Talaromyces columbinus sp. nov., and Genealogical Concordance Analysis in Talaromyces Clade 2a

During the course of mold surveys, a set of Talaromyces isolates were obtained that did not fit any described species. Phenotypic examination of these isolates showed that they were similar to T. piceus but differed in some growth characteristics. Multilocus DNA sequence data were obtained for the new isolates and some related species in the broader, more inclusive clade, and the data were analyzed using genealogical concordance. The new isolates are described as Talaromyces columbinus. From analysis of the related species, Penicillium rugulosum var. atricolum is given species status in Talaromyces as T. atricola. Penicillium tardum and P. chrysitis were showed to be synonyms of T. rugulosus. Penicillium scorteum and T. phialosporus were showed to be conspecific and under the rule of priority T. scorteus is the proper name for isolates previously known as T. phialosporus. Talaromyces wortmanii was showed to be distinct from Penicillium concavorugulosum and T. variabilis but the relationship of the latter two species remains unresolved. Examination of ITS sequences from GenBank showed that T. columbinus has previously been reported from human lung infections under the name Penicillium piceum.


Introduction
The genus Penicillium is widely known and studied because of its impact on human activities [1]. Historically, the broad sense of Penicillium has contained two very different groups, the first group containing monoverticillate, furcate and terverticillate species with broad ampuliform conidiogenous cells, and the second group containing biverticillate species with acerose conidiogenous cells. Where known, the two groups produced two different types of teleomorphs, either closed structures characterized by walls containing pseudoparenchymatous cells (Eupenicillium) or ascocarps having loosely woven hyphae forming the wall (Talaromyces). Several investigators e.g., [2] demonstrated the phylogenetic distinction between these two groups of Penicillium. Under the Melbourne nomenclatural code [3] dual naming was revoked in favor of using a single name for a single species. Samson et al. [4] formally recognized the phylogenetic distinction of the species with acerose conidiogenous cells and/or loosely woven ascocarp walls by placing them in the genus Talaromyces. Talaromyces is based on phylogenetic relationships, not whether the teleomorphic stage is known. Among the species placed in Talaromyces was Penicillium piceum under the new combination Talaromyces piceus.
Pitt [1] regarded Penicillium piceum as a relatively uncommon but widely distributed species. The species has been reported as causal organism in certain human lung infections [5,6] and has been examined as a potential source of extracellular catalase [7] and isochaetochromin [8]. P. piceum has also been investigated as an indicator species for use in forensic science [9]. New Talaromyces species are being discovered as the molecular tools for phylogenetics are more widely applied to this genus [10][11][12].
During the course of fungal surveys conducted by one of us (ZJ) certain isolates resembling T. piceus were found that did not perfectly fit the description [1] and were retained for further investigation. Subsequent detailed examination showed that these isolates were representative of a previously unknown species that we describe here as Talaromyces columbinus. Talaromyces piceus fits into a small clade within Talaromyces designated clade 2a by Samson et al. [4]. In assessing the positions of the species within this clade, a number of phenotype-based taxonomic hypotheses [1,[13][14][15] were also tested using genealogical concordance analysis [16].

Isolations
Isolation of strains from air was performed by using a single stage bio-aerosol impaction sampler (EMSL VP-400 Microbial Sampler) for the viable sampling of fungi. The sampler contains over 400 precision drilled holes and a base section for placement of the agar media. Air was passed through each sampler with a flow rate of 28.3 L/minute for 3 minutes. The media used for fungal isolations was malt extract agar.

Culture Methods
Isolates (Table 1) were grown on Czapek yeast extract agar (CYA), CYA with 20% sucrose (CY20S), malt extract agar (MEA), oatmeal agar (OA), dichloran-glycerol agar (DG18), Czapek yeast  [18]. The Ridgway [19] color guide was used to specify certain colors and those are referred to by plate number, e.g. (R47). Weighted regression analysis was used to model and compare the growth of 2 fungal species, Talaromyces piceus and T. columbinus, using the mean of 6 replications of colony diameter measurements (mm) as a function of temperature (uC). Standard weighting of 1/variance was used for each mean colony diameter at each temperature tested. A full and reduced model F-test was used to determine if the equations for each species were significantly different from one another. If a significant F-test value was obtained (indicating T. piceus was different from T. columbinus colony growth), regression coefficients were compared between the 2 species as well as colony diameters at each temperature using 95% confidence interval overlap.

