Susceptibility of Intact Germinating Arabidopsis thaliana to Human Fungal Pathogens Cryptococcus neoformans and C. gattii

Abstract

The fungus Cryptococcus contributes a large global burden of infectious death in both HIV-infected and healthy individuals. As Cryptococcus is an opportunistic pathogen, much of the evolutionary pressure shaping virulence occurs in environments in contact with plants and soil. The present studies investigated inoculation of intact seeds of the common weed Arabidopsis thaliana with fungal cells over a 21-day period. C. gattii was the more virulent plant pathogen, resulting in disrupted germination as well as increased stem lodging, fungal burden, and plant tissue colocalization. C. neoformans was a less virulent plant pathogen but exhibited prolonged tissue residence within the cuticle and vascular spaces. Arabidopsis mutants of the PRN1 gene, which is involved in abiotic and biotic signaling affecting phenylalanine-derived flavonoids, showed altered susceptibility to cryptoccocal infections, suggesting roles for this pathway in cryptococcal defense. The fungal virulence factor laccase was also implicated in plant pathogenesis, as a cryptococcal lac1Δ strain was less virulent than wild-type fungi and was unable to colonize seedlings. In conclusion, these studies expand knowledge concerning the ecological niche of Cryptococcus by demonstrating the pathogenic capacity of the anamorphic form of cryptococcal cells against healthy seedlings under physiologically relevant conditions. In addition, an important role of laccase in plant as well as human virulence may suggest mechanisms for laccase retention and optimization during evolution of this fungal pathogen.

INTRODUCTION

The basidiomycetous yeast Cryptococcus neoformans has emerged as one of the major causative agents of meningoencephalitis in immunocompromised hosts, such as persons with AIDS, organ transplant recipients, and patients receiving high doses of corticosteroid treatment. As rates of infection have diminished in developed countries, attention is increasingly being focused on high rates of cryptococcosis in the developing countries of Africa and Asia, where cryptococcosis has been estimated to account for as much as 17% of AIDS-related deaths—a disease burden surpassing that of tuberculosis in some regions of Africa (1, 2). Systemic infections can also occur in immunocompetent individuals; the fungus has been particularly problematic in a recently described outbreak of a second species of Cryptococcus, C. gattii, which has caused disease in the Pacific Northwest, centered on Vancouver Island in British Columbia, Canada (3, 4). In addition, C. gattii displays a more severe clinical course and is associated with increased growth in macrophages (5).

Since Cryptococcus acts as an opportunistic pathogen of humans, much of its evolutionary pressure has been exerted within ecological niches where retention and optimization of virulence factors is shaped by relationships within these environments. One example is the evolutionary pressure exerted by phagocytic free-living amoebae that inhabit fungal landscapes and may have shaped the evolution of virulence factors such as laccase (6). Accurate identification of the environmental ecology may thus help to further understanding and anticipation of pressures leading to virulence changes as well as lead to the identification of important infectious reservoirs for the pathogen. Since the first discovery of Cryptococcus in peach juice by Sanfelice in 1894 (7) and Staib's further characterization of growth on autoclaved plant material in the early 1970s (8), the fungus has been associated with an ability to grow on nonvital plant matter. A major reported environmental niche of the fungus is soil contaminated by pigeon guano (C. neoformans) or eucalyptus trees and decaying wood (C. gattii) (9). Historically, C. gattii had been reported to share a specific ecological niche with Eucalyptus camaldulensis and E. tereticomis, which produces a nutrient-rich leaf litter and decomposing branch fall (10, 11). However, examination of the outbreak on Vancouver Island identified high concentrations of C. gattii in soils without these specific plant species (12). Interestingly, environmental niches of the tropics where C. gattii has been isolated have shared aspects with parts of Vancouver Island in the Pacific Northwest. For example, both have high humidity and low sunlight penetrance from the overhanging canopy and are generally warm (above freezing temperatures) (13, 14). Recently, Cryptococcus species have been shown to invade wounded plants as an opportunistic infection through abrasions in mature plant surfaces (15) or after mating with a fungus of an opposite mating type (16). However, invasion of intact plants has not been described for the anamorphic form of the fungus that is the predominant form in the environment (17).

Thus, to further define the role of Cryptococcus within the plant environment and to investigate possible links between cryptococcal ecological fitness and human virulence, we tested for a role of the fungus, and of laccase, a principal cryptococcal virulence factor, in infection and colonization of the model plant Arabidopsis thaliana, found in all temperate regions of the world. Since Cryptococcus spp. are found predominantly in soils and rotting vegetation (18, 19), we propose that a model relevant to pathogenesis within the context of the leaf litter/forest floor would be to inoculate seeds at the germination period under conditions of dim light as well as bright light. In addition, since plant flavonoids form important antimicrobial defenses for plants, we tested fungal virulence against a mutant with a mutation in the AtPirin1 (PRN1) gene (the A. thaliana Atprn1 mutant accumulates increased amounts of these compounds; D. A. Orozco-Nunnelly, D. Muhammad, R. Mezzich, B.-S. Lee, L. Jayathilaka, L. S. Kaufman, K. M. Warpeha, submitted for publication) and the role of laccase in facilitating fungal growth in the presence of selected flavonoids in vitro.

