Functional Characterization of Secreted Aspartyl Proteases in Candida parapsilosis

Aspartyl proteases are present in various organisms and, among virulent species, are considered major virulence factors. Host tissue and cell damage, hijacking of immune responses, and hiding from innate immune cells are the most common behaviors of fungal secreted proteases enabling pathogen survival and invasion. C. parapsilosis, an opportunistic human-pathogenic fungus mainly threatening low-birth weight neonates and children, possesses three SAPP protein-encoding genes that could contribute to the invasiveness of the species. Our results suggest that SAPP1 and SAPP2, but not SAPP3, influence host evasion by regulating cell damage, phagocytosis, phagosome-lysosome maturation, killing, and cytokine secretion. Furthermore, SAPP1 and SAPP2 also effectively contribute to complement evasion.

silosis is increasing worldwide and C. parapsilosis is currently the second or third most common yeast species associated with invasive candidiasis in hospitals in Asian, European, and South American countries (3). C. parapsilosis is commonly associated with low-birth-weight neonate infections, invasive infections of hospitalized immunocompromised patients, and the receipt of parenteral nutrition or prolonged use of intravascular devices (4). Despite its clinical significance, the pathogenicity of C. parapsilosis and its virulence factors and interactions with the host are still poorly understood (5)(6)(7).
Aspartyl proteases are present in various organisms and are most active at acidic pH (pH 1.9 to 4.0), share a catalytic apparatus, and cleave dipeptide bonds between two hydrophobic amino acid residues (8). Fungal secreted aspartyl proteases are reported to directly mediate virulence (9)(10)(11)(12)(13). C. parapsilosis possesses three aspartyl acid protease-encoding genes, namely, SAPP1, SAPP2, and SAPP3. SAPP1 is duplicated in the species' genome (SAPP1a, SAPP1b) (14). A previously established Δ/Δsapp1a Δ/Δsapp1b strain, lacking SAPP1, was shown to be hypersusceptible to human serum (HS), caused attenuated host cell damage, and was phagocytosed and killed more efficiently by human monocytes and macrophages than the wild-type strain (15). In another study using reconstituted human oral epithelium (RHOE), levels of tissue damage caused by C. parapsilosis were significantly reduced in the presence of the Sapp inhibitor pepstatin, further highlighting the role of secreted proteases in the species' pathogenicity (16).
Upon superficial infection, epithelial cells trigger an inflammatory response by producing antimicrobial peptides and recruiting and activating innate immune cells, including macrophages and neutrophils (17)(18)(19). Candida species can efficiently avoid macrophage-mediated killing by host membrane rupture, secretion of proteases and lipases, and induction of pyroptosis and by nutrient competition with the host (20)(21)(22). Upon infection, the complement cascade also activates and plays a role in combating pathogens via enhancing chemotaxis, phagocytosis, or T and B cell differentiation (23). Pathogenic species have adopted several strategies to evade complement attack (24). In particular, C. albicans either recruits complement regulator proteins on its surface or cleaves complement proteins by secreting the proteases. C. parapsilosis can also bind to human complement proteins; however, the effect of this binding has not been fully resolved (25,26).
To date, multiple studies have shown that C. albicans aspartyl proteases have different abilities to damage epithelial cells, alter the host complement cascade, induce macrophage chemotaxis or cytokine production, and mediate NLRP3 inflammasome activation; less is known about the immune modulatory effects of aspartyl proteases in C. parapsilosis (12,27). Therefore, to elucidate the role of individual aspartyl proteases in the virulence of C. parapsilosis, SAPP mutant strains were generated. Functional characterization of these genes revealed that SAPP1 and SAPP2 (but not SAPP3) play an important role in C. parapsilosis pathogenicity.
All reintegrant mutant strains were established on the sapp1/2/3 Ϫ/Ϫ background to avoid cross-interference from each Sapp. Mutant strains were confirmed by colony PCR and Southern blotting (data not shown).
Expression levels of SAPP genes in the reintegrant mutant strains were determined using real-time PCR. Wild-type and mutant strains were cultivated in secreted-proteaseinducing medium (yeast carbon base [YCB] plus 0.2% bovine serum albumin [BSA]), and the levels of expression of SAPP1, SAPP2, and SAPP3 were monitored after 48 h of incubation. The levels of expression of genes SAPP1 and SAPP2 in reintegrant strains RI_SAPP1 and RI_SAPP2 were similar to what was observed in the wild-type strains, while the level of expression of SAPP3 was upregulated in the RI_SAPP3 strain by Ն4-fold (Fig. 1).
