Fungal Genetics and Biology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Fungal Genetics and Biology

journal homepage: www.elsevier.com/locate/yfgbi

Regular Articles A combined mitochondrial and nuclear multilocus phylogeny of the genus

a, b c Frank N. Martin ⇑, Jaime E. Blair , Michael D. Coffey a USDA-ARS, Salinas, CA, United States b Department of Biology, Franklin & Marshall College, Lancaster, PA, United States c Department of Plant Pathology & Microbiology, University of California, Riverside, CA, United States article info abstract

Article history: The most recent phylogenetic analysis of the genus Phytophthora was completed in 2008 (Blair et al., Received 23 September 2013 2008) and utilized 8.1 kb of sequence data from seven nuclear loci. Given the large number of species that Accepted 24 February 2014 have recently been described, this study was undertaken to broaden the available information on the Available online xxxx phylogeny of the genus. A total of 166 isolates representing 92 recognized species and 17 provisional spe- cies were analyzed, including many of the same isolates used in the nuclear multilocus study of Blair et al. Keywords: (2008). Four mitochondrial genes (cox2, nad9, rps10 and secY) were sequenced with a total of 2373 bp used in the analysis; the species relationships recovered with mitochondrial data were largely consistent Phytophthora with those observed previously in the nuclear analysis. Combining the new mitochondrial data with the Phylogeny Multispecies coalescent nuclear data from Blair et al. (2008) generated a dataset of 10,828 bp representing 11 loci, however res- olution of basal clade relationships was still low. We therefore implemented a modified multispecies coa- lescent approach with a subset of the data, and recovered increased resolution and moderate to high support for clade relationships. A more detailed analysis of species from clades 2 and 8 identified an addi- tional seven phylogenetic lineages that warrant further investigation to determine if they represent dis- tinct species. As has been reported in other phylogenetic studies of the genus, there was no consistent correlation between phylogenetic relatedness and morphological features or ecology. Published by Elsevier Inc.

1. Introduction and was recently reported to be approximately 117 species (Martin et al., 2012). With the recent descriptions of P. pluvialis (Reeser Species in the genus Phytophthora (Oomycetes) are capable of et al., 2013), P. mississippiae (Yang et al., 2013), P. cichorii, P. dauci infecting a wide range of plant species and can cause significant and P. lactucae (Bertier et al., 2013) and the hybrid species P.x damage to economically important crop plants worldwide. While serendipita and P.xpelgrandis (Man in‘t Veld et al., 2012) there some species have a narrow host range (e.g., P. infestans), others are at least 124 described species. Given the number of provisional are capable of infecting host species reflecting a wide range of species names used in the literature, this number will continue to plant genera (e.g., P. cinnamomi)(Erwin and Ribeiro, 1996). While increase in the future. There are several factors driving this resembling Eumycotan fungi with the production of hyphae, the expansion of species descriptions. The availability of DNA sequence genus is allied with the stramenopiles, a group more closely related data for multiple loci at websites like the Phytophthora to chromophyte algae and plants than to true fungi (Förster et al., Database (www.phytophthoradb.org, Park et al., 2008) and others 2000; Knoll, 1992; Baldauf and Palmer, 1993; Wainright et al., (www.phytophthora-id.org, www.q-bank.eu, www.boldsystems.org) 1993; Bhattacharya and Stickel, 1994; Weerakoon et al., 1998). In and the reduction in cost for generating such data has made it contrast to most Eumycotan fungi, Oomycetes are diploid through- easier for researchers to differentiate isolates from previously out their life cycle. described species, facilitating the identification of new lineages. Over the past 15 years there has been a significant expansion in Due to concerns with invasive species, there has also been an the description of species within the genus. In 1999 the number of increase in the number of surveys from previously under repre- species stood at approximately 55, but from 2000 to 2007 the sented ecosystems (such as streams and forests); in the process of number of valid species nearly doubled to 105 (Brasier, 2007) conducting these surveys a number of new species have been iso- lated and described. Given the increasing problems with invasive species such as P. ramorum, P. kernoviae and P. alni impacting com- Corresponding author. ⇑ E-mail address: [email protected] (F.N. Martin). mercial agriculture and natural ecosystems (Hansen et al., 2012), http://dx.doi.org/10.1016/j.fgb.2014.02.006 1087-1845/Published by Elsevier Inc.

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 2 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx and the quarantine restrictions put in place to prevent their spread, (http://phytophthora.ucr.edu). Many of these isolates were also it is imperative that molecular techniques be available to accu- used in the prior analysis of the genus by Blair et al. (2008). A total rately identify isolates to the species level as well as provide an of 92 recognized species (including 30 type cultures), 17 provi- easy means for identification of isolates representing new species. sional species, and seven distinct phylogenetic taxa are repre- Establishing a robust phylogenetic framework is therefore crucial sented in the analysis. Rather than give provisional names for for isolate identification and also for delimiting species boundaries. these phylogenetic taxa without having conducted a careful mor- A well resolved phylogeny will also assist in the analysis of geno- phological analysis, the following naming convention was used: mic sequencing studies as well as allow us to predict the behavior Phytophthora sp. affinis followed by the name of the most closely and ecological functions of new species based on the characteris- related species. In cases where more than one distinct entity is tics of their close relatives, which may be particularly important encountered a number is added sequentially to the name. when new invasive species are identified. To be consistent with Mostowfizadeh-Ghalamfarsa et al. (2010) Early studies on the phylogeny of the genus utilized a limited the naming of P. cryptogea isolates in this study followed a similar number of species and relied mainly on portions of the ribosomal convention. Isolate P1739 was in GI of Mostowfizadeh-Ghalamf- DNA (rDNA), primarily the internal transcribed spacer (ITS) region arsa et al. (2010) and was labeled as P. cryptogea in this study. In (Förster et al., 2000; Cooke and Duncan, 1997; Crawford et al., addition, the ITS sequences of isolates from Mostowfizadeh-Gha- 1996) or mitochondrial genes (Martin and Tooley, 2003a, 2003b). lamfarsa et al. (2010) were compared to isolates used in this cur- The first comprehensive examination of a wide range of species rent study; our P. cryptogea isolates (P1088, FJ801524) in the genus was reported by Cooke et al. (2000) using the ITS re- correspond to the P. cryptogea GI isolates (SCRP230, AY659441) gion to examine the phylogenetic relationship among 50 species. A of Mostowfizadeh-Ghalamfarsa et al. (2010). The P. sp. aff. crypto- total of 8 primary clades were identified with two additional gea GII isolates (P1380, FJ802061) corresponded to P. cryptogea clades, 9 and 10, consisting of P. macrochlamydospora and P. richar- GII (SCRP204, AY659422) while the P. sp. kelmania isolates diae as well as P. insolita, respectively. While there was some cor- (P10367, FJ801461) corresponded to P. cryptogea GIII (SCRP209, relation between phylogenetic grouping and morphological AY659426); this latter grouping was also supported by elongation features like sporangial shape, this was not absolute and many factor 1 a data (P10613, EU079607 vs. SCRP209, AY659519). The clades show more than one sporangial type. Using a similar num- grouping of P. cryptogea GI and GII isolates used in this study were ber of species, Kroon et al. (2004) expanded the analysis of the also supported by cox1 sequence data (data not shown; GenBank genus by using two nuclear (translation elongation factor 1a, accession numbers in Supplementary Table S1). b-tubulin) and two mitochondrial (cox1 and nad1) genes. While in A number of the cultures included in this analysis were part of general the results were congruent with those reported by Cooke other phylogenetic studies as well. For example, representatives of et al. (2000), there were some notable exceptions. For example, clade 2 species that were part of the P. citricola species complex in- the species in clade 9 and 10 of Cooke et al. (2000) were included clude isolate P10204, which is isolate P53 of Hong et al. (2011) that as part of clade 8 rather than as separate clades. Likewise, some was used in their resurrection of P. pini as a species. Isolate P10338 species were placed in different clades (P. tentaculata was placed in P. citricola clade E is IMI031372 in Jung and Burgess (2009) in clade 2a with P. multivesiculata rather than in clade 1 as reported where this phylogenetic species is described and isolates P6624 by Cooke et al., 2000). The most recent analysis of the genus was and P1321 were used as representatives of this entity in reported by Blair et al. (2008) using seven nuclear genes represent- Bezuidenhout et al. (2010). Isolates P7902, P10458 and P1817 were ing 8.1 kb of sequence data for 82 Phytophthora spp. This larger used as representatives of P. multivora in the analysis of Jung and multilocus analysis supported the observations of Cooke et al. Burgess (2009). Isolates P1819, P1822, and P1823 of P. capensis (2000) with 8 main clades plus two additional closely affiliated were included in the taxonomic description of this species by clades (clades 9 and 10) basal to the rest of the genus. Bezuidenhout et al. (2010), as were P. multivora P1817 and P. men- A number of new species have been described since Blair et al. gei P1165. Additional cox1 sequencing was used to confirm the (2008) and while some of these descriptions have included a com- identification of P. plurivora, P. multivora, P. pini, and P. cryptogea prehensive evaluation of phylogenetic relationships, for others isolates (data not shown; GenBank accession numbers in only limited information was provided. Also, while prior analysis Supplementary Table S1). The name of P. austrocedrae was changed (Cooke et al., 2000; Kroon et al., 2004; Blair et al., 2008) provided to P. austrocedri as suggested by Kroon et al. (2012). an excellent view of the terminal relationships among species Several isolates used in the prior study of Blair et al. (2008) were within the ten described clades, bootstrap support was not robust reclassified based on additional DNA sequence analysis, including enough to conclusively delineate the relationships among all the P. citricola (P7902), which was confirmed as P. multivora based on clades. The objective of this current work was to improve phyloge- cox1 sequence analysis (data not shown), P. gonapodyides netic resolution of the genus by including newly described species (P10337) was reclassified as P. lacustris, P. richardiae (P10811) and by adding mitochondrial gene data to the analysis. Many of the was reclassified as P. sp. aff. erythroseptica (since P6875 was iden- same isolates used in Blair et al. (2008) were included in the cur- tified by Buisman, who described this species, it was selected as a rent analysis, thereby allowing for concatenation of the mitochon- representative of P. richardiae) and P. porri (P10728) was reclassi- drial gene data with the previously reported nuclear data. In fied as P. sp. aff. brassicae-1 (isolate P7518 was selected as a repre- addition to traditional phylogenetic methods, we have also imple- sentative of P. porri). mented a modified multispecies coalescent approach to specifi- cally address the relationships among the ten Phytophthora clades. 2.2. Template amplification and sequencing

