Promiscuous mitochondria in Cryptococcus gattii

Abstract

Cryptococcus gattii is a primary pathogenic basidiomycetous yeast comprising four genotypic groups. Here we present data on two mitochondrial loci (MtLrRNA and ATP6). Two of the genotypic groups, namely amplified fragment length polymorphism (AFLP)5/VGIII and AFLP6/VGII, formed monophyletic lineages. The AFLP4/VGI genotypic group, however, possessed five different mitochondrial genotypes that did not form a monophyletic lineage. The majority of these isolates contained mitochondrial genomes that are partially identical to those found in isolates belonging to AFLP6/VGII, which is causing the ongoing and expanding Vancouver Island outbreak. Two out of four AFLP7/VGIV isolates contained an AFLP4/VGI allele of MtLrRNA. These observations are best explained by assuming a process of mitochondrial recombination. If this is true, mitochondrial recombination seems possible between cells belonging to different genotypic groups of C. gattii, especially between AFLP6/VGII or AFLP7/VGIV and AFLP4/VGI. We also have to assume that mitochondria, most likely, were transferred from cells belonging to AFLP6/VGII to AFLP4/VGI. As such a process of mitochondrial recombination is only possible after cell–cell conjugation, this may also allow the further exchange of genetic material, for example nuclear or plasmid in nature, between different genotypes of C. gattii. This may be relevant as it may provide a possible mechanism contributing to the modulation of virulence attributes of isolates, such as has been observed in the ongoing Vancouver Island outbreak of C. gattii.

Cryptococcus gattii, pathogen, mitochondria, recombination

Introduction

Cryptococcus gattii is a primary pathogenic basidiomycetous yeast that has previously been considered a variety of Cryptococcus neoformans (Kwon-Chung, 1982). Recently, C. gattii has been described as a distinct species because of differences in ecology, epidemiology, biochemical and molecular characteristics (Kwon-Chung, 2002; Kwon-Chung & Varma, 2006). Both C. gattii and C. neoformans may cause meningoencephalitis, but the primary pathogen C. gattii can cause disease in otherwise healthy individuals, whereas the opportunist C. neoformans primarily infects immunocompromised patients (Speed & Dunt, 1995). Cryptococcus gattii has been found on several tree species since its first isolation from Eucalyptus camaldulensis (Ellis & Pfeiffer, 1990; Callejas, 1998; Lazéra, 1998, 2000; Fortes, 2001; Krockenberger, 2002; Fraser, 2003; Randhawa, 2003; Granados & Castañeda, 2005; Escandón, 2006; Kidd, 2007), suggesting that trees might be the primary ecological niche of the species. In contrast, Kidd (2007) suggested that soil may be the principal reservoir for C. gattii. Cryptococcus gattii occurs predominantly in subtropical areas (Kwon-Chung & Bennett, 1984), but has also been isolated in Europe (Viviani, 2006) and in a temperate climate zone in Colombia (Escandón, 2006). Furthermore, C. gattii is responsible for the ongoing outbreak of cryptococcosis on Vancouver Island, Canada (Stephen, 2002; Hoang, 2004; Kidd, 2004) and the further spread of this outbreak to the Pacific Northwest of the United States (MacDougall, 2007; Upton, 2007). Taken together, these reports indicate that C. gattii also occur in more temperate climate zones.

Several molecular methods have identified four major genotypic groups within C. gattii (Ruma, 1996; Ellis, 2000; Chaturvedi, 2002; Biswas, 2003; Latouche, 2003; Meyer, 2003; Butler & Poulter, 2005; Diaz, 2005; Fraser, 2005; Kidd, 2005; Bovers, 2008b, c). An overview of these genotypic groups is given in Table 1. All these studies used nuclear regions to study the genotypic structure of C. gattii. Mitochondrial regions are useful genetic markers as well, because mitochondria evolve independently from the nuclear genome and thus provide an additional, independent dataset. In C. neoformans, mitochondria are usually uniparentally inherited from the MATa parent, although mitochondrial inheritance from the MATα parent may also occur at low frequency (Xu, 2000;; Yan & Xu, 2003; Toffaletti, 2004; Yan, 2004, 2007a, b). In our study, the partial nucleotide sequence of two mitochondrial regions, namely large ribosomal subunit RNA (MtLrRNA) and ATP synthase subunit 6 (ATP6), was investigated for 51 isolates representing all C. gattii genotypic groups. The sequences obtained were used for phylogenetic analyses and the resulting genealogies were compared with the genealogy obtained by analyses of six nuclear regions (Bovers, 2008b). Surprisingly, the analysis of the mitochondrial regions showed that the majority of amplified fragment length polymorphism AFLP4/VGI isolates studied possessed a mitochondrial genome that contained sequences that were partially identical to those found in AFLP6/VGII isolates. These findings are discussed and explained by assuming a process of mitochondrial recombination between the different genotypic groups of C. gattii.

1

Overview of Cryptococcus gattii genotypes as identified by several molecular methods

Serotype	AFLP genotype	Molecular genotype	IGS genotype	ITS genotype	Intein genotype
B/(C)	4A/4B	VGI	4a/4b/4c	3/7	ig-I
B/C	5A/5B/5C	VGIII	5	5	ig-III
B/(C)	6	VGII	3	4	ig-II
B/C	7	VGIV	6	6	ig-IV

*

Meyer (2003).

†

Barreto de Oliveira (2004).

‡

Boekhout (2001).

§

Kidd (2004).

¶

Litvintseva (2006).

∥

Diaz (2005).

**

Katsu (2004).

††

Butler & Poulter (2005).

‡‡

^‡‡Serotypes indicated between parentheses were not included in our study.

Materials and methods

Isolates

Fifty-one haploid isolates belonging to all C. gattii genotypic groups from clinical (59%), veterinary (8%), environmental (24%), laboratory (6%) and unknown (3%) origin were analyzed (Table 2). Cryptococcus neoformans isolates (125.91, H99, WM714, CBS6886 and JEC20) were used as outgroup in phylogenetic analyses, as C. neoformans is the species that is most closely related to C. gattii (Fell, 2000; Kwon-Chung, 2002).

