Molecular systematics of the cotton root rot pathogen, Phymatotrichopsis omnivora.

Cotton root rot is an important soilborne disease of cotton and numerous dicot plants in the south-western United States and Mexico. The causal organism, Phymatotrichopsis omnivora (= Phymatotrichum omnivorum), is known only as an asexual, holoanamorphic (mitosporic) fungus, and produces conidia resembling those of Botrytis. Although the corticoid basidiomycetes Phanerochaete omnivora (Polyporales) and Sistotrema brinkmannii (Cantharellales; both Agaricomycetes) have been suggested as teleomorphs of Phymatotrichopsis omnivora, phylogenetic analyses of nuclear small- and large-subunit ribosomal DNA and subunit 2 of RNA polymerase II from multiple isolates indicate that it is neither a basidiomycete nor closely related to other species of Botrytis (Sclerotiniaceae, Leotiomycetes). Phymatotrichopsis omnivora is a member of the family Rhizinaceae, Pezizales (Ascomycota: Pezizomycetes) allied to Psilopezia and Rhizina.

INTRODUCTION

A devastating disease of cotton in Texas, which caused large numbers of plants in affected areas to suddenly wilt and die, was first reported in the 1880s (Pammel 1888, 1889). The disease has been variably called cotton root rot (after the major crop host), Texas root rot (for the centre of distribution), or Ozonium or Phymatotrichum root rot (for the former names of the causal organism). It has since remained a considerable economic concern, causing up to $ 100 million in annual losses to the US cotton crop alone (based on disease loss estimates and price data for 1980–2008; provided by the National Cotton Council of America, www.cotton.org). The average loss of raw cotton fibre yield has been estimated to be 3.5 % in Texas and 2.2 % in Arizona, with losses ranging from 8–13 % in severely infested areas (Kenerley & Jeger 1992). The causal agent is a soilborne fungus known as Phymatotrichopsis omnivora or, more commonly, Phymatotrichum omnivorum (Streets & Bloss 1973, Kenerley & Jeger 1992, Kirkpatrick & Rothrock 2001; see below for taxonomic authorities). This species is capable of infecting more than 2 000 species of dicots (Streets & Bloss 1973), arguably the largest host range of any plant pathogen. It also causes severe losses in alfalfa, vegetable crops, grapes, and fruit and nut orchards throughout its range, which stretches from eastern Texas and southern Oklahoma west through Arizona and south into Mexico (Streets & Bloss 1973). Generally, infected plants quickly wilt in the summer, and almost inevitably die, usually in large circular patches in the field (Fig. 1a, b). Below ground, the taproots of wilted plants are rotted and usually covered with mycelial strands of the causal fungus (Fig. 1c).

Fig. 1

Phymatotrichum root rot and morphological characteristics of the causal fungus, Phymatotrichopsis omnivora. a. Disease foci in an alfalfa field (near Devol, OK); b. disease foci in a cotton field (near Austwell, TX); c. mycelial strands (arrows) on infected cotton root; d–f. mycelial strand showing acicular hyphae, cruciform hypha (arrow, inset e) and rectangular and polygonal cells (inset f); g. sporemat on soil surface; h–m. conidiophores and conidia borne on sporemat of Phymatotrichopsis omnivore; j. immature conidiophores produced from mycelial strand hyphae; k. botryoblastoconidia forming on conidiophores; l, m. ‘basidium-like’ conidiophores (arrows). — Scale bars: d = 100 μm; e = 50 μm; f, h = 25 μm; g = 5 mm; i = 20 μm; j–m = 10 μm.

Taxonomy

The confused taxonomic history of the cotton root rot fungus goes back more than a century. The causal agent was first identified by W.G. Farlow as Ozonium auricomum Link, based on nonsporulating mycelium associated with diseased roots (Pammel 1888). However, this name now applies to the asexual state of Coprinellus (Coprinus) domesticus and related species (Shear 1907, Orton & Watling 1979, Redhead et al. 2001). The cotton root rot fungus was described as a new species of Ozonium, O. omnivorum Shear (1907), again based on nonsporulating mycelium associated with diseased roots. Later, a conidial stage was found forming sporemats on soil surrounding diseased plants and was named Phymatotrichum omnivorum (Shear) Duggar (1916).

A hydnoid homobasidiomycete fruiting body was found associated with diseased plants and named Hydnum omnivorum Shear (1925), once again based on a different type specimen (C.L. Shear 5267, BPI 259732) from that of Ozonium omnivorum or Phymatotrichum omnivorum. Later, a corticioid homobasidiomycete fruiting body was discovered in a culture of Phymatotrichum omnivorum and identified as Sistotrema brinkmannii (Baniecki & Bloss 1969). Basidiospores of the Sistotrema failed to form the mycelium of Phymatotrichum, and Weresub & LeClair (1971) considered this report to be based on a homothallic culture contaminant.

The type species of Phymatotrichum, P. gemellum Bonord., was shown to be a member of Botrytis by Hennebert (1973). Hennebert (1973) believed that the name Phymatotrichum omnivorum should be attributed to Duggar alone since it was based on different specimens than examined by Shear (1907) when he described Ozonium omnivorum, and because the distinguishing features described by Duggar (the conidia) were not present in the type of Ozonium omnivorum (C.L. Shear 1447, BPI 455660). Phymatotrichum omnivorum was transferred to Phymatotrichopsis omnivora (Duggar) Hennebert and Phymatotrichum fimicola Dring to Pulchromyces fimicola (Dring) Hennebert.

The type specimen and cultures of Hydnum omnivorum were studied by Burdsall and Nakasone (1978) who transferred this species to Phanerochaete and distinguished it from Phymatotrichopsis omnivora and from Phanerochaete chrysorhiza on the basis of culture morphology. Phanerochaete omnivora has been found on dead stems and roots of angiosperm trees and shrubs in Arizona and Texas but has not been reported from cotton or most of the other hosts of Phymatotrichopsis omnivora (Burdsall & Nakasone 1978, Burdsall 1985). As of today, the name of this economically important plant pathogen is Phymatotrichopsis omnivora and, as far as is known, it is a holoanamorphic (solely asexual) fungus of unknown phylum (e.g., Ascomycota, Basidiomycota or Zygomycota).

