*Article* **Lack of Phylogenetic Differences in Ectomycorrhizal Fungi among Distinct Mediterranean Pine Forest Habitats**

**Irene Adamo 1,2,\*, Carles Castaño <sup>3</sup> , José Antonio Bonet 1,2, Carlos Colinas 2,4, Juan Martínez de Aragón 1,4 and Josu G. Alday 1,2**


**Abstract:** Understanding whether the occurrences of ectomycorrhizal species in a given tree host are phylogenetically determined can help in assessing different conservational needs for each fungal species. In this study, we characterized ectomycorrhizal phylogenetic composition and phylogenetic structure in 42 plots with five different Mediterranean pine forests: i.e., pure forests dominated by *P. nigra*, *P. halepensis*, and *P. sylvestris*, and mixed forests of *P. nigra-P. halepensis* and *P. nigra-P. sylvestris*, and tested whether the phylogenetic structure of ectomycorrhizal communities differs among these. We found that ectomycorrhizal communities were not different among pine tree hosts neither in phylogenetic composition nor in structure and phylogenetic diversity. Moreover, we detected a weak abiotic filtering effect (4%), with pH being the only significant variable influencing the phylogenetic ectomycorrhizal community, while the phylogenetic structure was slightly influenced by the shared effect of stand structure, soil, and geographic distance. However, the phylogenetic community similarity increased at lower pH values, supporting that fewer, closely related species were found at lower pH values. Also, no phylogenetic signal was detected among exploration types, although short and contact were the most abundant types in these forest ecosystems. Our results demonstrate that pH but not tree host, acts as a strong abiotic filter on ectomycorrhizal phylogenetic communities in Mediterranean pine forests at a local scale. Finally, our study shed light on dominant ectomycorrhizal foraging strategies in drought-prone ecosystems such as Mediterranean forests.

**Keywords:** DNA metabarcoding; phylogenetic structure; habitat filtering

#### **1. Introduction**

Ectomycorrhizal fungi are essential organisms in forests, as they form symbiotic relations with trees providing them nutrients in exchange for photosynthetic carbon [1–3]. Some ectomycorrhizal fungi are host specific [3–6] and are influenced by tree species as well as by soil abiotic factors such as pH and nutrient availability [7–10]. Therefore, host effect and abiotic soil parameters are often fundamental drivers of ectomycorrhizal community assembly [11–16]. Moreover, previous studies showed that ectomycorrhizal taxonomic community composition does not significantly change between Mediterranean congeneric pine species [17]. Nevertheless, how ectomycorrhizal fungi are phylogenetically structured among Mediterranean pine host species and whether at both taxonomic and phylogenetic level respond to similar abiotic factors has not been assessed yet. Previous studies showed that ectomycorrhizal responses to climate warming are modulated by host plant performance and nutrient availability [18–20]. Therefore, it is crucial to disentangle

**Citation:** Adamo, I.; Castaño, C.; Bonet, J.A.; Colinas, C.; Martínez de Aragón, J.; Alday, J.G. Lack of Phylogenetic Differences in Ectomycorrhizal Fungi among Distinct Mediterranean Pine Forest Habitats. *J. Fungi* **2021**, *7*, 793. https://doi.org/10.3390/jof7100793

Academic Editors: Anush Kosakyan, Rodica Catana and Alona Biketova

Received: 5 July 2021 Accepted: 17 September 2021 Published: 24 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

whether these drivers influence ectomycorrhizal phylogenetic composition and structure, to better understand forest ecosystem functioning [21].

Phylogenetic analyses are useful tools to estimate the relative importance of evolutionary and ecological forces structuring communities [12,22,23]. In this regard, phylogenetic indices have been implemented to calculate the phylogenetic relatedness of an observed community and compare the value to expectations of community assembly under neutral processes from a regional species pool [24]. Therefore, these indices enable us to characterize whether communities are more phylogenetically related (phylogenetic clustering) or less phylogenetically related (phylogenetic overdispersion) than expected by chance [22–24]. In general, habitat filtering is the dominant assembly process when closely related species that share similar traits are selected to coexist within the community (i.e., phylogenetic clustering). In contrast, competition processes occur when distantly related species with dissimilar traits are selected to co-occur within a community (i.e., phylogenetic overdispersion [22], while the random phylogenetic structure is detected when none of the above processes are inferred [22–25]. For example [26], observed phylogenetic clustering of Agaricomycotina communities (including mycorrhizal and saprotrophs) and observed that xeric oak-dominated forests acted as a filter for these communities. Likewise [27], found phylogenetically clustered arbuscular mycorrhizal communities along an altitudinal gradient and observed that environment was the primary ecological factor structuring these communities, either via changes in host plant or fungal niches. Although the ecological processes filtering communities have recently received criticism [28], investigating the communities' phylogenetic responses to the environment in different ecosystems is fundamental to understand the mechanisms that structure communities [29,30]. However, how ectomycorrhizal communities are phylogenetic structured in Mediterranean pine forests has not been studied yet.

The description of phylogenetic relations between ectomycorrhizal fungi might help to understand the evolutionary ecology of traits, species, and entire communities [31]. In this regard, exploration types of ectomycorrhizal fungi represent an important group of functional traits, which are defined according to the hyphal morphology, i.e., long distance, medium distance, medium distance fringe, short distance, or contact exploration types [32]. The hyphal morphology determines access to distinct nutrient sources, for example, nitrogen (N) [33–35]. Ectomycorrhizal species with short, contact, and medium smooth distance exploration types may preferentially use soluble inorganic forms of N close to the host roots due to the lack the enzymes to access organic N forms [33,36]. Conversely, some fungi have enzymes (i.e., fenton peroxidase) to access insoluble N substrates such as organic substrates and they usually show medium mat and long-distance exploration types [35–37]. However, long and medium fringe exploration types might demand higher carbon cost on the host than shorter distance exploration types [18,32], therefore species with shorter exploration types may be favored under stressful conditions [18,38]. In this regard, several studies have addressed ectomycorrhizal exploration types' responses to environmental drivers [39–42], however, the phylogenetic pattern of the trait in Mediterranean ecosystems has rarely been assessed. Thus, understanding the phylogenetic relationships between ectomycorrhizal species and the evolution of hyphal morphologies in the current climate change context might shed light on the future impacts on Mediterranean ecosystem functioning.

In this study, we aim to characterize the ectomycorrhizal phylogenetic composition and phylogenetic structure in 42 plots of five different Mediterranean pine forests: i.e., pure forests dominated by *P. nigra*, *P. halepensis*, and *P. sylvestris*, and mixed forests of *P. nigra-P. halepensis* and *P. nigra-P. sylvestris*. In line with the above premises, we hypothesized that:


phylogenetic composition [48]. Finally, among abiotic filters pH, P, and CN ratio strongly influenced ectomycorrhizal taxonomic community composition [17]. Here, we tested if these filters would act similarly over ectomycorrhizal phylogenetic composition.

• In Mediterranean ecosystems, soil N might not be limited due to warmer temperatures which may enhance N mineralization by increasing decomposition of the organic matter [17,49]. Therefore, short exploration types could uptake nutrients close to the host roots. Here, we expected that short and contact exploration types will be dominant, thus, both traits will be overrepresented and dispersed across the ectomycorrhizal phylogenetic tree in comparison with medium and long-term exploration types.

#### **2. Materials and Methods**

#### *2.1. Sites Selection*

The study was conducted in the mountainous pre-Pyrenees region of Catalonia, North-eastern Spain (Figure S1) in a set of long-term monitoring plots in which fungal fruiting has been recorded for ~20 years [50]. The region is under the influence of the Mediterranean climate, with a summer drought period from June to August and mean annual temperatures from 6 to 9 ◦C with most of the precipitation occurring in spring and autumn [51]. The 42 pine forests were randomly selected from the 579 sites included in the Forest Ecological Inventory of Catalonia carried out by Centre de Recerca Ecològica i Aplicacions Forestals (CREAF, Barcelona, Spain, 1992), trying to preserve even-aged forest. From the total 42 forests, 32 correspond to pure pine forests, with 14 plots corresponding to *P. nigra* and *P. sylvestris* species and 4 to *P. halepensis*, whilst 10 plots were mixed plots (7 mixed plots of *P. sylvestris* and *P. nigra* species and 3 plots dominated by *P. nigra* and *P. halepensis*). The main features of the study plots are summarized in Table 1 and Table S1.

**Table 1.** Table summarizing the main features of the study plots: BA (Basal area), Number of trees per hectare, Altitude, Slope, pH, CN ratio and P (Phosphorus). Ps: *P. sylvestris*, Pn: *P. nigra*, Ph: *P. halepensis*, Ps-Pn: *P. sylvestris-nigra*, Pn-Ph: *P. nigra-halepensis*.


#### *2.2. Soil Sampling*

Soils were sampled during the Autumn season (October and November) in 2009. In each of the selected forest stands, a 10 m × 10 m plot was established in the center for

long-term monitoring of fungal fruiting [51]. In each plot, we took four soil subsamples, i.e., one per plot-side [52], with a rectangular steel drill (a 30 cm depth and a 6 × 4.5 cm width). The four soil subsamples were pooled in the field and around 1 kg of the mixed sample was placed on ice and taken to the laboratory for fungal DNA extraction. A similar procedure was followed for soil samples to determine soil physico-chemical parameters.

#### *2.3. Soil Analysis*

Soil samples were analyzed using the methodology described in [53]. Each sample was air-dried, sieved (≤2 mm mesh), and soil texture (clay, sand, and lime proportions) was analyzed using the Bouyoucos—method [54]. Soil pH and electrical conductivity (EC) using a conductivity meter in a 1:2.5 soil:deionized water slurry [55]. Total nitrogen concentration using the Kjeldahl method [56]. Moreover, available phosphorus concentration using the Olsen method [57]; total organic matter and total carbon concentration using the Walkley-Black method [58]. Finally, exchangeable cations such as sodium (Na), potassium (K+) and magnesium (Mg2+) with atomic absorption spectroscopy after extraction with 1 N ammonium acetate (pH 7; [55–59]).

#### *2.4. Fungal Community and Bioinformatic Analysis*

Fungal DNA was extracted from 0.5 g of homogenized soil using the NucleoSpin® NSP soil kit (Macherey-Nagel, Duren, Germany) following the manufacturer's protocol. Fungal internal transcribed spacer 2 (ITS2) region was amplified in a 2720 Thermal Cycler (Life Technologies, Carlsbad, CA, USA) using the primers gITS7 [60], ITS4, and ITS4A [61,62]. We optimized the number of PCR cycles in each sample aiming for weak to medium PCR bands at the agarose gels, which was achieved in most of the samples by using 21–26 cycles. The final concentrations in the PCR reactions, PCR conditions, DNA purification and sequencing, and bioinformatics analyses were as explained by Adamo et al. (2021). Sequence data are archived at NCBI's Sequence Read Archive under accession number PRJNA641823 (www.ncbi.nlm.nih.gov/sra, accessed on 25 June 2020).

