The famous cultivated mushroom Bailinggu is a separate species of the Pleurotus eryngii species complex

The mushroom of the genus Pleurotus in western China, called Bailinggu, is a precious edible fungus with high economic value. However, its taxonomical position is unclear. Some researchers regard it as a variety of P. eryngii, namely P. eryngii var. tuoliensis, whereas others consider it to be a subspecies of P. eryngii, viz. P. eryngii subsp. tuoliensis. A total of 51 samples representing seven genetic groups of the genus Pleurotus were subjected to a phylogenetic analysis of partial sequences of the translation elongation factor 1 alpha gene (ef1a), the RNA polymerase II largest subunit gene (rpb1), the RNA polymerase II second largest subunit gene (rpb2) and nuc rDNA internal transcribed spacers (ITS). Our data indicate that the mushroom Bailinggu is a lineage independent of P. eryngii and should be lifted as its own species, namely P. tuoliensis. In addition, its known distribution range consists of both western China and Iran.

sequence analyses, Zervakis et al. 11 upgraded its taxonomic status to subspecies, and treated Bailinggu as P. eryngii subsp. tuoliensis (C.J. Mou) Zervakis & Venturella. Many studies have shown that single-copy protein encoding regions are more suited for revealing the relationships of closely related species 12 . Based on an analysis of ef1a and rpb2 sequence data, Rodriguez Estrada et al. 13 treated P. eryngii var. nebrodensis as an independent species, which is consistent with the viewpoint of Venturella 7 . The present study include a phylogenetic analysis of several genetic groups in the Pleurotus genus that was implemented using four nuclear DNA fragments (ef1a, rpb1, rpb2 and ITS) to infer the taxonomic status of Bailinggu from western China and its relationships with other related species. The phylogenetic species were then delimited in this study according to the genealogical concordance phylogenetic species recognition (GCPSR) criterion 14 .
Phylogenetic analysis and phylogenetic species recognition. Both ef1a (except CCMSSC 04235) and rpb2 (except CCMSSC 00929) were successfully amplified from 50 samples. A gene fragment of rpb1 was obtained from only 47 samples. After sequence alignment, editing and trimming, 525-bp, 1152-bp and 1093-bp segments, which contained 95, 307 and 102 parsimony informative sites, respectively, remained for phylogenetic analysis. The ITS dataset consisted of 50 sequences (with the exception of CCMSSC 00761) generated in this  study and 48 related ITS sequences retrieved from GenBank (Table S1). The sequence alignment comprised 577 nucleotide positions in the ITS region used for the phylogenetic analysis.
The phylogenetic trees that were reconstructed with three independent gene fragments (ef1a, rpb2 and rpb1) and inferred from a maximum likelihood (ML) analysis together with maximum likelihood bootstraps (LB), maximum parsimony bootstraps (PB) and Bayesian posterior probabilities (PP) are shown in Figs S1-S3, respectively. The three phylogenetic trees share the same topology. The phylogenetic tree obtained from the ML analysis with LB, PB and PP support based on the combined dataset (ef1a, rpb2 and rpb1) is shown in Fig. 3. Six major clades supported with high bootstrap values and posterior probabilities could be inferred, corresponding with samples of var. ferulae, var. eryngii, P. nebrodensis, P. tuoliensis (Bailinggu), P. ostreatus (Jacq.) P. Kumm. and P. pulmonarius (Fr.) Quél. Our results identified the mushroom Bailinggu as a monophyletic group supported by a bootstrap value of 100% and a posterior probability value of 1.00. According to the GCPSR criterion, the mushroom Bailinggu should be recognized as an independent phylogenetic species based on the fact that it is highly divergent from its sibling groups.