More recent work has provided some clues to the phylogenetic identity of Phymatotrichopsis omnivora. It is sensitive to the fungicide benomyl at rates of 5 mg/L (Hine et al. 1969, Lyda & Burnett 1970), a concentration to which most members of the Basidiomycota are tolerant, whereas members of the Ascomycota, excepting Pleosporales, are sensitive (Edgington et al. 1971). Gunasekaran et al. (1974) examined the hyphal walls of P. omnivora using transmission electron microscopy (TEM). Unfortunately, they did not study septa, which could have conclusively indicated whether P. omnivora is an ascomycete (simple septal pore with Woronin bodies) or basidiomycete (simple or dolipore septa lacking Woronin bodies) (Bracker 1967, Bartnicki-Garcia 1987). However, the hyphal walls of P. omnivora clearly possessed the bilayered structure typical of Ascomycota, with a thick, translucent inner layer and a thin, electron-dense outer layer (Gunasekaran et al. 1974). In contrast, hyphal walls of most Basidiomycota show multiple thin translucent and electron-dense layers (Bartnicki-Garcia 1987). Woronin bodies, diagnostic of filamentous Ascomycota, were discovered by Dong et al. (1981) in the hyphae of Phymatotrichopsis omnivora. Despite this strong evidence to indicate that P. omnivora is actually a member of the Ascomycota, the Dictionary of the Fungi (Kirk et al. 2001) lists Phymatotrichopsis as “? anamorphic Basidiomycota”. A preliminary phylogenetic analysis of the relationships among P. omnivora and other botryoblastosporic fungi using the nuclear ribosomal internal transcribed spacer (ITS) region was inconclusive (Riggs 1993). The purpose of the current study is to provide a more conclusive and precise systematic placement of the cotton root rot pathogen, Phymatotrichopsis omnivora, based on phylogenetic analyses of DNA sequence data from nuclear ribosomal DNA and protein-coding genes.

MATERIALS AND METHODS

Cultures

Phymatotrichopsis omnivora, Pulchromyces fimicola and Sistotrema brinkmannii were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and cultures of Phanerochaete omnivora and Phanerochaete chrysosporium from USDA-FPL (Madison, WI). Additional isolates of P. omnivora were obtained from Dr Mary Olsen, University of Arizona, Tucson (Table 1) or isolated from the roots of diseased cotton and alfalfa plants as previously described (Lyda & Kenerley 1992) and maintained on modified ATCC medium 1078 (M1078), containing per 1 000 mL distilled water: 1 g NH4NO3; 0.75 g MgSO4; 0.4 g KH2PO4; 0.9 g K2HPO4; 0.1 g CaCl2; 40 g glucose; 1 g yeast extract; 1 g peptone; 100 μL Vogel’s trace elements (Vogel 1964) and 18 g agar. Cultures collected for this study will be deposited at ATCC.

Table 1

Species used in molecular phylogenetic analyses, specimen information and GenBank accession numbers. New sequences generated for this study are indicated with GenBank numbers in bold.

