Introduction

Sporothrix stands for simply structured, single-celled conidia on clusters of denticles. This morphology is expressed in numerous species of the order Ophiostomatales. Main teleomorph genus is Ophiostoma, a large group of pathogens of woody plants characteristically associated with bark-beetles (Zhou et al. 2006; Zipfel et al. 2006; Roets et al. 2006). The slimy ascospores and (syn)anamorphic conidia classified in Sporothrix, Hyalorhinocladiella and Pesotum each have particular roles in this specific habitat (Zipfel et al. 2006). Among the few exceptions with an entirely different ecology within the Ophiostomatales is Sporothrix schenckii, a widespread pathogen of humans and other mammals (Guarro et al. 1999).

Sporothrix schenckii s.l. is responsible for sporotrichosis, a chronic, granulomatous, cutaneous or subcutaneous infection particularly occurring in humans and cats. The most common route of infection is via traumatic implantation into the skin of otherwise healthy individuals. In immunocompromised patients systemic and disseminated infections are observed, affecting multiple organs (Callens et al. 2006; Silva-Vergara et al. 2012). Sporothrix infections may take epidemic proportions (Dixon et al. 1991; Marimon et al. 2006; Verma et al. 2012). The source of human infection is environmental, mostly from soil and plant debris (Dixon et al. 1991; Mesa-Arango et al. 2002) and may also be transmitted from scratches by asymptomatic or infected cats (Schubach et al. 2004; Rosa et al. 2005; Barros et al. 2010).

The generic type species, S. schenckii is characterized by tear-shaped conidia on small, clustered denticles (de Hoog et al. 2000). The fungus is thermally dimorphic, i.e., at 37 °C a yeast-like phase is produced under appropriate conditions (Howard 1961; de Hoog 1974; Travassos and Lloyd 1980). Molecular phylogenetic analyses have shown that several species exist within the S. schenckii species complex (de Beer et al. 2003; Marimon et al. 2006, 2007; Madrid et al. 2010; Criseo and Romeo 2010). Multilocus sequence data proved to be supported by small phenotypic characters, which led to the description of the following novel clinically relevant species: S. brasiliensis, S. globosa, S. mexicana in addition to S. schenckii s. str. and S. luriei (Marimon et al. 2007, 2008). Recently described environmental species include S. stylites, S. lignivora, S. humicola (de Meyer et al. 2008), S. variecibatus (Roets et al. 2008), S. brunneoviolacea and S. dimorphospora (Madrid et al. 2010). The species differ significantly in virulence and predilection (Marimon et al. 2007; Arrillaga-Moncrieff et al. 2009; Fernández-Silva et al. 2012; Fernandes et al. 2013), and geographic distributions of some of the species are limited (Rodrigues et al. 2012).

The position of the S. schenckii complex amidst species with a rather consistent, entirely different ecology is puzzling. Some scattered case reports have been published in Ophiostoma, such as cutaneous and nail infections by O. stenoceras (Mariat 1971; Summerbell 1993) and a systemic infection in a leukemic patient by O. piceae (Bommer et al. 2009), but a human-pathogenic potential as in Sporothrix is absent from Ophiostoma.

Clinical Sporothrix species presently are classified with the aid of partial calmodulin sequences (Marimon et al. 2007). Sporothrix schenckii in a restricted sense still contains significant diversity (Marimon et al. 2006, 2007) compared to e.g. S. brasiliensis (Marimon et al. 2007; Rodrigues et al. 2012). This may reduce the barcoding gap, which is the ratio of intra- and interspecific variabilities, and hamper the development of specific barcode-identifiers (Heinrichs et al. 2012). The aim of the present study was to determine the validity of the internal transcribed spacer (ITS) region as a marker for species recognition in the S. schenckii complex. We evaluated an epidemiologically diverse strain panel of clinical and environmental isolates classified in 30 taxa with a global distribution and comprising both anamorph and teleomorph taxa.

