Three new Russula species in sect. Ingratae (Russulales, Basidiomycota) from southern China

Abstract Three new species of Russulasection Ingratae, found in Guizhou and Jiangsu Provinces, southern China, are proposed: R.straminella, R.subpectinatoides and R.succinea. Photographs, line drawings and detailed morphological descriptions for these species are provided with comparisons against closely-related taxa. Phylogenetic analysis of the internal transcribed spacer (ITS) region supported the recognition of these specimens as new species. Additionally, R.indocatillus is reported for the first time from China and morphological and phylogenetic data are provided for the Chinese specimens.


Introduction
Russula Pers. is a widespread genus that contains at least 2000, but possibly as many as 3000 species worldwide Adamčík et al. 2019;). Members of this genus form symbiotic relationships with a diversity of plant species in broad-leaved and coniferous forests, scrubland and meadows. The brightly tinged pileus, abundant sphaerocytes responsible for the fragile gills and stipe, amyloid spore ornamentation, gleocystidia staining in sulpho-aldehydes, lack of clamp connections and absence of a ramifying lactifer system ending in pseudocystidia are the main morphological features of this genus (Li et al. 2015a;Buyck et al. 2018;Looney et al. 2018). Due to frequent convergence or extreme plasticity of morphological features, precise identification of Russula species is difficult and establishing accurate taxonomy is challenging (Miller and Buyck 2002;Bazzicalupo et al. 2017).
Russula sect. Ingratae Quél. is characterised by tawny, ochraceous or ashy-grey to dark brown pileus with tuberculate striate margin, acute to subacute equal lamellae, flesh with a distinct fetid, spermatic or waxy odour, or like bitter almonds, cream-coloured spore print, spores partly showing inamyloid reaction in the suprahilar area, small-to medium-sized unicellular pileocystidia and articulated and branched hyphal ends in the pileipellis (Shaffer 1972;Romagnesi 1985;Sarnari 1998). The combination of these characters makes this section one of the more easily distinguishable groups in the Russula subgenus Heterophyllidiae Romagn. Recent multi-locus phylogenetic studies indicated that this morphologically well-defined group corresponded to the earlier subsections, Foetentinae, Pectinatinae and Subvelatae, representing a natural, well-supported monophyletic clade in phylogenetic topology of the northern temperate region (Looney et al. 2016;Buyck et al. 2018). The other easily distinguishable groups of subgenus Heterophyllidiae include subsections Amoeninae, Virescentinae and Substriatinae. Phylogenetic analyses also indicated it is more difficult to match a field aspect with a single monophyletic lineage (Wang et al. 2019;Deng et al. 2020;Wisitrassameewong et al. 2020).
Compared with Europe (Romagnesi 1985;Sarnari 1998), detailed analyses of Russula sect. Ingratae in Asia began relatively late. In southern China, several species were previously misidentified, based on morphological characters, with European or North American names, such as R. foetens Pers., R. grata Britzelm. (= R. laurocerasi Melzer) and R. pectinatoides Peck (Song et al. 2007;Li 2014). Rapid progress has been made in the past two decades, resulting in 15 new Russula species in Asian Ingratae, based on modern phylogenetic methods: R. abbotensis K. Das (Das et al. 2006(Das et al. , 2010(Das et al. , 2013Razaq et al. 2014;Li et al. 2015b;Jabeen et al. 2017;Lee et al. 2017;Song et al. 2018Song et al. , 2020Ghosh et al. 2020). The initial sequence data have supported the valid recognition of R. punctipes Singer and R. senecis Imai, but are still lacking for R. guangdongensis Z.S. Bi & T.H. Li and R. periglypta Berk & Broome (Lee et al. 2017;Song et al. 2018). Recent rDNA ITS phylogenetic analyses of R. sect. Ingratae in the Northern Hemisphere showed numerous unknown taxa and constant misidentifications of species in this group (Avis 2012;Melera et al. 2017;Park et al. 2017).
The importance of precise identification of Russula spp. in sect. Ingratae also comes from their economic value as several species are commonly sold as edible fungi in markets of southern China under the local name "You-la-gu (oily, acrid mushroom)". Several species of R. sect. Ingratae may cause gastrointestinal problems if not properly pre-cooked (Dai et al. 2010;Bau et al. 2014;Chen et al. 2016). During recent years, several field investigations have been carried out on campuses and, in parks, natural reserves and wild mushroom markets of south-western China to unveil the species diversity of sect. Ingratae in this region. A number of Russula taxa have been discovered as new to science, based on morphological and molecular phylogenetic evidence, of which three members of R. sect. Ingratae are described and illustrated herein. Additionally, we report R. indocatillus as a new record for China.