MATERIALS AND METHODS

Fungal strains, plasmids, and media.

Cryptococcus neoformans ATCC 208821 (H99) was a generous gift of J. Perfect, and the R265 strain of C. gattii was a generous gift of J. Heitman. The lac1Δ strain of C. neoformans was previously described (20). All strains were grown on YPD medium (1% yeast extract, 2% Bacto peptone, and 2% dextrose). Solid media contained 2% Bacto agar. Spot plates contained 20 g/liter agar, YNB (yeast nitrogen base), and 0.2% glucose plus the indicated chemical at 500 nM.

Seed accessions and plant growth protocol.

Matched seed lots of wild-type (wt) Columbia (col) Arabidopsis thaliana and seedling accessions, or lines, carrying T-DNA insertions within coding regions of Pirin1 (21) (SALK_006939) (21, 22) were obtained from the Arabidopsis Biological Resource Center (Columbus, OH) (23). Gene sequence accession numbers were obtained from GenBank (http://www.ncbi.nlm.nih.gov) and SIGnAL (http://signal.salk.edu). Seeds were surface sterilized and planted in 0.8% top agarose stabilized with 0.5× Murashige and Skoog (MS)/2-(N-morpholino)ethanesulfonic acid (MES) (pH 5.8) minimal media for plant growth (50 to 120 seeds per disk, depending on the experiment) on 0.5× MS/MES (pH 5.8) plates (50 ml of media/agarose per phytatray). Soil experiments used the same phytatrays with 1-cm-deep autoclaved dampened 0.5× MS Metromix (Scott's) soil on top, and seeds were sown in 0.5× MS broth (1 ml per disk of seeds). Sown seeds were subjected to a cold treatment (4°C) for 48 h to stimulate synchronized germination (24), but no light vernalization/treatment was performed. Sets of sown seeds after the 48-h cold period (vernalization) were grown in complete darkness for 48 h at 20°C and then incubated for the indicated periods in either dim light (10⁻¹ μmol m⁻² s⁻¹ for 16 h cycled with complete darkness for 8 h during each 24-h period) or in moderate summer daylight (10² μmol m⁻² s⁻¹ for 16 h followed by complete darkness for 8 h). All seedlings were maintained at 20°C for the indicated growth periods with sufficient moisture to avoid water stress. During dark incubation periods, sowing and inoculation of germinating seeds were performed under a dim green safelight.

Inoculation with Cryptococcus. (i) Seed infection model.

Seeds were inoculated with Cryptococcus culture of the indicated strains at a time after removal from the 4°C treatment prior to placement in the 20°C incubator. Cryptococcus strains were grown on YPD agar, 1 colony was selected, and yeast cells were dispersed in 0.5× MS/MES media (pH 5.8) and then read on a spectrophotometer for cell density. Viability of inocula was assessed on YPD agar and was found to exceed 90%. Culture was diluted to an optimal density (OD) at 595 nm of 0.2. The culture at 0.2 OD was diluted 1:10 (vol/vol), and 200 μl was applied evenly to the phytatray (9.5 by 8 cm) on top of the sown seeds (which were in a disk of ∼1-mm thickness on top of the plate) so that there was an even layer of Cryptococcus culture at a low concentration; on soil, the application was performed in the same volume with a perfume mister. After liquid culture was applied, the plates were maintained under growth protocol conditions and observed over time. Seedlings infected with green fluorescent protein (GFP)-expressing strains of C. neoformans (wt and lac1Δ) were photographed on a Zeiss Stereo Discovery V.8 microscope with a ×1 or ×8 zoom lens as indicated using Axiovision. Sets of 30 seedlings were examined in triplicate.

(ii) Leaf-wounding model.

An adaptation of the method of Tucker was used (25). Briefly, seeds of Arabidopsis genotypes were plated on phytatrays on 0.5× MS agarose. Both plant genotypes were permitted to grow in brighter light for 14 days. At 14 days, silicon carbide powder was applied to the first set of leaves at the leaf edge to abrade the cuticle, which was wiped by a sterile cotton-tipped swab dipped in 0.5× MS. Experimental sets of seedlings were inoculated with 200 μl of a 1:10 dilution of a culture suspension at 0.2 OD of the indicated fungal strain yeast cells or 0.5× MS (control) and seedlings maintained in light for 21 more days under the same conditions. Leaves were excised, mounted live (nonfixed) with a drop of sterile water, and examined by deconvolution microscopy as described for the seed infection model. Leaves were sectioned with apotome at ×200 and photographed using Zeiss Axiovision software v. 4.8.

Construction of a green fluorescent protein-expressing C. neoformans wt and lac1Δ strain.