#### *2.5. Taxonomic and Functional Identification*

We taxonomically identified the 600 most abundant OTUs, which represented 93% of the total sequences. We selected the most abundant sequence from each OTU for taxonomic identification, using PROTAX software [63] implemented in PlutoF, using a 50% probability of correct classification (called "plausible identifications") [63]. These identifications were confirmed and some of them improved using massBLASTer in PlutoF against the UNITE [64]. Taxonomic identities at species level were assigned based on >98.5% similarity with database references, or to other lower levels using the next criteria: genus on >97%, family on >95%, order on >92%, and phylum on >90% similarity. OTUs were assigned to the following functional guilds: (a) root-associated basidiomycetes, (b) rootassociated ascomycetes, (c) molds, (d) yeasts, (e) litter-associated basidiomycetes, (f) litterassociated ascomycetes, (g) pathogens, (h) moss-associated fungi, (i) soil saprotrophs (saprotrophic taxa commonly found in N-rich mineral soils), (j) unknown function, based on the UNITE database, DEEMY (www.deemy.de) or FUNGuild [65]. ECM species were assigned to exploration types according to the DEEMY database [66,67].

#### *2.6. Phylogenetic and Statistical Analyses*

The ghost-tree approach [68], which allows sequence data to be integrated into a single tree, was used to reconstruct the fungal phylogenetic tree. Foundation phylogeny at the family level was derived by (Treebase ID S20837) [69], following methodology and was based on the sequences of six genes 18sS rRNA, 28S rRNA, 5.8S rRNA, translation elongation factor 1-α (tef1α), and RNA polymerase II (two subunits: RPB1 and RPB2) [70].

Statistical analyses were implemented in the R software environment (version 3.6.1, R Development Core Team 2019). The *ape* package was used to load and manipulate the phylogenetic tree in newick format [71], while the *phyloseq* package was used to import

and handle OTU counts [72], taxonomic assignments, and associated phylogenetic tree. The *philR* package was used to analyze compositional data using the phylogenetic tree information [73]. The *picante* package was used to calculate the ectomycorrhizal phylogenetic structure indices (NRI and NTI) and Faith's phylogenetic diversity [74,75]. The *vegan* package was used for the multivariate analyses [76].

For all compositional analyses, the ectomycorrhizal species abundance matrix was previously filtered to exclude the taxa that were not seen in at least 10% of samples to eliminate random noise. We analyzed ectomycorrhizal phylogenetic community composition using philR which enables us to transform compositional data into an orthogonal unconstrained space with phylogenetic and evolutionary interpretation [73]. First, PhilR Isometric Log Ratio transformations were built from the phylogenetic tree utilizing a weighted reference [73], then a Euclidean distance matrix was built from the philR transformed data. After that, Redundancy analysis, (RDA function "*rda*") was used to visualize ectomycorrhizal phylogenetic compositional differences between tree host species. Moreover, ectomycorrhizal phylogenetic differences between tree host species were tested using permutational multivariate analyses of variance (PMAV, function "*adonis*") on the Euclidean dissimilarity matrix based on the philR transformed data. To test the phylogenetic structure of ectomycorrhizal communities between tree host species were calculated the standardized effect size of mean pairwise distances and mean nearest taxon distances using *ses.mpd* (Standardized effect of mean pairwise distances in communities) and *ses.mntd* (Standardized effect of nearest taxon index in communities) functions from *picante*. In each stand type, we compared the MPD and MNTD values with the MPD and MNTD distributions of random communities in order to identify whether communities were more over-dispersed or under-dispersed than expected by chance. We used the *independentswap* null model, which randomizes community data matrix with the independent swap algorithm maintaining species occurrence frequency and sample species richness, to construct from 9999 randomly assembled communities [77]. After calculating SES.MPD and SES.MNTD, the values were multiplied by −1 as these values are equivalent to −1 times NRI (net relatedness index) and NTI (nearest taxon index), respectively. Importantly, an increase in the NRI value indicates increasing phylogenetic clustering (or decreasing overall relatedness) of a set of species relative to the source pool [25]. On the other hand, the nearest taxon index (NTI) is a standardized measure of the mean phylogenetic distance to the nearest taxon in each sample/community [25]. The NRI measures the standardized effect size of the mean phylogenetic distance (MPD), which estimates the average phylogenetic relatedness between all possible pairs of taxa in a community. The NTI calculates the mean nearest phylogenetic neighbor among the individuals in a community. The ectomycorrhizal phylogenetic diversity comparisons between tree host species were done using Faith's PD phylogenetic diversity index with the function *pd*. Moreover, to assess the phylogenetic relationships among species change across space, we computed multiple-site phylogenetic turnover, nestedness, and phylo-beta diversity (Sorensen similarity index) per tree host species using "*phylo.beta.multi*" function in the betapart package [78].

Second, variation partitioning (function "*varpart*") was used to test the relative importance as variation sources of geographical distances, soil parameters, and stand structure in ectomycorrhizal phylogenetic composition (philR transformed data) and structure (NRI, NTI). To avoid multicollinearity before variation partitioning analysis highly correlated environmental variables were removed (*r* > 0.7). The geographical distances included were previously evaluated using principal coordinates of neighbors' matrices spatial eigenvectors (PCNM, *pcnm* function) based on UTM coordinates of the sampled stands with Euclidean distances. Thus, significant spatial eigenvectors were forward selected and the selected spatial eigenvectors were used as explanatory variables in the variation partitioning, together with soil (Sand content, K, Mg, organic matter, Na, N, P, water pH, and CN ratio) and stand structural variables (Tree species, Altitude, Slope, Trees per hectare, and Basal Area). The significance of each partition was tested using multivariate ANOVAs. Moreover, to evaluate the effect of pH, CN, and P on the ectomycorrhizal phylogenetic composition

we conducted a redundancy analysis (*rda* function). In addition, the *sm.density.compare* (*n* of permutations = 999) from the *sm* package was used to randomly assign pH values between the five tree hosts and estimate how different the densities were using a permutational test of density equality [79]. Lastly, to visualize if the ectomycorrhizal phylogenetic communities were clustering across the pH gradient, we performed a hierarchical cluster analysis on the ectomycorrhizal phylogenetic compositional data based on the Euclidian distance matrix using the function *hclust* in the *stats* package.

Finally, a binary data matrix was compiled with ectomycorrhizal exploration traits (contact, short, medium smooth, medium mat, medium-fringe, and long). Then, we calculated the trait *ses.mpd* and *ses.mntd* using the *independentswap* null model to assess trait structure following the same methodology for communities. Finally, to test for a phylogenetic signal to exploration types, K' Blomberg statistics were calculated for the presence of the traits using the function *MultiPhylosignal* in the *picante* package [74,80]. Moreover, the traits were visualized on the phylogenetic tree by plotting the exploration types at the tips of the phylogenetic tree following [74].

#### **3. Results**

#### *3.1. Ectomycorrhizal Phylogenetic Description*

The hybrid phylogenetic tree of ectomycorrhizal fungi was consistent with Mikryukov et al. (2020) (Figure 1). The families Sebacinaceae, Clavulinaceae, and Hydnaceae clearly formed a monophyletic group, while Bankeraceae Thelephoraceae, Russulaceae, and Albatrellaceae formed two distinct clades (Figure 1). Moreover, two other family groups were identified, one including Atheliaceae, Sclerodermataceae, Boletaceae, Gomphidiaceae, and Suillaceae, and the other including Tricholomataceae, Amanitaceae, Hydnangiaceae, Cortinariaceae, Hymenogastraceae, and Inocybaceae (Figure 1). Finally, the most abundant species in each tree host were indicated in Table S2.

#### *3.2. Ectomycorrhizal Phylogenetic Composition, Structure, and Diversity*

There were no significant differences in the ectomycorrhizal phylogenetic composition among tree host species (*r* <sup>2</sup> = 0.10, F(4,41) = 1.12, *p* = 0.281). The RDA and the sd-ellipses based on the philR Euclidean distance matrix clearly showed that all forest types were overlapping at the ordination center (Figure 2). Redundancy analyses resulted in two main axes that explained together 22% of the variance. However, *P. halepensis-nigra*, *P. halepensis*, and *P. nigra* communities were less spread (homogeneous), while, *P. sylvestris* and *P. sylvestris-nigra* communities were more overdispersed in the ordination space (heterogeneous). Regarding ectomycorrhizal phylogenetic structure, no significant difference was detected for NRI (F(4,41) = 0.26, *p* = 0.901) between tree host species. Positive mean values of NRI were detected in *P. halepensis* (0.51 ± 0.09), indicating ectomycorrhizal higher phylogenetic clustering. *P. nigra-halepensis* (0.14 ± 0.83), *P. sylvestris-nigra* (0.06 ± 0.25) and *P. sylvestris* (0.05 ± 0.27), and *P. nigra* (−0.02 ± 0.26) showed dispersion of NRI values positive and negative around 0 (Figure 3a). However, we detected significant differences in NTI values (F(4,41) = 2.96, *p* = 0.031) between tree host species. Mean positive NTI values were detected across all tree host species, except in *P. sylvestris-nigra* (−0.46 ± 0.37), indicating ectomycorrhizal phylogenetic clustering in *P. halepensis*, *P. nigra-halepenesis*, while *P. nigra*, *P. sylvestris* were not clearly defined, with values around 0, and a marginal phylogenetic overdispersion was detected in *P. sylvestris-nigra* (Figure 3b).

**Figure 1.** (**a**) The hybrid phylogenetic tree of ectomycorrhizal families based on the foundation phylogeny derived by Zhao et al., (2017), based on the sequences of six genes 18sS rRNA, 28S rRNA, 5.8S rRNA, translation elongation factor 1 α(tef1α) and RNA polymerase II. (**b**) Relative abundance of the most abundant ectomycorrhizal families. **Figure 1.** (**a**) The hybrid phylogenetic tree of ectomycorrhizal families based on the foundation phylogeny derived by Zhao et al. (2017), based on the sequences of six genes 18sS rRNA, 28S rRNA, 5.8S rRNA, translation elongation factor 1-α (tef1α) and RNA polymerase II. (**b**) Relative abundance of the most abundant ectomycorrhizal families.

tion of them.

*3.2. Ectomycorrhizal Phylogenetic Composition, Structure, and Diversity* 

logenetic overdispersion was detected in *P. sylvestris-nigra* (Figure 3b).

*nigra-halepensis* (Phylo beta.sim: 0.40; Phylo beta.sne: 0.14; Table S1).

