An overview of Agaricus section Hondenses and Agaricus section Xanthodermatei with description of eight new species from Pakistan

In a recent revision of the genus Agaricus, A. section Xanthodermatei was split into two sections A. sect. Hondenses and A. sect. Xanthodermatei. Our objectives were to investigate the species diversity of both sections in Pakistan and to give an overview of the major clades. Phylogenetic analyses based on the combined nucLSU, ITS and TEF1 dataset from 35 specimens of both sections revealed three major clades. Analyses based on ITS dataset and 106 specimens, including 33 from Pakistan, reveal eight new species and one new record species. These nine species are described in detail. It is noteworthy that intraspecific variability as well as interspecific variability between closely related species were very low in ITS sequences in many cases. In the case of the two new species A. xanthochromaticus and A. griseovariegatus, TEF1 sequence data were much more efficient than ITS to distinguish these species from each other. The other new species are A. atroumbonatus, A. fumidicolor, A. macropeplus, A. parviniveus, A. swaticus and A. bambusetorum. The latter is the only new species of A. sect. Hondenses in which it is morphologically atypical and also the unique (sub)tropical species. Agaricus gregariomyces is recorded for the first time in Pakistan. In addition, brief descriptions are provided not only for A. bisporiticus, A. endoxanthus and A. punjabensis, which are reported again in Pakistan, but also for A. californicus, which is reported for the first time in Spain and outside North America. In total 12 species of both sections were reported in Pakistan and half of them were from subtropical climatic areas, underlining the contribution of the climatic diversity to the high species richness in this country.

www.nature.com/scientificreports/ Morphological observations. Specimens were photographed at collection sites. Collected basidiomata were vouchered and dried by a fan heater. Macroscopic characters were noted such as color, size and shape of pileus, stipe and lamellae, generally following Largent 32 and Chen et al. 33 . The fresh sporocarp characters were carefully noted in the field viz. discoloration upon bruising, odor and Schaeffer's reaction. Color notations were indicated from the Munsell's Soil Color Charts 34 . For microscopic observations, slides were prepared in 5% aqueous KOH (w/v) and then stained in 1% aqueous Congo red solution (w/v). Microscopic features like, size and shape of basidiospores, basidia, cheilocystidia, pileipellis, stipitipellis and hyphae of lower surface of the annulus were measured using a light microscope (MX4300H, Meiji Techo Co., Ltd., Japan). Measurements of anatomical features (basidiospores, basidia and cheilocystidia) were presented based on at least 30 measurements. The dimensions of basidiospores are given in the form of (a) b-c (d) × (e) f-g (h), [avX, Qm, n = i × j] where b-c and f-g include the spore length and width respectively between the 5th percentile and the 95th percentile, (a) and (d) the extreme values of spore lengths recorded, (e) and (h) the extreme values of spores width recorded, avX the mean of length by width ± SD (standard deviation), Qm the mean of Q coefficient (length/width ratio), n is the total number (i) of spores measured of each collection and (j) number of total collections measured. Measurements of other microscopic structures (basidia, cheilocystidia, pileipellis, stipitipellis and hyphae of lower surface of annulus where the piece of annulus was available in dried samples) include the range between the extreme values measured in length and width. Basidia were measured excluding sterigmata.
Sampling for phylogenetic analyses and for comparisons between ITS sequences. For phylogenetic analyses, we used sequence data of 107 specimens of 72 species or putative species of A. sect. Xanthodermatei or A. sect. Hondenses of which 59 are named species. All 72 species except the eight new species described in this study and A. sp./LAPAM110 have been included in previous phylogenetic analyses 4,7,9,10,[16][17][18][19][20][35][36][37] . For the multi-gene analysis, 35 of the 107 specimens were used. They belong to 32 species of which 29 are named species. For A. malangelus Kerrigan, a representative from China with complete multigene dataset was used. However, we preferred to include the type specimen from North America to represent this species in phylogenetic analyses based on ITS dataset. Consequently, in the latter analyses, we used only 106 of the 107 samples of both sections listed in Table 1. In addition to the 107 specimens mentioned above, six specimens from six species belonging to different sections known to be closely related to A. sect. Xanthodermatei and A. sect. Hondenses, were used as outgroup. Taxon names (sections, subsections and species), specimens, and all GenBank accession numbers are indicated in Table 1.
