Sirtuins in the phylum Basidiomycota: A role in virulence in Cryptococcus neoformans

Abstract

Virulence of Cryptococcus neoformans is regulated by a range of transcription factors, and is also influenced by the acquisition of adaptive mutations during infection. Beyond the temporal regulation of virulence factor production by transcription factors and these permanent microevolutionary changes, heritable epigenetic modifications such as histone deacetylation may also play a role during infection. Here we describe the first comprehensive analysis of the sirtuin class of NAD+ dependent histone deacetylases in the phylum Basidiomycota, identifying five sirtuins encoded in the C. neoformans genome. Each sirtuin gene was deleted and a wide range of phenotypic tests performed to gain insight into the potential roles they play. Given the pleiotropic nature of sirtuins in other species, it was surprising that only two of the five deletion strains revealed mutant phenotypes in vitro. However, cryptic consequences of the loss of each sirtuin were identified through whole cell proteomics, and mouse infections revealed a role in virulence for SIR2, HST3 and HST4. The most intriguing phenotype was the repeated inability to complement mutant phenotypes through the reintroduction of the wild-type gene. These data support the model that regulation of sirtuin activity may be employed to enable a drastic alteration of the epigenetic landscape and virulence of C. neoformans.

Figure 1. C. neoformans has five candidate sirtuins.

(A) % identity scores (% similarity in parentheses) of predicted C. neoformans sirtuins with the S. cerevisiae proteins. (B) Maximum likelihood phylogenetic tree of the S. cerevisiae and C. neoformans predicted sirtuins created in MEGA7 with a WAG + G + I model and gamma shape parameters with 500 bootstrap replicates. Scale bar: 0.2 amino acid substitutions per site.

Figure 2. In vitro virulence factor production is decreased in only two of the five C. neoformans sirtuin mutants.

Spot dilution assays on L-DOPA, Christensen’s, and egg yolk agar were used to determine the level of melanin, urease, and phospholipase B production, respectively. Melanized strains appear brown-black, while urease and phospholipase B production appear as halos surrounding the colony. Plates were grown at both 30 and 37 °C.

Figure 3. Proteomic analyses reveal a significant number of protein abundance changes in the C. neoformans sirtuin mutants.

(A) Clustered heat map of normalized protein abundances for the wild-type and 5 mutants strains. “Red represents the most abundant proteins, yellow the least abundant”. (B) Number of proteins that were identified as significantly increased (up-arrow) or significantly decreased (down arrow) compared with wild type.

Figure 4. Sir2 association with the C. neoformans genome.

Representation of CHiP-seq data aligned to the rDNA array (A), a tRNA (B) and a miscRNA (C).

Figure 5. Virulence of sir2Δ in an invertebrate model of infection.

Survival of Bristol N2 strain nematodes co-cultured with C. neoformans on brain heart infusion or minimal media at 25 °C. p =< 0.005.

Figure 6. Virulence of sir2Δ in a murine model of infection.

Survival of female BALB/c mice infected with 5 × 10⁵ cells via nasal inhalation. Four sirtuin strains were subjected to a murine model of infection, with the original sir2∆ mutant (A), a second independent sir2∆ mutant (B), and two sir2∆ mutants created in the H99O background (C) all showing attenuation compared to wild type. Kill curves were considered statistically significant from wild-type when p =< 0.05; sir2∆-1 (p =< 0.001), sir2∆-1 + SIR2 (p =< 0.001), sir2∆-2 (p =< 0.001), sir2∆-2 + SIR2 (p =< 0.001), sir2∆-3 (p =< 0.001), sir2∆-3 + SIR2 (p =< 0.001), sir2∆-3 (p =< 0.001), sir2∆-3 + SIR2_2 kb (p =< 0.001), sir2∆-4 (p =< 0.001) and sir2∆-4 + SIR2_2 kb (p =< 0.001).

Figure 7. Virulence of hst2Δ, hst3Δ hst4Δ and hst5Δ in a murine model of infection.

Survival of female BALB/c mice infected with 5 × 10⁵ cells via nasal inhalation. Four sirtuin strains and their derived complemented strains were subjected to a murine model of infection. Kill curves were considered statistically significant from wild-type when p =< 0.05; hst3Δ (p =< 0.001), hst3Δ + HST3 (p =< 0.001), hst4Δ (p =< 0.001) and hst4Δ + HST4 (p =< 0.001).