Fine Identification and Classification of a Novel Beneficial Talaromyces Fungal Species from Masson Pine Rhizosphere Soil

Rhizosphere fungi have the beneficial functions of promoting plant growth and protecting plants from pests and pathogens. In our preliminary study, rhizosphere fungus JP-NJ4 was obtained from the soil rhizosphere of Pinus massoniana and selected for further analyses to confirm its functions of phosphate solubilization and plant growth promotion. In order to comprehensively investigate the function of this strain, it is necessary to ascertain its taxonomic position. With the help of genealogical concordance phylogenetic species recognition (GCPSR) using five genes/regions (ITS, BenA, CaM, RPB1, and RPB2) as well as macro-morphological and micro-morphological characters, we accurately determined the classification status of strain JP-NJ4. The concatenated phylogenies of five (or four) gene regions and single gene phylogenetic trees (ITS, BenA, CaM, RPB1, and RPB2 genes) all show that strain JP-NJ4 clustered together with Talaromyces brevis and Talaromyces liani, but differ markedly in the genetic distance (in BenA gene) from type strain and multiple collections of T. brevis and T. liani. The morphology of JP-NJ4 largely matches the characteristics of genes Talaromyces, and the rich and specific morphological information provided by its colonies was different from that of T. brevis and T. liani. In addition, strain JP-NJ4 could produce reduced conidiophores consisting of solitary phialides. From molecular and phenotypic data, strain JP-NJ4 was identified as a putative novel Talaromyces fungal species, designated T. nanjingensis.


Introduction
Rhizosphere fungi play roles in promoting plant growth and protecting plants from pests and pathogens. Phosphate-solubilizing fungi (PSF) are an important group of such fungi. Phosphate-solubilizing microbes in soil include PSF [1] and phosphate-solubilizing bacteria (PSB) [2]. Fungi and bacteria have their own advantages in adaptation in different environments. The variety and quantity of PSB were more than that of PSF [3], and the research studies on them are still in progress. Common PSF include Aspergillus, Penicillium, Trichoderma, and some mycorrhizal fungi. Phosphate-solubilizing fungi can be applied to a variety of crop ecosystems. For example, Aspergillus niger and Penicillium chrysogenum promote the growth and nutrient uptake of groundnut (Arachis hypogaea) [4]. Inoculation with the PSF Aspergillus niger significantly increases growth, root nodulation, and yield of soybean plants [5]. Phosphate-solubilizing fungi can also be applied to forest ecosystems. The fungal suspension and extracellular metabolites of Penicillium guanacastense have shown to increase the shoot length and root crown diameter of Pinus massoniana seedlings [6].
Penicillium is one of the most common genera of fungi worldwide. It is widely distributed in nature and primarily functions in breaking down organic matter to provide nutrients for its growth [7,8]. Since Link (1809) introduced the species concept of Penicillium [9] and Dierckx [10] introduced the subgenus classification system of Penicillium, studies on Penicillium have become increasingly popular. At the beginning of the 20th century, an early system of classification and identification based on colony characteristics and conidiophore branching patterns was proposed. The genus Talaromyces was first introduced by Benjamin (1955) as the sexual state of the genus Penicillium [11]. Stolk and Samson (1972) divided Talaromyces into four sections based on differences in their asexual states [12]. Later, more advanced and novel classification schemes based on conidiophore structure, branching pattern, and phialide shape, as well as strain growth characteristics, emerged. Pitt (1979) classified Penicillium into four subgenera: Aspergilloides, Biverticillium, Furcatum, and Penicillium, which contain 10 sections and 21 series. Since then, the modern concept of Penicillium sensu lato has emerged [13].
With the popularization of DNA-based phylogenetic studies of fungi, it has been gradually recognized that the subgenus Biverticillium within the genus Penicillium sensu lato is phylogenetically separate from other subgenera of Penicillium and is closely related to Talaromyces, the previously mentioned sexual morph of Penicillium. The subgenera Aspergilloides, Furcatum, and Penicillium originated from Penicillium sensu lato, together with the genus Eupenicillium, and other species now fall within Penicillium sensu stricto, whereas subgenus Biverticillium is synonymized under the current genus Talaromyces [14,15]. Today, section Talaromyces is not limited to sexual species, but it still contains most of the sexually reproducing species in the genus Talaromyces. Yilmaz et al. (2014) proposed a new sectional classification for the genus Talaromyces, placing the 88 accepted species into seven sections, namely, Bacillispori, Helici, Islandici, Purpurei, Subinflati, Talaromyces, and Trachyspermi [15]. Talaromyces flavus (Klöcker) Stolk and Samson (= T. vermiculatus (P.A. Dang.) C.R. Benj.) has always been the typus of genus in the Talaromyces and Talaromyces section Talaromyces through many revisions of the genus Talaromyces [12,13,15]. The current latest concept of species in Talaromyces section Talaromyces is consistent with what Stolk and Samson (1972) described. Stolk and Samson (1972) introduced the Talaromyces section to include species that produce yellow ascomata, which can occasionally be white, creamish, pinkish, or reddish and yellow ascospores. Conidiophores are usually biverticillate-symmetrical, with some species having reduced conidiophores with solitary phialides. Phialides are usually acerose, with a small proportion of species having wider bases [12]. Section Talaromyces species are commonly isolated from soil, indoor environments, humans with talaromycosis and food products. Common species include T. flavus, T. funiculosus, T. macrosporus, T. marneffei, T. pinophilus, and T. purpurogenus.