2

AFLP genotype, molecular type, serotype and origin of Cryptococcus gattii and Cryptococcus neoformans isolates

Isolate	Origin	Sero type	Mating type	Nuclear genotype*	AFLP geno type‡§	Mole cular type¶	MtLrRNA haplotype	ATP6 haplotype	References /sources	MtLrRNA	ATP6	ITS	IGS	RPB1	RPB2	CNLAC1	TEF1α
C. gattii AFLP4=VGI
48A (=CBS11230)	Lung of a goat, Spain	B	α	4A	4A	ND	2	4	Diaz (2000)	EF544832	EF544778	EF211195	AJ300934∥	EF211403	EF211523	EF211643	EF211758
503 2738 (=WM1251)	Human, Papua, New Guinea	B	α	4	4	VGI	2	1	Katsu (2004)	EF544833	EF544779	EF211196	EF211302	EF211404	EF211524	EF211644	EF211759
56A (=CBS11231)	Gut of a goat, Spain	B	α	4A	4A	ND	2	4	Diaz (2000)	EF544834	EF544780	EF211197	AJ300932∥	EF211405	EF211525	EF211645	EF211760
CBS883 (T)	Infected skin, syntype Cryptococcus hondurianus, Honduras	B	α	4	4B	ND	4	2	Boekhout (1997)	EF544835	EF544781	EF211198	EF211303	EF211406	EF211526	EF211646	EF211761
CBS919 (T)	Meningoencephalic lesion, type strain of Torulopsis neoformans var. sheppei	B	α	4	4B	ND	4	2	Boekhout (1997)	EF544836	EF544782	EF211199	AJ300928∥	EF211407	EF211527	EF211647	EF211762
CBS1622	Man, Tumour, Lille, France	B	α	4	4B	ND	2	3	Boekhout (1997)	EF102067**	EF102038**	EF102028**	EF102032**	EF102061**	EF102051**	EF102071**	EF102047**
CBS6289 (=CBS8273, =RV20186, =NIHB-3939)	Subculture of type strain of Cryptococcus gattii (RV 20186)	B	a	4A	4A	ND	2	4	Boekhout (1997)	EF544837	EF544783	EF211200	AJ300937∥	EF211408	EF211528	EF211648	EF211763
CBS6290	Man, Congo (Zaire)	B	α	4A	4A	ND	2	4	Boekhout (1997)	EF544838	EF544784	EF211201	AJ300930∥	EF211409	EF211529	EF211649	EF211764
CBS6992 (=NIH 17)	Man	B	α	4	4B	ND	2	3	Boekhout (1997)	EF102068**	EF102039**	EF102029**	AJ300923∥	EF102033**	EF102062**	EF102052**	EF102072**	EF102048**
CBS6998 (=NIH 365)	Human, Thailand	B	a	4	4	ND	2	4	Boekhout (1997)	EF544839	EF544785	EF211202	AJ300925∥	EF211410	EF211530	EF211650	EF211765
CBS7229 (T)	Meningitis, type strain Cryptococcus neoformans var. shanghaiensis, China	B	a	4	4B	ND	2	1	Boekhout (1997)	EF544840	EF544786	EF211203	AJ300926∥	EF211411	EF211531	EF211651	EF211766
CBS7748 (=IFM50902)	Air in Eucalyptus camaldulensis hollow, Balranald, SA, Australia	B	α	4	4B	VGI	4	2	Boekhout (1997)	EF544841	EF544787	EF211204	AJ300927∥	EF211412	EF211532	EF211652	EF211767
CBS8273 (=CBS6289, =RV20186, =NIH B-3939)	Subculture of type strain of Cryptococcus gattii (RV 20186)	B	a	4A	4	ND	2	4	Boekhout (1997)	EF544842	EF544788	EF211205	EF211304	EF211413	EF211533	EF211653	EF211768
E566 (=CBS11233)	E. camaldulensis tree #19 hollow 4, Renmark Australia	B	a	4	4	VGI	1	2	Halliday (2000)	EF544843	EF544789	EF211206	EF211305	EF211414	EF211534	EF211654	EF211769
RV20186 (=CBS6289, =CBS8273, =NIH B-3939) (T)	Cerebrospinal fluid, type strain of Cryptococcus gattii, Congo (Zaire)	B	a	4A	4A	ND	2	4	Gatti & Eeckels (1970)	EF544844	EF544790	EF211207	EF211306	EF211415	EF211535	EF211655	EF211770
RV54130	Second isolate of C. neoformans var. shanghaiensis, China	B	a	4	4B	ND	2	1	Boekhout (2001)	EF544845	EF544791	EF211208	EF211307	EF211416	EF211536	EF211656	EF211771
WM176	Eucalyptus citriodora,USA	B	α	4	4B	ND	2	1	Boekhout (2001)	EF544846	EF544792	EF211209	EF211308	EF211417	EF211537	EF211657	EF211772
WM179 (R)	Immunocompetent human, reference strain of molecular type VGI, Sydney, Australia	B	α	4	4	VGI	1	2	Meyer (2003)	EF544847	EF544793	EF211210	EF211309	EF211418	EF211538	EF211658	EF211773
WM276 (=CBS10510)	Eucalyptus tereticornis, Mt. Annan, NSW, Australia	B	α	4	4	VGI	2	1	Kidd (2005)	EF544848	EF544794	EF211211	EF211310	EF211419	EF211539	EF211659	EF211774
WM830	Immunocompetent, Human, Papua. New Guinea	B	a	4	4	VGI	2	1	Katsu (2004)	EF544849	EF544795	EF211212	EF211311	EF211420	EF211540	EF211660	EF211775
C. gattii AFLP5=VGIII
380C	Unknown	C	α	5	5C	ND	5	5	Boekhout (2001)	EF544850	EF544796	EF211213	EF211312	EF211421	EF211541	EF211661	EF211776
384C	Patient	C	α	5	5C	ND	5	8	Boekhout (2001)	EF544851	EF544797	EF211214	EF211313	EF211422	EF211542	EF211662	EF211777
CBS5758	Unknown	C	α	5	5C	ND	5	5	Boekhout (1997)	EF544852	EF544798	EF211215	AJ300929∥	EF211423	EF211543	EF211663	EF211778
CBS6955 (T) (=NIH 191)	Human, type strain of Cryptococcus bacillisporus, CA, USA	C	a	5	5C	ND	5	7	Boekhout (1997)	EF544853	EF544799	EF211216	AJ300940∥	EF211424	EF211544	EF211664	EF211779
CBS6993 (=NIH 18)	Human, CA, USA	C	α	5	5C	ND	5	5	Boekhout (1997)	EF544854	EF544800	EF211217	EF211314	EF211425	EF211545	EF211665	EF211780
CBS6996	Man	B	α	5	5C	ND	5	8	Boekhout (1997)	EF544855	EF544801	EF211218	AJ300939∥	EF211426	EF211546	EF211666	EF211781
CBS8755 (=HOO58-I-682)	Detritus of almond tree, Colombia	C	α	5	5A	ND	5	5	Boekhout (2001)	EF544856	EF544802	EF211219	EF211315	EF211427	EF211547	EF211667	EF211782
CN043 (=CBS11247)	Human, Auckland, New Zealand	B	α	5	5	VGIII	5	6	Katsu (2004)	EF544857	EF544803	EF211220	EF211316	EF211428	EF211548	EF211668	EF211783
