Targeted gene disruption reveals an adhesin indispensable for pathogenicity of Blastomyces dermatitidis

Abstract

Systemic fungal infections are becoming more common and difficult to treat, yet the pathogenesis of these infectious diseases remains poorly understood. In many cases, pathogenicity can be attributed to the ability of the fungi to adhere to target tissues, but the lack of tractable genetic systems has limited progress in understanding and interfering with the offending fungal products. In Blastomyces dermatitidis, the agent of blastomycosis, a respiratory and disseminated mycosis of people and animals worldwide, expression of the putative adhesin encoded by the WI-1 gene was investigated as a possible virulence factor. DNA-mediated gene transfer was used to disrupt the WI-1 locus by allelic replacement, resulting in impaired binding and entry of yeasts into macrophages, loss of adherence to lung tissue, and abolishment of virulence in mice; each of these properties was fully restored after reconstitution of WI-1 by means of gene transfer. These findings establish the pivotal role of WI-1 in adherence and virulence of B. dermatitidis yeasts. To our knowledge, they offer the first example of a genetically proven virulence determinant among systemic dimorphic fungi, and underscore the value of reverse genetics for studies of pathogenesis in these organisms.

Figure 1

(A) Strategy for targeting and replacing the WI-1 locus in B. dermatitidis. Anticipated homologous crossover leading to a gene replacement event at the WI-1 locus. Linearized pQWhph can direct the hph gene to the WI-1 locus through homologous crossover in two ways. ×'s depict crossing over at 1 kb of WI-1 5′ flanking sequence (all WI-1 coding, designated in black) and at 2.5 kb of WI-1 3′ flanking sequence (881 bp of WI-1 coding in black and 1395 bp of noncoding in horizontal lines). Alternatively, the broken line depicts crossing over at a short region of the 375-bp WI-1 minipromoter in front of the hph gene. Either crossover event would replace most of the WI-1 coding region with the hph cassette. (B) PCR analysis for targeted gene replacement of WI-1. Candidates isolated in colony screening for a WI-1 null phenotype as described (Materials and Methods) were further analyzed for targeted recombination by PCR amplification of a novel junction fragment spanning the WI-1 5′ region (not present on the targeting vector) and the hph gene. The forward primer (5′-TTGTTTGTCTCTGCCCCGTTTTC-3′ at −644 in the WI-1 5′ region) is present only in the endogenous locus, whereas the reverse primer (5′-CGTCGCGGTGAGTTCAGGCTTTTTC-3′ at +31) is present in the hph gene on transforming DNA. The resulting 675-bp fragment is amplified if WI-1 sequences have been interrupted by hph in the manner depicted by a dashed line in A, yielding an interrupted locus shown at the bottom of A and in B. The product's authenticity is confirmed by the presence of EcoRV and AatII restriction sites in the WI-1 upstream segment of the product, as shown in an accompanying agarose gel, and in accord with published WI-1 genomic sequence (8).

Figure 2

Phenotype and genotype of WI-1 knockout strains and reconstituted strains. Flow cytometry (A) and Western blot analyses (B) were done using anti–WI-1 mAb DD5CB4 (–10, 19). A and B show results for wild-type parental strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55; another parental strain, ATCC 60915, and its isogenic WI-1 knockout strain, 99, gave similar results that are not shown. (A) FACS^® analysis of WI-1 surface expression. WI-1 was stained with 1–2 μg of purified mAb per 10⁶ yeasts. Nonspecific mAb binding was assessed with IgG2A isotype control mAb used at the same concentration as anti–WI-1 mAb. Values for the mean channel fluorescence intensity of anti–WI-1 binding were: wild-type strain 26199 = 318; WI-1 reconstituted strain 4/55 = 403; and WI-1 knockout strain 55 = 5. (B) Western blot analysis of WI-1 in yeast cell extracts. Yeast cell proteins were extracted as described in Materials and Methods, using 10⁷ yeasts per strain, and loading an equivalent volume of cell-free extract per lane. Lane 1 is a positive control with 4 μg of purified WI-1. (C) Southern blot analysis of WI-1 targeting and replacement. Genomic DNAs were restricted with XbaI. Strains are wild-type parent strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55. ATCC 60915 and isogenic WI-1 knockout strain 99 are included to illustrate targeted gene replacement of WI-1 in a second strain of B. dermatitidis. The hph probe contains only hph coding sequence. The WI-1 probe is an AatII/EcoRI fragment from an upstream region of the gene (nucleotides −1091 to −388), which is not carried on the knockout vector QWhph.

