Molecular and Phenotypic Description of Coccidioides Posadasii Sp. Nov., Previously Recognized As the Non-California Population of Coccidioides Immitis
Total Page：16
File Type：pdf, Size：1020Kb
Mycologia, 94(1), 2002, pp. 73±84. q 2002 by The Mycological Society of America, Lawrence, KS 66044-8897 Molecular and phenotypic description of Coccidioides posadasii sp. nov., previously recognized as the non-California population of Coccidioides immitis M. C. Fisher1 dance (GCPSR) to detect genetically isolated groups The Institute of Zoology, Regent's Park, London NW1 by comparing the gene trees from a number of dif- 4RY, UK ferent loci (Avise and Ball 1990). Different genes will G. L. Koenig have different genealogies within a species due to re- T. J. White combination, but between species the genealogies Roche Molecular Systems, 1145 Atlantic Avenue, will be concordant due to the effects of genetic iso- Alameda, California 94501, USA lation and drift causing lineage sorting and coales- J. W. Taylor cence. Detecting the common branches between Department of Plant and Microbial Biology, University these gene trees is the key to GCPSR. of California at Berkeley, Berkeley, California 94720- GCPSR is ®nding increasing usage within both 3102, USA meiosporic and mitosporic fungal taxa (Taylor et al 2000), for instance the Gibberella fujikuroi complex (O'Donnell et al 1998), Ajellomyces capsulatus (Kwon- Abstract: Coccidioides posadasii sp. nov., formerly Chung) McGinnis and Katz (Kasuga et al 1999), As- known as non-California (non-CA) Coccidioides immi- pergillus ¯avus Link (Geiser et al 1998) and Coccidi- tis, is described. Phylogenetic analyses using single oides immitis Rixford and Gilchrist 1896 (Koufopanou nucleotide polymorphisms, genes, and microsatellites et al 1997, 1998). Here, we use GCPSR to demarcate show that C. posadasii represents a divergent, genet- barriers to gene ¯ow between individuals of the path- ically recombining monophyletic clade. Coccidioides ogenic fungus Coccidioides immitis. Our analysis and posadasii can be distinguished from C. immitis by nu- those of others clearly show the existence of two ge- merous DNA polymorphisms, and we show how ei- netically isolated and deeply divergent clades within ther of two microsatellite loci may be used as diag- C. immitis and we use this as the basis for describing nostic markers for this species. Growth experiments a new species, Coccidioides posadasii. Knowledge of show that C. posadasii has signi®cantly slower growth rates on high-salt media when compared with C. im- genetically de®ned species enables workers to look mitis, suggesting that other phenotypic characters closely for previously undetectable morphological may exist. and phenotypic differences. We use this approach to Key Words: allele, Coccidioidomycosis, microsat- show that C. immitis has a tendency to grow faster ellite, Onygenales, phylogeny, systematics than C. posadasii on high-salt media. This demon- strates that other, perhaps clinically important, char- acters may exist. Coccidioides immitis is a dimorphic pathogenic fun- INTRODUCTION gus found in the southwestern United States, Mexico, Species can be de®ned as groups of organisms that Central and South America (Pappagianis 1988). In share a common evolutionary history and, as a con- the saphrobic phase C. immitis is characteristically sequence, share exclusive characters. This is known found inhabiting the arid, sandy soils of the Lower as the evolutionary species concept (ESC; Simpson Sonoran Life Zone. Inhalation of arthroconidia caus- 1951, Simpson 1961, Wiley 1978) and is the most in- es a chronic pulmonary infection in humans and oth- clusive species concept to date (Mayden 1997). Mo- er vertebrates. In ca 0.5% of cases, secondary coccid- lecular genetics (Reynolds and Taylor 1991) and cla- ioidomycosis occurs, a serious disseminated infection distic analyses provide a method of diagnosing spe- that is often fatal (Rippon 1988). Immunity gener- cies under the ESC by describing shared exclusive ated from resolving the infection is speci®c and usu- characters (apomorphies) using an operational ally lifelong. method known as phylogenetic species recognition Coccidioidomycosis was originally described by (PSR). A subset of PSR uses genealogical concor- Alejandro Posadas (and later con®rmed by Robert Wernicke) from a soldier, Domingo Ezcurra, who ac- Accepted for publication June 4, 2001. 1 Corresponding author, phone: 020-7449-6617, Fax: 020-7586-2870, quired his infection in the Argentine pampas (Posa- Email: [email protected] das 1892, Wernicke 1892). Posadas and Wernicke rec- 73 74 MYCOLOGIA ognized the presence of an organism, likened to a This observation suggests that the species described protozoon of the order Coccidia. Formal description here are evolutionary species and could be recog- of C. immitis was performed by Rixford and Gilchrist nized as biological species, as well as phylogenetic from a case observed in California (Rixford and species, if a teleomorph were to be found. Gilchrist 1896). However, the parasite was then still Recently, the sampling of the C. immitis biogeo- thought to be a protozoan. The correct taxonomic graphic distribution was extended