Phylogenetic relationships among Bailinggu and its related species. ML, maximum parsimony (MP) and Bayesian algorithm (BA) analyses based on the ITS dataset yielded similar tree topologies with some differences in bootstrap and posterior probability values. The tree inferred from the ML analysis is shown in Fig. 4. A phylogenetic reconstruction based on the ITS dataset clustered the P. eryngii species complex samples into four major clades, which are supported with moderate bootstrap and high posterior probability values. One clade consists of the varieties eryngii, ferulae, elaeoselini Venturella, Zervakis & La Rocca, thapsiae Venturella, Zervakis & Saitta, and tingitanus Lewinsohn. The other three clades correspond to P. ferulaginis Zervakis, Venturella & Cattarossi from Italy, P. nebrodensis from Europe and Asia, and P. tuoliensis (Bailinggu) from Asia. The samples of Bailinggu form a monophyletic group in the ITS tree, which exhibits the furthest genetic distance from the other groups of the P. eryngii species complex. These results are consistent with those obtained based on each single-copy protein-encoding gene. The phylogenetic relationships among P. eryngii var. eryngii, P. eryngii var. ferulae, P. eryngii var. elaeoselini, P. eryngii var. thapsiae, and P. eryngii var. tingitanus obtained using the ITS dataset remain resolved.

Discussion
Bailinggu is one of the most widely cultivated mushrooms in China. Recently, this species has been involved in the researches of genetic diversity evaluation 15,16 , temperature response mechanism 17-20 , fructification mechanism 21 , and bioactive substance exploitation 22 . However, the most essential information on the taxonomic status of Bailinggu and its phylogenetic relationships with its sibling species remain uncertain.
The mushrooms from Pleurotus genus that grow on the roots and stems of Umbelliferae plants belong to the P. eryngii species complex. The morphological characteristics of Bailinggu from western China conform to those of the P. eryngii species complex 23 . The morphological differences between Bailinggu and its related species are shown in Table 1. The pileus color of P. eryngii var. eryngii ranges from brown and beige-brown to light beige, whereas the pileus color of P. eryngii var. ferulae from Europe ranges from grey-brown to slate grey to beige brown. The macro-morphological characteristics of Bailinggu are similar to those of P. nebrodensis, but the basidiospores of Bailinggu are slightly smaller than those of P. nebrodensis. The pileus color of P. eryngii var. ferulae from China is brown to white, therefore, it is not possible to distinguish Bailinggu from P. eryngii var. ferulae from China based exclusively on their macroscopic and microscopic characteristics.
The intersterility criterion is a derivative of the biological species criteria. Many cryptic species, such as the Armillaria mellea (Vahl) P. Kumm. 24 , have recently been recognized using this criterion. Previous mating compatibility tests of the P. eryngii species complex did not indicate any complete reproductive isolation among the genetic groups within the P. eryngii species complex. The mating rate between P. eryngii var. eryngii and P. eryngii var. ferulae was the highest, with a value of 98% 8 or 93% 10 , but those between P. nebrodensis and P. eryngii var. eryngii and between P. nebrodensis and were significantly lower, with values of 6-18% 25 . Very few mating tests have been performed between Bailinggu and other genetic groups. According to the previous studies, Chinese Bailinggu showed much higher compatibility with P. eryngii var. eryngii (65%) and P. eryngii var. ferulae (82%) than with P. nebrodensis (15%) and P. ferulaginis (11%) 8,11 . This indicates that Bailinggu might be closer to P. eryngii var. eryngii and P. eryngii var. ferulae. However, some evidence that many fungi genetically isolated in nature retain the ancestral character of interbreeding 14 . Hilber 26 found that P. eryngii var. eryngii and P. eryngii var. ferulae could mate with each other in the laboratory, but they appear to be reproductively isolated in the field and are associated with specific host plants.
The results of this study based on molecular data showed that Bailinggu is a separate phylogenetic species instead of a variety or subspecies of the P. eryngii complex according to the GCPSR criterion, although this mushroom retains high intercompatibility with P. eryngii var. eryngii and P. eryngii var. ferulae in the laboratory. A similar observation was found in a study of the P. ostreatus complex. Three intersterility groups or biological species (I, II, and VI) in the P. ostreatus complex were found to contain more than one phylogenetic species 27 . A phylogenetic reconstruction based on the ITS dataset and the combined dataset revealed that the genetic distance between Bailinggu and P. eryngii (var. eryngii, var. ferulae, var. elaeoselini, var. thapsiae, var. tingitanus) was greater than those between Bailinggu and P. ferulaginis and between Bailinggu and P. nebrodensis. This result is in conflict with the previous findings in the mating tests. Considering the geographical isolation of Bailinggu in nature, the results inferred from molecular data are more acceptable because DNA sequence divergence, be it allopatric or sympatric, might occur much earlier than the evolution of intersterility 28,29 .