WM161 (R)	Eucalyptus camaldulensis wood from hollow, reference strain of molecular type VGIII, San Diego, CA, USA	B	α	5B	5B	VGIII	5	5	Meyer (2003)	EF544858	EF544804	EF211221	EF211317	EF211429	EF211549	EF211669	EF211784
WM726	Eucalyptus citriodora, San Diego, CA, USA	B	α	5B	5B	VGIII	5	5	Boekhout (2001)	EF544859	EF544805	EF211222	EF211318	EF211430	EF211550	EF211670	EF211785
WM728	Eucalyptus sp. debris from car park of zoo, San Diego, CA, USA	B	α	5B	5B	VGIII	5	5	Boekhout (2001)	EF544860	EF544806	EF211223	EF211319	EF211431	EF211551	EF211671	EF211786
C. gattii AFLP6=VGII
A1M F2866 (=CBS11240)	Dead wild Dall's porpoise lymph node, Shores of Gulf Island, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544861	EF544807	EF211224	EF211320	EF211432	EF211552	EF211672	EF211787
A1M F2932 (=CBS11235)	Immunocompetent male, lung, Kelowna, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544862	EF544808	EF211225	EF211321	EF211433	EF211553	EF211673	EF211788
A1M R265 (=CBS10514)	Immunocompetent male, Duncan, Vancouver Island, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544863	EF544809	EF211226	EF211322	EF211434	EF211554	EF211674	EF211789
A1M R269 (=CBS11236)	Immunocompetent female, Victoria, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544864	EF544810	EF211227	EF211323	EF211435	EF211555	EF211675	EF211790
A1M R271 (=CBS11237)	Immunocompetent male, Nanoose Bay, Vancouver Island, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544865	EF544811	EF211228	EF211324	EF211436	EF211556	EF211676	EF211791
A1M R368 (=CBS11238)	Immunocompetent male, Victoria, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544866	EF544812	EF211229	EF211325	EF211437	EF211557	EF211677	EF211792
A1M R406 (=CBS11239)	Immunocompetent female, Nanaimo, Vancouver Island, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544867	EF544813	EF211230	EF211326	EF211438	EF211558	EF211678	EF211793
A1M R409	Immunocompetent female, Victoria, Canada	B	α	6	6A	VGIIa	1	2	Kidd (2004)	EF544868	EF544814	EF211231	EF211327	EF211439	EF211559	EF211679	EF211794
CBS1930	Sick goat, Aruba	B	a	6	6B	ND	1	2	Boekhout (1997)	EF544869	EF544815	EF211232	AJ300919∥	EF211440	EF211560	EF211680	EF211795
CBS6956 (=NIH 444, =ATCC32609)	Immunocompetent human, Seattle, WA, USA	B	α	6	6A	VGII	1	2	Boekhout (1997)	EF544870	EF544816	EF211233	AJ300920∥	EF211441	EF211561	EF211681	EF211796
CBS7750	Eucalyptus camaldulensis bark debris, San Francisco, CA, USA	B	α	6	6A	VGII	1	2	Boekhout (1997)	EF544871	EF544817	EF211234	AJ300922∥	EF211442	EF211562	EF211682	EF211797
CBS8684	Nest of wasp, Uruguay	B	α	6	6B	ND	1	2	Boekhout (2001)	EF544872	EF544818	EF211235	EF211328	EF211443	EF211563	EF211683	EF211798
HEC11102	Human, Rio de Janeiro, Brazil	B	a	6	6B	VGII	1	2	Katsu (2004)	EF544873	EF544819	EF211236	EF211329	EF211444	EF211564	EF211684	EF211799
ICB184 (=MTPI2, =CBS11251)	Hollow trees, Piaui, Brazil	B	α	6	6B	ND	1	2	Barreto de Oliveira (2004)	EF544874	EF544820	EF211237	EF211330	EF211445	EF211565	EF211685	EF211800
RAM2	Eucalyptus camaldulensis, Arnhemland, NT, Australia	B	α	6	6B	VGII	1	2	Katsu (2004)	EF544875	EF544821	EF211238	EF211331	EF211446	EF211566	EF211686	EF211801
WM178 (=IFM 50894) (R)	Immunocompetent human, lung, reference strain of molecular type VGII, Sydney, Australia	B	α	6	6A	VGII	1	2	Meyer (2003)	EF544876	EF544822	EF211239	EF211332	EF211447	EF211567	EF211687	EF211802
C. gattii AFLP7=VGIV
B-5742	Human, Punjab, India	C	α	7	7	VGIV	3	9	Katsu (2004)	EF544877	EF544823	EF211240	EF211333	EF211448	EF211568	EF211688	EF211803
B-5748	HIV positive patient, India	C	α	7	7	ND	3	9	Diaz & Fell (2005)	EF544878	EF544824	EF211241	EF211334	EF211449	EF211569	EF211689	EF211804
M27055	Clinical, Johannesburg, South Africa	C	α	7	7	VGIV	2	9	Latouche (2002)	EF544879	EF544825	EF211242	EF211335	EF211450	EF211570	EF211690	EF211805
WM779 (=IFM50896) (R)	Cheetah, reference strain of molecular type VGIV, Johannesburg, South Africa	C	α	7	7	VGIV	2	9	Meyer (2003)	EF544880	EF544826	EF211243	EF211336	EF211451	EF211571	EF211691	EF211806
C. neoformans var. grubii outgroup isolates
125.91 (=CBS10512)	Cryptococcal meningitis patient, Tanzania	A	a	1	1	ND	NA	NA	Lengeler (2002)	EF544827	EF544773	EF211129	EF211249	EF211337	EF211457	EF211577	EF211692
H99 (=CBS8710, =CBS10515) (T)	Patient with Hodgkin's disease, type strain of Cryptococcus neoformans var. grubii, NY, USA	A	α	1	1	VNI	NA	NA	Franzot (1999)	EF544828	EF544774	EF211146	EF211264	EF211354	EF211474	EF211594	EF211709
WM714	Cat paranasal, Australia	A	α	1B	1B	ND	NA	NA	Boekhout (2001)	EF544829	EF544775	EF211174	EF211289	EF211382	EF211502	EF211622	EF211737
C. neoformans var. neoformans outgroup isolates
CBS6886 (=NIH 430)	Dropping of pigeon	D	α	2	2	ND	NA	NA	Boekhout (1997)	EF544830	EF544776	EF211182	AJ300886∥	EF211390	EF211510	EF211630	EF211745
JEC20 (=CBS10511, =NIH-B4476)	Congenic pair with JEC21 that differs only in matingtype	D	a	2	2	ND	NA	NA	Kwon-Chung (1992)	EF544831	EF544777	EF211189	EF211296	EF211397	EF211517	EF211637	EF211752