to include previ- status of C. immitis as an ascomycete fungus was dem- ously unsampled populations from Southern Califor- onstrated by OphuÈls and Mof®t (1900) by culture on nia, Central and Southern Mexico, Venezuela, and arti®cial media of the fungal mycelia using arthro- Brazil, and analyzed using a suite of microsatellite spores isolated from laboratory infections of guinea markers (Fisher et al 1999, Fisher et al 2000b). Phy- pigs. The etiological relationship between C. immitis logenetic analyses showed that, despite the increased and coccidioidomycosis was also demonstrated by breadth and depth of sampling, the C. immitis phy- showing that arthroconidia cause infection in several logeny still contained two major clades (Fisher et al types of laboratory animal. The lack of any known 2000c). Here, we use our dataset of microsatellite al- meiosporic state in vitro or in vivo hampered further leles to show that these clades correspond to the pre- classi®cation until work by Sigler and Carmichael vious classi®cations of CA (Group II) and non-CA (1976) recognized the similarity between the asexual (Group I) C. immitis. Species rank is proposed for spores (arthroconidia) of C. immitis and those (aleu- the two clades. rioconidia) found in the mitosporic genus Malbran- chea Sacc., placing C. immitis in the order Onygena- ceae. This relationship was con®rmed by molecular MATERIALS AND METHODS phylogenetic methods (Bowman et al 1992, Pan et al One hundred and sixty-seven isolates identi®ed by clinical 1994, Bowman et al 1996), and Uncinocarpus reesii laboratories as Coccidioides immitis were used in this study Sigler and Orr was shown to be the sister group to (APPENDIX). These isolates represent the entire known C. immitis. geographical distribution of the pathogen and are cryogen- Research on the intraspeci®c relationships of C. ically preserved in the Roche Molecular Systems Culture immitis was ®rst attempted by Zimmerman et al Collection (RMSCC) for future reference (Roche Molecu- (1994), who compared RFLPs of total genomic DNA lar Systems, 1145 Atlantic Avenue, Alameda, California 94501, USA). Liquid cultures of each isolate were grown in and showed that 15 clinical isolates formed two a BL3 containment facility and total genomic DNA extract- groups, referred to as Group I and Group II. Group ed from lyophilized mycelia according to the protocol de- I contained the isolate `Silveira' that is extensively scribed by Burt et al (Burt et al 1995). Polymerase chain used in laboratory studies. Subsequent work by Burt reaction (PCR) ampli®cation of nine microsatellite-contain- et al (1997) using RFLPs of 10 DNA loci demonstrat- ing loci (GAC, 621, GA37, GA1, ACJ, KO3, KO7, KO1, ed the occurrence of highly signi®cant differences in KO9) was performed for each isolate using the ¯uores- allele frequencies between clinical isolates from Cal- cently labelled primers and conditions described previously ifornia, Arizona and Texas, the Californian popula- (Fisher et al 1999). The multilocus genotype of each isolate tion being the most divergent. This result was cor- was determined by electrophoresing the PCR products roborated by Koufopanou et al (1997, 1998) who through a 6% denaturing polyacrylamide gel using an au- used genealogies of ®ve nuclear genes to show that tomated sequencer (Applied Biosystems), the alleles pre- sent at each locus being determined by reference against a C. immitis consists of two non-interbreeding taxa, CA TAMRA-labelled internal size standard. A standardized (centered in California) and non-CA (represented by method for typing the alleles at each locus is available as a. clinical isolates from Arizona, Texas, Mexico, and Ar- pdf ®le at http://plantbio.berkeley.edu/;taylor/mf.html. gentina). The `Silveira' isolate was included in non- Phylogenetic analyses were performed using the micro- CA C. immitis, showing that Zimmerman's Group I satellite genetic distances DAS (Stephens et al 1992, Bowcock 2 and Koufopanou's non-CA were synonymous. Nucle- et al 1994) and (dm) (Goldstein et al 1995). Here, DAS 5 otide sequence divergence between CA and non-CA 1 2 (the total number of shared alleles at all loci / n) where showed that they had been reproductively isolated n is the number of loci compared. Pairwise distances cal- from one another for the past 11 million years, a culated using the mean character distance option in PAUP* result that was subsequently corroborated using a 4.0b1 (Swofford 1998) are identical to DAS and were used separate set of loci and C. immitis isolates (Fisher et here. The neighbor-joining algorithm in PAUP clustered these user-de®ned distances using the minimum evolution al 2000b). That independent loci were randomly as- option, and support for each clade was estimated by 1000 sorting with respect to one another within CA and neighbor-joining bootstrap replications of the dataset. Ge- non-CA showed that genetic recombination had oc- netic distance between populations was assessed using the curred between individuals within the two groups, microsatellite distance (dm)2. This measure more closely re- despite no teleomorph ever having been described ¯ects the genetic distance that has accrued within loci by for C.