There were no significant differences in the ectomycorrhizal phylogenetic composition among tree host species (*r*2 = 0.10, F(4,41) = 1.12, *p* = 0.281). The RDA and the sd-ellipses based on the philR Euclidean distance matrix clearly showed that all forest types were overlapping at the ordination center (Figure 2). Redundancy analyses resulted in two main axes that explained together 22% of the variance. However, *P. halepensis-nigra*, *P. halepensis,* and *P. nigra* communities were less spread (homogeneous), while, *P. sylvestris* and *P. sylvestris-nigra* communities were more overdispersed in the ordination space (heterogeneous). Regarding ectomycorrhizal phylogenetic structure, no significant difference was detected for NRI (F(4,41) = 0.26, *p* = 0.901) between tree host species. Positive mean values of NRI were detected in *P. halepensis* (0.51 ± 0.09), indicating ectomycorrhizal higher phylogenetic clustering. *P. nigra-halepensis* (0.14 ± 0.83), *P. sylvestris-nigra* (0.06 ± 0.25) and *P. sylvestris* (0.05 ± 0.27), and *P. nigra* (−0.02 ± 0.26) showed dispersion of NRI values positive and negative around 0 (Figure 3a). However, we detected significant differences in NTI values (F(4,41) = 2.96, *p* = 0.031) between tree host species. Mean positive NTI values were detected across all tree host species, except in *P. sylvestris-nigra* (−0.46 ± 0.37), indicating ectomycorrhizal phylogenetic clustering in *P. halepensis*, *P. nigra-halepenesis*, while *P. nigra*, *P. sylvestris* were not clearly defined, with values around 0, and a marginal phy-

The ectomycorrhizal phylogenetic diversity analysis showed no significant differences between tree host species (F(4,41) = 0.92, *p* = 0.458, Figure 3c), with PD mean values ranging from 6.6 of *P. nigra* and 8.3 *P. sylvestris* (Figure 3c). In addition, analysis of multiple-site phylogenetic similarities showed that total beta diversity values were similar across host tree species (Table S1), although, species turnover resulted strongly higher than species nestedness across the tree host species and with similar values, except for *P.* 

**Figure 2.** philR RDA ordination based on Euclidean distance matrix displaying ectomycorrhizal phylogenetic community composition of *P. halepensis*, *P. nigra-halepensis*, *P. nigra*, *P. sylvestris-nigra*  and *P. sylvestris* forest and the sd ellipses of each forest. **Figure 2.** philR RDA ordination based on Euclidean distance matrix displaying ectomycorrhizal phylogenetic community composition of *P. halepensis*, *P. nigra-halepensis*, *P. nigra*, *P. sylvestris-nigra* and *P. sylvestris* forest and the sd ellipses of each forest. *J. Fungi* **2021**, *7*, x FOR PEER REVIEW 9 of 17

**Figure 3.** Boxplots displaying (**a**) Net Relatedness Index (NRI) values (**b**) Nearest Taxon Index (NTI) (**c**) Faith's PD values between tree host species (Halepensis: *P. halepensis*, NigHal*: P. nigra-halepensis*, Nigra*: P. nigra*, SylNig*: P. sylvestris-nigra*  and Sylvestris: *P. sylvestris*). Means were compared using ANOVA and Tukey's HSD tests, with letters denoting significant differences between host species. **Figure 3.** Boxplots displaying (**a**) Net Relatedness Index (NRI) values (**b**) Nearest Taxon Index (NTI) (**c**) Faith's PD values between tree host species (Halepensis: *P. halepensis*, NigHal: *P. nigra-halepensis*, Nigra: *P. nigra*, SylNig: *P. sylvestris-nigra* and Sylvestris: *P. sylvestris*). Means were compared using ANOVA and Tukey's HSD tests, with letters denoting significant differences between host species.

*3.3. Main Drivers of Ectomycorrhizal Phylogenetic Composition and Structure*  When testing the relative importance of geographic distance, soil parameters, and stand structure on ectomycorrhizal phylogenetic composition, soil accounted for the greatest proportion of the total variance (4%) followed by geographic distance, however, these fractions were not significant (*p* > 0.05, Figure 4a). Moreover, stand structure, soil, and geographic distance shared 4% of the total variance. Conversely, when the phylogenetic structure was analyzed, stand structure, soil, and geographical distance shared an The ectomycorrhizal phylogenetic diversity analysis showed no significant differences between tree host species (F(4,41) = 0.92, *p* = 0.458, Figure 3c), with PD mean values ranging from 6.6 of *P. nigra* and 8.3 *P. sylvestris* (Figure 3c). In addition, analysis of multiple-site phylogenetic similarities showed that total beta diversity values were similar across host tree species (Table S1), although, species turnover resulted strongly higher than species nestedness across the tree host species and with similar values, except for *P. nigra-halepensis* (Phylo beta.sim: 0.40; Phylo beta.sne: 0.14; Table S1).

#### 8% proportion of variation, while stand structure accounted for 4% of the total variance (*p* > 0.05) (Figure 4b). Finally, the phylogenetic structure was marginally influenced by soil *3.3. Main Drivers of Ectomycorrhizal Phylogenetic Composition and Structure*

(2%) and not by geographic distance. When testing the relative importance of geographic distance, soil parameters, and stand structure on ectomycorrhizal phylogenetic composition, soil accounted for the greatest proportion of the total variance (4%) followed by geographic distance, however, these fractions were not significant (*p* > 0.05, Figure 4a). Moreover, stand structure, soil, and geographic distance shared 4% of the total variance. Conversely, when the phylogenetic

pH was the only significant soil predictor influencing phylogenetic composition (Variance = 0.55, F = 2.76, *p* = 0.002). Thus, the distribution of pH values was significantly different across tree hosts (*p* = 0.041). Moreover, when pH densities were compared only

**Figure 4.** Variance partitioning analyses for (**a**) ectomycorrhizal phylogenetic composition and (**b**) ectomycorrhizal phylogenetic structure (NRI, NTI indices) in response to stand structure, soil physiochemistry and geographic distance. Values show the fraction of variation explained by each group of parameters, as well as the shared contribution of each combinadifferences between host species.

*J. Fungi* **2021**, *7*, x FOR PEER REVIEW 9 of 17

**Figure 3.** Boxplots displaying (**a**) Net Relatedness Index (NRI) values (**b**) Nearest Taxon Index (NTI) (**c**) Faith's PD values between tree host species (Halepensis: *P. halepensis*, NigHal*: P. nigra-halepensis*, Nigra*: P. nigra*, SylNig*: P. sylvestris-nigra*  and Sylvestris: *P. sylvestris*). Means were compared using ANOVA and Tukey's HSD tests, with letters denoting significant

*3.3. Main Drivers of Ectomycorrhizal Phylogenetic Composition and Structure* 

structure was analyzed, stand structure, soil, and geographical distance shared an 8% proportion of variation, while stand structure accounted for 4% of the total variance (*p* > 0.05) (Figure 4b). Finally, the phylogenetic structure was marginally influenced by soil (2%) and not by geographic distance. netic structure was analyzed, stand structure, soil, and geographical distance shared an 8% proportion of variation, while stand structure accounted for 4% of the total variance (*p* > 0.05) (Figure 4b). Finally, the phylogenetic structure was marginally influenced by soil (2%) and not by geographic distance.

When testing the relative importance of geographic distance, soil parameters, and stand structure on ectomycorrhizal phylogenetic composition, soil accounted for the greatest proportion of the total variance (4%) followed by geographic distance, however, these fractions were not significant (*p* > 0.05, Figure 4a). Moreover, stand structure, soil, and geographic distance shared 4% of the total variance. Conversely, when the phyloge-

**Figure 4.** Variance partitioning analyses for (**a**) ectomycorrhizal phylogenetic composition and (**b**) ectomycorrhizal phylogenetic structure (NRI, NTI indices) in response to stand structure, soil physiochemistry and geographic distance. Values show the fraction of variation explained by each group of parameters, as well as the shared contribution of each combination of them. **Figure 4.** Variance partitioning analyses for (**a**) ectomycorrhizal phylogenetic composition and (**b**) ectomycorrhizal phylogenetic structure (NRI, NTI indices) in response to stand structure, soil physiochemistry and geographic distance. Values show the fraction of variation explained by each group of parameters, as well as the shared contribution of each combination of them.

> pH was the only significant soil predictor influencing phylogenetic composition (Variance = 0.55, F = 2.76, *p* = 0.002). Thus, the distribution of pH values was significantly different across tree hosts (*p* = 0.041). Moreover, when pH densities were compared only pH was the only significant soil predictor influencing phylogenetic composition (Variance = 0.55, F = 2.76, *p* = 0.002). Thus, the distribution of pH values was significantly different across tree hosts (*p* = 0.041). Moreover, when pH densities were compared only *P. sylvestris* and *P. nigra* differed significantly (*p* = 0.009) showing a larger left tail towards lower pH values (Figure S2). The hierarchical clustering of the ectomycorrhizal phylogenetic composition showed that communities were clustered into two main groups (Figure S3), Here, the group composed of *P. halepensis*, *P. sylvestris*, and *P. sylvestris-nigra* communities clustered at lower pH values (<7), while *P. nigra* and *P. nigra-halepensis* communities only occurred at higher pH values (>7) (Figure S3).

#### *3.4. Trait Evolution of the Exploration Types*

When the exploration traits were visualized on the phylogenetic tree, 59 OTUs out of 184 had short exploration types, up to 53 had contact exploration types, while 39 OTUs and 25 OTUs had medium fringe and medium smooth exploration types. Conversely, medium mat and long exploration types were the least abundant with 9 and 8 OTUs, respectively. Finally, we did not find any phylogenetic signal for any exploration type (0.25 < K < 0.77, *p* > 0.05), as exploration types were dispersed across the phylogenetic tree (Figure 5).

only occurred at higher pH values (>7) (Figure S3).

*3.4. Trait Evolution of the Exploration Types* 

**Figure 5.** The hybrid phylogenetic tree displaying the distribution of the exploration types (Contact, short, medium smooth, medium fringe and long). **Figure 5.** The hybrid phylogenetic tree displaying the distribution of the exploration types (Contact, short, medium smooth, medium fringe and long).

#### **4. Discussion 4. Discussion**

5).

The results of our phylogenetic study on ectomycorrhizal communities in Mediterranean pine forests showed that phylogenetic composition, structure, and diversity were similar among habitats with distinct pine tree hosts. However, significant differences were found in nearest taxon index values between *P. nigra-halepensis* and *P. sylvestris-nigra*, probably not directly caused by differences in tree hosts but due to higher differences in the local abiotic conditions in *P. sylvestris-nigra* than in *P. halepensis-nigra* sites. Moreover, we detected a weak abiotic filtering effect on the ectomycorrhizal phylogenetic compositional variation, being pH the only variable among soil variables that significantly influence the ectomycorrhizal phylogenetic community. This finding suggests that pH acts as a strong abiotic filter on the ectomycorrhizal community at both phylogenetic and taxonomic levels [17]. In contrast, ectomycorrhizal phylogenetic structure variation was marginally influenced only by the shared effect of stand structure, soil, and geographic distance. Therefore, the phylogenetic structure may be indirectly influenced by other processes (i.e., competition; [30]) not directly tested in this study. Finally, we identified that short and contact exploration types were the most abundant in these forest ecosystems. Conversely, long exploration types were the least abundant, although there was no phylogenetic signal since exploration types were dispersed across the phylogenetic tree. The results of our phylogenetic study on ectomycorrhizal communities in Mediterranean pine forests showed that phylogenetic composition, structure, and diversity were similar among habitats with distinct pine tree hosts. However, significant differences were found in nearest taxon index values between *P. nigra-halepensis* and *P. sylvestris-nigra*, probably not directly caused by differences in tree hosts but due to higher differences in the local abiotic conditions in *P. sylvestris-nigra* than in *P. halepensis-nigra* sites. Moreover, we detected a weak abiotic filtering effect on the ectomycorrhizal phylogenetic compositional variation, being pH the only variable among soil variables that significantly influence the ectomycorrhizal phylogenetic community. This finding suggests that pH acts as a strong abiotic filter on the ectomycorrhizal community at both phylogenetic and taxonomic levels [17]. In contrast, ectomycorrhizal phylogenetic structure variation was marginally influenced only by the shared effect of stand structure, soil, and geographic distance. Therefore, the phylogenetic structure may be indirectly influenced by other processes (i.e., competition; [30]) not directly tested in this study. Finally, we identified that short and contact exploration types were the most abundant in these forest ecosystems. Conversely, long exploration types were the least abundant, although there was no phylogenetic signal since exploration types were dispersed across the phylogenetic tree.