Next, we examined whether reintegration of SAPP genes altered the viability, morphology, or biofilm-forming ability of the mutant strains. No difference was observed between the levels of growth of the mutants in either yeast extract-peptonedextrose (YPD) or YCB liquid medium at 30°C and the levels seen with the wild-type strain (see Fig. S1A and B in the supplemental material), and the SAPP mutant strains produced elongated pseudohyphae to the same extent as the reference strain in YPD or RPMI medium supplemented with 10% fetal bovine serum (FBS) and spider liquid medium ( Fig. S2A to C). We observed no difference in colony morphologies (Fig. S3) or in biofilm-forming abilities (Fig. S4). We also tested the ability of the sapp1/2/3 Ϫ/Ϫ mutant to cope with stress by monitoring cell growth in the presence of several stressors (Table S3). The sapp1/2/3 Ϫ/Ϫ mutant strain showed no differences in growth in the presence of stressors (Fig. S5). These results demonstrate that the mutant strains retained the physiological attributes and stress responses of the parental strain.
Semiquantitative detection of extracellular protease activity of SAPP mutant strains. Candida secreted aspartyl proteases hydrolyze BSA present in agar plates. In order to examine the secreted protease activity of the established strains, the wild-type and SAPP mutant strains were spotted on plates containing YCB plus 0.2% BSA and, following amido black staining, the width of the clearance zone was measured. The C. parapsilosis wild-type strain showed a clear halo zone (7.3 mm in diameter) on BSAcontaining plates similar to the zones seen with strains RI_SAPP1 (5.78 mm) and RI_SAPP2 (5.76 mm). The RI_SAPP3 and sapp1/2/3 Ϫ/Ϫ strains, however, showed no proteolytic activity (Fig. 2). These results suggest that, in contrast to SAPP1 and SAPP2, reintegration of SAPP3 does not restore the aspartyl protease activity of the sapp1/2/ 3 Ϫ/Ϫ strain; thus, SAPP3 does not contribute to aspartyl protease secretion in this species.
C. parapsilosis RI_SAPP3 and sapp1/2/3 ؊/؊ strains are sensitive to human serum. To investigate the fungicidal effect of human serum on the examined strains, yeast cells were cultivated in the presence of normal human serum (NHS) and CFU determinations were performed at different time intervals. C. parapsilosis strains were also grown in the presence of 20% heat-inactivated serum (HiS). The viability of the RI_SAPP3 and sapp1/2/3 Ϫ/Ϫ strains was reduced significantly after 18 and 24 h of incubation in intact serum compared to the wild-type strain results, while the RI_SAPP1 and as RI_SAPP2 strains showed no sensitivity to NHS (Fig. 3A). However, no sensitivity was observed after HiS treatment (Fig. 3B). These data suggest that Sapp1 and Sapp2 are involved in protection against human serum proteins but that Sapp3 is not associated with this effect.
Secreted aspartyl proteases affect the adhesion capabilities of C. parapsilosis. We further examined whether SAPP genes influence the adhesion properties of C. parapsilosis by the use of biotic and abiotic surfaces. Results of the cell adhesion assays showed that all three reintegrated mutant strains had significantly reduced capabilities of adhesion to polystyrene surfaces compared to the reference strain (Fig. 4A). The highest reduction in adhesion was observed with the sapp1/2/3 Ϫ/Ϫ strain (approximately 40%), followed by RI_SAPP2 (25%), RI_SAPP3 (25%), and the RI_SAPP1 strain (20%).
A significant reduction in adhesion to cells of the TR146 human oral epithelial cell line was observed with strain RI_SAPP3, while a moderate decrease was detected in the case of the sapp1/2/3 Ϫ/Ϫ strain (Fig. 4B).