DNA was extracted from cultures as previously described (Blair 2. Materials and methods et al., 2008). The primers used for amplification and sequencing of mitochondrial loci (cox2, nad9, rps10 and secY) are listed in Table 2. 2.1. Cultures All amplifications were done with approximately 10 ng template

DNA, 0.5 lM forward and reverse primers, 2 or 3 mM MgCl2 The cultures used in this investigation are listed in Table 1 and (2 mM for cox1, 3 mM for others), 100 lM dNTP, 1x amplification with few exceptions were obtained from the World Ge- buffer and 1 unit of AmpliTaq (Applied Biosystems, Foster City, netic Resource Collection at the University of California, Riverside CA) in a volume of 25 ll. Templates were amplified in an ABI

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 3

Table 1 Isolates used in this study.

Isolate Details Phytophthora Isolate Identification Isolate Origins Waterhouse Cladea Species Localb Internationalc Host Country Group P. sp. ohioensis P16050 Oak forest soil Ohio I P. quercina (T) P10334 CBS782.95 Quercus robur Germany I 1 P. nicotianae P6303 Grammatophyllum sp. Indonesia II 1a P. cactorum P0714 ATCC10091, CBS231.30 Syringa vulgaris The Netherlands I P. sp. aff. cactorum P6625 Fragaria sp. Taiwan NAd P. hedraiandra P11056 Rhododendron sp. USA I P. idaei (T)e P6767 CBS971.95, IMI313728 Rubus idaeus UK I P. pseudotsugae (T) P10339 IMI331662 Pseudotsuga menziesii USA I 1b P. clandestina P3942 ATCC58715, CBS349.86 Trifolium subterraneum Australia I P. iranica (T) P3882 ATCC60237, CBS374.72, IMI158964 Solanum melongena Iran I P. tentaculata P8497 CBS552.96 Chrysanthemum leucanthemum Germany I 1c P. andina (T) P13365 Solanum brevifolium Ecuador IV P. infestans P10650 Solanum tuberosum Mexico IV P. ipomoeae P10225 Ipomoea longipedunculata Mexico IV P. mirabilis P3005 ATCC64068 Mirabilis jalapa Mexico IV P. phaseoli P10145 Phaseolus lunatus USA IV

2 P. bisheria (T) P10117 CBS122081 Fragaria sp. USA III P. capensis (T) P1819 Curtisia dentata South Africa III P1822 Stream water South Africa P1823 Olea campensis South Africa P. citricola (T) P0716 ATCC60440, CBS221.88, IMI21173 Citrus sinensis Taiwan III P1814 ATCC64811 Citrus sp. South Africa P. citricola clade Ef P1321 ATCC64809 Rubus sp. California NA P6624 ATCC66621 Fragaria x ananassa P. multivesiculata (T) P10410 CBS545.96 Cymbidium sp. The Netherlands IV P10327 CBS101593 Cymbidium sp. The Netherlands P10525 CBS101594 Cymbidium sp. The Netherlands P. multivora P1817 Medicago sativa South Africa III P7902 Pinus radiata California, USA P10458 Gardenia sp. California, USA P. pini P10204 P53g Rhododendron sp. West Virginia III P0767 Syringa sp. Canada P10762 Fagus grandifolia New York P10763 Fagus grandifolia New York P10764 Fagus grandifolia New York P. plurivora P1805 Humulus lupulus California, USA III P7491 IMI34289 Sambucus tenuifolium England P10623 Abies fraseri North Carolina P10627 Abies fraseri North Carolina P10679 Juglans regia New Zealand 2a P. botryosa (T) P3425 CBS581.69, IMI136915 Hevea brasiliensis Mylasia II P6945 IMI130422 Hevea brasiliensis Mylasia P. citrophthora P6310 Theobroma cacao Indonesia II P. colocasiae P6317 Colocasia esculenta Indonesia IV P. sp. aff. colocasiae-1 P0318 ATCC64851 Citrus sp. Australia NA P10341 IMI342898 Syringa sp. England P. sp. aff. colocasiae-2 P6262 Hevea brasiliensis India NA P. meadii P6128 ICRI-240 Elettaria cardamomum India II 2b P. capsici (T) P3605 ATCC52771, CBS128.23, IMI40502 Capsicum annuun New Mexico II P0253 ATCC46012 Theobroma cacao Mexico P. glovera P10619 Nicotiana tabacum Brazil III P. mexicana (T) P0646 ATCC46731, CBS554.88, IMI092550 Solanum lycopersicum Mexico II P. tropicalis (T) P10329 CBS434.91 Macadamia integrifolia USA (Hawaii) II P. sp. brasiliensis P0630 ATCC46705 Theobroma cacao Brazil NA P. mengei P1273 ATCC58657 Persea americana California III P1825 Vigna unguiculata Australia P1165 Persea americana Guatemala P. siskiyouensis P15123 Umbellularia californica Oregon III P. sp. aff. siskiyouensis P1200 ATCC64812 Theobroma cacao Brazil NA

3 P. ilicis P3939 ATCC56615 Ilex sp. Canada IV P. nemorosa P10288 Lithocarpus densiflorus USA IV P. pseudosyringae (T) P10437 CBS111772 Quercus robur Germany III P. psychrophila (T) P10433 CBS803.95 Quercus robur Germany IV

4 P. megakarya P8516 Theobroma cacao Sao Tome II P. palmivora P0255 ATCC26200 Theobroma cacao Costa Rica II (P. arecae) P10213 ATCC-MYA4039 Citrus sp USA P. quercetorum (T) P15555 ATCC-MYA4086, CBS 12119 Quercus rubra rhizosphere USA I

(continued on next page)

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 4 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

Table 1 (continued)

Isolate Details Phytophthora Isolate Identification Isolate Origins Waterhouse Cladea Species Localb Internationalc Host Country Group

5 P. heveae (T) P3428 CBS296.29, IMI180616 Hevea brasiliensis Malaysia II P10167 ATCC16701, IMI131373 soil USA P. katsurae P10187 ATCC MYA-4060 Castanea crenata Japan II P. sp. novaeguineae P1256 ATCC46427, IMI325912 Soil under Auracaria Papua New Guinea NA

6 P. sp. asparagi P10690 Asparagus officalis New Zealand VI P. sp. canalensis P10456 Canal water California NA P. sp. erwinii P3132 Banksia integrifolia Australia NA P. gemini (T) P15880 CBS123381 Zostera marina Netherlands V/VI P. gonapodyides DC-39 P501h Salix matsudana UK V/VI P. humicola (T) P3826 ATCC52179, CBS200.81, IMI302303 citrus grove soil Taiwan V P. sp. hungarica P10281 Alder root soil Hungary NA P. inundata P8478 IMI389750 Aesculus hippocastanum UK VI P. megasperma P3136 Brassica napus var. napus Australia V P. sp. personii P11555 Nicotiana tabacum USA V/VI P. sp. PgChlamydo P10669 Idesia polycarpa New Zealand V/VI P. pinifolia (T) P16100 CBS122924 Pinus radiata Chile V/VI P. taxon raspberry DC-36 P896h Soil, dying vegetation Tasmania V/VI P. rosacearum P3315 Prunus persica Ohio V P. sp. sulawesiensis P6306 Syzygium aromaticum Indonesia NA P. lacustris (T) P10337 PPIHAS-P566 Salix matsudana England V/VI