*P. sylvestris* and *P. nigra* differed significantly (*p* = 0.009) showing a larger left tail towards lower pH values (Figure S2). The hierarchical clustering of the ectomycorrhizal phylogenetic composition showed that communities were clustered into two main groups (Figure S3), Here, the group composed of *P. halepensis*, *P. sylvestris,* and *P. sylvestris-nigra* communities clustered at lower pH values (<7), while *P. nigra* and *P. nigra-halepensis* communities

When the exploration traits were visualized on the phylogenetic tree, 59 OTUs out of 184 had short exploration types, up to 53 had contact exploration types, while 39 OTUs and 25 OTUs had medium fringe and medium smooth exploration types. Conversely, medium mat and long exploration types were the least abundant with 9 and 8 OTUs, respectively. Finally, we did not find any phylogenetic signal for any exploration type (0.25 < K < 0.77, *p* > 0.05), as exploration types were dispersed across the phylogenetic tree (Figure

#### *4.1. Ectomycorrhizal Phylogenetic Description*

Our study allowed us to investigate the phylogenetic relationships between 256 OTUs using a multiple gene tree at family level as a foundation tree which allows us to build a better-supported tree (Figure 1a) [70]. Also, we were able to identify monophyletic groups of families, such as Sebacinaceae, Clavulinaceae, and Hydnaceae, and Atheliaceae, Sclerodermataceae, Boletaceae, Gomphidiaceae, and Suillaceae, however, this last clade formed a paraphyletic group with Russulaceae and Albatrellaceae. Moreover, two other family groups were identified, one including Bankeraceae, Thelephoraceae, Sclerodermataceae, Boletaceae, Gomphidiaceae, and Suillaceae, and the other including Tricholomataceae, Amanitaceae, Hydnangiaceae, Cortinariaceae, Hymenogastraceae, and Inocybaceae. Therefore, the resolved phylogenetic tree resulted in a strong backbone for the downstream analyses as the level of resolution allows us to perform reliable phylogenetic diversity analyses [81]. Finally, disentangling the ectomycorrhizal phylogenetic community structure in our study region, where the current climate change may lead to changes in ecosystems functioning, is crucial to predict the impacts on ectomycorrhizal taxonomic and phylogenetic community composition and diversity [82].

#### *4.2. Ectomycorrhizal Phylogenetic Composition, Structure, and Diversity*

Our results demonstrated that ectomycorrhizal phylogenetic community and diversity were not significantly different among pine tree host species or in NRI values, although there were differences in NTI values between *P. sylvestris-nigra* and *P. nigra-halepensis*. These results are in accordance with previous taxonomical studies on ectomycorrhizal communities between congeneric tree hosts [9,83], in which a lack of phylogenetic differences was observed. Similarly, ectomycorrhizal community composition was not different between phylogenetically related pines in China [84]. In contrast, several studies reported taxonomical differences between ectomycorrhizal communities between hosts of different families or genera [7,85]. Thus, it seems that at both taxonomic and phylogenetic levels, ectomycorrhizal communities are not varying significantly among phylogenetically close related tree hosts [14].

Similar phylogenetic studies detected phylogenetic clustering of Agaromycotina communities in xeric oak-dominated forests and concluded that oak acted as the main habitat filter [26]. Here, our results showed an opposite trend, with no significant differences in ectomycorrhizal phylogenetic structure and diversity between pine tree hosts. However, we observed significant differences in phylogenetic dispersion among habitats with distinct pines hosts. For example, ectomycorrhizal species in *P. halepensis-nigra* forest resulted in phylogenetic clusters, while in *P. sylvestris-nigra* were slightly more overdispersed at the tip of the phylogeny (Figure 3a), probably due to the low number of *P. nigra-halepensis* sites which may have caused underestimation of the differences between phylogenetic taxa. In this regard, the three *P. nigra-halepensis* sites showed similar soil properties (i.e., values range from pH: 8.18–8.38, CN: 11–151, P: 2–5, N: 0.12–0.16), which may have resulted in the occurrence of closely related species that are adapted to these similar abiotic conditions. These results may imply that Mediterranean pine host tree species are weak habitat filters for ectomycorrhizal fungi, probably due to a lack of host specificity among congeneric hosts. Thus, our results are in agreement with the hypothesis that the lack of phylogenetic composition, structure, and diversity between pine host species may be partially explained by possible conserved symbiosis between *Pinus* and ectomycorrhizal fungi [86].

Finally, we found high and similar turnover values in all the tree host species forest, while nestedness was significantly lower, except in the case of *P. nigra-halepensis* forest. It seems that both environmental filtering by soil and dispersal limitation may, to a certain extent, promote species replacement among sites [87]. However, in *P. nigra-halepensis* forest higher nestedness might indicate local species loss probably due to its soil site conditions that resulted in the occurrence of a locally adapted subset of species.

#### *4.3. Main Drivers of Ectomycorrhizal Phylogenetic Composition and Structure*

In this study, we observed that soil parameters influenced ectomycorrhizal phylogenetic composition, while phylogenetic structure variation was primarily influenced by the shared effect of the three environmental filters. However, these three fractions were not significant and explained a residual amount of variation, thus, the second hypothesis is not accepted. Although previous studies have identified that soil parameters are the main drivers of taxonomic ectomycorrhizal community variation in Mediterranean pine forest [17], here, soil parameters were marginally important in driving ectomycorrhizal phylogenetic composition. This may imply that at the phylogenetic level, the lack of strong abiotic gradients results in the occurrence of non-closely related species which are adapted to heterogeneous but not specific environmental conditions [88].

Soil properties have been widely described as a strong abiotic filter on taxonomic fungal communities at different spatial scales [14,62,89,90]. In contrast, we observed a weak abiotic filtering effect of soil physico-chemistry on phylogenetic community composition, with pH resulting in the only significant variable. The importance of pH as an influential variable over ectomycorrhizal community composition at local and regional scales has been widely described [87,90,91]. However, in view of our results, it seems that pH acts as an abiotic filter at both taxonomic and phylogenetic levels [87,92,93]. In addition, our

results showed that a left tail of *P. sylvestris* and *P. sylvestris-nigra* in the distribution of the pH values resulting in a wider niche for species adapted to low pH values (Figure S2). In this regard, phylogenetic fungal communities were more clustered at lower values of pH (<7), thus, it seems that lower pH values might result in the occurrence of only adapted fungal species that can grow and maintain cellular function in acidic environments [94], causing higher phylogenetic similarity [87,95,96].

Regarding phylogenetic structure, stand structure alone explained a proportion of its variation although this was not significant. However, stand structure, soil physicochemistry, and geographic distance shared an important proportion of variance. In Mediterranean ecosystems, the influence of stand structural variables on fungi has already been assessed over mushroom yields with important effects [81]. Although weakly, differences in stand structural variables may result in the occurrence of different less phylogenetically related fungal species that are better adapted to certain local forest conditions. In view of our results, we argue that ectomycorrhizal phylogenetic structure is more importantly influenced by the combined effect of all environmental variables. Hence, phylogenetic relatedness between species decreases with increasing geographic distance, differences in stand structure, and soil conditions. Finally, we hypothesize that there may be other processes influencing the ectomycorrhizal phylogenetic community, such as competition for space and resources [30]. that were not directly tested, therefore further studies are needed to further disentangle whether other processes influence phylogenetic structure in these ecosystems.

#### *4.4. Trait Evolution of the Exploration Types*

We observed that 51% of the ectomycorrhizal species had short and contact exploration types, 21% and 13% of the species had medium fringe and medium exploration types, respectively, and only 4% of the species had long exploration types. Similarly, [38] found that in P. pinaster-dominated Mediterranean forests, long distance were the least abundant exploration types, while short and contact types were dominating the community. Moreover, our results showed that traits were dispersed across the phylogenetic tree (Figure 5). Thus, the third hypothesis is accepted. In this regard, the dispersion of traits across the phylogenetic tree suggests that even more distant related species showed the same exploration type, resulting in a random trait pattern with a lack of phylogenetic signal. In addition, [38] found that mycorrhizal species with long distance exploration types were less abundant under drier conditions, whereas short-distance and contact type species increased. Recent studies suggest that drier conditions may favor short-contact types [18,38]. Similarly, based on our results, we argue that dispersion of short and contact exploration types might be an adaptation to the Mediterranean stress conditions where the limiting factor is water and not nutrients. Therefore, having medium mat and long exploration types might be a disadvantage due to their higher C demand on the host [18,32,38]. At the same time, in northern and temperate ecosystems, soil N is a limiting nutrient [81], and previous studies have shown that species with long medium mat exploration occurs in soils where N is limiting and patchily distributed [37,42], while short exploration types are more efficient in up-taking soluble inorganic N [36]. However, despite these observed trends and since exploration types of mycorrhizae represent a distinct set of fungal traits, the use of exploration types to study fungal trait responses to environmental changes can be often misleading, and further research should be addressed. In any case, previous work in this area showed a lack of N effect on mycorrhizal communities in Mediterranean pine forests [17], therefore, as N is not limiting it can be easily captured by ectomycorrhizal fungi close to the host with no need of investing in high biomass exploration types.

Finally, we acknowledge the accuracy of the ITS2 region in species identification and resolution [97], but also its limitation in phylogenetic applications due to its high variability [70,98]. However, the use of a backbone phylogenetic tree at family level constructed from multiple gene sequences provides a sufficient taxonomic resolution, thus can be an accurate predictor of phylogenetic diversity metrics [99]. Moreover, it is known

that the identification of basidiomycetes through ITS2 amplification is more efficient than in other taxa (i.e., Ascomycetes) [100]. Therefore, future studies aiming to disentangle fungal phylogenetic patterns in community structure should include a robust backbone phylogenetic tree and at least the whole ITS region.

#### **5. Conclusions**

In this study, we found no differences neither in ectomycorrhizal phylogenetic community composition nor structure and diversity, indicating that ectomycorrhizal communities at both phylogenetic and taxonomic levels do not change among phylogenetically closely related tree hosts. Moreover, soil parameters only had a marginal filtering effect on ectomycorrhizal phylogenetic variation as pH resulted in the only significant driver of the phylogenetic community. In this regard, our results showed that pH acts as the broadest abiotic filter of ectomycorrhizal communities at a local scale.

Conversely, the ectomycorrhizal phylogenetic structure was marginally influenced by the combined effect of soil, stand structure, and geographic distance, indicating that phylogenetic structure is mainly influenced by their combined effect.

Finally, short and contact distance were the dominant exploration types, as they may be favored under drought stress conditions but also under high nutrient availability. Our results shed light on the drivers of ectomycorrhizal phylogenetic community variation in Mediterranean pine forests, being fundamental to get a better insight on the drivers of community assembly and ecosystem functioning. Nevertheless, further research on ectomycorrhizal phylogenetic communities is needed to better understand how changes in deterministic processes will affect ectomycorrhizal communities and forest ecosystems' functioning.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/jof7100793/s1, Figure S1: Catalonia map displaying the location of the 42 plots, Figure S2: Density curves of pH values of the five tree hosts from the permutational test of density equality, Figure S3: Hierarchical clustering of ectomycorrhizal phylogenetic compositional data based on a Euclidean distance matrix, Table S1: Table summarizing the main texture and moisture properties of the study plots, Table S2: Most abundant ectomycorrhizal species detected in pure and mixed stands of *Pinus* spp., Table S3: Phylogenetic species turnover, nestedness and total beta diversity values across host tree species stands.