The cryptococcal shuttle vector pORA-KUT was used to express a fusion between the NAT protein and a synthetic green fluorescent protein (pCneo-GFP) utilizing C. neoformans codon usage (26). First, pORA-KUT containing the sequence of the EF-1α terminator was digested with BglII and EcoRI, and a PCR-amplified fragment of genomic DNA from H99 (obtained using primers GPD-1A-RI [5′-GCC GGA ATT CAT TGT ATT TAT GCA AGT ATA CTC] and GPD-865S-BglII [5′-GGA GAA GAT CTA GCA GAC AGT TGG G]) was digested with BglII and EcoRI inserted into compatible sites to produce pORA-KUG. The plasmid was recovered, verified by sequencing, and digested with PstI, and a PCR-amplified fragment of Cneo-GFP DNA (obtained using primers GFPmyc-PstIL [5′-TTA TAC TGC AGA TGT CCA AGG GTG AGG AGC TCT TCA CCG GT] and GFPmyc-PstIBamHIR [5′-TTA TAC TGC AGG GAT CCG AGG TCC TCC TCG GAG ATG AGC TTC TGC TC]) was digested with PstI and BamHI and inserted into compatible sites. The plasmid was recovered, the sequence was verified, the plasmid was digested with EcoRI, and a PCR-amplified fragment of the NAT gene (obtained using primers Nat-ORF-RI-s [5′-TTT CCG GAA TTC AGC GGC CGC CAC TCT TGA CGA C] and Nat-ORF-RI-a [5′-ACC GGA ATT CTT GGG GCA GGG CAT GCT CAT GTA G]) was digested and ligated into compatible sites to produce pORA-KUGNG. The plasmids were recovered, the sequences were verified, and the plasmids were linearized with SceI and transformed into C. neoformans H99 ura5 and lac1Δ ura5 cells by electroporation using standard methods (27). Stable transformants were selected by retained expression after serial plating on nonselective media.

Observation, experimental assessment, and histology.

Growth characteristics were observed 7, 14, 21, and 28 days after vernalization. At the indicated time points, sets of seedlings were observed under a dissecting microscope for morphological changes, including successful seed germination, growth, and, for older seedlings, stem lodging (defined as seedlings that had fallen over due to significant stress/damage which typically is not recoverable, indicating the death of the plant) and then sacrificed for fungal burdens. Seedlings were assessed for fungal burdens by harvesting aerial portions (50 to 100 seedlings) followed by homogenization in 1 ml of MS/MES media (pH 5.8) at 4°C and inoculation of aliquots onto YPD agar followed by maintenance at 20°C for up to 7 days, after which colony formation units (CFU) were assessed. The control (no plants) was achieved at the time of sowing by placing a Whatman 1 sterile filter (1 cm by 1 cm) on the soil or agar plate after inoculation and then applying the mixture to a new plate to measure CFU at the indicated time. For histology, harvested seedlings were washed gently, fixed in 2% formaldehyde, sectioned, and stained using hematoxylin and eosin or Gomori methenamine silver staining and standard methods (28).

Statistics.

The statistical significance of all seedling survival time data (Fig. 1) in comparison to the uninfected-control data was assessed by a log rank test (Mantel-Cox). Determinations of the rates of survival of seedlings at 21 days (see Fig. 4) were performed by a chi-square test using a contingency table, and fungal burden data (Fig. 2) were compared by t test (using Welch's modification). Differences in rates of survival of groups at 21 days (comparing each group to its untreated control) were determined using Fisher's exact test. Statistical analysis was conducted using GraphPad Prism software, version 4.03.

Fig 1.

Coincubation of seeds of Arabidopsis thaliana with fungal cells of C. gattii and C. neoformans results in seedling death. Seeds of either wt A. thaliana (wtAt) or prn1 mutants (At prn1) were incubated with wt strain C. gattii or C. neoformans as described in Materials and Methods and grown for the first 48 h in the dark and then moved to conditions of either dim light (A, B, E, and F) or bright light (C, D, G, and H). Seeds that never germinated or died in germination as determined by microscopy were scored as not surviving at day 1 (germination), and seedlings were monitored for stem lodging at day 14 and day 21. P values for comparisons between the indicated groups and uninfected seedlings are indicated (***, P < 0.001; **, P < 0.01; *, P < 0.05).

Fig 4.

Role of laccase in plant pathogenicity of C. neoformans. (A) Seedlings were inoculated with the indicated strains on agar as described for Fig. 1 and observed at 21 days for morphological changes and stem lodging/death. P values for comparisons between the indicated groups and uninfected seedlings are indicated (***, P < 0.001; **, P < 0.01; *, P < 0.05). (B) Leaves were photographed from above (left panels) or from the side (middle panel) and roots were photographed from the side (right panel) at 21 days postinfection. White arrows point to yellowed leaves and plant structures. Bars, 2 mm. (C) The indicated strains were spot plated at sequential 1:10 dilutions on agar containing yeast nitrogen base, 0.2% glucose, and a 500 nM concentration (each) of the flavonoids quercetin, transcinnamic acid, prephenic acid, sinapic acid, abscisic acid, and kaempferol (+ALL) or a 500 nM concentration of the indicated flavonoid.

