Evaluation of the Roles of Four Candida albicans Genes in Virulence by Using Gene Disruption Strains That Express URA3 from the Native Locus

In a previous article, using the ura blaster method of targeted gene disruption, we demonstrated that NOT5, a gene that is induced during thrush, encodes a newly identified factor important in candidal pathogenesis (2). Although the ura blaster method has been the conventional strategy for studying candidal virulence, mutant strains are not isogenic to control strains, since they do not express the selection marker URA3 from its native locus. It has recently been demonstrated that differences in URA3 position can alter virulence potential (1, 5, 6, 10, 11). In this study, we reinserted URA3 in its native locus in the not5 null mutant background, in order to systematically study the effects of URA3 position on morphogenesis, activity of orotidine 5′-monophosphate decarboxylase (OMP, the enzyme encoded by URA3), and virulence. In addition, we studied URA3 positional effects in null mutant strains of three other Candida albicans genes (NOT3, BUR2, and KEL1).

NOT5 (C. albicans open reading frame [ORF] IPF16198) encodes a protein of 662 amino acids (aa) that contains two transcription regulatory domains similar to those of the Saccharomyces cerevisiae Not protein family (Stanford DNA Sequencing and Technology Center websites http://www-sequence.stanford.edu/group/candida and http://genome-www.stanford.edu/Saccharomyces). In S. cerevisiae, the Not proteins comprise part of the CCR-NOT transcriptional regulatory complex (7). The C. albicans genome contains a gene closely related to NOT5, which is annotated as NOT3 in the CandidaDB website (http://genolist.pasteur.fr/CandidaDB/). Given the possibility that the proteins encoded by C. albicans NOT5 and NOT3 interact to regulate transcription, we included both genes in this study. In addition, we used in vivo induced antigen technology to identify ORFs IPF2971 and YGR238C as C. albicans genes that are expressed during thrush (2). ORF IPF2971 encodes a 526-aa protein that contains a cyclin domain (National Center for Biotechnology Information website http://www.ncbi.nlm.nih.gov) and exhibits 22% identity with S. cerevisiae Bur2p (Saccharomyces Genome Database [http://genome-www.stanford.edu/Saccharomyces]). On the basis of homology, we called the C. albicans gene corresponding to ORF IPF2971 BUR2. YGR238C (annotated as KEL1 in CandidaDB) encodes a 1,018-aa protein that contains a Kelch repeat domain and exhibits 26% identity with S. cerevisiae Kel1p (Saccharomyces Genome Database). The roles of C. albicans BUR2 and KEL1 in pathogenesis are unknown.

We used the ura blaster method to construct mutants of NOT3, BUR2, and KEL1 (Table 1); mutants of NOT5 were created in our earlier study. To reintroduce URA3 to its native locus, we removed URA3 from the second disrupted locus of the null mutants by selection on medium containing 5-fluoroorotic acid. We then purified a 4.8-kb PstI-BglII fragment containing URA3 from pUR3 (4) and used it to transform the ura3 mutants (4, 8, 9) (Table 1). URA3 position was verified by Southern analysis (data not shown). There was no difference in growth rates between any of the gene disruption strains and CAI-12 in liquid YPD (yeast extract-peptone-dextrose) medium at 37°C (data not shown).

We demonstrated that OMP activity correlated with URA3 position and copy number. OMP activity (expressed as units per milligram of protein) was significantly higher for C. albicans strain SC5314 (0.0050 ± 0.0002), which contains two copies of URA3, than for strains CAI-12 (0.0031 ± 0.0004; P = 0.002) and CAI-4 (0.0003 ± 0.0003; P < 0.0001), which contain a single copy or no copy, respectively. Furthermore, OMP activity was significantly lower for each gene disruption strain expressing URA3 from an ectopic locus than for the isogenic gene disruption strain with URA3 at the native locus (P < 0.05) (Table 2). Of note, the mutants with URA3 reintroduced at the native locus exhibited enzymatic activity at least as great as that of CAI-12.