Micromorphological features such as asexual sporulation structures (e.g., conidiophore) and sexual sporulation structures (e.g., cleistothecium) were of great significance for taxonomy. The branching pattern of conidiophores, namely the type of penicillus, is an important reference index for the traditional classification methods of Penicillium and Talaromyces fungi. The branching pattern generally includes Monoverticillate, Biverticillate, Terverticillate, Quaterverticillate, and Conidiophores with solitary phialides and Divaricate [13,[16][17][18]. Although the classification of Penicillium (and Talaromyces) based on these branching patterns is not completely consistent with the classification status of Penicillium (and Talaromyces) in modern taxonomy, an accurate description of these morphological and structural characteristics is still considered important. The important micromorphology characteristics of Penicillium and Talaromyces fungi include the following: all components of conidiophore (stipes, ramus, ramulus, metula, and phialide) and the sizes, wall texture/ornamentation, color of conidium, ascocarp, ascus, ascospore, and sclerotium. The penicillus includes four parts: ramus, ramulus, metula, and phialide. Sclerotium is produced only under certain conditions; if there are any, observe and record it.
A breakthrough period in the rapid development of classification systems came with the advent of DNA sequencing technology in the 1990s. The identification of Penicillium-group and filamentous fungi began to shift from observation of morphological characteristics to molecular phylogeny. Morphological features are the physical structures with which an organism operates and adapts to its environment, and some features may differ or may be affected by specific factors in the surrounding environment. The effects of medium preparation, inoculation techniques, and culture conditions can be minimized by using strictly standardized protocols [19][20][21]. Morphological identification still plays an irreplaceable role in the fine identification of strains, and a polyphasic approach using both techniques was finally adopted.
In Penicillium, Talaromyces, and many other genera of ascomycetes, internal transcribed spacer (ITS) sequences have been used to classify strains into species complexes or sections, as well as for species identification [15,18]. Due to the limitations of species barcoding based on the ITS region, secondary barcodes or identification markers are often required to identify isolated strains to the species level. Secondary barcodes should be easily amplified, able to distinguish closely related species, and come with a complete reference dataset (including representative gene sequences of all species). The following barcodes can generally be used for the identification of Talaromyces species. The Internal Transcribed Spacer (ITS) rDNA sequence is accepted as the official barcode for fungi [22]. β-tubulin (BenA) is used for the accurately identification of Penicillium species and can also be applied to Talaromyces species [15,18]. Trees have been constructed using other DNA barcode markers (Calmodulin (CaM), DNA-dependent RNA polymerase II (beta) largest subunit (RPB1), and DNAdependent RNA polymerase II (beta) second largest subunit (RPB2)). Among these, CaM, RPB1, and RPB2 exhibit the same potential as BenA and can be used as secondary barcodes for species identification. In recent years, usage of the CaM gene has gradually increased, and its reference dataset has become relatively complete. RPB1 and RPB2 have the added advantage of lacking introns in the amplicon, allowing for robust and easy alignment when used for phylogenetic analysis, but they may be difficult to amplify. At present, the reference dataset for the RPB2 gene of Talaromyces species is fairly robust, whereas that for the RPB1 gene is still being improved. During phylogenetic tree construction, in addition to the reference sequences of ex-types, other multiple collections from the same species should be considered to cover possible sequence variations. Comparing ITS, BenA, CaM, RPB1, and RPB2 sequences from a suspected new species with sequences of the same markers in related species can help to determine whether a species is new via genealogical concordance phylogenetic species recognition (GCPSR) [23]. This approach, which involved multigene phylogeny, morphological descriptions using macro-morphological and micromorphological characters and analysis of extrolites, has been used to develop the polyphasic species concept of filamentous fungi such as Penicillium and Talaromyces.
In our preliminary study, rhizosphere fungus JP-NJ4 was obtained from Masson pine rhizosphere soil and screened for phosphate solubilization and plant growth promotion [24]. Fungus JP-NJ4 has the potential to be used as an ecofriendly soil amendment for forestry and farming. With the aid of internal transcribed spacer (ITS) sequences, this strain was preliminarily identified as Penicillium pinophilum (which is now classified in the genus Talaromyces and has been renamed Talaromyces pinophilus). However, the variability of ITS sequences is insufficient to distinguish among closely related species [22]. To comprehensively investigate the function of fungus JP-NJ4, the classification status of this strain was investigated further. The identification process for strain JP-NJ4 involved many standard strains (type strains) that are currently stored at the Central Bureau of Fungal Cultures (Centraalbureau Voor Schimmelcultures (CBS)), which is part of the Royal Netherlands Academy of Arts and Sciences and was founded in 1904 by the Association Internationale des Botanistes [25]. Currently, CBS is one of the largest mycological research centers in the world, with more than 60,000 species in cultivation, including the type strains of many filamentous fungus and yeast species. Here, after reviewing the literature and observing the characteristics of fungus JP-NJ4, this strain was identified and described by referring to the standard research method (GCPSR) recommended in previous international research on filamentous fungal species such as Penicillium and Talaromyces, etc.

