Discovering Potential Pathogens among Fungi Identified as Nonsporulating Molds

Abstract

Fungal infections are increasing, particularly among immunocompromised hosts, and a rapid diagnosis is essential to initiate antifungal therapy. Often fungi cannot be identified by conventional methods and are classified as nonsporulating molds (NSM).We sequenced internal transcribed spacer regions from 50 cultures of NSM and found 16 potential pathogens that can be associated with clinical disease. In selected clinical settings, identification of NSM could prove valuable and have an immediate impact on patient management.

Fungal infections are increasing, particularly among immunocompromised hosts, and a rapid, accurate diagnosis is essential for the initiation of targeted antifungal therapy. Diagnosis of fungal infections usually depends on recovery of fungi from culture of clinical specimens, and their identification requires the presence of reproductive structures. Often fungi cannot be characterized fully because the mold does not sporulate, making identification by microscopic morphology not possible and potentially increasing the time to report an inconclusive result to 21 days. While many laboratorians and clinicians assume that these fungal isolates are environmental organisms and not clinically significant, to our knowledge no study has systematically attempted to classify these previously unidentifiable fungi in a clinical microbiology laboratory.

Amplification and sequencing of target regions within the ribosomal DNA gene complex has emerged as a useful, adjunctive tool for the identification of fungi and does not depend on mold sporulation for identification (3, 5, 9, 11). The internal transcribed spacer (ITS) regions 1 and 2 located between the highly conserved small (18S) and large (28S) ribosomal subunit genes in the rRNA operon are known to have sufficient sequence variability to allow identification to the species level for many fungi (2, 3, 5, 9, 11, 15). The goals of this study were to determine if sequencing the ITS 1 and 2 regions of nonsporulating molds (NSM) could identify fungi that were not identifiable by conventional methods and could serve as an approach to detect clinically relevant pathogens.

Sample selection.

Identification of molds directly from a specimen or submitted as an isolated culture to Associated Regional and University Pathologists, Inc., Laboratories (ARUP Laboratories) was attempted by growth on inhibitory mold agar, modified Sabouraud agars, or potato dextrose agar. Microscopic structures were observed on tease or tape preparations and slide cultures for up to 21 days. NSM were defined as molds without reproductive structures and that could not be further characterized. From 1 January 2005 to 1 January 2006, clinical isolates classified as NSM were randomly selected for gene amplification and sequencing.

DNA preparation.

After growth for 1 to 7 days on potato dextrose agar slants, lysates were prepared from approximately 1 cm² of mycelia with IDI lysis kits (GeneOhm Sciences, San Diego, CA). Briefly, in a biological safety cabinet, mycelia were collected by scraping the slant with a sterile stick in 1 ml of sterile, molecular-grade H₂O. The material was transferred to a 2-ml screw-cap tube. The tubes were centrifuged for 1 min at 6,000 × g. If the mycelia did not pellet, the material was contained with a pediatric blood serum filter (Porex Corp., Fairburn, GA). Supernatant was removed. The material was resuspended in 200 μl of IDI sample buffer and transferred to the lysis tube, which contained glass beads. Lysis tubes were vortexed on the highest setting for 5 min. The tubes were placed in a boiling water bath for 15 min. Tubes were centrifuged for 5 min at 16,000 × g. The supernatant was stored at −20°C until amplification.

Amplification and sequencing.

Real-time PCR with SYBR green DNA binding dye and amplicon melting temperature analysis was performed on the RotorGene 3000 (Corbett Robotics, Inc., Sydney, Australia). PCR mixtures contained the following: 1× Lightcycler FastStart DNA Master Hybridization Probes mixture (Roche Applied Science, Penzberg, Germany), which contained deoxynucleoside triphosphates, FastStart Taq DNA polymerase, and 1 mM MgCl₂ (additional MgCl₂ was added to a final concentration of 4.6 mM); 0.4 μM each of ITS1 forward (5′TCCGTAGGTGAACCTGCGG3′) and ITS4 reverse (5′TCCTCCGCTTATTGATATGC3′) primers (15); 1× SYBR green (Molecular Probes, Inc. Eugene, OR); and 3 μl template DNA. Thermal cycling parameters with RotorGene 3000 were 95°C for 10 min; 50 cycles of 95°C for 5 s, 60°C for 20 s, and 76°C for 30 s; and a final extension at 72°C for 2 min. The quality of the amplicon was determined using the derivative of the melt analysis curve (55°C to 99°C, 45-s hold at 55°C, 5 s/°C). The amplified product was processed for bidirectional sequencing using ExoSAP-IT (USB Corp., Cleveland, OH). Five microliters of Big Dye Terminator Ready Reaction Mix v. 1.1 (Applied Biosystems, Inc., Foster City, CA) was added to 4 μl of each primer (0.8 pmol/μl) and 3 μl of purified PCR product. Cycle sequencing was performed with a 9700 thermal cycler (ABI), using 25 cycles of 96°C for 10 s, 50°C for 5 s, and 60°C for 4 min. Sequencing reaction products were passed through a Sephadex G-50 fine column to remove unincorporated dye terminators. Purified sequencing reaction products were read on an ABI Prism 3100 Genetic Analyzer with a 50-cm capillary array.