To assess morphogenesis, we grew strains on medium 199 and Spider solid medium and calculated their filamentation ratios (FRs) (FR is the diameter of the colony plus lateral growth/diameter of colony). Among strains with URA3 at its native locus, only strain ΔNOT5d-U was unable to form hyphae on solid media. Strains ΔNOT5s-U, ΔNOT3s-U, ΔNOT3d-U, and ΔKEL1d-U formed hyphae, although their FRs were significantly shorter than that of CAI-12 (P < 0.05) (Table 2). We clearly demonstrated an association between URA3 position and morphogenesis, as the FRs for mutant strains with URA3 at its native locus were significantly longer than the FRs of strains expressing ectopic URA3 (P < 0.05) (Table 2). This association was most clearly demonstrated for the not3 null mutants. Strain ΔNOT3d was unable to form hyphae at all, whereas ΔNOT3d-U formed hyphae on both surfaces. Furthermore, strains ΔNOT3d, ΔNOT5s, and ΔNOT3s formed hyphae as readily in medium supplemented with uridine as strains ΔNOT3d-U, ΔNOT5s-U and ΔNOT3s-U, respectively. Since OMP is necessary for growth in environments lacking uridine, these finding suggest that morphogenesis was impaired in strains ΔNOT3d, ΔNOT5s, and ΔNOT3s due to insufficient OMP activity. To our knowledge, this is the first paper to demonstrate a relationship between URA3 position and morphogenesis.

To assess adherence, freshly harvested human buccal epithelial cells (BECs) were incubated with C. albicans strains for 1 h at 37°C, and the number of candidal cells attached to 100 BECs was counted (12). Among disruption strains with URA3 at its native locus, only strains ΔNOT5s-U and ΔNOT5d-U exhibited significantly less adherence to BECs than CAI-12 (P < 0.05) (Table 2). There was a clear association between URA3 position and adherence: strains ΔNOT5s-U, ΔNOT3d-U, ΔBUR2d-U, and ΔKEL1d-U exhibited greater adherence to BECs than did strains expressing ectopic URA3 (P < 0.05). Specifically, we showed that whereas ΔNOT3s, ΔNOT3d, and ΔKEL1d had significantly reduced adherence to BECs than CAI-12, adherence was no longer different from CAI-12 when URA3 was reintroduced into its native locus. This suggests that the attenuated adherence observed originally was due to the difference in URA3 position.

To assess pathogenicity in disseminated candidiasis, mice were infected via the lateral tail vein with 10⁶ CFU. Strain ΔNOT5d-U had attenuated virulence (Table 2): the time to mortality was significantly longer for ΔNOT5d-U than CAI-12. We demonstrated an association between URA3 position and virulence: the time to mortality for mice infected with ΔNOT5s-U, ΔNOT5d-U, ΔNOT3d-U, and ΔBUR2d-U was shorter than for mice infected with mutants in which URA3 was expressed ectopically (P < 0.05) (Table 2). Of note, ΔNOT5s-U, ΔBUR2d-U, and ΔKEL1d-U restored mortality to levels comparable to that of the control, suggesting that the attenuated mortality observed with ΔNOT5s, ΔBUR2d, and ΔKEL1d was caused by differences in URA3 position, rather than disruptions of the genes in interest.

By reintroducing URA3 to its native locus in the not5 null mutant background, we confirmed our earlier report that NOT5 plays a role in candidal virulence. More importantly, we showed that ectopic placement of URA3 in a variety of gene disruption backgrounds led to reductions in OMP activity, hyphal formation, adherence, and virulence. These differences altered our interpretation of the potential contributions of NOT3, BUR2, and KEL1 in the pathogenic process. It is therefore advisable that C. albicans genes previously ascribed a role in virulence using the conventional ura blaster method be retested, using strains expressing URA3 from its native locus or some other common site. Regardless of how the ura blaster method is modified, however, selectable markers that do not influence candidal phenotypes or growth within animal hosts would greatly facilitate future studies of pathogenesis.

• American Society for Microbiology