Chlorovibrissea korfii sp. nov. from northern hemisphere and Vibrissea flavovirens new to China

A new species, Chlorovibrissea korfii, is described and illustrated, which is distinct from other members of the genus by the overall pale greenish apothecia 0.8–2.0 mm in diam. and 0.6–1.5 mm high, J+ asci 70–83 × 4.5–5.5 μm, and non-septate ascospores 44–52 × 1.2–1.5 μm. This is the first species of Chlorovibrissea reported from northern hemisphere. Vibrissea flavovirens is reported from China for the first time. Sequence analyses of the internal transcribed spacer of nuclear ribosomal DNA are used to confirm the affinity of the two taxa.

In China, only two Vibrisseaceous fungi were recorded, i.e. V. cf. sporogyra (Ingold) A. Sánchez from Hainan Province (Zhuang et al. 2002) and V. truncorum (Alb. & Schwein.) Fr. from Guizhou Province (He 1988). During our studies of the helotialean fungi from China, two newly collected specimens fit the circumscription of Vibrisseaceae and represent two species. Based on morphology and molecular data, one is proposed as a new species of Chlorovibrissea, and the other is identified as V. flavovirens (Pers.) Korf & J.R. Dixon which is new to China.

Materials and methods
Specimens were collected from Beijing and Yunnan Province in 2016, and apothecial gross morphology when fresh was photographed by a Canon PowerShot G16 digital camera. Dried apothecia were rehydrated with distilled water and sectioned at a thickness of 15 μm with a Yidi YD-1508A freezing microtome (Jinhua, China). Measurements were taken from longitudinal sections and from squash mounts in lactophenol cotton blue solution using an Olympus BH-2 microscope (Tokyo, Japan). Iodine reactions of ascus apparatus were tested in Melzer's reagent and Lugol's solution with or without 3% KOH solution pretreatment according to Baral (2009). Images were captured using a Canon G5 digital camera (Tokyo, Japan) attached to a Zeiss Axioskop 2 Plus microscope (Göttingen, Germany) for anatomical structure. Voucher specimens were deposited in the Herbarium Mycologicum Academiae Sinicae (HMAS). Name of the new species was formally registered in the database Fungal Names, one of the three official nomenclatural repositories for fungal names (Redhead and Norvell 2012).
Genomic DNA was extracted from dried apothecia using Plant Genomic DNA Kit (TIANGEN Biotech. Co., Ltd., Beijing, China). Materials were crushed in liquid nitrogen before extraction. The internal transcribed spacer of nuclear ribosomal DNA (ITS) were amplified and sequenced using primer pair ITS1 and ITS4 (White et al. 1990) according to Zheng and Zhuang (2016). Newly generated sequences were deposited in GenBank, and other sequences used in the phylogenetic analyses were retrieved from GenBank (Table 1). Holwaya mucida (Schulzer) Korf & Abawi was selected as * Numbers in bold indicate sequences produced by this study outgroup taxon. Alignment was generated and manually edited with BioEdit 7.0.5.3 (Hall 1999). Maximum parsimony (MP) and Neighbor-joining (NJ) analyses were carried out using PAUP*4.0b10 with parameters used by Zheng and Zhuang (2014). The topological confidence of the NJ and MP trees was assessed with bootstrap analysis using 1,000 replications, each with 10 replicates of random stepwise addition of taxa. The resulting trees were viewed in TreeView 1.6.6 (Page 1996).

Phylogenetic analyses
The ITS alignment consisted of 574 characters including gaps, of which 292 were constant, 52 were variable and parsimony-uninformative, and 230 were parsimonyinformative. Forty-five equally most parsimonious trees were yielded by MP analysis (Tree length = 1085, Consistency index = 0.4719, Homoplasy index = 0.5281, Retention index = 0.6739, Rescaled consistency index = 0.3180) and one of them was shown in Fig. 1. MP and NJ bootstrap proportions (BP) greater than 50% were labeled at the nodes. In the phylogenetic tree ( Fig. 1), species of Acephala, Chlorovibrissea, Phialocephala and Vibrissea formed four well-supported clades corresponding to each genus, and three of them further clustered together with high supporting values except for Chlorovibrissea, which showed a distant relationship with others. The undescribed species appeared as a distinct terminal lineage within Chlorovibrissea. ITS sequences of V. flavovirens from the Chinese, North American and New Zealand materials were of high similarity (99%) and formed a well-supported terminal branch. Etymology. The specific epithet is in memory of the late distinguished mycologist Dr. R.P. Korf.
Notes. This is the first report of V. flavovirens from China. The fungus was originally described from France and currently known in Denmark, Germany, Madeira, New Zealand, UK and USA (Korf 1974;Iturriaga 1995;Sandoval-Leiva et al. 2014). The Chinese collection agrees with the description of V. flavovirens by Iturriaga (1995). The ITS sequence of the Chinese specimen shared high similarity (99%) with those of North American and New Zealand materials, and the sequences of materials from different geographic regions formed a strongly supported terminal branch (Fig. 1).

Discussion
The three sexual genera in Vibrisseaceae are mainly differentiated by color of apothecia and structure of ectal excipulum. The excipular cells of Vibrissea are more or less angular to globose and lying at a high angle to the receptacle surface, while those of Chlorovibrissea and Leucovibrissea are rectangular and axes of cells are nearly parallel to the receptacle surface. Chlorovibrissea is different from Leucovibrissea in having greenish apothecia instead of the whitish ones. Differences between Chlorovibrissea and Vibrissea were found in ascal apex, which is round to somewhat truncate and with the apical ring placed subapically in the former genus, while that of the latter is acute with apical ring placed at the tip (Sandoval-Leiva et al. 2014). The ascal apex of C. korfii is somewhat conical and with apical ring placed apically, broader at tip and narrower downward (Fig.  2g), unlike that of other known Chlorovibrissea species. Further studies are needed to evaluate whether ascal apical apparatus is phylogenetically informative in Vibrisseaceae.
Chlorovibrissea was assumed to have a southern hemisphere distribution because species of the genus was never reported from north hemisphere (Kohn 1989;Sandoval-Leiva et al. 2014). The discovery of C. korfii from China expands the distribution of the genus to north hemispheres, which might break the assumption "the origin of the southern hemisphere vibrisseaceous fungi could be independent from the northern hemisphere representative" (Wang et al. 2006).
In the family Vibrisseaceae, sequence data of Leucovibrissea are not available. The two asexual genera Vibrisseaceae are Phialocephala and Acephala. Due to the heterogeneity of Phialocephala, only the type species, P. dimorphospora, and two closely related taxa were included in our phylogenetic analyses. Vibrissea, Acephala and Phialocephala clustered as a highly supported group (MPBP/NJBP = 100%/100%, Fig. 1), while Chlorovibrissea were distantly related (Fig. 1). The results coincided with those of the previous studies (Wang et al. 2006;Sandoval-Leiva et al. 2014), in which Chlorovibrissea and Vibrissea appeared as two separate clades. As to relationships among species of Chlorovibrissea, C. albofusca and C. chilensis were very closely related (MPBP/NJBP = 100%/100%), and C. korfii is associated with the rest species of the genus, which did not receive a reasonable support (MPBP = 81%). The interspecific relationships of Vibrissea were hardly demonstrated ( Fig. 1) since very few species were sampled.
The two asexual genera are recognizable and associated each other with low supports (Fig. 1). The two existing species of the genus Acephala were sisters ( Fig. 1, MPBP/NJBP = 99%/99%). Acephala was thought to be congeneric with Phialocephala by some authors, and differentiated only by the lack of observed sporulation in culture (Grünig et al. 2009;Münzenberger et al. 2009;Tanney et al. 2016). Phialocephala species are commonly isolated as dark-septate endophytes from coniferous tree roots or from decomposing wood (Menkis et al. 2004), and attributed to Vibrisseaceae mainly based on sequence data (Wang et al. 2006;Kirk et al. 2008). Connections of some Phialocephala species with Mollisia or mollisioid sexual states were reported recently (Grünig et al. 2009;Tanney et al. 2016). It seems that a lot of work needs to be done to establish the generic concepts.
In conclusion, Vibrisseaceae established based on morphology is quite possibly polyphyletic. Sequence data of Leucovibrissea are desirable to get a more comprehensive outline of the family. Phialocephala s.l. is heterogeneous. Its generic concept needs to be clarified. As more sequences of vibrisseaceous fungi are available, the circumscription of the family will become monophyletic. Day MJ, Hall JC, Currah RS (2012) Phialide arrangement and character evolution in the helotialean anamorph genera Cadophora and Phialocephala.