Materials and methods

Fungal isolates

A total of 124 isolates identified as Sporothrix and Ophiostoma species (68 clinical, 56 environmental) by morphology and/or partial β-tubulin (BT2) sequence were included in the study (Table 1). Strains were provided by reference collections of ATCC, CBS, CMW, CNM-CM, IPEC and KMU. Ex-type strains were included for all species.

Table 1 Ophiostoma/Sporothrix isolates included in the study
Full size table

DNA extraction

Isolates were grown for 10 day on potato dextrose agar (PDA). DNA was extracted following the Quick CTAB extraction: Two mL screw-capped tubes were filled with 490 μL CTAB-buffer 2× and 6–10 acid-washed glass beads, 1–10 mm3 fungal material was added, and 10 μL Proteinase K; the material was mixed thoroughly on a MoBio vortex for 10 min. After incubation for 60 min at 60 °C. 500 μL chloroform: isoamylalcohol (24:1) was added and shaken for 2 min. Tubes were centrifuged at 14,000 r.p.m. for 10 min. The upper layer was collected in a fresh tube. To ~400 μL DNA sample 2/3 vol (~270 μL) of ice cold iso-propanol was added and spun again at 14,000 r.p.m. for 10 min and samples were dissolved in 1 mL ice-cold 70 % ethanol. Tubes were spun again at 14,000 r.p.m. for 2 min, samples were air-dried and resuspended in 50 μL TE-buffer. Samples were stored −20 °C until use. Quality of DNA was verified by running 2–3 μL on a 0.8 % agarose gel.

PCR amplification

Sporothrix/Ophiostoma ITS regions were directly amplified from genomic DNA with primers ITS5 5′-GGA AGT AAA AGT CGT AAC AAG G-3′ and ITS4 5′-TCC TCC GCT TAT TGA TAT GC-3′ as described by Madrid et al. (2010). BT2 region was amplified using the primers Bt2-F 5′-GG (CT) AACCA (AG) AT (ATC) GGTGC (CT) GC (CT)-3′ and Bt2-R 5′-ACC CTC (AG) GTG TAG TGA CCC TTG GC-3′ according to Marimon et al. (2006). The reaction mixture (25 μL final vol) contained PCR buffer (10×) 2.5 μL, water 15 μL, dNTP mix (1 mM) 2.5 μL, 1 μL of each primer (10 pmol), Taq polymerase (0.5 U) 1 μL, BSA (Bovine serum albumin) 1 μL, and DNA 1 μL [100 ng/μL]. PCR reactions were performed in a Hybaid Touchdown PCR machine (Hybaid, Middlesex, U.K.). PCR conditions were: one cycle of 5 min at 95 °C, followed by 35 cycles of 35 s at 95 °C, 30 s at 52 °C (ITS) or 60 °C (BT2) and 1 min at 72 °C, followed by one cycle of 6 min at 72 °C. PCR products were visualized by electrophoresis on a 1 % (w/v) agarose gel. PCR products were purified with the High Pure PCR kit (Roche, Mannheim, Germany). Both strands of the PCR fragments were sequenced with the above-mentioned primers. The ABI PrismH Big DyeTM Terminator v. 3.0 Ready Reaction Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, U.S.A.) was used for sequencing PCR. Sequences were determined with an ABI PRISM™ 3,100 Genetic Analyzer (Applied Biosystems). DNA sequences of opposite strands were edited with Sequence Navigator version 1.0.1 (Applied Biosystems). All sequences were aligned with MAFFT v. 5.667 (Katoh et al. 2002).