Aspartyl proteases promote intracellular survival of C. parapsilosis by altering phagosome-lysosome maturation. A previous study reported that Candida cells can replicate and survive within macrophages, either by diverting the normal process of phagosome maturation, causing physical damage, or by withstanding the hostile environment of the mature phagosome-lysosome (30). Here, we aimed to examine if C. parapsilosis aspartyl proteases influence phagosome-lysosome maturation in human PBMC-DMs. We analyzed the phagosome-lysosome maturation after coincubating pHrodo-stained Candida cells with PBMC-DMs for 2 h. Interestingly, PBMC-DMs infected with the wild-type strain, mutant strain RI_SAPP1, and mutant strain RI_SAPP2 showed and HiS (B) was examined by determination of CFUs at 0, 6, 24, and 48 h. Data were obtained from three independent experiments. Differences between groups were considered statistically significant at P Ͻ 0.05. *, P Ͻ 0.05; **, P Ͻ 0.01. a lower rate of phagosome-lysosome fusion (16.66% Ϯ 0.5732%, 20.76% Ϯ 0.7194%, and 13.78% Ϯ 1.216%, respectively) than was seen with RI_SAPP3 (29.52% Ϯ 2.719%) and sapp1/2/3 Ϫ/Ϫ (28.70% Ϯ 2.025%), indicating that Sapp1 and Sapp2 (but not Sapp3) may promote intracellular survival of C. parapsilosis in human macrophages (Fig. 8).  C. parapsilosis Sapp proteins regulate the cytokine response of host macrophages. In order to examine if the cytokine responses triggered by the wild-type strain, the RI_SAPP mutants, and strains sapp1/2/3 Ϫ/Ϫ differed significantly, we stimulated human PBMC-DMs for 24 h with each strain and measured interleukin-1␤ (IL-1␤), tumor necrosis factor alpha (TNF-␣), IL-6, and IL-8 responses. The obtained results indicated that PBMC-DMs stimulated with either the wild-type strain or the RI_SAPP1 and RI_SAPP2 strains produced similar IL-1␤, IL-8, and TNF-␣ levels. In contrast, macrophages stimulated with strain sapp1/2/3 Ϫ/Ϫ produced significantly less IL-1␤ and IL-6 and moderately but not significantly less IL-8 than the wild-type strain (Fig. 9). PBMC-DMs stimulated with RI_SAPP3 produced significantly lower IL-8 and moderately low  IL-6 levels; however, no significant differences were observed in the production of IL-1␤ and TNF-␣ compared to wild type.
Sapp1p and Sapp2p have differential cleavage capacities against human complement proteins. C. albicans secreted aspartyl proteases can cleave components of human serum, including complement proteins (such as complement component 3b [C3b], C4b, and C5 and the complement regulator FH) and other microbicidal plasma proteins (31,32). Therefore, to test if C. parapsilosis Sapp proteins are also able to cleave human complement proteins, we incubated C3b and C4b and complement regulatory proteins with the purified Sapp proteins. Our results indicated that the cleavage efficiency of Sapp1p against C3b was higher (shown with stronger cleavage fragment) than that of Sapp2p, which may suggest a difference in the substrate preferences of the two proteases (Fig. 10A). Moreover, Sapp1p and Sapp2p were also able to cleave human C4b (Fig. 10B). Purified C3b and C4b were incubated without Sapp proteins for the same 3-h time period and used as negative controls; cleavage of C3b and cleavage of C4b by factor I in the presence of its cofactors were included as positive controls. To investigate if C. parapsilosis Sapp1p and Sapp2p can cleave complement regulators of the FH protein family, we measured the capacity of Sapp1p and Sapp2p to degrade FH, FHL-1, FHR-1, and FHR-5. Coincubation of Sapp1p or Sapp2p with FHL-1 or FHR-1 revealed that the proteases were not able to cleave these human complement proteins, as visualized by Western blotting (Fig. S6). However, FH was cleaved by both fungal proteases after 15 h of incubation. Interestingly, Sapp2p but not Sapp1p was able to cleave FHR-5 at the early time point of 3 h, further indicating a difference in the substrate preferences of C. parapsilosis Sapp proteins (Fig. 11).
Since attachment of opsonic complement proteins to pathogens enhances CR3mediated phagocytosis by macrophages and C. albicans cleaves CR3 and CR4 on macrophages (31), we also tested whether C. parapsilosis Sapp1p and Sapp2p can cleave complement receptors CR3 and CR4; however, we did not find substantial differences in the levels of expression of CR3 and CR4 receptors on macrophages after protease treatment (Fig. S7).
Fungal burden and Galleria mellonella survival. CFU recovery data show that RI_SAPP1 produced CFU numbers similar to those seen with wild-type C. parapsilosis in G. mellonella larvae (Fig. 12A). In contrast, the virulence of the other mutants was attenuated compared to that of the parental strain.
Overall, larvae infected with wild-type and mutant strains showed no significant difference in survival after the 7 days of infection (Fig. 12B).