Morphological analyses
Specimens were collected in Guizhou, Jiangxi and Jiangsu Provinces from July to September in 2017 and 2018. The majority of the samplings are from Guizhou Province of southwestern China. This mountainous Province lies in the eastern end of the Yungui Plateau. This region has a humid subtropical monsoon climate and is mostly covered by subtropical evergreen forests (Editorial Board of Vegetation in China 1980;Chen et al. 2020). Each of the specimens was collected from different patches of forest to avoid duplications from a single mycelium. Photographs of fresh basidiocarps were taken using a Canon Powershot G1 X Mark II digital camera in the field. Macroscopic characters were recorded at the same time under daylight. The colour codes and names from Ridgway (1912) were employed in descriptions. Specimen desiccation was accomplished in a Fruit FD-770C food dryer at a constant temperature of 65 °C over 12 h. Small tissue pieces of lamellae and pileipellis for microscopic observations were taken from dried specimens, sectioned by hand with a Dorco razor blade and rehydrated in 5% potassium hydroxide (KOH). Microscopical characters were observed using a Nikon Eclipse Ci-L photon microscope and Olympus BH2 with a drawing tube. Staining of basidiospores, mycelia and cystidia were performed by chemical reaction with Melzer's Reagent and sulphovanillin (SV). Measurements and line drawings of basidiospores (exclusive of apiculus and spore ornamentation) and elements in hymenium, pileipellis and stipitipellis were executed from microphotographs taken at 1600× magnification with a Cossim U3CMOS14000 camera. A JSM-IT300 cold-field scanning electron microscope was used for examination of basidiospore ornamentation. At least 20 observation data were employed for each morphological character of every analysed collection. The format, α/β/γ, represented the numbers of basidiospores, basidiocarps and specimens that were measured microscopically. For those basidiospore dimensions, these were indicated as (a-) b-c (-d), the extremes of the measured values (a and d) are displayed in brackets. The values of b and c are 5 th and 95 th percentiles when observed readings were arranged from small to large. Q is the ratio of basidiospore length to width. The Q in bold is the mean value of Q plus or minus standard deviation. The pileipellis was vertically sectioned at the edge and centre of the pileus. Shapes and sizes of basidia, cystidia and hypha were observed, measured and illustrated. For other details on microscopic observation and measurement, see Li (2014) and Adamčík et al. (2019). Exsiccatae of these new species are preserved in the Macrofungus Section, Mycological Herbarium of Guizhou Academy of Sciences (HGAS-MF), Herbarium of Hebei Agricultural University (HBAU) and Herbarium of Fungi, Jiangxi Agricultural University (HFJAU).

DNA extraction, polymerase chain reaction (PCR) and sequencing
Tissue samples from dried specimens were ground in centrifuge tubes using abrasive rods attached to an electric drill. DNA extractions were performed using a modified CTAB method as in Li (2014). PCR reactions were carried out in a Dragonlab TC1000-G 96-well thermocycler. Sequences in the ITS region were amplified with primers ITS5 and ITS4 (White et al. 1990) using the reaction conditions of Li et al. (2019). PCR products were separated by electrophoresis on 1.2% agarose gels and stained with Biotium GelRed. The concentrations of extracted DNA and PCR products were determined by a ThermoFisher Scientific NanoDrop One spectrophotometer. Nucleotide concentration > 50 ng/μl was used as the criterion of a qualified PCR product for Sanger sequencing by GENEWIZ Inc. An ABI 3730XL DNA analyser and an Applied Biosystems Sanger sequencing kit were used following manufacturer's procedures by Biomed Gene Technology Company (Beijing, China).

