Candida albicans END3 Mediates Endocytosis and Has Subsequent Roles in Cell Wall Integrity, Morphological Switching, and Tissue Invasion.

The role of endocytosis in Candida albicans secretion, filamentation, and virulence remains poorly understood, despite its importance as a fundamental component of intracellular trafficking. Given that secretory mutants display defects in endocytosis, we have focused our attention on endocytic mutants to understand the interconnection between endocytosis and other secretory pathways. Using a reverse-genetic approach based upon CRISPR-Cas9 mediated gene deletion, we studied the functions of the gene END3, which plays a key role in clathrin-based endocytosis. In the end3Δ/Δ null mutant, clathrin-mediated endocytosis was substantially reduced. While in vitro growth, cell morphology, and vacuoles appeared normal, the mutant was impaired in actin patch formation, filamentous growth, biofilm formation, cell wall integrity, and extracellular protease secretion. In addition, susceptibility to various antifungal agents was altered. Consistent with the inability to form hyphae, in an in vitro keratinocyte infection model, the null mutant displayed reduced damage of mammalian adhesion zippers and host cell death. Thus, C. albicans END3 has a role in efficient endocytosis that is required for cell wall integrity, protein secretion, hyphal formation, and virulence-related processes. These findings suggest that impaired endocytosis subsequently affects other secretory pathways, providing evidence of the interconnection between these processes. IMPORTANCE Candida albicans is a fungal commensal organism that can cause serious opportunistic infections in immunocompromised patients leading to substantial complications and mortality. A better understanding of the microbe's biology to develop more effective therapeutic and diagnostic tools is required as invasive candidiasis is a problem of continued clinical importance. This study focuses on endocytosis, an important but incompletely understood cellular mechanism needed to uptake nutrients and communicate with a cell's environment. In this study, we have assessed the role of endocytosis in cell wall integrity, biofilm formation, and tissue invasion in C. albicans. These findings will improve our understanding of cellular mechanisms underlying endocytosis and will inform us of the interconnection with other intracellular transport processes.

INTRODUCTION

Endocytosis is an essential component of the secretory pathway in eukaryotic cells and is needed for many aspects of cellular functioning, such as uptake of nutrients, extracellular signaling, and for the regulation of plasma membrane as well as cell wall composition. This highly orchestrated process begins at the cell wall and plasma membrane with vesicle formation that transports cargo into the cell. The best-characterized and most selective endocytic process is clathrin-based endocytosis (CME) which allows receptor-mediated cargo to enter the cell in an energy dependent manner (1, 2).

CME has been extensively studied in Saccharomyces cerevisiae (Sc) and requires more than 50 proteins (reviewed in Goode et al., 2015 [1]). The early-coat proteins in this process define endocytic-site location and cargo recruitment (3) while the middle and late-coat proteins induce membrane invagination and facilitate vesicle scission, in part by inducing actin assembly (4). The early-coat proteins include clathrin, the AP-2 adaptor protein, Ede1p, Syp1p, and Pal1p. Ede1p, an Eps15 homology (EH) domain protein, is one of the earliest proteins recruited to endocytic sites and facilitates site initiation and stabilization (5). While Ede1p and Syp1p remain associated with the plasma membrane, clathrin, AP-2p, and Pal1p are internalized along with the clathrin coat (6). The recruitment of Sla2 plays a key role in the transition to form the middle coat. Its ANTH domain binds to the ENTH domain of the epsins Ent1p and Ent2p. These epsins facilitate membrane bending via interactions with the polarity-establishing GTPase Cdc42p in the presence of phosphotidylinositol 4,5-bisphosphate (PIP2) (7–9). The middle-coat proteins recruit the central late-coat proteins setting the stage for actin assembly that facilitates membrane invagination and vesicle scission (10). Key late-coat proteins include Pan1p, Sla1p, Prk1p, End3p, and Las17p. Their assembly is concluded when Las17p is recruited to endocytic sites partially through its interaction with Sla1p (11). The late coat assembly provides the foundation for subsequent membrane invagination and endocytosis. Pan1p is the only endocytic protein essential for viability in this pathway and provides a major contribution to actin cytoskeletal organization by binding Arp2/3p, F-actin, and type-I myosin motors with its EH domains (12–14).

The 40-kDa protein Sc End3p plays a central role in the coat assembly process and facilitates the interaction with the actin cytoskeleton during CME (15). Sc End3p has two EH domains at the N-terminus that provide a binding site for Sc Sla1p (14). The EH1 domain also includes a putative EF-hand Ca(2+)-binding site and part of a consensus sequence for the binding of PIP2 (16). Neither of these two sites are found to be required for End3p function in Sc (15). The C-terminal end has a coiled-coil structure known to form a central complex with Sc Pal1p and facilitate localization to cortical actin patches and Pan1p phosphorylation (14).The tandem repeats of the coiled-coil region appear to act independently of one another and provide redundant binding sites for Pan1. Mutations of the End3 repeats revealed that the amino acid residues that are conserved between these repeats are the actual binding sites for Pan1p (14). End3p and Pan1p are likely recruited to endocytic sites via interactions with Sla2p and Ent1p. Pan1p has been found to bind Sla2p and Ent1p in vitro (10, 17, 18) The Pan1-End3p complex is connected with Sla1p through interactions with the End3 N-terminus (13). End3p plays the crucial role in this recruitment as Pan1p is redundant in this aspect (13, 19).