Source of the Strain
The strain JP-NJ4 was a phosphate-solubilizing fungus isolated from rhizosphere soil of Pinus massoniana (yellow brown soil) in the back mountain of Nanjing Forestry University. The strain is now stored in the China Center for Type Culture Collection (CCTCC) (http: //www.cctcc.org, accessed on 18 January 2022). Holotype with the preservation number M 2012167 was stored in a metabolically inactive state by cryopreservation [26,27].

DNA Extraction, PCR Amplification, and Sequencing of Strain JP-NJ4
Strain JP-NJ4 was cultured on malt extract agar (MEA) culture medium at 25 • C for 7-14 days. Genomic DNA was extracted and purified according to the method of Cubero et al. [28], and the extract was stored at −20 • C. The DNA barcode markers required for the identification of JP-NJ4 strain included the ITS region and BenA, CaM, RPB1, and RPB2 genes [29][30][31][32][33][34][35][36][37][38]. The primers needed for the amplification of these genes are shown in Table  S1. All primers and polymerase chain reaction (PCR) amplification sequences needed for the experiment were synthesized and sequenced by the Shanghai Sangon Company (http://www.sangon.com, accessed on 18 January 2022).
In this study, a 50.0 µL DNA amplification thermal cycling reaction mixture system was selected, and the formula of 20.0 µL reaction system was also provided. The volumes of the components in the system are as follows: premix Taq™ solution 25.0 µL, DNA template (10 ng/µL) 2.5 µL, forward primer 2.5 µL, reverse primer 2.5 µL and dd H 2 O 17.5 µL for 50.0 µL system; premix Taq™ solution 10.0 µL, DNA template (10 ng/µL) 1.5 µL, forward primer 1.0 µL, reverse primer 1.0 µL, and dd H 2 O 6.5 µL for 20.0 µL system. Premix Taq™ (Ex Taq ™ Version 2.0 plus dye) is a 2x concentration mixed reagent of DNA polymerase, buffer mixture, and dNTP mixture required for PCR reactions purchased from Takara company (https://takara.company.lookchem.cn/, accessed on 18 January 2022). The concentration of the ingredients in the Premix Taq™ solution is as follows: Ex Taq Buffer (2×conc.) with Mg 2+ at a concentration of 4 mM (mmol/L); highly efficient amplification DNA polymerase (TaKaRa Ex Taq) at a concentration of 1.25 U/25 µL; the dNTP (deoxy-ribonucleoside triphosphate) Mixture (2×conc.), with a concentration of 0.4 mM (mmol/L) for each base; additional pigment markers (Tartrazine/Xylene Cyanol FF), specific gravity additaments, and stabilizers were included. The reagent is stored at −20 • C. The total amount of DNA template can be 10-100 ng, and it can be added according to the experimental requirements. The concentration of primer prepared in accordance with the operational guidelines is 100 µmol/L, diluted 10 times to 10 µmol/L for use.
The DNA amplification thermal cycling programs for each gene is as follows: Standard PCR was selected for general ITS, BenA, and CaM, with initial denaturing 94 • C for 5 min, cycles 35 of denaturation 94 • C for 45 s, annealing 55 • C (52 • C) for 45 s, elongation 72 • C for 60 s, final elongation 72 • C for 7 min, and rest period 10 • C, ∞. Touch-down PCR was selected for RPB1, with 5 cycles of 30 s denaturation at 94 • C, followed by primer annealing for 30 s at 51 • C, and elongation for 1 min at 72 • C; followed by 5 cycles with annealing for 30 s at 49 • C and 30 cycles for 30 s at 47 • C, finalized with an elongation for final 10 min at 72 • C, rest period 10 • C, ∞ (the denaturation and elongation conditions of the second and third cycles are the same as those of the first cycle). Touch-up PCR (= step-up PCR) was selected for RPB2, with initial denaturing 94 • C for 5 min, followed by 5 cycles of 45 s denaturation at 94 • C, primer annealing for 45 s at 50 • C (48 • C), and elongation for 1 min at 72 • C; followed by 5 cycles with annealing for 45 s at 52 • C (50 • C) and 30 cycles for 45 s at 55 • C (52 • C), finalized with an elongation for final 7 min at 72 • C, rest period 10 • C, ∞ (the denaturation and elongation conditions of the second and third cycles are the same as those of the first cycle). The values in parentheses refer to alternative reaction conditions.

