Phylogeny and morphological analyses of Penicillium section Sclerotiora (Fungi) lead to the discovery of five new species

Phylogeny of Penicillium section Sclerotiora is still limitedly investigated. In this study, five new species of Penicillium are identified from the samples collected from different places of China, and named P. austrosinicum, P. choerospondiatis, P. exsudans, P. sanshaense and P. verrucisporum. The conidiophores of P. austrosinicum and P. exsudans are monoverticillate like most members of the section, while the rest species are biverticillate similar to the only two species P. herquei and P. malachiteum previously reported in the section Sclerotiora. The phylogenetic positions of the new taxa are determined based on the sequence data of ITS, BenA, CaM and RPB2 regions, which reveals that all the species with biverticillate condiophores form a well-supported subclade in the section. The new Penicillium species clearly differ from the existing species of the genus in culture characteristics on four standard growth media, microscopic features, and sequence data. Morphological discrepancies are discussed between the new species and their allies.

Although the three strains of P. choerospondiatis were from different sites, they are morphologically identical. As to their sequence divergences, very few variations were detected. Their ITS and RPB2 regions are basically the same, six bp differences are found in the BenA gene, and three bp divergences exist in the CaM region among collections, which are treated as infraspecific variations.
Notes: In the phylogenetic tree (Fig. 1), P. verrucisporum appears as an independent terminal branch in subclade II. Among species of the subclade, P. choerospondiatis is similar in comparatively short stipes, but differs in no growth on CZ at 25 °C, less metulae per verticil, larger metulae and phialides, larger and finely roughened conidia, and slower growth on CYA, MEA and YES at 25 °C. The morphological differences among the related taxa are detailed in Table 1.

Discussion
Members of Penicillium section Sclerotiora are characterized by the pigmented mycelia in shades of yellow and/or orange in culture, reverse view of colony yellow, orange or red, and sclerotia and cleistothecia if present bright-colored 12 .
The conidiophores of the species in this section were mostly monoverticillate, and conidiophore branching pattern has not been considered of phylogenetic importance. In this work, we add three new species bearing biverticillate conidiophores to the section. Based on the four-gene phylogeny, three subclades in the section are recognized (Fig. 5). Species with monoverticillate conidiophores are in subclades I and III, while all those possessing biverticillate conidiophores are in subclade II. They join the species P. herquei and P. malachiteum recognized previously with a similar conidiophore branching pattern, which reveals that conidiophore branching pattern is phylogenetically informative. The result highlights that the morphological feature of conidiophore branching pattern is in accordant with the phylogenetic analysis and may be used as a reliable character in taxonomy of the section Sclerotiora. Most species in Penicillium section Sclerotiora bearing monoverticillate conidiophores have vesiculate conidiophore apices, except for P. adametzii, P. angulare and P. viticola 11,35 . Although these three species do not form vesiculate conidiophores, they are scattered among or intertwined with those producing vesiculate conidiophores in the phylogenetic tree (Fig. 5). The former two species are located in subclade III and the latter is in subclade I. For the new species in subclade I, the conidiophores of P. austrosinicum and P. exsudans are vesiculate. The feature vesiculate conidiophore might not be phylogenetically informative for the group.
Sclerotia have been produced by seven members of the section Sclerotia 28 . The color of sclerotia varies among species. For example, P. austrosinicum gives rise to cream to yellow sclerotia, and P. hirayamae, P. sclerotiorum and P. vanoranjei form orange ones on CYA at 25 °C 11,34 . In fact, sclerotia appear sporadically in Penicillium species across sections 34 , such as P. corvianum in section Canescentia producing brown sclerotia 26 , P. macrosclerotiorum in section Gracilenta giving rise to the white ones 36 , and P. salamii in section Brevicompacta having orange ones 21 . The production of sclerotia, as a morphological feature, does not reflect the phylogenetic relationships among species of the genus.
Species in section Sclerotia have been isolated from diverse substrates including soil, plants, and insects 28,33,37 . Apart from the species dominant in soil, the species collected from plant materials comprise a substantial proportion. Three of our five new species are from plant debris: P. choerospondiatis infecting the fruits of Choerospondias axillaris, and P. austrosinicum and P. exsudans from the rotten fruits of unidentified plants. Furthermore, P. herquei is on a leaf of Agauria pyrifolia, P. hirayamae is on cereals 11 , P. viticola infects grape 35 Table 2. Fungal species and sequences used in phylogenetic analyses. T: ex-type strains. n.a.: data not available. * The accessions in bold are newly obtained in this study.
been isolated from nuts of Juglans nigra and Carya ovata 34 . Some species are fungicolous: P. angulare is from a wood-decaying polypore 38 . Some species are living as plant endophytes [8][9][10] . Along with the identifications of new species, a broader range of substrates will be expected. Penicillium herquei can be planted by the nonsocial leaf-rolling weevil Euops chinesis [39][40][41] , which indicates that the fungus evolved the ability to adapt divergent niches. The genus Penicillium has been established for more than 200 years. New species of this genus have been increasingly found from different regions of the world, especially during the past two decades. The results of this study broaden our knowledge of the species diversity of this group. It is undoubted that more Penicillium species will be found in the unexplored areas of China as well as in other regions of the world based on the integrated or comprehensive studies of morphology, cultural characteristics and sequence data. Morphological observations. Morphological characterization of each sample was conducted following the standardized methods established by Visagie et al. 17 . Four standard growth media were used including the Czapek yeast autolysate agar (CYA, yeast extract Oxoid), malt extract agar (MEA, Amresco), yeast extract agar (YES) and Czapek's agar (CZ). The methods for culture inoculation, incubation, microscopic examinations and digital recordings were described in our previous study 42 . DNA extraction, PCR amplification and sequencing. Fungal cultures were grown on the potato dextrose agar (PDA) medium for 7 d and then harvested for DNA extraction using the Plant Genomic DNA Kit (DP305, TIANGEN Biotech, Beijing, China). The fragments of the internal transcribed spacer region (ITS), beta-tubulin (BenA), calmodulin (CaM) and the second largest subunit of RNA polymerase II (RPB2) genes were amplified by PCR using the primers reported by Visagie et al. 17 . The products were purified and subject to sequencing on an ABI 3730 DNA Sequencer (Applied Biosystems).

Phylogenetic analyses.
The sequences obtained in this study have been deposited in GenBank. The accessions and those retrieved from GenBank 17 are listed in Table 2. The sequences of each gene (i.e., ITS, BenA, CaM or RPB2) were aligned using the program MAFFT (ver. 7.221) 43 , and subsequently processed with BioEdit (ver. 7.1.10) 44 . The individual or concatenated gene data sets were used to generate the respective Maximum-Likelihood (ML) trees using the software MEGA (ver. 6.0.6) 45 withthe most suitable nucleotide substitution model and 1,000 replicates of bootstrap tests. Bayesian Inference (BI) analysis was performed with MrBayes (ver. 3.2.5) 46 using a Markov Chain Monte Carlo (MCMC) algorithm. Appropriate nucleotide substitution models and parameters were determined by using the program Modeltest (ver. 3.7) 47 . Four MCMC chains (one cold and three heated) were run for one million generations with the trees sampled every 100 generations. The first 25% trees were excluded as the burn-in phase of the analyses, and the posterior probability (PP) values were estimated with the 75% remaining trees. The consensus trees were viewed in FigTree (ver. 1.3.1; http://tree.bio.ed.ac.uk/software/ figtree/). The species Penicillium levitum in section Lanata-Divaricata was used as an outgroup.