**Author Contributions:** C.C. (Carlos Colinas), J.A.B., J.M.d.A. and J.G.A. planned and designed the fungal research. J.M.d.A. sampled the soils and together with C.C. (Carlos Colinas), measured the environmental variables. C.C. (Carles Castaño) performed the lab works and the bioinformatic analyses. I.A. analysed the data and wrote the manuscript, with inputs of J.G.A. and C.C. (Carles Castaño). All authors provided inputs on the last version. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project has received funding from the European Union's H2020 research and innovation programme under Marie Sklodowska-Curie grant agreement No 801586. This work was partially supported by the Spanish Ministry of Science, Innovation and Universities, grants RTI2018-099315- A-I00. I.A. was supported by a H2020-Marie Slodowska Curie Action Cofund fellowship (801596), J.G.A. was supported by Ramon y Cajal fellowship (RYC-2016-20528) and J.A.B. benefitted from a Serra-Húnter Fellowship provided by the Generalitat of Catalunya.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Available upon reasonable request.

**Conflicts of Interest:** The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.

#### **References**


**Xin Meng 1,2,3, Geng-Shen Wang 1,2,3, Gang Wu 1,2, Pan-Meng Wang 1,2,3, Zhu L. Yang 1,2,\* and Yan-Chun Li 1,2,\***


**Abstract:** *Leccinum* is one of the most important groups of boletes. Most species in this genus are ectomycorrhizal symbionts of various plants, and some of them are well-known edible mushrooms, making it an exceptionally important group ecologically and economically. The scientific problems related to this genus include that the identification of species in this genus from China need to be verified, especially those referring to European or North American species, and knowledge of the phylogeny and diversity of the species from China is limited. In this study, we conducted multilocus (nrLSU, *tef1-α*, *rpb2*) and single-locus (ITS) phylogenetic investigations and morphological observisions of *Leccinum* from China, Europe and North America. Nine *Leccinum* species from China, including three new species, namely *L. album*, *L. parascabrum* and *L. pseudoborneense*, were revealed and described. *Leccinum album* is morphologically characterized by the white basidioma, the white hymenophore staining indistinct greenish blue when injured, and the white context not changing color in pileus but staining distinct greenish blue in the base of the stipe when injured. *Leccinum parascabrum* is characterized by the initially reddish brown to chestnut-brown and then pale brownish to brown pileus, the white to pallid and then light brown hymenophore lacking color change when injured, and the white context lacking color change in pileus but staining greenish blue in the base of the stipe when injured. *Leccinum pseudoborneense* is characterized by the pale brown to dark brown pileus, the initially white and then brown hymenophore lacking color change when injured, and the white context in pileus and stipe lacking color change in pileus but staining blue in stipe when bruised. Color photos of fresh basidiomata, line drawings of microscopic features and detailed descriptions of the new species are presented.

**Keywords:** boletes; taxonomy; morphology; phylogeny; new taxa

#### **1. Introduction**

The genus *Leccinum* Gray is a species-rich genus of Boletaceae and is characterized by a whitish or yellow hymenophore, a white to cream context unchanging or staining blue or red when injured, a brown to blackish scabrous to dotted squamules on the surface of the stipe, and comparatively long and smooth basidiospores. Generally, most species of the genus are widely spread in the subarctic, boreal, temperate and Mediterranean regions, with a few secondary expansions to the neotropics [1–12]. Species in *Leccinum* are both ecologically and economically important. Most species of this genus exhibit mycorrhizal host specificity. Species of *Leccinum* sect. *Scabra* Smith & Thiers are associated with plants of *Betula*, while species of *L.* sect. *Fumosa* (A.H. Smith, Thiers & Watling) Gelardi are associated with plants of *Populus*. In *L.* sect. *Leccinum*, species are found exclusively associated with plants of *Populus* (e.g., *L. albostipitatum* den Bakker & Noordel. and *L. insigne*

**Citation:** Meng, X.; Wang, G.-S.; Wu, G.; Wang, P.-M.; Yang, Z.L.; Li, Y.-C. The Genus *Leccinum* (Boletaceae, Boletales) from China Based on Morphological and Molecular Data. *J. Fungi* **2021**, *7*, 732. https://doi.org/ 10.3390/jof7090732

Academic Editors: Anush Kosakyan, Rodica Catana and Alona Biketova

Received: 28 May 2021 Accepted: 31 August 2021 Published: 6 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

A.H. Sm., Thiers & Watling), *Betula* (e.g., *L. atrostipitatum* A.H. Sm., Thiers & Watling), Pinaceae (e.g., *L. vulpinum* Watling and *L. piceinum* Pilát & Dermek) and Ericaceae that form arbutoidmycorrhizas (e.g., *L. manzanitae* Thiers and *L. monticola* Halling & G.M. Muell.). However, there are species in section *Leccinum* that are not host specific, i.e., *L. aurantiacum* (Bull.) Gray. This species is associated with plants of *Betula*, *Populus*, *Quercus*, *Salix* and sometimes with *Tilia* [13,14]. Some species of this genus are well-known edible mushrooms, such as *L. quercinum* (Pilát) E.E. Green & Watling, *L. scabrum* (Bull.) Gray and *L. versipelle* (Fr. & Hök) Snell, which are collected in China during the mushroom season.

The genus *Leccinum* was established by Gray in 1821 [13], based on the type species *L. aurantiacum*. Subsequently, more and more mycologists noticed the morphological distinctness and described many new species of this genus. As currently circumscribed, the genus comprises roughly 150 species [1–3,6–56]. North America is the species diversity center of this genus, and in total 118 species have been recorded from this area [19]. Some of the most important works are the serial works of Smith and Thiers [1,15–17], in which three sections of this genus were proposed (*L.* sect. *Leccinum* Smith & Thiers, *L.* sect. *Luteoscabra* Smith & Thiers and *L.* sect. *Scabra*), with 68 species described from Michigan. Twelve species from Central America were described: one species from Belize, eight species from Costa Rica and three species from Colombia [20–24]. In Europe, Singer divided species of this genus into four sections, including two known sections, *L.* sect. *Luteoscabra* and *L.* sect. *Leccinum*, and two newly proposed sections, *L.* sect. *Roseoscabra* and *L.* sect. *Eximia* [3]. In Singer's infrageneric classification, *L.* sect. *Scabra*, established by Smith and Thiers, was merged to *L.* sect. *Leccinum*. Recent molecular phylogenetic evidence has revealed that species of *L.* sect. *Luteoscabra*, *L.* sect. *Roseoscabra* and *L.* sect. *Eximia* belong to divergent clades of Boletaceae and represent many new genera (32,52–54). Thus, the genus *Leccinum* is restricted to the section *Leccinum* (Singer's infrageneric classification) [3]. den Bakker and Noordelos revised the European *Leccinum* species based on morphology and nrLSU sequences and documented sixteen species [14]. In their subsequent study, they treated the three subclades revealed by den Bakker et al. in *L.* section *Leccinum* [33,57] as three subsections (viz. *L.* subsect. *Leccinum*, *L.* subsect. *Fumosa* A.H. Sm., Thiers & Watling and *L.* subsect. *Scabra* Pilat & Dermek) [14]. This infrageneric subdivision was followed in the treatment of the genus in this study. In the Southern Hemisphere, four species have been reported, including one from New Zealand and three from Australia [27–29].

In Asia, six species of *Leccinum* have been reported from Malaysia [6]; ten species from Japan [7–10]; and a total of 31 species have been reported from China based on an extensive literature review [34–36,38–52,56]. Among these Chinese species, twelve species, viz. *L. albellum* (Peck) Singer, *L. chromapes* (Frost) Singer, *L. crocipodium* (Letell.) Watling, *L. eximium* (Peck) Singer, *L. extremiorientale* (Lar. N. Vassiljeva) Singer, *L. griseum* (Quél.) Singer, *L. hortonii* (A.H. Sm. & Thiers) Hongo & Nagas., *L. nigrescens* (Richon & Roze) Singer, *L. rubropunctum* (Peck) Singer, *L. rubrum M. Zang*, *L. rugosiceps* (Peck) Singer and *L. subglabripes* (Peck) Singer have been transferred to other genera [5,11,35,52–55]; eight species, viz. *L. duriusculum* (Schulzer ex Fr.) Singer, *L. intusrubens* (Corner) Høil., *L. oxydabile* (Singer) Singer, *L. quercinum*, *L. rufum* (Schaeff.) Kreisel, *L. subleucophaeum* E.A. Dick & Snell, *L. subradicatum* Hongo and *L. variicolor* Watling were reported without specimen support [39–43,49,51]; and eleven species, viz. *L. ambiguum* A.H. Sm. & Thiers, *L. atrostipitatum* A.H. Sm., Thiers & Watling, *L. aurantiacum*, *L. holopus* (Rostk.) Watling, *L. olivaceopallidum* A.H. Sm., Thiers & Watling, *L. potteri* A.H. Sm., Thiers & Watling, *L. roseofractum* Watling, *L. scabrum*, *L. subgranulosum* A.H. Sm. & Thiers, *L. subleucophaeum* var. *minimum* C.S. Bi and *L. versipelle* were reported with specimen citations [34,38,44–48]. Among these eleven species reported with specimen citations, only *L. subleucophaeum* var. *minimum* was originally described from China, and the remaining species were identified as species originally described from Europe and North America based on general morphological similarities. Indeed, a few species described from Europe and North America do occur in China, especially in northeastern and northwestern China. However,

most species found in China have evolved independently in the southern part of China. Thus, identification of the Chinese *Leccinum* species needs to be reconfirmed.

In this study, we used both morphological data and molecular sequences from the nuclear ribosomal internal transcribed spacer (ITS), the large subunit of the nuclear ribosomal RNA (nrLSU), the translation elongation factor 1-alpha (*tef1-α*) and the RNA polymerase II second largest subunit (*rpb2*), together with ecological data to (1) elucidate species diversity of *Leccinum* in China; (2) evaluate the phylogenetic relationships of species within *Leccinum*; (3) make morphological and ecological comparisons between closely related species.

#### **2. Materials and Methods**

#### *2.1. Taxon Sampling*

Nineteen specimens of the genus *Leccinum* from China were examined. For each collection, a part of the basidioma was dried with silica gel for DNA extraction. The remaining materials were then air-dried at 45–50 ◦C using an electric food dehydrator. Specimens studied in this work were deposited in the Herbarium of the Kunming Institute of Botany, Chinese Academy of Sciences (KUN). Genera are abbreviated as follows: *L.* for *Leccinum*, *Le.* for *Leccinellum*, *O.* for *Octaviania*, *R.* for *Rossbeevera*, *Ru.* for *Rugiboletus*, *T.* for *Turmalinea*, Ca. for *Castanopsis*, *Li*. for *Lithocarpus*, *P.* for *Pinus* and *Q.* for *Quercus*.