*

Bovers et al. (2008c). †Boekhout (2001).

‡

Barreto de Oliveira (2004).

§

F. Hagen, (unpublished data).

¶

Meyer (2003).

∥

Diaz (2000).

**

Bovers (2008a).

MtLrRNA and ATP6 haplotypes are indicated for Cryptococcus gattii isolates and GenBank accession numbers are given for all sequenced regions.

ND, not determined; NA, not applicable; (T), type strain; (R), reference strain.

Cultivation, DNA extraction, PCR reaction and sequencing

Cultivation and DNA extraction of C. neoformans and C. gattii isolates was carried out using an optimized protocol of Bolano (2001), which has been described previously (Bovers, 2006). PCR reactions were performed in a total volume of 50 μL containing 10 × PCR buffer (10 mM Tris-HCl, 50 mM KCl, 1.5 mM MgCl₂, 0.01% gelatin, 0.1% Triton X-100, pH 8.3), 0.2 mM dNTPs, 0.6 μM of primer, 1.0 U Taq DNA polymerase (Gentaur, Brussels, Belgium) and 1–2 μL genomic DNA. Amplification of MtLrRNA was carried out using primers MtLrRNA-F (5′-GACCCTATGCAGCTTCTACTG-3′) and MtLrRNA-R (5′-TTATCCCTAGCGTAACTTTTATC-3′) (White, 1990). Amplification conditions were 94 °C for 3 min, followed by 35 cycles of 94 °C for 1 min, 50 °C for 1 min and 72 °C for 1 min, with a final extension step of 72 °C for 5 min. Mitochondrial ATP6 was amplified using primers ATP6-F (5′-ATTACATCTCCACTAGAACAATTC-3′) and ATP6-R (5′-AGTTCAATGGCATCCTTGATATAG-3′). Amplification conditions were 94 °C for 5 min, followed by 35 cycles of 94 °C for 1 min, 54 °C for 1 min and 72 °C for 2 min, with a final extension step of 72 °C for 5 min. Amplicons were purified with the GFX™ PCR DNA and Gel Band Purification Kit (Amersham Biosciences, Piscataway, NY) and used for sequencing. Sequencing reactions were performed with the BigDye v3.1 Chemistry kit (Applied Biosystems, Foster City, CA) using primers that had been used in the initial PCR reactions. Sequencing reactions were purified with Sephadex G-50 Superfine columns (Amersham Biosciences) and a MultiScreen HV plate (Millipore, Billerica, MA). An ABI 3700XL DNA analyzer (Applied Biosystems) was used to determine the sequences. GenBank accession numbers are listed in Table 2. Three isolates (CBS6289, CBS8273 and RV20186) were independently included in the analysis, but could be traced back to the same original isolate, and accordingly were found to have the same mitochondrial haplotypes. To confirm the correctness of our sequence data, especially related to the observed chimeric sequences, single colonies were selected, and processed for DNA extraction, PCR and sequence analysis. Furthermore, in an independent research project, a number of isolates from each lineage that were included in the present study were independently processed for the analysis of ATP6 sequences. Both gave the same sequences; hence, we are confident that we have been working with correct sequences (F. Hagen & T. Boekhout, unpublished data).

Alignment and phylogenetic analyses

The program seqman 5.03 (DNASTAR, Madison, WI) was used to assemble consensus sequences and these were checked manually. Sequences were aligned using clustalx version 1.8 (Thompson, 1997) and visually corrected using genedoc version 2.5.000 (http://www.nrbsc.org/downloads/). Neighbor-Joining (NJ) and Maximum Parsimony (MP) phylogenetic analyses were performed using paup^* (phylogenetic analysis using parsimony) version 4.0b10 (Swofford, 2000). NJ analyses were carried out with the uncorrected (‘p’), Jukes–Cantor, Kimura two-parameter and HKY85 substitution models. Any ties that were encountered were broken randomly. Bootstrap analysis (Hillis & Bull, 1993) with 1000 replicates was used to determine the significance of branches. MP analyses were carried out (heuristic search, stepwise addition, random taxon addition, 1000 maximum trees) with tree bisection and reconstruction as the branch-swapping algorithm. All characters were unordered and of equal weight; gaps were treated both as missing and as a new character state. Bootstrap analysis was performed with 1000 replicates.