Even though Sc End3p is important for efficient CME based internalization of endocytic vesicles (15), the end3Δ mutant is viable and exhibits a normal axial budding pattern, but cytokinesis is inhibited. Calcofluor White (CW) staining of Sc end3Δ reveals distorted chitin and cell wall distribution. Cell size and actin organization are altered in the Sc end3Δ mutant with fewer cortical actin patches (15). These phenotypic characteristics are tightly linked to complex formation with Pan1p and Sla1p because preventing these interactions, as well as the deletion of any of these genes, results in similar phenotypes (20).

In this study we have characterized the End3 homolog in Candida albicans (Ca) and have identified similar phenotypes similar to that seen in the Sc end3 mutant. While endocytosis and cell wall integrity defects were very prominent in the Ca end3Δ/Δ deletion strain, cell morphology appeared to be normal. Moreover, the subsequent defects in endocytosis also produce alterations in characteristics specific to C. albicans such as filamentation and other virulence-related attributes. Our findings support the concept of existing biological intersections between endocytosis and other secretory pathways.

RESULTS

Structural characterization and identification of C. albicans END3 as the ortholog of S. cerevisiae END3.

A search of the C. albicans genome database for orthologs of Sc END3 revealed a 1,188 bp intron-less open reading frame (ORF) (Ca orf19.1711) whose deduced protein product is predicted to be a 395-aa protein having overall 36% identity and 52% similarity to Sc End3p (Fig. 1). The homology is strongest in two regions in the N-terminal region identified by Simple Modular Architecture Research Tool (SMART) (https://smart.embl.de/) as having similarity to the protein-protein interaction modules called epsin homology (EH) domains. Comparing Ca and Sc End3p, EH domain 1 has 60% identity and 73% similarity, whereas EH2 is less well conserved. The first EH domain also has a weak match to consensus EFhand (EFh), a calcium binding motif found in both Sc (15) and Ca. These motifs are generally found in pairs and utilize the oxygens of the conserved glutamate or aspartate residues for binding calcium. Only a single EFh is discernible in End3p in either Ca or Sc. Sc also has a putative PIP2 binding site (15), the sequence of which is partially conserved (50% similarity) in Ca.

FIG 1

C. albicans End3p bears strong homology to S. cerevisiae End3p. The alignment of S. cerevisiae End3p with C. albicans End3p is shown. Identical amino acid residues are shown in black, similar residues in blue, and mismatched residues in red. The black boxes show the conserved epsin homology domains (EH). The green box encloses a putative EFh domain, with the invariant acidic residue marked by the symbol (#). The purple box outlines putative PIP2 binding domain in S. cerevisiae (7). C-terminal coiled coil domains of Ca align with Sc repeats marked in blue. Asterisks below the Sc sequence indicate the residues in S. cerevisiae that were found to bind Pan1 (8).

The carboxyl half of the Ca End3p has regions of low complexity and coiled coils, the latter are predicted to be involved with protein-protein interactions. There is some similarity (up to 62% in the second repeat) between the species that align with regions identified in Sc as two C-terminal repeats (Fig. 1) shown to be essential for End3p function (15) and involved in binding to Pan1p (14) In Sc, the repeats are found to be redundant with respect to their ability to bind to Pan1p (14). The residues that were identified in Sc to participate in binding to Pan1p are only partially conserved in Ca.

Construction of a C. albicans end3Δ/Δ mutant.

To determine the function of C. albicans END3 we used a reverse genetics approach to analyze gene function. The CRISPR-Cas9 method was used to generate a C. albicans end3Δ/Δ null mutant (KO) and corresponding complemented, or “knock-in” (KI) strain, where both copies of the wild-type allele were reintroduced into the end3Δ/Δ null mutant (23). Correct strain construction was confirmed by Southern blotting and PCR (sFig.1).

Contribution of END3 to growth, viability, and cell morphology.

The end3Δ/Δ null mutant (KO) grows normally compared to wild-type and the complemented control strain (wt, KI; Fig. 2A and B). This finding was confirmed on yeast extract peptone dextrose (YPD) plates (Fig. 2A, 30°C) and in a growth curve in liquid media (Fig. 2B). The doubling times of KO, KI, and wt were 3.6 ±0.24, 3.6 ±0.32 and 3.4 ±0.18 h, respectively. Next, we detected increased temperature sensitivity at 42°C in the KO compared with controls (Fig. 2A). The end3Δ/Δ mutant’s cell and vacuolar morphology appeared normal when cells were observed via differential interference contrast (DIC) microscopy (Fig. 2C). Cells were stained with CW and visualized using fluorescence microscopy which revealed that the end3Δ/Δ null mutant had an enhanced peripheral stain which suggests altered cell wall chitin deposition in the mutant (Fig. 2C and D). These findings indicate that END3 has a role in cell wall integrity.

FIG 2

The end3Δ/Δ null mutant strain grows comparably to control strains. The end3Δ/Δ null mutant strain (KO), wild type (wt), and knock in strain (KI) were depicted on YPD plates incubated for 48 h at (A) 30°C to 42°C, and visualized (B) in a growth curve, (C) by DIC and fluorescence microscopy with Calcofluor White (CW) staining, and (D) in a quantitative analysis for CW staining. The bar is 5 μm. Results shown in (A and C) are a representative of three different experiments. Error bars in (B, D) are the standard deviation of three different data sets. Each individual data point was run in triplicate and was averaged. Statistical significance in (D) between control strains (wt, KI) and KO was determined with the Student’s t test (wt/ko P = 0.0008; KI/KO P = 0.001).