DISCUSSION
Aspartyl proteases are present in a diverse range of microorganisms and play a crucial role in nutrition acquisition and pathogenesis. The presence of aspartyl pro- teases in pathogenic Candida species and their absence in nonpathogenic fungal species such as Saccharomyces cerevisiae suggests their role in pathogenesis. Previously, we showed that C. parapsilosis ⌬/⌬sapp1a, ⌬/⌬sapp1b, and ⌬/⌬sapp1a-/⌬/⌬sapp1b deletion mutant strains are less virulent than the wild-type strain, demonstrating that Sapp1p plays a role in pathogenesis regulation. To date, the roles of SAPP2 and SAPP3 in C. parapsilosis virulence have not been investigated. Therefore, in the present study, we aimed to delineate their roles in pathogenicity using a secreted aspartyl proteasedeficient strain (sapp1/2/3 Ϫ/Ϫ ) and mutant strains that express each SAPP gene individually under the control of a constitutive promoter (pCaTDH3).
In C. albicans, secreted proteins play an important role in morphology and biofilm formation (33)(34)(35)(36). Hence, we first determined the corresponding effects of SAPP proteins in C. parapsilosis. In contrast to C. albicans, C. parapsilosis SAPP proteins do not affect either of these properties. SapII, SapV, and SapVI were previously reported to play a role in tissue adhesion C. albicans in addition to their role in biofilm formation (37). Furthermore, C. albicans Sap1p, Sap2p, Sap3p, and Sap9p were previously reported to be involved in adherence to epithelial cells (10,38,39). In the present study, we showed that Sapp1p, Sapp2p, and Sapp3p in C. parapsilosis also contribute to adhesion, although possibly to differing degrees.
As shown by examining the effect of cell wall-perturbing agents, disruption of the SAPP genes did not affect the mutant strain's survival, indicating that C. parapsilosis aspartyl proteases do not influence the species' fitness and viability.
On the other hand, disruption of SAPP1 and SAPP2 but not SAPP3 resulted in serum sensitivity. These results suggest that only the former two proteases are required for serum survival in this species. This observation is consistent with a previous finding according to which enhanced Sapp1p production was detected in C. parapsilosis cells in the presence of serum albumin (28).
Pathogenic fungi have been previously reported to overcome the fungicidal effects of human serum via actively secreting aspartyl proteases to neutralize proteins with antimicrobial effects (15,40). For instance, complement proteins have diverse functions that include opsonization of microbes to facilitate phagocytosis, activation of cellular responses, initiation of inflammation, and direct lysis of microbial cells (41,42). The protective effects of Sapp1p and Sapp2p mentioned above might be the result of their ability to cleave complement components. Therefore, we further aimed to examine the complement cleavage activity of purified Sapp1p and Sapp2p proteins. Complement component 3 (C3) plays a central role in all three complement pathways. Following its cleavage by C3 convertase, the resulting C3b fragment forms the C5 convertases that are necessary for the progression of the complement cascade. Our results suggest that C. parapsilosis is able to escape such complement-mediated attacks through the activity of its secreted aspartyl proteases, as both Sapp1p and Sapp2p are able to efficiently degrade the active complement C3b and C4b fragments required for convertase functioning and opsonization, similarly to the degradation and thus inactivation in the host mediated by serine protease factor I, a complement control protein (CCC).
FH and FHL-1 inhibit complement activation in the host but also do so when sequestered from serum by pathogenic microbes as an immune escape mechanism. FH and FHL-1 bind to microbial ligands through specific domains that are partially conserved among other members of the FH protein family, i.e., the FHR proteins (43,44). FHRs were also reported previously to be involved in complement cascade regulation, although this is a controversial issue (43,45,46). FHR-1 was reported to inhibit C5 and the terminal pathway, whereas FHR-2 inhibits the alternative pathway and activation of the terminal pathway. FHR-5 displays weak cofactor activity and inhibits the C3 convertase and was recently reported to inhibit C5 conversion (47)(48)(49)(50)(51). On the other hand, FHR-1, FHR-4, and FHR-5 were shown to support alternative pathway activation at the C3 level by binding C3b and allowing the formation of the C3 convertase (52)(53)(54). Thus, FHRs-due to the presence of conserved domains-may competitively inhibit FH/FHL-1 binding to microbes and enhance opsonization (50,55). According to our results, neither FHL-1 nor FHR-1 is cleaved by C. parapsilosis Sapp1p or Sapp2. Furthermore, a difference in substrate preference is also evident, as Sapp2p, but not Sapp1p, is able to cleave FHR-5. The cleavage of FHR-5 but not FHL-1 and FHR-1 suggests that Sapp2p presumably cleaves at locations near complement control protein (CCP) domains 3, 4, 5, 6, and 7, which are absent in FHL-1 and FHR-1 but present in FHR-5 and FH, although further studies are needed to confirm this hypothesis. These data suggest that the secreted aspartyl proteases of this species show a substrate preference for complement proteins involved in activation of the cascade, rather than for complement control proteins (e.g., factor H family proteins).