Figure 2

Phenotype and genotype of WI-1 knockout strains and reconstituted strains. Flow cytometry (A) and Western blot analyses (B) were done using anti–WI-1 mAb DD5CB4 (–10, 19). A and B show results for wild-type parental strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55; another parental strain, ATCC 60915, and its isogenic WI-1 knockout strain, 99, gave similar results that are not shown. (A) FACS^® analysis of WI-1 surface expression. WI-1 was stained with 1–2 μg of purified mAb per 10⁶ yeasts. Nonspecific mAb binding was assessed with IgG2A isotype control mAb used at the same concentration as anti–WI-1 mAb. Values for the mean channel fluorescence intensity of anti–WI-1 binding were: wild-type strain 26199 = 318; WI-1 reconstituted strain 4/55 = 403; and WI-1 knockout strain 55 = 5. (B) Western blot analysis of WI-1 in yeast cell extracts. Yeast cell proteins were extracted as described in Materials and Methods, using 10⁷ yeasts per strain, and loading an equivalent volume of cell-free extract per lane. Lane 1 is a positive control with 4 μg of purified WI-1. (C) Southern blot analysis of WI-1 targeting and replacement. Genomic DNAs were restricted with XbaI. Strains are wild-type parent strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55. ATCC 60915 and isogenic WI-1 knockout strain 99 are included to illustrate targeted gene replacement of WI-1 in a second strain of B. dermatitidis. The hph probe contains only hph coding sequence. The WI-1 probe is an AatII/EcoRI fragment from an upstream region of the gene (nucleotides −1091 to −388), which is not carried on the knockout vector QWhph.

Figure 2

Phenotype and genotype of WI-1 knockout strains and reconstituted strains. Flow cytometry (A) and Western blot analyses (B) were done using anti–WI-1 mAb DD5CB4 (–10, 19). A and B show results for wild-type parental strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55; another parental strain, ATCC 60915, and its isogenic WI-1 knockout strain, 99, gave similar results that are not shown. (A) FACS^® analysis of WI-1 surface expression. WI-1 was stained with 1–2 μg of purified mAb per 10⁶ yeasts. Nonspecific mAb binding was assessed with IgG2A isotype control mAb used at the same concentration as anti–WI-1 mAb. Values for the mean channel fluorescence intensity of anti–WI-1 binding were: wild-type strain 26199 = 318; WI-1 reconstituted strain 4/55 = 403; and WI-1 knockout strain 55 = 5. (B) Western blot analysis of WI-1 in yeast cell extracts. Yeast cell proteins were extracted as described in Materials and Methods, using 10⁷ yeasts per strain, and loading an equivalent volume of cell-free extract per lane. Lane 1 is a positive control with 4 μg of purified WI-1. (C) Southern blot analysis of WI-1 targeting and replacement. Genomic DNAs were restricted with XbaI. Strains are wild-type parent strain ATCC 26199, isogenic knockout strain 55, and WI-1 reconstituted strain 4/55. ATCC 60915 and isogenic WI-1 knockout strain 99 are included to illustrate targeted gene replacement of WI-1 in a second strain of B. dermatitidis. The hph probe contains only hph coding sequence. The WI-1 probe is an AatII/EcoRI fragment from an upstream region of the gene (nucleotides −1091 to −388), which is not carried on the knockout vector QWhph.