Species	Vouchers, Isolates, Strains (Herbarium1)2	GenBank Accession Numbers
		SSU	ITS	LSU	RPB2	β-tub3
Aleuria aurantia	OSC 100018	AY544698	–	AY544654	DQ247785	–
Anthracobia sp.	OSC 100026	AY544704	–	AY544660	–	–
Ascobolus carbonarius	KH 00.008 (C) (dubl. OSC 100079)	AY544720	–	AY500526	–	–
Ascobolus crenulatus	KH.02.005 (C) (dubl. OSC 100082)	AY544721	–	AY500527	–	–
Ascodesmis nigricans	CBS 389.68	–	–	DQ168335	–	–
Ascodesmis sphaerospora	RK 95.55 (O)	U53372	–	–	–	–
Balsamia magnata	JMT 13020 (OSC)	U42656	–	U42683	–	–
Barssia oregonensis	RF 533 (OSC)	U42657	–	U42684	–	–
Boudiera acanthospora	ARON 2167 (O)	U53373	–	–	–	–
Boudiera tracheia	Rana 79.049 (C)	–	–	AY500530	–	–
Byssonectria terrestris	SSU: UME 29218, LSU: KS-94-4 (C)	Z30241	–	AY500531	AY500504	–
Caloscypha fulgens	DJ053103-2	DQ247807	–	DQ247799	DQ247787	–
Cazia flexiascus	JMT 12993 (OSC)	U42666	–	U42694	–	–
Cheilymenia stercorea	KH04282003-4 (dubl. OSC 100034)	AY544705	–	AY544661	DQ471123	–
Chorioactis geaster	SSU: mh 694 (FH), LSU: H.W. Keller & K.C. Rudy s.n. (FH)	AF104340	–	AY307944	–	–
Choiromyces venosus	JMT 7014 (OSC)	U42661	–	U42688	–	–
Cookeina tricholoma	SSU: mh 686 (FH), LSU: 1D-D5 (FH)	AF006311	–	AY945860	–	–
Desmazierella acicola	SSU: ‘Norway’ (FH), LSU: RK 95.12 (Herb. Roy Kristiansen)	AF104341	–	AY945854	–	–
Dingleya verrucosa	JMT 12617 (OSC)	U42659	–	U42686	–	–
Discina macrospora	NSW 4498 (MICH)	U42651	–	U42678	–	–
Disciotis venosa	OSC 100045 (dubl. NRRL 22213)	U42643/AY544711	–	U42670/AY544667	DQ470892	–
Donadinia sp.	mh 669 (FH)	AF104342	–	DQ220329	–	–
Eleutherascus lectardii	CBS 626.71	DQ062997	–	DQ168334	DQ470918	–
Fischerula subcaulis	JMT 1889 (OSC)	U42646	–	U42673	–	–
Galiella rufa	mh 101 (FH)	AF004948	–	AY945850	–	–
Genea harknessii	Trappe 11775 (FH, dubl. OSC)	DQ646526	–,DQ220335	–	–
Geopora cf. cervina	KH.03.61 (FH)	DQ646527	–	DQ220344	–	–
Geopora cooperi f. gilkeyae	Trappe 18034 (FH, dubl. OSC)	DQ646528	–	DQ220342	–	–
Geopyxis carbonaria	SSU: _ (FH), LSU: C F-49793 (C)	AF104665	–	DQ168336	–	–
Glaziella aurantiaca	PR-5954 (FH)	DQ062996	–	DQ220351	–	–
Gyromitra californica	OSC 100068	AY544717	–	AY544673	DQ470891	–
Gyromitra esculenta	NRRL 20925 (dubl. CBS 335.73)	U42648	–	U42675	AY641045	–
Gyromitra melaleucoides	NSW 7196 (OSC)	U42653	–	U42680	–	–
Helvella cf. compressa	OSC 100019 (OSC)	AY544699	–	AY544655	DQ497613	–
Humaria hemisphaerica	KH.03.100 (FH)	DQ646529	–	DQ220353	–	–
Hydnotrya cerebriformis	NSW 6494 (OSC)	U42649	–	U42676	–	–
Iodophanus carneus	SSU: ARON 2102, LSU+RPB2: JHP 00.027 (C)	U53380	–	AY500534	AY500506	–
Iodowynnea auriformis	18510 PAN (FH)	DQ646530	–	AF335118	–	–
Labyrinthomyces varius	JMT 14825 (OSC)	U42662	–	U42689	–	–
Lamprospora ascoboloides	KH.03.54 (FH)	DQ646531	–	DQ220358	–	–
Lasiobolidium orbiculoides	CBS 344.73	DQ063000	–	DQ062995	–	–
Lasiobolidium spirale	CBS 782.70	DQ646533	–	DQ220363	–	–
Lasiobolus ciliatus	KS-94-005 (C)	DQ646532	–	DQ167411	–	–
Leucangium carthusianum	JMT 7205 (OSC)	U42647	–	U42674	–	–
Marcelleina persoonii	KH.00.07 (C)	DQ646534	–	AY500536	–	–
Marcelleina tuberculispora	All-94-8 (C)	DQ646535	–	AF335120	–	–
Melastiza contorta	KH.01.06 (C)	DQ646536	–	AY500539	–	–
Melastiza cornubiensis	KH.03.43 (FH)	DQ646537	–	DQ646524	–	–
Miladina lecithina	KH.03.156 (FH)	DQ646538	–	DQ220371	–	–
Morchella elata	SSU+LSU+RPB2: NRRL 25405, SSU+LSU: NRRL 22447 (dubl. OSC 100042)	U42641/AY544709	–	U42667/AY544665	AF107810	–
Morchella esculenta	SSU: NRRL 22335, SSU+LSU: MV3 (dubl. OSC 100041), LSU+RPB2: ATCC 10968	U42642/AY544708	–	AY544664/AF279398	AY641054	–
Nanoscypha tetraspora	mh PR61 (FH)	AF006314	–	DQ220374	–	–
Neolecta vitellina	SSU: UME 29192 (U), LSU: JP 176 (F)	Z27393	–	AF279401	–	–
Neottiella rutilans	SSU: ARON 2690 (U), LSU: KH.03.55 (FH)	AF061720	–	DQ220377	–	–
Neournula pouchetii	NSW 6435 (OSC)	AF104666	–	AY307940	–	–
Octospora hygrohypnophila	KH.03.30 (FH)	DQ646539	–	DQ220379	–	–
Orbicula parietina	C F-24441 (C)	DQ062998	–	DQ062988	–	–
Orbilia auricolor	CBS 547.63 (dubl. OSC)	DQ471001	–	DQ470953	DQ470903	–
Otidea onotica	SSU: mh 685 (FH), LSU: KH-98-107 (C)	AF006308	–	AF335121	–	–
Pachyella clypeata	FH No. 387 (FH)	DQ646540	–	AY500542	–	–
Pachyphloeus melanoxanthus	SSU: 1255 (UP), LSU: Gardner & Healy 195 (FH)	AF054899	–	DQ191674	–	–
Parascutellinia carneosanguinea	KH.03.34 (FH)	DQ646541	–	DQ220388	–	–
Paurocotylis pila	SSU: UME 30230, LSU: Trappe 12583 (OSC)	U53382	–	DQ168337	–	–
Peziza arvernensis	SSU+LSU: ALTA 9353, RPB2: KH-98-12 (C)	AF133175	AF133175	AF133162	AY500497	–
Peziza badiofusca	KH-98-113 (C)	DQ646542	–	AF335132	–	–
Peziza echinispora	SSU: DHP #136 (C), LSU+RPB2: Jukka Vauras 9110F (TURA)	AF006309	–	AF335138	AY500496	–
Peziza gerardii	KH-97-90 (C)	DQ646543	–	AF335143	–	–
Peziza lobulata	KH.03.157 (FH)	DQ646544	–	AY500548	–	–
Peziza michelii	TL-5692 (C)	DQ646545	–	AY500549	–	–
Peziza polaripapulata	KH-96-11 (C)	DQ646546	–	AY500551	–	–
Peziza quelepidotia	NRRL 22205	U42665	–	U42693	–	–
Peziza subisabellina	SSU: ALTA 9029, LSU: Winterhoff 8844 (herb. Winterhoff)	AF133144	–	AF335164	–	–
Peziza succosa	SSU: UME 29567 (U), LSU: KH-98-07 (C)	U53383	–	AF335166	–	–
Peziza vesiculosa	SSU: OSC 100074 (OSC), SSU+LSU: OSC 126 (OSC), LSU+RPB2: JV 95-652 (C)	AFTOL-202/DQ470995	–	AY500552/DQ470948	AY500489	–
Phillipsia domingensis	SSU: mh 688 (FH)	AF006315	–	–	–	–
Phillipsia crispata	LSU+RPB2: T. Læssoe AAU-44895a (AAU, C)	–	–	AY945845	DQ017599	–
Phymatotrichopsis omnivora	ATCC 22316	EF441991	EF441991 / EF494042	EF441991	–	–
	ATCC 28960	EF441992	EF441993	–	–	–
	ATCC 32445	EF494052	EF494043	EF494060	EF494070	–
	ATCC 32446	EF441994	EF441994	EF441994	–	–
	ATCC 32448	EF494048	EF494038	EF494056	–	–
	ATCC 48084	EF441995	EF441995	–	–	–
	M Olsen #1	–	EF441996	–	–	–
	M Olsen #2	EF441997	EF441997	EF441997	–	–
	M Olsen #3	–	EF441998	–	–	–
	M Olsen #4	EF441999	EF441999	–	–	–
	M Olsen #5	EF442000	EF442000	–	–	–
	PC04	EF494049	EF494039	EF494057	–	EF494064
	PP04	EF494047	EF494037	EF494055	–	EF494063
	TAMDC04	EF494050	EF494040	EF494058	–	EF494065
	NFAlf	EF494051	EF494041	EF494059	EF494069	EF494066
	OKAlf8	EF494045	AY549456	EF494053	EF494067	EF494061
	TXCO3-9	EF494046	AY549455	EF494054	EF494068	EF494062
	BMD Type sporemat (GLH 2868) (FH)	–	FJ013259	–	–	–
Pseudombrophila guldeniae	Kongsv. 85.10B (C)	DQ063001	–	DQ062993	–	–
Pseudombrophila theioleuca	C F-70057 (C)	DQ062999	–	DQ062989	–	–
Pseudopithyella minuscula	SSU: mh 673 (FH), LSU: mh 675 (FH)	AF006317	–	AY945849	–	–
Pseudoplectania nigrella	SSU: ‘Japan’, LSU: KH-97-28 (FH)	AF104345	–	AY945852	–	–
Psilopezia cf. nummularialis	TL-11785 (QCNE, dubl. C)	EU722510	–	EU722509	–	–
Psilopezia deligata	KH-99-13 (FH)	DQ646547	EF494044	DQ220390	EF494071	–
Psilopezia juruensis	T. Læssøe AAU 44912 (QCA, dubl. C)	DQ646548	–	DQ220391	–	–
Pulchromyces fimicola	ATCC 18595 (dubl. CBS 127.69, CUP 49531)	EF442001	EF442001	EF442001	–	–
	ATCC 36770 (dubl. IFAS-F 316)	EF442002	EF442002	EF442002	–	–
Pulvinula archeri	SSU: DAOM 195928, LSU: BAP 458 (FH)	U62012	–	DQ220392	–	–
Pyronema confluens	TL-11685 (QCNE, dubl. C)	DQ646549	–	DQ220397	–	–
Pyronema domesticum	SSU: ARON 1766, LSU+RPB2: CBS 666.88 (dubl. OSC 100503)	U53385	–	DQ247805	DQ247795	–
Reddellomyces donkii	JMT 13292 (OSC)	U42660	–	U42687	–	–
Rhizina undulata	SSU: NRRL 22168, LSU: KH.02.44 (FH)	U42664	–	DQ220410	–	–
Rhodotarzetta rosea	KH.03.107 (FH)	DQ646550	–	DQ220413	–	–
Sarcoscypha austriaca	SSU: mh 667 (FH), LSU: mh 670 (FH)	AF006318	–	AY945856	–	–
Sarcoscypha coccinea	spat 03-02 (dubl. OSC 100003)	AY544691	–	AY544647	AY544755	–
Sarcosphaera coronaria	SSU+LSU: OSC 100049, SSU: ALTA 9605, LSU: KS-94-24A (C), RPB2: KS-94-19 (C)	AY544712/AF133157	–	AY544668/AY500555	AY500523	–
Scabropezia scabrosa	Pfister 13.8.83 (FH)	AF133158	–	AF133173	–	–
Scutellinia scutellata	SSU: ARON 2188, SSU+RPB2: KH03212003-1 (dubl. OSC 100015), LSU: KS-94-035H (C)	U53387/DQ247814	–	DQ220421	DQ247796	–
Sowerbyella imperialis	CL2004-105 (C)	DQ646551	–	DQ220427	–	–
Sphaerosporella brunnea	LSU: KH.03.04 (FH) SSU: UME 31147	U53388	–	DQ220433	–	–
Strobiloscypha keliae	SSU: NSW 7333 (OSC), LSU: NSW 6387 (OSC)	AF006310	–	DQ220437	–	–
Tarzetta catinus	SSU: UME 29731, LSU: KS.94.10A (C)	U53389	–	DQ062984	–	–
Terfezia arenaria	SSU: 1217-1 (UP)	AF054898	–	–	–	–
Terfezia claveryi	LSU: Trappe 3195 (FH, dubl. OSC)	–	–	AY500558	–	–
Tricharina praecox	KH.03.101 (FH)	DQ646552	–	DQ646525	–	–
Trichophaea hybrida	SSU: UME 29738, LSU: KH.04.39 (FH, dubl. DBG)	U53390	–	DQ220454	–	–
Trichophaea woolhopeia	KH.01.33 (C)	DQ646553	–	DQ220460	–	–
Trichophaeopsis bicuspis	SSU: ARON 2222 (O), LSU: NSW 8316 (OSC)	U53391	–	DQ220461	–	–
Tuber gibbosum	NSW 7049 (OSC)	U42663	–	U42690	–	–
Underwoodia columnaris	Kanouse 1951 (MICH)	U42658	–	U42685	–	–
Urnula craterium	SSU: mh 671 (FH, dubl. DEB #278082), LSU+RPB2: DHP 04-511 (FH)	AF104347	–	AY945851	DQ017595	–
Verpa bohemica	NRRL 20858 (dubl. CBS 551.72)	U42645	–	U42672	–	–
Verpa conica	NRRL 20856 (dubl. CBS 407.81)	U42644	–	U42671	–	–
Wilcoxina mikolae	SSU: ATCC 52684, LSU: WS 36 (SFSU)	U62014	–	DQ220468	–	–
Wolfina aurantiopsis	SSU: –, LSU: DHP 04-599 (FH)	AF104664	–	AY945859	–	–
Wynnella silvicola	NSW 6219 (OSC)	U42655	–	U42682	–	–

