm species identification, especially when working with species complexes. Additional gel based techniques, such as PCR-RFLP, SSCP, random amplified polymorphic DNAs (RAPDs), amplified fragment length polymorphisms (AFLP) and simple sequence repeat (SSR) analysis, for species identification and population analysis are reviewed in Martin et al. (2012).
Recommended genetic markers
The following genetic markers have been found to amplify well across all species and provided a similar level of phylogenetic resolution as a concatenated dataset of seven nuclear and four mitochondrial genes (Martin et al. 2014). Information on amplification and sequencing primers for these genes may be found at the Phytophthora Database (www.phytophthoradb.org).
Nuclear genes–LSU, β-tubulin
Mitochondrial–cox2, nad9, rps10
Sequence alignments of the seven nuclear and four mitochondrial markers used in Martin et al. (2014) and Fig. 20 may be downloaded at TreeBASE (http://purl.org/phylo/treebase/phylows/study/TB2:S14595). A table with additional information on isolates used in the analysis may be found in Martin et al. (2014) with GenBank accession numbers listed in the supplementary material of this citation. These sequences can also be downloaded from the Phytophthora Database (www.phytophthoradb.org).