Figure 3

WI-1 mediated binding of B. dermatitidis to macrophages and lung tissue. (A) Binding and entry of macrophages by B. dermatitidis wild-type and knockout yeast. The upper portion of the panel shows binding of yeasts to macrophage J774.16 cells. The lower portion shows entry of yeasts into the macrophages. Binding and entry were quantified after incubation for varying intervals as shown. Unopsonized yeasts were added to wells of chamber slides containing 2.5 × 10⁴ macrophages at a ratio of four yeast cells per macrophage. The association index is defined as the number of attached and ingested yeasts per macrophage. The ingestion index is defined as the number of yeasts internalized per macrophage. Results are mean ± SEM of three independent experiments. The two strains differ significantly in binding and entry of macrophages for all time points after 1 h of incubation (P < 0.01). (B) Complement reverses defects in binding and entry in the WI-1 knockout strain. Assays were done as in A, except that they included 10% fresh mouse serum (as a source of complement) or heat-inactivated mouse serum (as a control). Results are the mean ± SEM for three separate experiments. Wild-type and knockout yeast differ significantly from each other in binding and ingestion both in the absence of serum and in the presence of heat-inactivated serum (P < 0.001 for each comparison), but not in the presence of fresh serum. (C) Defective binding and entry of macrophages by WI-1 knockout yeasts is reversed after reexpression of WI-1. Unopsonized yeasts were incubated with macrophages for 6 h; results are the mean ± SEM of three separate experiments. Wild-type yeasts and WI-1 reconstituted yeasts bind and enter macrophages significantly better than do knockout yeasts (P < 0.001 for each comparison). (D) Binding of yeasts to lung tissue ex vivo. Lungs of normal mice were removed and cryopreserved and 6-μm sections were applied to glass slides as in Materials and Methods. The number of unopsonized yeasts attached to lung tissue sections was quantified. The extent of binding is defined as the number of yeasts attached per 0.01 mm² of lung tissue, which was determined by inspecting a high power field at ×600 and counting 30 fields per strain in each experiment. Results are the mean ± SEM of six independent experiments. Wild-type yeasts and WI-1 reconstituted yeasts bind to lung tissue significantly better than do knockout yeasts (P < 0.001 for each comparison).

Figure 3

WI-1 mediated binding of B. dermatitidis to macrophages and lung tissue. (A) Binding and entry of macrophages by B. dermatitidis wild-type and knockout yeast. The upper portion of the panel shows binding of yeasts to macrophage J774.16 cells. The lower portion shows entry of yeasts into the macrophages. Binding and entry were quantified after incubation for varying intervals as shown. Unopsonized yeasts were added to wells of chamber slides containing 2.5 × 10⁴ macrophages at a ratio of four yeast cells per macrophage. The association index is defined as the number of attached and ingested yeasts per macrophage. The ingestion index is defined as the number of yeasts internalized per macrophage. Results are mean ± SEM of three independent experiments. The two strains differ significantly in binding and entry of macrophages for all time points after 1 h of incubation (P < 0.01). (B) Complement reverses defects in binding and entry in the WI-1 knockout strain. Assays were done as in A, except that they included 10% fresh mouse serum (as a source of complement) or heat-inactivated mouse serum (as a control). Results are the mean ± SEM for three separate experiments. Wild-type and knockout yeast differ significantly from each other in binding and ingestion both in the absence of serum and in the presence of heat-inactivated serum (P < 0.001 for each comparison), but not in the presence of fresh serum. (C) Defective binding and entry of macrophages by WI-1 knockout yeasts is reversed after reexpression of WI-1. Unopsonized yeasts were incubated with macrophages for 6 h; results are the mean ± SEM of three separate experiments. Wild-type yeasts and WI-1 reconstituted yeasts bind and enter macrophages significantly better than do knockout yeasts (P < 0.001 for each comparison). (D) Binding of yeasts to lung tissue ex vivo. Lungs of normal mice were removed and cryopreserved and 6-μm sections were applied to glass slides as in Materials and Methods. The number of unopsonized yeasts attached to lung tissue sections was quantified. The extent of binding is defined as the number of yeasts attached per 0.01 mm² of lung tissue, which was determined by inspecting a high power field at ×600 and counting 30 fields per strain in each experiment. Results are the mean ± SEM of six independent experiments. Wild-type yeasts and WI-1 reconstituted yeasts bind to lung tissue significantly better than do knockout yeasts (P < 0.001 for each comparison).