Previous research using sequence analyses of ITS and IGS1 showed that Bailinggu is a phylogenetic sister group to P. eryngii 11 . However, our study indicates that P. ferulaginis is much more similar to P. eryngii in terms of not only morphology, distribution, and ecology but also DNA divergence. The phylogenetic analysis revealed that Bailinggu is a sister group to the eryngii-ferulaginis-nebrodensis clade and is not closely related to the other genetic groups of the P. eryngii species complex.
Reproductive isolation caused by host specialization is often observed in basidiomycetes, particularly plant pathogenic fungal species 30 . To the best of our knowledge, the P. eryngii species complex has developed a certain degree of host specificity. To detect whether the relationships among the genetic groups of the species complex correlate with those among their hosts, the phylogeny of the relevant hosts was reconstructed based on ITS1 and ITS2 sequences retrieved from GenBank (Fig. S4). The results showed that the eryngii, ferulae, elaeoselini, thapsiae, and tingitanus varieties are so closely related genetically that they could not be distinguished by ITS analysis, but the relationship among hosts of P. eryngii var. eryngii, P. ferulaginis, and P. nebrodensis is markedly closer. In contrast, the genetic relationships of Bailinggu with the ferulae, elaeoselini, thapsiae, and tingitanus varieties are distant, but the genetic relationships among their hosts are close, indicating that hosts might not be the main reason for the divergence of Bailinggu from other genetic groups. Its long geographical isolation might be the main reason for the distant genetic relationship among Bailinggu and other genetic groups.
Pleurotus eryngii, P. ferulaginis and P. nebrodensis are mainly distributed in the Mediterranean and surrounding areas, whereas recent studies found that P. eryngii and P. nebrodensis also occur in Asia 11 . The distributions of the two mushrooms are wide and continuous, but there is very limited information on the distribution of Bailinggu. The samples of Bailinggu used in the present study were mostly from western China, and partly from Iran 11,31 . The main distribution area of Bailinggu in China is located far from the distribution areas of other genetic groups with the exception of P. eryngii var. ferulae from China. There are no obvious differences in morphological characteristics or habitat between Bailinggu and P. eryngii var. ferulae from China. However, a sequence analysis showed a remarkable difference between them in terms of DNA sequence, which is consistent with previous results 6 . What efficient prezygotic barriers that maintain the separation of both gene pools will require further study. The pileus color of P. eryngii var. ferulae from China is different from that of P. eryngii var. ferulae from Europe. Moreover, the phylogenetic analysis showed that they cluster according to their geographical origins even though they still belong to the same genetic group. Geographical isolation and differences in biotope would likely lead to increasing divergence of an individual population to enhance differentiation 32,33 .

Conclusion
This study, which involved multiple or independent DNA gene fragment analyses in combination with a morphological analysis, showed that Bailinggu is highly divergent from its related groups at the DNA level but presents no significant differences in morphology or mating incompatibility. According to the GCPSR criterion, Bailinggu is an independent phylogenetic species in the P. eryngii complex, and based on its geographical isolation in nature, P. eryngii var. tuoliensis or P. eryngii subsp. tuoliensis should be upgraded to an independent species, and P. tuoliensis should be the scientific name for this mushroom. The taxonomic treatment is as follows: Pleurotus Specimen examined: GDGM 27082 ( Table 2).

Materials and Methods
Taxon sampling. Fifty-one specimens representing seven different genetic groups of Pleurotus were used in this study ( Morphological observation. The morphological characteristics of fresh fruitbodies were observed and recorded in the field. The samples were dried at 40-50 °C, and microscopic features were observed with a light microscope. The size of the basidiospores was described in the form of (a)b-c(d), and 90% of the measurements were within the range of b and c; a and d (in the parentheses) are the minimum and maximum of the measurements, respectively, whereas the quotient (Q) of their dimensions was calculated as the ratio of the spore length (arithmetic average of all spores) to the spore width (arithmetic average of all spores).