DNA sequence analyses

In order to evaluate the global ITS diversity in the S. schenckii complex we collected nucleotide sequences from Sporothrix/Ophiostoma isolates from different regions of the world from GenBank. Methods used as well as the number of sequences retrieved in the search are exemplified in the Supplementary Fig. 1. Phylogenetic analysis included sequences previously published in the literature and in GenBank (Table 1) as well as newly generated sequences originated from clinical isolates of Sporothrix and other closely related environmental Ophiostoma and Sporothrix species. Phylogenetic analyses were conducted with MEGA 5 (Tamura et al. 2011) with Maximum Likelihood and Neighbor-joining methods. Evolutionary distances were computed using the Kimura 2-parameter method (Kimura 1980) with a discrete Gamma distribution for the ITS dataset and Tamura 3-parameter method (Tamura 1992) for BT2 dataset, both using 1,000 bootstrap replicates (Felsenstein 1985).

Results

In the present study we used a Boolean search method to retrieve Sporothrix sequences deposited in GenBank. Using ‘Sporothrix’ as a query we were able to retrieve approximately 907 entries. Using an exclusive search strategy, we recovered approximately 130 nucleotide sequences that matched the ITS region (Supplementary Fig. 1). From these entries, only a few sequences were long enough (> 500 bp including ITS1 + 5.8S + ITS2) to be used for confident alignment.

A final alignment was created with 124 sequences, including 25 and 99 sequences of Ophiostoma and Sporothrix, respectively. Aligned ITS sequences were 637 bp long, including 374 invariable characters, 213 variable parsimony-informative (33.4 %), and 42 singletons. Positions containing gaps and missing data were eliminated. Indel regions were evaluated considering the sequence FJ545232 from the type strain of S. schenckii (CBS 359.36) from Maryland, USA (Fig. 1a). The variation within the phylogenetically related species of clinical interest (68 sequences), including S. schenckii, S. brasiliensis and S. globosa is shown in Fig. 1b. A substantial number of 17 polymorphic sites was noted differentiating the species S. brasiliensis from its sister taxon S. schenckii (14 of them were parsimony-informative). The ITS1 region had a higher degree of mutations than ITS2 (ITS1/ITS2 ratio = 2,4).

Fig. 1
figure 1

Polymorphisms in ITS1/2 + 5.8S nucleotide sequences of Sporothrix and Ophiostoma. a Mutations at each position in the aligned sequences (n = 124) including environmental and clinical isolates in Sporothrix and Ophiostoma complex. b Sequence comparison among 68 strains in the clinical clade including S. brasiliensis, S. schenckii and S. globosa. All comparisons were done relative to S. schenckii type strain CBS 359.36 (FJ545232)

Full size image

An ITS phylogenetic tree was constructed using Maximum likelihood (model K2-P + G) with 1,000 bootstrap replications (Fig. 2). The 124 sequences were distributed into 30 taxa described in previous studies (de Beer et al. 2003; Zhou et al. 2006; Zipfel et al. 2006; Roets et al. 2006, 2008; de Meyer et al. 2008; Madrid et al. 2010). An unambiguous separation between species of clinical importance and strictly environmental species was observed, with strains from human and animal sources being concentrated in the clades S. brasiliensis, S. schenckii, S. globosa and S. luriei (A; Fig. 2).

Fig. 2
figure 2

Phylogenetic relationships inferred from maximum likelihood analysis of ITS sequences of 124 strains belonging to Sporothrix and Ophiostoma. Numbers close to the branches represent indices of support based on 1,000 bootstrap replications. Branches with bootstrap support higher than 70 % are indicated in bold. Sporothrix brasiliensis (b), S. schenckii (c and d) and S. globosa (e) frequencies are calculated from clinical isolates collected worldwide and available at GenBank. Isolates were listed according the geographical origin from the Americas (AM), Europe (EU), Africa (AF), Asia (AS), or Australia (AU)

Full size image

The geographical distribution of isolates in the clinical clade varied with the species. The clade identified as S. brasiliensis (B) had a high incidence in South America, being restricted to Brazil (Fig. 2). The phylogenetic group identified here as S. schenckii s.str. was subdivided into two clusters. A first group (C) adjacent to S. brasiliensis prevailed in the Americas (61 %), followed by Asia (17 %), Africa (11 %) and Europe (11 %). A second set of S. schenckii strains (D) harboring the type strain CBS 359.36 prevailed in the Americas (50 %), followed by Africa (38 %), Asia (6 %) and Europe (6 %). Sporothrix globosa (E) is present with high frequency in Asia (56 %) and Europe (28 %), followed by the Americas (11 %) and Africa (5 %).