Contribution of END3 to stress tolerance.

We next characterized the growth of the end3Δ/Δ mutant (KO) in response to cell wall and osmolar stress as well as antifungal agents to characterize the functionality of the cell wall and membrane. When the strains were grown on YPD plates for 48 h at 30°C containing the cell walls stressors SDS (0.05%), CW (50 μg/mL), or Congo red (140 μg/mL), the C. albicans end3Δ/Δ mutant demonstrated a dramatic growth defect in comparison to control strains (Fig. 3A). Next, strains were assessed in conditions of high osmolar and/or cation stress (500 mM NaCl, KCl, and 100 mM LiCl) on YPD plates for 48 h at 30°C. In conditions with high NaCl concentrations, the end3Δ/Δ mutant grew slower when compared to controls strains, and not at all when exposed to LiCl (Fig. 3B). In contrast, on YPD with high KCl concentrations, there was no growth defect.

FIG 3

The end3Δ/Δ mutant strain has a decreased tolerance to cell wall and ion stress. end3Δ/Δ null mutant (KO), wild type (wt), and knock in (KI) strains were grown for 48 h on YPD plates containing (A) cell wall stressors SDS (0.05%), Calcofluor White (50 μg/mL), and Congo red (140 μg/mL), and (B) cationic cell stress conditions including 500 mM KCl, 500 mM NaCl, or100 mM LiCl. Experiments shown in are a representative of three different experiments.

We then analyzed growth on plates with YPD media containing antifungal agents to address whether the defects in cell wall integrity and response to osmotic stress correlated with altered sensitivity to antifungal agents. Three commonly used antifungal drugs were selected, fluconazole, caspofungin, and amphotericin B. Fluconazole function requires cellular internalization, while caspofungin and amphotericin B are active directly on the cell wall and membrane, respectively. When grown on YPD plates for 48 h at 30°C containing different concentrations of caspofungin and amphotericin B, the C. albicans end3Δ/Δ mutant exhibited a substantial growth defect at all concentrations (0.025 to 0.1 μg/mL) and the highest concentration (1.0 μg/mL) tested, respectively (Fig. 4A and B). Conversely, in the presence of fluconazole, the end3Δ/Δ mutant was less sensitive to the drug than wt and KI (Fig. 4C). While control strains exhibited a dramatic growth defect at 2 and 4 μg/mL fluconazole, growth of the KO strain was not affected. The results suggest that a compromised cell wall and membrane renders the KO more vulnerable to caspofungin and amphotericin B. The resistance of the KO to fluconazole may be related to defects in endocytosis and warrants further mechanistic investigation.

FIG 4

The end3Δ/Δ mutant strain has increased susceptibility to amphotericin B and caspofungin and reduced susceptibility to fluconazole. The end3Δ/Δ null mutant (KO), wild type (wt) and knock in (KI) strains were grown for 48 h on YPD plates containing antifungal drugs (A) amphotericin B at concentrations of 0.11 μg/mL, 0.33 μg/mL, 1.0 μg/mL; (B) caspofungin at concentrations of 0.025 μg/mL, 0.05 μg/mL, 0.1 μg/mL; and (C) fluconazole at concentrations of 1.0 μg/mL, 2.0 μg/mL, 4.0 μg/mL. Experiments shown in (A to C) are a representative of three different experiments.

The C. albicans end3Δ/Δ mutant is defective in endocytosis and actin patch formation.

In order to define the role of END3 in C. albicans endocytosis, we used the lipophilic fluorescent dye FM4-64 and microscopy, as described in (42) with some modifications. FM4-64 initially binds the plasma membrane and is then actively internalized by clathrin-based endocytosis followed by transport to the vacuole via endocytic intermediates. Mutants with defects in endocytosis exhibit a time delay in FM4-64 endocytosis and therefore staining of the vacuole and other intracellular membranes.

To reduce the binding of FM4-64 in the cell periphery of the end3Δ/Δ mutant (KO) during the endocytosis assay, we followed a protocol described in (43) and incubated the cells with FM4-64 on ice. Visualization with fluorescence microscopy demonstrated that the end3Δ/Δ mutant strain did not accumulate FM4-64 in the vacuolar membrane compared to control strains (Fig. 5A) under these conditions. We occasionally observed a very bright diffuse signal of FM4-64 that was comparable with the result when cells were incubated on ice, a condition that inhibits endocytosis (Fig. 5A, Fig. S2). Although vacuoles in the end3Δ/Δ mutant appeared normal, there was clearly a time delay in vacuolar staining when compared to control strains indicating a marked, but likely not complete, defect in endocytosis. Our quantitative analysis of vacuolar staining confirmed the microscopy data (Fig. 5B).

