A phylogeny for the plant pathogen Piptoporellus baudonii using a multigene data set

Piptoporellus baudonii is proposed as a new combination for Laetiporus baudonii in the Polyporales (Basidiomycota) based on morphological and molecular features. This parasitic macrofungus attacks cashew trees, Eucalyptus , cassava, Tectona , and some indigenous trees in southern regions of Tanzania and poses a serious threat to agroforestry and livelihood conditions in the area. Phylogenetic trees were produced from partial sequences of three rDNA gene regions and a portion of translation elongation factor 1-alpha ( TEF1 ) gene of Laetiporus baudonii for compar-isons with samples from the antrodia clade. Our results reveal a strongly supported group of L. baudonii with Piptoporellus in Fomitopsidaceae. Piptoporellus baudonii shares many morphologi- cal features with other members of Piptoporellus but differs from them in having broadly ellipsoid or rarely ovoid basidiospores. Both morphological and phylogenetic evidence justify the placement of L. baudonii in Piptoporellus together with the three other known species in the genus.


INTRODUCTION
Laetiporus baudonii (Pat.) Ryvarden (Ryvarden 1991) is a large parasitic polypore fungus well known in Africa. Reports suggest that the species afflicts a wide range of economically important plants in various parts of Africa (Van der Westhuizen 1973), including Congo (Patouillard 1914), on Manihot in Madagascar (Heim 1931), on Cassia siamea, Khaya senegalensis, and Citrus species in Ghana (Ofosu-Asiedu 1975), in tea plantations in Malawi (Rattan and Pawsey 1981), on Eucalyptus in South Africa (Van der Westhuizen 1973), on Acacia tortilis in Kenya, on Anacardium occidentale (Cashew tree) in Tanzania, and on various indigenous plants (Sijaona 2007). Death of the host plants is associated with the appearance of large bright orangeyellow poroid basidomata at the bases of trunks and on the ground among the infected trees. These fruit bodies develop from a pseudosclerotial tissue composed of a mat of coarse creamy-yellowish mycelium and sand grains that cover the roots and underground parts of the stems of afflicted trees (Van der Westhuizen 1973). Root infection causes early symptoms of loss of the deep green color of leaves and the yellowing of leaves on individual branches, followed by frequent and rapid wilting of the leaves (Rattan and Pawsey 1981).
Since it was first described from Congo, Polyporus baudonii Pat. (Patouillard 1914) has been less well understood in a systematic framework that has resulted in the naming of several taxonomic synonyms (Westerhuizen 1973). In a study investigating the morphology of fruit bodies and cultures isolated from a parasitic fungus found in South Africa, and comparison with the descriptions of the type specimens of Polyporus baudonii and Phaeolus manihotis R. Heim, Van der Westhuizen (1973) observed morphological similarities between the studied parasitic fungus and Phaeolus manihotis, with the former generally having larger basidiomata than the latter. Van der Westhuizen (1973) confirmed the morphological resemblance between Phaeolus manihotis and Polyporus baudonii as suggested by Browne (1968). The study also referred to the fungus as Polyporus baudonii due to the observation of a dimitic hyphal system that differed from that of the type species of Phaeolus (Donk 1960), which has a monomitic hyphal system. Furthermore, Van der Westhuizen (1973) observed complex morphological features that did not match those of any known genus at that time. Because of this, Ryvarden described the new genus Pseudophaeolus Ryvarden, with Ps. baudonii (Pat.) Ryvarden as the type (Ofosu-Asiedu 1975). Later, Ryvarden (1991) accepted P. baudonii in Laetiporus as L. baudonii (Pat.) Ryvarden.
The association of a polyporoid macrofungus with wilting of cashew nut trees, cassava, and Eucalyptus trees in the Mtwara and Lindi regions of southern Tanzania has been reported (Sijaona 2003(Sijaona , 2007. Morphologically, this fungus resembled L. baudonii. In order to ascertain its identity, and given the inconsistencies in classifications based on morphological features, we incorporated molecular data for an accurate placement. This study aims to ascertain the phylogenetic position of L. baudonii using a four-gene region data set based on analyses of partial sequences from three nuc rDNA gene regions and translation elongation factor 1-alpha (TEF1).