Microscopy
Microscopic examination was performed by gently pressing a ca 2065 mm piece of transparent tape onto a colony, rinsing the tape with one or two drops of 70% ethanol and mounting the tape in lactic acid with fuchsin dye. Additional microscopic samples were made by teasing apart a small amount of mycelium in a drop of water containing 0.5% Tween 20. A Leica DM 2500 microscope with bright field, phase contrast and DIC optics was used to view the slides. A Spot camera with Spot imaging software was mounted on the microscope and used for photomicrography. A Nikon digital SLR camera with a D70 lens was used for colony photography. Photographs were re-sized and fitted into plates using Microsoft PowerPoint 2010.

Phylogenetic Analysis
Strains used in phylogenetic analysis (Table 1) were grown in 25 mL of 2% malt extract broth in 125-mL Erlenmeyer flasks shaken at 200 rpm (25-28uC). Mycelium was harvested after 1-2 days growth by vacuum filtration over Whatman #1 filter paper and then placed loosely in microfuge tubes, frozen and freeze dried. The freeze-dried biomass was ground to a fine powder and rehydrated with 0.5 mL CTAB buffer [20]. Proteins were extracted by the addition of 0.5 mL chloroform. After brief emulsification, the aqueous phase was separated from the organic phase by centrifugation. The aqueous phase was transferred to a clean tube and DNA was precipitated by the addition of 0.5 mL isopropanol. The precipitate was collected by centrifugation and rinsed with 70% ethanol. The pellet was rehydrated with 0.1 mL TE buffer and stored at 220uC until used.
Beta tubulin (BT2), calmodulin (CF), nuclear internal transcribed spacer region (ITS), DNA replication licensing protein (Mcm7), RNA polymerase II (RPB2) and ribosome biogenesis protein (Tsr1) were amplified from 1:10 diluted genomic DNA using previously described primers and conditions [21]. RNA polymerase II largest subunit (RPB1) sequences were obtained for select isolates using published methods [4]. Initial results indicated that the BT2 primers (BT2a, BT2b) were amplifying paralogous genes. The primers (BT2f and T22) and procedures of Hubka and Kolarik [22], developed to resolve this problem in Aspergillus japonicus, were used to conduct repeated amplification and sequencing of the suspect products. This procedure also produced the paralogous gene products.
Amplified DNA was prepared for sequencing reactions with ExoSapit [23]. Sequencing reactions were performed using DyeDeoxy terminator kits (v 3) following the manufacturer's instructions and subsequent analysis on an ABI 3730 DNA analyzer [24]. Sequencing was performed in both directions and any conflicts were resolved using SEQUENCHER 5 [25] to visualize the sequences. Sequences were carefully reviewed for quality and experiments were repeated if there was doubt about the reliability of sequence reads. Sequences were subjected to BLAST search against the GenBank nr database.
Three DNA sequence datasets were aligned using CLUSTALW [26]. One was composed of a wide array of species in the genus Talaromyces clade 2A [4] including only RPB2 data, whose purpose was to show the overall position of the new species in the clade. The other alignments were sequences from T. piceus and T. columbinus isolates for each of the sequenced genes and alignments of T. rugulosus and T. phialosporus and putative synonyms. Aligned datasets were analyzed using maximum parsimony [27] and 1000 bootstraps. Conditions for parsimony analysis were random sequence addition, MAXTREES = 5000 and swap = TBR (tree bisection and reconnection). For bootstrap calculations ADDSEQ = ASIS, MAXTREES = 100 and swap = TBR were set. Talaromyces proteolyticus was used as the out-group species for the analysis based on the study by Samson et al. [4]. Tree files were visualized and converted to emf format using TREEVIEW [28] and redrawn for publication using CorelDraw X6 [29]. Genealogical concordance [16] was assessed through visual comparison of the individual locus trees.

