Figure

Figure. Bayesian phylogenetic analyses of 8 isolates of Cryptococcus gattii sensu stricto from the southeastern United States. We used BEAST 1.8.4 software (http://beast.community) to produce calibrated phylogenies with the mean estimates of...

We identified single-nucleotide polymorphisms (SNPs) and conducted phylogenetic analyses as previously described (10). We identified 43,731 total SNPs among 8 isolates, based on a genome quality breadth of 16,991,136 bases. The maximum-likelihood tree (Figure) had a consistency index of 1.0, and all branching bootstrap values equaled 100 for 100 replicates. Because all isolates in the southeastern United States group contain only the α mating type and the phylogenies have a perfect consistency index (i.e., demonstrating no homoplasy), we assume a clonal expansion for this population (i.e., lack of multiple mating types limits opportunity for recombination), although cryptic recombination cannot be ruled out. We therefore applied the clonal mutation rate obtained from the prior Bayesian molecular clock analysis (using BEAST software [10]) of the Pacific Northwest C. gattii species complex, specifically for VGIIa and VGIIc (1.59 × 10⁻⁸ SNPs/base/y). We determined the best-fitting clock and demographic model combination by implementing path and stepping stone sampling marginal-likelihood estimators as described previously (5). We implemented a general time-reversible nucleotide substitution model and an uncorrelated log-normal molecular clock with a Bayesian skyline model across 4 chains with 3 billion iterations and achieved across- and within-chain convergence.

Employing these previously described methods with a log-normal Bayesian skyline model on the genomes from the 8 isolates from the southeastern United States, BEAST provided an estimated range of time to most recent common ancestor (tMRCA) of approximately 9,000–19,000 years. A general recombining population mutation rate for Cryptococcus was previously estimated at 2.0 × 10⁻⁹ (11), which is an order of magnitude slower than the estimated clonal rate; therefore, applying this rate provides for a tMRCA that is nearly an order of magnitude greater (77,000–148,000 years).

The clonal mutation rate appears to be more appropriate than the general mutation rate for recombining populations for estimating tMRCA, showing that the timing of C. gattii emergence and subsequent divergence in the southeastern United States is clearly much older than the Pacific Northwest emergence. The sharp contrast in the time of arrival compared with C. deuterogattii in the Pacific Northwest is considerable. Although C. gattii has apparently been in the southeastern United States for thousands more years than C. deuterogattii has been in the Pacific Northwest, the number of cases detected in the southeastern United States is far fewer. Infections in the South are probably not regularly missed, given that C. gattii in the southeastern United States causes primarily a devastating meningitis or meningoencephalitis (6,9). Questions of whether subacute cases might be going undetected, whether the distribution of the fungus in the environment might be lower, or whether the niche is not readily accessible by humans remain unanswered. We clearly know less about the Cryptococcus species that has been in the United States for thousands of years (C. gattii sensu stricto) than we know about the one that has only recently arrived (C. deuterogattii).

The Pacific Northwest emergence has been hypothesized to be related to the opening of the Panama Canal, enabling more shipping from eastern ports of South America to the West Coast of North America (5). The estimated timing of emergence of C. gattii in the southeastern United States occurred long before industrial shipping. Although humans might have populated the southeastern United States by the time of the dispersal, most of the human movement was north to south rather than the reverse (12). This relatively recent emergence during the Pleistocene epoch is different from what has been hypothesized for other endemic mycoses, a much older fungal species dispersal resulting from mass migration of animal populations between continents (13). The idea of recent emergence also departs sharply from what has been hypothesized as a speciation attributable to prehistoric land movements, as has been proposed for larger Cryptococcus species separations (14). C. gattii sensu stricto has been found in and likely originates from South American locales (15). Besides anthropogenic means, possible dispersal mechanisms out of South America might have included animal, bird, or insect species migration or even ocean detritus brought by Caribbean currents or hurricanes. However, given the lack of population structure, dispersal probably occurred through a discrete event or through limited events from the same originating population. No matter the mechanism of arrival, C. gattii has been hiding, mostly undetected, in the southeastern United States for millennia.