ABSTRACT
The use of molecular identification techniques has revealed an increasing number of new species within Aspergillus section Terrei. We phenotyped a set of 26 clinical isolates that showed genetic differences from Aspergillus terreus sensu stricto by analyzing sequences from PCR-amplified β-tubulin and calmodulin genes and the internal transcribed spacer region. Since the isolates were phylogenetically and morphologically different from all of the members of Aspergillus section Terrei, they are described here as a new species, Aspergillus citrinoterreus, so named because it produces a diffusible yellowish pigment in agar. A. citrinoterreus isolates were significantly more susceptible to itraconazole, voriconazole, and posaconazole than A. terreus sensu stricto isolates were; in contrast, the amphotericin B MICs for both species were high. A. citrinoterreus was found in clinical samples from patients with proven or probable invasive aspergillosis and colonized patients, none of whom had hematological malignancies as predisposing conditions. However, they did have other underlying conditions such as chronic obstructive pulmonary disease, cirrhosis, and cancer or had received a solid organ transplants and presented not only with invasive pulmonary aspergillosis but also with mediastinitis. A. citrinoterreus isolates were detected for the first time in 2002. In all cases of invasive aspergillosis, A. citrinoterreus was found to be a copathogen, mostly with A. fumigatus.
INTRODUCTION
Invasive aspergillosis affects patients with hematological and nonhematological conditions such as chronic obstructive pulmonary disease (COPD) (1–3). Most cases of invasive aspergillosis are caused by Aspergillus fumigatus and A. flavus. A. terreus is the third most common cause of invasive aspergillosis and a particularly prevalent microorganism in some geographic areas (4–8).
Molecular tools can provide an accurate picture of the epidemiology of invasive aspergillosis and have revealed the presence of cryptic Aspergillus species frequently missed by conventional techniques (9–11). The number of newly described species within the section Terrei has grown during the last few years and includes A. alabamensis, A. allahabadii, A. ambiguous, A. aureoterreus, A. carneus, A. floccosus, A. hortai, A. microcysticus, A. neoafricanus, A. neoindicus, A. niveus, A. pseudoterreus, and A. terreus sensu stricto (12). To date, most of these species, with the exception of A. terreus sensu stricto, have not been reported to cause invasive aspergillosis in humans.
We previously used molecular techniques to identify a set of Aspergillus section Terrei isolates collected from clinical samples from patients admitted to a general teaching hospital (13). Most of the isolates were identified as A. terreus sensu stricto, but a clade of isolates showed some remarkable genetic differences. The isolates comprising the clade have been phenotypically characterized and are reported here as representatives of a proposed new species within Aspergillus section Terrei, namely, Aspergillus citrinoterreus.
MATERIALS AND METHODS
Hospital description.This study was carried out at Hospital Gregorio Marañón, a large tertiary-care hospital serving a population of approximately 715,000 inhabitants of Madrid, Spain. The institution cares for all types of patients at risk of acquiring aspergillosis, including solid organ and bone marrow transplant recipients and patients with hematological malignancies, HIV infection, and COPD.
Study population and fungal isolates.We studied 26 A. terreus sensu lato isolates from the lower respiratory tract (n = 21), wounds (n = 4), and an abscess (n = 1) from 18 patients with proven (n = 1) or probable (n = 5) invasive aspergillosis or Aspergillus colonization (n = 12) admitted to Hospital Gregorio Marañón (Madrid, Spain) (Table 1). The isolates, which differed genetically from A. terreus sensu stricto, comprised 18 strains collected from 2002 to 2010 and studied in our previous investigation (13) and 8 additional strains collected from 2010 to 2012. Patients were classified according to the revised criteria of the EORTC (14); patients with COPD fulfilled Bulpa's criteria (15). We included 10 clinical isolates of A. terreus sensu stricto as controls.
Clinical sources and years of isolation of A. citrinoterreus isolates
DNA extraction, amplification, and sequencing.Genomic DNA was extracted from conidial suspensions of the isolates with the QIAamp DNA minikit (Qiagen, Heidelberg, Germany). The internal transcribed spacer (ITS) region and fragments of the β-tubulin (Tub) and calmodulin (Cal) genes were amplified and sequenced as previously described (13). Amplicons were purified with illustra GFX PCR DNA and Gel Band Purification kits (GE Healthcare UK Limited, Little Chalfont, Buckinghamshire, United Kingdom).