Fig 2.

Coincubation of seeds of Arabidopsis thaliana with fungal cells of Cryptococcus results in significant cryptococcal fungal burdens. The indicated seeds were coincubated with indicated fungal strains on either agar or soil in dim light as described for Fig. 1. At 14 days, intact seedlings were recovered, separated from root material, washed extensively, and homogenized and fungal CFU burdens determined per mg of plant tissue.

RESULTS

Coincubation of Cryptococcus gattii with seeds from Arabidopsis thaliana reduces germination rates and leads to stem lodging with greater virulence under dim-light conditions.

To simulate an environment in which plant and fungal communities coexist, we sowed seeds of Arabidopsis thaliana, either within an experimental agar surface or in humidified soil, and sowed a suspension of fungal cells of Cryptococcus in the plant medium at the same time and observed effects on germination and growth of seedlings at 7, 14, and 21 days. Seedlings were incubated under conditions of either dim light, to simulate conditions under an overhead canopy, or brighter light, to simulate sunlight in open spaces. Seedling viability was measured by loss of stem integrity (stem lodging) in crops as well as in this species (29). Seeds that never germinated or died in germination as determined by microscopy were scored as not surviving at day 1 (germination; see black arrow in Fig. 1A), and seedlings were monitored for stem lodging at day 14 and day 21. The C. gattii strain used was obtained from the recent Vancouver Island outbreak (4). As shown in Fig. 1A, inoculation of C. gattii onto agar containing wt plant seeds under dim-light conditions resulted in significant plant virulence, with reductions in germination (fungus-treated group, 21 failures; untreated group, 3 failures; P < 0.001) as well as poor survival over the 21-day period (fungus-treated group, 74.2% mortality; untreated group, 5.8% mortality; P < 0.0001). C. gattii also infected intact seedlings grown in soil, with almost half of the seedlings killed over a 21-day period in dim light (fungus-treated group, 44% mortality; untreated group, 1% mortality; P < 0.0001) (Fig. 1B). Seedlings were less susceptible when exposed to an equivalent inoculum of C. gattii in bright light (light, 30%; dark, 74.2%; P < 0.001), and the effect was also present for seedlings grown in soil (light, 19%; dark, 44%; P < 0.001) (Fig. 1C and D). These data suggest that anamorphic C. gattii has the potential to be a significant plant pathogen of intact germinating seedlings.

Coincubation of Cryptococcus neoformans with Arabidopsis thaliana seeds leads to less but significant stem lodging during early growth, and virulence is independent of light conditions.

Because of the global distribution of C. neoformans, the second fungal strain studied was C. neoformans serotype A strain H99, first isolated from a patient with Hodgkin's disease presenting with meningoencephalitis (30). Inoculation of viable wild-type (wt) plant seeds in agar with this strain also resulted in reductions in seedling viability versus untreated seedlings over the 21-day period in bright light (fungus-treated group, 4.2% mortality; untreated group, 0% mortality; P < 0.05) (Fig. 1G) and a trend toward reduced viability in dim light (fungus-treated group, 8.4% mortality; untreated group, 3.8% mortality) (Fig. 1E), although the result in dim light did not reach statistical significance. Seedlings inoculated in soil showed susceptibility to the fungus in dim light (fungus-treated group, 9% mortality; untreated group, 2% mortality; P < 0.001) (Fig. 1F) that was equivalent to that seen with seedlings incubated in bright light (fungus-treated group, 9% mortality; untreated group, 0% mortality; P < 0.05) (Fig. 1H). We also tested fungal inoculation on a strain with a mutation of the AtPirin1 gene (Atprn1). The A. thaliana Atprn1 mutant is defective in Pirin1 (PRN1), a member of the cupin superfamily which is of interest because it is involved in germination (31) and abiotic and biotic signaling (21, 22) and was originally identified in plants (tomato) as being involved in cell stress and apoptosis (32). The AtPRN1 protein also possesses quercetinase activity, where the A. thaliana Atprn1 mutant accumulates antimicrobial phenylpropanoid flavonoids derived from phenylalanine, notably quercetin (Orozco-Nunnelly et al., submitted). Interestingly, the A. thaliana Atprn1 mutant was markedly more susceptible in experimental agar to fungal inoculation than the wt A. thaliana seedlings under conditions of both dim light (fungus-treated group, 24.2% mortality; untreated group, 8.4% mortality; P < 0.001) and bright light (fungus-treated group, 25% mortality; untreated group, 2.5% mortality; P < 0.001), suggesting a role for AtPRN1 in protection of seedlings from C. neoformans. This increased susceptibility of the A. thaliana Atprn1 mutant was reproduced in soil with approximately the same rate of time-dependent killing (Fig. 1F and D). Interestingly, the A. thaliana Atprn1 mutant was more resistant to the effects of C. gattii inoculation than the wt seedlings (fungus-treated group [Atprn1], 14.8% mortality; A. thaliana wt, 74.2% mortality; P < 0.001) (Fig. 1A), suggesting differing relationships of the two cryptococcal species with their plant host. In summary, these data suggest the capacity for C. neoformans to infect intact seedlings and a role for plant PRN1-dependent modulating pathways in susceptibility to cryptococcal plant infections.