#### *2.2. Morphological Observation*

The macroscopic descriptions are based on the detailed field notes and photographs of fresh basidiomata. Color codes of the form "4B2" indicate the plate, row, and color block from Kornerup and Wanscher [58]. For microscopic studies, the microscopic features of each part of the basidioma were observed under microscope (Leica DM2000, Leica Microsystems, Wetzlar, Germany), including basidiospores, basidia, cheilocystidia, pleurocystidia and pileipellis, using 5% KOH as a mounting medium to revive the dried materials. Microscopic studies follow Zhou et al. [59]. In the description of Basidiospores, the abbreviation n/m/p means n basidiospores measured from m basidomata of p collections in 5% KOH solution. The notation of the form (a) b–c (d) stands for the dimensions of the basidiospores; the range b–c contains a minimum of 90% of the measured values, a or d given in parentheses stands for extreme values. Q is used to mean "length/width ratio" of a basidiospore in a side view; Q<sup>m</sup> means average Q of all basidiospores ± sample standard deviation. Measurements of basidospores, cystidia, basidia and terminal cells in pileipellis are presented as length × width. All microscopic structures were drawn freehand from rehydrated material under the microscope with 10× eyepiece and 100× objective (the total magnification is 1000×).

#### *2.3. Molecular Procedures*

Genomic DNA was extracted from silica gel dried materials or herbarium specimens using the CTAB (Cetyltrimethyl ammonium bromide) method [60]. Polymerase chain reactions (PCRs) were performed to amplify partial sequences of nrLSU, *tef1-α, rpb2* and ITS using the extracted DNA. The nrLSU region was amplified with primers LROR/LR5 and LROR/LR3 [61]; *tef1-α* was amplified with primer pair EF1-983F and EF1-1567R [62]; *rpb2* was amplified with primers bRPB2-6F and bRPB2-7.1R [63] and ITS was amplified with primer pair ITS1 and ITS4 [64]. Protocols for the polymerase chain reactions (PCRs) and sequencing followed those in Wu et al. [65] and the references therein.

#### *2.4. Sequence Alignments and Phylogenetic Analyses*

The newly generated sequences of each locus were blasted in GenBank, and the most closely related sequences (nucleotide identities >95%) were downloaded for further alignment. Sequences were aligned separately for each of the loci using MAFFT v7.130b with the E-INS-I strategy and manually optimized on BioEdit v7.0.9 [66,67]. Two datasets, the ITS dataset and the multi-locus (nrLSU + *tef1-α* + *rpb2*) dataset, were analyzed using RAxML and Bayesian methods, respectively. For the multi-locus dataset, single-gene analyses were conducted to assess incongruence among individual genes using the ML

method (results not shown). Because no well-supported bootstrap value (BS > 70%) [55] conflict was detected among the topologies of the three genes, their sequences were then concatenated together for further multi-locus analyses.

For ML analyses, the multi-locus and ITS datasets were analyzed using RAxML (https://www.phylo.org/, accessed on 26 August 2021) under the model GTRGAMMA [68]. Statistical supports for the phylogenetic analyses were determined using nonparametric bootstrapping with 1000 replicates. For BI analyses, the parameter model was selected by the Akaike information criterion (AIC) as the best-fit likelihood model with Modeltest 3.7 (Free Software Foundation, Boston, MA, USA) [69]. The models employed for each of the four loci were GTR + I + G for ITS, nrLSU and *tef1-α*, and SYM + I + G for *rpb2*. Posterior probabilities (PP) were determined twice by running one cold and three heated chains in parallel mode, saving trees every 1000th generation. Other parameters were kept at their default settings. Runs were terminated once the average standard deviation of split frequencies went below 0.01 [70]. Chain convergence was determined using Tracer v1.5 (http://tree.bio.ed.ac.uk/software/tracer/, accessed on 26 August 2021) to confirm sufficiently large ESS values (>200). Subsequently, the sampled trees were summarized after omitting the first 25% of trees as burn-in using the 'sump' and 'sumt' commands implemented in MrBayes.

#### **3. Results**

#### *3.1. Molecular Phylogenetic Analysis*

A total of 57 sequences, including fifteen for nrLSU, fifteen for *tef1-α*, fourteen for *rpb2* and thirteen for ITS, were newly generated in this study and aligned with sequences downloaded from GenBank. Sequences retrieved from GenBank and obtained in this study for the multi-locus phylogenetic analyses are listed in Table 1. The multi-locus dataset (Supplementary File S1) contained 122 sequences (49 for nrLSU, 41 for *tef1-α*, 32 for *rpb2*), representing 51 samples, and the alignment contained 2195 nucleotide sites, of which 530 were parsimony informative. *Borofutus dhakanus* Hosen & Zhu L. Yang and *Spongiforma thailandica* Desjardin, Manfr. Binder, Roekring & Flegel were chosen as the outgroup [71,72]. ML and Bayesian analyses produced very similar estimates of tree topologies, and thus only the tree inferred from ML analysis is displayed (Figure 1). The monophyly of *Leccinum* was highly supported (BS = 100% and PP = 1) in our analyses. Four main clades were recovered, and three of them correspond to the three known subsections, viz. *L.* subsect. *Leccinum*, *L.* subsect. *Fumosa* and *L.* subsect. *Scabra* of *L.* sect. *Leccinum* [14]. Three new species, namely *L. album*, *L. parascabrum* and *L. pseudoborneense*, were revealed in our multi-locus phylogenetic analyses. *Leccinum parascabrum* formed the remaining clade with BS = 100% and PP = 1, while *L. pseudoborneense* and *L. album* nested in *L.* subsect. *Scabra* and clustered together with *L. flavostipitatum* E.A. Dick & Snell, *L. subradicatum* and *L. variicolor* with low supported lineage (BS = 54%).


**Table 1.** Information on specimens used in multi-locus phylogenetic analyses and their GenBank accession numbers. Sequences newly generated in this study are indicated in bold.


**Table 1.** *Cont.*

**Figure 1.** Maximum-likelihood phylogenetic tree generated from a three-locus (nrLSU + *tef1-α* + *rpb2*) dataset. BS > 50% in ML analysis and PP > 0.95 in Bayesian analysis are indicated as RAxML BS/PP above or below supported branches. Species of this genus from China and type species of this genus (*L. aurantiacum*) are indicated in bold. Voucher specimens and localities where the specimens were collected are provided behind the species names. AU = Austria, BE = Belgium, CN = China, CR = Costa Rica, FI = Finland, FR = France, GER = Germany, IT = Italy, JP = Japan, TAI = Thailand, UK =United Kingdom, USA = United States of America and UZ = Uzbekistan.

For the ITS dataset, as revealed by den Bakker et al. [57] and our primary analysis, the ITS1 region contains a minisatellite, which is characterized by the repeated presence of CTATTGAAAAG and CTAATAGAAAG core sequences and mutational derivatives. Moreover, some species contain a minisatellite in the ITS2 region, e.g., the newly described species *L. album* (GenBank Acc. No.: MZ392872 for clone 1 and MZ392873 for clone 2), with a region of 212 bp that consists of tandem repeats (see Supplementary Material for details). Though there is length variation in either the ITS1 or ITS2 spacers, it can also provide some phylogenetic signals. We performed phylogenetic analyses of the ITS dataset. In this dataset (Supplementary File S2), 51 samples were included. The length of the dataset was 1416 bp, of which 377 were parsimony informative. *Leccinellum albellum* (Peck) Bresinsky & Manfr. Binder was chosen as outgroup. ML and Bayesian analyses also produced very similar estimates of tree topologies, and only the tree inferred from ML analysis is displayed (Figure 2). The monophyly of *Leccinum* was also well supported (BS = 100% and PP = 1) in our analyses. Three new species viz. *L. album*, *L. parascabrum* and *L. pseudoborneense*) were revealed. *Leccinum album* is closely related to *L. pseudoborneense* yet without statistical support, while *L. parascabrum* forms an independent lineage. Species to which *L. parascabrum* is phylogenetically related remain as yet unknown.

**Figure 2.** Maximum-likelihood phylogenetic tree generated from ITS dataset. BS > 50% in ML analysis is indicated above or below supported branches. Species of this genus from China and type species of this genus (*L. aurantiacum*) are indicated in bold. Voucher specimens, localities and GenBank numbers are provided behind the species names. AT = Austria, CAN = Canada, CN = China, CR = Costa Rica, FR = France, GL = Greenland, JP = Japan, NL = The Netherlands, PL = Poland, SW = Sweden, UK =United Kingdom, USA = United States of America and UZ = Uzbekistan.

> Our ML and Bayesian analyses of ITS and multi-locus datasets revealed the existence of eight *Leccinum* species from China, including five known species viz. *L. melaneum* (Smotl.) Pilát & Dermek, *L. quercinum*, *L. scabrum*, *L. schistophilum* and *L. versipelle* and three new species viz. *L. album*, *L. parascabrum*, and *L. pseudoborneense*. The final alignments of both datasets were deposited in TreeBASE (S27490).

#### *3.2. Taxonomy*

*Leccinum album* X. Meng, Yan C. Li & Zhu L. Yang, sp. nov., (Figures 3g–h and 4). MycoBank: MB 838917.

*Diagnosis*: This species differs from other species in *Leccinum* in the combination of the entirely white pileus, the white pileal context not changing color when injured, the white hymenophore staining indistinct greenish blue when hurt, the white stipe coarsely covered with initially white and then darkened verrucose squamules, and the white stipe context always staining greenish blue at the base when injured.

*Holotype*: CHINA. Hunan Province: Chenzhou, Zhanghua County, Mangshan National Forest Park, E 112◦920 , N 24◦940 , alt. 850 m, associated with *Castanopsis fissa*, *Cyclobalanopsis glauca*, *Lithocarpus glabra* and *Pinus kwangtungensis*, 3 September 2007, Y.C. Li1072 (KUN-HKAS53417, GenBank Acc. No.: MZ392872 and MZ392873 for ITS, MW413907 and HQ326880 for nrLSU, MW439267 and HQ326861 for *tef1-α*, and MW439259 and MW439260 for *rpb2*).

**Figure 3.** Basidiomata of *Leccinum* species. (**a**–**c**) *Leccinum parascabrum* (KUN-HKAS99903, holotype); (**d**–**f**) *Leccinum pseudoborneense* ((**d**) from KUN-HKAS110157; (**e**,**f**) from KUN-HKAS110156, holotype); (**g**,**h**) *Leccinum album* (KUN-HKAS53417, holotype).

*Etymology*: Latin "*album*" means white, referring to the color of the basidiomata. Basidiomata small to medium-sized. Pileus 3–5.5 cm in diam., hemispherical when young, subhemispherical to convex or plano-convex when mature, white (1A1) when young, white to cream (2B2–3) when mature; surface covered with concolorous farinose to pubescent squamules; context 5–10 mm thick in the center of pileus, taste mild, white (1A1) to pallid, not changing color when bruised; Hymenophore adnate when young, adnate to slightly depressed around apex of stipe; surface white (1A1), staining indistinct greenish blue (25B5–7) when injured; pores subangular to roundish, 0.3–1.5 mm wide; tubes up to 5 mm long, white to dirty pinkish (13A2), not changing color when bruised. Stipe 8–10 × 0.8–1.2 cm, clavate to subcylindrical, always enlarged downwards; surface white (1A1), densely covered with white (1A1) verrucose squamules, staining light greenish blue at base when injured; context whitish (1A1), staining blue at base when injured; basal mycelium white (1A1), lacking color change when injured.

**Figure 4.** Microscopic features of *Leccinum album* (KUN-HKAS53417, holotype). (**a**) Basidiospores. (**b**) Basidia and pleurocystidium. (**c**) Cheilocystidia and pleurocystidia. (**d**) Pileipellis. Bars = 10 µm. Drawings by Y.-C. Li.