Results

ATP6

Analysis of partial ATP6 sequences showed that among 611 analyzed nucleotides, 34 nucleotides were polymorphic within C. gattii. Nine haplotypes could be identified (Fig. 1a). Haplotype 1 was present in AFLP4/VGI isolates 503 2738, CBS7229, RV54130, WM176, WM276 and WM830. All AFLP6/VGII isolates and AFLP4/VGI isolates WM179, E566, CBS883, CBS919 and CBS7748 possessed haplotype 2. Haplotype 3 was identified in AFLP4/VGI isolates CBS1622 and CBS6992. Haplotype 4 was found in AFLP4/VGI isolates 48A, 56A, CBS6289, CBS6290, CBS6998, CBS8273 and RV20186. AFLP5/VGIII isolates 380C, CBS5758, CBS6993, CBS8755, WM161, WM726 and WM728 possessed haplotype 5, whereas haplotype 6 was present in AFLP5/VGIII isolate CN043. Haplotype 7 was found in AFLP5/VGIII isolate CBS6955 and haplotype 8 was present in AFLP5/VGIII isolates 384C and CBS6996. All AFLP7/VGIV isolates possessed haplotype 9. The ATP6 haplotype of each isolate is indicated in Fig. 1a and Table 2. The alignment of the partial ATP6 sequences is given in Supporting Information, Fig. S1.

1

Phylogenetic tree of Cryptococcus gattii obtained by analysis of (a) partial ATP6 sequences. Presented is one of the four most parsimonious trees (length 83, consistency index 0.819, retention index 0.974) computed with gaps treated as missing data. Data consisted of 611 characters of which 64 characters were parsimony informative; (b) partial MtLrRNA sequences. Presented is one of three most parsimonious trees (length 15, consistency index 1.000, retention index 1.000) computed with gaps treated as missing data. Data consisted of 299 characters of which 13 characters were parsimony informative. Bootstrap values (1000 replicates) are indicated for the main branches. h, haplotype number; M, mitochondrial genotype. Haplotypes of both ATP6 and MtLrRNA are indicated by the colors given in Fig. 2. In addition, those not included in Fig. 2 are as follows: ATP6 – h5, grey; h6, ochre; h7, orange; h8, brown; h9, black; MtLrRNA – h3, black; h5, grey.

NJ and MP analyses showed a division of C. gattii into two clades. One clade consisted of haplotypes 1–3, and was found to be supported by bootstrap values ranging from 84% to 91% for NJ and MP analyses, respectively. The other clade contained haplotypes 4–9 and was found to be strongly supported by bootstrap values ranging from 93% to 99% for NJ and MP analyses. Haplotype 1 formed a separate cluster supported by bootstrap values ranging from 94% to 99% for NJ and from 75% to 76% for MP analyses. The cluster consisting of haplotypes 2 and 3 was strongly supported by bootstrap values ranging from 95% to 100% for NJ and MP analyses. Haplotype 4 formed a separate cluster supported by bootstrap values ranging from 97% to 99% for NJ and from 74% to 75% for MP analyses. Haplotypes 5–8 formed a cluster that was supported by bootstrap values ranging from 92% to 94% for NJ and MP analyses. Haplotype 5 formed a separate cluster in NJ analyses, supported by bootstrap values ranging from 87% to 97% in NJ and 94% in MP analyses. Haplotypes 6–8 also formed a strongly supported cluster with bootstrap values ranging from 93% to 95% for NJ and from 82 to 83% for MP analyses. A separate haplotype 8 cluster was strongly supported by bootstrap values ranging from 98% to 99% for NJ and MP analyses. Haplotype 9 formed a separate cluster strongly supported by bootstrap values ranging from 97% to 100%.

MtLrRNA

Analysis of partial MtLrRNA sequences resulted in the identification of five C. gattii haplotypes (Fig. 1b, Table 2). The alignment of the five haplotypes is shown in Fig. S2. Four polymorphic nucleotides, which are indicated within brackets, were present among 299 analyzed nucleotides. Haplotype 1 (CTAG) was present in all AFLP6/VGII isolates and in AFLP4/VGI isolates WM179 and E566; haplotype 2 (CTAC) was present in most of the AFLP4/VGI isolates and in AFLP7/VGIV isolates M27055 and WM779; haplotype 3 (CTTC) was present in AFLP7/VGIV isolates B5742 and B5748; haplotype 4 (CGAC) was present in AFLP4/VGI isolates CBS883, CBS919 and CBS7748; and haplotype 5 (GTAA) was present in all AFLP5/VGIII isolates. An overview of the MtLrRNA haplotype of all isolates is given in Table 2. Both NJ and MP analyses positioned haplotype 1 basal to the other haplotypes, but the topology of the tree of haplotypes 1–4 was not well supported. Haplotype 5 formed a separate cluster in NJ and MP analyses with bootstrap values ranging from 80% to 86%.