7a P. alni subsp. alni (T) P16203 IMI392314 Alnus glutinosa Netherlands VI P10568 Alnus sp. Hungary VI P. alni subsp. uniformis (T) P16206 IMI392315 Alnus sp. Sweden VI P. cambivora P0592 ATCC46719 Abies procera USA VI P. europaea (T) P10324 CBS109049 Quercus rhizosphere France V P. fragariae P3821 ATCC36057 Fragaria sp. UK V P. rubi P3289 Rubus sp. New York V P. uliginosa P10328 CBS109055 Quercus robur rhizosphere Germany V 7b P. cajani P3105 ATCC44388 Cajanus cajan India VI P. cinnamomi (T) P2110 ATCC46671, CBS144.22, IMI22938 Cinnamomum burmannii West Sumatra VI P2159 ATCC46678, IMI157799 Vitis vinifera South Africa P. parvispora P8495 CBS413.96 Beaucamea sp. Germany VI P. cinnamomi var. robiniae P16350 ATTC 56194 Robinia pseudoacacia China NA P. melonis P10994 Trichosanthes dioica India VI P. sp. niederhauserii P10617 Thuja occidentalis North Carolina VI P. pistaciae P6197 ATCC62268 Pistacia vera Iran V P. sojae P3114 Glycine max USA V P. vignae P3019 Vigna unguiculata Australia VI

8a P. cryptogea GI P1088 ATCC46721, CBS290.35 Callistephus chinensis California VI P1739 IMI45168 Solanum lycopersicum New Zealand P7788 IMI328662 Daucus carota England P10672 Olearia paniculata New Zealand P. cryptogea GII P1380 Vitis vinifera South Africa P2001 Malus sylvestris Australia P3103 ATCC52402 Solanum marginatum Ecuador P. sp. kelmania P10613 Abes fraseri North Carolina, USA V/VI (P cryptogea GIII) P1810 Prunus sp. California, USA P. drechsleri P10331 Gerbera jamesonii New Hampshire, USA P11637 ATCC46724, CBS292.25 Beta vulgaris California, USA P. erythroseptica P1699 ATCC36302, CBS956.87 Solanum tuberosum Ohio, USA VI P10382 Solanum tuberosum Idaho, USA P10385 Solanum tuberosum Idaho, USA P. sp. aff. erythroseptica P3876 ATCC46734, CBS240.30 Zantedeschia aethiopica USA NA P10811 Zantedeschia aethiopic Japan P10359 Zantedeschia sp. Japan P10358 Zantedeschia sp. Japan P10355 Zantedeschia sp. Japan P. medicaginis P10683 Medicago sativa California, USA V P0127 Medicago sativa Australia P1678 Malus sylvestris Oregon, USA P. sansomeana P3163 Silene latifolia subsp. alba New York, USA V P. trifolii P7010 Trifolium sp. Mississippi, USA V 8b P. brassicae P10414 CBS113350 Brassica oleraceae The Netherlands IV P3273 CBS212.82 Brassica oleraceae Netherlands P3828 ATCC56354, CBS112967, Brassica oleraceae United Kingdom P10153 IMI195524 Brassica oleraceae Netherlands P10155 CBS686.95 Brassica oleraceae Netherlands P. porri P7518 CBS181.87 Allium porrum Netherlands III P. sp. aff. brassicae-1 P10728 Daucus carota France NA

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 5

Table 1 (continued)

Isolate Details Phytophthora Isolate Identification Isolate Origins Waterhouse Cladea Species Localb Internationalc Host Country Group P. sp. aff. brassicae-2 P6207 CBS112968 Allium cepa Switzerland NA P. primulae P10333 CBS620.97 Primula acaulis Germany III P10224 CBS172.95 P. austrocedri (T) P16040 MYA-4074 Austrocedrus chilensis Argentina IV P15132 MYA-4073 Austrocedrus chilensis Argentina P. syringae P10330 CBS110161 Rhododendron sp. Germany III 8c P. foliorum P10969 Rhododendron sp. CA, USA III P. hibernalis P3822 ATCC56353, CBS114104, IMI134760 Citrus sinensis Australia IV P0647 ATCC32995, CBS 953.87, IMI325911 Citrus sinensis California P. lateralis P3888 ATCC28512 Chamaecyparis lawsoniana USA V P1728 Chamaecyparis lawsoniana P. ramorum P10301 CBS101329 Rhododendron sp. The Netherlands IV

9 P. captiosa (T) P10719 ICMP15576 Eucalyptus saligna New Zealand VI P. sp. cuyabensis P8213 rainforest soil Ecuador V P. fallax P10725 NZFS1719 Eucalyptus fastigata New Zealand P. hydropathica P16857 Stream baiting West Virginia, USA P. irrigata (T) P16861 ATCC MYA-4457, 23J7 Irrigation reservoir West Virginia, USA P. insolita (T) P6195 ATCC38789, CBS691.79, IMI288805 soil Taiwan VI P. sp. lagoariana P8223 rainforest soil Ecuador NA P. macrochlamydospora P10267 Glycine max Australia VI P. sp. napoensis P8221 rainforest Ecuador NA P. parsiana (T) P15164 IMI395329 Ficus carica Iran VI P. polonica P15005 Alnus glutinosa rhizosphere Poland V P. quininea P3247 ATCC56964, CBS406.48 Cinchona officinalis Peru V P. richardiae P6875 ATCC60353, CBS240.30 Zantedeschia aethiopica USA VI P. sp. thermophilum P10457 Canal water California NA

10 P. boehmeriae (T) P6950 CBS291.29, IMI180614 Boehmeria nivea Taiwan II P. kernoviae P10681 Annona cherimola New Zealand II P. gallica P16826 CBS111474 V/VI

Outgroup Pythium undulatum P10342 IMI337230 Larix sp. UK Phytopythium vexans P3980 ATCC12194, CBS340.49, IMI32044 NA NA

a Molecular clade as shown in Fig. 1. b Local identification numbers from the World Phytophthora Genetic Resource Collection (P). c International identification abbreviations: ATCC, American Type Culture Collection, USA; CBS, Centraalbureau fur Schimmelcultures, The Netherlands; IMI, CABI Bio- sciences, UK; PPIHAS, Plant Protection Institute of the Hungarian Academy of Science; ICMP, International Collection of Microorganisms from Plants, Auckland, NZ; NZFS, New Zealand Forest Service. d NA = not available. e Type isolate of the species (T). f Clade E of Jung and Burgess (2009). g From Hong et al. (2011). h From Brasier et al. (2003).

9600 thermal cycler with the following cycling conditions: 1 inter- P. plurivora (P10679), P. sp. sulawesiensis (P6303) and P. sp. aff. val of 95 °C for 3 min; 35 cycles of 95 °C for 1 min, 1 min annealing cryptogea group II (P3103) using previously reported procedures for the indicated temperature (Table 2) and extension at 72 °C for (Blair et al., 2008). Heterozygous bases were coded with the 2 min; 1 interval of 72 °C for 5 min followed by a 10 °C hold. After appropriate IUPAC codes and consensus sequences were used in confirming template amplification by running samples on an aga- subsequent phylogenetic analyses. rose gel, sequencing templates were prepared by treatment with ExoSap-IT (USB, Cleveland, OH) in accordance with the manufac- 2.3. Phylogenetic analysis turer’s instructions and sent to the Penn State Genomics Core Facil- ity of the Huck Institute for the Life Sciences (University Park, Sequences for the mitochondrial loci were aligned with DS Gene Pennsylvania) for Sanger sequencing (Applied Biosystems v2.5 (Accelrys, San Diego, CA); alignments were adjusted manually 3730XL) with the amplification primers unless otherwise noted as needed in MacClade v4.02 (Sinaur Associates, Sunderland, MA). (Table 2). Each template was sequenced in both directions to gen- Nuclear loci were aligned individually using ClustalX v2.1 (Larkin erate a consensus sequence based on complementary strands in et al., 2007) and alignments were adjusted manually as needed Sequencher 4.7 (Gene Codes, Ann Arbor, MI). Heterozygous bases in MEGA v4.1 (Tamura et al., 2007); heterozygous bases were trea- were not observed in this sequence data. All sequences have been ted as missing data. Three datasets were constructed for phyloge- deposited in GenBank (Supplementary Table S1); mitochondrial netic analysis: mitochondrial data only, nuclear data only, and a sequences for additional isolates may be downloaded from the concatenation of mitochondrial and nuclear data for those isolates Phytophthora Database (www.phytophthoradb.org). present in both datasets. Sequences from two different isolates of Sequences for the nuclear loci (60S ribosomal protein L10, ß- P. alni subsp. alni were used in the combined analysis as shared tubulin, enolase, heat shock protein 90, large subunit rDNA, TigA data were not available (P10568 for the nuclear data, P16203 for gene fusion, translation elongation factor 1a) were the same data- the mitochondrial data). Each dataset was analyzed using Model- set used in Blair et al. (2008) with additional sequence data gener- Test v3.7 (Posada and Crandall, 1998), and the general time revers- ated for P. siskiyouensis (P15123), P. sp. brasiliensis (P0630), ible nucleotide substitution model with gamma-distributed rate

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 6 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

Table 2 Primers used for amplification and sequencing.