Seedling fungal burden is associated with fungal virulence in Arabidopsis thaliana seedlings.

Next, we assessed whether rates of stem lodging were associated with seedling fungal burdens. At 14 days after inoculation, seedlings were harvested and analyzed for tissue fungal burden by culture after careful washing and subsequent tissue homogenization. As a control, an equivalent filter square obtained from the surface of inoculated agar or soil was washed and sampled after equivalent fungal inoculation in the absence of seeds (no plants). As shown in Fig. 2, after infection of wt plant seeds, C. gattii showed increased fungal colony counts compared to the C. neoformans strain (1,152 versus 75 CFU/mg plant tissue; P < 0.001), consistent with its greater pathogenicity toward plant seedlings as exhibited in Fig. 1. (The weight of seedlings in these experiments at 14 days was approximately 0.7 mg/plant; thus, infection with C. gattii showed a calculated fungal burden of approximately 800 CFU/plant.) Colony counts of the C. gattii strain were reduced in the A. thaliana Atprn1 seedlings compared to wt seedlings (199 and 79.5 CFU/mg plant tissue; P < 0.001 for both), correlating with the reductions in killing observed for these fungal strain and seedling combinations. In contrast, for C. neoformans infections, the A. thaliana Atprn1 mutant seedlings showed increased fungal burden compared to wt seedlings (P < 0.001), corresponding to increases in virulence toward the mutant plant seedlings as exhibited in Fig. 1. Increased fungal burden was also present in seedlings grown in soil, with an increase in both C. neoformans and C. gattii fungal CFU (744 and 287 CFU/mg for the plant tissue and 158 and 103 CFU/mg for the no-plant control, respectively; P < 0.001). Again, the A. thaliana Atprn1 mutant displayed increased fungal burdens for C. neoformans, but not C. gattii, commensurate with susceptibilities of this plant mutant to these fungal strains. This implies a role for plant-associated fungal proliferation in plant pathogenesis by the two different species of Cryptococcus. In addition, increases in fungal burdens over those seen with soil (no plants) suggest that environmental fungal burdens are increased by successful cryptococcal seedling infections over those seen with soil residence alone.

Tissue invasion of seedlings by C. neoformans and C. gattii after coinoculation of fungal cells and seeds of Arabidopsis thaliana.

Since killing of seedlings within the 21-day time frame could have been due to indirect effects of fungal products, we undertook a microscopic study to determine whether fungal invasion was present after inoculation. Seedlings were gently washed 14 days after inoculation, fixed, embedded, and sectioned. While 14-day seedling invasion could be observed by microscopy in wet preparations (data not shown), sectioning was performed to minimize the likelihood that tissue colocalization was the result of superimposition rather than fungal residence within plant tissue. An early time point was chosen to evaluate early invasive events. As shown in Fig. 3A, after infections by both species, numerous Cryptococcus yeast cells (see arrows) were observed within plant tissue as revealed by either Gomori methamine silver staining (GMS; left panels) or hematoxylin-eosin staining (H & E; right panels). The presence of attached fungal daughter cells, exhibited best in the inset showing the C. gattii infection of wt A. thaliana, suggested that fungi were actively growing during infection. A few yeast cells could be detected outside plant tissue, which could represent colonization or a sectioning artifact. Interestingly, the combinations of strains that yielded low rates of plant killing and low fungal burdens (C. neoformans/A. thaliana wt or C. gattii/A. thaliana Atprn1) tended to have infections localized in adjacence to the leaf primordia, which are zones of meristematic cells with no observable plant response. In these germinating/young seedlings, the protective plant cuticle composed of long-chain fatty acids and waxes is not yet fully formed but instead represents a procuticle that presents less of a barrier to infection (29, 33, 34). In contrast, combinations that yielded high rates of plant killing and high fungal burdens (C. neoformans/A. thaliana Atprn1 or C. gattii/A. thaliana wt) exhibited widespread invasion of stems in addition to primordial areas, suggesting either a nonselective route of invasion or widespread dissemination after infection.

Fig 3.

Coincubation of Arabidopsis thaliana with C. neoformans or C. gattii results in fungal invasion of plant tissue. (A) The indicated seeds and fungal cells were coincubated in dim light as described for Fig. 1. At 14 days, seedlings were harvested, fixed, and subjected to tissue embedding and sectioning, followed by staining with either Gomori silver (left panels) or hematoxylin-eosin (right panels). Arrows show fungal cells. (B) The indicated strains were spot plated at sequential 1:10 dilutions on agar containing yeast nitrogen base, 0.2% glucose, and a 500 nM concentration (each) of the flavonoids quercetin, transcinnamic acid, prephenic acid, sinapic acid, abscisic acid, and kaempferol (+ALL) or a 500 nM concentration of the indicated flavonoid.

