Adhesive and Mammalian Transglutaminase Substrate Properties of Candida albicans Hwp1

See allHide authors and affiliations

Science  05 Mar 1999:
Vol. 283, Issue 5407, pp. 1535-1538
DOI: 10.1126/science.283.5407.1535


The pathogenesis of candidiasis involves invasion of host tissues by filamentous forms of the opportunistic yeast Candida albicans. Morphology-specific gene products may confer proinvasive properties. A hypha-specific surface protein, Hwp1, with similarities to mammalian small proline-rich proteins was shown to serve as a substrate for mammalian transglutaminases. Candida albicansstrains lacking Hwp1 were unable to form stable attachments to human buccal epithelial cells and had a reduced capacity to cause systemic candidiasis in mice. This represents a paradigm for microbial adhesion that implicates essential host enzymes.

Candida albicans causes severe oropharyngeal and esophageal mucositis in patients with human immunodeficiency virus and hematogenously disseminated disease in iatrogenically compromised hosts (1). Loss of immune host defenses is primarily responsible for susceptibility to opportunistic candidiasis arising from endogenous C. albicans in the normal flora. However, proadhesive and proinvasive factors of C. albicans contribute to disease by mediating penetration of host tissues. Filamentous forms, particularly true hyphae, embed themselves within the superficial, keratinized layer of stratified squamous epithelium and grow by apical extension (2). True hyphae are extensions of germ tubes that emerge from yeasts. To explore the role of surface proteins in tissue invasion, we studied the function ofHWP1, a developmentally regulated gene expressed in germ tubes and true hyphae (2). HWP1 encodes an outer mannoprotein, Hwp1, with a cell surface–exposed NH2-terminal domain and COOH-terminal features conferring covalent integration into cell wall β-glucan (2,3).

Unlike most pathogens that form various types of weak interactions with host cells, C. albicans germ tubes form nondissociable complexes with human buccal epithelial cells (BECs) that are characteristic of transglutaminase (TGase)-mediated reactions in stability and in being prevented by TGase inhibitors (4). A possible role for Hwp1 in formation of stable complexes with BECs was suggested by the amino acid sequence of the NH2-terminal domain, which resembles mammalian TGase substrates (2), particularly the head and central domains of small proline-rich (SPR) proteins (5). Perhaps Hwp1 contributes to proliferation ofC. albicans in keratinized epithelium by interacting with keratinocyte TGases (6), enzymes that create a host barrier defense by cross-linking SPR proteins and other proteins through covalent Nɛ-(γ-glutamyl)lysine isodipeptide bonds. Cross-linking of epithelial cell proteins is essential. Mice lacking keratinocyte TGase die within a few hours of birth (7).

To determine if the SPR-like region of Hwp1 is a substrate for TGase, we examined a recombinant protein, rHwp1ΔC37, that encompasses the NH2-terminal proline- and glutamine-rich domain (8) for the ability to incorporate [14C]putrescine in the presence of TGase2. TGase2, also known as tissue TGase, is widely distributed in the body and is commonly used to characterize TGase substrates (9,10). Levels of radioactivity associated with the recombinant protein were equivalent to those of casein, a known TGase substrate, and were fourfold greater than for bovine serum albumin (BSA), a negative control (Fig. 1A). Examination of proteins after TGase reactions (Fig. 1, B and C) showed no radioactivity associated with BSA; however, the recombinant protein was similar to other TGase2 substrates in the generation of multiple species of radiolabeled reaction products including monomers displaying increased migration (11), species of high molecular weight (10, 12), and dimers bridged by putrescine. The production of all radiolabeled forms depended on the presence of active TGase. In addition to [14C]putrescine, TGase2 catalyzed the incorporation of another TGase cosubstrate, monodansylcadaverine (13), into rHwp1ΔC37 (14). The behavior of rHwp1ΔC37 in interactions with TGase2 matched that of SPR proteins and other host TGase substrates and implicated the NH2-terminus of Hwp1 in cross-linking reactions with primary amines through Nɛ-(γ-glutamyl)lysine isodipeptide bonds.

Figure 1

(A) Incorporation of [14C]putrescine by rHwp1ΔC37, casein, and BSA mediated by TGase (25). The values are the means (±SD) of two experiments performed in triplicate. (B) Autoradiograph of reactions with rHwp1ΔC37, [14C]putrescine, and TGase after SDS–polyacrylamide gel electrophoresis and fluorography (26, 27). Labeled products were not formed when reactions were supplemented with TGase inhibitors (24), including 20 mM EGTA to chelate Ca2+ (lane 2) and 20 mM iodoacetamide (lane 3), or in the absence of TGase (lane 5) or rHwp1ΔC37 (lane 4). Control reactions lacking putrescine showed that formation of the 59kD species represented dimeric rHwp1ΔC37 bridged by putrescine (14). (C) Immunoblotting to detect rHwp1ΔC37 in TGase reactions with affinity-purified rabbit antibodies to rHwp1 (2, 28) and horseradish peroxidase–conjugated goat antibodies to rabbit immunoglobulin G (Zymed), developed with ECL reagents (Amersham). Lane 6, rHwp1ΔC37 alone. TG, Tgase; iodo., iodoacetamide.