Introduction
Molecular (DNA-based) species identification is the process by which newly generated DNA sequences are examined for taxonomic affiliation and sometimes functional aspects by comparison to reference sequences of firmly established taxonomic origin. It is a powerful tool to identify organisms, particularly those with few or no discriminatory morphological characters and those with cryptic or inconspicuous life styles (Pečnikar and Buzan 2014). Fungi are one such group (Stajich et al. 2009). Molecular exploration of substrates such as soil, water, and even household dust from the built environment has revealed a spectacular diversity of hitherto unrecognized fungal lineages (Grossart et al. 2016, Nilsson et al. 2016, and recent estimates put the number of extant fungal species at upwards of 6 million (Blackwell 2011, Taylor et al. 2014). The number of recognized, validly described species, in contrast, stands at ~135,000 (http://www.speciesfungorum.org, July 2017). Fruiting bodies or other tangible somatic structures are not known for any of the 40 previously unknown fungal lineages examined by Tedersoo et al. (2017), and at present DNA-based methods represent the only way to approach the taxonomic affiliation of these and other lineages. Several factors combine to make molecular identification of fungi complicated. In addition to the lack of reference sequences for more than 99% of the estimated number of extant species of fungi, technical complications such as chimera formation and low read quality may introduce noise and bias to such efforts (Hyde et al. 2013, Kõljalg et al. 2013. To some extent, software tools are available to exercise some degree of control over these complications (e.g., Edgar et al. 2011, Bengtsson-Palme et al. 2013. Furthermore, many -perhaps even most -researchers seem to be aware of the need to approach existing as well as newly generated sequences in a critical way (e.g., Nilsson et al. 2012, Alm Rosenblad et al. 2016, which nevertheless does not appear to prevent substandard entries from being deposited in the databases of the International Nucleotide Sequence Database Collaboration (INSDC: GenBank, ENA, and DDBJ, Schoch et al. 2014, Cochrane et al. 2016. Such substandard INSDC entries may skew research efforts through, e.g., BLAST sequence similarity searches (Altschul et al. 1997) or inclusion in multiple sequence alignments and phylogenetic analyses.
One aspect of sequence reliability that remains largely unexplored is quality trimming of the distal (approximately 25 bases at the very 5' and 3') ends of Sanger sequences. Owing to the nature of the Sanger sequencing process, the very first bases are often hard to resolve due to the presence of un-incorporated nucleotides and leftover primers. Similarly, the signal-to-noise ratio typically drops with the length of the amplicon in that it becomes increasingly difficult to separate amplicons of near-identical lengths from each other on the electrophoresis gel. Thus, an important part of Sanger sequencing is to inspect the resulting chromatograms and remove any noisy distal sequence parts in the newly generated sequence data. This step is, however, sometimes overlooked. When working with INSDC data for fungal molecular identification and sequence analysis purposes, we regularly come across entries whose distal ends appear to be very poorly trimmed. They may feature extended homopolymer regions (e.g., AAAAAAAAA…) or stretches of seemingly random bases that are not found in other conspecific sequences (Nilsson et al. 2012). These potentially noisy sequence ends make it difficult to judge BLAST results: are the mismatches in the distal ends of sequences due to actual biological (nucleotide) differences, or is the reason for the mismatches simply low read quality owing to poor trimming of the reference sequences? There is no direct way of knowing, although clues can perhaps be gleaned from the BLAST alignment and comparison with other conspecific sequences. We fear that, in many cases, researchers will not ponder this question, but will rather assume (or will use automated sequence processing tools that assume) that the mismatches observed are of a biological nature. This will translate into compromised molecular identification, suboptimal assignment of taxonomic affiliations, and unsatisfactory use of sequence data.
The problem is of particular concern for the nuclear ribosomal internal transcribed spacer (ITS) region, the formal fungal barcode and the most popular genetic marker for assessing the taxonomic composition of fungal communities (Schoch et al. 2012, Lindahl et al. 2013). This marker is used by hundreds of studies annually, such that the ramifications of poorly trimmed reference sequences could taint the results of numerous studies each year (cf. Gilks et al. 2002). In an effort to assess the extent of poor sequence trimming in the public sequence repositories, we compared the ITS sequences from 86 fungal (draft) genomes with the public fungal ITS sequences from the same species in the INSDC. We found that in many cases, researchers do not seem to have applied stringent sequence trimming; indeed, in many cases, researchers do not appear to have inspected the chromatograms at all before depositing the sequence data in the INSDC. We conclude by offering a set of observations and recommendations to alleviate the sequence trimming problem in present and future molecular research efforts.

Retrieval of reference ITS sequences from genomes
The ribosomal operon is regularly left out from genome sequencing efforts due to assembly difficulties (Schoch et al. 2014, Hibbett et al. 2016, such that there is no straightforward way to obtain the ITS region from all existing fungal genomes (as has been reported for other genes, Bai et al. (2015)). We therefore used BLAST in the NCBI Whole Genome Shotgun database (https://www.ncbi.nlm.nih.gov/genbank/ wgs/) to identify fully assembled ribosomal regions, using the very conserved 5.8S gene of the ITS region as the BLAST seed. Of the 130 matches returned, 86 were found to represent full-length ITS regions of distinct species that were also represented by at least one reasonably full-length Sanger-derived ITS sequence in the INSDC. In addition to the full ITS region, we kept 50 bases of the upstream nuclear ribosomal small subunit (nSSU/18S) gene and 50 bases of the downstream nuclear ribosomal large subunit (nLSU/28S) gene in the genome-derived ITS sequences to guide the subsequent alignment step.

Retrieval of INSDC sequences
For each of the 86 species (spanning 3 fungal phyla and 29 orders, Suppl. material 1), we downloaded all reasonably full-length ITS sequences from the INSDC using the NCBI query phrase "Species name[ORGN] AND 5.8S[TITL] AND 200:900[SLEN]". We were specifically interested in sequences generated using the traditional ITS1/ITS1F and ITS4/ITS4B primer sets (cf. Tedersoo et al. 2015) since sequences of this coverage are frequently used in DNA barcoding and systematics efforts (Lindahl et al. 2013). Each set of conspecific sequences (the genome-derived sequence plus the conspecific INSDC sequences) was aligned separately in MAFFT 7.307 (Katoh and Standley 2013), and sequences found to contain more than 50 bases of SSU or more than 50 bases of LSU were excluded. Similarly, sequences found to lack more than 50 bases of the 5' end of the ITS1 region, or more than 50 bases of the 3' end of the ITS2 region, were discarded. Alignments were adjusted manually, as needed, following Hyde et al. (2013). Sequences found to be chimeric, taxonomically misidentified, or the subject of other severe technical complications were removed from the alignments prior to statistical analysis. Wherever we found evidence of significant taxonomic variation (e.g., cryptic species) in the alignments, we removed all sequences (alleles) that we deemed to come from a different cryptic species/allele compared to the genome-derived sequence in question. In this study we sought to compare sequence variation in the context of poor sequence trimming rather than in the context of major sequencing artifacts, cryptic species, or allelic divergence.