Phylogenetic Tree Construction of Strain JP-NJ4
Sequences of five genes from strain JP-NJ4 have been sequenced and deposited in GenBank (Table 1). By conducting a Basic Local Alignment Search Tool (BLAST) search in National Center for Biotechnology Information (NCBI) database (https://www.ncbi. nlm.nih.gov, accessed on 18 January 2022), the results showed the best matched DNA sequences for each gene/region. In order to make phylogenetic trees, the type strains of Talaromyces species were added. For monogenic and polygenic phylogeny, ITS, BenA, CaM, RPB1, and RPB2 sequence data were compared and aligned using ClustalW software included in the MEGA package version 6.0.6 [39]. All datasets (DNA sequences) were concatenated in MEGA and the BioEdit Sequence Alignment Editor software (Version 7.0.9.0) [40]. The aligned data sets were analysed using both Maximum Likelihood (ML) and Bayesian inference (BI) methods, and ML phylogenetic trees were constructed for each gene/region and concatenated polygenic sequences. According to the results of Akaike Information Criterion (AIC) calculated in MEGA package, the best model for ML phylogenetic tree construction is selected. The ML analysis is performed, and the trees were constructed by calculating the initial tree (constructed by the BioNJ method), selecting the Nearest-Neighbour-Interchange (NNI) option for the following heuristic search. Bootstrap analysis was performed on 1000 repetitions to calculate the support at the node. Bayesian Inference phylogenies were inferred using PhyloSuite v1.2.1 [41]. ModelFinder was used to select the best-fit model (2 parallel runs, 2,000,000 generations) using Bayesian Information Criterion (BIC) for BI [42]. The sample frequency was set at 100, with 25% of trees removed as burn-in. Bayesian inference posterior probabilities (BIpp) values and bootstrap values are labelled on nodes.    [14,15], the type strains and other related strains were selected. The current sectional classification information of Talaromyces species was also marked in Table 1.

Observation on the Morphological Characteristics of Strain JP-NJ4
Important features used to describe the large group of Penicillium and its related fungi are as follows: Macromorphology, including colony texture, mycelium growth and color, shape, color, abundance, and texture of conidia, the presence and color of soluble pigments and exudates, the reverse color of the colony and the acid production of the strain on creatine sucrose agar (CREA) [43], etc. Micromorphology, including asexual sporulation structures (e.g., conidiophore) and sexual sporulation structures (e.g., cleistothecium), etc. To comprehensively investigate the growth of strain JP-NJ4 on different media, we formulated the following supplemented-medium types (see Table S2) from common media, these media can be used to observe other taxonomic characteristics of strains.
Czapek yeast autolysate (CYA) [13] and malt extract agar (MEA) [8] are two standard media recommended for species identification of Penicillium and related filamentous fungi.
Czapek's agar (CZ) [16] CZ is the medium used by Raper and Thom (1949) and Ramírez (1982) in taxonomic studies [44]; this also includes Blakeslee's Malt extract agar (MEAbl) of Blakeslee (1915) [45]; yeast extract sucrose agar (YES) [43]; and YES as the recommended medium for the analysis of species' extracellular secretions (extrolites). Oatmeal agar (OA) [8] and Hay infusion agar (HAY) medium [46] were also included. Sexual reproduction of fungal strains most often occurs on OA and HAY media, which can provide valuable information for taxonomy. Use oatmeal/flakes for OA and dry straw for HAY. Creatine sucrose agar (CREA) is the production of acid that can be observed by color reactions (ranging from purple to yellow) in CREA, which are often useful for distinguishing closely related species. Dichloran 18% Glycerol agar (DG18) [47] and Czapek Yeast Autolysate agar with 5% NaCl (CYAS) [18] were used. DG18 and CYAS were used to detect the growth rate of the strain under low water activity.
Preparation for macromorphology observation: The strain JP-NJ4 was inoculated in Potato dextrose agar (PDA) medium and cultured at 15 • C for 25 days to collect conidia. Conidia were washed with distilled deionized water (dd H 2 O) and diluted with a semi-solid agar solution containing 0.2% agar and 0.05% Tween 80 to prepare the conidia suspension, which was stored at 4 • C for standby use [13]. Conidia suspension was extracted with a micropipette (Eppendorf) and inoculated in three points (1 µL per point) [18]. All media were incubated at a constant temperature of 25 • C for 7 days; each formula of medium is shown in Table S2. In addition, the Czapek Yeast Autolysate agar (CYA) was cultured at 30 • C and 37 • C, and Malt Extract agar (MEA) was cultured at 30 • C, and the data were recorded for the species identification of strain JP-NJ4. After 7 and 14 days of strain culture, the criss-cross method was used to measure the colony diameter.