To study the TGase substrate properties of native Hwp1 on germ tube surfaces, we created isogenic strains with or without Hwp1 (15) for comparison in TGase assays. Disruption ofHWP1 was verified by Southern (DNA) and Northern (RNA) blotting (Fig. 2). CAH7-1A lackedHWP1 mRNA (Fig. 2C) and protein (16), whereas heterozygous and complemented strains, CAH7 and CAHR3, had about half the amount of HWP1 mRNA found in strains with unaltered HWP1 genes, although the amount of Hwp1 was indistinguishable from that of wild-type strains (14). Germ tubes of strains with one or both copies of HWP1 displayed equivalent levels of TGase2 substrates, whereas parent blastoconidia were negative (Fig. 3A). Identical results were found with four clinical isolates of C. albicans from dental patients, and with other laboratory strains of C. albicans (14). TGase substrates were not detected in the presence of EGTA (Fig. 3A) or iodoacetamide, nor was endogenous TGase activity of C. albicans detected in whole organisms or in broken cell walls (14). In contrast to strains with Hwp1, TGase substrates were not detectable on germ tubes of the homozygous hwp1/hwp1 mutant CAH7-1A (Fig. 3A), indicating the requirement of Hwp1 for TGase substrate activity. Thus, the NH2-terminal domain of native Hwp1 appears to be a surface-exposed TGase substrate available for interactions with exogenous mammalian TGases.

Figure 2

Disruption of HWP1(15). (A) Genomic HWP1DNA open reading frame (arrow), plasmid insert of pGBHWP1 (shaded rectangle), disruption fragment from pHWP1URA3, and the rescue fragment (1.7 kb). (B) Southern blot of Eco RI–digested genomic DNA probed with the rescue fragment; HWP1/HWP1(SC5314, CAI4, and UnoPP-1) (3.8 kb) (lanes 1, 2, and 6), hwp1/HWP1 (CAH7) (3.8, 5.3, and 2.3 kb) (lane 3), hwp1/hwp1 (CAH7-1A) (5.3 kb, 5.1-kb doublet, and 2.3 kb) (lane 4), revertant CAHR3 (5.1 and 3.8 kb, 3.7-kb doublet) (lane 5). (C) Northern blot probed withHWP1 and ENO (internal control) (2). Although CAHR3 harbored excess HWP1 DNA (3.7-kb fragment),HWP1 mRNA levels [quantified with ImageQuant Software (Molecular Dynamics)] were equivalent to those of CAH7, and developmental regulation was preserved.

Figure 3

(A) Detection of TGase substrates on germ tube surfaces of hwp1/hwp1 mutant and wild-type strains of C. albicans. Washed, M199-germinatedC. albicans cells (2 × 107 per milliliter) (2), TGase2 (8.5 μg) (25), and 5-(biotinamido)pentylamine (30 μM) (Pierce) were incubated in reaction buffer 1 (400 μl) [100 mM tris-HCl (pH 7.5), 5 mM CaCl2, 1 mM DTT, 2 mM EDTA] for 15 min at 37°C. EDTA (10 mM) was added to stop the reaction. Staining with avidin–fluorescein isothiocyanate (1:100) (Zymed) detected biotin groups on germ tubes, and BSA-rhodamine (1:30) (Difco) in PBS served as a counterstain. SC5314 (wild-type), CAH7-1A (hwp1/hwp1), CAHR3 revertant (HWP1/hwp1), SC5314 + EGTA (SC5314 cells with 20 mM EGTA). Bar, 5 μm. (B and C) Adherence to human BECs (17) of heterozygous hwp1/HWP1 (CAH7) and homozygoushwp1/hwp1 (CAH7-1A) mutants, and HWP1revertant (CAHR3), relative to the homozygous HWP1/HWP1strain [UnoPP-1, a CAI4 derivative made Ura+ by disruption of an enolase gene with URA3 (20)]. (B) Stabilized adhesion. Adherence of UnoPP-1 was set at 100%, corresponding to counts per minute of 3100 to 6200, resulting in germ tube BEC ratios of 1 to 2:1 in a typical experiment. Asterisks indicate the presence of iodoacetamide (*) and monodansylcadaverine (**). The values are the means (±SD) of two experiments performed in duplicate. Statistically significant differences were determined by Student's t test. CAH7-1A versus CAH7 (P = 0.009), CAH7-1A versus UnoPP-1 with iodoacetamide or monodansylcadaverine to inhibit TGase (P = 0.936 and 0.142, respectively), and CAHR3 versus CAH7 (P = 0.977). (C) Overall adhesion. CAH7-1A versus CAH7 (P = 0.032), and CAHR3 versus CAH7 (P = 0.513). (D) BEC envelope-associated [14C]rHwp1ΔC37 after incubations under adhesion assay conditions (19) with (*) or without 10 mM iodoacetamide followed by boiling in sample buffer for 5 min. Results of two experiments performed in duplicate (±SD) are shown. Statistically significant differences were determined by Student'st test. Uninhibited samples versus samples with iodoacetamide (P = 0.025), rHwp1ΔC37 versus BSA (P = 0.025), and samples with iodoacetamide versus BSA (P = 0.855).