Multiple sequence alignment and analysis
We went through each position in each of the alignments, starting from the 50th-tolast base of the SSU to the 50th base of LSU, and noted the proportion of INSDC sequences that produced a different nucleotide base from that of the corresponding genome-derived ITS sequence. All three of DNA base mismatches, gaps, and DNA ambiguity symbols (Cornish-Bowden 1985) were counted as mismatches. For each sequence in the alignment, we calculated the dissimilarity (proportion of mismatches) as a function of its relative position. The dissimilarities of the 86 species were then combined using a weighted average with weights proportional to the total number of available sequences for each species. The standard errors were calculated based on the corresponding weighted sample standard deviation. To examine the average age (NCBI date of last modification) of the poorly trimmed sequences, all sequences with at least 5% average dissimilarity among the 5% of its first bases or 5% of its last bases were classified as "potentially poorly trimmed", and their date of NCBI modification was assessed. The association between year and proportion of "potentially poorly trimmed" sequences was examined using overdispersed Poisson rate regression. The relative number of "potentially poorly trimmed" sequences was used as the response variable and time (year) as covariate. All statistical analyses were done in R 3.2.1 (R Core Team 2017).

Multiple sequence alignment
The 86 multiple sequence alignments, each covering at most 50 bases of the SSU, the full ITS region (minus at most 50 bases of the 5' end of ITS1 and/or 50 bases of the end of ITS2), and at most 50 bases of the LSU, are provided in Suppl. material 2. The average length of the alignments was 648 bases (SD: 90, min: 416, max: 941), and the average number of sequences was 123 (SD: 219, min: 1, max: 1586).

Read quality variation
The plotting of disagreements with respect to the genome-derived sequences revealed that insufficient trimming of sequence data seems to be a widespread problem (Figs 1-2). An average of 13.1% of the sequences (SD: 19.6%) in each alignment were classified as "potentially poor trimmed", i.e. they showed at least 5% dissimilarity compared to the corresponding genome sequence over the first or last 5% of the aligned bases. For the remaining (non-distal) bases, those values were down to 0.22% (SD: 0.90%). The dissimilarity was found to be 7.9% and 5.3% in the 5' and 3' ends, respectively ( Fig. 2b-c). The proportion of potentially poorly trimmed sequences was consistently high over the years 1997-2016, with a weak but significantly increasing trend (p=0.0291, Fig. 3).

Discussion
We provide data to suggest that many public DNA sequences are poorly trimmed in their distal parts. The fact that poorly trimmed sequences continue to be deposited through 2016 furthermore suggests that this problem will not go away by itself over time. We hope that the present paper will serve as an eye-opener, both for researchers who risk using the poorly trimmed data for molecular identification and for researchers generating and depositing sequences in public sequence repositories. The way it is now, these sequences may confound sequence similarity searches by falsely suggesting that two sequences (biological entities) are less similar than what really is the case. This reduces the precision in taxonomic and functional assessment -whether manual or carried out through some software pipeline -of newly generated sequences. Other kinds of sequence analysis, such as phylogenetic analyses, will similarly be distorted by poorly trimmed sequences.
Fortunately, managing read quality in Sanger sequences is fairly straightforward. The chromatograms, indicating the relative signal strength for each of the four purines/ pyrimidines C, T, A, and G for each position in the sequence, are a key resource in this pursuit. Brief guidelines for how chromatograms should be processed are available in Hyde et al. (2013) and through various textbooks, online tutorials, and troubleshooting guides (e.g., Kearse et al. 2012, Green andSambrook 2012). Trying to squeeze out Figure 1. Example of poorly trimmed sequences (sequence four and on) from the species Setosphaeria turcica. The 5' end of the alignment is shown, and the poorly trimmed sequences cover the last ~5 bases of SSU and the immediate start of ITS1. The topmost sequence is genome-derived, and sequences two and three are regular Sanger sequences retrieved from the INSDC from other studies than the one with the poorly trimmed sequences (sequences four and on). SeaView v. 4 (Gouy et al. 2010) was used to visualize the alignment. extra information from chromatograms by progressing too far in the 5' or 3' ends is not a good idea, and researchers should make it a habit to crop sequence ends aggressively. Generally speaking, habitually trying to salvage sequences with chromatograms of modest overall quality is not likely to be in the best interest of science. In most cases, it would appear to be better to re-process and re-sequence the material using other DNA extraction protocols, primers, or PCR conditions (cf. Larsson and Jacobsson 2004, Young et al. 2014, Lorenz et al. 2017. Finally, sequence similarity searches using BLAST may be used to get an idea of the technical quality of newly generated sequences (cf. Hyde et al. 2013), including at least cursory inspection of whether the distal ends of sequences are trimmed well enough. BLAST is, however, a somewhat blunt tool when it comes to assessing the read quality of sequence ends and we recommend it as a complement to, rather than as a replacement of, manual inspection of chromatograms. NCBI recently launched a unified system for multiple rRNA submission types, the Submission Portal (https://submit.ncbi.nlm.nih.gov/). This includes an ITS submission wizard specifically tailored to provide various verification steps that should decrease the likelihood of low quality submissions. This includes the use of ITSx (Bengtsson-Palme et al. 2013) to improve annotation, vector screening and automatic trimming, plus/minus mis-assembly checks, and trimming or removal of sequences with a high number of ambiguities. Hopefully, this will raise the awareness on part of sequence authors of the need to screen sequence data for quality issues prior to deposition.
In this study we show that incomplete (or lack of ) trimming of sequence ends remains abundant in molecular mycology. Although this was expected based on our experience, this is the first study to provide at least an initial estimate of the magnitude of the problem. We used genome-derived ITS sequences from 86 fungal species from 29 different orders in our pursuit, such that we think that it is reasonable to extrapolate our findings to the fungal kingdom at large. Furthermore, we cannot think of any reason why this would be a uniquely fungus-specific problem, and we consider that our findings in fact may hold true for Sanger sequences from all genes and groups of organisms, possibly excluding groups and genes that only a few meticulous researchers have worked on. We would, however, like to stress that we provide estimates rather than hard facts. Our approach relied on genome-derived ITS sequences, and we quantified deviations from the genome sequences among conspecific ITS sequences in the INSDC as assessed through species names (Latin binomials). However, some degree of deviation from the genome-derived sequences is expected, since intraspecific ITS variation may reach 3% or in some cases more (Schoch et al. 2012, Garnica et al. 2016. Similarly, the multicopy nature of the ITS region is a potential complication in that we may inadvertently have used a rare and perhaps deviant genome ITS copy and compared it to more common ITS copies (cf. Lindner et al. 2013). That said, such intraspecific or intragenomic variation is not known to be limited to the very start and end of the ITS region or other genetic markers and should not be able to produce the pattern seen in Figs 1 and 2a. In addition, we explicitly sought to avoid comparing sequences across different alleles and cryptic species by excluding sequences that did not match the respective genome-derived sequence closely.
In conclusion, we have shown beyond reasonable doubt that there is room for improvement in the way the mycological community -and to some degree the scientific community at large -trim their DNA sequences. The poor sequence trimming leaves a mark on all subsequent studies that make use of those sequences through BLAST searches or otherwise. Mycology faces enough challenges as it is without having to worry about the burden of poorly trimmed sequences (cf. Pautasso 2013), and we hope that this study will serve as a wake-up call when it comes to trimming of sequence entries in mycology and elsewhere.