To compare the variability within species and between very closely related species, ITS sequences of 28 specimens were aligned. Their GenBank accession numbers are in Table 1 GenBank accession numbers: AJ418776,  AY484689, AY899271, AY899272, DQ182529, DQ182534, DQ185563, DQ185564, DQ185565, KJ609510,  KM657923, KM657925, KT824788, KT824789, KX098651, KX098652, KX098653, MH752460. ITS sequence data of ZRL2012474, ZRL2012582, and ZRL2012616 were used in Zhou et al. 22 where these three samples were neither described nor identified. They are incorrectly identified as A. deardorffensis in GenBank (KM657893, KM657894, and KM657896 respectively). In addition, although the ITS sequences of ZRL2012474 and ZRL2012582 are highly similar, their translation elongation factor 1-α gene (TEF1) sequences available in GenBank differ at 13 positions. We included these two samples in our Table of comparison of ITS sequences but until these data would be confirmed we prefer not to include them in our phylogenetic analyses. www.nature.com/scientificreports/ DNA extraction, PCR and sequencing. DNA from dried specimens was extracted using a modified 2% CTAB protocol as described by Zhao et al. (2011). PCR amplification of the nuc rDNA internal transcribed spacer region (ITS1-5.8S-ITS2 = ITS), nuc rDNA large subunit fragment (LSU) and translation elongation factor 1-α gene fragment (TEF1) was performed using the combination of primers ITS1F/ITS4 38,39 , LROR/LR5 40  and EF1-983F/EF1-1567R  Taxon-specific ITS markers. Markers characterizing species, clades, sub-sections or sections were identified in the alignment of 106 ITS sequences of specimens belonging to A. sect. Xanthodermatei or A. sect. Hondenses. These markers are indicated by uppercase letters and are given along with flanking sequences. IUPAC codes, such as Y, do not indicate ambiguity (C or T) but heteromorphism (C and T) likely reflecting allelic polymorphism in the heterokaryotic (n + n) basidiomata. Insertions or deletions (indel) are indicated in square brackets. The places of the markers are numbered according to their position either in the 5'-3' ITS1-5.8S-ITS2 DNA sequence (beginning by tygaatt) of the characterized species, or in the alignment deposited in TreeBASE for the characterization of clades or higher taxa.

Results
Phylogenetic analyses. For phylogenetic analyses, 57 sequences were newly generated (35 ITS, 11 LSU, and 11 TEF1) from 37 specimens. Among these, 33 specimens belong to 12 species from Pakistan including eight new species described in this study, one specimen was from Dominican Republic (LAPAM110 as A. sp.), and three from Europe: LAPAG608 (A. californicus), LAPAG611 (A. biannulatus) and LAPAG259 (A. pseudopratensis). ITS sequences of the latter two specimens were previously obtained, while the 35 remaining specimens were included for the first time in phylogenetic analyses. All samples with their GenBank accession numbers are listed in Table 1.
The alignment for multi-gene analyses consisted of sequences (ITS/LSU/TEF1) of 40 specimens and 2099 characters. The percentage of missing data was 6% (7/120) due to seven non-obtained sequences. The five samples used as outgroup belong to five different species of four different sections: A. sect. Bivelares (A. bisporus and A. cupressicola), A. sect. Agaricus (A. campestris), A. sect. Crassispori (A. variicystis) and A. sect. Trisulphurati (A. trisulphuratus). The 35 remaining samples belonged to 32 species or putative species of A. sect. Xanthodermatei or A. sect. Hondenses (Table 1).
The alignment for ITS dataset consisted of sequences of 108 specimens and 709 characters. Two samples of species belonging to A. sect. Agaricus were used as outgroup (A. campestris and A. langei), while the 106 remaining sequences belonged to 72 species or putative species of A. sect. Xanthodermatei and A. sect. Hondenses.
For both multi-gene and ITS analyses, the topologies of the trees using the Maximum likelihood (ML) or the Bayesian inference (BI) methods were highly similar. The bootstrap support values (BS) and Bayesian posterior probabilities (PP) are reported in the ML phylograms of Fig. 1 for multi-gene analyses, and Fig. 2 for ITS analyses.
Despite we completed multi-gene sequence dataset for many new samples and some previous samples it remains many species only having ITS sequence data. For this reason we present first the phylogeny of the sections and major clades based one multi-gene-analyses. The ITS analysis remains useful mainly at species level to circumscribe new taxa and identify their most closely related taxa.
Comparing the ITS barcode sequences of 72 species or putative species of clades Xan I, Xan II and Xan III, we found three clade-specific markers. Two markers are indel mutations (insertion or deletion), which are sufficient to characterize the three clades: www.nature.com/scientificreports/ crgaacCryttgy@590, the nucleotide C characterizes Xan I. In the current stage of knowledge, these molecular clade-specific markers at ITS positions 148, 590 and 684-690 allow to classify the species in the three clades Xan I, Xan II and Xan III and comfort the existence of these clades. These clades were revealed in multi-gene phylogenetic analyses despite the used algorithms did not take indel in consideration.
Placement of 33 specimens from Pakistan in the phylogenetic tree. The 72 species or putative species included his study were distributed in sections or major clades as follows: eight in A. sect. Hondenses (clade Xan I) and 64 in A. sect Xanthodermatei; 16 in Xan II and 48 in Xan III; 1, 11, and 36 in Xan III A, Xan III B, and Xan III C, respectively.