Figure 4

Targeted gene replacement of WI-1 reduces the pathogenicity of B. dermatitidis. (A) Survival after infection with wild-type strain 26199 and WI-1 knockout strain 55. Male BALB/c mice (n = 15 mice/group) were infected intranasally with yeast cells of each strain, in varied doses. Survival was monitored for 72 d after infection. The two groups differ significantly (P < 0.001) at each infectious dose tested. The experiment shown is representative of three independent experiments. The phenotype of knockout yeasts was stable; no revertants were identified among yeasts grown from mice infected with strain 55. (B) Gross and microscopic pathology of mice infected with ATCC 26199 wild-type yeasts or WI-1 knockout yeasts. Mice were analyzed 3 wk after infection. Lungs were stained with hematoxylin and eosin to assess inflammation, and with Gomori methenamine silver to visualize yeasts. The arrow denotes an isolated granuloma surrounded by normal lung tissue in a mouse infected with the WI-1 knockout strain 55. (C) Survival after infection with wild-type strain 26199, WI-1 knockout strain 55, and WI-1 reconstituted strain 4/55. Survival experiments were done as in A, using a dose of 10⁴ yeasts to establish infection. The wild-type and WI-1 reconstituted strains were significantly different from the WI-1 knockout strain (P < 0.001 for each comparison).

Figure 4

Targeted gene replacement of WI-1 reduces the pathogenicity of B. dermatitidis. (A) Survival after infection with wild-type strain 26199 and WI-1 knockout strain 55. Male BALB/c mice (n = 15 mice/group) were infected intranasally with yeast cells of each strain, in varied doses. Survival was monitored for 72 d after infection. The two groups differ significantly (P < 0.001) at each infectious dose tested. The experiment shown is representative of three independent experiments. The phenotype of knockout yeasts was stable; no revertants were identified among yeasts grown from mice infected with strain 55. (B) Gross and microscopic pathology of mice infected with ATCC 26199 wild-type yeasts or WI-1 knockout yeasts. Mice were analyzed 3 wk after infection. Lungs were stained with hematoxylin and eosin to assess inflammation, and with Gomori methenamine silver to visualize yeasts. The arrow denotes an isolated granuloma surrounded by normal lung tissue in a mouse infected with the WI-1 knockout strain 55. (C) Survival after infection with wild-type strain 26199, WI-1 knockout strain 55, and WI-1 reconstituted strain 4/55. Survival experiments were done as in A, using a dose of 10⁴ yeasts to establish infection. The wild-type and WI-1 reconstituted strains were significantly different from the WI-1 knockout strain (P < 0.001 for each comparison).

Figure 4

Targeted gene replacement of WI-1 reduces the pathogenicity of B. dermatitidis. (A) Survival after infection with wild-type strain 26199 and WI-1 knockout strain 55. Male BALB/c mice (n = 15 mice/group) were infected intranasally with yeast cells of each strain, in varied doses. Survival was monitored for 72 d after infection. The two groups differ significantly (P < 0.001) at each infectious dose tested. The experiment shown is representative of three independent experiments. The phenotype of knockout yeasts was stable; no revertants were identified among yeasts grown from mice infected with strain 55. (B) Gross and microscopic pathology of mice infected with ATCC 26199 wild-type yeasts or WI-1 knockout yeasts. Mice were analyzed 3 wk after infection. Lungs were stained with hematoxylin and eosin to assess inflammation, and with Gomori methenamine silver to visualize yeasts. The arrow denotes an isolated granuloma surrounded by normal lung tissue in a mouse infected with the WI-1 knockout strain 55. (C) Survival after infection with wild-type strain 26199, WI-1 knockout strain 55, and WI-1 reconstituted strain 4/55. Survival experiments were done as in A, using a dose of 10⁴ yeasts to establish infection. The wild-type and WI-1 reconstituted strains were significantly different from the WI-1 knockout strain (P < 0.001 for each comparison).