Most other specimens were collected from Hunan Province; only one specimen was collected from Huang Mountain, Anhui Province. The fresh basidiomata were dried using an electric dryer EVERMAT operated at 45 °C for 10 h. The dried specimens, along with the holotype of the newly described species, were deposited in the Mycological Herbarium of Hunan Normal University (MHHNU), Changsha, China. A small piece of fresh basidioma was also dried with silica gel for molecular analysis.

Specimens were photographed in situ using a Sony digital camera (LICE-7, Sony, Tokyo, Japan). The macromorphological characters of fresh mushrooms were recorded as soon as possible after collection. Color codes were described following Kornerup and Wanscher (1978). Microscopic structures were studied from dried materials mounted in 5% aqueous KOH, and Congo red was used as a stain when necessary. All the measurements were performed at 1000× magnification, and a minimum of 20–30 basidiospores from each basidioma were measured in side view. Micromorphological investigations were performed by means of a Nikon Eclipse 50i microscope (Nikon, Tokyo, Japan). The measurement methods followed those of Fan et al. (2013). Dimensions of basidiospores and Q values were given as (a) b–c (d), where “b–c” cover a minimum of 90% of the measured values, and “a” and “d” represent extreme values; Q is the ratio of length to width in an individual basidiospore. Qm is the average Q of all basidiospores ± sample standard deviation. The descriptive terms are in accordance with Fan and Bau (2020), Horak et al. (2015), and Matheny et al. (2012). SEM images of basidiospores were obtained using a scanning electron microscope JSM-6380LV (JEOL Ltd., Tokyo, Japan).

DNA was extracted from dried basidiomata using a fungal DNA extraction kit manufactured by Omega Bio-Tek (Norcross, GA, USA). The following primer pairs were used for PCR amplification and sequencing: ITS5 and ITS4 for the internal transcribed spacer (ITS) region (White et al. 1990); LR0R and LR5 for the nuclear ribosomal large subunit (nrLSU) region (Vilgalys and Hester 1990); and bRPB2-6F and bRPB2-7.1R for RNA polymerase II second largest subunit (rpb2) region (Matheny 2005). PCR protocols for ITS and nrLSU were as described in White et al. (1990), and for rpb2, as described in Matheny (2005). PCR products were purified and sequenced by TsingKe Biological Technology Co., Ltd. (Beijing, China).

Thirty-six sequences (12 for ITS, 12 for nrLSU and 12 for rpb2) were newly generated for this study and deposited in GenBank (Table 1). The new sequences were subjected to a BLAST search and relevant related sequences retrieved from GenBank (Table 1).

The sequences were aligned using MAFFT v7.310 (Katoh and Standley 2013) and manually edited using BioEdit v7.0.5 (Hall 1999). Maximum likelihood (ML) analysis was performed using RAxML v7.9.1 (Stamatakis 2006) under the GTR + GAMMA + I nucleotide substitution model and performing nonparametric bootstrapping with 1000 replicates. Bayesian inference (BI) was performed in MrBayes v3.2 (Ronquist et al. 2012). The optimal substitution model was determined using the Akaike information criterion (AIC) as implemented in MrModeltest v2.3 (Nylander 2004). The selected substitution model for the three partitions was as follows: General Time Reversible + Gamma (GTR + G) for ITS, and General Time Reversible + Proportion-Invariant + Gamma (GTR + I + G) for nrLSU and rpb2. The BI analysis was conducted with the following parameters: four simultaneous Markov chains (MCMC), each with two independent runs and trees summarized every 1000 generations. The analyses were completed after 1,000,000 generations when the average standard deviation of split frequencies was 0.009808 for the analysis, and the first 25% generations were discarded as burn-in. The phylograms from ML and BI analyses were visualized with FigTree v1.4.3 (Rambaut 2009).

The procedure of toxin extraction and detection followed Xu et al. (2020) with slight modifications. A 0.05 g powdery sample of dried mushroom pileus was mixed with 2 mL of a methanol-water solution (7:3 v/v) and vortexed for 30 min at room temperature. The mixture was treated in an ultrasonic bath for 30 min. After centrifugation at 10,000 rpm for 5 min, the supernatant was purified using a QuCHERS–PP column. Subsequently, the extract was mixed with acetonitrile to a final volume of 1.0 mL. The obtained sample solution was centrifuged at 21,000 rpm for 2 min before UPLC–MS/MS analysis. Lentinula edodes was used as a blank sample.

UPLC–MS/MS analysis was carried out with a Waters ACQUITY I-Class UPLC system coupled with a Waters Xevo TQ-S MS/MS system (Waters, Milford, MA, USA). The chromatographic separation was conducted using an ACQUITY UPLC Amide column (2.1 × 100 mm, 1.7 μm; Waters). A gradient elution system used the mobile phase A (acetonitrile) and the mobile phase B (0.05% formic acid aqueous solution) at a flow rate of 0.6 mL/min. The gradient program was as follows: (1) 70–10% A for 1 min, (2) 10% A for 0.5 min, (3) 10–70% A for 0.5 min, and (4) 70% A for 3 min. The analytical column was set to 40 °C, and the injection volume was 2.0 μL. The muscarine content was estimated in the mushroom extract by using standard muscarine (Sigma-Aldrich, St. Louis, MO, USA, Chemical purity ≥ 98%).

