Experience with the MicroSeq D2 Large-Subunit Ribosomal DNA Sequencing Kit for Identification of Filamentous Fungi Encountered in the Clinical Laboratory

Described herein is our experience with the MicroSeq D2 large-subunitrDNA sequencing kit for the identification of filamentous fungiencountered in the mycology laboratory at the Mayo Clinic. Atotal of 234 filamentous fungi recovered from clinical specimenswere used in the evaluation. All were identified by using phenotypiccharacteristics as observed macroscopically and microscopicallyon any medium or a combination of media, which included Sabouraud'sdextrose, inhibitory mold, cornmeal, Czapek-Dox, potato dextrose,and V8 juice agars; all isolates were sequenced using the MicroSeqD2 large-subunit rDNA sequencing kit. Of the of 234 isolates,158 were correctly identified to the appropriate genus or genusand species by using nucleic acid sequencing. Sequences for70 (29.9%) of the isolates (27 genera) were not included inthe MicroSeq library. Of the 80 dematiaceous and 154 hyalinefungi sequenced, 65 and 51.2%, respectively, gave results concordantwith those determined by phenotypic identification. Nucleicacid sequencing using the MicroSeq D2 large-subunit rDNA sequencingkit offers promise of being an accurate identification system;however, the associated library needs to include more of theclinically important genera and species.

INTRODUCTION

Top

Introduction

The identification of filamentous fungi in the clinical laboratorycan be challenging; success in this endeavor depends on theorganism and the experience of the microbiologist or technologist.For workers to become familiar with the fungi commonly seenin the local area is expected since the fungi are normal inhabitantsof the environment and may often be found in the respiratorytract secretions of healthy individuals. However, it is importantto remember that all are potentially pathogenic in the immunocompromisedpatient. Nucleic acid probes to identify the dimorphic pathogensBlastomyces dermatitidis, Coccidioides immitis, and Histoplasmacapsulatum are commercially available; these are more oftenused by larger laboratories with substantial experience. Manyof the hyaline and dematiaceous fungi do not sporulate, or elsethey are uncommon enough that phenotypic methods are not alwayshelpful for identification. The Biolog system (Biolog, Hayward,Calif.) offers a commercially available phenotypic array thatcan identify 618 different filamentous fungi; however, no clinicalevaluation has been done.

Newer antifungal agents have a narrower spectrum of activity,and whether identification to the species level will have aneffect on therapeutic success is not known. Until this questionis resolved, it is necessary for the laboratory to provide themost accurate possible identification of fungi so that the clinical,microbiologic, and treatment outcomes can be determined.

Nucleic acid sequencing appears to provide more objective separationof genera and species than that provided by conventional techniques.Nucleic acid sequencing has already become an important toolthat is useful for the identification of aerobic and anaerobicbacteria, mycobacteria, and fungi (including yeasts) (1, 2,4, 6). Turenne et al. (6) showed that sequencing of the ITS2region has great potential for the identification of yeastsand several of the filamentous fungi, including some speciesof Aspergillus, Zygomycetes, and dermatophytes. A recent reviewby Iwen et al. (3) also discussed the internal transcribed spacerregions as potential targets to identify fungal pathogens inclinical culture specimens. Ninet et al. (5) demonstrated thatthe dermatophyte species could easily be identified by usingthe MicroSeq D2 large-subunit ribosomal DNA (rDNA) fungal sequencingkit and their custom database of sequences derived from well-characterizedisolates.

To our knowledge, this is the first report of an experiencewith the commercially available MicroSeq D2 large-subunit rDNAfungal sequencing kit (Applied Biosystems, Foster City, Calif.)used in a clinical microbiology laboratory to identify filamentousfungi.

MATERIALS AND METHODS

Top

Materials and Methods

Filamentous fungus isolates. A total of 234 filamentous fungi were used in the evaluation(see Tables 1 to 4); all were recovered from clinical specimensat the Mayo Clinic or were referred by other laboratories toMayo Medical Laboratories for identification.

View this table:
[in this window]
[in a new window]
TABLE 1. Comparison of dematiaceous filamentous fungi identification using conventional methods and nucleic acid sequencing ( $<=$ 1%)

View this table:
[in this window]
[in a new window]
TABLE 4. Comparison of hyaline filamentous fungi identification using conventional methods and nucleic acid sequencing ( $>=$ 1%)

Identification of cultures by phenotypic methods. All cultures of filamentous fungi were identified on the basisof a combination of macroscopic and microscopic morphologicalfeatures. Media used to induce sporulation included Sabouraud'sdextrose, inhibitory mold, cornmeal, Czapek-Dox, potato dextrose,and V8 juice agars. Since fungi were not routinely identifiedto the species level by the clinical laboratory, a determinationof how well the library identified the filamentous fungi tospecies was not made in most instances.