Markera Amplification primers Amplicon sizeb (bp) Annealing temperature (°C) cox2 + spacer FM 35 CAGAACCTTGGCAATTAGG 1041 54 Phy10b GCAAAAGCACTAAAAATTAAATATAA nad9 Nad9-F TACAACAAGAATTAATGAGAAC 1024 61 Nad9-R GTTAAAATTTGTACTACTAACAT rps10 Prv9-F GTATACTCTAACCAACTGAGT 601 59 Prv9-R GTTGGTTAGAGTAAAAGACT secY SecY-F TCTATCGTGTTTACCAATTTC 946 61 SecY-R TAACAAATGGATCTTCTTTAAAA cox1 FM84 TTTAATTTTTAGTGCTTTTGC 1208 56 FM77 CACCAATAAAGAATAACCAAAAATG or FM83 CTCCAATAAAAAATAACCAAAAATG

Internal sequencing primers cox2 + spacer FM82 TTGGCAATTAGGTTTTCAAGATCC FM80 AATATCTTTATGATTTGTTGAAA cox1 FM50 GTTTACTGTTGGTTTAGATG FM85 AACTTGACTAATAATACCAAA

a FM35 was previously reported in Martin (2000), Phy10b in Martin et al. (2004), secY (ymf16) in Martin (2008) and nad9 and rps10 in Martin and Coffey (2012) and cox1 in Martin and Tooley (2003a). b Size may vary slightly depending on the isolate.

variation and a proportion of invariable sites (GTR + I + G) was the birthRate and popMean parameters (mean = 1.0, initial va- identified as the most appropriate model for each dataset based lue = 1.0). Preliminary analyses were run for ten million genera- on both the hierarchical likelihood ratio test and the Akaike infor- tions to evaluate prior settings; the final analysis was run twice mation criterion. Maximum likelihood analyses were performed for 50 million generations and evaluated in Tracer. The first five with GARLI v1.0 (http://garli.googlecode.com); parameter values million generations were removed as burn-in, and the maximum were estimated via two initial searches, then fixed for a bootstrap clade credibility tree was calculated in TreeAnnotator v1.7.5. analysis with 1000 replicates. Maximum parsimony analyses were conducted with the PHYLIP package (Felsenstein, 2005); 500 boot- 3. Results strap replicates were generated with SEQBOOT and analyzed in DNAPARS. Bayesian analyses were performed with MrBayes The four mitochondrial loci analyzed in this study showed var- v3.1.2 (Ronquist and Huelsenbeck, 2003); two analyses were run iation in amplicon length as well as in level of sequence diver- simultaneously for two million generations (four million for the gence. The average pairwise distance among isolates (Kimura-2 mitochondrial only dataset) with three heated chains (tempera- parameter model) for the cox2, nad9, rps10 and secY loci was ture = 0.2) and one cold chain. Flat Dirichlet priors were used for 0.071, 0.075, 0.096, and 0.104, respectively. Variation in length of rate parameters and nucleotide base frequencies; a uniform prior the genes was observed with cox2 ranging from 678 to 687 bp (0, 1) was used for the gamma shape parameter and the proportion (clade 3 species showed a 6 bp deletion before the termination co- of invariable sites. Log files were viewed in Tracer v1.4 (Rambaut don relative to other species while species in the P. brassicae clade and Drummond, 2007) to evaluate chain mixing and convergence, had a 3 bp insertion relative to the other species immediately be- and to estimate the appropriate burn-in (10% for each analysis). fore the termination codon of the gene), nad9 ranged from 558 to Majority-rule consensus trees were calculated after removing 567 bp (some species had a 9 bp deletion relative to other species the burn-in. Analyses limited to the Clade 2 and Clade 8 species just before the termination codon), rps10 ranged from 318 to were conducted as described above. All datasets have been depos- 348 bp (clade 9 and 10 species had a 9 bp deletion relative to other ited in TreeBASE (http://purl.org/phylo/treebase/phylows/study/ species between base 73–81 while P. asparagi had an insertion just TB2:S14595). before the termination codon) and secY ranged from 741 to 762 bp A modified multispecies coalescent approach was implemented (due primarily to insertions just before the termination codon in * in BEAST v1.7.5 (Drummond et al., 2012) using the BEAST option various species). The concatenated alignment totaled 2373 bp. Se- to specifically address the relationships among clades. Two nuclear quence data for a total of seven nuclear loci representing 8455 bp loci (ß-tubulin and large subunit rDNA) and the concatenated were also used in the analysis. There was a greater level of se- mitochondrial data were used in this analysis as they each showed quence divergence for the mitochondrial data compared to the nu- strong support for the monophyly of individual clades. XML-for- clear data (average pairwise distance of 0.086 vs. 0.068, matted input files were prepared using BEAUti v1.7.5; clade affin- respectively), which is also reflected in the branch lengths ob- ity was indicated for each isolate under the Traits tab and the served in the individual analyses (Figs. 1 and 2). GTR + I + G model was specified for each locus (with the mitochon- The topologies of the mitochondrial and nuclear gene trees drial data treated as a single locus given their assumed linkage). were generally congruent, with both datasets showing strong sup- Uniform priors were used for rate parameters (0, 10) and the pro- port for two main monophyletic lineages within Phytophthora, the portion of invariable sites (0, 1); an exponential prior (mean = 1.0) grouping of Clades 1 through 8 and a separate grouping of Clades 9 was used for the gamma shape parameter. Initial values for each and 10. Clades 1 through 5 were found to form a monophyletic prior were set to the ModelTest estimates for each locus. A strict group in both datasets, but the positions of Clades 6, 7, and 8 var- clock was used to model rate variation with an exponential prior ied, with clade 8 being the most basal of the major clades in the nu- on the rate (mean = 0.1, initial value = 0.1), and a Yule process clear analysis and clade 7 being most basal in the mitochondrial was used as the species tree prior with an exponential prior on analysis. The combined analysis of both the mitochondrial and nu-

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 7 clear datasets (Fig. 3), totaling 11 loci and 10,828 bp, generated a similar topology, with low support for relationships among Clades 1 through 8. The use of a modified multispecies coalescent analysis on a subset of the data produced a topology with improved resolution among the clades (Fig. 4), showing strong support for the monophyly of Clades 1 through 5, 1 through 8, 9 and 10, and a grouping of Clades 6 and 7. In all analyses, P. quercina and P. sp. ohioensis formed a well-supported group with affinities to Clade 1, however additional information will be needed to unambigu- ously assign these two species to a specific clade. A mitochondrial multilocus analysis focusing on clade 2 species (Fig. 5) revealed several distinct phylogenetic groupings that did not correspond to described species. In clade 2a there were two such groupings; P. sp. aff. colocasiae-1 and -2, which were initially labeled as P. inflata or P. citrophthora and P. meadii, respectively. In clade 2b there was also a distinct phylogenetic group, P. sp. aff. sis- kiyouensis, that was basal to P. siskiyouensis. The position of the provisional species P. sp. brasiliensis remained ambiguous, as it was found to be basal to P. capsici, P. glovera and P. mexicana in the mitochondrial analysis whereas in the nuclear analysis it was basal to P. siskiyouensis; bootstrap support for these relationships was generally low. A mitochondrial multilocus analysis focusing on clade 8 species (Fig. 6) identified four putative new species. There was strong bootstrap support separating P. sp. aff. erythro- septica from P. erythroseptica and P. sp. aff. cryptogea GII from P. cryptogea. Two putative new species were identified that were ba- sal to P. brassicae, P. sp. aff. brassicae-1 and P. sp. aff. brassicae-2.