1For herbaria abbreviations see Index Herbariorum (http://sciweb.nybg.org/science2/IndexHerbariorum.asp).

2When different isolates were used as sources for different genes, the respective gene is indicated prior to the isolate designation, i.e. ‘Gene: Isolate’.

3When different sequences were used for rDNA or rDNA+RPB2 trees, two sets of sequences for the same species will be listed.

Sporemats were recovered from pots of Phymatotrichopsis omnivora-inoculated plum trees grown in Houston black clay and were identified based on morphology and ITS-rDNA sequences amplified using Phymatotrichopsis omnivora-specific primers (PoITSA 5’-CCTGCGGAAGGATCATTAAA-3’ and PoITSB 5’-GGGGGTTTTCTTTGTTAGGG-3’; developed in this study). Hand-sectioned sporemats were mounted in lactoglycerol and examined using a Nikon Eclipse E800 microscope with PlanFluor objectives and a CCD camera (Qimaging, Burnaby, Canada). Digital micrographs were contrast-adjusted, cropped and scale bars inserted in Photoshop (Adobe Systems Inc., San Jose, USA).

Specimens of P. omnivora at the Farlow Herbarium (Harvard University, Cambridge, MA) studied and described by Duggar (1916) were examined microscopically and small fragments excised for DNA isolations. Specimens examined were labelled as follows:

“Phymatotrichum omnivorum (Shear) on soil in cotton field, Paris, Texas, Sept. 18, 1915, BMD, Received from Missouri Bot. Garden June 1916 (sporemat on soil peds mounted in slide box; insert: Ostracoderma omnivorum, comb. nov. ined., TYPE SPECIMEN for the conidial state, Examinavit G.L. Hennebert 2868, Nov. 1961)”;

“Phymatotrichum omnivorum (Shear) on Cultv. Cotton, Petty, Texas, Sept. 12, 1902, BMD, “Ozonium” stage, Recv. from Missouri Bot. Garden, June 1916 (insert 1: Shear Bull Torr. Bot Club 34: 305 1907, on root of cotton; insert 2: Ozonium state of Ostracoderma omnivorum, comb. nov. ined., Examinavit G.L. Hennebert 2869, Nov. 1961)”; and

“Phymatotrichum omnivorum (Shear) Paris, Texas, Sept. 18, 1915, BMD, “Ozonium” stage on Cotton, Recd from Missouri Bot. Garden, June, 1916, See also Box (insert: Ozonium state of Ostracoderma omnivorum, comb. nov. inedit., Examinavit G.L. Hennebert 2870, Nov. 1961)”.

Herbarium specimens will be referred to by the examination numbers given by G.L. Hennebert (e.g. GLH #2868, GLH #2869, and GLH #2870).

