False Identification of Coccidioides immitis: Do Molecular Methods Always Get It Right?

Abstract

rRNA sequence analysis of a partial region of the 18S and 5.8S-internal transcribed spacer 2 (ITS2) region of Chrysosporium keratinophilum highlights its potential molecular misidentification as Coccidioides immitis. Molecular identification of medically important fungi should not be based solely on sequence analysis of the 18S rRNA gene but should be confirmed by sequence analysis of an additional rRNA gene locus, such as the ITS region(s).

Recently, several groups have advocated the use of molecular methods for the identification of medically important fungi, employing only the 18S rRNA gene locus as the PCR target for sequence analysis (3, 5). Choosing the 18S rRNA gene locus has been popular due to (i) the extensive number of entries to GenBank of 18S rRNA sequence data, (ii) the highly conserved nature of this gene, and (iii) the taxonomic and phylogenetic importance of the 18S rRNA. However, we have recently noted a potential problem with the molecular identification of an isolate of Chrysosporium keratinophilum: the organism could have been mistakenly identified as Coccidioides immitis, with important downstream implications as described below. A white filamentous fungal isolate that had undergone few phenotypic, colonial, or mycological observations was submitted as a known fungus as part of a national laboratory quality control scheme. This isolate was forwarded for molecular identification by using a combination of the 18S and 5.8S rRNA gene loci, as well as with the internal transcribed spacer 1 (ITS1) and ITS2 regions. All gene loci were amplified by PCR (3, 4), and the resulting amplicons were subsequently sequenced in accordance with the method of Millar et al. (6). The resulting novel partial 18S rRNA sequence (GenBank accession number AY234824) was aligned with closely matching sequences with the aid of the BLASTn tool and showed a 99.8% identity (434 of 435 bases) with Coccidioides immitis (GenBank accession number X58571) (Fig. 1A). Further analysis of the 5.8S rRNA-ITS2 region (GenBank accession number AY234823) showed a sole 100% identity (458 of 458 bases) with Chrysosporium keratinophilum (GenBank accession number AJ131681), whereas a Coccidioides sp. (Coccidioides posadasii; GenBank accession number U18360) had an 84.7% identity score, with only 416 of 491 bases identical out of the 458 bases analyzed. Thirty-three gaps were noted within the query sequence (Fig. 1B). Subsequently, phenotypic examination of the filamentous fungus confirmed the isolate to be Chrysosporium keratinophilum, in accordance with previously published characteristics (1). Until now, there has been no sequence information on the 18S rRNA gene of Chrysosporium keratinophilum, and hence those identification methodologies which rely solely upon this target could have potentially misidentified the unknown fungus as its closest phylogenetic neighbor with a high identity score.

FIG. 1.

(A) Sequence alignment of a partial region of the 18S rRNA gene with the query (Chrysosporium keratinophilum; GenBank accession number AY234824) isolate and Coccidioides immitis (GenBank accession number X58571). The shaded region indicates sequence divergence. (B) Sequence alignment of a partial region of the 5.8S rRNA-ITS2 region with the query (Chrysosporium keratinophilum; GenBank accession number AY234823) isolate, a previously submitted sequence of Chrysosporium keratinophilum (GenBank accession number AJ131681) and Coccidioides posadasii (GenBank accession number U18360). The shaded regions indicate sequence divergence. SEQ, sequence.

In the present scenario, the effects of making an incorrect identification through the singular use of the 18S rRNA PCR sequencing approach would have been profound, whereby the presence of a relatively harmless soil-living fungi, Chrysosporium sp., could have been misidentified as the fungal pathogen, Coccidioides immitis. More recently, there has been much concern over the potential dangers of Coccidioides immitis as a bioterrorism agent (2). Should clinical specimens, particularly sputum, be examined for this agent in potential bioterrorism attack situations, molecular methods may be an attractive means of screening, thereby reducing the need for propagating large biomasses of infectious conidia in biosafety level 3 facilities.

In conclusion, one should note that the 18S rRNA gene database is complete neither for medically important fungi nor for environmental fungi and that laboratory identification for which this gene locus is employed as the sole means of identification may present potential misidentifications. We therefore recommend employment of a polyphasic approach, namely, a combination of phenotypic examination and characterization of the fungal isolate in series with a molecular approach that uses the rRNA gene loci. Although conventional mycological techniques would clearly differentiate Coccidioides immitis from Chrysosporium keratinophilum, phenotypic examination of cultures is potentially dangerous to the laboratory worker. Danger also exists when the numbers of specimens examined would constitute loading the laboratory with considerable numbers of fungal spores of a category 3 fungal agent. Because of these factors, diagnostic laboratories may wish to switch to a molecule-based identification method. In such cases, as with the former, an additional rRNA gene locus, such as the 28S or the ITS1 or ITS2 regions, should be sequenced in addition to the 18S rRNA gene; the resulting identifications should then be compared before definitive identifications are made for unknown fungal isolates.