Large-subunit ribosomal DNA sequencing. The MicroSeq D2 large-subunit rDNA fungal sequencing kit iscomposed of PCR and cycle sequencing modules, identificationand analysis software, and a library of fungal nucleic acidsequences.

DNA was extracted from fungal cells by placing a 1.0-µlloopful of organism into a 2.0-ml microcentrifuge tube containing200 µl of PrepMan Ultra sample preparation reagent (AppliedBiosystems). Tubes were vortexed for 10 to 30 s, followed byheating at 100°C for 10 min in a heat block. Lysates werestored at -20°C if testing was not performed immediately.

The D2 large-subunit rDNA fragment was amplified by adding 25µl of diluted (1:50) genomic DNA to 25 µl of mastermix consisting of forward and reverse primers to the PCR module.Conditions for PCRs were as follows: 95°C for 10 min, 35cycles each at 95°C for 30 s, 53°C for 30 s, 72°Cfor 1 min, and a final extension step at 72°C for 10 min.

Ten microliters of amplicon was loaded onto a 2% E-Gel, subjectedto electrophoresis, and viewed according to manufacturer's instructions(Invitrogen Life Technologies, Carlsbad, Calif.) to determinewhether PCR products were present.

Purification of the PCR product to remove excess primers andnucleotides was performed using shrimp alkaline phosphatase(2.0 U/µl) and Exonuclease I (10.0 U/µl) (USB Corporation,Cleveland, Ohio). The enzymes were activated for 15 min at 37°C,followed by inactivation at 80°C for 15 min.

After removal of dyes with Sephadex 650, cycle sequencing wasperformed using the sequencing module reagents as follows: 25cycles of 96°C for 10 s, 50°C for 5 s, and 60°Cfor 4 min. Labeled amplicon was placed on an ABI 3100 16 capillarygenetic analyzer (Applied Biosystems).

Analysis of sequence data. All sequence sample files were assembled, edited, and comparedto those in the MicroSeq D2 fungal library, which contained1,072 entries, including 788 species of filamentous fungi (version1.4.2, February 2002). A distance score of 0.00% (100% match)to $<=$ 1.00% (99% match) was used as a guide for identificationsince no cutoff value has yet been determined. The organismchoice giving the closest match was considered the most likelycorrect identification. Organisms having a distance score ofmore than 1.00% were considered to be unique isolates that weremost closely related to the closest database match present inthe library; however, sequences for many of the fungi were notincluded in the D2 library. Sequencing results for the filamentousfungi were available within 24 h (2).

Resolution of discrepant isolates. For instances in which the conventional and sequencing identificationsdiffered, cultures were examined microscopically again for characteristicmorphological features. Isolates with distance scores of $>=$ 1.0 were not reidentified.

Nucleotide sequence accession numbers. Our clinical mycology laboratory has constructed a databaseof additional organisms that is not included in the MicroSeqlibrary and of species that exhibit genetic diversity. Thesemay be found in GenBank and are listed sequentially under accessionnumbers AY234870 to AY235033.

RESULTS

Top

Results

As shown in Tables 1 to 4, 55.9% (131 of 234) of all filamentousfungi included in this evaluation were correctly identifiedto the appropriate genus or genus and species (there were severalinstances where two choices were given) using the MicroSeq D2library. Note that sequences for 70 (29.9%) of the isolates(from 27 genera) studied were not included in the library.

A total of 80 dematiaceous fungi representing 36 genera and/ordifferent species were sequenced, and 52 (65%) gave resultsthat were concordant with those of conventional phenotypic identification.Of these, 39 had a distance score of $<=$ 1% (99% similarity), and31 (79.5%) gave concordant results. The remaining 41 had a distancescore of $>=$ 1%, and 21 (51.2%) gave concordant results(Table 2). Sequences for 12 (15%) of the dematiaceous fungistudied (11 genera) were not included in the MicroSeq library.Dematiaceous fungi with a distance score of $<=$ 1% were placed intothe correct genus using nucleic sequencing with few exceptions,including Chalara (1), Epicoccum (2), and Sporothrix (1) species.However, 14 isolates (11 genera) with a distance score of $>=$ 1% gave discordant results, and 13 (92.9%) were notincluded in the MicroSeq library.

View this table:
[in this window]
[in a new window]
TABLE 2. Comparison of dematiaceous filamentous fungi identification using conventional methods and nucleic acid sequencing ( $>=$ 1%)

A total of 154 hyaline fungi representing 45 genera and/or differentspecies were sequenced, and 88 (57.1%) had a distance scoreof $<=$ 1%; 59 (67.0%) gave results concordant with those of conventionalphenotypic identification (Table 3). The remaining six isolateshad a distance score of $>=$ 1%, and 20 (30.3%) gaveconcordant results (Table 4). Sequences for 58 (37.7%) of thehyaline fungi studied (16 genera) were not included in the MicroSeqlibrary.