DNA Sequences
DNA sequences determined in this study are deposited in GenBank and accession numbers are found in Table 1.
BLAST searches of GenBank using ITS sequences revealed that T. piceus ex type isolates held by NRRL (Agricultural Research Service culture collection Peoria, IL), CBS (Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands) and ATCC (American Type Culture Collection, Manassas, VA) that originated from the same parental culture were different. CBS 361.48 (GenBank JN899370 ) differs from NRRL 1051 (GenBank KF196893) by six gap positions, all deletions of T residues relative to NRRL 1051; ATCC 10519 (GenBank DQ666826) differs from NRRL 1051 by insertion of two A residues and one T residue relative to NRRL 1051. Five isolates of T. piceus examined in this study (Table 1) shared an identical and unique ITS sequence. The RPB2 sequence of CBS 391.48 (JF417433) should be identical to NRRL 1048 BLAST searches of our new species against the GenBank nr (nonredundant sequence) database provided a 100% similarity of our new species to the ITS sequence from IMI 392509 (DQ666824) (IMI, International Mycological Institute, Egham, UK), a fungus isolated from a human lung infection in Buenos Aires, Argentina [6]. Other gene sequences from our new species (BT2, CF, Mcm7, RPB2, Tsr1 and RPB1) produced no high homology to any sequences in the nr database.
Seven of eight T. columbinus isolates shared an identical ITS sequence (e.g. KF196896) and one isolate had a single transition (KF196900) versus the others. T. piceus isolates shared the same ITS sequence (e.g., KF196893). The common species sequences differed by seven transition base differences, one transversion and one indel. In our analysis, barcode identification [30] of these species is practical.
The RPB2 based phylogeny of Talaromyces clade 2a is shown in Fig. 1. Phylogenetic relationships and topology of single locus trees of the isolates are shown in Fig. 2. The product of BT2 primer amplifications placed isolates of T. piceus and T. columbinus in four statistically supported distinct clades. That tree is notably discordant from the CF, Mcm7 and RPB2 loci (Fig. 2) that place the same isolates in two strongly supported clades. The majority rule of congruence analysis indicates these are distinct species [16].
The phylogenetic relationships of T. phialosporus and Penicillium scorteum, and T. atricola, T. rugulosus, P. chrysitis and P. tardum are shown in Fig. 3. T. phialosporus and P. scorteum are conspecific; T. rugulosus P. tardum and P. chrysitis are conspecific, with strong support from concordance analysis and T. atricola is a distinct species.
Temperature and medium dependent colony growth:-Full and reduced model comparison analysis of temperature dependent growth rates (Fig. 4) showed that the equations for T. piceus and T. columbinus were significantly different from each other at a = .01 and.05 levels. Confidence interval overlap on the regression coefficients a, b, and c showed that intercepts (a) and slopes for X (b) and X 3 (c) aren't different between the equations for the 2 species. It appears that T. columbinus and T. piceus show the same colony growth at the The BT2 tree divides T. piceus and T. columbinus isolates into four statistically supported clades while the other three loci place all T. piceus isolates on a single branch and all T. columbinus isolates on another branch. The BT2 primer set appears to be amplifying different gene fragments or the locus may have undergone rearrangements that make the analysis appear paralogous. Genealogical concordance is seen in a majority of the trees, which supports T. columbinus and T. piceus as distinct species. doi:10.1371/journal.pone.0078084.g002 lower and higher temperatures extremes tested (10 and 50uC) with growth lagging for T. columbinus at 14.5 through 30uC, then surpassing T. piceus at temperatures 35 through 48uC. Talaromyces columbinus isolates showed no growth below 20uC, while T. piceus isolates all produced small colonies at 14.5uC. Maximum growth appears at ca 35uC, where the growth of T. columbinus exceeds that of T. piceus although the individual isolates are variable. At 44.5uC T. piceus isolates showed little or no growth while T. columbinus isolates' colonies were 10-20 mm diam.