Six of the 33 specimens newly collected in Pakistan belong to four previously described species of A. sect. Xanthodermatei in the ITS tree (Fig. 2). Two of the six (KH72 and PU252) are grouped in Xan II. The ITS sequence of the specimen KH72 is identical to that of the type specimen of A. gregariomyces known from China and is the first report of this species in Pakistan. The ITS sequence of the specimen PU252 is identical to that of the type specimen of A. bisporiticus, a species already known from Pakistan and Thailand 18 . The four remaining specimens match with two species in clade Xan III C. The sequences of the three specimens PU265, L11 and CM200 are identical or differ at one position from the type specimen of A. punjabensis, a species previously described from Pakistan. The specimen J8 has a sequence that differs at one position from the sequences of the cosmopolitan species A. endoxanthus. It is not rare to find one or even more differences between sequences of specimens of species such as A. punjabensis and A. endoxanthus that exhibit large genetic variability with numerous heteromorphisms and polymorphic positions 16 . Phylogenetic data support the belonging of these six new specimens to four species previously reported from Pakistan.
The 27 remaining specimens from Pakistan are grouped in eight clades in the ITS tree (Fig. 2), which represent eight newly described species in this study. Five of these clades are well supported (BS > 80 and PP > 0.95). The clade of the new species A. swaticus is relatively well supported (BS = 86 and PP = 0.87) and sister to A. langensis   35 . We note that despite the ITS sequences of all three specimens of A. swaticus differ at only two positions from the sequence of A. langensis, the divergence node between the two species is well supported. The two remaining clades correspond to the two sister species A. griseovariegatus and A. xanthochromaticus. Their ITS sequences differ at only one position. Because such a low level of divergence between species is infrequent in the genus Agaricus, we included two specimens of each of these two species in the multi-gene analysis. In the multi-gene tree, the clades corresponding to A. griseovariegatus and A. xanthochromaticus are well supported and have longer branches due to differences at six additional positions in TEF1 sequences and one in LSU sequences. In addition, the node of divergence between these two species and, more generally, between the eight new species and their closest relative are well supported except for A. macropeplus. Indeed, the latter is not clearly related to any species in the ITS tree but appears closely related to A. karstomyces in multi-gene tree. Inter-and intra-specific variability. Among ITS sequences of 27 specimens belonging to the eight new species described in this study, and also among LSU and TEF1 sequences of four specimens belonging to two of these species, we noted a quasi-absence of intraspecific variability. The only exception is A. parviniveus, in which four heteromorphisms were found in samples B40 and PU257, at three and one positions of their ITS sequences, respectively. In addition, we noted a low level of genetic divergence between closely related species such as A. griseovariegatus and A. xanthochromaticus, which differ at only one position of their ITS sequences. Based on these preliminary observations of a low variability among ITS sequences of our new specimens, we extended this investigation to other species of A. sect. Xanthodermatei. On the one hand, we examined the intraspecific variability among numerous available ITS sequences of A. xanthodermus, the type species of A. sect. Xanthodermatei. On the other hand, we examined the inter-and intra-specific variability of the ITS region and of the spore size in a group of ten closely related named or putative species in the clade Xan III C, which includes three of our new species (A. griseovariegatus, A. macropeplus and A. xanthochromaticus).
To know if the high genetic homogeneity observed in our new species also applies to A. xanthodermus, which is probably the most represented species of the section in GenBank, we retrieved 30 ITS sequences having been deposited as A. xanthodermus. We first discarded 12 sequences of misidentified specimens which belong to more or less closely related species such as for example three sequences of specimens from China (JX434654, EU326208 and EU273511), which are 100% identical to sequences of A. sinoplacomyces Callac & R.L. Zhao and differ at 8 positions from A. xanthodermus. A second example consists of two samples from Ethiopia (KP229414 and KP229415; Sitotaw et al. 48 ), which are closely related to the tropical species A. volvatulus Heinem. & Gooss.-Font. (1-3 differences) described from Africa and which differ at 11 positions from A. xanthodermus. We found that the 18 remaining sequences are 100% identical except VM101/KT824788, which differs at one position (T insertion). These 18 samples are from Europe, Iran, USA or Canada. Therefore, based on ITS data, A. xanthodermus is a highly homogeneous species although it is distributed in different continents. We also note that various species of the section are frequently misidentified as A. xanthodermus.
The alignment of 28 ITS sequences belonging to ten closely related species revealed 18 variable nucleotide sites. Molecular characters at these 18 positions are indicated in Table 2. In this table, the 28 sequences are arranged as far as possible in the order in which they would have evolved from the most consensual sequence of A. deardorffensis and following the most parsimonious sequence of events. This order agrees with the order in which these species appear in the ITS tree where the position of A. deardorffensis is ancestral-like relatively to this group of species. A similar order is observed for the six species of this group that are represented in the multi-gene tree. In this group of species, the intraspecific variability is quasi-absent except in A. leptocaulis and except two heteromorphisms found in a sample of A. tibetensis and in a sample of A. placomyces Peck. In addition, the number of differences between closest related species is very low since it varies from 1 to 3.