Sequences were analyzed with the SmartGene Integrated Database Network (SmartGene Inc., Raleigh, NC) software version 3.2.3 vr. SmartGene is a web-based software and database system with reference sequences derived from the National Center for Biological Information (NCBI) GenBank repository. Phylogenetic trees and alignments of reference sequences with ≥98% identity to NSM isolate sequences were created using Clustalx v1.83 (University of British Columbia Bioinformatics Centre).

Nucleotide sequences with match lengths of ≥400 bp were analyzed. Sequence-based identifications were defined by percent identity: species, ≥99%; genus, 93 to 99%; and inconclusive, ≤93%. Molds were classified as fungi often associated with clinical disease (potential pathogens), emerging pathogens with more than three cases reported (emerging pathogens), or plant/soil-associated fungi with no or few published cases of human disease. Sequence information was independently analyzed by the Mycotic Diseases Branch of the Centers for Disease Control and Prevention (CDC) specifically to confirm microorganism identification.

The number of NSM reported was 215, which comprised 3.27% of the total number of fungal isolates identified at ARUP in 2005. A representative subset of 50 isolates reported as NSM (23%) was randomly selected for gene sequencing. The sources of these isolates were mostly respiratory (38 isolates) or skin, hair, and nails (11 isolates), with the source for 1 isolate not provided (Table 1).

TABLE 1.