During cryptococcal invasion of plants, the fungus would be expected to encounter intracellular plant defenses. Since flavonoids represent an important form of antifungal defense for plants (35, 36), we assayed for potential growth inhibition of C. gattii and C. neoformans against a combination of phenylalanine-derived flavonoids (31), including quercetin, which accumulates in the A. thaliana Atprn1 mutant to higher levels than in the wt (Orozco-Nunnelly et al., submitted), as well as transcinnamic acid, sinapic acid, abscisic acid, kaempferol, and prephenic acid, a phenylalanine precursor, and abscisic acid, a plant stress hormone that stimulates flavonoids (Fig. 3B), at 500 nM concentrations typical of plant extracts (37). Interestingly, we found both fungal species to be resistant to these compounds either in combination (+All) or singly (the latter demonstrated by the quercetin and transcinnamic acid data in Fig. 3B), consistent with an ability to survive in plant tissue.

Laccase mutants of C. neoformans show attenuated virulence toward Arabidopsis seedlings.

To investigate possible roles for the mammalian virulence factor laccase as a plant virulence factor against Arabidopsis seedlings, we compared plant killing and fungal burdens of a C. neoformans lac1Δ mutant strain to those of its congenic wt strain. Seedlings were grouped at a higher density to simulate germination from a point location, which increases susceptibility to plant pathogens (38). As shown in Fig. 4A, inoculation with wt C. neoformans resulted in increased death of seedlings grown in agar at 21 days (98 survived in the untreated group versus 84 in the fungus-treated group; P < 0.01; n = 100 seeds), with less killing by the C. neoformans lac1Δ strain (97 survived in the C. neoformans lac1Δ-treated group versus 100 in the wt fungus-treated group; P < 0.01; n = 100 seeds). Seedlings grown in soil also showed susceptibility to fungal infection, although reductions in killing were not evident in the C. neoformans lac1Δ strain. The A. thaliana Atprn1 mutant was again more susceptible than the A. thaliana wt to the C. neoformans wt strain, and fungal virulence was reduced after LAC1 deletion. Examination of the leaf morphology of the A. thaliana wt or Atprn1 mutant seedlings that survived after infection in soil (Fig. 4B) demonstrated yellowing of leaves in seedlings inoculated with the wt fungal strains but not in those inoculated with the C. neoformans lac1Δ strain, suggesting that the fungal disease was more widespread that was evident by failure to germinate or plant lodging alone.

Because of the role of flavonoids in plant defenses, we again assayed for a potential role for laccase in protection from quercetin, transcinnamic acid, prephenic acid, sinapic acid, abscisic acid, and kaempferol during growth in YNB–0.2% glucose over 2 days at 30°C (Fig. 4C). Interestingly, we found increased sensitivity of the C. neoformans lac1Δ mutant to a combination of these compounds. Individual inhibition assays demonstrated laccase-dependent growth inhibition by quercetin and transcinnamic acid, but not the other 4 compounds, suggesting a role for laccase as a fungal protectant against these plant flavonoids.

Fungal persistence of C. neoformans within surviving plants.

Sublethal infection of plants suggests that C. neoformans may act to colonize intact plants in the anamorphic state that could provide a stable ecological niche for the fungus. To test this hypothesis, a codon-optimized green fluorescent protein was expressed under the control of a constitutive promoter (GPD [glycerol-3-phosphate dehydrogenase]) in either the wt strain (GFP-C. neo) or a lac1Δ strain (GFP-lac1Δ) of C. neoformans and used to infect germinating seeds. As shown in Fig. 5 (left panels), after 7 days, fluorescent fungal cells could be observed in healthy A. thaliana wt and Atprn1 mutant seedlings underneath the cuticle layer (black arrows). The brightness of the fungal fluorescence was evident in the reflective cuticle layer external to the fungus (white arrows). However, the bright-field microscopy could not distinguish the cuticle layers, so a less likely possibility is that the fungus was contained within the cuticle, as the outer transparent layer is visible. Persistence of fungal cells within living plants, predominantly within the vascular layers of the mature plants, was again evident at up to 31 days after infection in both yellowed and morphologically normal plants (Fig. 5, right panels). In contrast, we were not able to observe persistence of the C. neoformans GFP-lac1Δ strain at either time point (data not shown), suggesting a role for laccase in protection of the fungus against plant defenses during colonization.

Fig 5.