Molecular methods

Genomic DNA was isolated following Zolan & Pukkila (1986). Some DNA preparations required further cleaning using glass milk (Gene Clean II, Bio101, La Jolla, California) or electrophoresis in 0.7 % agarose gels in Tris acetate EDTA (TAE) buffer followed by electroelution (GeBA flex-tube micro-dialysis kit, Gene Bio-Application Ltd, Kfar-Hanagid, Israel). Genomic DNA was also isolated from homogenized mycelia using a glass filter-based kit (UltraClean Microbial DNA, MoBio Laboratories, Inc., Carlsbad, CA). DNA was isolated from Farlow Herbarium specimens using a E.Z.N.A. Forensic DNA Extraction Kit (Omega Biotek, Doraville, GA) with the manufacturer’s dried blood protocol with the following modifications: intact dried herbarium tissue (3–30 mm3 piece) was incubated in 200 μL Buffer STL and 25 μL OB protease solution 45 min using a Thermomixer (Eppendorf, Westbury, NY), frozen over liquid nitrogen and thawed at 60 °C, twice, and incubated at 60 °C shaking at 500 rpm for 20 h. An additional 100 μL Buffer STL and 10 μL OB protease solution were added to each extraction tube, freeze-thawed as before and incubated at 60 °C shaking at 500 rpm for 20 h more. Softened herbarium tissue was then crushed with a sterile pestle in the lysis buffer and DNA isolated according to manufacturer’s instructions with solution volumes adjusted for the additional 110 μL lysis buffer (STL + OB protease).

Nuclear rDNA (SSU, ITS and 5’ LSU regions) was PCR amplified using the following primer pairs SSJ and NS8, NS1 and NS8 (for SSU), ITS4 and ITS5 (for ITS), PoITSA and ITS2 (for herbarium material), LROR and LR7 (for LSU) or SSG and LR5 (for SSU to LSU) (Vilgalys & Hester 1990, White et al. 1990, Hausner et al. 1993). Two successive PCR reactions were used to amplify the ITS region from the P. omnivora herbarium specimen. For the first PCR, 50 μL reactions were denatured at 95 °C for 3 min, followed by 41 cycles of 94 °C for 30 s, 50 °C for 45 s and 72 °C for 45 s and a final extension of 72 °C for 7 min. After observing a faint band by gel electrophoresis, 1 μL from each of the first PCRs were used as templates for a second 50 μL PCR with an initial denaturation of 95 °C for 3 min, 20 cycles of 94 °C for 30 s, 50 °C for 45 s, and 72 °C for 45 s, and a final extension of 72 °C for 7 min. Using the thermocycler program and reverse primers of Liu et al. (1999), sequences spanning conserved regions 3–11 in RPB2 from P. omnivora isolates were amplified in two overlapping segments using the primer pairs RPB2-Ds3F (5’-WSYGARAAGGTHYTBATYGCRCAAGAGCG-3’) and fRPB2-7cR, and RPB2-Ds6F (5’-TGGGGWYTSGTHTGYCCWGC-3’) and fRPB2-11aR. A region of the β-tubulin gene spanning three introns was amplified and sequenced with primers Bt2a and Btspect (Glass & Donaldson 1995, Paolocci et al. 2004). Sequences were obtained in an automated sequencer (ABI 377) using dye-terminator technology and the following primers: SSJ, NS1, NS2, NS3, NS4, NS5, SSG, NS8, ITS1, ITS4, ITS5, LS1R, LS1, LR3R, LR7, LR16, NL1, NL4 and LR3 for rDNA (Vilgalys & Hester 1990, White et al. 1990, Hausner et al. 1993); and RPB2-Ds3F, fRPB2-5F, fRPB2-5R, RPB2-Ds6F, fRPB2-7cF, fRPB2-7cR, RPB2-980F, RPB2-1014R, RPB2-1554R, RPB2-1599F, RPB2-2488F, RPB2-2568R and fRPB2-11aR for RPB2 (Liu et al. 1999, Reeb et al. 2004). Complementary strand sequences were aligned and corrected in SeqEd (ABI Software) or ChromasPro (Technelysium Pty Ltd) and combined with most similar sequences from GenBank determined using BLASTn (Altschul et al. 1990, McGinnis & Madden 2004). All newly derived sequences have been deposited in GenBank as accession numbers EF441991–EF442000, EF494037–EF494070 and FJ013259 (Table 1).

Phylogenetic analyses

Large subunit and SSU rDNA sequences from Phymatotrichopsis omnivora, Pulchromyces fimicola and an additional species of Psilopezia, Ps. cf. nummularialis, were added to a data matrix containing 99 species of Pezizales (Hansen & Pfister 2006) by hand using the software Se-Al v. 2.0a11 (Rambaut 2002). The sequences represent all known sublineages within Pezizales, 82 genera and 14 families (out of c. 164 genera and 16 families; Table 1). Neolecta vitellina was used as outgroup. To substantiate the placement of Phymatotrichopsis omnivora and Pulchromyces fimicola within Pezizales, a data matrix including an additional gene, RPB2, was compiled representing a subset of the taxa from the combined LSU and SSU dataset. Amino acid sequences of RPB2 were deduced using a combination of BLASTx (Altschul et al. 1997) and the ExPASy translate tool (http://us.expasy.org/tools/dna.html). Multiple sequence alignments were generated using ClustalX (Thompson et al. 1997) or Muscle (Edgar 2004). The final alignments are available from TreeBASE (S2105).

Individual and combined analyses of the data matrices were performed using PAUP v. 4.0b10 (Swofford 2002) and MrBayes v. 3.1.1 (Huelsenbeck & Ronquist 2001, Ronquist & Huelsenbeck 2003) on Macintosh computers. Maximum parsimony (MP) analyses with heuristic searches consisted of 1 000 or 5 000 (for the subset LSU-SSU-RPB2 datasets) random sequence addition replicates with tree bisection-reconnection (TBR) branch swapping, MULPARS in effect and saving all equally most parsimonious trees (MPTs). All characters were equally weighted and unordered. In MP analyses of the individual, larger SSU rDNA data matrix a two-step search was performed (due to an exceedingly large number of trees generated), as follows: First, 1 000 heuristic searches were performed with random sequence addition and TBR branch swapping, with MAXTREES unrestricted, and keeping only up to 15 trees per replicate. Second, exhaustive swapping was performed on all the MPTs discovered with MAXTREES set to 15 000. Robustness of individual branches was estimated by parsimony bootstrap proportions (BP), using 500 (LSU-SSU dataset) or 1000 (LSU-SSU-RPB2 dataset) bootstrap replicates, each consisting of a heuristic search with 100 random addition sequence replicates, TBR branch swapping, and MAXTREES set at 100 (LSU-SSU) or unrestricted (LSU-SSU-RPB2).