In the remaining tree, Sporothrix species were flanked by Ophiostoma species which were mainly derived from soil, plants or found in association with bark beetles. Despite the good taxonomic resolution achieved for clinical species, in the environmental clade some taxa were not easily differentiated using the ITS region. Sporothrix mexicana was located amidst the environmental Sporothrix species (S. humicola, S. stylites, S. pallida, and S. nivea) which had identical ITS sequences. The same was found with several clusters of Ophiostoma species. Sporothrix brunneoviolacea and S. lignivora took remote positions; S. lignivora was used to root the tree.

In order to calibrate the ITS-based phylogeny, some isolates from each taxa evaluated previously were chosen for a second analysis using the BT2 region (including the type strains for each species). This region was selected because it has been widely used for taxonomy of environmental Ophiostoma/Sporothrix species (Roets et al. 2006, 2008; Zipfel et al. 2006). The BT2 complete alignment included 71 sequences (Table 1). Aligned sequences of BT2 matched 607 characters, including 189 invariable characters, 307 variable parsimony-informative (50,5 %), and 37 singletons. Positions containing gaps and missing data were eliminated. All taxa were clearly separated using this locus. A strong separation between clinical and environmental clades was observed for BT2, coinciding with the bipartition found in ITS. In agreement with ITS phylogeny, the environmental species O. phasma was the nearest taxon to the S. schenckii complex. This topology is in agreement with previous studies (Roets et al. 2006, 2010; Madrid et al. 2010). The topologies of trees of BT2 and ITS showed strict correspondence, with all clinical clades being recognized using both genes.

Discussion

Our data provide a representation of the S. schenckii complex as it is embedded in the phylogeny of Ophiostomatales (de Beer et al. 2003; de Meyer et al. 2008). An Ophiostoma teleomorph has been predicted for S. schenckii on the basis of morphological similarity with anamorphs of e.g. O. stenoceras and O. nigrocarpum as supported by sequence data (Berbee and Taylor 1992; de Beer et al. 2003). In our analysis, Ophiostoma phasma (CMW 20676) from Protea in South Africa appears most closely related (Fig. 2a). However, S. schenckii is still located at relatively large distance. By combined phenotypic and molecular characters Marimon et al. (2007) introduced S. brasiliensis, S. globosa and S. mexicana as new species, next to S. luriei, a variety of S. schenckii which was attributed species status. Earlier described species such as S. pallida and S. inflata were confirmed to be distinct taxa. Although Marimon et al. (2007) made a substantial contribution towards understanding the relationships within the species complex, their study was limited mainly to clinical isolates from few geographical origins. The present study expands the investigation of Sporothrix species in ecological origins and global representation of strains analyzed. In our expanded comparison it is apparent that S. mexicana is remote, being a member of the S. pallida complex (Fig. 2). This species presents a mild potential pathogenicity to humans (Rodrigues et al. 2012), which is exceptional outside the S. schenckii complex.

The S. schenckii complex is presently restricted to four species: S. schenckii s.str., S brasiliensis, S. globosa and S. luriei, which deviate not only phylogenetically from the remainder of Ophiostomatales, but also by their virulence to mammals. Species distinction within the complex is presently based on partial calmodulin sequences (Marimon et al. 2007; Oliveira et al. 2011; Romeo et al. 2011). The topology of the BT2 tree (Fig. 3) is broadly concordant with that of the ITS tree (Fig. 2), and both are essentially similar to that derived from partial calmodulin sequences in Oliveira et al. (2011) and Romeo et al. (2011). Several of the ITS clades, including the S. schenckii clade are statistically supported with high bootstrap values (Fig. 2). Other groups of species such as those around S. pallida have identical ITS sequences, and similar clusters of closely related species are noted e.g. with Ophiostoma proteae and O. africanum.