Combined analysis of mitochondrial data

All AFLP6/VGII isolates contained MtLrRNA haplotype 1 and ATP6 haplotype 2, whereas AFLP7/VGIV isolates possessed MtLrRNA haplotypes 2 or 3 that differed by one nucleotide, and ATP6 haplotype 9. The AFLP5/VGIII isolates contained MtLrRNA haplotype 5 and ATP6 haplotypes 5–8 that formed one cluster. AFLP4/VGI isolates possessed MtLrRNA haplotypes 1–4 and ATP6 haplotypes 1–4. Surprisingly, these haplotypes did not cluster together. The AFLP4/VGI isolates could be divided into five mitochondrial genotypes based on the presence of ATP6 and MtLrRNA alleles (viz. M1–M5). Genotype AFLP4/VGI-M1 (MtLrRNA haplotype 2 and ATP6 haplotype 4, i.e. isolates 48A, 56A, CBS6289, CBS6290, CBS6998, CBS 8273, RV 20186) corresponded to the previously described genotypic subgroup AFLP4A (Boekhout, 2001). This genotype formed a clade with AFLP5/VGIII and AFLP7/VGIV isolates in the ATP6 analyses (Fig. 1a) similar to the topology obtained by analyses of six concatenated nuclear regions (see Bovers, 2008b, Fig. 3). Genotype AFLP4/VGI-M1 thus appears to be the core AFLP4/VGI genotype, because they contain AFLP4/VGI-typical alleles of both the nuclear and mitochondrial loci, and in the phylogenetic analysis of the nuclear genes these isolates appeared basal. Genotype M2 of AFLP4/VGI (MtLrRNA haplotype 2 and ATP6 haplotype 1, i.e. isolates 503 2738, CBS7229, RV54130, WM176, WM276 and WM830) possessed the core AFLP4/VGI MtLrRNA sequence. Interestingly, the ATP6 sequence of this genotype, i.e. haplotype 1, was found to be a chimerical sequence in all the isolates studied. This sequence was partially identical to the sequence found in AFLP6/VGII isolates (i.e. h2) and partially identical to the sequence of isolates with a core AFLP4/VGI mitochondrial sequence (i.e. h4, Fig. S3). Genotype AFLP4/VGI-M3 (MtLrRNA haplotype 2 and ATP6 haplotype 3, i.e. isolates CBS1622, CBS6992) possessed the core AFLP4/VGI MtLrRNA sequence and an ATP6 sequence that differed only at one nucleotide position from the ATP6 sequence found in AFLP6/VGII isolates. Genotype AFLP4/VGI-M4 (MtLrRNA haplotype 4 and ATP6 haplotype 2, i.e. isolates CBS883, CBS919, CBS7748) possessed a MtLrRNA sequence that differed one nucleotide from the core AFLP4/VGI MtLrRNA sequence and an ATP6 sequence identical to that of AFLP6/VGII isolates. Genotype AFLP4/VGI-M5 (MtLrRNA haplotype 1 and ATP6 haplotype 2, i.e. isolates E566, WM179 possessed MtLrRNA and ATP6 sequences identical to those found in AFLP6/VGII isolates. Isolate WM179 and isolate E566 that both belong to the mitochondrial genotype M5, have MATα and MATa, respectively. This indicates that the presence of the AFLP6/VGII allele in AFLP4/VGI isolates does not depend on the MAT locus. Thus, the mitochondrial genome of part of the AFLP4/VGI isolates contained sequences that were partially identical to those found in AFLP6/VGII isolates. The incongruence between the mitochondrial genotypes of AFLP4/VGI and the topology of the AFLP4/VGI clade as derived from the analysis of six concatenated nuclear regions (Bovers, 2008b) is illustrated in Fig. 2.

2

Comparison of mitochondrial genotypes and a phylogenetic tree of genotype AFLP4/VGI of Cryptococcus gattii obtained by analysis of six concatenated nuclear regions (RPB1, RPB2, CNLAC1, TEF1α, IGS1 and internal transcribed spacer). Presented is one of the 60 most parsimonious trees (length 847, consistency index 0.902, retention index 0.961) computed with gaps treated as missing data. Data consisted of 3932 characters of which 459 characters were parsimony informative. Bootstrap values (1000 replicates) are indicated. The mitochondrial genotype as well as the ATP6 and MtLrRNA haplotypes are shown. T, type strain; R, molecular type reference strain (Meyer, 2003). Haploid isolates are underlined.

Phylogenetic analyses of partial ATP6 sequences were also carried out without the presence of the AFLP4/VGI isolates that contained sequences that are partially identical to those found in AFLP6/VGII isolates. These AFLP4/VGI isolates were excluded because their presence might obscure the genotypic structure of C. gattii. The topology obtained is depicted in Fig. 3. The cluster that contained isolates with core AFLP4/VGI alleles and AFLP7/VGIV isolates formed a sister group to the AFLP5/VGIII isolates, and this AFLP4/VGI-AFLP7/VGIV-AFLP5/VGIII clade was strongly supported with bootstrap values ranging from 98% to 99% for NJ and from 94% to 95% for MP analyses. The AFLP6/VGII isolates clustered basal to the other C. gattii genotypic groups in both the mitochondrial and nuclear analyses with strong support.

3

Phylogenetic tree of Cryptococcus gattii obtained by analysis of (a) partial ATP6 sequences. AFLP4/VGI isolates that possessed mitochondrial sequences identical to sequences found in AFLP6/VGII isolates were excluded from the analysis. Presented is the most parsimonious tree (length 79, consistency index 0.861, retention index 0.978) computed with gaps treated as missing data. Data consisted of 611 characters of which 64 characters were informative of parsimony; (b) six concatenated nuclear regions (RPB1, RPB2, CNLAC1, TEF1α, IGS1 and internal transcribed spacer). Presented is one of the 60 most parsimonious trees (length 847, consistency index 0.902, retention index 0.961) computed with gaps treated as missing data. Data consisted of 3932 characters of which 459 characters were informative of parsimony. Bootstrap values (1000 replicates) are indicated for the main branches. T, type strain; R, molecular type reference strain (Meyer, 2003).

Discussion

Phylogenetic analyses of partial mitochondrial sequences of C. gattii showed that AFLP6/VGII isolates clustered basal to all other C. gattii genotypic groups. These results correspond to the topology that was found after the analysis of six concatenated nuclear regions (Bovers, 2008b; Fig. 3).

Surprisingly, two AFLP4/VGI isolates (AFLP4/VGI-M5) were found that possessed MtLrRNA and ATP6 sequences identical to those found in AFLP6/VGII isolates. In addition, three AFLP4/VGI isolates (AFLP4/VGI-M4) possessed a MtLrRNA sequence that differed by one nucleotide from the core AFLP4/VGI sequence, but the ATP6 sequence was identical to the sequence found in all AFLP6/VGII isolates. Furthermore, two AFLP4/VGI isolates (AFLP4/VGI-M3) possessed the core AFLP4/VGI MtLrRNA sequence, whereas the ATP6 sequence differed by one nucleotide from the sequence found in AFLP6/VGII isolates. In addition, six AFLP4/VGI isolates (AFLP4/VGI-M2) possessed the core AFLP4/VGI MtLrRNA sequence, but here the ATP6 sequence appeared to be a chimera between core AFLP4/VGI and AFLP6/VGII sequences.