The spore sizes of the seven named species of this group are compared in the Fig. 3. For the five species known from Asia, the spore sizes were measured on the same 18 specimens that were used in the Table of comparison of ITS sequences, while only mean spore sizes reported in Kerrigan (2016) are represented for the two species from North America. Except for A. tibetensis, of which the mean spore size appears higher than those of the other species (5.75 × 3.80 µm) but to a lesser extent than the value (6.8 × 4.2 µm) reported by Zhou et al. 22 , spore sizes are not useful to reliably distinguish the species of this group from each other. Original description: Pileus 2.5-3.5 cm in diam., hemispheric to convex, surface white (5.8 PB 9/2.9), smooth (scales not observed on any specimen), shiny (more pronounced in NWL322). Margin entire in young sporocarps, splitting radially at maturity, appendiculate, not exceeding the lamellae. Lamellae 3-5 mm in diam., pinkish-brown (2.1R 5.6/2) to dark brown (7.5YR 3/1.9), free, crowded, intercalated with lamellulae and with entire edges. Stipe 3-9.5 × 0.5-1 cm, cylindrical with bulbous base (1.5-2.5 cm wide), usually curved in the lower half, fistulose, provided with annulus in its upper part, above and below the annulus smooth and white (4.6 PB 8.7/3) with brown (6.5R 6.7/0.9) tinge with age and becoming brown immediately when rubbed/touched. Annulus superous, simple, narrow, membranous, fragile, white (4.9 PB 9/3.1) and smooth in both sides. Context when cut white, with a faint yellow discoloration appearing slowly. Odor faint, typical mushroom-like, pleasant.
Basidiospores  Table 1. Those of the nine remaining samples are given in Material and Methods. b Variable characters are numbered according to their positions in an alignment including the 28 ITS sequences used in the table.  www.nature.com/scientificreports/   Notes: Morphologically, in A. sect. Hondenses, only A. hondensis sometimes has a white pileus surface but this species is very different from A. bambusetorum by having larger basidomata (pileus 8-15 cm), a broad annulus, a yellow discoloration of the context, a strong odor of phenol and smaller spores (5.1 × 3.4 µm on average). Many species in A. sect. Xanthodermatei also have a white pileus surface but their annulus is always membranous, thick and complex with a squamose lower surface or a double or triple edged margin. One among these, A. pseudopratensis (Bohus) Wasser, can sometimes have an annulus with simple margin but it differs by having a more robust habit, a context first discoloring pale yellow becoming blood red with time, an odor of phenol and predominantly tetrasporic basidia. Agaricus parviniveus (described in this study) also has a white pileus surface and predominantly bisporic basidia but it differs by having a squat habit, a triple edged annulus margin and shorter spores. Phylogenetically, the most closely related species A. biannulatus    Etymology: The specific epithet is a compound Latin word where "atro" means dark colored and "umbonatus" refers to the umbonate shape of the pileus.
Basidiospores (4.0-) 4.5-5 (-5.5) × (2.5-) 2.7-2.9 (-3.2) µm, [avX = 4.8 ± 0.58 × 2.8 ± 0.21 µm, Q m = 1.72, n = 2 × 30], ellipsoid to ellipsoid-elongate, light to dark brown in KOH, smooth with a prominent apiculus, without apical pore and with granular content. Basidia 10-17 × 5.5-8 µm, broadly clavate or slightly truncate at the apex, hyaline in KOH, bisporic to tetrasporic. Cheilocystidia absent. Pleurocystidia absent. Underside of the annulus consisting of two types of hyphae some cylindrical, not or slightly constricted at the septa, composed of elongated elements, 3-13 µm in diam., terminal elements with rounded ends; other consisting of globose to subglobose, ovoid or slightly clavate elements, narrow at septa, up to 17 µm wide. Pileipellis consisting of cylindrical hyphae 3-13.5 µm in diam., frequently septate and branched, very few subglobose cells also observed, the wider the most constricted at the septa, in KOH some hyaline and some with an internal brown vacuolar or diffuse pigment, terminal elements with rounded ends. Stipitipellis hyphae 3.5-10.5 µm in diam., cylindrical hyphae mixed with broader hyphae, parallel, hyaline in KOH, branched.  Fig. 2). Morphologically A. daliensis only differs from A. atroumbonatus by exhibiting yellow discoloration at the stipe base when cut. In addition, the ITS sequences of these two species differ at eight positions. To a lesser extent, Agaricus atrodiscus L.J. Chen, Callac, R.L. Zhao & K.D. Hyde is also phylogenetically related to A. atroumbonatus, but this species is tropical and differs by having an strong odor of phenol, bigger spores on average ( Original description: Pileus 8-11.5 cm in diam., hemispherical then convex with depressed centre at maturity, entire and blackish grey (5YR2/2) at disc, outside the disc densely covered with concolorous punctiform squamules progressively scattered towards the margin, giving a smoky appearance to the pileus, on a white background. Surface dry and dull. Margin entire, slightly