Preparation for the micromorphology observation: Colonies of strain JP-NJ4 cultured on MEA for one to two weeks in a dark environment at 25 • C were used for micromorphology observation, and OA and HAY medium were used when ascomata were not observed on MEA. OA and HAY media are often used for the observation of ascocarp, ascus and ascospore [31,[48][49][50] and may be cultured for up to 3 weeks if required for ascocarp production. Then, the colonies used for micromorphology observation were rinsed with 2mL 0.1 mol/L phosphate-buffered saline (PBS) three times. Lactic acid (60%) was used as the fixative. Since most species produce large amounts of hydrophobic conidia, 70% ethanol is usually used to flush out excess conidia and prevent air from getting trapped in lactic acid between the slide and cover. The characteristics of strain JP-NJ4 were observed with a compound microscope (Axio Imager M2.0; Zeiss, Germany) equipped with a digital camera (AxioCam HRc; Zeiss, Germany). The colonies were dehydrated in a graded ethanol solution and dried with liquid carbon dioxide at a critical point (EmiTech K850). After gold spraying (Hitachi E-1010), the Micromorphology of strain JP-NJ4 (conidiophore, conidium, ascocarp, ascus, and ascospore) was observed by scanning electron microscope (SEM) (FEI Quanta 200, FEI, USA).

Phylogeny-Based Species Identification
With the help of concatenated phylogenetic trees based on five gene regions, including the internal transcribed spacer region, BenA, CaM, RPB1, and RPB2, we investigated the taxonomic position of strain JP-NJ4. Figures 1-3 and Figures S1-S4 show the phylogenetic relationships among strain JP-NJ4 and representative species of Talaromyces. Concatenated phylogenetic trees of five (ITS, BenA, CaM, RPB1, and RPB2) and four (ITS, BenA, CaM, and RPB2) gene regions and individual phylogenetic trees of each gene region were constructed using the maximum-likelihood method. Talaromyces dendriticus (CBS_660.80_T) was chosen as an out-group for Talaromyces section Talaromyces. Trichocoma paradoxa (CBS_788.83_T) was chosen as the out-group for the Talaromyces genus. Bootstrap values obtained from 1000 replications are shown at the nodes of the tree, and bootstrap support lower than 50 is not shown. In the multi-gene phylogenetic analysis (five gene region), strain JP-NJ4 clustered with T. liani (Figure 1) in Talaromyces section Talaromyces (orange area), with bootstrap values of 100% (BIpp = 1). The concatenated phylogeny of five gene region shows that strain JP-NJ4 and T. liani differ in their genetic distance from other species of Talaromyces. Phylogenetically, the results of four genes indicate that strain JP-NJ4 T. nanjingensis is close to T. brevis, with bootstrap values of 93% (BIpp = 1) (Figure 2).
Among the phylogenetic trees obtained from each DNA gene region, that of the ITS region was less clearly resolved; although most species formed monophyletic groups in the strict consensus trees, several had low bootstrap support values. The ITS sequence of strain JP-NJ4 clustered with those of nine other strains of T. liani and three strains of T. brevis (bootstrap = 32%, BIpp = 0.61) ( Figure S1). The CaM sequence of strain JP-NJ4 clustered well with two strains of T. brevis (bootstrap = 65%, BIpp = 0.99) (one strain of T. brevis was deleted because its sequence was shorter) and seven other strains of T. liani (bootstrap = 99%, BIpp = 0.96) ( Figure S2). The RPB1 sequence of strain JP-NJ4 clustered with that of the type strain of T. liani (CBS_225.66_T) (bootstrap = 85%, BIpp = 0.99) ( Figure S3). The RPB2 sequence of strain JP-NJ4 clustered perfectly with three strains of T. brevis (bootstrap = 99%, BIpp = 1) ( Figure S4). The phylogenetic tree of the CaM gene region shows that strain JP-NJ4 and T. liani differed little in their genetic distance from other species of Talaromyces. However, the phylogenetic trees of the BenA, RPB1, and RPB2 gene regions show that strain JP-NJ4 and T. liani differed markedly in their genetic distance from type strain of T. liani and other multiple collections of T. liani.       (Figure 3). This indicates that T. nanjingensis is still genetically different from its genetic relatives T. liani and T. brevis. The sequences of T. nanjingensis and 'T. liani' (voucher KUC21412) were significantly similar in BenA gene. However, T. nanjingensis and 'T. liani' (voucher KUC21412) differ from T. liani and T. brevis in BenA gene by more than ten bases, and half of their base arrangement pattern is similar to T. liani and the other half is similar to T. brevis ( Figure S8a). This could mean they should be new species. It also explains the low bootstrap values. This is also due to the continued discovery of new species, filling gaps in the evolutionary trees, and the lack of some transitional species, of which the T. nanjingensis is one, which has characteristics common to both T. liani and T. brevis. T. nanjingensis is more similar to T. brevis in acid production. T. liani (CBS_118885) is the only acid-producing strain of T. liani that is genetically closest to T. nanjingensis. The phenotypic information of these species may also hint at evolutionary continuity. After a detailed search, we found that T. liani strain T2C1 is equivalent to 'T. liani' (voucher KUC21412) and ITS sequence was also obtained. 'T. liani' (voucher KUC21412) has at least two base differences with T. nanjingensis, T. liani, and T. brevis in the ITS region, and the front-end of the sequence is similar to that of T. liani and T. brevis (CBS_141833_T). In particular, it has a distinctive differential base A at the end of its sequence ( Figure S8b). Therefore, 'T. liani' (voucher KUC21412) is also different from T. nanjingensis. 'T. liani' (voucher KUC21412) is described by Heo et al. as one of the microorganisms selected from intertidal mudflats and abandoned solar salterns that can produce bioactive compounds [52]. The strain voucher KUC21412 was described as 'T. liani' (with quotation marks) in this manuscript, as its current species identity may be in some doubt. 'T. liani' (voucher KUC21412) is a comparable species, although only ITS (MN518409.1) and BenA sequences (MN531288.1) have been submitted to NCBI, the quality of the sequences is reliable, which proves that the BenA sequence of T. nanjingensis is reliable. Although T. nanjingensis had little difference with T. brevis in ITS, CaM, RPB1, and RPB2 genes, it had great difference with T. brevis in BenA gene ( Figure S8a). BenA is the secondary barcode with the highest reliability in filamentous fungi; thus, the classification results obtained by using this gene are relatively more referential and accurate [15,18]. According to the Bayesian tree (ITS + BenA + CaM + RPB2 and single gene BenA) and the ML tree with deletion of T. brevis and 'T. liani' (voucher KUC21412), it can be more obvious that T. nanjingensis has a long genetic distance from T. brevis and T. liani (the small diagrams in Figures 2 and 3).