Phylogenetic analyses.Sequences from each locus were aligned by MEGA version 5.05 (16) and ClustalW (17) and manually improved when necessary. For patients with multiple samples from which the same fungus was isolated, only the isolate found in the first sample (n = 18) was included in the phylogenetic analyses. Neighbor-joining analyses using the Kimura two-parameter model were applied to each data partition individually in order to compare and check the stability of the individual phylogenies. Since no incongruence was found, the three loci were combined into a single data set. Phylogenetic reconstructions of the combined data set were made on the basis of maximum-likelihood (ML) analysis with Mega 5.05 and Bayesian inference (BI) analysis with MrBayes version 3.1.2 (18). For ML analysis, nearest-neighbor interchange was used as a heuristic method, gaps were treated as partial deletions with a 95% site coverage cutoff, and the robustness of the branches was estimated with a 1,000-generation ML bootstrapped data set (bootstrap values), for which a value ≥70% was considered significant. For BI analysis, two parallel runs of four incrementally heated Markov chains were performed for 800,000 generations with a sample frequency of 1,000 generations. The 50% majority rule consensus tree and Bayesian posterior predictive values were calculated after removing the first 25% of the samples; Bayesian posterior predictive values of ≥0.95 were considered significant. The best-fit model for each data partition (GTR+G for ITS and SYM+G for Tub and Cal) was estimated with MrModelTest version 2.3 (19). In addition, 65 sequences representing 19 type and reference strains were retrieved from GenBank and included in the phylogenetic analyses (Table 2).
GenBank accession numbers of clinical strains of A. citrinoterreus and A. terreus sensu stricto and reference sequences used in this study
Morphological and physiological study.Colony features and growth rates were determined for all of the isolates grown on creatine agar (CREA), Czapek yeast autolysate agar (CYA), malt extract agar (MEA), or oatmeal agar (OA) after 7 days of incubation at 25, 37, 40, or 45°C in darkness (12). The color notations used in the descriptions are those of Kornerup and Wanscher (20). Morphological observations and measurements were recorded for all of the isolates grown on MEA after 7 days at 25°C mounted in 85% lactic acid. Photographs of the microscopic structures were made with a Zeiss Axio Imager M1 light microscope with Nomarski differential interference contrast and phase-contrast optics (Zeiss, Oberkochen, Germany) and recorded with a DeltaPix Infinity X digital camera.
Antifungal susceptibility testing.Susceptibilities to the antifungals itraconazole (ITC), voriconazole (VRC), posaconazole (PSC), and amphotericin B (AMB) were determined by the CLSI M38-A2 procedure (21). The final concentrations of the antifungal agents tested ranged from 0.003 to 8 μg/ml. The antifungal susceptibilities of the 26 isolates were compared with those of 72 clinical isolates of A. terreus sensu stricto from our collection. Pairwise comparisons of both species were performed by using the Mann-Whitney test to calculate differences in antifungal susceptibility.
Accession numbers.Information about Aspergillus citrinoterreus has been submitted to MycoBank and assigned accession number MB810584. Newly determined sequences for clinical strains of A. citrinoterreus and A. terreus sensu stricto were submitted to GenBank under accession numbers KP175260 to KP175266, KP175270 to KP175278, KP175284 to KP175295, and LN680657 to LN680712 (Table 2).
RESULTS
Phylogenetic analysis of the combined data set (Fig. 1) included 1,423 bp (ITS, 485 bp; Tub, 395 bp; Cal, 543 bp). The ingroup consisted of 49 strains from 14 taxa, and Aspergillus neoniveus CBS 261.73 and Aspergillus flavipes CBS 260.73 were used as the outgroups. Aspergillus section Terrei was well delimited (bootstrap support value of 84, Bayesian posterior predictive value of 1.00) and consisted of 14 well-supported lineages, 13 of which corresponded to species currently accepted as members of this section. The 18 clinical strains and 4 additional reference strains previously assigned to A. terreus sensu stricto (NRRL 260, NRRL 1913, UOA/HCPF 9927, and UOA/HCPF 10158-2) were grouped in a fully supported clade (bootstrap support value of 100, Bayesian posterior predictive value of 1.00) that was a sister to A. terreus sensu stricto (97.7% sequence similarity). Since the former set of isolates was phylogenetically and morphologically different from all of the members of this section, they are described below as a new species, Aspergillus citrinoterreus. This clade was also consistently formed with a high level of statistical support in the individual Tub and Cal phylogenies (data not shown). In contrast, the analysis of the single ITS region showed insufficient resolution for many of the species of Aspergillus section Terrei, including the new species A. citrinoterreus.