4. Discussion

Overall the phylogenetic relationships among species observed in the mitochondrial gene analysis and the combined mitochon- drial and nuclear datasets were similar to the nuclear analysis pre- viously reported (Blair et al., 2008), however, some minor differences were observed. One particular relationship to note is the position of P. alni subsp. alni; there was strong bootstrap sup- port for placement of the type isolate, together with the P. alni subsp. uniformis type isolate, basal to other clade 7a species in the mitochondrial analysis, while in the nuclear analysis another isolate of P. alni subsp. alni grouped closely with P. cambivora and P. fragariae. Given that this species is a hybrid (Brasier et al., 1999) this difference in position likely reflects uniparental inheri- tance of the mitochondria, which is a factor that should be kept in mind when evaluating mitochondrial phylogenies. In their study of P. alni subspecies, Ioos et al. (2006) suggested that P. alni subsp. alni may have arisen from hybridization events between P. alni subsp. uniformis and P. alni subsp. multiformis due to the presence of three alleles in the nuclear loci of P. alni subsp. alni (two similar to multiformis and one similar to uniformis). The authors also found that some isolates of P. cambivora shared alleles with P. alni subsp. uniformis for three of the four nuclear loci they examined (Ioos et al., 2006). Our results may support this hypothesis for the origin of P. alni subsp. alni as the isolate used in our nuclear analysis (P10568) was found to have the multiformis mitochondrial haplo- type (Paa93, M0) by Ioos et al. (2006). While heterozygous bases were present in our nuclear data for this isolate, the proportion was similar to other unrelated isolates and did not suggest the presence of three alleles. However, additional sequence data, espe- cially from P. alni subsp. multiformis, is needed to determine the number and order of hybridization events that led to the formation Fig. 1. Genus wide phylogeny for Phytophthora using four mitochondrial loci (2373 of these lineages. Similarly, a recent analysis of both mitochondrial nucleotides). Maximum likelihood branch lengths shown. Numbers on nodes and nuclear data from the hybrid species P. andina suggested that represent bootstrap support values for maximum likelihood (top), maximum two separate hybridization events had occurred, as two distinct parsimony (middle) and Bayesian posterior probabilities as percentages (bottom). mitochondrial haplotypes were present among isolates of this spe- Nodes receiving significant support (>95%) in all analyses are marked with an * cies (Blair et al., 2012). asterisk ( ). Scale bar indicates number of substitutions per site.

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 8 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

Fig. 2. Genus wide phylogeny for Phytophthora using seven nuclear loci (8455 nucleotides). Maximun likelihood branch lengths shown. Numbers on nodes represent bootstrap support values for maximum likelihood (top), maximum parsimony (middle) and Bayesian posterior probabilities as percentages (bottom). Nodes receiving significant support (>95%) in all analyses are marked with an asterisk (*). Scale bar indicates number of substitutions per site.

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 9

Fig. 4. Topology of Phytophthora clades estimated using a modified multispecies coalescent approach (two nuclear and four mitochondrial loci, 4815 nucleotides). Numbers on nodes represent Bayesian posterior probabilities.

Similar to Blair et al. (2008), our results also support a basal posi- tion to subclades 1b and 1c, although bootstrap support varied among datasets. In the nuclear analysis, there was strong support for the basal position of P. cinnamomi and P. parvispora within clade 7b; the mitochondrial analysis showed these two species as basal to the entire clade 7, although with low bootstrap support. Our re- sults also clearly support the classification of P. parvispora (Scanu et al., 2013) and P. cinnamomi var. robiniae as species distinct from P. cinnamomi. A multilocus analysis of a larger number of isolates is currently in progress to clarify the taxonomic classification of P. cinnamomi var. robiniae (M.D. Coffey, F.N. Martin and M. A. Mans- field, unpublished). As noted in Mostowfizadeh-Ghalamfarsa et al. (2010), the P. richardiae isolate used in Blair et al. (2008) (P10811) was misidentified as it grouped in clade 8a. Isolate P6875, which was identified as P. richardiae by Buisman, the descriptor of the species, clustered within clade 9 as seen in other phylogenetic analyses (Cooke et al., 2000; Mostowfizadeh-Ghalamfarsa et al., 2010; Kroon et al., 2004). While combining the four mitochondrial and seven nuclear gene loci into a single 10,828 bp dataset increased bootstrap sup- port for some relationships within clades and increased support for the monophyly of clades 1–8 and clades 9–10, it did not im- prove the phylogenetic resolution among clades 1 through 8. We therefore implemented a modified multispecies coalescent ap- proach, which reconstructs the most likely species tree from a set of gene trees (Degnan and Rosenberg, 2009), by treating clade membership as the trait of interest (so each clade represented a Fig. 3. Genus wide phylogeny for Phytophthora using seven nuclear and four ‘‘species’’ in the analysis). This analysis supported the grouping mitochondrial loci (10,828 nucleotides). Maximum likelihood branch lengths of clades 1 through 5, the basal position of clade 8 to clades 1 shown. Numbers on nodes represent bootstrap support values for maximum through 7 and the grouping of clades 9 and 10 basal to clades likelihood (top), maximum parsimony (middle) and Bayesian posterior probabilities 1–8 (Fig. 4). In addition, there was strong support for a grouping as percentages (bottom). Nodes receiving significant support (>95%) in all analyses are marked with an asterisk (*). Scale bar indicates number of substitutions per site. of clades 6 and 7, which was observed in the analysis of nuclear (Fig. 2) but not mitochondrial loci or the concatenated mitochon- drial and nuclear dataset (Fig. 3). In all analyses P. quercina and Other minor differences observed among the analyses were the the provisional species P. sp. ohioensis grouped together with some positions of P. nicotianae and P. cinnamomi. While there was unam- affinity to clade 1, but there was poor bootstrap support for a biguous support for the inclusion of P. nicotianae within clade 1, its placement within any of the established clades. exact affinity remains unclear. Previous studies were unable to Due to problems encountered when sequencing the nad9 locus resolve the relationship of P. nicotianae to the three well-supported for P. bisheria, P. alticola and P. frigida these species were not in- subclades, with Cooke et al. (2000) suggesting an affinity with cluded in the mitochondrial multilocus analysis; running a com- subclade 1b, Kroon et al. (2004) showing no resolution, and Blair bined analysis with cox2, rps10 and secY indicated P. bisheria et al. (2008) suggesting a basal position to subclades 1b and 1c. (P10117) was closely affiliated with P. frigida (P16053, P16059)

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 10 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

Fig. 5. Clade 2 phylogeny for Phytophthora using four mitochondrial loci (2373 nucleotides). Maximum likelihood branch lengths shown. Numbers on nodes represent bootstrap support values for maximum likelihood (top), maximum parsimony (middle) and Bayesian posterior probabilities as percentages (bottom). Nodes receiving significant support (>95%) in all analyses are marked with an asterisk (*). Scale bar indicates number of substitutions per site. and basal to P. multivesiculata (data not shown). This close affilia- Bezuidenhout et al., 2010; Hong et al., 2011) with support for P. tion with P. multivesiculata was observed in the ITS analysis in citricola being basal to P. pini, P. plurivora and P. citricola clade E, fol- the original species description of P. bisheria (Abad et al., 2008) lowed by P. capensis and P. multivora basal to P. citricola. The and the ITS and b-tubulin analysis of P. frigida (Maseko et al., mitochondrial gene data supports the separation of P. pini (Hong 2007). In the cox2 analysis P. alticola (P16052) was closely affiliated et al., 2011) and P. citricola clade E (Bezuidenhout et al., 2010) as with P. quercetorum/P. palmivora (data not shown), which is consis- distinct species. tent with the ITS and b-tubulin analysis reported by Maseko et al. Several additional clade 2 species that were formally described (2007). since the nuclear phylogeny of Blair et al. (2008) were added to the New clade 2 species that were once part of the P. citricola com- mitochondrial multilocus analysis, including the clade 2b species P. plex were included in our mitochondrial multilocus analysis, glovera (Abad et al., 2011), P. mengei (Hong et al., 2009), and P. sis- including P. multivora, P. capensis, P. plurivora, P. pini and a distinct kiyouensis (Reeser et al., 2007). In the mitochondrial multilocus phylogenetic lineage previously identified as P. citricola clade E analysis the placement of P. glovera between P. capsici and P. trop- (Bezuidenhout et al., 2010). The general relationship among these icalis was also observed in the ITS, ß-tubulin and translation elon- and other clade 2 species were similar to what has been previously gation factor 1a analysis of Abad et al. (2011). P. siskiyouensis and P. reported for nuclear loci (Scott et al., 2009; Jung and Burgess, 2009; mengei grouped together in subclade 2b in the mitochondrial