Identification of 48 nonsporulating molds by ITS sequencing

Source	ARUP identification (GenBank accession no.)	Identity (%)	Sequence length (bp)	Match length (bp)	No. of species with >98% homology	CDC identification	Microorganism classification
BAL	Inonotus cuticularis (AF237730)	99.73	753	747	1	Filamentous basidiomycete	Plant/soil
BAL	Ceriporiopsis aneirina (AY219362)	91.41	660	489	0	Filamentous basidiomycete	Plant/soil
BAL	Polyporus sp. (AF516545)	94.43	659	646	0	Filamentous basidiomycete	Plant/soil
Scalp	Trichophyton verrucosum (AF168126)	99.85	655	655	1	Trichophyton verrucosum	Pathogen
Lung	Irpex lacteus (AB079265)	99.69	651	651	1	Filamentous basidiomycete	Plant/soil
Sputum	Irpex lacteus (AB079265)	99.85	650	648	1	Filamentous basidiomycete	Plant/soil
Bronchi	Pleurotus ostreatus (AY636055)	99.53	644	640	5	Filamentous basidiomycete	Plant/soil
Bronchi	Oxyporus latemarginatus (AF232721)	99.83	633	576	2	Filamentous basidiomycete	Plant/soil
BAL	Coprinus sp. (AF345808)	98.08	625	624	1	Filamentous basidiomycete	Potential pathogen
BAL	Pycnoporus sp. (AF363759)	94.36	623	618	0	Filamentous basidiomycete	Plant/soil
BAL	Pycnoporus sp. (AF363759)	94.35	619	618	0	Filamentous basidiomycete	Plant/soil
Sputum	Amauroderma subresinosum (AJ627583)	91.16	612	611	0	Filamentous basidiomycete	Plant/soil
Ankle	Exophiala jeanselmei (AJ866273)	99.67	608	607	4	Exophiala sp.	Pathogen
BAL	Phellinus robinae (AY558646)	99.17	607	603	1	Filamentous basidiomycete	Plant/soil
BAL	Trametes versicolor (AY309018)	99.84	607	607	6	Filamentous basidiomycete	Plant/soil
BAL	Pycnoporus sp. (AF363759)	93.73	606	606	0	Filamentous basidiomycete	Plant/soil
BAL	Hyphodontia flavipora (AF455399)	99.48	603	572	1	Filamentous basidiomycete	Plant/soil
BAL	Schizophyllum commune (AF348142)	99.33	603	597	1	Schizophyllum commune	Potential pathogen
Sinus	Schizophyllum commune (AF348142)	99.33	603	600	1	Schizophyllum commune	Potential pathogen
BAL	Thanatephorus cucumeris (AF455463)	99.00	600	599	2	Filamentous basidiomycete	Plant/soil
Sinus	Phlebia subserialis (AB084620)	99.83	598	572	1	Filamentous basidiomycete	Plant/soil
BAL	Oxyporus latemarginatus (AF232721)	99.83	592	573	2	Filamentous basidiomycete	Plant/soil
Bronchi	Schizophyllum commune (AY573544)	99.66	588	588	1	Schizophyllum commune	Potential pathogen
Sinus	Schizophyllum commune (AF348142)	99.32	585	585	1	Schizophyllum commune	Potential pathogen
Sinus	Schizophyllum commune (AF348142)	99.32	584	584	1	Schizophyllum commune	Potential pathogen
Scalp	Aphanoascus fulvescens (AF038357)	99.13	575	575	2	Aphanoascus sp.	Plant/soil
Toe	Lewia infectoriaa (AF397239)	99.65	574	574	2	Alternaria infectoriaa	Pathogen
Sputum	Eutypella scoparia (AF373064)	99.48	573	573	1	Eutypella scoparia	Plant/soil
BAL	Phlebia sp. (AY787680)	81.75	572	572	0	Filamentous basidiomycete	Plant/soil
Lung	Phanerochaete chrysosporium (AF475147)	88.93	572	572	0	Filamentous basidiomycete	Plant/soil
Bronchi	Oxyporus latemarginatus (AF232721)	99.46	570	557	2	Filamentous basidiomycete	Plant/soil
Sputum	Arthrinium sp. (AY425967)	100.00	570	570	2	Arthrinium sp.	Plant/soil
Nail	Botryosphaeria obtusa (AY259094)	99.82	564	560	2	Coelomycete fungus	Plant/soil
Sinus	Myrothecium sp. (AJ619957)	99.28	563	556	1	Possible coelomycete fungus	Plant/soil
Sputum	Arthrographis cuboidea (AB213444)	99.29	563	563	1	Arthrographis sp.	Plant/soil
Skin	Aureobasidium pullulans (AY225167)	99.82	559	556	1	Aureobasidium pullulans	Pathogen
Bronchi	Phlebia subserialis (AB084620)	99.66	558	588	1	Filamentous basidiomycete	Plant/soil
Lung	Ceriporiopsis sp. (AY781250)	99.46	557	557	1	Filamentous basidiomycete	Plant/soil
Nail	Coniothyrium sp. (AY157492)	100.00	542	516	2	Coelomycete fungus	Potential pathogen
Other	Pestalotiopsis sp. (AF409979)	99.08	541	541	10	Pestalotiopsis sp.	Plant/soil
Pleural fluid	Botryosphaeria rhodinab (AY568635)	100.00	537	504	1	Filamentous basidiomycete	Potential pathogen
BAL	Oxyporus latemarginatus (AF232721)	99.80	536	512	2	Filamentous basidiomycete	Plant/soil
BAL	Amanita tenuifolia (AF085492)	99.39	517	494	1	Filamentous basidiomycete	Plant/soil
Lung	Fusarium sp. (U61695)	99.22	516	515	1	Fusarium sp.	Pathogen
Nail	Pestalotiopsis sp. (AF409967)	99.60	498	498	20	Pestalotiopsis sp.	Plant/soil
BAL	Ajellomyces capsulatusc (U18363)	100.00	485	467	1	Histoplasma capsulatumc	Pathogen
Nail	Phialophora sp. (AF083206)	96.98	430	464	1	Phialophora lignicola	Pathogen
Scalp	Exophiala dermatitidis (AY213650)	100.00	404	404	3	Exophiala sp.	Pathogen

^aLewia infectoria teleomorph of Alternaria infectoria.

^bBotryosphaeria rhodina teleomorph of Lasiodiplodia theobromae.

^cAjellomyces capsulatus teleomorph of Histoplasma capsulatum.

Forty-eight of the isolates had ITS sequence and match lengths of ≥400 bp, with an average sequence length of 572.3 ± 74.9 bp and an average match length of 560.3 ± 86.5 bp. The two isolates with sequence lengths of <400 bp were not included in the analysis. For the 48 sequences, gene sequencing identified 44 (92%) to genus level and 38 (79%) to species level, with 4 (8%) being inconclusive.

Sixteen isolates had reference sequences that shared ≥98% identity for more than one species, involving 12 genera. As a representative example, two isolates identified as being in the Pestalotiopsis genus had 10 and 20 reference sequences sharing ≥98% identity with multiple, different species, respectively. Phylogenetic tree construction for Pestaloptiopsis spp. also demonstrated a high degree of interspecies similarity and an absence of well-characterized strains within this genus. Similar results were found for 11 other genera.