MATERIALS AND METHODS
Study site.-The material studied originated from the Mtwara region in southern Tanzania (ca. 10°10′-11°30′S to 37°58′-40°25E), where cashew is a major cash crop that is frequently affected by this fungus. The region is mostly characterized by a high relative humidity (87% in Feb-Jun, which is usually the long rainy season), whereas the lowest relative humidity of 63% occurs in Sep-Oct. Temperatures vary with cold months (Jun-Sep) with an average of 19.5 C, whereas hot months reach above 30 C (Sep-Dec). The soils are sandy loam or red clay soil, receiving an annual rainfall of 900-1200 mm (Sijaona 2003;Sijaona and Shija 2005).
Field sampling.-An unidentified fungus was first observed by the staff at the Naliendele Agricultural Research Institute (NARI). A field trip was arranged by the first author during the rainy season of 2011. Basidiomata were collected, and each collection locality was recorded using Global Positioning System (GPS) (Tibuhwa et al. 2010). Prior to collecting, mushrooms were photographed in situ (FIG. 1A-D). Field identification features such as sporocarp shape, color, smell, and color changing on bruising were recorded.
Collected samples were brought to the Department of Molecular Biology and Biotechnology Laboratory at the University of Dar es Salaam (UDSM) where parts of the mushrooms were oven dried at 50 C for 8 h. Vouchers were deposited at UDSM and UPS herbaria, the latter registered in Index Herbariorum (Thiers [continuously updated]). Color descriptions were based on Kornerup and Wanscher (1962).
Four collections of Pseudophaeolus baudonii (Pat.) Ryvarden (one from Zimbabwe, one from Senegal, one from Ghana, and one from Uganda) were obtained from the herbarium at Oslo University.
Microscopy.-Microscopic characters were recorded from specimens preserved by dehydration using silica gel and later observed in a 10% ammonium solution in an aqueous solution of Congo red. Twenty measurements of basidiospores and basidia were analyzed statistically, with the results presented as (min)A −SD-A+SD(max), where min is smallest observed value, A is the arithmetic average, SD is the standard deviation, and max is largest observed value for the measured specimen. Basidiospore characters, which include their size, shape, and reactions to Melzer's reagent, were observed. The hyphal system (monomitic, dimitic, or trimitic) was studied as well as the type of septa (simple septa or clamped septa in generative hyphae).
Molecular study.-This part of the study was carried out at the Department of Organismal Biology, Uppsala University. Total DNA was extracted from the inner part of the basidiomata, preferably from the hymenium to avoid contamination, following the protocol of the Plant Genomic DNA Extraction Kit (Qiagen, Hilden, Germany). Diluted (10 −1 -10 −3 ) and undiluted DNA was used for polymerase chain reaction (PCR) amplifications. For herbarium material, nested PCR was employed due to low yields from initial PCR. The 5′ end of the nuc 28S rDNA (28S), nuc rDNA internal transcribed spacer region ITS1-5.8S-ITS2 (ITS), a portion of nuc 18S rDNA (18S), and a portion of TEF1 were amplified. Primers used for amplification and sequencing included LR0R, LR5, and LR7 (Vilgalys and Hester 1990) for 28S; ITS1F (Gardes and Bruns 1993), 5.8S (Vilgalys and Hester 1990), ITS3, ITS4, and ITS5 (White et al. 1990) for ITS; NS1, NS7, and NS4 for 18S (White et al. 1990); and EF1-983F, EF-2218R, and EF1-1567R for TEF1 (Rehner and Buckley 2005). For PCR amplification, we used the AccuPower PCR PreMix (Bioneer, Daejeon, Korea), adding 3 µL diluted or undiluted DNA, 1.5 µL of each primer (10 µM), and water to a total volume of 20 µL. Thermal cycling parameters were as described in Savić and Tibell (2009) and touchdown PCR as in Matheny (2005) for TEF1. Amplification products were visualized on 1.5% agarose gels stained with gel red, and PCR products were purified using Illustra ExoStar buffer (GE Healthcare, UK) diluted 10×, following the manufacturer's protocol. Sequencing was carried out by Macrogen (www.macrogen.com).
Alignments and phylogenetic analyses.-Additional sequences from GenBank (TABLE 1) were chosen to reconstruct a multigene alignment including as much taxonomic coverage of Fomitopsidaceae and Laetiporaceae following Han et al. (2016) and Song et al. (2018). Alignments were performed using MAFFT 7 online (https//mafft.cbrc.jp/ alignment/server/; Katoh et al. 2019) and viewed and manually adjusted, where necessary, using AliView 1.26 (Larsson 2014). Ambiguously aligned regions were excluded prior to phylogenetic analyses. We retained only the 5.8S region of ITS for the combined data set, since the neighboring regions (ITS1 and ITS2) were poorly aligned. A conflict among single-locus data sets was considered significant if a well-supported monophyletic group (posterior probability [PP] ≥0.95) was found to be well supported as nonmonophyletic when different loci were used. Further analyses were carried out after concatenation using SequenceMatrix (Vaidya et al. 2011).
The best-fit model of DNA evolution for the analyses, for both individual codon positions and genes, was obtained using Akaike information criterion (AIC) as implemented in MrModeltest 2.3 (Nylander 2004). The GTR+I+G model was employed across sites for 28S, 18S, and the 2nd codon of TEF1. For the 1st codon of TEF1, the model F81+I+G was applied, whereas HKY+G was  implemented for both the 3rd codon of TEF1 and 5.8S. Bayesian inference (BI) was conducted using MrBayes 3.2.6, and branch support was estimated by the posterior probability (PP) (Ronquist and Huelsenbeck 2003;Ronquist et al. 2012). Two independent runs were executed, each with four Markov chains for 10 million generations, sampling trees every 100 generations. A 25% threshold was used as the burn-in. The Markov chain Monte Carlo (MCMC) analysis converged well in advance of the burn-in threshold, and chain mixing was found to be satisfactory as assessed by using Tracer 1.5 (Drummond et al. 2012). Maximum likelihood (ML) estimates were carried out by RAxML 8.2.10 using the GTR+G+I model of site substitution (Stamatakis 2014). Branch support was obtained by bootstrapping 1000 replicate data sets (Hillis and Bull 1993). Sequence alignments and phylogenetic trees were deposited in TreeBASE (submission ID 26318).