Phylogenetic analyses
Bidirectional sequencing results were assembled with MegAlign in DNAStar LaserGene 7.1 (https://www.dnastar.com). Low quality nucleotide sites at both ends of the sequences were trimmed. All new sequences from this study were deposited in GenBank (http:// www.ncbi.nlm.nih.gov/nuccore/). The BLAST algorithm was used to search the similar sequences and for the new species. Table 1 contains closely matched ITS sequences of the new species (percent identities over 97%) retrieved from GenBank and UNITE (https:// unite.ut.ee/) databases. Sampling for the phylogenetic backbone of Russula section Ingratae referred to Melera et al. (2017), Park et al. (2017 and Song et al. (2018). These sequences were combined with those of the new species and aligned in Mafft 7.428 with L-INS-I strategy applied (Nakamura et al. 2018). Five sequences from species of the other sections of Russula subgenus Heterophyllidiae, R. cyanoxantha (Schaeff.) Fr., R. grisea Fr., R. heterophylla (Fr.) Fr., R. ilicis Romagn. and R. substriata J. Wang et al., were chosen as out-group. The matrix file was manually optimised using BioEdit 7.0.5 (Hall 1999) and deposited in TreeBASE repository with study ID S28207 (http://purl.org/phylo/treebase/phylows/ study/TB2:S28207?x-access-code=cda6b439c0eada24d5199bc264971fb5&format=ht ml). Phylogenetic analyses were executed using Bayesian Inference (BI), Maximum Likelihood (ML) and Maximum Parsimony (MP) methods. Bayesian analysis was performed in MrBayes 3.2.7a (Ronquist et al. 2012). Best evolutionary model selection was carried out with MrModeltest 2.4 operated on PAUP* 4.0a165 through Akaike's Information Criteria (AIC) calculation (Nylander 2004). The calculation of posterior probabilities (PP) parameters was performed through the Markov chain Monte Carlo (MCMC) algorithm. The sampling frequency of the trees was set as every 100 th generation. One cold and three hot Markov chains were run for 2 ´ 10 6 generations. The analysis ceased when the average standard deviation was maintained below 0.01. A percentage of 25% trees were discarded as burn-in before the construction of the 50% majority rule consensus tree. MP analysis was conducted in PAUP* 4.0a167 (Swofford 2004). The tree bisection-reconstruction (TBR) was carried out with a heuristic search. A total of 1000 replicates were set for bootstrap support (Felsenstein 1985). The setting of maxtrees was 5000. Branches collapsed when minimum length was zero. A Kishino-Hasegawa (KH) test (Kishino and Hasegawa 1989) was executed to determine whether trees were significantly different. The consistency index (CI), homoplasy index (HI), retention index (RI), rescaled consistency index (RC) and tree length (TL) were performed in MP analysis. ML analysis was performed in raxm-lGUI 1.5b3 with 1000 replicates (Silvestro and Michalak 2012). Trees were displayed and exported in FigTree 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/). Names of species in Fig. 1 and Table 1 were cited from source databanks. Definitions for clades and complexes were also presented in Fig. 1.

Phylogenetic analyses
A total of 112 ITS sequences (107 of sect. Ingratae and 5 of out-groups), including 13 newly-generated ones, were analysed in this study. The alignment for ITS phylogenetic analyses was composed of 543 characters including gaps. Of these characters in the matrix, 266 were variable, 201 were parsimony-informative, 65 variable characters were parsimony-uninformative. The parameters of MP analysis were CI 0.444, HI 0.784, RI 0.784, RC 0.348 and TL 869. The most suitable model for BI and MP analyses is GTR+I+G.
Distribution. China (Guizhou) and India (Uttarakhand). Notes. The Chinese collections fit well with the original description of Ghosh et al. (2020), except for a few differences. The Indian specimens have longer basidia, 35-60 × 9-11 μm. The original description of R. indocatillus also noted that the type specimen was collected in a temperate mixed forest with Myrica, Quercus and Rhododendron. The coniferous tree species in this habitat were not mentioned. The Chinese collection is from a subalpine coniferous forest of subtropical region dominated by Pinus spp. with the main undergrowth species of Berberis cavaleriei, Corylus yunnanensis, Elaeagnus umbellata and Rosa sweginzowii .
Habit and habitat. Single to scattered on yellow brown soil in coniferous forest dominated by Pinus armandii and P. massoniana at 1100-1400 m altitude.
Notes. This new species is similar to R. pseudopectinatoides in its brownish-yellow pileus, slightly acrid taste, cream spore print, spores with low, subreticulate ornamentation and gelatinous pileipellis. It is notable that basidiomata of R. subpectinatoides were collected from a forest of introduced coniferous tree species. Cedrus deodara is native in the western Himalayas, while Pinus parviflora and P. thunbergii are naturally distrib-uted in the Japanese archipelago and Korean peninsula. Therefore, this new taxon may also occur in these introduced areas with its accompanying trees.