The GTR+I+G model of nucleotide substitution was found to fit each of the rDNA datasets best using a hierarchical likelihood ratio test as implemented in the program MrModeltest v. 2.2 (Nylander 2004). In Bayesian analyses of the LSU-SSU-RPB2 combined dataset, rDNA nucleotide data and RPB2 amino acid data were specified as distinct partitions to allow the use of the GTR+I+G model of evolution for SSU and LSU sequences and an empirical amino acid model (Whelan & Goldman 2001) for RPB2 sequences. Bayesian analyses for the larger LSU-SSU dataset consisted of two parallel searches each run for 5 000 000 generations, whereas analyses of the LSU-SSU-RPB2 dataset consisted of two searches run for 2 000 000 generations. An incremental heating scheme for analyses used the default settings in MrBayes (i.e. three heated chains and one cold chain). For the LSU-SSU dataset, trees sampled prior to the chains reaching a split deviation frequency of 0.05 were discarded as the ‘burn-in’, while the remaining trees were used to calculate the Bayesian posterior probabilities (PP) of the clades. For the LSU-SSU-RPB2 dataset, trees prior to stabilizing at < 0.01 average standard deviation between chains were discarded as ‘burn-in’ and the remaining trees were used to calculate the Bayesian PPs of the clades.

Based upon the phylogenetic analyses, constraint parsimony analyses of the combined LSU-SSU-RPB2 dataset were constructed in which Phymatotrichopsis or Rhizinaceae were forced into monophyly with alternative distinct lineages or outside the Pezizomycetes (Table 2). Constraint topologies were manually specified in PAUP v. 4.0b10 and heuristic searches of 1 000 replicates, saving only those trees in agreement with the forced constraint, were conducted using the same settings as the parsimony searches described above. The resulting trees were compared using the nonparametric comparison test of Templeton (Templeton 1987).

Table 2

Impact of phylogenetic constraints on the position of Phymatotrichopsis omnivora (Po) within a 31-taxon dataset (Fig. 3) on the resulting tree scores (#MPTs = number of equally most parsimonious trees; CI = consistency index; p = probability from a non-parametric two-tailed test (Templeton 1987), where trees with p < 0.05 are rejected as significantly worse.

Constraint	#MPTs	Length (steps)	CI	p
None	18	3123	0.601	best
Rhizinaceae with lineage B	6	3123	0.601	0.995
Rhizinaceae with lineage C	6	3125	0.600	0.637–0.732
Rhizinaceae and Caloscypha within lineage B	18	3128	0.600	0.535–0.603
Rhizinaceae with lineage A	3	3163	0.593	0.0003
Rhizinaceae with Pezizaceae	15	3176	0.591	< 0.0001
Po only with lineage A	6	3342	0.561	< 0.0001
Po only with lineage B	3	3331	0.563	< 0.0001
Po only with lineage C	3	3409	0.550	< 0.0001
Po only with Caloscypha	6	3320	0.565	< 0.0001
Po only outside Pezizomycetes	9	3341	0.562	< 0.0001

RESULTS

Phymatotrichopsis omnivora isolates

Besides isolates from ATCC, several isolates were cultured from alfalfa and cotton fields displaying characteristic symptoms (Fig. 1a, b) and signs of Phymatotrichum root rot. Mycelial strands were often observed on infected cotton roots (Fig. 1c), but were less conspicuous on alfalfa roots (not shown). Under magnification, mycelial strands were hirsute with acicular hyphae (Fig. 1d), some of which displayed cruciform branching (Fig. 1e). Though strands were rhizomorphic in appearance, with a melanised rind consisting of polygonal plectenchymatous cells (Fig. 1f), no obvious apical meristems were observed, and so would be better termed ‘mycelial cords’ (Kirk et al. 2001). One isolate, OKAlf8, formed typical sporemats on the surface of black clay (Fig. 1g), in which OKAlf8-inoculated plum trees had been potted. These sporemats developed the characteristic globose conidiophores with botryose blastoconidia borne singly on denticles (Fig. 1h–k). In a few cases, clavate or moniliform conidophores with apically borne conidia formed (Fig. 1l, m), similar in appearance to the ‘basidia’ observed previously (Baniecki & Bloss 1969). Examined herbarium specimens from FH of P. omnivorum possessed either characteristic hirsute mycelial cords (‘Ozonium’ stage) on cotton roots (GLH #2869 and GLH #2870) or crustose sporemats adhering to peds of black clay (GLH # 2868). Upon microscopic examination, excised pieces from the sporemat were not found to possess any readily apparent conidiophores; however, characteristic hirsute mycelial cords were observed ramified throughout the soil underlying the sporemats (data not shown).

Molecular data

Fifty six new sequences were determined in this study from Phymatotrichopsis omnivora, Pulchromyces fimicola, Psilopezia cf. nummularialis and Psilopezia deligata (Table 1). Efforts to amplify RPB2 from Ps. nummularialis were unsuccessful. The six β-tubulin sequences from P. omnivora were determined to not be phylogenetically informative (data not shown) and thus not included in phylogenetic analyses. From the three herbarium specimens of P. omnivorum, a partial ITS sequence was amplified only from the sporemat specimen (GLH #2868) using one of four primer pairs attempted (data not shown). Based on the alignment of this sequence with ITS sequences from over one hundred other P. omnivora isolates, the herbarium specimen sequence was most similar to P. omnivora isolates from El Campo, TX (100 % identity, 302/302) and the ATCC 48084 isolate (99 % identity, 302/303), which belong to an ITS haplotype common in southern Oklahoma and throughout eastern and central Texas (data not shown).

LSU and SSU gene tree

No supported conflict (BP ≥ 75 %, PP ≥ 95 %) was detected between the individual LSU and SSU gene trees. The combined dataset consisted of 2 743 characters of which 774 were parsimony informative. Parsimony analyses resulted in 6 equally most parsimonious trees (MPTs). The strict consensus tree of all MPTs was nearly completely resolved, except for a trichotomy of the three species of Psilopezia (indicated with an asterisk in Fig. 2). Nevertheless, many of the deeper branches have only low BP support. Bayesian analyses reached an average standard deviation of split frequencies below 0.05 after approximately 377 000 generations and the first 3 770 trees were excluded as the ‘burn-in’. Bayesian PPs supported many of the terminal relationships in the phylogeny with confidence but, as with BPs, failed to support some of the deeper nodes.

Fig. 2

Phylogenetic relationships of Phymatotrichopsis omnivora and Pulchromyces fimicola among a broad sampling of Pezizomycetes inferred from combined analyses of LSU and SSU rDNA. One of 6 most parsimonious trees is shown here. Terminal taxa represent individual specimens (see Table 1). Only one branch, indicated with an asterisk, collapses in the strict consensus tree of all MP trees. Numbers by branches are MP bootstrap proportions ≥ 70 %. Thickened branches indicate Bayesian posterior probabilities ≥ 95 %, obtained from a 50 % majority rule consensus tree of the 46 230 trees sampled from a Bayesian MCMC analysis. The three primary lineages are labelled A, B and C for discussion.