Two novel species of Calonectria isolated from soil in a natural forest in China
QianLi Liu 1 , ShuaiFei Chen 1

Introduction
Calonectria species include many notorious plant pathogens and are widely distributed in tropical and subtropical areas of the world (Crous 2002, Lombard et al. 2010d, Aiello et al. 2013, Alfenas et al. 2015. These species can cause serious plant epidemics on a wide range of plant hosts (Peerally 1991, Crous 2002, and result in considerable economic losses to agriculture and forestry. Example include shoot blight on Pinus spp. in South African nurseries (Crous et al. 1991), root rot on Myrtus communis in Tunisia , and leaf blight on Buxus sempervirens in Iran (Mirabolfathy et al. 2013). In addition, members of the genus Calonectria are responsible for red crown rot of Glycine max (soybean) in Japan (Yamamoto et al. 2017), fruit rot of Nephelium lappaceum (rambutan) in Puerto Rico (Serrato-Diaz et al. 2013) and root rot of Arbutus unedo (strawberry) in Italy ). As an important fast-growing tree species, Eucalyptus plays a significant role in the global pulpwood supply. Previous research showed that Calonectria leaf blight (CLB), associated with several species of Calonectria, is considered to be one of the most prominent Eucalyptus leaf diseases that has occurred in numerous countries such as Brazil (Alfenas et al. 2015, Lombard et al. 2016), China (Zhou et al. 2008, Chen et al. 2011), Colombia (Rodas et al. 2005), India (Sharma et al. 1984) and Vietnam (Old et al. 1999).
Other fungal diseases of Eucalyptus spp. caused by Calonectria species include dampingoff, shoot blight, and root rot, which have been observed in Brazil (Ferreira 1989) and South Africa (Crous et al. 1991), and these diseases have received considerable attention.
Calonectria spp. are soil-borne fungi, they can form microsclerotia in soil and infected plant roots, stem and leaves as primary inoculum. After diseased tissues decompose or the plants are harvested, microsclerotia are released into the soil, which allows them to survive for extended periods even up to 15 years or more (Sobers andLittrell 1974, Crous 2002). Species of Calonectria are also rapidly dispersed via aerial dissemination and water movement, which leads to the transmission of Calonectria disease . Based on previous studies, at least 145 Calonectria species have been identified using molecular data and have been described worldwide (Crous 2002, Crous et al. 2004, 2015, Lombard et al. 2010a, b, c, 2011, 2015, 2016, Chen et al. 2011, Xu et al. 2012, Alfenas et al. 2013a, Gehesquière et al. 2015. Sixty species were isolated from soil samples collected in subtropical or tropical regions (Crous 2002, Crous et al. 2004, Lombard et al. 2010a, b, c, 2015, 2016, Chen et al. 2011, Xu et al. 2012, Alfenas et al. 2015.
In China, Calonectria has a relatively high species diversity, and to date, 28 Calonectria species have been identified and described. Based on previous studies, Calonectria species have been reported in nine provinces and one Special Administrative Region (SAR), which with the exception of LiaoNing and ShanDong Provinces belong to temperate regions (Luan et al. 2006, Li et al. 2010). Most Calonectria have been isolated from agronomic crops or forestry plantations in subtropical and tropical regions, including FuJian, GuangDong, GuangXi, GuiZhou, HaiNan, JiangXi and YunNan Provinces, as well as Hong Kong SAR (Crous et al. 2004, Lombard et al. 2010a, 2015, Chen et al. 2011, Gai et al. 2012, Xu et al. 2012, Pei et al. 2015. China has large areas of plantation and natural forests. To date 27 Calonectria species have been isolated from Eucalyptus tissues with CLB/leaf rot symptoms or from soils originating from Eucalyptus plantations in tropical or subtropical areas in Fu-Jian, GuangDong, GuangXi and HaiNan Provinces (Crous et al. 2004, Lombard et al. 2010a, 2015, Chen et al. 2011). However, little information is known about the species diversity of Calonectria in natural forests. In this study, a number of soil samples were collected from a natural forest in the temperate region of central China, and baited with alfalfa seeds for Calonectria. The aim of the current study was to identify these isolates using a combination of phylogenetic analyses and morphological characteristics and to gain a preliminary understanding of the species diversity of Calonectria in natural forests in China.

Fungal isolates
In April 2016, 17 soil samples were collected from a natural forestry area in central China. The collected soils were baited with surface-disinfested (30 s in 75% ethanol and washed several times with sterile water) Medicago sativa (alfalfa) seeds using the method described by Crous (2002). After one week, sporulating conidiophores were produced on infected alfalfa tissue. Using a dissection microscope AxioCam Stemi 2000C (Carl Zeiss, Germany), conidial masses were selected and scattered onto 2 % malt extract agar (MEA) (20 g malt extract powder and 20 g agar powder per liter of water: malt extract powder was obtained from Beijing Shuangxuan microbial culture medium products factory, Beijing, China; the agar powder was obtained from Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) using sterile needles. After incubation at 25 °C for one day, germinated spores were individually transferred onto fresh MEA under the dissection microscope and were incubated at 25 °C for one week.
Single conidial cultures were deposited in the Culture Collection of the China Eucalypt Research Centre (CERC), Chinese Academy of Forestry (CAF), ZhanJiang, GuangDong Province, China. Representative isolates were stored in the China General Microbiological Culture Collection Center (CGMCC), Beijing, China. The specimens (pure fungal cultures) were deposited in the Collection of Central South Forestry Fungi of China (CSFF), GuangDong Province, China.

DNA extraction, PCR and sequence reactions
Single conidial cultures grew on MEA for one week at 25 °C, after which actively growing mycelium was scraped using a sterilized scalpel and transferred into 2 mL Eppendorf tubes. Total genomic DNA was extracted following the protocols "Extraction method 5: grinding and CTAB" described by Van Burik et al. (1998). The extracted DNA was dissolved in 30 μL TE buffer (1 M Tris-HCl and 0.5 M EDTA, pH 8.0), and a Nano-Drop 2000 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to quantify the concentration.
Amplified fragments were sequenced in both directions using the same primer pairs used for amplification by the Beijing Genomics Institute, Guangzhou, China. Sequences were edited using MEGA v. 6.0.5 software (Tamura et al. 2013). All sequences of the isolates obtained in this study were submitted to GenBank (http://www.ncbi. nlm.nih.gov) ( Table 1).

Phylogenetic analyses
The sequences generated from this study were added to other sequences of closely related Calonectria species downloaded from GenBank for phylogenetic analyses. All sequences used in this study were aligned using the online MAFFT v. 7 (http://mafft. cbrc.jp/alignment/server) with the alignment strategy FFT-NS-i (Slow; interactive refinement method). The aligned sequences were manually edited using MEGA v. 6.0.5 and were deposited in TreeBASE (http://treebase.org).
Phylogenetic analyses were conducted on individual tef1, his3, cmdA and tub2 sequence datasets and on the combined datasets for the four gene regions, depending on the sequence availability. Two methods, maximum parsimony (MP) and maximum likelihood (ML) were used for phylogenetic analyses.
MP analyses were performed using PAUP v. 4.0 b10 (Swofford 2003), gaps were treated as a fifth character, and characters were unordered and of equal weight with 1000 random addition replicates. A partition homogeneity test (PHT) was conducted to determine whether data for the four genes could be combined. The most parsimonious trees were acquired using the heuristic search option with stepwise addition, tree bisection, and reconstruction branch swapping. MAXTREES was set to 5,000, and zero-length branches were collapsed. A bootstrap analysis (50% majority rule, 1,000 replicates) was carried out to determine statistical support for internal nodes in trees. | GenBank accession no. indicated in bold were generated in this study ¶ "N/A" represents information that are not available.
The tree length (TL), consistency index (CI), retention index (RI) and homoplasy index (HI) were used to assess phylogenetic trees (Hillis and Huelsenbeck 1992). ML analyses were performed using PHYML v. 3.0 (Guindon and Gascuel 2003), and the best evolutionary model was obtained using JMODELTEST v. 2.1.5 (Posada 2008). In PHYML, the maximum number of retained trees was set to 1,000, and nodal support was determined by non-parametric bootstrapping with 1,000 replicates.
Based on the morphological characteristics, datasets were separated into two groups: the Prolate Group and the Sphaero-Naviculate Group (Lombard et al. 2010b), and therefore phylogenetic analyses were performed with two separate sequence datasets. Calonectria hongkongensis (CBS 114711 and CBS 114828) and Ca. pauciramosa (CMW 5683 and CMW 30823) represented the outgroup taxa for the Prolate Group and Sphaero-Naviculate Group, respectively. The phylogenetic trees were viewed using MEGA v. 6.0.5 for both MP and ML analyses.