If Hwp1 proteins serve as connections to BECs through interactions with TGases, then the stable attachments described previously (4) should be greatly reduced in strains lacking Hwp1. To test this hypothesis, we compared isogenic strains with or without Hwp1 in stabilized adhesion assays, which involved treatment of germ tube:BEC complexes with heat and the anionic detergent SDS to dissociate weak, noncovalent bonds (4, 17). Strains with one or both copies of HWP1 displayed equivalent levels of stabilized adhesion (Fig. 3B), indicating a lack of gene copy number effect and correlating well with apparent wild-type levels of Hwp1 protein and TGase substrates on germ tubes in these strains (Fig. 3A). Stabilized adhesion was prevented by monodansylcadaverine, a competitive inhibitor of TGase-mediated protein cross-linking reactions (13, 18), and by iodoacetamide, supporting the involvement of BEC TGases (Fig. 3B). The hwp1/hwp1 mutant strain CAH7-1A was greatly impaired in the ability to form stable attachments to BECs in that stabilized adhesion was only 23% of the other strains and was equivalent to values for other strains when TGase inhibitors were added. To provide further support for the role of Hwp1 in mediating stabilized adhesion, we radiolabeled rHwp1ΔC37 and incubated it with BECs in the presence or absence of iodoacetamide followed by treatment with heat and SDS (19). The results were consistent with stabilized adhesion assays in that rHwp1ΔC37 associated with BECs was sixfold greater when TGase was not inhibited. Adhesion of the control protein, BSA, was minimal (Fig. 3D). We have not identified the BEC TGases and TGase substrates participating in interactions with C. albicans, but rHwp1ΔC37 forms stable attachments to BECs and Hwp1 is required for formation of stable complexes between C. albicans germ tubes and BECs.

Hwp1 had a minor effect on overall adhesion of germ tube:BEC complexes not subjected to heat and detergent treatments (Fig. 3C). Overall adhesion of CAH7-1A was 45% that of UnoPP-1, lower than the expected value of 65% on the basis of a stabilized to overall adhesion ratio of 35 ± 5.75% for UnoPP-1. Hwp1 may contribute to typical, noncovalent adhesive forces between germ tubes and BECs, but it is more likely that the primary role of Hwp1 is in the formation of stable attachments and that the in vitro assay does not permit maximal cross-linking of germ tubes to BECs. Hyphae complexed to human BECs in vivo in specimens of pseudomembranous candidiasis of the buccal mucosa were not dissociated by heat and detergent treatments (up to 30 min) used in the stabilized adhesion assays (14), indicating that stable attachments predominate in candidiasis. Long incubation periods (18 hours) are required for maximal cross-linking of SPR2 protein by epithelial cell TGase3 (10). The similarity of Hwp1 to SPR proteins suggests that interactions of BEC TGases with SPR proteins are indicative of interactions with Hwp1 and that the primary role of Hwp1 is in the formation of stable attachments.

The importance of Hwp1 in candidiasis is supported by experiments showing that, despite similar growth rates and germ tube production [(15) and Fig. 3A], CAH7-1A was a poor inducer of systemic candidiasis in mice (Fig. 4). Of the six mice injected with CAH7-1A, five survived to the 30-day end point, two having cleared the infection. In the other groups, only 2 of 18 mice given HWP1-expressing strains survived. Colony-forming units of C. albicans were detected in homogenized brain, kidney, liver, and spleen of all infected animals, and numerous hyphae were found in the kidneys. CAH7-1A hyphae appeared less invasive in that organisms were limited to the renal pelvis, whereas other strains invaded the parenchyma (14). Thus, Hwp1 may promote internal candidiasis through TGases present in nonsuperficial tissues.

Figure 4

Survival of mice (CBA/J H-2khaplotype) intravenously injected with HWP1 mutant strains. Four groups (six per group) of mice were injected with stationary-phase yeasts (29) (2 × 105 cells per mouse in 0.2 ml of PBS) of SC5314 (wild-type), CAH7 (hwp1/HWP1), CAH7-1A (hwp1/hwp1), or CAHR3 (HWP1/hwp1, revertant). The experiment was terminated at 30 days. The authenticity of strains taken from organs of infected mice was verified by assessment of the presence of Hwp1 on germ tubes by indirect immunofluorescence. Statistical analysis showed survival differences (P < 0.01) by the Wilcoxon rank sum test. (*) Significance of survival differences relative to thehwp1/hwp1 strain, CAH7-1A, by the log rank test. Differences between strains expressing HWP1 were not significant.

Germ tubes and hyphae of C. albicans exhibit highly polarized, apical growth and require mechanisms for anchoring within and penetration of host tissues. In mimicking mammalian TGase substrates, Hwp1 forms stable attachments between germ tubes and mammalian cells and is important in the pathogenesis of candidiasis.

  • * Present address: Grays Harbor Dental Specialists, 105 South Broadway, Aberdeen, WA 98520, USA.

  • To whom correspondence should be addressed. E-mail: sundstrom.1{at}


View Abstract

Navigate This Article