ML tree inferred from combined ITS, Tub, and Cal sequences of Aspergillus isolates from section Terrei. Branch lengths are proportional to phylogenetic distances. ML bootstrap support values of ≥70% and Bayesian posterior predictive values of ≥0.95 are shown above the branches. Strongly supported branches (bootstrap support value of 100, Bayesian posterior predictive value of 1.00) are in bold. Sequences of Aspergillus neoniveus and Aspergillus flavipes were used to root the tree. A superscript T indicates a type strain. CBS, culture collection of the CBS-KNAW Fungal Biodiversity Center, Utrecht, The Netherlands; GM, clinical strains stored at Hospital Gregorio Marañón; UOA/HCPF, University of Athens/Hellenic Collection of Pathogenic Fungi.
Taxonomy.Aspergillus citrinoterreus Guinea, Sandoval-Denis, Escribano, Bouza & Guarro, sp. nov. (Fig. 2).
Images of Aspergillus citrinoterreus sp. nov. (CBS 138921). Panels: a to c, colonies on MEA, CYA, and CMD, respectively, after 7 days at 25°C; d to f, conidiophores; g, vesicle, metulae, and phialides; h and i, conidia; j and k, accessory conidia. Scale bars: d to f, 10 μm; g to k, 5 μm.
Etymology.So named because it produces a diffusible yellowish pigment.
Diagnosis.The new species is closely related to A. terreus sensu stricto, from which it differs by producing acid on CREA and forming light-colored colonies with abundant intense diffusible yellow pigment, septate stipes, globose to subglobose yellowish conidia, and smaller obovoid accessory conidia.
On CYA, colonies were velvety to dusty with a slightly floccose center, flat or slightly folded, at first white and then becoming brownish orange to grayish brown (5C3 to 5D3), reaching 33 to 35 mm in diameter at 7 days. Colonies in the reverse were grayish yellow to brownish orange (4B6 to 5C5) with abundant diffusible pastel yellow to light yellow pigment (3A4 to 4A4). On MEA, colonies were velvety to felty with a floccose center and flat with a white, dusty, and regular margin. They were greenish orange (5B3 to 5B6) in color, reaching 23 to 25 mm in diameter at 7 days. The reverse side of colonies was light yellow (4A4) with abundant yellow diffusible pigment. On OA, colonies were sandy to dusty, with abundant submerged mycelium, flat, and pastel yellow to light yellow (3A4 to 4A4). Colonies in the reverse were pastel yellow with abundant pastel yellow (2A4 to 3A4) diffusible pigment. On CREA, colonies were felty to fluffy, flat, greenish orange (5B3), reaching 35 to 40 mm in diameter at 7 days, with slight acid production at 25 and 37°C, except for two strains (GM 3967 and GM 4016) that produced large amounts of acids, leading to complete agar acidification and a color change after 3 days of incubation at 25 and 37°C. Conidiophores were hyaline and short (100 to 500 μm), widening from 2 to 5 μm at the base to 4 to 8 μm at the apex. They were also septate, smooth, and thick walled, gradually swelling to a globose to subglobose vesicle measuring 12 to 22 μm in diameter. Vesicles were typically biseriate at maturity, although some were monoseriate in young cultures. Metulae were cylindrical (5.5 to 9 by 1.5 to 3 μm), smooth, and thin walled, covering two-thirds of the vesicle surface. Conidiogenous cells were cylindrical with an apical constriction (5 to 9 by 1 to 2.5 μm), smooth, and thin walled. Conidia were globose to subglobose (2 to 3 by 1.5 to 3 μm), light yellow, smooth, thin walled, and arranged in compact columns. Accessory conidia were obovoid to ellipsoidal (3.5 to 4.5 by 3.5 to 4 μm), smooth, and thin walled and formed directly on hyphae or from short stalks (0.5 to 1 by 1 to 1.5 μm).
Holotype.Spain, from human sputum, 2006, T. Peláez (CBS H-22005; ex-type cultures CBS 138921 = GM 228).
A. citrinoterreus isolates were significantly more susceptible to ITC, VRC, and PSC than isolates of A. terreus sensu stricto were. In contrast, AMB MICs were high for both species (Table 3). A. citrinoterreus was found in clinical samples from patients with proven or probable invasive aspergillosis or from patients with colonization, none of whom had hematological malignancies as an underlying condition. Characteristics of the patients with invasive aspergillosis are shown in Table 4. Patients had predisposing conditions such as COPD, cirrhosis, and cancer or had received solid organ transplants and presented with invasive pulmonary aspergillosis and extrarespiratory involvement such as mediastinitis. Isolates of A. citrinoterreus were found for the first time in 2002. In all cases, A. citrinoterreus was found to be a copathogen, mostly with A. fumigatus (Table 4).