DNA extraction, amplification and sequencing. Total DNA was extracted using a DP305-Plant
Genome Extraction Kit (Tiangen, China). PCR amplifications were conducted using the following primer pairs: EF595F/EF116OR for the portion of the ef1a gene, fRPB2 5F/bRPB2 7.1R, b6.9F/b11R1 for the fragment of the rpb2 gene 12 , RPB1 2F (5′ ATTGCGGGCGACTAAAGG 3′ ) and RPB1 5R (5′ CTGCTCAAACTCGGAGATAA 3′ ) for the part of the rpb1 gene, and ITS1/ITS4 34 for the ITS region. Each amplification reaction system contained approximately 20 ng of DNA template, 0.2 mM dNTPs, 0.5 mM each primer, and 1 U of Ex Taq DNA polymerase (TaKaRa, Japan) in a final volume of 20 μ L. The PCR was conducted using the following program: 94 °C for 4 min followed by 35 cycles of 94 °C for 50 s, 55 °C for 50 s, and 72 °C for 1 min. The reaction was completed by incubation at 72 °C for 10 min. The amplified products were separated by electrophoresis on 1.2% agarose gels and stained with ethidium bromide. Sequencing was performed by BGI Co., Ltd (Beijing, China). The PCR products from each sample that failed to yield direct sequencing results were cloned using a pGEM-T easy cloning kit (Promega, USA) and transformed into DH5α component Escherichia coli cells. Ten random transformed E. coli colonies were selected for sequencing, and the sequence data were deposited in GenBank (Table 2).
Sequence alignments. Each DNA sequence was assembled and edited manually if needed. Sequence alignments were performed using the MUSCLE algorithm in MEGA 5.0 35 . Different alignments were performed for different analytical purposes. Multiple or independent DNA gene fragments (ef1a, rpb2 and rpb1) were used to reconstruct the phylogeny of mushrooms of the genus Pleurotus to infer the taxonomic status of the Chinese Bailinggu. For the Pleurotus genus, more ITS sequences than sequences of the other three genes were readily available in GenBank, and the relationships among Bailinggu and its sibling species were further investigated using the ITS dataset.
Phylogenetic analysis. Phylogenetic reconstructions using ef1a, rpb1, rpb2, the combined data set of the three genes, and ITS were performed using MP, ML and BA. The MP analyses were performed with PAUP* 4.0b10 36 . Heuristic searching with TBR branch swapping was implemented with 1000 random-addition sequence replicates. The bootstrap analysis was conducted with 1,000 replicates using the heuristic search 37 . ML analyses were conducted in PHYML3.0 38 , and the bootstrap analysis was performed with 1000 replicates. BA analyses were run using MrBayes3.1.2 39 . The Markov Chain Monte Carlo (MCMC) algorithm 40 was utilized to calculate the Bayesian posterior probabilities. Four Markov chains were run for 5,000,000 generations with the trees sampled every 1000 th generation. The average standard deviation of split frequencies was restricted to less than 0.01. The first 25% trees were discarded as burn-in. The optimum substitution model for each dataset was estimated by jModelTest 41 according to the Corrected Akaike Information Criterion (AIC) 42 for the ML analyses and the Bayesian information criterion (BIC) 43 for the Bayesian analyses. For the ML analyses, the optimal substitution models for the four partitions determined using the AIC were as follows: TIM1 + G for ef1a and rpb1, SYM + G for rpb2, and TPM2uf + G for ITS. The Bayesian analyses were performed with the following selected substitution models: TrNef + G for ef1a and rpb2, TPM1 + G for rpb1, and TPM2uf + G for ITS. The samples without available sequences were not used in the phylogenetic reconstructions.
Phylogenetic species determination. The phylogenetic species were delimited in this study according to the genealogical concordance phylogenetic species recognition (GCPSR) criterion. Using this method, phylogenetic species were recognized as genealogically exclusive under GCPSR if they were concordantly supported by multiple independent loci 14 .