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 11

Fig. 6. Clade 8 phylogeny for Phytophthora using four mitochondrial loci (2373 nucleotides). Maximum likelihood branch lengths shown. Numbers on nodes represent bootstrap support values for maximum likelihood (top), maximum parsimony (middle) and Bayesian posterior probabilities as percentages (bottom). Nodes receiving significant support (>95%) in all analyses are marked with an asterisk (*). Scale bar indicates number of substitutions per site. analysis, a relationship that was also observed in the ITS analysis identified in the mitochondrial multilocus analysis that, based on reported by Hong et al. (2009) but in contrast to the placement clear delineation from other species, represent phylogenetically of P. siskiyouensis intermediate between P. capsici and P. tropicalis distinct species (Fig. 5; P. sp. aff. colocasiae -1, P. sp. colocasiae -2 in the ITS analysis of Reeser et al. (2007). Given the support of and P. sp. aff. siskiyouensis) and need careful morphological evalu- the multilocus analysis, the phylogenetic placement for P. siskyou- ation and additional nuclear loci sequenced to clarify their taxo- ensis observed in Fig. 3 is believed to be more accurate. The sepa- nomic status. ration of P1165 from the other P. mengei isolates suggests a more Historically it has been a challenge to morphologically delineate careful analysis of this isolate may be in order to confirm if it rep- species in the P. cryptogea/P. drechsleri species complex. While re- resents variation within the species or a separate lineage. This iso- cent work on these species has been providing phylogenetic sepa- late did not have a close grouping with another P. mengei isolate in ration based on DNA sequence data, the taxonomic classification of the three gene analysis (two nuclear and cox1) of Bezuidenhout some lineages remains unclear. In a multilocus analysis with four et al. (2010). Cultures of the provisional clade 2b species P. sp. bra- nuclear and one mitochondrial locus, Mostowfizadeh-Ghalamfarsa siliensis have been described as P. capsici in the past, but based on et al. (2010) clearly distinguished isolates of P. drechsleri and iden- this analysis and a prior, less comprehensive analysis (Bowers tified three distinct clades within P. cryptogea, classified as GI, GII et al., 2007), isolates recovered from Theobroma cacao in Brazil and GIII. While there was consistent support for a closer relation- clearly represent a phylogenetically distinct species. Efforts are ship between GI isolates and the homothallic P. erythroseptica, currently underway with a larger collection of isolates to formally there was little resolution among GII and GIII isolates; the presence describe this species (M.D. Coffey, F.N. Martin and M. A. Mansfield, of heterozygous bases within the nuclear loci suggests introgres- unpublished). Several additional putative new clade 2 species were sion has occurred among isolates from different P. cryptogea groups

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 12 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

(Mostowfizadeh-Ghalamfarsa et al., 2010). Using a parsimony- (Hansen et al., 2009). The phylogenetic placement of P. gallica based analysis of ancestral recombination, Olson et al. (2011) also (Jung and Nechwatal, 2008); P. austrocedri (Greslebin et al., 2007); reported evidence of recombination events among haplotypes of P. P. capensis, P. multivora, P. plurivora (Bezuidenhout et al., 2010; cryptogea; these authors suggest that P. drechsleri and P. cryptogea Jung and Burgess, 2009) and P. pini (Hong et al., 2011) agreed with represent incipient species still in the process of diverging, those previously reported. although they acknowledge that this interpretation may be unique As has been previously suggested (Cooke et al., 2000; Kroon to their samples from greenhouse floriculture hosts. Our combined et al., 2004, 2012; Martin et al., 2012) there was no consistent cor- analysis (Fig. 3) and expanded mitochondrial analysis of Clade 8 relation between phylogenetic grouping and morphological fea- (Fig. 6) also showed significant molecular divergence between P. tures. While some clades show primarily papillate (clade 4, 5, cryptogea and P. drechsleri. However, contrary to Mostowfizadeh- and 10), semipapillate (clade 3) or nonpapillate sporangia (clades Ghalamfarsa et al. (2010), our results suggest that the molecular 6, 7, and 9), other clades show combinations of these features divergence observed among the P. cryptogea lineages may justify (clade 1, 2 and 8). Although there was some grouping of this spo- their redescription as separate species. P. cryptogea isolate P1739 rangia feature in the modified multispecies coalescent analysis was isolate SCRP207 (GI) of Mostowfizadeh-Ghalamfarsa et al. (clades 4 and 5 for papillate; 6 and 7 for nonpapillate sporangia), (2010) and was clearly separate from the GII P. cryptogea isolates other distinct clades also contain species that exhibit this charac- as well as the P. sp. kelmania isolates in the mitochondrial (Figs. 1 teristic as well (clade 10 for papillate and clade 9 for nonpapillate). and 6) and combined mitochondrial and nuclear analysis (Fig. 3). Likewise, oogonial ornamentation, homothallic/heterothallic mat- These groupings were also evident in isozyme analysis (Mills ing, and type of antheridial attachment were polyphyletic in our et al., 1991); GI isolates (P1088, P1739) were in isozyme group B analyses. while GII were in isozyme group D (P3103) and E (1380, P2001), A total of 17 provisional species were included in this analysis. and P. sp. kelmania (P1810) and a GIII isolate (P1811) of Mostowfi- While P. sp. PgChlamydo, P. taxon raspberry, P. citricola clade E, P. zadeh-Ghalamfarsa et al. (2010) were in isozyme group C. P. sp. sp. kelmania, P. sp. niederhauserii and P. sp. brasiliensis have been kelmania is a provisional species initially noted in an abstract in previously discussed (Abad et al., 2002; Abad and Abad, 2003; 2002 (Abad et al., 2002) but not yet formally described; Mostowfi- Bowers et al., 2007; Brasier et al., 2003; Jung and Burgess, 2009), zadeh-Ghalamfarsa et al. (2010) suggested this provisional species only limited information is available for many of the others we may be conspecific with either GII or GIII. In our multilocus nuclear have examined here. P. sp. ohioensis was recovered from soils col- analysis, this isolate (P10613) showed a higher level of heterozy- lected around a declining mature white oak tree (Quercus alba) in gosity ( 6% of variable sites) as compared to P. drechsleri, P. ery- Ohio (Y. Balci, personal communication); it was associated with  throseptica, and other P. cryptogea isolates (0–2%; data not oak roots and not aboveground portions of the tree and is homo- shown). Additional analyses will be needed with a larger group thallic with papillate sporangia. P. sp. hungarica is a homothallic of isolates from diverse hosts in order to fully resolve the evolu- species recovered primarily from riparian ecosystems with apl- tionary history and taxonomic status of the distinct molecular lin- erotic oospores with a mean diameter of 41 lm and predominately eages within P. cryptogea. paragynous antheridia, whereas P. sp. personii does not reproduce The isolates labeled P. sp. aff. erythroseptica in Fig. 6 were all sexually, forms chlamydospores and internally proliferating spo- recovered from Zantedeschia sp. and were distinct in comparison rangia (Jung et al., 2011). P. sp. napoenis, P. sp. cuyabensis and P. of cox1 sequences with the P. erythroseptica isolates used in Most- sp. lagoariana are clade 9 species recovered from primary rainfor- owfizadeh-Ghalamfarsa et al. (2010) (AY659580, AY659585, est in the Cuyabeno Faunistic Reserve in Ecuador while P. sp. AT65957, AY659588, AY659589); there was a close affiliation but novaeguineae was recovered from soil under Auracaria in Papua they differed by 4 SNPs (data not shown). In the mitochondrial New Guinea (M. Coffey, unpublished data). multilocus and the combined nuclear and mitochondrial analysis With the resources currently available in public databases to ac- there is strong bootstrap support separating P. sp. aff. erythroseptica cess curated sequence data tied to vouchered specimens, it is an from P. erythroseptica, indicating further analysis is needed to clar- easier task today compared to the not so distant past to identify ify if these isolates represent a new species or a subpopulation of P. unique sequences that may be indicative of a new species. While erythroseptica. Likewise, additional analysis of the clade 8b phylo- it is important to have some sort of label attached to this entity genetic species P. sp. aff. brassicae -1 (P10728, recovered from Daucus so others can use it as a reference point, the best approach to deriv- carota)andP. sp. aff. brassicae -2 (P6207, recovered from Allium cepa) ing this label is an open question. Here we have chosen to use the is needed to determine their relationship to the recently described abbreviation affinis (related to another species) for this purpose clade 8b species P. cichorii, P. dauci, P. lactucae and provisional species although there are limitations to this terminology. For example, P.taxonparsleyandP.taxoncastitis(Bertier et al., 2013). the addition of other species to an analysis may change the phylo- The analysis presented herein provides clarification of phyloge- genetic relatedness of the taxa, thereby changing the affinis classi- netic relationships for some of the more recently described species. fication. While assigning a provisional species name may be The original description of P. irrigata placed it between P. richardiae desirable as well, this can lead to the long term use of these names and P. fallax/P. captiosa (Hong et al., 2008), but inclusion of addi- without a formal description being published; this may also lead to tional clade 9 species in the mitochondrial gene analysis places it confusion as it is not uncommon to see provisional names mistak- closer to the provisional species P. sp. napoensis and P. sp. cuyaben- enly used as formal names. Another option is to label the isolate sis (Fig. 1). Likewise, the expanded mitochondrial gene analysis according to the molecular clade or sub-clade with which it groups, placed P. parsiana close to the provisional species P. sp. thermophi- although this may not always provide the desired level of resolu- lum and P. sp. lagoariana rather than basal to P. insolita and P. pol- tion. Whatever naming convention is used by the research commu- onica separate from other clade 9 and 10 species (Mostowfizadeh- nity, it is crucial that this be a temporary approach, followed Ghalamfarsa et al., 2008); we have also clarified the placement of P. rapidly by a formal taxonomic description. hydropathica (Hong et al., 2010) within clade 9. Our analyses also Similarly, the expanded use of DNA sequence data for the iden- placed P. rosacearum between P. asparagi and the rest of the clade tification of new species has led to an increased use of the phylo- 6 species rather than just basal to P. megasperma and P. gonapody- genetic species concept in delineating species boundaries. This ides (Hansen et al., 2009); and placed P. sansomeana as the basal can sometimes lead to conflicts in taxonomic assignment, as taxa species in clade 8a rather than basal to P. drechsleri but separate traditionally ‘‘lumped’’ into a single species based on morphologi- from P. trifolii and P. medicaginis in ITS phylogenetic analysis cal and ecological traits are shown to encompass multiple distinct