FIG 5

The end3Δ/Δ mutant strain has delayed endocytosis and altered actin patch localization. Fluorescent images of the end3Δ/Δ null mutant (KO) and control strains, wild type (wt) and knock in (KI) strains stained with the lipophilic fluorescent dye FM4-64 are shown after incubation of 10, 20, and 30 min at RT (A), and with rhodamine-phalloidin to label actin (C). (B) represents a quantitative analysis of cells with vacuolar staining and (D) depicts the statistical analysis of the number of actin patches. Experiments shown in (A) and (C) are a representative of three different experiments. The bars are 5 μm. Error bars in (B) and (D) are the standard deviation of three different experiments. Each individual data point was run in triplicate and was averaged.

Because End3p’s putative interaction partner Pan1p is a major contributor to actin organization and recruitment during CME we tested whether actin organization is altered in the end3Δ/Δ mutant. Rhodamine phalloidin staining revealed fewer actin patches in the end3Δ/Δ mutant when compared with controls (Fig. 5C and D). We determined that wild-type (wt) and the reintegrant strain (KI) had an average of 5.3 ±0.3 and 4.7 ±0.25 patches/cell, respectively while end3Δ/Δ revealed only 1.5 ±0.25 actin patches/cell.

The C. albicans end3Δ/Δ mutant is defective in filamentation.

Hyphal formation requires efficient endocytosis and a proper cell wall composition. When we tested the ability of the end3Δ/Δ mutant to form hyphae compared with controls, the null mutant displayed a reduction in filamentation when grown on solid filamentation media such as RPMI 1640 and M199 (Fig. 6A). When the strains were grown on YPD with 10% fetal calf serum (FCS) or Spider medium, some filamentous growth was seen in end3Δ/Δ on the edges of the spotted colony similar to controls. When we grew the strains in liquid RPMI 1640 media in strong inducing conditions, we saw a delay in filamentation in the end3Δ/Δ mutant, with a predominance of pseudo-hyphal growth while the controls showed true hyphal formation (Fig. 6B).

FIG 6

The end3Δ/Δ mutant is defective in filamentation and forms an aberrant biofilm. The end3Δ/Δ null mutant (KO), wild type (wt) and knock in (KI) strains were assayed (A) on agar plates grown for 72 h at 37°C with different media known to induce hyphal growth. In (B), these strains were grown in liquid RPMI 1640 media overnight and were visualized by DIC and fluorescence microscopy after staining with Calcofluor White. The bar indicates 5 μm. (C) Biofilms were grown on plastic surfaces with RPMI 1640 media for 48 h and assayed for biofilm metabolic activity. (D) These biofilms were visualized by light microscopy using a DIC filter. The bar indicates 50 μm. Experiments shown in (A), (B), and (C) are a representative of three different experiments. Error bars in (D) are the standard deviation of three different data sets. Each individual data point was run in triplicate and was averaged. Statistical significance in (C) between control strains (wt, KI) and KO was determined with the Student's t test (wt/KO P = 0.0002; KI/KO P = 0.0002).

To test whether the inability to form true hyphae influenced biofilm formation in the end3Δ/Δ mutant, we performed detailed studies of biofilm formation using the XTT-reduction assay and light microscopy (Fig. 6C, D). The end3Δ/Δ mutant formed a sparse, patchy, poorly adherent biofilm, which lifted easily with minimal disturbance, unlike the wild-type and complemented control (Fig. 6C). Biofilm metabolic activity was reduced by 65% compared with controls (Fig. 6D).

The C. albicans end3Δ/Δ mutant is defective in secretion of extracellular proteases.

Secretion of conditionally regulated proteins, such as invertase (which is secreted in response to environmental conditions), and trafficking of carboxypeptidase Y to the vacuole is normal in the S. cerevisiae end3Δ/Δ mutant (44). However, it is not known whether there are any secretory defects in the C. albicans end3Δ/Δ null mutant. Because the Saps are the most extensively studied C. albicans secretory proteins (45), we compared the abilities of the C. albicans end3Δ/Δ mutant to secrete catalytically-active proteases out of the cell by using bovine serum albumin (BSA) plate assays for proteinase secretion. The diameter of an opaque zone around the spotted colony provides an indirect assay of aspartyl protease secretion. In comparison to the control strains (wt, KI), the end3Δ/Δ mutant had a smaller zone of BSA degradation. This finding suggests reduced secretion of secreted aspartyl proteases in end3Δ/Δ when compared with the controls (Fig. 7C). To further explore this finding, we then assessed the mRNA expression level of Sap4 to 6 in the end3Δ/Δ mutant and the control strains under conditions of hyphal growth. Our findings revealed a marked decrease of Sap4 to 6 mRNA expression in the null mutant when compared with wt and KI suggesting that END3 contributes to secretion of Sap4 to 6.

FIG 7

The dissolution of cell–cell adhesions in human VK-2 cells is impaired in the presence of C. albicans end3Δ/Δ. Human VK-2 cells infected with the end3Δ/Δ null mutant (KO), wild type (wt) and knock in (KI) strains for 6 h and 24 h were visualized (A) by fluorescence microscopy using an antibody against E-Cadherin and DAPI for labeling the nucleus, and (B) were tested by Western blotting for the proteins E-cadherin and Tubulin (loading control). (C) These C. albicans strains were further tested in a protease agar plate assay for their ability to lyse BSA after an overnight incubation at 30°C. (D) RNA expression levels of Sap4 to 6 are depicted in end3Δ/Δ and control strains after growth in hyphal inducing conditions for 4 h. Experiments shown in (A to D) are a representative of at least three different experiments. The bar in (A) is 10 μm. Statistical significance in (B) between control strains (wt, KI) and KO was determined with the Student’s t test (wt/KO P = 0.014; KI/KO P = 0.019).