Diagnosis: Microscopically similar to T. piceus, producing colonies on CYA with a distinct dark bluish gray color, and growing well at 44.5uC, versus minimal or absent growth at 44.5uC for T. piceus; T. columbinus isolates do not grow on CYA-5% NaCl while T. piceus isolates grow moderately well.
Talaromyces scorteus (Nakazawa, Takeda Neotype: BPI 892679 dried colony preparation of NRRL 1129, isolated from military equipment, designated here. Nakazawa and associates, in accordance with then current rules did not designate a type, but distributed a type isolate (CBS 340.34) which was later obtained by Thom and accessioned in the Agricultural Research Service culture collection as NRRL 1129. This isolate, as dried colonies is designated as the neotype.
Raper and Thom [14] reported the authors of P. scorteum as Takedo, Suematsu and Nakazawa. This error has been perpetuated in the literature e.g., [1,14,31]. An interpretation of the original article, which was written mostly in Japanese, reveals that the authors are Nakazawa, Takeda and Suematsu, fide Prof. Junta Sugiyama.

Discussion
The sequence differences between the three type isolates of T. piceus could arise from sequence reading errors, mishandling of the cultures so that they are no longer identical or on rare occasion copies of the rDNA repeat units have different sequences [32]. The multiple gaps distinguishing the ITS sequences of NRRL 1051 (KF196893), CBS 342.48 (JN899370) and ATCC 10519 (DQ666826) are most easily explained as sequence reading errors. The single base difference between the RPB2 sequences of two T. rugulosus type isolates (NRRL 1048, KF196975 and CBS 391.48, JF417433) appears to be a simple sequence reading error. The ca 20% sequence difference between the RPB1 sequence from two type isolates of P. tardum (NRRL 1073, KF196957 and CBS 258.37, JN899293) is too great to be simple sequencing error and most likely resulted from mishandling of sequence data or cultures. Samson et al. [4] indicated some concern about their sequence from this isolate.
Talaromyces columbinus isolates were collected in air samplers in several widely separated states of the USA and from a sample of corn grits in Illinois. It is widely dispersed in the US. The 100% similarity of the ITS sequence from the lung infection fungus of a patient in Buenos Aires (IMI 392509, DQ666824) with T. columbinus (KF196900) suggests that this infective species is T. columbinus and that the species is present in Argentina, although we have not seen the actual isolate. The patient in Argentina [6] had underlying chronic granulomatous disease that appears to make patients susceptible to opportunistic fungal infection [33]. Another lung infection fungus from a patient in Bonn, Germany [5] was identified by ITS sequence and phenotype. The phenotype suggested P. piceum and the ITS sequence was reported to be similar to that of P. piceum CBS 361.48 but it was not deposited in a public database. The authors of that study [5] made the sequence available to us and it is a 100% match to the sequence from T. columbinus. On this basis we are confident that the German lung infection isolate was T. columbinus. On the basis of 100% ITS sequence similarity, T. columbinus is known from North America, South America and Western Europe.
The phenotypic distinction of T. piceus and T. columbinus relies most notably on the failure of T. columbinus isolates to grow on CYA medium supplemented with 5% NaCl, and by the color of colonies grown at 37uC on CYA (compare Figs. 5,6). The lack of growth on NaCl supplemented medium appears to be a reaction to the salt rather than an osmotic effect since T. piceus and T. There is no growth on NaCl amended CYA medium. Cultures incubated at temperature between 30uC and 46uC produce a dark brown soluble pigment in the agar. The color of the CYA culture at 37uC is the basis of the epithet. E. Conidia mostly globose 2.5-3.5 mm diam and smooth walled with occasional much larger and ovoid individuals, bar = 10 mm. F, G. Penicillus structure with vesiculate stipe, metulae and phialides. Conidia often form conical aggregations apically on the penicillus, bar = 10 mm. doi:10.1371/journal.pone.0078084.g005 columbinus showed moderate growth on DG18 medium (18% glycerol) and CY20S (20% sucrose). Both species grow above 37uC, but T. columbinus grows well at 44uC while T. piceus isolates mostly fail to grow at this temperature. Colony diameters in the middle area of the temperature growth range show differences by species (Fig. 4) but the variation seen in the isolates of each species makes reliance on this one character dubious.