Sexual compatibility
Based on multi-gene phylogenetic analyses, isolates of each identified Calonectria species were crossed with each other in all possible combinations. Crosses were performed on minimal salt agar (MSA; Guerber and Correll 2001) on the surface of the medium using three sterile toothpicks. Isolates crossed with themselves were regarded as controls. These crosses were used to determine whether the identified species had a heterothallic or a homothallic mating system. The cultures were incubated at 25 °C for six weeks. When isolate combinations produced extruding viable ascospores, crosses were considered successful.

Morphology
To determine the morphological characteristics of the asexual morphs, representative isolates identified by DNA sequence comparisons were selected. Agar plugs from the periphery of actively growing single conidial cultures were transferred onto synthetic nutrient-poor agar (SNA; Nirenburg 1981) and incubated at 25 °C for one week (there were five replicates per isolate). Asexual structures that emerged on the surface of the SNA medium were mounted in one drop of 80% lactic acid on glass slides and examined under an Axio Imager A1 microscope (Carl Zeiss Ltd., Munchen, Germany) and an AxioCam ERc 5S digital camera with Zeiss Axio Vision Rel. 4.8 software (Carl Zeiss Ltd., Munchen, Germany). Sexual morphs were studied by transferring perithecia obtained from the sexual compatibility tests into a tissue-freezing medium (Leica Biosystems, Nussloch, Germany) and were hand-sectioned using an HM550 Cryostat Microtome (Microm International GmbH, Thermo Fisher Scientific, Wall-dorf, Germany) at -20 °C. The 10-μm sections were mounted in 80% lactic acid and 3% KOH.
Fifty measurements were made for each morphological structure of the isolates selected as the holotype specimen, 30 measurements were made for the isolates selected as the paratype specimen. Minimum, maximum and average (mean) values were determined and presented as follows: (minimum-) (average -standard deviation) -(average + standard deviation) (-maximum).
The optimal growth temperature of the Calonectria species was determined by transferring the representative isolates to fresh 9 mm MEA Petri dishes, which were incubated under temperatures ranging from 5 to 35 °C at 5 °C intervals in the dark (there were five replicates per isolate). Colony colors were determined by inoculating the isolates on fresh MEA at 25 °C in the dark, after seven days incubation, a comparison was performed using the colour charts of Rayner (1970).

Fungal isolates
A total of 40 isolates with the typical morphological of Calonectria species were obtained from the infected alfalfa tissue cultivated in the soil samples. Based on preliminary phylogenetic analysis of the tef1 gene region (data not shown), 16 isolates from all soil samples were selected for further study (Table 1).

Phylogenetic analyses
Sequences for the 78 ex-type and other strains of 48 Calonectria species closely related to isolates obtained in this study were downloaded from GenBank (Table 1). For the 16 isolates collected in this study, nine resided in the Prolate Group, and seven were clustered in the Sphaero-Naviculate Group. Phylogenetic analyses of individual tef1, his3, cmdA and tub2 and the combined sequence datasets were conducted using both MP and ML method. For both the Prolate and Sphaero-Naviculate Groups, although the related position of some Calonectira species were slightly different between the MP and ML trees, the overall topologies were similar, and the ML trees were exhibited.
For the Prolate and Sphaero-Naviculate Groups, the PHT comparing the combined tef1, his3, cmdA and tub2 gene datasets generated P values of 0.141 and 0.333, respectively, which indicated that no significant difference existed between these datasets. These datasets were consequently combined and subjected to phylogenetic analyses. For each of the two groups, the sequence alignments of tef1, his3, cmdA, tub2 and the combination of the four genes were deposited in TreeBASE (TreeBASE No. 21357). The number of parsimony informative characters, the statistical values for the phylogenetic trees of the MP analyses, and the parameters for the best-fit substitution models of ML analyses are shown in Table 2.
Phylogenetic analyses of each of the individual and combined sequence datasets indicated that in the Prolate Group, the nine isolates resided in the Ca. colhounii species complex and were closely related to Ca. colhounii, Ca. eucalypti, Ca. fujianensis, Ca. nymphaeae, Ca. paracolhounii and Ca. pseudocolhounii. In the his3 and cmdA phylogenetic trees, the nine isolates and Ca. fujianensis were clustered in the same clade (Suppl. materials 2, 3), while in the trees based on the tef1 and tub2 sequences, the nine isolates formed an independent clade (Supplementary Figures 1, 4). Based on the phylogenetic analyses of the combined sequences of the four genes, the nine isolates formed a new, strongly defined phylogenetic clade that was distinct from other Calonectria species and was supported by high bootstrap values (ML = 94%, MP = 93%) (Figure 1). Fixed unique single nucleotide polymorphisms (SNPs) were identified in the new phylogenetic clades of the nine isolates and their phylogenetically closed Calonectria species (Table 3). The total number of SNP differences between the new clade and the other closely related species varied between 10-34 for all four gene regions combined (Table 4). The results of these phylogenetic and SNP analyses indicate that the nine isolates in the Prolate Group represent a distinct, undescribed species.
Phylogenetic analyses of each of the individual and combined datasets indicated that in the Sphaero-Naviculate Group, the seven isolates were clustered in the Ca. kyotensis species complex and were closely related to Ca. canadiana. In the tef1 phylogenetic trees, the seven isolates were grouped in the same clade with Ca. canadiana (Suppl. material 5). In the phylogenetic trees based on the his3, cmdA and tub2 sequences, the seven isolates formed an independent clade distinct from Ca. canadiana and other species in the Ca. kyotensis species complex (Suppl. materials 6, 7 and 8). Based on the combined sequences of the four genes, the seven isolates formed a strongly defined phylogenetic clade that was distinct from Ca. canadiana and was supported by high bootstrap values (ML = 100%, MP = 100%) ( Figure 2). The seven isolates obtained in this study were distinguished from Ca. canadiana using SNP analyses for each of the tef1, his3, cmdA and tub2 gene region sequences (Tables 5). The total number of SNP differences between the seven isolates and Ca. canadiana for all four genes was 51 (Table 6). The results indicate that the seven isolates in the Sphaero-Navivulate Group represent a novel species.

Sexual compatibility
After a six-week mating test on MSA, all 16 isolates and the crosses of isolates of each identified species failed to yield sexual structures, indicating that they were either selfsterile (heterothallic) or had retained the ability to recombine to produce fertile progeny.   Table 3.

Species
Isolate no.