Dissolution of cell–cell adhesions is impaired in the C. albicans end3Δ/Δ mutant.

The destruction of tight junctions between host cells is a hallmark of C. albicans virulence and is facilitated by the proteases Sap4 to 6 degrading the adhesion molecule E-cadherin (46, 47). To assess the capacity of the end3Δ/Δ mutant to compromise tight junctions, we infected the vaginal human keratinocyte cell line VK-2 with KO and the control strains (wt, KI) for 6 h and 24 h, and evaluated the cells over time for the presence of E-cadherin. Co-culture with C. albicans KO, wt, or KI after 6 h of incubation revealed E-cadherin labeled puncta at tight junctions between VK-2 cells. After 24 h, no staining was observed using the anti-E-cadherin antibody in VK-2 cells cocultured with wt and KI (Fig. 7A). In contrast, in VK2-cells exposed to the KO strain, the E-cadherin labeled tight junctions remained intact (Fig. 7A).

The presence of E-cadherin in VK-2 cells during coculture with C. albicans was analyzed by Western blotting (Fig. 7B). While at 6 h of incubation, E-cadherin levels were similar between cocultures, by 24 h, E-cadherin was undetectable in cells cocultured with wt and KI strains. In contrast, the E-cadherin level remained stable over time in VK-2 cells incubated with the KO strain (Fig. 7B).

Taken together, these findings suggest reduced secreted protease activity in the end3Δ/Δ mutant when compared with control strains, resulting in decreased virulence in this in vitro model, as measured by the ability to disrupt cell–cell adhesion of vaginal keratinocytes.

The C. albicans end3Δ/Δ mutant lacks the ability to kill host cells.

To define the ability of the end3Δ/Δ strain to compromise host cell viability, we tested the strain in a host cell killing assay. Using a microplate Live/Dead assay (Invitrogen, Waltham, MA) as described in (28), we evaluated human VK-2 cell growth after 6 h and 24 h of infection with the end3Δ/Δ mutant and the control strains KI and wt. After 6 h of infection, 95% of VK-2 cells were alive in cultures with KO and the control strains (Fig. 8A). After 24 h, the KO killed about 25% of VK-2 cells, while wt and KI strains destroyed 80% of host cells (Fig. 8B). This finding underscores the substantially reduced virulence of the end3Δ/Δ strain, when assayed in vitro.

FIG 8

The C. albicans end3Δ/Δ mutant has a reduced ability to kill human VK-2 cells in vitro. VK-2 cells infected with the end3Δ/Δ null mutant (KO), wild type (wt), and knock in (KI) strains for (A) 6 h and (B) 24 h were tested in a LIVE/DEAD assay. Error bars in (A) and (B) are the standard deviation of three different data sets. Each individual data point was run in triplicate and was averaged. Statistical significance between control strains (wt, KI) and KO was determined with the Student’s t test (wt/KO P = 0.0003; KI/KO P = 0.0003).

DISCUSSION

In prior work on secretion in C. albicans, we noted that several pre-vacuolar and other secretory mutants had defects not only in exocytic pathways (2, 24, 25, 48, 49), but also appeared to have defects in endocytosis as well (46, 47). These findings suggested an intersection between these secretory pathways. To further understand this interconnection, we have now turned our focus to analyzing key components of clathrin-based endocytosis, a pathway that remains under-studied in C. albicans. A few genes involved in endocytosis such as EDE1, SLA2, and SLA1 have been characterized in C. albicans, and we have recently analyzed the early endocytosis gene ENT2 (28). Null mutants of Ca SLA2, SLA1, and ENT2 are defective in clathrin-mediated endocytosis whereas the null mutant of Ca EDE1 does not appear to have this defect (50).

Here we have identified Ca END3 to be a non-essential gene. The corresponding null mutant is defective in efficient endocytosis and actin patch formation. While this finding is consistent with the (Sc) end3Δ null mutant there are some substantial morphological and growth differences. Unlike in the Sc end3Δ mutant, we were unable to identify a bipolar budding pattern, an abnormal enlarged cell shape, and altered vacuolar morphology in the Ca end3Δ/Δ mutant. On the other hand, the null mutant of the putative End3p binding partner Sla1p, has a reduced growth rate and forms enlarged, globular mother cells due to delays in cell cycle progression. Our recent studies of the null mutant of Ca ENT2, which encodes another potential interacting partner of End3p, revealed a slight growth defect but a normal cell shape and vacuolar morphology of the yeast form. These variations are likely a result of different functions in CME.

Consistent with Sc, the Ca end3Δ/Δ mutant has a strong defect in cell wall morphology and function (15). We observed substantial impairments in cell wall integrity when challenged by a variety of cell wall stressors and antifungal agents that impact the cell wall directly (caspofungin, amphotericin B; Fig. 5A and B). We also detected increased peripheral staining in the end3Δ/Δ mutant using dyes such as CW and FM4 to 64 (Fig. 2 and 5A, B, respectively). Because CW binds chitin, this result suggests that there is increased chitin content of the cell wall, potentially as a compensatory response to other defects in cell wall composition or architecture. Why enhanced initial binding of FM4-64, which attaches via membrane insertion, was detected on the cell surface of the end3Δ/Δ mutant is not clear and warrants further investigation.