Genealogical concordance analysis of multilocus DNA sequence data from isolates of Talaromyces rugulosus, Penicillium chrysitis and P. tardum revealed that they are conspecific. Penicillium rugulosum, the basionym of T. rugulosus was published in 1910 (MB 210907) and has priority over P. chrysitis Biourge published in 1923 (MB 260588) and P. tardum Thom published in 1930 (MB 279778). Samson et al. [4] placed P. tardum in their tree diagram in a different clade along with T. pinophilus. We carefully checked our data and fungal isolates and are confident that our placement of P. tardum in synonymy with T. rugulosus is correct.
Penicillium rugulosum var. atricolum Thom (MB 274357) was found in concordance analysis to represent a distinct species sibling to T. rugulosus and has been elevated to species status in Talaromyces as T. atricola.
Fungal nomenclature is based on type specimens and publication of names based on those types. If two people should describe the same new species under different names, the name proposed by the author first to publish is chosen as the legitimate name using the principle of priority. There are provisions in the nomenclatural code for conserving a name that may be synonymous with an earlier published name. Generally a name can be conserved if it is associated with significant industrial processes or very widely embedded in literature. A strong case for the advantages of keeping a later synonym as the legitimate name must be made and ruled on [34]. Frisvad et al. [35] proposed preserving the names Penicillium chrysogenum (penicillin producer), Aspergillus niger (enzyme and citrate producer) and Aspergillus nidulans (a model genetic system) and protected status was granted [34]. Slightly later Pitt and Samson [31] generated a comprehensive list of commonly used names in Aspergillus and Penicillium and asked for protected status for all those names. Protection for the names was not granted [34], but following the suggestions of Pitt and Samson [31] was recommended to promote taxonomic stability.
Penicillium scorteum is listed by Raper and Thom [14], Pitt [1] and Pitt et al. [13] as a synonym of P. rugulosum. However as our data analysis (Fig. 3) shows P. scorteum is conspecific with T. phialosporus. Publication of P. scorteum in 1934 (MB 492647) predates the publication of P. phialosporum in 1959 (MB 302415). Talaromyces phialosporus is accepted in the list of names in common use (NCU) [31]. However, Talaromyces phialosporus is not a commonly encountered species [36] and no widely known industrial or medical process is associated with this name. Penicillium scorteum is also a name not often encountered. In this case we adhere to the rule of priority rather than the NCU as neither P. scorteum nor T. phialosporus are commonly used or reported names and this name change will not have widespread disadvantageous effect.
Samson et al. [4] presented a tree diagram based on RPB1 sequences that portrayed T. variabilis, P. concavorugulosum and T. wortmanii as identical. In Fig. 1 there is a clear distinction of each of these species based on RPB2 sequences. Strong statistical support shows that T. wortmanii is a distinct species, but additional analysis based on more isolates and more loci is required to establish the relationship of T. variabilis and P. concavorugulosum under the genealogical concordance paradigm.
The electronic version of this article in Portable Document Format (PDF) in a work with an ISSN or ISBN will represent a published work according to the International Code of Nomenclature for algae, fungi, and plants, and hence the new names contained in the electronic publication of a PLOS ONE article are effectively published under that Code from the electronic edition alone, so there is no longer any need to provide printed copies.
In addition, new names contained in this work have been submitted to MycoBank from where they will be made available to the Global Names Index. The unique MycoBank number can be resolved and the associated information viewed through any standard web browser by appending the MycoBank number contained in this publication to the prefix http://www.mycobank. org/MycoTaxo.aspx?Link = T&Rec = . The online version of this work is archived and available from the following digital repositories: PubMed Central, LOCKSS.