MycoBank MB821349
Diagnosis. Calonectria montana can be distinguished from the phylogenetically closely related species Ca. canadiana by the size of macroconidia.
Substratum. Soil under the natural forest. Notes. Calonectria montana is a new addition to the Ca. kyotensis complex and is phylogenetically closely related to Ca. canadiana (Crous 2002, Crous et al. 2004, 2016. The macroconidia of Ca. montana (av. 43.2 × 4.6 μm) are shorter and wider than those of Ca. canadiana (av. 50 × 4 μm).

Discussion
This study identified two novel species of Calonectria from soil in a natural forest in the temperate region of central China. The identification of the fungi was supported by DNA sequence comparisons and morphological features. The two species were named Calonectria lichi and Ca. montana.
Calonectria lichi is a new addition to the Ca. colhounii complex that belongs to the Prolate Group. Based on phylogenetic analyses of four gene sequences, Ca. lichi formed a distinct and well-supported phylogenetic clade closely related to Ca. fujianensis, Ca. nymphaeae and Ca. paracolhounii, but it can be distinguished from these species by its larger macroconidia. To date, 10 species in the Ca. colhounii complex have been identified and described. Other than Ca. lichi described in this study, the other species include Ca. colhounii, Ca. eucalypti, Ca. fujianensis, Ca. macroconidialis, Ca. monticola, Ca. nymphaeae, Ca. paracolhounii, Ca. parva and Ca. pseudocolhounii (Crous 2002, Lombard et al. 2010b, 2016, Chen et al. 2011, Xu et al. 2012. Of these species, Ca. colhounii, Ca. eucalypti, Ca. fujianensis, Ca. nymphaeae and Ca. pseudocolhounii have been shown to be homothallic and always produce bright yellow perithecia (Crous 2002, Lombard et al. 2010b, Chen et al. 2011, Xu et al. 2012. In China, four species in the Ca. colhounii complex have been reported: except for Ca. lichi, which was isolated from a natural forest in the temperate zone in central China, the other species, including Ca. fujianensis, Ca. pseudocolhounii and Ca. nym-phaeae, were previously isolated from tropical or subtropical regions in southern China (Chen et al. 2011, Xu et al. 2012.
Calonectria montana adds a new species to the Ca. kyotensis complex that belongs to the Sphaero-Naviculate Group. Phylogenetic analyses showed that Ca. montana, which formed an independent clade with a high bootstrap value, is closely related to Ca. canadiana. Morphological differences were observed between Ca. montana and Ca. canadiana, especially with respect to the size of the macroconidia and the shape of the vesicles (Kang et al. 2001, Crous 2002. Species in the Ca. kyotensis complex are characterized by having sphaeropedunculate vesicles with lateral stipe extensions on a conidiogenous apparatus (Crous et al. 2004, Lombard et al. 2010b, 2015, 2016. No lateral stipe extensions were produced by Ca. montana, indicating that this species is different from other species in the Ca. kyotensis complex. In this study, Ca. montana was isolated from soil in central China, 14 species residing in the Ca. kyotensis complex were previously reported in China, and all of them were isolated from soil in southern China (Crous et al. 2004. The results from this study suggest that more species in Ca. kyotensis complex have yet to be discovered from China.
Species of Calonectria are important plant pathogens that can cause devastating diseases on various plant hosts worldwide, especially on horticultural, agronomic and forestry crops (Polizzi et al. 2001, Crous 2002, Saracchi et al. 2008, Chen et al. 2011. In China, Calonectria species have been reported as pathogens of various important agronomic and forestry crops. In agriculture, the Fabaceae and Arecaceae plant families are susceptible to infection by Calonectria species, including Ca. ilicicola, which causes black rot (CBR) of Arachis hypogaea (peanut) and Medicago sativa (Gai et al. 2012, Pei et al. 2015, Ca. ilicicola causes red crown rot of Glycine max (soybean) (Guan et al. 2010), and Ca. colhounii and Ca. pteridis cause leaf spot on Phoenix canariensis and Serenoa repens, respectively (Luo et al. 2009, Yang et al. 2014. In forestry, leaf blight caused by Calonectria species is considered as one of the most serious threats to Eucalyptus plantations and nurseries in southern China (Zhou et al. 2008, Lombard et al. 2010a, Chen et al. 2011. The leaf inoculations showed that all tested Calonectria species were pathogenic to the tested Eucalyptus clones, including the clones that are widely planted in southern China (Chen et al. 2011, Li et al. 2014a. These research results suggest that species of Calonectria need to be monitored carefully, both in agronomic crops and forests. Accurate diagnosis of plant diseases and identification of their casual agents provide the foundation for developing effective disease management strategies (Booth et al. 2000, Crous 2002, Old et al. 2003. Based on previous research results, the majority of Calonectria species identified and described in China were isolated from diseased plant tissues or soil under forestry plantations in subtropical and tropical regions (Crous et al. 2004, Lombard et al. 2010a, 2015, Chen et al. 2011. In this study, two novel Calonectria species were described, and they were isolated from soil in a natural forest in the temperate zone. The results from this study suggest that more extensive surveys need to be conducted to collect Calonectria in more geographic regions with different climate zones, which will help to clarify the species diversity of Calonectria in China. shown above branches, with bootstrap values below 60 % marked with an *, and absent analysis values are marked with -. Isolates representing ex-type material are marked with "T", isolates highlighted in bold were sequenced in this study and novel species were covered in blue. The tree was rooted to Ca. hongkongensis (CBS 114711 and CBS 114828 living within healthy plant roots, stems and leaves (Mukherjee et al. 2013). Species of the genus belong to one of the most useful groups of microbes to have had an immense impact on human welfare. Some species are widely used as effective biocontrol agents for several soil-borne plant pathogens (Harman et al. 2004, Hasan et al. 2012, Liu et al. 2012, producers of enzymes, antibiotics and heterologous proteins for food, feed, textile and biofuel industries (cellulases, hemicellulases) (Samuels 1996, Almeida et al. 2007, Cheng et al. 2012, Lopes et al. 2012, Mukherjee et al. 2013. Many members are treated as agents for improving seed germination and nutrient use efficiency, breaking of seed dormancy, as well as source of transgenes and herbicides, and are long known to improve plant growth through the production of phytohormones and certain secondary metabolites (Harman et al. 2004, Shoresh et al. 2010, whereas others are causal agents of opportunistic infections of humans and animals (Samuels 1996, Kuhls et al. 1999, Kredics et al. 2003, and due to association of certain species with economically significant production losses in commercial mushroom farms (Samuels et al. 2002, Park et al. 2006, Kim et al. 2012a, 2012b. The genus Trichoderma was established in 1794 including four species (Samuels 1996). In recent years, the number of Trichoderma species increases dramatically. Bissett et al. (2015) presented a list of 254 names of species and two names of varieties in Trichoderma with name or names against which they are to be protected, following the ICN (Melbourne Code, Art. 14.13). More recently, In a large-scale survey of Trichoderma from rotten wood and soil in China, Zhuang (2016a, b, c, d, e, 2017) published 27 new species based on the collections producing ascomata or asexual morphs on woody substrates; Chen and Zhuang (2016)  During our investigation of the diversity of Trichoderma species in China, two species were found to represent undescribed new taxa, on the basis of both morphological and cultural characters and DNA sequence analyses of partial nuc translation elongation factor 1-α encoding gene (TEF1-α) and the gene for nuc RNA polymerase II second largest subunit (RPB2). Differences between the new species and their close relatives are discussed, and a phylogenetic analysis is provided.

Specimens and strains
Specimens were collected from Henan and Fujian provinces, China, and deposited in the Mycological Herbarium of Jilin Agricultural University (HMJAU). Strains were obtained either by single ascospore isolation from fresh stromata of sexual morphs or by direct isolation from asexual morphs on the substrates. Cultures are deposited in the China General Microbiological Culture Collection Center (CGMCC).