In Sc, the Pan1-End3p complex binds to Sla2p and Ent1/2p which couples the complex to the clathrin coat. Consistent with this, we identified phenotypic similarities between the Ca end3Δ/Δ and Ca ent2Δ/Δ mutants regarding cell wall integrity, hyphal formation, and endocytosis defects. Another group has found that the Ca sla2Δ/Δ mutant is defective in all polarized secretion including CME and shows mislocalization of lipid rafts that are usually found at the hyphal tip. They further identified mislocalization of Rvs167p, a lipid raft-associated protein, and the cell wall remodeling enzyme Gsc1p, which is a b-1,3-glucan synthase (51). The actin binding Arp2/3 complex that interacts with the PalI-End3p complex is also required for cell wall integrity (38). These findings suggest that defects in cell wall integrity may be a common feature in endocytosis mutants.

Improperly functioning endocytic transport in the end3Δ/Δ mutant has consequences for the secretion of extracellular components. As C. albicans proteases Sap4 to 6 are responsible for degrading E-cadherin in host cells (52), the inability of the end3Δ/Δ mutant to break down E-cadherin in a vaginal epithelial cell line, and reduced extracellular BSA degradation (Fig. 7A to C) suggests that their secretion is reduced in the end3Δ/Δ strain. Our studies of Ca ent2Δ/Δ revealed a deficiency in secretion of extracellular proteases as well. These findings provide direct and indirect evidence for impairments in cellular trafficking other than endocytosis and suggest intersection of these pathways.

We also found that the end3Δ/Δ mutant is impaired in proper hyphal formation and only produced a weak biofilm (Fig. A to D), which is likely related to the dramatic changes of the cell wall and the need for massive secretion of extracellular components during these processes. The end3Δ/Δ mutant in strong hyphal-inducing conditions had a predominantly pseudo-hyphal appearance. Similar findings were seen in binding partners of End3p. For instance, the Ca sla1Δ/Δ mutant produced shorter filaments in hyphal-inducing conditions (37) and the Ca ent2Δ/Δ mutant was unable to form hyphae and biofilms all together (28). Interestingly, while the Ca actin-binding Arp2/3p complex is required for hyphal and biofilm formation it does not have an essential function for endocytosis (38, 39).

The impaired hyphal growth, cell wall integrity, and transport defects in the end3Δ/Δ mutant were associated with a greatly reduced virulence-related phenotypes in our in vitro assays. The Live/Dead assay revealed only a limited ability of the end3Δ/Δ mutant to kill host cells, a characteristic that was found in Ca ent2Δ/Δ as well (28). The pseudo-hyphal growth and the limited ability to secrete proteases that degrade host cell E-cadherin provide a mechanistic explanation for this greatly reduced ability to damage host cells in this in vitro model.

Our study suggests that C. albicans End3p provides a connection between endocytosis and other intracellular transport pathways, as well as a contribution to cell wall integrity and hyphal formation. It demonstrates that an impairment in endocytosis leads to serious implications for cell function and virulence-related properties in C. albicans. These roles are partially shared with the late coat proteins, Ent2p and Sla1p, suggesting that late coat assembly is central for a functional endocytic process. Further investigations of the role of CME in virulence including in vivo studies, are therefore warranted.

MATERIALS AND METHODS

Identification of C. albicans END3.

The predicted protein and DNA sequences for Sc and Ca END3 were retrieved from the Saccharomyces Genome Database (http://www.yeastgenome.org) and Candida Genome Database (http://www.candidagenome.org), respectively. A single potential ortholog to Sc End3p was identified by a BLASTp search of the Candida Genome Database utilizing the functions on these websites. Alignments of the protein sequences was done with the SnapGene (SnapGene.com) protein align function, using the Smith-Waterman method (21). The web resource, SMART, available at https://smart.embl.de, (22) was utilized for the identification and annotation of protein domains.

Deletion of C. albicans END3 and Southern blotting.

Table 1 includes the list of oligonucleotides and primers used for gene deletion and Southern blotting experiments. We generated a C. albicans end3Δ/Δ null mutant using the CRISPR-Cas9 protocol developed by the Hernday laboratory (23). Alleles of END3 were then re-integrated via homologous recombination into the end3Δ/Δ null mutant to generate a complemented KI strain. Correct strain construction was demonstrated by Southern blotting. Briefly, 5 μg genomic DNA was digested with SphI and SpeI overnight and loaded onto a 1% agarose gel. The blot was performed using standard protocols. A DNA probe was generated by PCR using DIG-labeled dNTP (Roche, Basel, Switzerland). Detection of the probe was visualized with a DIG Luminescent Detection Kit (Roche, Basel, Switzerland).

TABLE 1

Strain table

Name	Description/ genotype	Reference
AHY940	SC5314 LEU2 heterozygous knockout a/α leu2Δ/LEU2	(23)
CREnd3KO	SC5314 end3 homozygous knockout a/α end3Δ/end3Δ	This study
CREnd32KI	SC5314 end3 homozygous knock in a/α end3Δ/end3Δ END3/END3	This study

Preparation of genomic DNA and plasmid isolation.