Infection of seedlings of Arabidopsis thaliana with GFP-expressing strains of C. neoformans demonstrates internal colonization of infected seedlings. Seeds were infected with Cryptococcus neo-GFP using the seed infection model described in Materials and Methods and photographed at the indicated times using deconvolution microscopy. Panels shown are merged fluorescent/bright-field channels of seedling regions. (A and B) Seedling stem approximately 1 to 2 mm inferior to the apical meristem at 7 days (7d) after infection. White arrows indicate the clear outer cuticle and black arrows the cryptospheres. (C and D) Images of areas 3 to 5 mm within stem tissue, located approximately 1 to 2 mm from the apical meristem. Bars: 20 μm (A); 6 μm (B); 20 μm (C); 25 μm (D).

Role of laccase in a wounding model of cryptococcal plant pathogenicity.

To assess a role for laccase in a second model of plant pathogenesis, we inoculated fungal strains after abrading leaf edges of young plants 14 days after germination using a wounding model (25). As shown in Fig. 6A, inoculation of wounded leaves of plants with wt fungal cells caused significant death/lodging of plants whereas little virulence was exhibited in this model by the C. neoformans lac1Δ strain. In addition, using the GFP-expressing fungal strains to quantitate the fungal load in individual leaves, extensive fungal burden was observed after infection of either A. thaliana wt plants or Atprn1 mutants. In contrast, few C. neoformans GFP-lac1Δ strains were observed in leaves, consistent with its reduced ability to cause plant virulence. Fluorescence microscopy also showed invasion of leaf tissue from the outer edge of the leaf (upper right corner) toward the inner leaf tissue (lower right corner). These data thus confirm a role for laccase as a significant plant virulence factor in this second model of plant pathogenesis.

Fig 6.

C. neoformans demonstrates successful infection of wt and Atprn1 mutants of Arabidopsis thaliana in a wounding model. GFP-expressing strains of C. neoformans were inoculated on leaves of the indicated seedlings 14 days after germination as described in Materials and Methods, and, at 21 days, leaves were excised and fungal burdens determined by fluorescence microscopy scoring of 30 leaves and plants scored for death or lodging at 21 days (A) and observed by deconvolution fluorescence microscopy at the 7th layer into the leaf parallel to the adaxial surface (B). GFP and merge with bright field are shown. Bar, 20 μm. ***, P < 0.001.

DISCUSSION

Cryptococcus species have long been known to coinhabit the plant ecosystem since the fungus was first isolated from peach juice by Sanfelice in 1894 (7). More recently, the fungus has been shown to exhibit plant pathogenesis under restricted conditions. For example, plant virulence was exhibited during mating with the assistance of plant-derived products such as myo-inositol (16). Plant wounding also results in colonization at the localized wound site and is dependent on extracellular fibrils (15). Another related basidiomycete, Ustilago maydis, was found to infect Arabidopsis seedlings, but only in the filamentous heterokaryon state (39). However, infection of intact plants by the anamorphic form of the fungus, which is the predominant form in the environment, has not been reported. Here we report successful infection of intact germinating seedlings with the inoculation of Cryptococcus yeast cells dispersed either in agar planting media or soil under conditions simulating low and moderate daylight levels. Yeast forms of the organism were also found to colonize seedlings and persist in the mature plant by survival adjacent to the plant vasculature. Successful infections by the anamorphic form of the fungus thus create the potential for a significant expansion of the ecological niche and suggest a role as a plant endophyte that both colonizes plants and causes disease.

Pathogenic mechanisms of endophytic fungi are thought to involve a balance of antagonisms between the pathogen and the plant host (40). In the present case, a shift toward successful pathogenesis and internal persistence was most likely due to the immature state of the plant's immune system during germination. In contrast, for example, Springer et al. (15), who found infection only after wounding, used Arabidopsis leaves of mature seedlings (4 to 6 leaves from each of several 3-week-old, light-grown plants) grown in a growth chamber under full-light conditions. Typically, older, light-grown plants with mature outer cuticles are resistant to fungal invasion, which then requires specialized invasive structures for successful pathogenesis, such as the appressorium of the rice blast fungus Magnaporthe grisea (41). In very dim lighting, however, such as that found in temperate or tropical rainforests, where only 2% sunlight reaches the forest floor, development of the photosynthetic apparatus and protective cuticles, suberin-coated roots, and extracuticular structures that prevent pathogenic attachment and penetration into cells can be delayed (42). Indeed, this effect of light exposure in Arabidopsis giving protection against other plant pathogens, such as the turnip crinkle virus, has been previously reported (43). In addition, many of the infective foci observed during intact seeding invasion were within the apical meristem (Fig. 3A), a region of rapidly dividing cells that is a vulnerable target for many of the flower-infecting fungi that cause systemic infection through the apical meristem (44). Thus, while the exact molecular mechanisms of cryptococcal plant invasion require more study, infection of seeds during germination appears to be an attractive infective route for both the C. neoformans and C. gattii pathogens that typically inhabit plant-rich soils (19), sometimes at high concentrations (12). While these modeling studies require confirmation in the field, increased fungal burden in the environment from resultant seedling infections, as well as persistence within intact seedlings, suggests that plant pathogenesis by anamorphic forms of the fungus could represent an important ecological niche for both fungal spread to mammalian hosts and evolutionary selection pressure. Indeed, it has been hypothesized that the diverse environments and requirements of endophytes drive the phenotypic plasticity that has been implicated as a “motor of evolution” (40).