More obviously, the phylogenetic tree of BenA gene with long sequence version ( Figure 3B) was better than that of the BenA gene with short sequence version ( Figure 3A) to show the true classification status of strain JPNJ4. The results showed that strain JPNJ4 was quite different from T. brevis and T. liani in phylogeny ( Figure 3B). The long sequence version of the phylogenetic tree ( Figure 3B) only reduced or lost information on some other species, but this did not affect the accurate identification of JP-NJ4, because this version included T. brevis and T. liani. New species resources are undoubtedly important, as are innovations in identification methods and fine-delineation of species. Taxonomy of these species of Talaromyces are similar to the results obtained by Samson et al. (2011) and Yilmaz et al. (2014) [14,15]. These phylogenetic results suggest that strain JP-NJ4 is a potential novel species.

Species Identification Based on Macromorphology and Micromorphology
By combining these results with morphological observations, the taxonomic position of strain JP-NJ4 can be further elucidated. The macromorphology of strains, including the morphology and diameter of colonies on specific media, is an important trait for species identification. Based on the preliminary results on CYA and MEA, the macromorphology of the strain may be observed on several other culture media for more accurate identification.
We selected as many media as possible to observe the strains in more detail. Czapek (CZ) medium was used in early taxonomic studies of Penicillium and was selected for comparison with CYA medium. Blakeslee's MEA, which has been widely used historically, was compared with the MEA culture medium used by the CBS-KNAW Fungal Biodiversity Centre in Utrecht (Table S2). Medium with the addition of hay (HAY) was compared with oatmeal agar for observing the sexual reproduction of fungal strain JP-NJ4. However, strain JP-NJ4 did not grow on this recommended medium.
To clearly observe the colony morphology of strain JP-NJ4 on various culture media, we obtained photographs with black and white background colors. In this paper, we provided colony morphology photographs of strain JP-NJ4 grown at 25 • C for 7 (Figures 4 and S5) and 14 d (Figures 5 and S6) on 10 different media, as well as image data for strain JP-NJ4 grown at 25 • C, 30 • C, and 37 • C in CYA medium ( Figure S7).  Reverse colonies of Talaromyces species on CYA and MEA media commonly produce yellow or red soluble pigments. Numerous species of Talaromyces, including T. albobiverticillius, T. amestolkiae, T. atroroseus, T. cnidii, T. coalescens, T. marneffei, T. minioluteus, T. pittii, T. purpurogenus, T. ruber, and T. stollii produce soluble red pigments. In T. nanjingensis strain JP-NJ4, weak production of yellow and orange soluble pigments was observed on CYA and MEA in colonies grown at 25 • C for 7 d (Figure 4 and Figure S5). The reverse colony color on MEA was similar to those of T. albobiverticillius, T. minioluteus, and T. purpurogenus. Strong and stable red soluble pigment production occurred on MEA in colonies grown at 25 • C for 14 d ( Figure 5 and Figure S6). The reverse colony color on MEA in colonies grown at 25 • C for 14 d was similar to those of T. amestolkiae, T. coalescens, and T. marneffei grown on MEA at 25 • C for 7 d. T. nanjingensis strain JP-NJ4 could produce acid on CREA at 25 • C for 7 d (present strong, the color reaction showed a marked shift from purple to yellow) and at 25 • C for 14 d (present very strong; the color reaction appears to be intense yellow).