Genomic DNA was extracted from yeast cells using the MasterPure Yeast DNA purification kit (Epicentre Biotechnologies, Madison, WI) following to manufacturer’s instructions. Competent Escherichia coli DH5α cells (Invitrogen, Waltham, MA) were used to generate and maintain plasmids. E. coli cells carrying plasmids of interest were grown in LB medium (1% tryptone, 0.5% glucose, and 1% NaCl) with 100 μg/mL ampicillin at 37°C. Plasmid DNA was extracted from overnight cultures using the Qiagen Plasmid Miniprep system (Qiagen, Germantown, MD) according to the manufacturer’s instructions. Plasmids used and generated in this study are listed in Table 2.

TABLE 2

Plasmids

Name	Description	Reference
pADH137	C.alb LEUpOUT CAS9 expression plasmid	(23)
pADH118	C.alb LEUpOUT “blank” gRNA plasmid	(23)
pCREnt2	C.alb LEUpOUT “blank” gRNA plasmid with End3 guide RNA	This study

Strains, media, and cell culture.

The C. albicans strains utilized in this study are listed in Table 3. YPD (1% yeast extract, 2% peptone, 2% glucose) or (yeast nitrogen base [YNB], 6.7g, plus amino acids, and 2% glucose) were used as a standard growth media. Unless otherwise stated, liquid cell cultures were established by incubation at 30°C with shaking at 250 rpm.

TABLE 3

Primer

Name	Sequence	Reference
Hu EcadherintandF	gtc tgt agg aag gca cag cc	(40)
Hu EcadherintndR	tgc aac gtc gtt acg agt ca	(40)
Hu GAPDHF	gga cct gac ctg ccg tct a	(41)
Hu GAPDHR	tgc tgt agc caa att cgt tg	(41)
End3 inner F	cca atg tcc cac aag aat tac c	This study
End3 inner R	ttg atg gtt ttg ctt gag cc	This study
End3 outer F/donor KI	ata aag gtc gtg tgc ttt gtg t	This study
End3 outer R/donor KI	gtt ctg gat gtt act gct ga	This study
End3 probe Southern F	gac tca ctt act tgc cgc ct	This study
End3 probe Southern R	cgt gac ta caca aag cac acg	This study
End3 KO donor U.90	aat tat tta tac atc ttc gtc tct tcc cta ttt acg aaa aat gcg aat ttt taa atc caa aga atc aag aat tat taa aat taa atc aac	This study
End3 KO donor L.90	gtt gat tta att tta ata att ctt gat tct ttg tag tta aaa att cgc att ttt cgt aaa tag gga aga gac gaa gat gta taa ata att	This study
End3 guide RNA oligo	cgt aaa cta ttt tta att tga ttt ctg att ctt cta aac ggt ttt aga gct aga aat agc	This study
AHO1098	caaattaaaaatagtttacgcaag	(23)
AHO1099	gtt tta gag cta gaa atg caa gtt	(23)

VK-2/E6E7 (ATCC CRL-2616, Manassas, VA), a human vaginal keratinocyte cell line, was grown in Keratinocyte-SFM media (Invitrogen, Waltham, MA) in a cell culture incubator at 37°C with 5% C02. Following the manufacturer’s instructions, cells were split once a week and fed every 2 days.

Cell growth assay and agar plate assays for response to environmental stress and filamentation.

To assay growth curves, cells from an overnight culture in YPD at 30°C were counted and diluted in YPD to a cell concentration of 1 × 106 cells/mL. Then, 100 μL of this cell dilution were placed in 96-well plates, in quadruplicate. Cell growth was determined over 21 h at 30°C using a microplate reader (BioTek Synergy H1, Winooski, VT) that recorded the extinction at OD 600 nm every 30 min.

Growth rates were assessed on agar plate assays as described (24, 25). Five μL of cells from 1:5 serial dilutions were spotted onto agar plates. Growth was assayed after 48 h in response to 30°C, 37°C, and 42°C, cell wall stressors (0.2% SDS in YPD, 140 μg/mL Congo red in YNB, 50 μg/mL CW in YNB, all from Sigma-Aldrich, St. Louis, MO), potassium chloride, lithium chloride, and sodium chloride (all Sigma-Aldrich, St. Louis, MO) in YPD, BSA with YNB (26) as well as antifungal drugs (amphotericin B, caspofungin, fluconazole, all Sigma-Aldrich, St. Louis, MO, in YPD). Filamentation was determined after 72 h at 37°C on solid YPD with 10% (vol/vol) fetal calf serum (FCS, Fisher Scientific, Waltham, MA), Medium 199 supplemented with l-glutamine (M199, Sigma-Aldrich, St. Louis, MO), RPMI 1640 pH 7.0 (Fisher Scientific, Waltham, MA), and Spider medium, as described previously (27).

Analyses of endocytosis and fluorescence-activated cell sorting with FM4-64.

The lipophilic membrane dye FM4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenylhexatrienyl) pyridiniumdibromide] (EMD Millipore, Temecula, CA), which is actively endocytosed, was used to assay membrane-related endocytosis as described (28). Briefly, FM4-64 at a final concentration of 2 mM was added and incubated on ice for 20 min to ice cold strains that have been grown to a mid-log phase. The cells were washed in ice cold PBS and were incubated at room temperature for 10, 20, and 30 min. Membrane transport at each time point was stopped with ice cold 12 mM sodium azide (Sigma-Aldrich, St. Louis, MO). Staining of the vacuoles was visualized using a Zeiss Axio Imager M1 fluorescence microscope. Images were taken using standard Texas Red filters for FM4-64 and analyzed for the ability to stain the vacuolar membrane.