Basidiospores (40/2/1) 15–19 × 5–7 µm, Q = 2.5–3, Q<sup>m</sup> = 2.75 ± 0.15, subfusiform to narrowly ellipsoid in side view with slight suprahilar depression, subcylindrical to fusiform in ventral view, smooth, somewhat slightly thick-walled (up to 0.5 µm thick), hyaline to yellowish in KOH, brownish yellow to olivaceous brown in Melzer's Reagent. Basidia 23–33 × 10–13 µm, clavate, 4-spored, hyaline to yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Hymenophoral trama boletoid, hyphae subcylindrical, 4–10 µm wide, hyaline to yellowish in KOH, yellowish to yellow in Melzer's Reagent. Cheilo- and pleurocystidia 42–60 × 11–17.5 µm, abundant, subfusiform to fusiform, thinwalled, yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Pileipellis a trichoderm, composed of more or less vertically arranged 5–10 µm wide hyphae, hyaline to yellowish in KOH, yellowish to yellow in Melzer's Reagent. Pileal trama made up of 6–12 µm wide filamentous hyphae, thin-walled, yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Clamp connections absent in all tissues.

*Habitat and distribution*: Solitary or scattered in tropical forests dominated by plants of the families Fagaceae (*Castanopsis fissa*, *Cyclobalanopsis glauca* and *Lithocarpus glabra*) and Pinaceae (*Pinus kwangtungensis* or *P. armandii*); on acidic, humid and loamy soils; distribution insufficiently known, rather rare in China and currently found in central and southeastern China (Hunan and Fujian Provinces).

*Additional Specimen examined*: CHINA. Fujian Province: Jianning County, E 116◦840 , N 26◦830 , alt. 900 m, associated with *Castanopsis fissa*, *Cyclobalanopsis glauca* and *Pinus armandii*, 16 July 1971, N.L. Huang 716 (KUN-HKAS39522).

*Commentary*: *Leccinum album* is characterized by the white pileus, the white hymenophore staining indistinct greenish blue when hurt, the white stipe densely covered with initially white and then darkened scabrous squamules, the white context in pileus not changing color when injured, and the white context in stipe unchanging or only staining distinct greenish blue at base when injured. Morphologically, *L. album* is close to *L. holopus*, *L. cyaneobasileucum* Lannoy & Estadès and *Le. albellum* (Peck) Bresinsky & Manfr. Binder in similar pileus colors. However, *L. holopus*, originally described from Europe (Germany), differs from *L. album* in its medium to large basidiomata (pileus 4–10 cm wide), becoming more viscid pileus with age, pure white or dirty white to pale buff or pale pallid pileus always with a glaucous green tinge, long hymenophoral tubes measuring 9–15 mm long, narrow and subcylindrical hymenial cystidia measuring 30–50 × 7.5–12.5 µm, narrow pileipellis hyphae measuring 3.5–5 µm wide, and association with trees of the genus *Betula* (Betulaceae) [80–82]. *Leccinum cyaneobasileucum*, originally described from France, is different from *L. album* in its white or greyish brown to light brown pielus, woolly stipe surface, slender basidiospores with Q<sup>m</sup> ≥ 3, relatively narrow hymenial cystidia measuring 32–44 × 5.5–7.5 µm, narrow pileipellis hyphae measuring 2–6.5 µm wide, and association with trees of the genus *Betula* [83]. *Leccinellum albellum*, originally described from New York, is characterized by its basidiomata not changing color when bruised and narrow basidiospores measuring 13–20 × 4–6 µm [16,17,30].

Phylogenetically, *L. album* is related to *L. variicolor* and *L. pseudoborneense* in the analyses of the multi-locus and ITS datasets, respectively (Figures 1 and 2). However, *L. variicolor* differs from *L. album* in its white to grey or cream pileal context staining vinaceous to brown when bruised, white stipe context staining pink to coral red in the upper part and greenblue in the lower part when bruised and association with plants of *Betula* [81]. *Leccinum pseudoborneense* is different from *L. album* in its pale brown to dark brown pileus, white context in pileus and stipe staining blue when bruised, narrow basidiospores measuring (11) 12–19 (20) × 4–5 (6) µm, narrow hymenial cystidia measuring 28–40 × 4–10 µm, and distribution in southwestern China.

*Leccinum parascabrum* X. Meng & Yan C. Li & Zhu L. Yang, sp. nov., (Figures 3a–c and 5). MycoBank: MB 838916.

*Diagnosis*: This species differs from other species in *Leccinum* by its initially reddish brown to chestnut-brown and then brown to pale brownish or even dirty white pileus, white pileal context lacking color change when injured, white to pallid and then light brown hymenophore lacking color change when injured, and the white stipe context staining greenish blue at the base when injured.

*Holotype*: CHINA. Hunan Province: Chenzhou, Zhanghua County, Mangshan National Forest Park, E 112◦920 , N 24◦940 , alt. 1100 m, associated with *Castanopsis fissa*, *Lithocarpus glabra* and *Pinus kwangtudgensis*, 12 September 2016, G. Wu 1784 (KUN-HKAS99903, GenBank Acc. No.: MZ392874 for ITS, MW413911 for nrLSU, MW439271 for *tef1-α*, and MW439264 for *rpb2*).

*Etymology*: Latin "*parascabrum*" refers to its similarity to *L. scabrum*.

Basidiomata small to medium-sized. Pileus 2.5–12.5 cm in diam., hemispherical when young, subhemispherical to convex or applanate when mature, reddish brown (12E8) to chestnut-brown (8C7–8) when young, brown (6C6) to pale brownish (7D7–8) or even dirty white (6A2) when mature; surface tomentose; context 6–13 mm thick in the center, white (1A1), not changing color when bruised; Hymenophore adnate when young, adnate to slightly depressed around apex of stipe; surface white to pallid (1A1) when young, and becoming light brown (6B4) when mature, not changing color when injured; tubes 6–14 mm long, 0.5–1.5 mm wide, creamy white (1A1), not changing color when bruised.

Stipe 12–14 × 1.1–2.2 cm, clavate, swollen downwards, always staining greenish blue at base when injured; surface white (1A1), covered with initially white (1A1) to light beige (5A4) and then brownish (7D8) squamules; context white (1A1), staining greenish blue (25B6–7) at base when injured; basal mycelium white (1A1).

**Figure 5.** Microscopic features of *Leccinum parascabrum* (KUN-HKAS99903, holotype). (**a**) Basidiospores. (**b**) Basidia and pleurocystidium. (**c**) Cheilocystidia. (**d**) Pleurocystidia. (**e**) Pileipellis. Bars = 10 µm. Drawings by Y.-C. Li.

Basidiospores (80/2/2) 16–20 (–21) × 5–6 µm, Q = 3.2–3.8, Q<sup>m</sup> = 3.43 ± 0.18, subfusiform to fusiform, slightly thick-walled (up to 0.5 µm thick), yellowish to yellowish brown in KOH, yellow to yellow-brown in Melzer's Reagent. Basidia 24–33 × 8–12 µm, clavate, 4-spored, hyaline to yellowish in KOH, yellowish to yellow in Melzer's Reagent. Hymenophoral trama boletoid, hyphae cylindrical, 3–7 µm wide, hyaline to yellowish in KOH, yellowish to yellow in Melzer's Reagent. Cheilo- and pleurocystidia 34–68 × 7.5–16 µm, abundant, subfusiform to fusiform, thin-walled, yellowish to pale yellowish brown in KOH, yellowish brown to brown in Melzer's Reagent. Pileipellis a trichoderm, composed of 5–9 µm wide filamentous hyphae, yellowish to pale brownish in KOH. Pileal trama made

up of 5–10 µm wide filamentous hyphae, thin-walled, hyaline to yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Clamp connections absent in all tissues.

*Habitat and distribution*: Solitary or scattered in tropical forests dominated by plants of the families Fagaceae (*Lithocarpus glabra, Castanopsis fissa* and Ca. *hystrix*) and Pinaceae (*Pinus kwangtudgensis* or *P. yunnanensis*.); on acidic or slightly alkaline, loamy soils; distribution insufficiently known, rather rare in China, currently known from central and southwestern China (Hunan and Yunnan Provinces).

*Additional Specimen examined*: CHINA. Yunnan Province: on the way from Tengchong County to Longling County, E 98◦590 , N 24◦810 , alt. 2010 m, associated with *Lithocarpus glabra*, *Castanopsis hystrix* and *Pinus yunnanensis*, 19 July 2009, Y.C. Li 1700 (KUN-HKAS59447, GenBank Acc. No.: MZ392875 for ITS, MW413912 for nrLSU, MW439272 for *tef1-α*, and MW439265 for *rpb2*).

*Commentary: Leccinum parascabrum* is characterized by the initially reddish brown to chestnut-brown and later brown to pale brownish or even dirty white pileus, the white pileal context not changing color when injured, the white to pallid and then light brown hymenophore not changing color when injured, the white stipe context with greenish blue color change at the base when injured, and the relatively large basidiospores measuring 16–20 (–21) × 5–6 µm, Q = 3.2–3.8. *Leccinum parascabrum* generally shares the similar colors of pileus and hymenophore, and the similar slender stems with *L. duriusculum*, *L. griseonigrum* A.H. Sm., Thiers & Watling, *L. scabrum*, *L. uliginosum* A.H. Sm. & Thiers and *Le. pseudoscabrum* (Kallenb.) Mikšík. However, *L. duriusculum*, originally described from Europe, can be distinguished from *L. parascabrum* by its pale grey-brown to dark greyish or reddish brown pileus, white context staining violaceous pink when bruised but yellow-green to blue-green in the base of stipe, relatively small basidiospores measuring 11.5–15.5 × 4.5–6 µm [84]. *Leccinum griseonigrum*, originally described from North America, differs from *L. parascabrum* in its avellaneous to dingy cinnamon-buff pileus, white pileal context staining blue when bruised, relatively small basidiospores measuring 13–16 × 4–5.5 µm, and association with trees of the genus *Populus* [16]. *Leccinum scabrum* differs from *L. parascabrum* in its wrinkled pileus, pale white hymenophore, pinkish discoloration when injured, and never bluish color change at the base of stipe [13,14,81]. *Leccinum uliginosum*, originally described from North America, is different from *L. parascabrum* in its dark fuscous to drab-grey pileus, white context in pileus becoming reddish and then fuscous when bruised, relatively small basidiospores measuring 14–18 × 3.5–5 µm, and small and inconspicuous hymenial cystidia [17]. *Leccinellum pseudoscabrum* differs from *L. parascabrum* in its initially red to purplish brown and then blackish brown context color change when injured, and the palisadoderm pileipellis composed of subglobose cells [14]. *Leccinum parascabrum* also shares the similar colors of pileus and hymenophore and the bluish color change at the base of stipe with *L. variicolor*. However, *L. variicolor* is different from *L. parascabrum* in its white to grey or cream pileal context staining vinaceous to brown when bruised, white stipe context staining pink to coral red in the upper part and green-blue in the lower part when bruised, relatively small basidiospores measuring (10) 13.5–17.5 (–20.0) × 5.0–6.5 (7.0) µm with Q = 2.4–3.1, and association with plants of *Betula* sp. [81]. In our phylogenetic analysis of the multi-locus and ITS datasets (Figures 1 and 2), *L. parascabrum* formed independent clades within *Leccinum*. It might represent a distinct section or subsection. However, formal change of the infrageneric division of this clade should await more molecular and morphological data from additional taxa. Species to which it is phylogenetically related remain as yet unknown.