Morphological study
Dried stromata were rehydrated and longitudinal sections through ascomata were made with a freezing microtome (Leica CM1950) at a thickness of 5-10 μm. Agar media employed were cornmeal dextrose agar (CMD, Difco, Sparks, MD, USA, with dextrose 20 g/L), potato dextrose agar (PDA, Solarbio, Beijing, CHINA) and synthetic low nutrient agar (SNA, Nirenberg 1976, pH adjusted to 5.5). Colonies were incubated in 9 cm diam Petri dishes at 25 °C with alternating light/darkness (12/12 h) at 20 °C, 25 °C, 30 °C and 35 °C and measured daily until the dishes were covered with mycelium. The characteristics of asexual and sexual states were described following the methods of Jaklitsch (2009) and Zhu and Zhuang (2015). Photographs were taken using a Leica DFC450C digital camera (Tokyo, Japan) connected to a Zeiss Axioskop 2 Plus microscope (Göttingen, Germany) for anatomical structures and to a Zeiss Stemi 2000C stereomicroscope for gross morphology.

DNA extraction, amplification and sequencing
Genomic DNA was extracted from mycelium harvested from colonies on PDA after 1-2 wk with a NuClean Plant Genomic DNA Extraction Kit (CoWin Biosciences, Beijing, China) according to the manufacturer's protocol. Fragments of the nuc rDNA internal transcribed spacers (ITS1-5.8S-ITS2 = ITS), TEF1-α and RPB2 were amplified with the primer pairs ITS4 and ITS5 (White et al. 1990), EF1-728F (Carbone and Kohn 1999) and TEF1LLErev (Jaklitsch et al. 2005), fRPB2-5f and fRPB2-7cr (Liu et al. 1999), respectively. PCR products were purified with the PCR Product Purification Kit (Biocolor BioScience and Technology Co., Shanghai, China) and cyclesequenced on an ABI 3730 XL DNA Sequencer (Applied Biosciences, Foster City, Calofornia) with the same primer in fragments amplification for ITS and primers reported by Jaklitsch (2009) for TEF1-α, and RPB2 at Beijing Tianyihuiyuan Bioscience and Technology, China. The strains and the NCBI GenBank accession numbers of DNA sequences used in this work are listed in Table 1.

Phylogenetic analyses
Sequences were assembled, aligned and manually adjusted when needed with BioEdit 7.0.5.3 (Hall 1999). NEXUS files were generated with Clustal X 1.83 (Thompson et al. 1997). To identify the phylogenetic positions of Trichoderma fujianense and T. zonatum, RPB2 and TEF1-α sequences were combined for the analyses. Thirty-three sequences representing 30 Trichoderma taxa were selected for analyses, with Nectria eustromatica and N. berolinensis selected as outgroup taxa. Alignments are deposited in TreeBASE accession number 21272.
Maximum parsimony (MP) analysis was performed with PAUP 4.0b10 (Swofford 2002) using 1000 replicates of heuristic search with random addition of sequences  (Nylander 2004). gtr+i+g was estimated as the best-fit model for combined sequences. Four MCMC chains were run from random trees for 2 000 000 generations and sampled every 100 generations. The first 5000 trees were discarded as the burn-in phase of the analyses, and Bayesian inference posterior probability (BIPP) was determined from the remaining trees. Trees were visualized in TreeView 1.6.6 (Page 1996).

Phylogenetic analyses
The partition homogeneity test (P = 0.01) of RPB2 and TEF1-α sequences indicated that the individual partitions were generally congruent (Cunningham 1997). Phylogenetic positions of the new species were determined by analyses of the combined RPB2 and TEF1-α dataset containing 33 taxa and 2396 characters, of which 1304 characters were constant, 366 variable characters were parsimony-uninformative and 726 were parsimony-informative. Five most-parsimonious trees with the same topology were generated (Figure 1) (tree length = 3178, CI = 0.4685, HI = 0.4572, RI = 0.5493 and RC = 0.2982).
Thirty-three sequences representing 30 green-spored Trichoderma species and two outgroup taxa Nectria berolinensis and N. eustromatica were used to construct the phylogenetic tree (Figure 1). All the green-spored species formed a monophyletic group (100 % MPBP/BIPP), which is basically consistent with the previous study by Jaklitsch (2009) and Zhu and Zhuang (2015).
Colony radius on PDA after 72 h 7.5-8.5 mm at 20 °C, 8.5-10 mm at 25 °C, 0.5-1 mm at 30 °C, no growth at 35 °C, mycelium covering the plate after 2 wk at 25 °C. Colony circular, compact with distinctly zonate, with commonly lobed or coarsely wavy margin, centre dense, green, margin relatively looser, whitish. Conidiation noted around the plug after 3-4 d, effuse, spreading from the centre over the entire colony surface. No distinct odor, no diffusing pigment observed.
Habitat and distribution. On the surface of rotten wood in humid forests of east China.
Etymology. The epithet "fujian", indicating occurrence of the fungus in Fujian province.
Remarks. Morphologically, the new species is most similar to Trichoderma costaricense in conidiophore character and phialide shape and size; while the latter fungus produces abundant chlamydospores on CMD, has relatively larger conidia (5.2-6.0 × 3.2-4.0 μm) and faster growth on PDA and SNA, and grows well and sporulates at 35 °C (Chaverri and Samuels 2003). Furthermore, sequence similarity of ITS and RPB2 between these species was only 90.1% and 92.1%, with 60 bp and 68 bp differences among 606 bp and 864 bp, respectively. Among the species with green ascospores, T. gelatinosum, T. nigrovirens, T. chromospermum and T. thelephoricola also generated gliocladium to verticillium-like conidiophores, but they are not phylogenetically closely related.
On CMD colony radius after 72 h 30-43 mm at 20 °C, 32-46 mm at 25 °C, 17-34 mm at 30 °C, no growth at 35 °C. Colony hyaline, circular, loose, forming obvious zonate, covering the plate after 5-7 d at 25 °C. Aerial hyphae radially arranged. Conidiation at 25 °C noted after 3 d, first effuse, soon followed by formation of granules or pustules, particularly along the margin, spreading from the centre across the entire plate. No distinct odor, no diffusing pigment observed.
On PDA after 72 h 38-48 mm at 20 °C, 55-62 mm at 25 °C, 28-30 mm at 30 °C, no growth at 35 °C; mycelium covering the plate after 8 d at 25°C. Colony circular, conspicuously dense, becoming zonate with broad, slightly downy zones and narrow, well-defined, convex, white to green farinose zones. Aerial hyphae numerous, mostly short, becoming fertile from the centre. Conidiation at 25 °C starting after 2 d, green after 4 d, first simple, mostly on short aerial hyphae concentrated in the centre and in denser zones, later abundant in pustules. Autolytic activity lacking or inconspicuous, no coilings seen. No diffusing pigment, no distinct odour noted.
Habitat and distribution. On the surface of rotten wood in humid forests of south central and east China.
Species of the Chlorosporum clade usually produce pale yellow or pale green, semi-translucent stromata, globose to subglobose ascospores and gliocladium-like or verticillium-like conidiophores Samuels 2003, Zhu andZhuang 2015). Trichoderma zonatum is characterized by pulvinate, pale yellow to light brown stromata with densely disposed dark green to black ostioles, monomorphic ascospores, simple trichoderma-like conidiophores, green, (sub)globose to pyriform conidia. Morphologically, stromata of T. zonatum are not typical of the Chlorosporum clade and differ from all other species by relatively larger and non-transparent. It is most similar to T. chromospermum in gross stromata morphology, while the latter fungus clearly differs by much shorter asci [(78-)85-90(-102) μm], gliocladium-like conidiophores and ellipsoidal to cylindrical conidia Samuels 2003, Zhu andZhuang 2015).