A protonated molecular ion ([M + H]⁺ = 174.2) was chosen as the parent ion as well as two daughter ions of 57.0 and 97.0, which were used for qualitative and quantitative detection, respectively. The MS/MS conditions were as follows: ESI⁺ mode; cone, 18 V; collision, 16 V; capillary, 3.0 kV; desolvation temperature, 500 °C; source temperature, 150 °C; desolvation gas flow, 1000 L/Hr; cone gas flow, 150 L/Hr; and collision gas flow, 0.19 mL/min. All the gases were 99.999% pure. Other parameters were used with default values. The product ion confirmation (PIC) was set as follows: scan function; daughter scan; activation threshold level, 500× background noise; minimum activation threshold, 5000 counts; reset threshold level, 50% of act threshold; mass above parent, 100 Da; minimum mass, 50 Da; centroid; scan speed at 5000 amu/s; PIC duration, 0.5 s; and collision energy, 20 V. The analytical results were reported as X ± U (k = 2, p = 95%), where X is the analytical content and U is the expanded measurement uncertainty (Eurachem 2012).

The combined dataset (ITS, nrLSU, and rpb2) contained 1987 total characters and included 58 sequences. The topologies of ML and BI phylogenetic trees obtained in this study were practically the same and the only ML tree with branch lengths and support values is shown in Figure 1. All members of Inocybe s. str. in the dataset formed a monophyletic lineage with strong support (MLB = 85%, BPP = 1). Ten specimens of I. squarrosolutea from China (MHHNU8536, MHHNU8984, MHHNU31006, MHHNU31042, MHHNU31173, MHHNU31427, MHHNU31434, MHHNU31445, MHHNU31875, MHHNU32151) and two samples labeled as “I. sphaerospora” from China (ZRL20151281) and Japan (60-774) grouped together in a well-supported lineage (MLB = 100%, BPP = 1.0). The new species, I. squarrosofulva, formed a well-supported distinct lineage from I. squarrosolutea (MLB = 100%, BPP = 1.0) and is sister to the lineage of I. squarrosoluta with significant support (MLB = 100%, BPP = 1.0).

Figure 1. 

Phylogenetic relationship and placement of Inocybe squarrosofulva and I. squarrosolutea inferred from the combined dataset (ITS, nrLSU, and rpb2) using ML. Bootstrap values ≥80% and Bayesian posterior probabilities ≥0.95 are reported on the branches. Sequences generated in this study are shown in bold. The new species is indicated in red. The red branch indicates the confirmed presence of muscarine, the gray branch indicates ambiguous for muscarine, and the black branch indicates a lack of muscarine.

Through UPLC–MS/MS detection, we found that both I. squarrosolutea and I. squarrosofulva contained muscarine (Figs 7, 8). In the qualitative analysis, muscarine was identified by comparing the retention time (0.91 min) and relative deviation (0.6%) within the allowable relative range of 25%. The calibration curve in the matrix blank extract given by Y = 69369X + 6849.33, R² = 0.9990 (X is injection volume, Y is peak area, and R² is correlation coefficient) for muscarine concentration was in the range of 0.5–20 ng/mL. The contents of in I. squarrosolutea and I. squarrosofulva were 136.4 ± 25.4–1683.0 ± 313 mg/kg dry weight and 31.2 ± 5.8–101.8 ± 18.9 mg/kg dry weight, respectively (Fig. 9). Recovery of muscarine ranged from 72.2 to 93.6%; the average recovery was 83.0%.

Figure 7. 

Total ion current (TIC) chromatogram of muscarine in Inocybe squarrosolutea (MHHNU31427).

Figure 8. 

Total ion current (TIC) chromatogram of muscarine in Inocybe squarrosofulva (MHHNU31548).

Figure 9. 

Relative muscarine concentrations measured by UPLC–MS/MS.