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx 13 phylogenetic lineages. The appropriate ‘‘splitting’’ of these lineages Abad, Z.G., Abad, J.A., Coffey, M.D., Oudemans, P.V., Man in‘t Veld, W.A., De Gruyter, in order to best reflect their evolutionary history will require ro- H., Cunnington, J., Louws, F.J., 2008. Phytophthora bisheria sp. nov., a new species identified in isolates from the rosaceous raspberry, rose and strawberry in three bust molecular evidence, including evidence from more complex continents. Mycologia 100, 99–110. methods such as multispecies coalescent analyses and ancestral Abad, Z.G., Ivors, K.L., Gallup, C.A., Abad, J.A., Shew, H.D., 2011. Morphological and recombination approaches. These types of analyses will be partic- molecular characterization of Phytophthora glovera sp. nov. from tobacco in Brazil. Mycologia 103, 341–350. ularly important when examining possible species complexes such Baldauf, S.L., Palmer, J.D., 1993. Animals and fungi are each other’s closest relatives: as P. cryptogea (as described above). In our analyses presented here, congruent evidence from multiple proteins. Proc. Natl. Acad. Sci. USA 90, we have used seven nuclear and four mitochondrial loci to evaluate 11558–11562. Bertier, L., Brouwer, H., de Cock, A.W.A.M., Cooke, D.E.L., Olsson, C.H.B., Hofte, M., phylogenetic relationship among Phytophthora species. However, 2013. The expansion of Phytophthora clade 8b: three new species associated we have identified a subset of loci that provide a similar level of with winter grown vegetable crops. Persoonia 31, 63–76. resolution, and may be more practical for continued phylogenetic Bezuidenhout, C.M., Denman, S., Kirk, S.A., Botha, W.J., Mostert, L., McLeod, A., 2010. Phytophthora taxa associated with cultivated Agathosma, with emphasis on the analysis of the genus or for the description of new species. Based P. citricola complex and P. capensis sp. nov. Persoonia 25, 32–49. of the ease of amplification and the level of phylogenetic informa- Bhattacharya, D., Stickel, S.K., 1994. Sequence analysis of duplicated actin genes in tiveness, the LSU and b-tubulin regions are good candidates for the Lagenidium giganteum and Pythium irregulare (Oomycota). J. Mol. Evol. 39, 56–61. analysis of nuclear loci. While the elongation factor 1-alpha locus Blair, J.E., Coffey, M.D., Park, S.Y., Geiser, D.M., Kang, S.C., 2008. A multi-locus phylogeny for Phytophthora utilizing markers derived from complete genome has also been used by us and others (Kroon et al., 2004; Mostowfi- sequences. Fungal Genet. Biol. 45, 266–277. zadeh-Ghalamfarsa et al., 2010; Olson et al., 2011), analysis of the Blair, J.E., Coffey, M.D., Martin, F.N., 2012. Species tree estimation for the late blight available genomic data for Phytophthora species indicates that this pathogen, Phytophthora infestans, and close relatives. PLoS ONE 7, e37003. http://dx.doi.org/10.1371/journal/pone.0037003. gene is duplicated (J. Blair, unpublished data); divergence between Bowers, J.H., Martin, F.N., Tooley, P.W., Luz, E.D.M.N., 2007. Genetic and duplicates may bias the phylogenetic signal of this locus. While the morphological diversity of temperate and tropical isolates of Phytophthora ITS is useful for identification of isolates by barcoding (along with capsici. Phytopathology 97, 492–503. Brasier, C.M., 2007. Phytophthora biodiversity: how many Phytopthora species are the cox1 locus), its utility as a phylogenetic locus is limited to clo- there? In: Goheen, E.M., Frankel, S.J. (Eds.), USDA, Forest Service. Pacific sely related species within a clade or sub-clade; due to length var- Southwest Research Station, pp. 101–115. iation it is impossible to generate an unambiguous ITS alignment Brasier, C.M., Cooke, D.E.L., Duncan, J.M., 1999. Origin of a new Phytophthora pathogen through interspecific hybridization. Proc. Nat. Acad. Sci. USA 96, across the entire genus, leading to limited resolution and generally 5878–5883. low bootstrap support for relationships. For mitochondrial loci, Brasier, C.M., Cooke, D.E.L., Duncan, J.M., Hansen, E.M., 2003. Multiple new while the secY gene amplifies well and was the most divergent of phenotypic taxa from trees and riparian ecosystems in Phytophthora gonapodyides–P. megasperma ITS Clade 6, which tend to be high-temperature the four examined (average pairwise distance of 0.104), obtaining tolerant and either inbreeding or sterile. Mycol. Res. 107, 277–290. clean sequencing reads can be problematic for some species. The Cooke, D.E.L., Duncan, J.M., 1997. Phylogenetic analysis of Phytophthora species rps10 was the next most divergent mitochondrial locus (0.096) fol- based on ITS1 and ITS2 sequences of the ribosomal RNA gene repeat. Mycol. Res. lowed by nad9 (0.075) and cox2 (0.071); all three of these loci 101, 667–677. Cooke, D.E.L., Drenth, A., Duncan, J.M., Wagels, G., Brasier, C.M., 2000. A molecular amplified and sequenced well. A concerted effort by the Phytoph- phylogeny of Phytophthora and related oomycetes. Fungal Genet. Biol. 30, 17– thora community to sequence these loci, perhaps in addition to 32. the traditional ITS locus, will increase the availability of data for fu- Crawford, A.R., Bassam, B.J., Drenth, A., Maclean, D.J., Irwin, J.A.G., 1996. Evolutionary relationships among Phytophthora species deduced from rDNA ture evolutionary analyses as well as for the development of appro- sequence analysis. Mycol. Res. 100, 437–443. priate diagnostic and monitoring tools. Degnan, J.H., Rosenberg, N.A., 2009. Gene tree discordance, phylogenetic inference and multispecies coalescent. Trends Ecol. Evol. 24, 332–340. Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012. Bayesian phylogenetics Acknowledgments with BEAUti and the BEAST 1.7. Mol. Phylogen. Evol. 29, 1969–1973. Erwin, D.C., Ribeiro, O.K., 1996. Phytophthora Diesases Worldwide. APS Press, St. The senior author would like to thank Lorien Radmer for her Paul, MN. Felsenstein, J., 2005. PHYLIP version 3.6. . supported by USDA Agriculture and Food Research Initiative Plan Förster, H., Cummings, M.P., Coffey, M.D., 2000. Phylogenetic relationships of Biosecurity Competitive Grant Nos. 2007-55605-17835 and 2008- Phytophthora species based on ribosomal ITS I DNA sequence analysis with emphasis on Waterhouse groups V and VI. Mycol. Res. 104, 1055–1061. 55605-18773. JEB acknowledges support from grants to Franklin Greslebin, A.G., Hansen, E.M., Sutton, W., 2007. Phytophthora austrocedrae sp. nov., a & Marshall College from the Commonwealth of Pennsylvania, Ben new species associated with Austrocedrus chilensis mortality in Patagonia Franklin Technology Development Authority, and from the Howard (Argentina). Mycol. Res. 111, 308–316. Hansen, E.M., Wilcox, W.F., Reeser, P.W., Sutton, W., 2009. Phytophthora rosacearum Hughes Medical Institute’s Undergraduate Science Education Pro- and P. sansomeana, new species segregated from the Phytophthora megasperma gram. Mention of trade names or commercial products in this man- ‘‘complex’’. Mycologia 101, 129–135. uscript is solely for the purpose of providing specific information Hansen, E., Reeser, P., Sutton, W., 2012. Phytophthora beyond agriculture. Ann. Rev. Phytopathol. 50, 359–378. and does not imply recommendation or endorsement by the U.S. Hong, C., Gallegly, M.E., Richardson, P.A., Kong, P., Moorman, G.W., 2008. Department of Agriculture. USDA is an equal opportunity provider Phytophthora irrigata, a new species isolated from irrigation reservoirs and and employer. rivers in Eastern United States of America. FEMS Microbiol. Lett. 285, 203–211. Hong, C.X., Gallegly, M.E., Browne, G.T., Bhat, R.G., Richardson, P.A., Kong, P., 2009. The avocado subgroup of Phytophthora citricola constitutes a distinct species, Appendix A. Supplementary material Phytophthora mengei sp. nov. Mycologia 101, 833–840. Hong, C.X., Gallegly, M.E., Richardson, P.A., Kong, P., Moorman, G.W., Lea-Cox, J.D., Ross, D.S., 2010. Phytophthora hydropathica, a new pathogen identified from Supplementary data associated with this article can be found, in irrigation water, Rhododendron catawbiense and Kalmia latifolia. Plant. Pathol. the online version, at http://dx.doi.org/10.1016/j.fgb.2014.02.006. 59, 913–921. Hong, C., Gallegly, M.E., Richardson, P.A., Kong, P., 2011. Phytophthora pini Leonian resurrected to distinct species status. Mycologia 103, 351–360. References Ioos, R., Andrieux, A., Marcais, B., Frey, P., 2006. Genetic characterization of the natural hybrid species Phytophthora alni as inferred from nuclear and Abad, Z.G., Abad, J.A., 2003. Advances in the integration of morphological and mitochondrial DNA analysis. Fungal Genet. Biol. 43, 511–529. molecular characterization in the genus Phytophthora: The case of P. Jung, T., Burgess, T.I., 2009. Re-evaluation of Phytophthora citricola isolates from niederhauseria sp. nov. Phytopathology 93 (Suppl. 6), S1 (Abstract). multiple woody hosts in Europe and North America reveals a new species, Abad, Z.G., Abad, J.A., Creswell, T., 2002. Advances in the integration of Phytophthora plurivora sp. nov. Persoonia 22, 95–110. morphological and molecular characteristization in Phytophthora genus: The Jung, T., Nechwatal, J., 2008. Phytophthora gallica sp. nov., a new species from case of P. kelmania and other putative new species. Phytopathology 92 (6 rhizosphere soil of declining oak and reed stands in France and Germany. Suppl.), S1. Mycol. Res. 112, 1195–1205.