Discussion
Phylogenetic analyses of Trichoderma species with green spores based on sequences of RPB2 and TEF1-α were performed by Chaverri and Samuels (2003). In the more recent study by Zhu and Zhuang (2015) a phylogenetic tree with 45 species having green-spored was inferred from RPB2 and TEF1-α sequences. In our study analyses of the combined sequences of the same genes of 30 related Trichoderma species were carried out to ascertain the phylogenetic positions of our new species. The tree topology is basically consistent with previous researches (Chaverri and Samuels 2003, Jaklitsch 2009, Zhu and Zhuang 2015. The study of Chaverri and Samuels (2003) suggested that phenotypic characters, alone are usually not useful in understanding phylogenetic relationships in Trichoderma, because teleomorph characters, for example, the tissue structure of the stroma, the size and character of the perithecia, asci and ascospores, are generally highly conserved and anamorph characters tend to be morphologically divergent within monophyletic groups, clades or species complex. Based on the results of the present study, we conclude that similarity in teleomorphic characters is not indicative of close phylogenetic relationships, holomorphs must be studied in order to effectively determine both life cycles and species concepts.

A dynamic, web-based resource to identify rust fungi (Pucciniales) in southern Africa
Introduction are few reference sequences of rust fungi publically available. For example, of the 4,000 described species of Puccinia, approximately 200 (0.05%) of these have either an ITS or LSU sequence on GenBank (Marin-Felix et al. 2017). Molecular identification from a barcode marker is more common for well-studied species of rust fungi. A web-based resource to identify rust fungi by host and morphology in southern Africa is introduced in the present study. The resource is based around a Lucid key, freely available to all users. The key is dynamic, and can be updated according to taxonomic changes or the discovery of new taxa. The scientific community is invited to contribute specimens and images to the development of this key.

Taxon selection and identification
The first 50 species of rust that accompany the release of this key were collected from the Gauteng, KwaZulu-Natal, Limpopo and Mpumalanga provinces in South Africa, and from Botswana and Swaziland. Specimens were usually collected during field surveys of forestry plantations as well as in adjacent native or farmed vegetation. These 50 species are commonly encountered or important pathogens of trees in natural and planted forests, including Austropuccinia psidii, Phakopsora myrtacearum, Ravenelia macowaniana, Uromyces aloës and Uromycladium acaciae. Specimens were identified on the basis of their host and morphology of spores. In some cases molecular barcodes were used for identification. This is described below for the identification of Puccinia porri, and was published for identifications made in prior studies McTaggart et al. 2015b;Roux et al. 2013).

Morphology and image capture
Spore stages, such as aeciospores, urediniospores and teliospores, were removed from host material with a scalpel, then mounted in clear lactic acid (100% v/v) on a microscope slide and gently heated. Slides were examined with a Leica DM 2500 compound microscope using differential interference microscopy and images were taken with a Leica DFC550 camera. Measurements of each examined spore stage were made from a minimum of 20 spores per specimen.
The approach to stacking multiple images follows that of Shivas et al. (2014). Composite images were made with image stacking software Helicon Focus (Helicon Soft, Kharkov). For example, teliospores shown on the website are montaged from two to four images taken through different focal planes. Images of spore stages with ornamented walls were captured in two focal planes, one through the equator of the spores, and the other through the upper surface of the wall. The roll-over Java Script used by Shivas et al. (2014), to simulate focusing through a microscope, was incorporated for spore stages with ornamented walls.
Host symptoms were photographed with hand-held digital cameras, for example a Coolpix Nikon S9300. Host symptoms of fresh leaf material were scanned on an Epson Perfection V700 flatbed scanner with a minimum resolution of 300 dpi. Images that were finally used for the website were selected based on their quality and diagnostic potential.

Key development
An interactive key, the Rust Fungi of Southern Africa, was built using Lucid 3.5.32 (http://www.lucidcentral.org). The dataset used for rust fungi had 93 features and 320 character states, which included the morphological features of all spore stages present on the examined specimens.

Results
The key has been made publicly available at the following URL: http://collections.daf. qld.gov.au/web/key/africarust.
There are 50 taxa uploaded to the website. Two of these are species of Ravenelia that may represent new taxa. There are 18 genera on the website, of which Aecidium and Uredo are anamorphic genera used for species with unknown telial stages that have uncertain phylogenetic positions. The website contains 190 images, of which 38 are field shots, 48 are scanned host symptoms and 104 are spore stages taken from a light microscope.
A comprehensive list of rust fungi reported in southern Africa since Doidge (1950) is included in the 'references and records' page of the website. The list includes references that have described or reported new taxa, and changed taxonomic names of rust fungi in southern Africa.
One taxon included in the Rust Fungi of Southern Africa is a new addition for the region. Puccinia porri, which was taxonomically resolved by McTaggart et al. (2016a), was found on Allium porrum in South Africa. This identification was confirmed by an ITS-LSU sequence that had 100% identity over 1646 characters to specimens on GenBank identified as P. porri by McTaggart et al. (2016a). This sequence has been deposited in GenBank (KY849820) and the specimen can be viewed on the Rust Fungi of Southern Africa (collections.daff.qld.gov.au/web/key/africarust/Media/Html/pucciniaporri.html). Rust fungi on species of Allium in South Africa were previously identified as P. allii (Doidge 1950), which is a species complex.

Discussion
Identification of rust fungi is challenging for a number of reasons, including their complex lifecycles, multiple species on one host, multiple hosts and the fact that there are few contemporary resources with information about their biology and morphology. Furthermore, identification based on a molecular barcode is not always possible, as many species have not been sequenced. The Rust Fungi of Southern Africa is a webbased, interactive resource that allows users to identify taxa based on host range and morphology. The identification is supported by comparison to images of symptoms and spore stages made from reference specimens. It further acts as a real-time list of rust fungi in southern Africa.  recorded about 546 species of rust fungi in Botswana, Namibia and South Africa. The literature indicates that there are 572 species of rust fungi in southern Africa, which we have listed in the Rust Fungi of Southern Africa. Many of these species will certainly represent the same organism, for example, independently described aecial or uredinial stages of teleomorphic species ). There are 90 species of Aecidium and 53 species of Uredo in the list of taxa, and these will likely belong to other genera such as Puccinia (discussed by McTaggart and Shivas in Marin-Felix et al. 2017).
Further diversity may be expected from cryptic species, which have been found in multiple genera of rust fungi on hosts in the Annonaceae (Beenken 2014), Fabaceae (Doungsa-ard et al. 2015;McTaggart et al. 2015a) and Poaceae (Demers et al. 2017;Liu and Hambleton 2013). Doidge (1950) recorded one species of rust, Uromyces aloës, on 18 different host species, and this may represent a taxon with cryptic diversity.
Two rust fungi were recently described in southern Africa from agricultural and forestry hosts, namely Macruropyxis fulva on Saccharum and Phakopsora myrtacearum on Eucalyptus Martin et al. 2017). It is interesting that two new rusts were found on introduced, well-studied plants in southern Africa. Host jumps were found to be one of the main drivers of speciation for rust fungi (McTaggart et al. 2016b), and host shifts or jumps from native plant species in South Africa to introduced species may explain the observed new taxa on exotic, well-studied hosts.
The Rust Fungi of Southern Africa is the second publicly released Lucid key to identify rust fungi. The Rust Fungi of Australia (available at: http://collections.daff.qld.gov. au/web/key/rustfungi) currently contains 122 species (Shivas et al. 2014). The broader scientific and non-scientific communities are invited to contribute images and specimens to the authors and help build these resources. Submissions for the resource will be acknowledged as a contribution on the home page of the website.