RESULTS
Phylogenetic analyses.-A total of 12 new sequences from three specimen vouchers were generated (TABLE 1). BLAST results from GenBank (accessed 23 Apr 2019), using BLASTn with "discontiguous megablast" (for cross-species comparison, searching with coding sequences) with sequences in this study, showed the highest sequence similarity to Piptoporus soloniensis (ITS query cover 97% and identity 76%).
The analyses contained a total of 249 sequences representing 64 species of the "antrodia clade," including the additional sequences downloaded from GeneBank and two species from the "core polyporoid" clade as outgroups. No significant incongruence among single gene trees was detected; hence, the four matrices were concatenated. The concatenated data matrix contained 2859 unambiguously aligned sites. The phylogeny of the "antrodia clade" and the position of Piptoporellus baudonii were inferred from four data sets: 58 sequences of 5.8S, 66 of 28S sequences, 63 of 18S, and 62 sequences of TEF1. The combined analysis recovered a phylogeny with two distinct clades representing Fomitopsidaceae and Laetiporaceae (FIG. 2). Clade annotations follow Han et al. (2016) Basidiomata 20-35 cm wide, pileate, stipitate, and seldom effused-reflexed forming shelves clustered in rosettes up to 85(-110) cm wide, total weight of up to 15 kg; upper surface of pileus bright orange-yellow (5A5-6) when young and fresh, rusty brown (7DE7) upon aging, surface soft without crust, with a few faint concentric brown zones. When aging the basidiomata undergo autolysis. Hymenophore poroid, pore surface concolorous with the upper surface of the pileus or slightly paler. Context fleshy, light ochraceous (2A4-5).
Ecology and distribution: On a wide range of hosts (TABLE 2); for additional host trees, see Van der Westhuizen (1973) and Rattan and Pawsey (1981). Basidiomata occurring during the rainy season (Nov-May). Widely distributed in Africa and Madagascar (Kile 2000); also recorded from Yemen (Al-Fatimi et al. 2005).

DISCUSSION
Laetiporus baudonii has had a controversial taxonomic history regarding its generic placement and species recognition as reported previously by Van der Westhuizen (1973). Here, we infer the phylogenetic position of the species using molecular data from four gene regions. In our analyses, L. baudonii from Zimbabwe clustered with two isolates from Tanzania to form a highly supported and distinct species-level lineage within Piptoporellus of the antrodia clade (FIG. 2). Phylogenetic evidence thus justifies the incorporation of Laetiporus baudonii in Piptoperellus along with other sampled members of the genus-P. soloniensis, P. hainanensis, and P. triqueter. Morphologically, P. baudonii is similar to other Piptoporellus in having annual, pileate, or stipitate basidiomata, similar pileus coloration, a dimitic hyphal system, thin-walled basidiospores, and presence of clamp connections. However, P. baudonii differs from the other Piptoporellus species by spore shape (Han et al. 2016).
For more than 10 years, farmers in Mtwara, Tanzania, have noticed fungal infections in plantations of cashew. Farmers also observed that when the fungus was found beneath a tree, the leaves of the tree were chlorotic and completely wilted. Defoliation and wilting of the attacked trees continued until the whole tree succumbed, with the wilt spreading to adjacent trees. However, P. baudonii has low host specificity. During our study, we observed its occurrence on a wide range of trees, including Eucalyptus, cashew tree, cassava trees, Tectona trees, and indigenous trees, in southern Tanzania (see TABLE 2). Researchers at the Naliendele Agricultural Research Institute, Mtwara, have confirmed these observations. A clear indication of local fungal infection was seen in plantations of cultivated Tectona grandis and Eucalyptus trees, where distinct patches of dry and dying trees were noted. Pathogenicity studies of P. baudonii and possible mitigation measures are ongoing.
Piptoporellus baudonii has a deep impact on agroforestry and ecology of the affected areas. Cashew is the main cash crop for more than 380 000 households in southeastern Tanzania, whereas cassava is their major staple food. Infection by P. baudonii, which attacks both their main cash crop and main staple food apart from other cultivated and wild forest trees, is of great concern in this region and a threat to agroforestry. There is concern that a collapse in this system would entail complex and multifarious socioeconomic problems. Mitigation procedures are essential to inhibit spread of the disease caused by P. baudonii.