Talaromyces species have positive impacts in the medical field. The members of this genus can produce a variety of antibiotics and antibacterial substances, such as the rugulosin produced by T. rugulosus [11,63]. Other extrolites of the genus (e.g., erythroskyrine, etc.) have anti-tumor [64], anti-malignant cell proliferation (antiproliferative), and anti-oxidant properties [65]. Talaromyces fungi also have a strong ability to produce enzymes, including that of β-rutinosidase and phosphatase [66,67], endoglucanase and cellulase [68], cellulase [69][70][71], and others. These fungi have also been investigated for functions in plant disease resistance, such as T. flavus [72][73][74][75] and T. pinophilus [76]; moreover, this includes the plant growth promotion of T. pinophilus [77]. In the present study, fungal strain JP-NJ4, which was isolated from the rhizosphere soil of Pinus massoniana, exhibited abilities of phosphate solubilization and plant growth promotion [24], and it was identified as a novel species in genus Talaromyces, section Talaromyces, using the polyphasic approach in this manuscript.
The fungal genera of Penicillium and Talaromyces have many similarities in morphology, such as asexual sporulation structures (e.g., conidiophore), the branching pattern of conidiophores, namely the type of penicillus, and sexual sporulation structures (e.g., cleistothecium). Mistakes are easily made when distinguishing between them. Therefore, we can use molecular methods to conduct preliminary identification of species in these genera. It should be noted, for the modern taxonomic identification of a species, morphological characteristics and molecular phylogenetic results are equally important. Professional recommendations regarding appropriate phylogenetic and morphological data in species delineation are necessary to avoid taxonomic discrepancies [78]. Phylogenetic trees of species are constructed using extensive data obtained through searches and literature review. Normally, the first step is to input a nucleotide sequence obtained through PCR and sequencing technology into the NCBI website for comparison using the nucleotide BLAST. Using the default settings, we obtained 100 sequences that are most similar to the target sequence. The purpose of this step is to roughly determine the genus of the unknown strain. In this paper, BLAST analysis was conducted using ITS, BenA, CaM, RPB1, and RPB2 sequences of strain JP-NJ4.
During the process of collecting and collating the sequences needed to build phylogenetic trees, we encountered the following problems. Among the sequences submitted to NCBI, for the same gene from the same strain of the same species, sequences were uploaded under multiple sequence numbers. By conducting BLAST analysis of the sequence and preliminary phylogenetic tree construction, we found that some of the sequences were consistent with the earliest submitted sequences, whereas others did not cluster with the type strains of their species. This difference may be due to misidentification by later sequence submitters or mislabeling of different strains as the same strain. Therefore, when selecting sequences to construct phylogenetic trees, if two sequences are obtained with differing base compositions, we used the sequences submitted earlier or those referenced in the authoritative literature. Only using validated sequences is also reliable. If the sequences were identical, they were all retained in the tables used to build the phylogenetic tree ( Table 1).
The ITS region is the most commonly used molecular marker for fungal identification. In T. liani, NRRL 1014 and NRRL 1015 are equivalent to NRRL 1009, and the base sequences of the ITS region and the other four specific genes in the three strains are identical. Therefore, when constructing the ITS phylogenetic tree for strain JP-NJ4, NRRL 1009 was selected to represent all three strains [79]. In addition, nine other T. liani strains were added. By conducting sequence alignment analysis of the CaM gene, we found notable differences in the composition and arrangement of the bases in this gene among species in different sections of genus Talaromyces. This may result in the deletion of too many bases in order to ensure sequence alignment in the tree constructing of strain JP-NJ4 at the genus level, resulting in loss of information. Therefore, in order to ensure the length of a CaM sequence in tree construction and improve the accuracy of species identification, the CaM gene phylogenetic tree was constructed at the level of section Talaromyces. We further determined the taxonomic status of strain JP-NJ4 by evaluating the taxonomic relationships among these highly similar species within the genus Talaromyces. In addition, during the Alignment-Align process of ClustalW in MEGA software (Version 6.0 and 7.0), inaccuracies may be introduced into the alignment results when large differences exist among the sequences. Therefore, the best comparison results can be obtained through multiple repeated comparisons.
We also found that the gene sequences of RPB2 from some type strains of Talaromyces species could not be retrieved from the NCBI database. By performing comparison and analysis of the RPB2 sequences of other species in genus Talaromyces, we found that the gene sequence data of RNA polymerase (RNA polymerase gene, partial cds) downloaded for these type strains included the gene sequence of RPB2. Therefore, these RNA polymerase gene sequences can be used to complement the construction phylogenetic trees based on the RPB2 gene. Moreover, in previous studies of Penicillium and Talaromyces [14,15,18,31,80], the precedent of using the RNA polymerase gene sequence for constructing a RPB2 phylogenetic tree has been established (e.g., JX315698 Talaromyces amestolkiae DTO 179F5_T). Using this method, the taxonomic status of unknown species can be further refined. The sequences used for this analysis include the following: KX961275 Talaromyces 78. Some specific genes, such as Translation elongation factor (Tef ) and mitochondrial Cytochrome c oxidase 1 (Cox1), have not been universally used in Talaromyces, and relatively few sequences for these genes are available from the NCBI database. Currently, although phylogenetic trees of the genus Talaromyces constructed from these remain imprecise, the genes have been used for identifying Penicillium species [6].