Our results show that AFLP4/VGI isolates exist that possess mitochondrial genomes that consist completely of AFLP6/VGII sequences or that contain a combination of AFLP4/VGI and AFLP6/VGII sequences. It could be argued that the presence of AFLP6/VGII mitochondrial sequences in AFLP4/VGI isolates is a retained ancestral character state, because genotype AFLP6/VGII was found to be the ancestral lineage (Bovers, 2008b). However, in that case AFLP6/VGII mitochondrial sequences would be expected to occur in other genotypic groups of C. gattii as well, especially in AFLP7/VGIV as this genotypic group is most closely related to AFLP6/VGII (Bovers, 2008b). We have to keep in mind, however, that the number of available AFLP7/VGII isolates is small (i.e. only four strains are currently know to belong to this group); thus, it is possible that nucleotide variation in this lineage remains to be detected in the future when more isolates will become available. In the case that ancestral mitochondrial sequences remained preserved in the more derived lineages, viz., AFLP7/VGIV, AFLP5/VGIII and AFLP4/VGI, one would expect that the same or similar sequences would be present in the phylogenetically intermediate genotypes because the patterns of divergence of the various genotypic groups is well supported by a phylogenetic analysis of six nuclear genes (Bovers, 2008b). Moreover, one would expect that some polymorphisms would be present due to ongoing evolution since the initial divergence of the various genotypic groups of C. gattii. In contrast with these suppositions, we observed only AFLP4/VGI isolates that possessed ATP6 sequences identical to those found in AFLP6/VGII isolates. These sequences did not occur in any of the other, phylogenetically intermediate, genotypes, viz., AFLP7/VGIV and AFLP5/VGIII. This suggests that the presence of AFLP6/VGII mitochondrial sequences in AFLP4/VGI isolates is not a retained ancestral character state but, more likely, is a consequence of mitochondrial recombination. The likelihood of mitochondrial recombination is further supported by the presence of the MtLrRNA haplotype 2 allele of AFLP4/VGI in two isolates of AFLP7/VGIV. If this would be the result of retaining ancestral sequences, this process would contradict the well-supported phylogeny based on six nuclear genes in which genotype AFLP4/VGI is more derived than AFLP7/VGIII (Bovers, 2008b).

AFLP4/VGI and AFLP6/VGII isolates have been isolated from the same geographic regions (Sorrell, 1996; Kidd, 2003, 2005, 2007; Campbell, 2005a, b), which indicates that these genotypic groups occur in the same areas and may physically interact. Interestingly, AFLP4/VGI isolates that possessed mitochondrial genomes with (partial) AFLP6/VGII sequences were all isolated from countries near the Pacific Ocean, whereas isolates with core AFLP4/VGI sequences originated from Africa and Europe. Although the number of isolates included in our study is small, this may indicate that the atypical AFLP4/VGI isolates originated in the region of the Pacific Ocean.

If mitochondrial recombination has occurred, one has to assume a process of somatic fusion or mating to occur between cells of the recombining genotypes. To the best of our knowledge, somatic fusion without mating has never been reported in C. neoformans or in C. gattii. Mating usually occurs between isolates of opposite mating type (MATa and MATα) (Kwon-Chung, 1975, 1976a, b), but has also been reported between isolates of identical mating type (Lin, 2005). During mating, the nucleus of the MATα cell migrates into the conjugation tube and the recipient MATa cell generates a hypha (McClelland, 2004). Previous reports indicated that mitochondria in C. neoformans were uniparentally inherited from the MATa parent (Xu, 2000; Yan & Xu, 2003), probably because the MATa cell produces the hypha. However, recent studies have shown that in some cases mitochondrial inheritance from the MATα may occur as well (Yan & Xu, 2003; Toffaletti, 2004; Yan, 2004, 2007a, b) resulting in the presence of MATα parental mitochondria in the progeny. Recent studies on other fungal species have shown that strictly uniparental or biparental mitochondrial inheritance is rare (Yan & Xu, 2005), and it has therefore been suggested to treat mitochondrial inheritance as a quantitative rather than a qualitative trait (Birky, 1995). In cells of the mitochondrial genotype M5, which combine a nuclear genome of the genotype AFLP4/VGI with the presence of the mitochondrial alleles ATP6 and MtLrRNA from AFLP6/VGII, cells of both MATa (i.e. isolate E566) and MATα (i.e. isolate WM179) seem to have functioned as a donor for mitochondrial genes. Therefore, in these cases, the transfer of mitochondria does not seem to be regulated by the mating type. Mitochondrial inheritance following AFLP4/VGI × AFLP6/VGII mating may have resulted in a cell with both types of mitochondria. Subsequent mitotic divisions and mitochondrial selection could have resulted in isolates that possessed a complete AFLP6/VGII mitochondrial genome. In addition, recombination between the two types of mitochondria that were present may have resulted in mitochondrial genomes with both AFLP4/VGI and AFLP6/VGII sequences. It is important to note that recombination between mitochondria has been observed in other fungi under laboratory conditions, for example Agaricus bisporus (De la Bastide & Horgen, 2003), Agrocybe aegerita (Barroso & Labarère, 1997), Neurospora intermedia (Yang & Griffiths, 1992), Saccharomyces cerevisiae (Nakagawa, 1992), as well as in natural populations, for example A. aegerita (Barroso, 1995), Armillaria gallica (Saville, 1998), Candida albicans (Anderson, 2001; Jacobsen, 2008) and Neurospora crassa (Taylor, 1986). In addition, mitochondrial recombination has been observed in laboratory crossings of C. neoformans isolates (Toffaletti, 2004; Yan, 2007b).