Fig. 3
figure 3

Phylogenetic relationships of Sporothrix and Ophiostoma inferred from β-tubulin sequences by Neighbor-joining algorithm based on the Tamura 3-parameter model. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) is shown next to the branches (Bootstrap support values > 80 are indicated in bold). GenBank accessions numbers are indicated next to strain code

Full size image

In the highly supported, derived clade of S. schenckii and relatives the ITS differences are large enough to distinguish all four presently recognized taxa using this widely applied barcoding gene. Three of the four species of this group are known from pseudoepidemics: S. brasiliensis in Brazil (Rodrigues et al. 2012), S. schenckii s.l. in South Africa (Vismer and Hull 1997), and S. globosa in China (Han et al. 2006; Li et al. 2007; Zhang and Lin 2008; Mei et al. 2011; Wang et al. 2012). Although Sporothrix species are primarily environmental, their traumatic inoculation into human hosts e.g. by wood splinters from pine wood (South Africa) or from scratches of cat paws in Brazil (Schubach et al. 2005, 2008) is highly efficient. This is quite a remarkable feature in the fungal Kingdom, where (pseudo)epidemics are exceptional. Outside the S. schenckii complex, only accidental, unlinked cases are known, e.g. O. stenoceras (Summerbell 1993), O. piceae (Bommer et al. 2009), and S. mexicana (Rodrigues et al. 2012). Isolate CMW 7613, previously identified as S. schenckii by de Beer et al. (2003) grouped within the S. pallida complex, a typically environmental clade, although this isolate was derived from a human case of sporotrichosis in South Africa. These examples show that mammal-pathogenicity also occurs outside the S. schenckii complex, but it remains highly exceptional with scattered cases. In the S. schenckii complex a high degree of virulence is constitutive, and in this sense the group deviates considerably from the remaining Ophiostomatales. It is recommended that the group remains separate, and that it is not merged with Ophiostoma.

Utilizing length polymorphisms within the ITS region we were able to detect epidemiological differences among clinical Sporothrix strains. In agreement with previous studies (Marimon et al. 2006, 2007), S. brasiliensis remains restricted to the Brazilian territory (Fig. 2). Most of the sequences representing the first group (C) of S. schenckii are predominant in the Americas and Asia, whereas those of the second group (D) are present in the Americas followed by Africa. Europe has a low incidence of sporotrichosis. Surprisingly the pathogenic species S. globosa (E) is predominant in Asia and Europe, but rare in the Americas and Africa. This is in agreement with Madrid et al. (2009) that show S. schenckii is more common in Americas than S. globosa. Figure 2 also shows that strains originating from a restricted geographic area mostly do not constitute monophyletic lineages. For example, strains from South Africa scattered in at least 3 major clades, and the same was found for the strains from Japan. Our increased geographic sampling for the ITS dataset did not affect delimitation success in the clinical clade. Unfortunately we were unable to compare strains and sequences from Australian epidemics of sporotrichosis; S. schenckii was identified as the causal agent of the Australian epidemics using morphological characters and the pulsed-field gel electrophoresis (PFGE) technique (O'Reilly and Altman 2006; Feeney et al. 2007).

ITS is the most widely applied gene for routine identification, and has been recommended as a fungal barcoding gene (Seifert 2009; Schoch et al. 2012; Toju et al. 2012). The four current species of the complex are all distinguishable by fixed mutations. The smallest barcoding gap is between S. schenckii and S. brasiliensis, at a minimum distance of 4.44 % and mean interspecific divergence of 0.0079. We conclude that ITS is sufficient for routine species distinction of all clinically relevant Sporothrix-like species including the occasional agents in Ophiostomatales that are remote from the S. schenckii complex.