C. albicans attachment and subsequent colonization are necessary to induce inflammatory responses in epithelial cells (56). Activation of epithelial cells also shapes the responses of monocytes, macrophages, and other immune cells during a fungal infection. Professional antigen-presenting cells such as macrophages connect the innate and adaptive arms of the host's immune responses by processing and presenting antigens to other effector cells and actively eliminating pathogens. Thus, we next examined if disruption of any of the C. parapsilosis SAPP genes would have an effect on macrophage activity. Our results indicate that human PBMC-DMs were able to phagocytose and eliminate sapp1/2/3 Ϫ/Ϫ and RI_SAPP3 cells more efficiently than the wildtype and RI_SAPP1 or RI_SAPP2 strains.
The aspartyl proteases of C. albicans induce proinflammatory cytokine responses to differing degrees. For instance, SapI, SapII, and SapVI significantly induce IL-1␤, TNF-␣, and IL-6 production, while SapIII is able to stimulate IL-1␤ and TNF-␣ secretion (57). Besides inducing low levels of host cell damage, the sapp1/2/3 Ϫ/Ϫ and RI_SAPP3 strains also induced lower levels of proinflammatory cytokines (IL-1␤, IL-6, and IL-8) than the parental and RI_SAPP1 or RI_SAPP2 strains. These results, together with the data gathered from G. mellonella infection (an invertebrate model commonly applied to mimic basic cellular and humoral mammal-like immune responses in vivo [58]), further suggest differences in the contribution of C. parapsilosis Sapp proteins to virulence.
In conclusion, we demonstrated in the present study that C. parapsilosis Sapp proteins did not affect formation of pseudohyphae or biofilm. However, Sapp1p and Sapp2p play roles in adhesion to epithelial cells and in host cell damage and might promote survival within macrophages. Sapp-mediated cleavage of complement proteins also suggests that C. parapsilosis might also interfere with human complement attack. In summary, Sapp1p and Sapp2p, but not Sapp3p, are the major and fully functional aspartyl proteases in C. parapsilosis that actively affect the species' pathogenicity.

MATERIALS AND METHODS
Strains and growth conditions. The strains used in the present study and their abbreviations are listed in Table S1 in the supplemental material. Strains were cultured overnight in YPD broth at 30°C, with shaking. Cells from overnight cultures were collected by centrifugation and washed twice with sterile 1ϫ PBS (phosphate-buffered saline), and the number of cells was adjusted as indicated in descriptions of the respective experiments. For growth assays and gene expression studies, the wild-type and mutant strains were cultivated in YCB (yeast carbon base) medium supplemented with 0.2% BSA (bovine serum albumin) at 30°C. Escherichia coli DH5␣ was grown in LB (Luria-Bertani broth) or on LB plates supplemented with ampicillin (0.1 mg/ml) for plasmid construction and propagation.
Mutant strains expressing the individual SAPP genes were generated using the SAPP1-SAPP2-SAPP3 (sapp1/2/3 Ϫ/Ϫ ) null mutant strain. Solely SAPP1-, SAPP2-, and SAPP3-expressing mutants were established using a replacement cassette targeting the Neut5l locus and containing the respective SAPP open reading frames (ORFs) under the control of the CaTDH3 constitutive promoter. In each case, nourseothricin was used as a selection marker. C. parapsilosis strains were transformed by electroporation as described previously (59). The transformants were confirmed by colony PCR and Southern blot analysis.
Gene expression studies. Total RNA was isolated from C. parapsilosis wild-type cells grown in YCB plus 0.2% medium for 48 h using a RiboPure RNA purification kit according to the manufacturer's instructions. A 500-ng volume of RNA was subjected to reverse transcription using a RevertAid firststrand cDNA synthesis kit according to the protocol provided by the manufacturer. Quantitative PCR (qPCR) was performed using the primers listed in Table S2. The amplification conditions were as follows: one cycle of denaturation for 3 min at 95°C; denaturation at 95°C for 10 s; 49 cycles of annealing at 60°C for 30 s, and elongation at 65°C for 30 s; with a final extension step at 72°C for 30 s. TUB4 was used as an internal control.