When constructing phylogenetic trees, it is necessary to delete redundant and irrelevant sequences. In the BLAST comparison results, the sequences related to some species did not include the corresponding type strains. In such cases, the sequence information should be validated, as the sequences might have been misidentified (wrongly identified as another species). For example, in Table 1, species marked with a yellow background color did not cluster with the type strains of the corresponding species, and phylogenetic results indicate that these species may be new species-Talaromyces_stollii (blue font) ( Figure S4). This discrepancy is due to the fact that not all sequences in the NCBI database have been verified. Therefore, type strains of these species should be added as references for molecular identification and construction of phylogenetic trees. Here, we selected sequences from the type strain of T. pinophilus and other related strains, and some sequences of T. pinophilus that were not relevant to our study were removed.
In addition, when building phylogenetic trees, if the sequences used to construct the tree are not sufficiently comprehensive, the strain to be identified will only cluster with the sequences of similar species, rather than the sequence of the closest species. This problem occurs because the sequences closest to that of the strain to be identified at the genetic level may not be included in the NCBI-BLAST results due to differences in the length of the uploaded sequences or differences in gene coverage, resulting in an inaccurate phylogenetic tree. Specifically, analysis of NCBI-BLAST results revealed that most sequences included only partial sequences of a gene (not all the bases of the gene). The uploaded gene sequences are inconsistent in length, and each sequence contains a different region of the full-length gene. These differences result in the common phenomenon of sequences that appear most similar in the alignment results not being those that are actually most similar to the destination sequence (i.e., the results are inaccurate).
In summary, in previous international research on filamentous fungal species such as Penicillium and Talaromyces, the standard research method (GCPSR) was recommended. This polyphasic approach, which involved multigene phylogeny, morphological descriptions using macro-morphological and micro-morphological characters. To build an accurate phylogenetic tree based on NCBI-BLAST sequences, it is essential to refer to sequences provided in the authoritative literature. For the gene sequences of type strains, the selected sequences should be validated or verified. Using ITS and four specific gene sequences in various Talaromyces species, we constructed two phylogenetic trees (tree 1: ITS, BenA, CaM, RPB1, and RPB2 ( Figure 1); tree 2: ITS, BenA, CaM, and RPB2 ( Figure 2)) based on combinations of multiple genes. In the genus Talaromyces, combinations of three or four genes are more common, whereas analyses of five genes have been rare. At present, ITS, BenA, CaM, RPB1, and RPB2 are the most authoritative and reliable genes for the identification of Talaromyces species. The preliminary phylogenetic tree construction results indicate that the species most closely related to strain JP-NJ4 is T. liani. The concatenated phylogenies of five (or four) gene regions and single gene phylogenetic tree (BenA, RPB1, and RPB2 genes) all also show that T. nanjingensis strain JP-NJ4 and T. liani clustered together but differ markedly in their genetic distance from type strain of T. liani and other multiple collections of T. liani. The morphology of JP-NJ4 (M 2012167) largely matches the characteristics of T. liani, but the rich and specific morphological information provided by its colonies was different from that of T. liani. In addition, strain JP-NJ4 could produce reduced conidiophores with solitary phialides. From molecular and phenotypic data, strain JP-NJ4 was identified as a putative novel Talaromyces fungal species, designated T. nanjingensis. T. nanjingensis also can produce yellow, orange, and red soluble pigments in their mycelium, including diffusing pigments, similar to other species of the genus [81,82]. Due to the rich and specific morphological information provided by colonies, additional colony morphology photographs of this strain growing at 25 • C for 14 days on 10 different media were captured. We believe that it is essential to apply this information as part of the general method of strain identification. Future research will focus on the ecological function of T. nanjingensis JP-NJ4 and its impacts on the environment in terms of ecological security will also be assessed.
The information of the culture preservation institutions involved is as follows (alphabetically):  Figure S1: Maximum likelihood phylogeny of ITS regions for strain JP-NJ4 and other species classified in Talaromyces sect. Talaromyces. Talaromyces dendriticus (CBS_660.80_T) was chosen as out-group. Support in nodes is indicated above branches and is represented by posterior probabilities (BI analysis) and bootstrap values (ML analysis). Full support (1.00/100%) is indicated with an asterisk (*). Bootstrap values lower than 50 is hidden. Best-fit model of Bayesian Inference phylogeny according to BIC: GTR+F+I+G4; Best-fit model of Maximum likelihood phylogeny according to AIC: Tamura 3-parameter (T92) +G+I; alignment, ITS 467 bp. Scale bar: 0.0020 substitutions per nucleotide position. T indicates ex type. The strain with red font is the strain JP-NJ4 to be identified. Figure Figure S8: Base alignment and differences in specific genes of Talaromyces nanjingensis, T. brevis and T. liani. A. BenA gene, alignment, 348 bp; B. ITS region, alignment, 448 bp. Table S1: Primers for amplification and sequencing of ITS and specific genes in strain JP-NJ4.