Here we suggest the occurrence of mitochondrial recombination among environmental, veterinary and clinical isolates of C. gattii. Recombination between two mitochondrial genomes from different genotypes is only possible after cell–cell conjugation occurs. However, in that case one might expect the occurrence of hybrid nuclear genomes within the populations, and this is not the case with the loci that we have studied (Bovers, 2008b). Hybrid isolates between the two C. neoformans varieties and between C. neoformans and C. gattii have been described (Tanaka, 1999; Cogliati, 2000; Boekhout, 2001; Bovers, 2006, 2008c), as has DNA exchange between serotype A and serotype D isolates of C. neoformans (Kavanaugh, 2006). The resulting hybrid isolates are diploid or aneuploid (Tanaka, 1999; Cogliati, 2000; Lengeler, 2001; Bovers, 2006, 2008c) and possess alleles from both parents. However, all AFLP4/VGI isolates that had been studied using flow cytometry were found to be haploid (M. Bovers & H. Hoogveld, unpublished data) and nuclear AFLP6/VGII sequences have not been found in these AFLP4/VGI isolates (Bovers, 2008b), which indicates that the isolates, most likely, are not hybrids. Possibly, mating between different C. gattii genotypic groups does not result in the formation of hybrid nuclear genomes, but may lead to transfer of mitochondria and subsequent mitochondrial recombination. Moreover, the predominant uniparental inheritance of mitochondria as well as the small size of the mitochondrial genome increases the chance of incorporation of foreign alleles by drift and selection (Martinsen, 2001; Funk & Omland, 2003; Ballard & Whitlock, 2004). The predominant uniparental inheritance and the haploid nature of mitochondria both decrease the effective population size and thereby increase the chance of fixation by drift (Funk & Omland, 2003; Ballard & Whitlock, 2004). In addition, the small genome size decreases the probability of counter selection because negative gene interactions are less likely to occur (Martinsen, 2001; Ballard & Whitlock, 2004). These processes could explain why mitochondrial regions, but not nuclear regions, were incorporated in a different genetic background. Another possibility is that smaller regions of the AFLP6/VGII genome have been incorporated in parts of the nuclear genome of AFLP4/VGI isolates, but that these have escaped detection.

Mitochondria play a central role in energy production and in responses to stresses (Osiewacz & Kimpel, 1999; Ahkter, 2003). These processes play a key role in the fitness of an organism, and the mitochondrial genotype may therefore influence fitness as well. A different mitochondrial genotype may have a positive effect on fitness, as was observed in S. cerevisiae when mitochondria of wine yeasts were transferred to a laboratory yeast, resulting in an increased viability and increased ethanol and temperature tolerance (Jiménez & Benítez, 1988). However, the mitochondrial genotype does not always affect fitness, and no difference in virulence or high-temperature growth rate could be found for strains with mitochondria derived from both C. neoformans varieties (Toffaletti, 2004). In addition, the mitochondrial genotype may also have a negative effect on fitness. Several mitochondrial proteins are produced in the nucleus (Osiewacz & Kimpel, 1999; Burton, 2006) and these nuclear encoded proteins interact with mitochondrial encoded proteins, for example in the electron transport system. The presence of a mitochondrial genotype different from the nuclear genotype may therefore negatively affect fitness as both genomes have coevolved (Burton, 2006). Cryptococcus gattii isolates with different mitochondrial genotypes should be further studied to investigate whether the mitochondrial genotype affects their fitness.

Our results indicate that conjugation may have occurred between AFLP4/VGI and isolates of both AFLP6/VGII and AFLP7/VGIV leading to the generation of cells with an AFLP4/VGI nuclear genotype and a partial AFLP6/VGII or AFLP7/VGIV mitochondrial genotype. Our study is the first to report on mitochondrial inheritance in C. gattii and provides evidence for mitochondrial recombination occurring in nature. Furthermore, the high percentage of AFLP4/VGI isolates studied that possessed a recombined mitochondrial genome (65%) indicates that recombination occurs frequently in nature. Recently, recombination has been demonstrated to occur in natural populations of genotype AFLP4/VGI of C. gattii using AFLP analysis (Saul, 2008). Future research is needed to elucidate the genetic mechanism of the intergenotypic transfer of mitochondria in C. gattii and the effect this has on phenotypic properties. The occurrence of mitochondrial recombination between isolates of genotypes AFLP4/VGI and AFLP6/VGII may also be important, because strains of the latter are involved in the ongoing Vancouver Island outbreak of C. gattii (Kidd, 2004; MacDougall, 2007). Plasmogamy between cells belonging to different genotypes may allow the exchange and introgression of genetic elements in either genetic background, as has also been observed between isolates of serotype A and serotype D of C. neoformans (Kavanaugh, 2006), and such processes may contribute to the formation of C. gattii strains that differ in virulence. It is likely that such genetic transfer events may have an effect on the virulence properties, and this alternative hypothesis concurs with the same sex mating hypothesis (Fraser, 2005; Lin, 2005) that has been proposed to explain the origin of hypervirulence of isolates involved in the ongoing Vancouver Island outbreak of C. gattii.

Acknowledgements

Isolates were kindly donated by the following researchers: J.W. Kronstad, K.J. Kwon-Chung, M.S. Lazera, D. Swinne, W. Meyer and J.M. Torres-Rodríguez. We would like to thank Cañan Bayrakdar and Hans Hoogveld for technical assistance. Work of M.B. was supported by the ‘Odo van Vloten fonds’. E.E.K. and F.H. were funded by the Strategy Fund of the Royal Netherlands Academy of Arts and Sciences (RNAAS-KNAW).

Authors' contributions

M.B. and F.H. equally contributed to the work, M.B. and F.H. performed the work, T.B. designed the research, and M.B., F.H., E.E.K. and T.B. wrote the paper.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Alignment of Cryptococcus gattii partial ATP6 sequences.

Fig. S2. Alignment of Cryptococcus gattii partial MtLrRNA sequences.

Fig. S3. Alignment of Cryptococcus gattii partial ATP6 sequences for haplotypes 1, 2 and 4.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Author notes

Present address: Eiko E. Kuramae, Netherlands Institute of Ecology, Centre for Terrestrial Ecology, 6666 ZG Heteren, The Netherlands.

Editor: Cletus Kurtzman