splitting at maturity. Lamellae 3-6 mm in diam., at first light brown, then dark brown (7.5YR1/2), free, crowded, with intercalated lamellulae and with entire edges. Stipe 8.4 × 1.4-1.7 cm wide at maturity, cylindrical, slightly bulbous at base when mature, sometimes curved, solid, provided with an annulus in its upper part, above the annulus smooth and white, below smooth, light greyish brown in its lower half and white in the extreme base, rhizomorphs present. Annulus superous, broad, fragile, upper side smooth, underside with perimarginal linear dark grey squames (apparent in young basidiomata), www.nature.com/scientificreports/ Notes: As for the species described above (A. atroumbonatus), an unequivocal morphological differentiation with all taxa of clade Xan III is not possible. In the ITS tree (Fig. 2), A. fumidicolor clusters with three other tropical or subtropical species in a well-supported clade (BS = 99%, PP = 1). In this clade A. fumidicolor is sister to A. punjabensis and is also related but to a lesser extent to A. endoxanthus and but these two species differ from  , conical in youngsporocarps, then truncate conical, completely covered with greyish brown scales radially arranged. Surface dry and dull. Margin slightly recurved not exceeding the lamellae. Lamellae free, crowded, intercalated with lamellulae, dark brown. Stipe 6-10.5 × 0.5-1.5 cm, clavate to cylindrical slightly bulbous at base, stuffed, provided with rhizomorphs at the base and an annulus in its upper third, smooth to slightly fibrillose and white above and below the annulus, discoloring light brown when touched towards the base. Annulus superous, double, white, flaring, broad, upper side smooth, underside floccose, with a double-edged margin. Context discoloring faint yellow when bruised. Odor of phenol, mild.
Habit, habitat and distribution: Gregarious on grassy ground with leaf litter near pine trees. It is known from China and Pakistan. Notes: Agaricus gregariomyces, originally described from China from a single collection (ZRL2012624), is reported for the first time from Pakistan (KH72). Macroscopic characters of both collections match except for the odor described as pleasant in the Chinese collection and mild phenol-like in the Pakistani collection. Microscopically, spore size is similar in both collections, however only tetrasporic basidia were observed in ZRL2012624 while frequent bisporic or tetrasporic basidia but rare trisporic basidia were observed in KH72. In addition, cheilocystidia were not observed in ZRL2012624 while they are abundant, broadly clavate to sphaeropedunculate shaped in KH72. Agaricus gregariomyces shares most of the morphological characters with other members of A. sect. Xanthodermatei. Therefore, molecular methods are crucial to identify this species. In this clade, the most related species are A. buckmacadooi Kerrigan   Etymology: The specific epithet 'griseovariegatus' refers to the variegated grey color of pileus. Species-specific ITS markers-None.
Original description: Pileus 8-13.5 cm in diam., ovate to broadly parabolic when young, then convex to plano-convex, umbonate, covered with variegated olive-grey (5Y 3/2) appressed squamules in young stages www.nature.com/scientificreports/ turning dark olive grey (2.5Y 2/2) at maturity on a white background (10YR9/2), in the collections SA194, SA232, KH167 broadly radially fissured at maturity. Surface dry and dull. Margin entire in young sporocarps, then ruptured at maturity, slightly appendiculate, vaguely incurved, slightly exceeding the lamellae, yellowishbrown when rubbed. Lamellae white to light pinkish (8.3RP 7/3.8) at young stage and finally brown (7.5YR7/2) at maturity, free, crowded, intercalated with lamellulae and with entire edges. Stipe 10-15 × 2.0-2.6 cm at the apex, 2.5-3.0 cm at the bulbous base, cylindrical, stuffed, provided with annulus in its upper half close to lamellae, above and below the annulus smooth, white when young discoloring rusty brown by rubbing, becoming slightly brownish at maturity with rhizomorphs at the base in most specimens. Annulus superous, double edged, broad, pendant, membranous, upper side smooth, underside floccose, white with light brown perimarginal linear squamules. Context white, discoloring very faint yellow when cut in the upper half of the stipe (only in KH167). Odor of phenol. Basidiospores Notes: Morphologically, A. griseovariegatus is characterized by its pileus with appressed squamules with olivaceous tones, and by its broadly radially fissured pileus margin at maturity. These two combined characters are not noticed neither in its sister species A. xanthochromaticus (Figs. 1 and 2) nor in other species of clade III. In clade II, the collection JBSD127423 from Dominican Republic similarly exhibits a pileus completely covered with dark greyish-brown squamules with olivaceous tones, but it is neither variegated nor radially fissured and lacks odor of phenol. Etymology: The specific epithet "macropeplus" means broad annulus. From the Latinized Greek words makros (long, large, broad) and peplos (robe, tunic).