The Asian species of sect. Ingratae already recognizable by their long slender stipe, such as R. gelatinosa, R. guangdongensis, R. punctipes, R. senecis, R. subpunctipes and R. tsokae and cannot be confused with our new species, even more so because they have basidiospores composed of long wings, 2 μm high or more (Song et al. 2018(Song et al. , 2020. A similarly-winged spore ornamentation also differentiates species of the R. grata lineage which, moreover, usually have a distinct bitter almond smell. The more golden yellow pileus of species in the R. foetens or R. subfoetens lineages also avoids confusion with our new species and because many of these are distinctly very acrid. The strong yellowish stipe base that turns immediately red with KOH easily allows one to distinguish the few species of the R. insignis lineage. In the R. granulata lineage, the Asian species R. rufobasalis has reddish tinged stipe base, pleurocystidia with mucronate or appendiculate apices and longer terminal cells, up to 60 μm (Song et al. 2018). Finally, the typically very acrid taste allows us to eliminate most species of the R. amoenolens lineage, notwithstanding their sometimes quite similar colouration. The same very acrid taste also differentiates R. obscuricolor, which was described from the Indian Himalayas ) and showed close affinity to some Southern Hemisphere Ingratae in our phylogeny.

Discussion
The modern taxonomy of Russula calls for a combination of detailed microscopic observations with universal and specific standard, multi-gene phylogenetic analyses and accurate symbiotic plant species information Adamčík et al. 2019). The ITS phylogenetic analyses are the most common for practical identification of Russula species, because ITS is regarded as an adequate single gene DNA barcode for this genus ) and it has the largest number of available referential sequences in open databases (Schoch et al. 2012). A combination of morphological and ITS phylogenetic analyses supported the three new species amongst Asian Ingratae: R. straminella, R. subpectinatoides and R. succinea. The results of this study also indicate that R. indocatillus may have a wider distribution, from the Himalayan region to south-western China. The four species discussed here have distinct morphologies that allow each one to be differentiated from the others: • R. subpectinatoides and R. indocatillus possess the more or less inflated, shortcelled chains of hyphal ends, typical for most species in the subgenus Heterophyllidiae (Figs 4, 5, 8 and 9). These are abundant in R. subpectinatoides, but less so in R. indocatillus and absent in both other species which possess very dense, intricate and strongly branching, narrow ends in the pileipellis, more or less cemented in mucus that make microscopic examination of these hyphal ends very difficult. Compared to R. straminella, hyphal ends in the pileus centre of R. succinea have a more wavyundulate form (Figs 6, 7, 10 and 11).
All four species have similar pileocystidia, but in R. indocatillus, they are smaller overall at the pileus surface compared to the other three species (Figs 4 and 5), while in R. straminella, they are often more or less thick-walled (Figs 6 and 7).
When comparing basidiospores, R. subpectinatoides stands out because of the low subreticulate ornamentation (Figs 3 and 8), whereas the other species have more developed, higher warts or ridges that are much less interconnected, while R. indocatillus has almost completely isolated warts (Figs 3 and 4).
Some European members of section Ingratae, viz. R. amoenolens Romagn., R. pectinata Fr., R. pectinatoides Peck and R. sororia (Fr.) Romell may have been confused morphologically with some of these new species (Wu 1989;Ying and Zang 1994), but more recent diversity analyses indicated that some Chinese specimens, identified as R. amoenolens and R. insignis Quél., have broad morphological similarities, but also considerable difference (ca. 2%) in the ITS sequence compared to European samples of these species (Li 2014;Liu et al. 2017;Cao et al. 2019). Whether these Chinese specimens represent unknown taxa or intraspecific geographically-separated populations is still debatable (Wang 2020). The factual presence of these species of European and North American origin in China have been analysed in recent years (Li 2014;Zhang 2014;Wang 2019;Liu 2019) and symbiotic host plants were found to be very similar between north-eastern China, Europe and North America (Wu 1979).