View this table:
[in this window]
[in a new window]
TABLE 3. Comparison of hyaline filamentous fungi identification using conventional methods and nucleic acid sequencing ( $<=$ 1%)

Isolates of Aspergillus niger, Aspergillus versicolor, Fusariumspecies, Mucor species, Paecilomyces lilacinus, Penicilliumspecies, Rhizopus species, Scopulariopsis species, and Trichodermaspecies were correctly characterized to the genus level whenthe distance score was $<=$ 1%. Isolates of Aspergillus fumigatus(10), Aspergillus terreus (7), Blastomyces dermatitidis (8),Coccidioides immitis (4), Onychocola canadensis (1), and Penicilliummarneffei (1) gave discordant results since they were not includedin the MicroSeq library. Arthrinium phaeospermum, Aureobasidiumpullulans, Beauveria species, Cokeromyces recurvatus, Engyodontiumalbum, Exophiala jeanselmei, Nigrospora species, Pithomycesspecies, Malbranchea species, Mucor species, Paecilomyces species,and Penicillium species were placed into the appropriate genusdespite having distance scores of $>=$ 1%. Of the 40hyaline fungi (16 genera) with a distance score of $>=$ 1%that gave discordant results, 38 (82.6%) were not included inthe MicroSeq library.

DISCUSSION

Top

Discussion

The Clinical Microbiology Laboratories of the Mayo Clinic usenucleic acid sequencing to identify selected isolates of mycobacteria,yeasts, and aerobic and anaerobic bacteria. The Clinical MycologyLaboratory has the D2 rDNA fungal identification system componentavailable for use. As described previously, the accompanyinglibrary of fungal sequences contains 788 entries that represent205 genera of filamentous fungi, including the type culturesfor most.

The filamentous fungi included in this evaluation were representativeof those seen routinely in clinical microbiology. The organismssent to the Mayo Clinic mycology laboratory were usually thosethat could not be readily identified by the referring laboratories.We sought to determine how well nucleic acid sequencing wouldperform in terms of the identification of these organisms.

Overall nucleic acid sequencing identified 67.5% of the filamentousfungi to the correct genus or genus and species levels. No standardcutoff point is available for interpreting the distance score;however, most of the species with concordant phenotypic scoresof <1% identification had a score of less than 1.00%. Futurestudies will determine if this value is valid, but what is morelikely is that some genetic diversity within some species willbe observed, and the distance scores will have a range.

As has been seen in other areas of clinical microbiology, molecularmethods have provided more accurate identification of organismsand, in some instances, changes in taxonomy. Additional workwill be needed in order to determine which gene or combinationof genes is needed for complete separation of genera and species.Some organisms may require sequencing of more than one targetbefore a definitive identification can be made.

Nucleic acid sequencing will perhaps be of the greatest benefitto the laboratory for the identification of organisms that arenot commonly seen. The MicroSeq D2 library is not inclusiveof all clinically important fungi and should be expanded soas to include them. However, the library's flexibility is whatallows each laboratory to construct a custom database; thisability will make the system even more useful and complete.

The MicroSeq D2 large-subunit rDNA sequencing kit appears tobe accurate and useful for the identification of filamentousfungi—even those that are relatively uncommon—thatare seen in the clinical laboratory. However, the library doesnot include some of the common aspergilli, i.e., Blastomycesdermatitidis and Coccidioides immitis and other environmentalflora that cause disease in immunocompromised patients. Nucleicacid sequencing identification of dimorphic pathogens is notcritical since they may be identified by the use of nucleicacid probes. Sequencing can best be used for the identificationof organisms that cannot be fully identified by their microscopicmorphological features, particularly those that do not sporulatewithin 48 h of subculture. A cost analysis comparing both phenotypicand nucleic acid sequencing for identification showed that nodifference existed in the charges. How often nucleic acid sequencingwill be used by the clinical laboratory must be determined.However, a shortened turnaround time is important to the clinicianand, ultimately, to the patient.

ACKNOWLEDGMENTS

We express our gratitude to David T. Lynch for his technicalexpertise and to Laura Onken for her outstanding secretarialskills. All of the technologists in the clinical mycology andsequencing laboratories contributed much to the success of thisstudy; for this, we express our sincere appreciation. This workcould not have been done without them.

FOOTNOTES

* Corresponding author. Mailing address: Division of Clinical Microbiology, Mayo Clinic, 200 First St. SW, Rochester, MN 55905. Phone: (507) 284-3704. Fax: (507) 284-4272. E-mail: roberts.glenn{at}mayo.edu.

REFERENCES

Top