Eight NSM were classified as fungi with pathogenic potential; these were Trichophyton verrucosum, Ajellomyces capsulatus (anamorph Histoplasma capsulatum), Aureobasidium pullulans, Exophiala dermatitidis, Exophiala jeanselmei, Fusarium sp., Lewia infectoria (anamorph Alternaria infectoria), and Phialophora sp. Eight NSM were classified as potential emerging pathogens: Botryosphaeria rhodina (anamorph Lasiodiplodia theobromae) (one isolate), Coniothyrium sp. (one isolate), Coprinus sp. (one isolate), and Schizophyllum commune (five isolates). Thirty-two isolates (66.7%) were identified as plant- and soil-associated fungi, mostly as filamentous basidomycetes.

Independent analysis of the NSM sequences by the CDC using the GenBank database correlated with our identifications. All eight potential pathogens were identified by the CDC, as were the five isolates identified as Schizophyllum commune, an emerging pathogen. Six additional fungi were identified to the genus or species level by the CDC method. For the remaining 29 isolates, the CDC reported them as filamentous basidomycetes (26 isolates) or coelomycetes (3 isolates) without further identification, which corroborated our classifications.

Molds are ubiquitous organisms in the environment and often transiently colonize the respiratory, integument, and gastrointestinal systems of healthy hosts. Traditionally, NSM recovered in the laboratory have been dismissed as insignificant environmental organisms without further testing. Additionally, for those laboratories that attempt to augment sporulation specifically for identification, the process can require up to 3 weeks of incubation and often is without success. To date, sequencing of NSM has been limited to case reports. To our knowledge, this study is the first to systematically evaluate a large number of NSM clinical isolates with ITS sequencing for further species identification. With sequencing the ITS 1 and ITS 2 regions, we identified the majority of NSM that could not be identified by conventional mycological methods and classified over one-third of these molds as well-recognized or emerging pathogens with potential for invasive disease in the appropriate clinical setting. As a national reference laboratory, we were unable to obtain clinical histories for these patients, but we presume that for the majority of these NSM, clinicians felt that the isolates had clinical relevance and referred them to our laboratory for fungal identification. We acknowledge that the classification of fungal pathogenicity with clinical relevance was difficult in the absence of patient data, but we based our classifications on standard practices and the current literature (1-3, 6-11, 13, 14). For example, ITS sequencing identified Histoplasma capsulatum (from bronchoalveolar lavage [BAL]), and although clinical correlation was not possible, the significance of this pathogen is well recognized. Similarly, we identified pathogens such as Schizophyllum commune (3, 10), Coniothyrium (6, 7), Coprinus (8, 13), and Botrysphaeria rhodina (1, 14), which can be considered emerging pathogens capable of causing disease when there is histopathological or other corroborating evidence. For many isolates identified by ITS sequencing, we classified them as environmental fungi because these microorganisms are commonly encountered in the clinical laboratory and usually have no clinical importance in healthy subjects (12). Yet, for immunocompromised hosts, the pathogenicity of such environmental fungi has not been well defined, and the capacity to characterize these organisms by ITS sequencing may prove useful when there is a high index of suspicion that the environmental NSM is a putative pathogen.

The taxonomy of many fungi is evolving, and names of organisms may change rapidly, while obsolete nomenclature may exist in databases; this may contribute to the results that we and others have observed in fungal identification using ITS sequences (4, 5). In attempts to minimize the risk for misidentification (2), we specifically selected a conservative match length of ≥400 bp and a 93% identity cutoff for identification to the genus level. The 93% cutoff was empirically determined to allow identification to the genus level of additional plant/soil-associated fungi. Despite these criteria, we found some clinical isolates sharing a high percent identity for multiple species, which suggests inadequate variation in the ITS regions and thereby prevention of species discrimination. When definitive fungal identification is clinically indicated, additional targets may be necessary to identify certain genera and species (5, 9). Four NSM were not conclusively identified by ITS sequencing in our study; these may represent novel species that do not have representative ITS reference sequences in GenBank.

In summary, ITS sequencing provides a fast alternative to conventional identification for nonsporulating molds. For specific specimen sources (e.g., tissue or BAL) and in the appropriate clinical setting, sequencing the ITS region for nonsporulating molds that cannot be conclusively characterized by conventional means could serve as a valuable tool to identify clinically relevant pathogens and enable the timely initiation of appropriate antifungal treatment.