*Leccinum pseudoborneense*X. Meng & Yan C. Li & Zhu L. Yang, sp. nov., (Figures 3d–f and 6).

**Figure 6.** Microscopic features of *Leccinum pseudoborneense* (KUN-HKAS110156, holotype). (**a**) Basidiospores. (**b**) Basidia and pleurocystidia. (**c**) Pleurocystidia. (**d**) Cheilocystidia. (**e**) Pileipellis. Bars = 10 µm. Drawings by Y.-C. Li.

#### MycoBank: MB 838915.

*Diagnosis*: This species differs from other species in *Leccinum* in its nearly glabrous and pale brown to dark brown pileus, white context in pileus lacking color change when injured, white context in stipe staining blue when bruised, initially white and then brown hymenophore not changing color when injured, white stipe covered with ochraceous to dark brown squamules, and trichodermal pileipellis composed of 3–6 µm wide interwoven hyphae.

*Holotype*: CHINA. Yunnan Province: Xishuangbanna, Menghai County, Bada Town, E 100◦120 , N 21◦830 , alt. 1900 m, associated with *Castanopsis calathiformis*, *Castanopsis indica* and *Lithocarpus truncatus*, 22 June 2020, G.S. Wang 947 (KUN-HKAS110156, GenBank Acc. No.: MZ412902 for ITS, MW413908 for nrLSU, MW439268 for *tef1-α*, and MW439261 for *rpb2*)

*Etymology*: Latin "*pseudo*" = false, "*borneense*" = *L. borneense*, "*pseudoborneense*" is proposed because this species is similar to the species *L. borneense* originally described from Malaysia.

Basidiomata small to medium-sized. Pileus 4–10 cm diam, subhemispherical to convex or plano-convex; surface nearly glabrous, viscid when wet, pale brown (6D6–C5) to dark brown (6F6–E5); context 5–10 mm thick in the center, white (1A1), not changing color when bruised; Hymenophore adnate to depressed around apex of stipe; white (1A1) to pallid when young and becoming brown (6B5) when mature, not changing color when injured. Tubes 4–10 mm long, creamy white (1A1) when young, and becoming brownish yellow (5C7–8) when mature, not changing color when bruised; pores fine, no more than 1 mm wide. Stipe 10–15 × 2.1–2.9 cm, clavate, always swollen downwards; surface white (1A1), covered with ochraceous (2B3–5) to dark brown (6E7) squamules, staining asymmetric blue (23E7) when injured; context white (1A1), staining blue (23E7) when injured; basal mycelium white (1A1).

Basidiospores (100/5/5) (11–) 12–19 (–20) × 4–5 (–6) µm, Q = (2.75–) 3–3.58 (–3.6), Q<sup>m</sup> = 3.31 ± 0.16, subfusiform to ellipsoid, slightly thick-walled (up to 0.5 µm thick), yellowish brown to olive brown in KOH, yellow-brown to dark olive-brown in Melzer's Reagent. Basidia 18–30 × 8–9 µm, clavate, 4-spored, hyaline to yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Hymenophoral trama boletoid, hyphae cylindrical, 3–6 µm wide, hyaline to yellowish in KOH, yellowish to brownish yellow in Melzer's Reagent. Cheilo- and pleurocystidia 28–40 × 4–10 µm, abundant, subfusiform to fusiform, thin-walled, yellowish to brownish yellow in KOH, brownish to yellow-brown in Melzer's Reagent. Pileipellis a trichoderm, composed of more or less vertically arranged 5–12 µm wide filamentous hyphae, yellowish brown to brownish in KOH, brown to dark brown in Melzer's Reagent. Pileal trama made up of 6–12 µm wide filamentous hyphae, thin-walled, hyaline to yellowish in KOH, yellowish to yellow in Melzer's Reagent. Clamp connections absent in all tissues.

*Habitat and Distribution*: Scattered in tropical forests dominated by plants of the families Fagaceae (*Castanopsis calathiformis*, Ca. *orthacantha*, Ca. *indica*, *Lithocarpus truncatus*, *Li. mairei* and *Quercus griffithii*); on acidic, loamy or mossy, humid soils; moderately common in southwestern China (Yunnan Province).

*Additional specimens examined*: CHINA. Yunnan Province: Xishuangbanna, Menghai County, Bada Township, E 100◦130 , N 21◦840 , alt. 1900 m, associated with *Castanopsis calathiformis*, Ca. *indica* and *Lithocarpus truncatus*, 22 June 2020, G.S. Wang 960 (KUN-HKAS110157, GenBank Acc. No.: MZ412903 for ITS, MW413909 for nrLSU, MW439269 for *tef1-α*, and MW439262 for *rpb2*), the same location, 22 June 2020, G.S. Wang 965 (KUN-HKAS110158, GenBank Acc. No.: MZ412904 for ITS, MW413910 for nrLSU, MW439270 for *tef1-α*, and MW439263 for *rpb2*); Nanjian County, Gonglang Town, Huangcaoping, E 100◦300 , N 24◦540 , alt. 1200 m, associated with *Castanopsis orthacantha*, *Lithocarpus mairei* and *Quercus griffithii*, 30 June 2015, K. Zhao 773 (KUN-HKAS92401, GenBank Acc. No.: MZ536632 for nrLSU, MZ543307 for *tef1-α*, and MZ543309 for *rpb2*); Jinghong County, Dadugang Town, E 100◦250 , N 21◦260 , alt. 600 m, associated with *Castanopsis indica* and *Lithocarpus truncatus*, 30 June 2014, K. Zhao 476 (KUN-HKAS89139, GenBank Acc. No.: MZ536631 for nrLSU, MZ543306 for *tef1-α*, and MZ543308 for *rpb2*).

*Commentary: Leccinum pseudoborneense* is characterized by the nearly glabrous and pale brown to dark brown pileus, the white context in pileus not changing color when injured, the white context in stipe staining blue when bruised, the initially white and then brown hymenophore not changing color when injured, the white stipe covered with ochraceous to dark brown squamules, and the trichodermal pileipellis composed of 3–6 µm wide interwoven hyphae. *Leccinum pseudoborneense* is similar to *L. borneense* (Corner) E. Horak, originally described from Malaysia, in that they share a brown pileus, bluish color change of the context in stipe when bruised, and similar size of basidiospores. However, *L. borneense* differs from *L*. *pseudoborneense* in its yellow to olive yellow hymenophore staining blue when hurt, pale yellow to yellow pileal context staining blue when hurt, and deep yellow

context in stipe staining blue but sometimes with reddish tint at base when injured [6,84]. *Leccinum pseudoborneense* is phylogenetically close to *L. album* in our phylogenetica analyses (Figures 1 and 2). However, *L. album* has a white basidioma, white hymenophore staining indistinctly greenish blue when hurt, white context in pileus not changing color when injured, white context in stipe unchanging or only staining distinctly greenish blue at base when injured, and relatively broad basidiospores measuring 15–19 × 5–7 µm.

#### **4. Discussion**

The genus *Leccinum* was defined and recognized variously by different mycologists. In an early molecular study, *Leccinum* was shown to be polyphyletic and proposed to be restricted to the sections *Leccinum* and *Scabra* by Binder and Besl [4]. Subsequently, Bresinsky and Besl [32] erected a genus *Leccinellum* Bresinsky & Manfr. Binder, to accommodate *L.* section *Luteoscabra*, including species with yellow hymenophores and/or context. In this study, the phylogenetic inferences based on the multi-locus dataset of nrLSU, *tef1-α* and *rpb2* largely coincide with those of Binder and Besl [4], Bresinsky and Besl [32] and den Bakker et al. [33]. Thus, we adopt the treatment of Bakker et al. [33] and treat *Leccinum* in a strict circumscription, which only includes species of *L.* sect. *Leccinum* (Singer's infrageneric classification with *L.* sect. *Scabra* merged to this section). Species in *Leccinum* are characterized by the white context lacking color changes or staining blue, gray or reddish tints when injured and the cutis-like pileipellis composed of interwoven filamentous hyphae

Eleven *Leccinum* species with specimen citations have been reported from China before this study, of which five species (*L. ambiguum*, *L. atrostipitatum*, *L. olivaceopallidum*, *L. potteri* and *L. subgranulosum*) were originally described from North America, five species (*L. aurantiacum*, *L. holopus*, *L. roseofractum*, *L. scabrum* and *L. versipelle*) were originally described from Europe, and only one taxon (*L. subleucophaeum* var. *minimum*) was originally described from China. Our molecular phylogenetic analyses along with morphological studies identified the existence of *L. quercinum*, *L. scabrum*, *L. subleucophaeum* var. *minimum* and *L. versipelle* in China. The distribution of other reported species have not yet been found, based on morphological and/or molecular data. In addition, three species new to science (*L. album*, *L. parascabrum* and *L. pseudoborneense*) and two species new to China (*L. melaneum* and *L. schistophilum*) were revealed in our study, based on molecular and morphology evidence. In conclusion, there are nine species of *Leccinum* in China.

Most species of *Leccinum* exhibit strong mycorrhizal host specificity. The host specificity along with climate type and edaphic factors appear to be important factors determining the distribution of different species. In China, *L. melaneum*, *L. scabrum*, *L. schistophilum* and *L. versipelle* are found in temperate forests and associated with plants of *Betula platyphylla* on acidic soils. *Leccinum album*, *L. parascabrum, L. pseudoborneense* and *L. subleucophaeum* var. *minimum* are found in tropical forests and associated with plants of Fagaceae (*Castanopsis calathiformis*, *Ca. hystrix*, *Ca. indica*, *Ca. orthacantha*, *Cyclobalanopsis glauca*, *Lithocarpus mairei*, *Li. truncatus* and *Quercus griffithii*) and/or Pinaceae (*Pinus kwangtudgensis* and *P. yunnanensis*) on acidic soils. It is noteworthy that *L. parascabrum* can be found in acidic or slightly alkaline habitats. *Leccinum versipelle* is found in subtropical forests and is associated with plants of *Populus yunnanensis* on acidic soils. The combination of the color of basidioma, the morphology of pileal surface, the size of basidiospores, the morphology of stipe, the color changes when injured, the climate type, the edaphic factors and the host preferences is very important in distinguishing species in this genus.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/jof7090732/s1, Alignment S1: alignment of multi-locus dataset; Alignment S2: alignment of ITS dataset.

**Author Contributions:** Conceptualization: Z.L.Y., Y.-C.L. and X.M.; field sampling: Z.L.Y., Y.-C.L., G.W. and G.-S.W.; molecular experiments and data analysis: X.M. and P.-M.W.; original draftwriting: X.M.; original draft—review and editing: Y.-C.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundation of China (Nos. 31872618, 31750001, 32070024, 31970015), the Natural Science Foundation of Yunnan Province (2018FB027), the Ten Thousand Talents Program of Yunnan (YNWR-QNBJ-2018-125), and the Key Research Program of Frontier Sciences, CAS (QYZDY-SSW-SMC029).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Publicly available datasets were analyzed in this study. This data can be found here: https://www.ncbi.nlm.nih.gov/; http://www.mycobank.org/; https://www. treebase.org/treebase-web/home.html, accessed on 26 August 2021.

**Acknowledgments:** The authors thank X. H. Wang, Z. W. Ge, B. Feng and Q. Cai, Kunming Institute of Botany, CAS, for providing samples and/or related literatures.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