Basidiospores Notes: Agaricus macropeplus is morphologically very similar to many species of clade Xan III and therefore an unequivocal morphological differentiation with all taxa is not possible. Phylogenetically, Agaricus macropeplus clusters in the same clade as A. griseovariegatus, A. sinoplacomyces and A. xanthochromaticus (Figs. 1 and 2) from which it differs in the ITS sequence at 3, 5 and 6 nucleotide sites, respectively ( Table 2). Agaricus griseovariegatus differs by having a pileus surface covered with brown squamules with olicaceous tones progressively scattered towards the margin but it is very difficult to distinguish A. macropeplus from A. sinoplacomyces and A. xanthochromaticus. Agaricus sinoplacomyces was described with a pleasant odor (but this feature needs confirmation) and A. xanthochromaticus has an intense odor of phenol and slightly smaller spores on average (4.9 × 3.3 µm). Etymology: The specific epithet derives from the two Latin words "parvi" meaning small and "niveus" referring to the white color of pileus.
Basidiospores ( Notes: Basidiomata of A. swaticus have been collected from different areas of Swat district. Morphologically some differences have been observed, possibly due to contrasting climatic conditions. Pileus surface rimose, umbonate, dull and dry, dark greyish brown, with incurved and wavy margin was noticed in the samples collected from the dry temperate regions (SJ53 and SJ60), and a fleshy pileus having brown scales on surface, indistinctly umbonate, slightly recurved and split margin in specimen SA111 collected from the moist temperate region. However, there are no molecular differences between the ITS sequences of the three specimens.
In our phylogenetic analyses A. swaticus is sister to A. langensis and is also phylogenetically related to A. parvitigrinus, and A. californicus Peck. The ITS sequences of A. langensis and A. swaticus differ at two positions from each other, but their LSU sequences also differ at one position, which is much more significant because this DNA region is much less variable than the ITS region. Morphologically A. langensis differs from A. swaticus by its white unchanging context when cut, its larger spores (7.2 × 4.4 µm on average) and the lack of cheilocystidia. The fact that both molecular and micro-morphological differences with A. langensis are consistent among our three specimens of A. swaticus collected in different locations, supports our taxonomic treatment of A. swaticus as a distinct species. However, the original description of A. langensis by He et al. (2018) was based on a single collection and therefore the morphological variability of A. langensis is still unknown.
Agaricus parvitigrinus differs by having a pileus with white margin, a narrow annulus, a faint odor of phenol, slightly smaller spores (5.83 × 3.73 µm on average) and basidioliform scarce cheilocystidia, and A. californicus by having a pileus with white margin, unchanging context when cut, and spores shorter and wider on average  MycoBank: MB808224 A new specimen (PU252, Fig. 19D) of this species of A. sect. Xanthodermatei (clade Xan II) was collected in Lahore, which is in a subtropical region of Pakistan. Morphological, molecular and phylogenetic data completely agreed with those previously reported for samples of this species from Pakistan and Thailand by Thongklang et al. (2014). We have examined also the type of A. bisporiticus (MCR25) provided by LAH herbarium and our observations agree with those published in the original description since bisporic, trisporic, and tetrasporic basidia were observed frequently, spores are rather similar, measuring (4.5-) 5.7-6. Three new specimens (L11, PU265 and CM200; Fig. 19) of this species of A. sect. Xanthodermatei (clade Xan III) were collected from Changa Manga forest (CM200) and Lahore (L11 and PU265). With three polymorphic ITS-nucleotide sites, and high morphological variability, the subtropical species A. punjabensis appears to be quite variable (Fig. 19). For example, the pileus of the specimen PU265 is fissured in an irregular pattern forming various small patches of dark brownish grey squames. The pileus diameter of CM200 is bigger (11.5 cm) than those described by Chen et al. (2016). In the previous report of A. punjabensis, the discoloration was not recorded. A bright yellow discoloration and a strong odor of phenol at stipe base was noted in our new collections.    Fig. 19F. A specimen (JBSD127423, duplicate in LAPAM 110) belonging to A. sect. Xanthodermatei (clade Xan II) was collected in a mixed forest with broadleaf trees and Pinus occidentalis in Jarabacoa, La Vega, Dominican Republic. This specimen belongs to a new species which is not formally named because only a single incomplete basidiomata, lacking the stipe base, was collected. This specimen is remarkable for its pileus surface with olivaceous tones, a color never noticed in any member of clade Xan II. Pileus 3 cm in diameter, hemispherical, dark grayish-brown with olivaceous tones, entire at center and covered elsewhere by small scales, punctiform near the center and triangular towards the margin on a whitish background, with margin slightly exceeding the lamellae. Annulus attached to the pileus margin, 2 mm broad, with scales in the lower margin. Lamellae 3 mm broad, dark brown with whitish edge. Stipe 3 mm wide, smooth, whitish with a brown apex. Context white, with a very faint yellow discoloration towards the base when broken, without a distinctive odor.