Phymatotrichopsis omnivora and Pulchromyces fimicola were nested within the Pezizales (Fig. 2). Phymatotrichopsis omnivora formed a monophyletic group with Rhizina undulata and three species of Psilopezia (Rhizinaceae), although with only low support (BP 56 %, PP 72 %). The lineages B (Morchellaceae–Discinaceae–Helvellaceae–Tuberaceae) and C (Pyronemataceae–Ascodesmidaceae–Glaziellaceae–Sarcoscyphaceae–Sarcosomataceae–Chorioactidaceae), Rhizinaceae and Caloscyphaceae formed a strongly supported monophyletic group (BP 93 %, PP 100 %). Parsimony analyses suggested that Caloscyphaceae was a sister group to a clade of the lineages B and C and Rhizinaceae (BP 78 %). Lineage C was strongly supported (BP 96 %, PP 100 %), whereas the relationships between Rhizinaceae and the lineages B and C were without support. Pulchromyces fimicola was nested within lineage C, but its placement among members of Pyronemataceae and Ascodesmidaceae was uncertain (Fig. 2). LSU and SSU rDNA sequences from Phymatotrichopsis omnivora showed several substitutions or deletions (17/1404 bp in the LSU region (1.21 %), 18/1741 bp in the SSU region (1.03 %)). The two available isolates of Pulchromyces fimicola had identical sequences through 2 989 bases of the SSU, ITS, and 5’-LSU regions.

Combined LSU, SSU genes and RPB2 protein tree

Overall no supported conflict (BP ≥ 70 %, PP ≥ 90 %) was detected between the individual trees constructed from LSU and SSU rDNA and RPB2 amino acid sequences. The combined dataset consisted of 6 194 characters of which 757 were parsimony informative. Parsimony analyses resulted in 18 MPTs (Fig. 3). The strict consensus tree of all MPTs was highly resolved and the majority of nodes were well supported by BP. Bayesian analyses reached an average standard deviation of split frequencies below 0.01 after approximately 180 000 generations and the first 2 000 trees were excluded as the ‘burn-in’. Bayesian PPs supported many of the terminal, as well as, deep nodes in the phylogeny with confidence.

Fig. 3

Phylogenetic relationships of Phymatotrichopsis omnivora with selected Pezizomycetes based on DNA sequences of SSU and LSU rDNA and deduced amino acid sequences of RPB2. One of 18 most parsimonious trees is shown here. Branch support at nodes are MP bootstrap proportions ≥ 70 % (number before ‘/’) and Bayesian posterior probabilities ≥ 95 % (number after ‘/’). Branches that collapsed in a strict consensus of the MP trees or the trees retained in the Bayesian analysis are indicated by ‘*’. Orbilia auricolor (Orbiliomycetes) was used as the outgroup to root the tree (James et al. 2006). The three primary lineages are labelled A, B and C and the Rhizinaceae is shaded yellow for discussion.

Parsimony analyses of the combined LSU-SSU-RPB2 dataset recovered the same major lineages, with high BP support, as those found with support in analyses of the LSU-SSU alignment. Phymatotrichopsis omnivora was strongly supported within the family Rhizinaceae (BP 86 %, PP 100 %). Bayesian analyses suggested that Rhizinaceae was a sister group to the lineages B and C (PP 100 %), whereas the relationship between Rhizinaceae and lineages B and C was unresolved in MP analyses (Fig. 3). As in analyses of the LSU-SSU alignment, the Ascobolaceae and Pezizaceae were not supported as a distinct lineage (A). Nevertheless, the two families were resolved as sister taxa or successive sister taxa to the rest of the Pezizales (Fig. 2, 3).

Parsimony trees resulting from constraint analyses that forced Phymatotrichopsis omnivora to group outside of Rhizinaceae, with either lineage A, B or C, Caloscyphaceae, or outside Pezizomycetes, or with Rhizinaceae and lineage A were strongly rejected using the Templeton test (P < 0.0001; Table 2). However, those trees recovered from analyses forcing Rhizinaceae to form a monophyletic group with Morchellaceae–Discinaceae– Helvellaceae (lineage B), as seen in MP analyses of the LSU-SSU dataset (Fig. 2), could not be rejected (p = 0.995). Forcing Rhizinaceae with lineage C or with Caloscyphaceae and lineage B also could not be rejected (p = 0.637–0.732 or p = 0.535–0.603, respectively).

DISCUSSION

Neither Sistotrema brinkmannii nor Phanerochaete omnivora represent the teleomorph of the cotton root rot pathogen. Phymatotrichopsis omnivora is not a member of the phylum Basidiomycota. Instead, Phymatotrichopsis omnivora is an anamorphic (mitosporic) member of the phylum Ascomycota, class Pezizomycetes (order Pezizales, operculate discomycetes). Our phylogenetic analyses place Phymatotrichopsis omnivora in Rhizinaceae with Psilopezia and Rhizina. Rhizinaceae was resurrected as a monotypic family based on molecular data (O’Donnell et al. 1997), and recently, species of Psilopezia were suggested to belong to the family (Hansen & Pfister 2006). Whether Rhizinaceae represents an independent lineage within Pezizomycetes, as suggested by Hansen & Pfister (2006) and our Bayesian analyses (Fig. 3), is still uncertain, as we are unable to reject constraint topologies that force Rhizinaceae to group with lineage B (with or without Caloscyphaceae) or lineage C. Based on SSU and LSU sequences, Pulchromyces fimicola (formerly Phymatotrichum fimicola) is also a member of the class Pezizomycetes, but is clearly not congeneric with Phymatotrichopsis. Instead, it is closely related to members of the C-lineage, possibly in Pyronemataceae or Ascodesmidaceae. Pulchromyces has been found on the dung of mice, otters, bats and shrews, in temperate and tropical regions, in Ghana, Panama and the United States (Pfister et al. 1974). A number of genera shown to be closely related to Pulchromyces, namely Ascodesmis, Lasiobolidium, Lasiobolus and Pseudombrophila (Fig. 2), are similarly fimicolous, although the fimicolous habit has been multiply derived throughout the Pezizomycetes and many other groups of fungi. A better taxon sample of these minute Pezizomycetes and related anamorphs will be required to settle the taxonomic position of Pulchromyces at the family level.

The anamorphic morphology of Phymatotrichopsis omnivora partially supports its placement in the Pezizomycetes. The botryoblastoconidia produced by Phymatotrichopsis omnivora are also observed in many of the pleomorphic Pezizomycetes in which anamorph–teleomorph associations have been determined. For example, the anamorphic genera Chromelosporium, Oedocephalum, Ostracoderma, Glischroderma and Dichobotrys, are associated with the Pezizomycetes meiosporic genera, Peziza (first four) and Trichophaea (Paden 1972, Hennebert 1973, Hansen et al. 2001). However, botryoblastosporic reproduction occurs in several classes of both the Ascomycota and Basidiomycota. Such anamorphic genera are found in the Leotiomycetes (inoperculate discomycetes), in Botrytis, Streptobotrys, Amphobotrys, and Veruccobotrys, and in the Agaricomycetes (Homobasidiomycetes), in Spiniger (Hennebert 1973, Stalpers 1974, Kiffer & Morelet 2000). Thus, botryoblastosporic patterns of conidiogenesis arose several times during fungal evolution and may have limited value for taxonomic classifications above genus.