Based on the morphological characteristics, the mushroom was identified as I. squarrosolutea, which was first described from Cameron Highlands of Malaysia (Horak 1979). According to the original description, this species is characterized by a large-sized basidiomata, a bright yellow coloration and a scaly pileus and orange fibrillose veil remnants on the stipe. Our Chinese materials fit well with the original description in basidiomata size, outwards appearances, and the shape and size of micro-features. Meanwhile, there are some tiny difference between them. The holotype of I. squarrosolutea has longer scales (up to 4 mm) in pileus, smaller basidiospores (4–8 × 5–6 μm), finer basidia (18–26 × 5–7 μm), thicker hymenial cystidia (30–60 × 14–25 μm) (Horak 1979). This species is a close relative of I. lutea which, by contrast, has a smaller fruiting body, a smooth pileus, and distinctly smaller basidiospores (Kobayasi 1952; Horak 1979). It is easily for people to confuse I. squarrosolutea and I. sphaerospora because of their similar appearance. In fact, they can be easily distinguished by their basidiospores. The basidiospores of I. squarrosolutea are nodulose, while those of I. sphaerospora are globose (Kobayasi 1952; Horak et al. 2015). In phylogenetic analysis (Fig. 1) the specimens of I. sphaerospora identified by Horak et al. (2015) formed a monophyletic lineage with strong support (MLB = 100%, BPP = 1), and was distinct from I. squarrosolutea. However, the two materials labeled as I. sphaerospora from China (ZRL20151281) and Japan (60-774), cluster together with I. squarrosolutea in the phylogenetic tree, indicating an inaccurate identification of these two materials.

Inocybe squarrosofulva is characterized by its orange brown to dark brown pileus with squarrose scales, distinctly filamentous annulate cortina in stipe, stipe pruinose only near the apex, nodulose basidiospores with six hemispheric knobs, and its odor like raw potatoes. Phylogenetic analyses revealed that I. squarrosofulva is an independent lineage in Inocybe s. str. and is sister to I. squarrosolutea. However, I. squarrosolutea differs in having primrose yellow to bright yellow pileus with less squarrose scales, no distinctly filamentous annulus cortina in the stipe, a subbulbous to bulbous stipe base, a less nodulose basidiospores, and smaller hymenial cystidia. Microscopically, I. lutea is similar to new species in shape and size of pleurocystidia and basidiospores, but the pileus of I. lutea covered with radially fibrils and pruinate all over the stipe (Kobayasi 1952; Horak 1979). Lastly, a Papua New Guinea material described as Inocybe luteifolia (E. Horak) Garrido 1988 (non Inocybe luteifolia A.H. Sm. 1941), which is an illegitimate species name, resembles the new species in macromorphology, but it has smaller basidiomata, larger cheilocystidia and pleurocystidia (55–85 × 10–20 μm), no caulocystidia on the stipe, and a fish-like odor (Horak 1979).

Kuyper (1986) recognized two groups on the (informal) level of “supersection”, viz. Cortinatae and Marginatae, according to the different development mode and, hence, absence or presence of a cortina, and the nature of stipe covering. Due to their presence of a cortina and pruinose at the apex of the stipe, both I. squarrosolutea and I. squarrosofulva might be classified in supersection Cortinatae. The morphological characteristics corresponding to the phylogenetic branches are not yet clear (Matheny 2005; Matheny et al. 2020), so the infrageneric framework of Inocybe s. str. is still unknown and its characterization requires more research.

According to the literature, muscarine was first isolated and identified from Amanita muscaria, but the actual muscarine content of A. muscaria is very low (usually around 0.0003% of the fresh weight) (Waser 1961). Conversely, muscarine concentrations are much higher in Inocybe s. l. spp. (Malone et al. 1962). Brown et al. (1962) detected the muscarine contents of 34 species of Inocybe s. l. by paper chromatographic method, ranging from 0.01 to 0.80% in approximately 75% of them. Kosentka et al. (2013) used liquid chromatography–tandem mass spectrometry (LC–MS/MS) to determine whether muscarine was present in 30 new samples of Inocybe s. l. Of the 30 species they assayed, eleven species tested positive for presence of muscarine, ranging from ca. 0.00006% to 0.5%. Xu et al. (2020) determined the muscarine content of I. serotina by UPLC-MS/MS, and its muscarine content was 324.0 ± 62.4 mg/kg. In our study, the toxin content in each sample was determined using a linear regression equation according to the peak area of the UPLC–MS/MS analysis chromatogram of the test sample (Figs 7, 8). The results showed that both species contained muscarine; the content of muscarine in I. squarrosolutea ranged from 136.4 ± 25.4 to 1683.0 ± 313 mg/kg dry weight and the content in I. squarrosofulva was generally lower, ranging from 31.2 ± 5.8 to 101.8 ± 18.9 mg/kg dry weight (Fig. 9). Calculated on a dry-weight basis, the percentage concentrations were 0.01–0.17% for I. squarrosolutea and 0.003–0.01% for I. squarrosofulva, which is in range of previous reports.

There are some differences in the muscarine content of different poisonous Inocybe spp., even within a particular species. The capacity of Inocybe species to accumulate muscarine may be influenced by certain hereditary (infraspecific races) or environmental factors (Brown et al. 1962). In this study, the differences in muscarine content among specimens of I. squarrosolutea may be related to region and climate. I. squarrosofulva MHNNU31548 and I. squarrosofulva MHNNU31927 were collected in the same place in different years. The weather was sunny at the time of the collection of I. squarrosofulva MHNNU31548, and there was heavy rain at the time of the collection of I. squarrosofulva MHNNU31927, so it is presumed that the difference in muscarine content may be related to rainwater washing.