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006 14 F.N. Martin et al. / Fungal Genetics and Biology xxx (2014) xxx–xxx

Jung, T., Stukely, M.J.C., St. Hardy, G.E., White, D., Paap, T., Dunstan, W.A., Burgess, Mostowfizadeh-Ghalamfarsa, R., Cooke, D.E.L., Banihashemi, Z., 2008. Phytophthora T.I., 2011. Multiple new Phytophthora species from ITS clade 6 associated with parsiana sp. nov., a new high-temperature tolerant species. Mycol. Res. 112, natural ecosystems in Australia: evolutionary and ecological implications. 783–794. Persoonia 26, 13–39. Mostowfizadeh-Ghalamfarsa, R., Panabieres, F., Banihashemi, Z., Cooke, D.E.L., 2010. Knoll, H.A., 1992. The early evolution of eukaryotes: a geological perspective. Phylogenetic relationship of Pethybr. & Laff and P. Science 256, 622–627. drechsleri Tucker. Fungal Biol. 114, 325–339. Kroon, L.P.N.M., Bakker, F.T., Van Den Bosch, G.B.M., Bonants, P.J.M., Flier, W.G., Olson, H.A., Carbone, I., Benson, D.M., 2011. Phylogenetic history of Phytophthora 2004. Phylogenetic analysis of Phytophthora species based on mitochondrial and cryptogea and Phytophthora drechsleri isolates from floriculture crops in North nuclear DNA sequences. Fungal Genet. Biol. 41, 766–782. Carolina greenhouses. Phytopathology 101, 1373–1384. Kroon, L.P.N.M., Brouwer, H., de Cock, A.W.A.M., Govers, F., 2012. The genus Park, J., Park, B., Blair, J.E., Geiser, D.M., Isard, S., Mansfield, M.A., Nikolaeva, E., Park, Phytophthora anno 2012. Phytopathology 102, 348–364. S.Y., Russo, J., Kim, S.H., Greene, M., Ivors, K.L., Balci, Y., Peiman, M., Coffey, M.D., Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, Jung, K., Lee, Y.H., Rossman, A., Farr, D., Cline, E., Grhnwald, N.J., Luster, D.G., H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J., Schrandt, S., Martin, F., Makalowska, I., Kang, S., 2008. Phytophthora database: a Higgins, D.G., 2007. ClustalW and ClustalX version 2.0. Bioinformatics 23, 2947– cyberinfrastructure supporting the identification and monitoring of 2948. Phytophthora. Plant Dis. 92, 966–972. Man in‘t Veld, W., Rosendahl, K.C.H.M., Hong, C., 2012. Phytophthora x serendipita sp. Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of DNA substitution. nov. and P. x pelgrandis, two destructive pathogens generated by natural Bioinformatics 14, 817–818. hybridization. Mycologia 104, 1390–1396. Rambaut, A., Drummond, A. J., 2007. Tracer v1.4, . Martin, F.N., 2000. Phylogenetic relationships among some Pythium species inferred Reeser, P.W., Hansen, E.M., Sutton, W., 2007. Phytophthora siskiyouensis, a new from sequence analysis of the mitochondrially encoded cytochrome oxidase II species from soil, water, myrtlewood (Umbellularia californica) and tanoak gene. Mycologia 92, 711–727. (Lithocarpus densiflorus) in southwestern Oregon. Mycologia 99, 639–643. Martin, F.N., 2008. Mitochondrial haplotype determination in the oomycete plant Reeser, P., Sutton, C.A., Hansen, E., 2013. Phytophthotra pluvialis, a new species from pathogen Phytophthora ramorum. Curr. Genet. 54, 23–34. mixed tanoak-Douglas-fir forests of western Oregon, USA. North Am. Fungi. 8. Martin, F.N., Coffey, M.D., 2012. Mitochondrial haplotype analysis for differentiation http://dx.doi.org/10.2509/naf2013.008.007. of isolates of Phytophthora cinnamomi. Phytopathology 102, 229–239. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference Martin, F.N., Tooley, P.W., 2003a. Phylogenetic relationships among Phytophthora under mixed models. Bioinformatics 19, 1572–1574. species inferred from sequence analysis of mitochondrially encoded Scanu, B., Hunter, G.C., Linaldeddu, B.T., Franceschini, A., Maddau, L., Jung, T., cytochrome oxidase I and II genes. Mycologia 95, 269–284. Denman, S., 2013. A taxonomic re-evaluation reveals that Phytophthora Martin, F.N., Tooley, P.W., 2003b. Phylogenetic relationships of Phytophthora cinnamomi and P. cinnamomi var. parvispora are seaparate species. For. Path. ramorum, P. nemorosa and P. pseudosyringae, three species recovered from http://dx.doi.org/10.1111/efp.12064. areas in California with sudden oak death. Mycol. Res. 107, 1379–1391. Scott, P.M., Burgess, T.I., Barber, P.A., Shearer, B.L., Stukely, M.J.C., Hardy, G.E.S.J., Martin, F.N., Tooley, P.W., Blomquist, C., 2004. Molecular detection of Phytophthora Jung, T., 2009. Phytophthora multivora sp. nov., a new species recovered from ramorum, the causal agent of sudden oak death in California, and two additional declining Eucalyptus, Banksia, Agonis and other plant species in Western species commonly recovered from diseased plant material. Phytopathology 94, Australia. Persoonia 22, 1–13. 621–631. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary Martin, F.N., Abad, G., Balci, Y., Ivors, K., 2012. Identification and detection of genetic analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599. Phytophthora: reviewing our progress, identifying our needs. Plant Dis. 96, Wainright, P.O., Hinkle, G., Sogin, M.L., Stickel, S.K., 1993. Monophyletic origins of 1080–1103. the metazoa: an evolutionary link with fungi. Science 260, 340–342. Maseko, B., Burgess, T.I., Coutinho, T.A., Wingfield, M.J., 2007. Two new Weerakoon, N.D., Roberts, J.K., Lehnen Jr, L.P., Wilkinson, J.M., Marshall, J.S., Phytophthora species from South African Eucalyptus plantations. Mycol. Res. Hardham, A.R., 1998. Isolation and characterization of the single b-tubulin gene 111, 1321–1338. in Phytophthora cinnamomi. Mycologia 90, 85–95. Mills, S.D., Forster, H., Coffey, M.D., 1991. Taxonomic structure of Phytophthora Yang, X., Copes, W.E., Hong, C., 2013. Phytophthora mississippiae sp. nov., a new cryptogea and P. drechsleri based on isozyme and mitochondrial DNA analysis. species recovered from irrigation reservoirs at a plant nursery in Mississippi. J. Mycol. Res. 95, 31–48. Plant Pathol. Microbiol. 4. http://dx.doi.org/10.4172/2157-7471.1000180.

Please cite this article in press as: Martin, F.N., et al. A combined mitochondrial and nuclear multilocus phylogeny of the genus Phytophthora. Fungal Genet. Biol. (2014), http://dx.doi.org/10.1016/j.fgb.2014.02.006