Rhizomorph-like, mycelial strands are formed by both Phymatotrichopsis omnivora (Lyda & Kenerley 1992) and, proposed confamilial, Rhizina undulata (Booth & Gibson 1998). Conspicuous mycelial strands are often found on the infected roots of host plants and are often used by plant pathologists to diagnose the root rots caused by either fungus. Besides soilborne dissemination, the mycelial strands connect the reproductive structures, sporemats of Phymatotrichopsis omnivora or apothecia of Rhizina undulata, to nutritional sources. The root-like nature of the apothecial mycelial strands of Rhizina was the namesake character of the genus (Fries 1822). The mycelial strands of Phymatotrichopsis eventually form long-lived, hypogeous sclerotia (King & Loomis 1929, Neal 1929, King et al. 1931), while sclerotia have not been reported for Rhizina, which survives as thick-walled ascospores that are stimulated to germinate by fire (Jalaluddin 1967b).

The majority of the Pezizomycetes traditionally have been considered saprobic, but the trophic strategies of most species are not well-studied and remain undocumented. The inclusion of the Tuberales, which are assumed to be mainly mycorrhizal, in the Pezizales (Trappe 1979, Læssøe & Hansen 2007) and molecular studies identifying numerous other Pezizomycetes as ectomycorrhizal associates (Dahlstrom et al. 1999, Fujimura et al. 2005, Tedersoo et al. 2006) has revealed mycorrhizae as a major ecological niche of many pezizalean fungi. On the other hand, the ecology of Phymatotrichopsis omnivora, a mostly hypogeous plant pathogen with an extensive dicotyledonous host range (Lyda 1978), is relatively rare among the Pezizomycetes. Rhizina undulata is also a plant pathogen that infects a wide range of conifers (Gremmen 1971). Other plant pathogenic Pezizomycetes include the conifer seed pathogen Caloscypha fulgens (Paden et al. 1978) and the Strumella canker fungus, Conoplea globosa (= Strumella coryneoidea; mitosporic Urnula) (Kopcke et al. 2002, Wang et al. 2005). Also, species of Octospora, Lamprospora and Neottiella form obligate associations with numerous bryophytes, which have been interpreted as parasitic (Döbbeler 1979, Benkert 1993, Davey & Currah 2006). Both Phymatotrichopsis and Rhizina also colonise dead plant debris in field situations, acting as facultative saprobes, and utilise these substrates for reproduction (Jalaluddin 1967a; Rush & Gerik 1989).

Very few similarities in apothecia morphology support a close relationship of Psilopezia with Rhizina (Hansen & Pfister 2006), and no obvious mitosporic or somatic similarities support a confamilial relationship with Phymatotrichopsis. The little that is known about the natural history of Psilopezia suggests a saprobic life style on wet, rotted wood (Pfister 1973), while Rhizina and Phymatotrichopsis are plant pathogens with a facultative saprobic phase. Nevertheless, based on our phylogenies of combined rDNA and RPB2 sequences, the monophyly of the Rhizinaceae, including Rhizina undulata, Phymatotrichopsis omnivora and Psilopezia deligata, was highly supported (BP 86 %, PP 100 %) and constraint topologies that forced Phymatotrichopsis to group outside Rhizinaceae were rejected. The relationships among Psilopezia, Rhizina and Phymatotrichopsis were, however, not resolved with confidence (the branch collapses in the strict consensus tree of all MPTs, and PP 90 %). Psilopezia may possess an as yet unrecognised pathogenic phase, or represents a saprotrophic sister group to a derived parasitic clade of Rhizina and Phymatotrichopsis. More members of the Rhizinaceae must be identified and characterized before further inferences on the evolution of their nutritional strategies can be clarified.

Knowledge of the correct phylogenetic placement of the cotton root rot pathogen as a member of the Pezizomycetes (Ascomycota), and not Agaricomycetes (Basidiomycota), will have significance in detecting the pathogen in the field and in developing methods of chemical or biological control. Also, it will facilitate current efforts to assemble and annotate the genome sequence of Phymatotrichopsis omnivora strain OKAlf8 (http://www.genome.ou.edu/fungi.html) through comparative genomics with related ascomycetes. In addition to Phymatotrichopsis, genomic projects of two other Pezizomycetes, Tuber melanosporum and T. borchii, are ongoing (Poma et al. 2006, Lazzari et al. 2007; http://mycor.nancy.inra.fr/IMGC/Tuber-Genome/index.html). The insights into the genetic underpinnings of this fascinating, but understudied, class of fungi should prove fruitful.

Nomenclature and typification

Given the economic importance of Phymatotrichopsis omnivora and the presence of ITS sequence variation among strains of this species (data not shown), it is important that a consensus is reached as to the correct author citation and (therefore) typification of this species. Duggar (1916) explicitly transferred the species Ozonium omnivorum Shear to the genus Phymatotrichum because of the presence and nature of conidia in specimens of what he believed to be the same species as described by Shear (1907) and thus did not designate a type specimen among the various collections he referred to. The decision of Hennebert (1973) to attribute the name solely to Duggar therefore left the species without a type specimen. The relevant sections of the International Code of Botanical Nomenclature (ICBN; McNeill et al. 2006) are Art. 7.4, 48.1 and 59.6. Article 7.4 states that “a new name formed from a previously published legitimate name (stat. nov., comb. nov.) is, in all circumstances, typified by the type of the basionym”, unless the author(s) explicitly excluded the type of the basionym (Art. 48.1) or explicitly described a new morph, simultaneously meeting all the requirements for description of a new species (Art. 59.6) (McNeill et al. 2006). The decision by Hennebert (1973) rests on a narrow definition of Art. 59.6, that a conidial form should represent a new ‘morph’ separate from the ‘sterile’ mycelium that produced it, and goes against the growing consensus among mycologists of the principle of ‘one fungus – one name’ (Hennebert 1993). We therefore choose to treat the decision by Hennebert (1973) to attribute the basionym of Phymatotrichopsis omnivora to Duggar as an error to be corrected under Art. 33.6, resulting in the authorities for the combination of Phymatotrichopsis omnivora (Shear) Hennebert and the restitution of Shear’s type specimen (C.L. Shear 1447, BPI 455660) as holotype. The living culture, strain OKAlf8 (ATCC MYA-4551; isolated from infected alfalfa roots growing near Belleville, OK by S. Marek, August 2003), which is currently the basis of genome sequencing (http://www.genome.ou.edu/fungi.html), provides a sound anchor for future molecular studies.