Actin staining.

Strains were grown to exponential phase in YPD and fixed in 3% paraformaldehyde (Sigma-Aldrich, St. Louis, MO) for 90 min. Actin staining was performed with rhodamine phalloidin after permeabilization with Triton X-100 using the protocol described in (29). For the analysis, the number of actin patches of each individual cell was counted in approximately 100 cells per strain. The total number of actin patches was determined and averaged by the counted cell number. The standard deviation was calculated from the average of the three experiments for each strain.

Analysis of adherence and biofilm formation in C. albicans endocytosis mutants.

Biofilms were visualized via DIC light microscopy on a Nikon Eclipse Ti inverted microscope, and metabolic activity was measured by XTT-reduction, in triplicate, according to published methods (30). Briefly, the XTT-reduction assay was performed after 48 h of biofilm growth at 37°C. Films were washed and incubated with the XTT-menadione substrate at 37°C for 3 h. Supernatant was transferred to a fresh 96-well plate, and absorbance was read at 490 nm in a microplate reader (BioTek Synergy H1, Winooski, VT). Experiments were performed in triplicate, each with eight replicates per strain. Statistical significance was assessed using the Student’s t test in Microsoft Excel.

Immunofluorescence microscopy of C. albicans strains.

Cell morphology of yeast and hyphal forms was assessed via fluorescence microscopy as described (28, 31). Cells from an overnight culture in YPD at 30°C (yeast growth) or 37°C in RPMI 1640 (filamentous growth) were mixed 20:1 with CW (Sigma-Aldrich, St. Louis, MO), placed on a slide and visualized with a 63× lens and DAPI filter of a Nikon Eclipse 80i microscope.

Quantitative plate assay of C. albicans strains CW staining.

For the quantitative plate assay, 1.0 × 106 cells from each strain were mixed 20:1 with CW and incubated for 30 min. Cells were washed six times with ice cold PBS, resuspended in 300 μL and placed in triplicates of 100 μL solutions into 96-well plates. The strains were analyzed with a microplate reader (BioTek Synergy H1, Winooski, VT) using excitation and emission filter for CW (380 nm and 475 nm, respectively) and graphed using Microsoft Excel.

Immunofluorescence microscopy of Co-cultured VK-2 cells with C. albicans strains.

VK-2 cells growing on glass coverslips were infected with C. albicans end3Δ/Δ or control strains for 6 h and 24 h in keratinocyte SFM media at 37°C and 5% CO2 as described (28, 32). The MOI was 0.01 for C. albicans strains in these experiments. VK-2 cells were then stained for 1 h with a 1:100 dilution of a E-cadherin primary antibody (R&D, Minneapolis, MN) in 0.2% gelatin PBS then washed and incubated with an anti-mouse-GFP secondary antibody (Invitrogen, Waltham, MA) for another hour. Cover slides were mounted on slides with an antifade mounting solution containing DAPI. A Zeiss Axio Imager M1 microscope was used to visualize the cells using standard GFP and 4′,6-diamidino-2-phenylindole (DAPI) filters.

RNA extraction, reverse transcription, and real-time qPCR.

From an overnight culture in YPD, C. albicans strains were diluted 1:50 in SFM media and grown for 4 h at 37°C. Cells were harvested and RNA was isolated using the RNeasy-Plus kit (Qiagen, Germantown, MD). A reverse transcription reaction from 100 ng RNA was performed with SuperScript III Reverse Transcriptase according to the manufacturer’s instructions (Invitrogen, Waltham, MA). In addition, 1 μL of this reaction was used as a template for a 20 μL real-time qPCR with a Sso Advanced Universal SYBR Green Supermix (Bio-Rad, Hercules, CA) and primers for actin (33) and Sap4 to 6 (34, 35) as described. Reactions were set up in triplicates for each strain and amplified gene.

Western blotting and detection of E-cadherin.

Western blotting was performed with a standard protocol as described in (36). Protein lysate was used from VK-2 cells infected with the C. albicans end3Δ/Δ mutant or control strains for 6 h and 24 h in keratinocyte SFM media at 37°C and 5% CO2. Primary antibodies to E-cadherin (R&D systems, Minneapolis, MN) and tubulin (Invitrogen, Waltham, MA) were incubated overnight and were used at 1:500 and 1:1,000 dilutions, respectively. Anti-mouse secondary antibody (Invitrogen, Waltham, MA) was incubated for 1 h in a 1:10,000 dilution. Tubulin was used as a loading control.

Live/Dead viability assay.

VK-2 cells were grown in 96-well plates in a serum-free SFM Media to a 30% confluence. C. albicans strains were grown overnight in a YPD and were cocultured with VK-2 cells at a concentration of 5 × 105 cell/mL for 6 h and 24 h. The Live/Dead viability assay from Invitrogen (Waltham, MA) was performed using the protocol provided by the manufacturer and as described previously (28). The resulting data were analyzed according to the manufacturer’s instructions using Excel (Microsoft).