Antifungal susceptibilities of 26 A. citrinoterreus and 72 A. terreus sensu stricto isolates to ITC, VRC, PSC, and AMB
Clinical characteristics of the six patients with invasive aspergillosis from whom A. citrinoterreus was isolated
DISCUSSION
We describe A. citrinoterreus, a new species within Aspergillus section Terrei. A. citrinoterreus was found in samples from patients without hematological malignancies and in patients with and without invasive aspergillosis. In addition, it was slightly more susceptible to azoles than A. terreus sensu stricto is.
In recent years, with the aid of phylogenetic analyses, several new species belonging to Aspergillus section Terrei have been described (12, 22). Samson et al. (12) demonstrated that numerous taxa of section Terrei, which in the past were reduced to varietal status among A. terreus on the basis of phenotypic criteria, corresponded to distinct but closely related phylogenetic species, many of which were almost indistinguishable by morphological criteria. Similarly, although subtle, the differences in morphological, physiological, and antifungal-susceptibility features between A. citrinoterreus and its closest relative, A. terreus, are consistent with the results of the individual and combined three-gene phylogenetic analyses that allowed us to propose A. citrinoterreus as a new species. In 1934, Blochwitz described Aspergillus boedijni, a species that, like A. citrinoterreus, differed from A. terreus by forming yellowish conidia and a diffusible pigment (23). However, A. boedijni was considered a morphological variety of A. terreus (24) and was later considered to be synonymous with A. terreus (25). A. citrinoterreus, however, can be differentiated from the original description of A. boedijni by its larger and hyaline conidiogenous cells, larger conidia, and abundant production of a diffusible yellow pigment. In contrast, A. boedijni produces intense yellow conidiogenous cells and a diffusible pigment that is at first yellow but turns light brown to reddish with age. Although there is no type or authentic strain available of A. boedijni, a reference strain considered by Thom and Raper in 1945 to belong to the latter species (CBS 594.65) (24) clustered within A. terreus sensu stricto, which confirmed it to be distinct from the new species A. citrinoterreus.
Arabatzis and Velegraki (26) recently described the sexual morph of A. terreus on the basis of mating crosses and multilocus sequence analyses based on the same genetic loci applied in our study. Their phylogenetic analyses revealed a topology similar to that observed in our results, with several internal branches. However, the findings were not statistically significant. Interestingly, according to our phylogenetic analysis, most of the fertile strains in that study showed 99% sequence similarity to A. citrinoterreus, while only a single strain grouped within A. terreus sensu stricto. Accordingly, there is evidence that the teleomorph described corresponds to the sexual state of A. citrinoterreus. The use of mating crosses to infer species limits can be misleading because of the fertility of closely related phylogenetic species, indicating that phylogenetic divergence can precede reproductive isolation (27, 28).
A. citrinoterreus was found in samples from patients with and without invasive aspergillosis. In the six patients with invasive aspergillosis, A. citrinoterreus was found to be a copathogen, mainly with A. fumigatus. Therefore, it is difficult to assess the clinical significance of this new species. Further studies to clarify the association between invasive aspergillosis and A. citrinoterreus should be carried out. Of particular interest is the fact that we isolated A. citrinoterreus only from patients with nonhematological predisposing conditions, such as COPD; this group of patients has received greater attention during the last few years (3).
The AMB MICs for isolates of Aspergillus section Terrei are commonly high, and the outcome of patients treated with VRC is better than that of patients treated with a polyene (7, 29, 30). The Infectious Diseases Society of America guidelines recommend the use of VRC instead AMB for the treatment of infections caused by A. terreus (31). The low number of patients included in our series and the fact that A. citrinoterreus was a copathogen make it difficult to draw conclusions about the appropriate antifungal treatment for patients infected with this species. In addition, the observed lower azole MICs than for A. terreus sensu stricto should be studied in the future.
In conclusion, we describe A. citrinoterreus, a new species within Aspergillus section Terrei that was found in samples from patients with nonhematological predisposing conditions. Further studies are required to determine the potential pathogenic role of this new species, which is easily misidentified as the well-known pathogen A. terreus.
ACKNOWLEDGMENTS
We thank Thomas O'Boyle for editing the manuscript.
This work was supported by grant CP09/00055 from the Fondo de Investigación Sanitaria (FIS, Instituto de Salud Carlos III, Plan Nacional de I+D+I 2008-2011, FEDER support). J.G. (MS09/00055) and P.E. (CD09/00230) are supported by the FIS.
This study does not present any conflicts of interest for us.
FOOTNOTES
- Received 29 October 2014.
- Returned for modification 11 November 2014.
- Accepted 5 December 2014.
- Accepted manuscript posted online 10 December 2014.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