While C. neoformans was able to infect germinating seeds and colonize young seedlings, C. gattii was more successful as a lethal plant pathogen under these conditions, and that success was facilitated by low-light conditions. This suggests relative differences in the plant pathogenicity of the two species and perhaps different infective strategies. For example, light-independent infection by C. neoformans might suggest an adaptation to plant defenses induced during light exposure and could suggest infection of seedlings heavily exposed to light. It is notable that, due to the integrity of its cell wall matrix, Cryptococcus neoformans is particularly resistant to photodynamic killing (45), suggesting success in high-light environments. Reduced plant killing by this species, compared to C. gattii, could also enable C. neoformans to be a more effective colonizer of surviving plants, enabling the fungus to avoid environmental pathogens such as free-living amoebae (6, 46). In addition, differing susceptibilities of the A. thaliana Atprn1 mutant to the two fungal strains also suggest differences in the infective strategies of the two strains. Previous studies have suggested that A. thaliana Atprn1 mutants defective in quercetinase accumulate flavonoids derived from phenylalanine, particularly quercetin (22) (Orozco-Nunnelly et al., submitted). Flavonoids are broad-spectrum antimicrobials constituting part of the antifungal defenses of plants and are found in large amounts in angiosperms that are prevalent under high-light conditions (47–49). Greater susceptibility of plants that are higher in phenylpropanoids like that encoded by Atprn1 could thus reduce barriers to infection, facilitating infection/colonization with C. neoformans under high-light conditions. However, the A. thaliana Atprn1 mutant did not display increased susceptibility to C. gattii, suggesting that accumulated phenylpropanoids may not be as important in C. gattii infections or that other factors may play a role. Other roles of Pirin proteins in general, such as effects on germination (31), apoptosis (32), or seedling growth (22) (Orozco-Nunnelly et al., submitted), could also have played a role in these differing plant susceptibilities. Interestingly, the understory of Vancouver Island and the Pacific Northwest, where C. gattii is prevalent, is fairly nondiverse, being composed largely of moss and ferns and, due to the low light and dampness, few angiosperms. Moss and ferns produce small amounts of phenylpropanoids, whereas angiosperms, represented by species such as wild ginger and young seedling trees, possess larger amounts of quercetin and other phenylpropanoids. Prolonged residence within a restricted niche of low-phenylpropanoid plants may help to explain why C. gattii does not favor the high phenylpropanoid content of the A. thaliana Atprn1 mutant, whereas a more light-tolerant species such as C. neoformans may favor such plants. It is thus interesting to speculate whether the reduced plant biodiversity of the Pacific Northwest promoted optimization of C. gattii as a plant pathogen, facilitating its emergence as a human pathogen. Indeed, recent work has shown a correlation between reduced biodiversity and the emergence of infectious pathogens (50). In summary, these data, while generated in the laboratory, show the diverse abilities of C neoformans and C. gattii to infect intact germinating seedlings.

These experiments also provide insights into the intersecting roles of fungal laccase, plant virulence, and the evolution of mammalian pathogenesis. The laccase enzyme is a major virulence factor of C. neoformans for mammalian infection (51), where its presence has been used as a marker of pathogenic species of Cryptococcus. The cryptococcal laccase product melanin has also been shown to be required for protection from killing by free-living amoebae (6), providing evolutionary pressure to optimize virulence factor expression. In mammalian pathogenesis, laccase has been implicated in prostaglandin E2-mediated immunosuppression (52), reductions in fenton reagents in macrophages (53), and production of melanin pigments (54). The present studies, as well as previous work, further suggest that evolutionary pressures within the plant-fungal ecological niche could have facilitated optimized virulence factor expression. Fowler et al. (55) recently reported that the commonly found plant flavonoids are substrates for laccase, resulting in the formation of a defensive melanin-like cell wall coating. Such coated fungal cells have an increased resistance to cell death caused by oxidants produced by UV radiation and macrophages (56). The present studies expand knowledge of the relationship of laccase with flavonoids by showing that the virulence factor laccase has a role in facilitating fungal growth in the presence of these plant compounds. In addition, mutation of AtPRN1 led to increased susceptibility to C. neoformans infection, and a dependence of this virulence on the presence of the cryptococcal LAC1 gene suggests a common feature between laccase and products affected by expression of AtPRN1.

In summary, these data show an expansion of the ecological niche of Cryptococcus to include infection of intact seedlings by yeast forms of the fungus, resulting in both lethal infections and persistent colonization. They also suggest a complex relationship between these infections and the role of laccase as a plant virulence factor that has the potential to shape the evolution of virulence of this important mammalian pathogen.