Discussion
A brief overview on the relationships between A. sect Hondenses and A. sect. Xanthodermatei. This study focuses on A. sect. Xanthodermatei and A. sect. Hondenses. These two sections share morphological characteristics, which allow to distinguish them from the other eleven sections in A. subg. Pseudochitonia. In recent studies of Zhao et al. 7 and Parra et al. 4 , A. sect. Hondenses appears closer to A. sect. Bivelares than to A. sect. Xanthodermatei in multi-gene trees but without support value. In contrast, A. sect. Hondenses and A. sect. Xanthodermatei are sister to each other in Zhou et al. 22 , and in the present study with a relatively good support (BS = 78, PP = 1). This relationship will have to be confirmed in the future in a larger multi-gene analysis including the 13 sections of the subgenus and a higher proportion of samples with multi-gene data. However, it should be noted that, for example, the number of species of these two sections with TEF1 sequence data already increased from 9 in Zhao et al. 7 to 29 in the present study.

Comparison of the major clades and interest of indel markers.
Despite our efforts to compare the clades Xan I, Xan II, and Xan III, we did not find any morphological feature reliably characterizing these clades. According to Table 1, temperate species are clearly preponderant in Xan I (7/8) and Xan III B (8/11), while tropical or subtropical species are preponderant in Xan II (11/16) and Xan III A (1/1). The clade Xan IIIC is complex and include the half of all the named or unnamed species (36/72) with some remarkable subclades. For example, a well-supported subclade includes four (sub)tropical species of which three are found in Pakistan (A. endoxanthus, A. fumidicolor and A. punjabensis), while another subclade comprises four temperate species including the type species of the section A. xanthodermus.
Comparing the ITS barcode sequences of all the species, we report three discriminating markers between the three clades Xan I, Xan II and Xan III. Two of them are clade-specific indels, which are sufficient to classify all the species in the three clades. A risk when using clade or taxon-specific marker is that its specificity can be lost when new species are discovered. It is what happened for certain clade-specific markers previously reported by Thongklang et al. 18,19 . Apparently, indel markers remained more reliable than nucleotide markers possibly due to a lower homoplasy 49 . In addition, such indels are phylogenetically informative and could be included in the analyses, however for our current purpose this was not necessary since even without them, the major clades were phylogenetically well-supported in our multi-gene tree. Nevertheless, in the tree only based on ITS sequences (Fig. 2), the sections and major clades are generally present and as well-or less well-supported than in the multi-gene tree with the exception of Xan II and Xan III A. Indeed, Xan II appears as a paraphyletic group and, in addition, the single species of clade Xan III A (A. flavidodiscus) is mixed with species of the clade Xan II. In fact, using only ITS sequence, the placement of A. flavidodiscus is unstable. For example, it appeared in Xan III in both ITS and multigene trees of Parra et al. (2018) but this was well-supported only in the multigene tree. Phylogeny of clades Xan II and Xan III A remains improperly resolved in ITS tree.
In conclusion, multi-gene analyses were necessary to establish the major clades and their phylogeny. However, two indels in ITS alignment are efficient to classify the species in the three major clades. Despite Zhao et al. 7 proposed the taxon A. sect. Hondenses for the clade Xan I, we did not try to propose taxa for the other clades, considering that it is presently not possible to reliably characterize them morphologically.
Low level of intra-and interspecific variability in ITS sequences of A. sect. Xanthodermatei. We were surprised by the absence of variability among the ITS sequences of seven of the eight new species proposed in this study. Since this is generally not so frequent in genus Agaricus, we decided to examine on the one hand the intraspecific variability in A. xanthodermus, which is the type of A. sect. Xanthodermatei and one of the most documented species of this section in GenBank, and on the other hand the intra-and interspecific variabilities in a group of ten closely related species.
In the case of A. xanthodermus, we found that the 18 ITS sequences retrieved from GenBank and which are from three different continents are identical with the exception of a difference at a single nucleotide site in only one of them. Since we observed this low variability in most of or new species, one conclusion is that the intraspecific variability in the ITS sequences is frequently low in A. sect. Xanthodermatei. However, it is not a rule since, for example, ITS sequences in A. endoxanthus are highly variable 19 . www.nature.com/scientificreports/ In the case of the group of closely related species (Table 2), we found that the intraspecific variability was quasi-absent in most of these species and that certain species differ from each other at a single ITS nucleotide site. This is the case of A. deardorffensis from North America and A. tibetensis from Asia. More surprisingly, this is also the case for A. xanthochromaticus and A. griseovariegatus which are both new species described from Pakistan. Such a low level of interspecific variability has been previously observed in this section for example between ITS sequences of A. xanthodermus and A. moelleri Wasser, which differ from each other at only at two positions 16 . We can also note that the new species A. swaticus and the recently described species A. langensis differ also at two positions from each other. It is noteworthy that in genus Agaricus when a single difference is observed between ITS sequences of two samples, they are generally considered as belonging to the same species. However, when several specimens are available and when this difference is correlated with morphological differences and/or, with differences in other sequenced genes, the two divergent entities are considered as distinct species. This is the case for the two new species A. xanthochromaticus and A. griseovariegatus, which not only are morphologically relatively well-distinguishable but also differ at 6 positions of their TEF1 sequences. This suggests that an efficient process of concerted evolution occurred for ITS. In other respects, contrarily to other groups of closely related species such as in A. sect. Arvenses (Konrad & Maubl.) Konrad & Maubl., there is no indication of reticulate evolution of hybridization between species or distant populations. This suggests that efficient reproductive barriers take place during the speciation process. Geographical and climatic factors may have contributed to these speciation events since these species are distributed in North America or in Asia and only one of them adapted to subtropical conditions (A. sinoplacomyces). However, three of them are temperate and allopatric species newly described (A. griseovariegatus, A. macropeplus and A. xanthochromaticus).