The topology of the ITS phylogram ( Fig. 1) in this study largely corresponds to that of Park et al. (2017). Of the three subsections in sect. Ingratae, the majority of subsect. Pectinatinae Bon (type species R. pectinata) with species that are typically more greyishbrown to greyish-cream is distributed over clades C and H (Bon 1988), while Subvelatae (Singer) Singer (type species R. subvelata Singer) with members that have velar rudiments consisting of loosely, arachnoid-pulverulent floccons on pileus surface (Singer 1986), forms the highly-supported clade I. The species R. indocatillus, newly-recorded from China in this study, is located in Clade H. This well-supported clade also contains the R. amoenolens complex from Europe and R. cerolens and allies from North America. The African species complex of R. oleifera Buyck in subsect. Oleiferinae Buyck (type species R. oleifera Buyck) with species that sometimes present an annulus, corresponds to Clade D (Sanon et al. 2014). This clade was a sister clade to the remainder of sect. Ingratae in the multilocus phylogenetic analysis of Buyck et al. (2018). The large majority of European species that cluster around R. foetens compose clade F, a clade highly supported by Bayesian analysis only. The latter clade is typically composed of yellowish-brown to orange brown species and roughly corresponds to species traditionally placed in subsect. Foetentinae (Melzer & Zvara) Singer (type species R. foetens), of which it is characterised by dull, ochraceous or pallid coloured pileus, often with pectinate-sulcate to tuberculate-sulcate and distinctly subacute to acute margin, context odour of nitrobenzene, oily, fish, iodoform, or of other unpleasant smells (Singer 1986). Clade F also contains two of our new species, R. straminella and R. succinea, which share a similar pileipellis structure. Clade E received higher support in ML and MP analyses and shared with Clade F that two of the three species were also yellowish-to orange brown. This clade harbours three species: R. rufobasalis from Asia and the North American R. granulata Peck and R. ventricosipes. The results of our phylogenetic analyses, based on ITS sequences, indicate that more unknown subsections may exist in sect. Ingratae. More complex multi-gene analyses are urgently needed to clarify the phylogenetic relationships amongst species in this section.
Compared with previous analyses (Melera et al. 2016;Lee et al. 2017), more gasteroid species of sect. Ingratae were included in our study. The majority of gasteroid taxa clustered as two branches in Clade F. The other gasteroid species were mainly scattered in clades of agaricoid taxa. The phylogenetic topologies and low supported branches within sequestrated complex 2 may indicate an urgent need to study the type material of these gasteroid species for clarification of synonyms. Lee et al. (2017) summarised the general patterns observed for spores in the four clades of sect. Ingratae by showing a trend for basidiospore size to increase, while the shape changes from ellipsoid to spherical and for species that have smaller spores to have more ellipsoid spores and vice versa. However, these patterns were less clear when gasteroid species of this section were taken into account (Table 2). These gasteroid species suggest that the patterns, proposed in Lee et al. (2017), do not fit well with all members of the sect. Ingratae. Gasteroid taxa are known to have typically more globose and larger spores, because there are no evolutionary pressures of asymmetrical spores with hilar appendages for ballistospory in agaricoid species (Wilson et al. 2011). According to statistics, exceptions that do not follow these general patterns are common in sect. Ingratae. Over 40% (5/11) of counted gasteroid species of this section have sub-globose to broadly ellipsoid, even ellipsoid spores. In simple terms, a significant portion of gasteroid species have larger, but still more ellipsoidal spores. The authors suggested that these exceptions may be ascribed to the multiple and irreversible evolutions of gasteromycetation (Miller et al. 2001;Hibbett 2007). Ancestor genotype, divergence time and environmental factors all may exert different influences on this phenotype.
Spore ornamentations consisting of winged ridges are regarded as one of the most distinctive morphological characters for some members of sect. Ingratae. These species include R. grata, R. fragrantissima and R. illota from Europe and northern China, R. mutabilis from North America, R. gelatinosa, R. punctipes, R. subpunctipes and R. senecis from eastern and southern Asia. A majority of these species and R. foetens formed a not highly supported clade in phylogenetic analyses of Lee et al. (2017). As more samplings and species of sect. Ingratae were involved, the monophyly of winged-spore species was not supported in this analysis. Close phylogenetic relationships were detected in strongly-supported clades of R. grata-R. fragrantissima, R. mutabilis-R. illota and R. punctipes-R. subpunctipes-R. senecis. This phylogenetic inconsistency called for a further multi-gene analysis.
The habitats of the four species of this study show a common feature of coniferous forests dominated by Pinus spp. The current altitudes of distributions of R. indocatillus and R. succinea indicate a habitat of subalpine climate. These two species may have wider distributions than current records because the corresponding ectomycorrhizal symbiotic trees are representative and widespread species in Sino-Japanese and Sino-Himalayan floral subregions (Wu 1980;Chen et al. 2020). For R. straminella and R. subpectinatoides which were collected from reforested plantations and transplanted botanic gardens, intensive samplings on initial areas of symbiotic trees are needed for clarifying the types of habitats.