In conclusion, in A. sect. Xanthodermatei, we hypothesize that efficient reproductive barriers take place in speciation process and that efficient concerted evolution frequently occurs for ITS. We frequently observe a low level of intraspecific variability and a low level of interspecific variability between closely related species. Samples that differ at only one position in the ITS sequence might belong to different species more frequently than in other sections. In this case, it is prudent to have several collections and, as far as possible, to compare them with other markers than ITS like TEF1. Some unexpected features. In A. sect. Hondenses, the new species A. bambusetorum described with its smooth white small sized cap, its narrow simple annulus, its pleasant odor and its distribution in subtropical climatic area, is completely atypical in this section. In other respects, it was surprising to find bisporic basidia, sometimes abundantly, in all the new species. They were also observed in the new collection of A. gregariomyces and were abundant in the type as in the new collection of A. bisporiticus. Before interpreting these unexpected observations, it should be necessary to observe fresh specimens on dry microscopic preparation without cover glass to better estimate the proportions of n-spored basidia, not only for the species collected in Pakistan but for all the species of A. sect. Xanthodermatei in the future. Variable proportion of bisporic basidia are reported in different taxa of A. sect. Bivelares, which is closely related to A. sect. Xanthodermatei, such as in A. bisporus 50 and more recently in A. sinodeliciosus Z.R. Wang & R.L. Zhao 51 . The proportion of bisporic basidia could be misestimated due to non-representative or too small sampling in dried specimens. In addition, this trait can be sensitive to environmental factors 52 . Spores of bisporic basidia could be heterokaryotic and in this case the life cycle is pseudohomothallic, which could strongly impact the population structure of these species and their evolution.
A summary of the increase of the number of species and of the species richness in Pakistan. The number of named species with available ITS data set currently reaches 8 in A. sect. Hondenses and 53 in A. sect. Xanthodermatei. Among these 61 species, 16 were described before 2000, 33 from 2000 to 2018 and listed in 2 with one correction (A. freirei Blanco-Dios must be placed in A. sect. Hondenses instead of A. sect. Xanthodermatei), 3 in 2019 35,53 , and 8 in this study. All 61 species were included in our analyses except the most recently described species A. rubripes J.F. Zheng & L.H. Qiu (holotype K17071201; ITS/GenBank MH220318; Guangdong Province China). For phylogenetic purpose, it was not necessary to update our analyses with this species because its ITS sequence differs at only one nucleotide site from those of A. sp./ZRLWXH3092, which was already included in the analyses and placed in clade Xan II (Figs. 1 and 2).
Among the 61 species identifiable with ITS barcode in both sections, 12 are reported from Pakistan including 8 new species and a new report from this country in the present study. This represents a high species richness compared to the only four species reported from Iran (Mahdizadeh et al. 2017) in the two sections, which, at this time, were still grouped in A. sect. Xanthodermatei. Species reported from Iran are all temperate and previously reported from Europe. In contrast, 9 of the 12 species found in Pakistan are endemic, only two are reported from other Asian countries (China and Thailand), and the remaining one is the cosmopolitan tropical species A. endoxanthus. None of these 12 species are known in Europe except occasional cases of introduction of A. endoxanthus in glass-houses. Half of them are tropical or subtropical and they are distributed in the different major clades as follows: 1 in Xan 1, 2 in Xan II, 9 in Xan III of which 2 are in Xan III B and 7 in Xan III C. This remarkable biodiversity can be partly explained by the diversity of the climates and biotopes in Pakistan.
For the sake of completeness 14 species are described. Eight new species and one new record for Pakistan are described in detail. New collections of three other species previously reported from Pakistan are briefly described. The two remaining brief descriptions do not concern samples from Pakistan but (i) a sample from Spain which formally represents the first report in Europe and outside North America of A. californicus, (ii) a sample of an unnamed species from Dominican Republic that we considered useful for its contribution to the clade Xan II in the phylogenetic analysis.
We predict that the number of species will continue to quickly increase and we recommend, as much as possible and more specifically for A. sect. Xanthodermatei, to use several specimens, to get TEF1 sequences, and to www.nature.com/scientificreports/ use the BLAST but with caution. Indeed, the percentage of identity can be incorrect due to heteromorphisms and many sequenced samples are misidentified in GenBank. In other respects, discoloration, odor and structure of the annulus should be carefully examined and reported from specimens in the field and it must be never forgotten that the species of these two sections are presumably poisonous.