Previous studies have focused on the adherence properties of the Als proteins [23
]. The different Als proteins mediate adherence to broad variety of host substrates, and this adherence is likely critical for C. albicans
to infect host surfaces. In this study, we report that an additional function of Als3 is that it can act as an invasin protein and induce endocytosis by normally nonphagocytic host cells. Furthermore, we show that this endocytosis occurs as a result of the N-terminal region of Als3 binding to either N-cadherin or E-cadherin. These conclusions are supported by the following results. First, an als3
Δ null mutant was endocytosed poorly by endothelial cells and two different oral epithelial cell lines. Second, latex beads coated with rAls3-N were efficiently and specifically endocytosed by both endothelial cells and oral epithelial cells. Third, the rAls3-N–coated beads were avidly endocytosed by CHO cells expressing either N-cadherin or E-cadherin, but not by CHO cells that did not express these cadherins. These current results are also consistent with our previous data that S. cerevisiae
expressing Als3 are endocytosed by endothelial cells [23
In our prior studies, we have determined that C. albicans
must be endocytosed by either endothelial cells or oral epithelial cells to cause damage to these cells [4
]. We found that the als3
Δ null mutant caused virtually no damage to either endothelial cells or FaDu oral epithelial cells. Similarly, Zhao et al. [24
] reported that an als3
Δ null mutant had markedly reduced capacity to damage reconstituted human epithelium. Our results suggest that the reduced capacity of the als3
Δ null mutant to damage endothelial cells and oral epithelial cells in vitro is due to its defect in invading these cells.
The invasion and damage defects of the als3
Δ null mutant suggest that Als3 is important for virulence during hematogenously disseminated and oropharyngeal candidiasis. Investigations to determine the role of Als3 in C. albicans
virulence are currently ongoing. However, we have found that vaccination of mice with rAls3 provides significant protection from lethal hematogenously disseminated candidiasis and ameliorates the course of oropharyngeal and vaginal candidiasis [40
]. Thus, Als3 is a useful therapeutic target.
In our assays, the als3
Δ null mutant had significantly reduced adherence to endothelial cells, but normal adherence to oral epithelial cells. It has been reported by others that an als3
Δ null mutant had reduced adherence to endothelial cells, buccal epithelial cells, and FaDu epithelial cells [24
]. Also, we found that S. cerevisiae
expressing Als3 bound to both endothelial cells and FaDu oral epithelial cells [23
]. The discrepancy between the previous and current results is likely due to differences in the methodology of the assays. Specifically, the longer incubation time and the use of a 24-well tissue culture plate (rather than a 6-well plate) in the current endocytosis assay make it less sensitive to differences in adherence among strains.
We also found that the als1Δ/als1Δ mutant was endocytosed normally by oral epithelial cells, whereas latex beads coated with rAls1-N induced some endocytosis by these cells. The probable explanation for these apparently conflicting results is that the als1Δ/als1Δ mutant still expressed Als3 on its surface. Thus, the presence of Als3 masked the effects of the absence of Als1 when the endocytosis assay was performed using whole organisms. Studies performed with latex beads coated with rAls1-N indicated that Als1 is capable of inducing epithelial cell endocytosis by itself. Collectively, these results demonstrate the utility of using both mutant analysis and recombinant proteins to assess the contribution of an individual Als protein to host cell interactions.
Previously, we determined that the N-terminal region of Als proteins contains the substrate-binding site [23
]. This conclusion is supported by our current finding that latex beads coated with just the N-terminus of Als3 were able to induce endocytosis. However, it was notable that coating the latex beads with rAls3-N resulted in a much greater increase in endocytosis than adherence. Also, coating beads with rAls1-N resulted in little to no increase in adherence, even though previous studies have clearly demonstrated that Als1 is an adhesin [36
]. The likely explanation for the relative lack of adherence of the beads is that they were coated with fragments of Als1 or Als3 that lacked the tandem repeats present in the central portion of the full-length proteins. Previously, we have found that the adhesive function of Als1 is dramatically reduced when some or all of the tandem repeats are eliminated [26
]. Similarly, Oh et al. [35
] found that versions of Als3 with fewer tandem repeats mediated less adherence than did versions of Als3 with more tandem repeats. In addition, Rauceo et al. [38
] reported that the absence of tandem repeats reduced the binding affinity of Als5. Collectively, these results support a model in which the tandem repeats influence the conformation and/or the substrate accessibility of the N-terminal region of Als3 and thereby enhance its binding affinity for host substrates. However, even the weak binding of the N-terminal portion of Als3 to N-cadherin or E-cadherin is sufficient to induce endocytosis.
The results with the als3
Δ null mutant and the latex beads coated with rAls3-N demonstrate that endothelial cell N-cadherin and epithelial cell E-cadherin are two host cell ligands for Als3. Furthermore, binding of the N-terminus of Als3 to either of these cadherins is sufficient to induce endocytosis. Previously, we have reported that N-cadherin is one of the endothelial cell surface proteins that mediate endocytosis of C. albicans
]. E-cadherin is known to mediate the endocytosis of Listeria monocytogenes
by epithelial cells [41
], but it was not previously known to mediate the endocytosis of any fungi. Our results with C. albicans
raise the interesting possibility that perhaps L. monocytogenes
can utilize N-cadherin to invade endothelial cells.
Although the N-termini of Als1 and Als3 share considerable homology at the amino acid level, Als3 induced endocytosis much more efficiently than did Als1. The results of computer-assisted modeling provide a possible structural explanation for these functional differences. A striking finding was that the 3-D structures of the N-terminal regions of Als1 and Als3 are predicted to have recurrent β barrel domains and a global negative surface potential similar to the N-terminal regions of N-cadherin and E-cadherin. Even though N-terminal regions of Als1 and Als3 share general features, they are predicted to differ both in their 3-D structure and their interactions with N- and E-cadherin. While Als1 appears to favor an extended conformation, Als3 tends toward a more folded structure that forms a more distinct molecular cleft.
Our molecular modeling results also suggest that Als1 binds more weakly to N- and E-cadherin than does Als3. These results are consistent with our finding that beads coated with rAls3-N were endocytosed more avidly by CHO cells expressing N- or E-cadherin than were beads coated with rAls1-N. Computer-assisted modeling also predicted that Als1 and Als3 interact with N- and E-cadherin via different domains, and the geometry of Als3 binding to E-cadherin is similar to that of E-cadherin binding to itself. Collectively, these findings support the hypothesis that Als3 functions as a molecular mimic of mammalian cadherins, thereby facilitating C. albicans invasion of endothelial and oral epithelial cells.
The N-terminal regions of Als1 and Als3 are also predicted to share structural homology to bacterial adhesins and invasins. The conformation of Als1 is similar to that of the pfam02368 family of bacterial immunoglobulin-like domain surface protein/adhesins, which includes the invasin of Y. pseudotuberculosis
]. In contrast, the cleft motif of Als3 is more similar to that of the COG4886 family of leucine-rich repeat invasins/intimins [42
]. Interestingly, this family of proteins includes internalin A of L. monocytogenes
, which mediates epithelial cell invasion by binding to E-cadherin [41
The models presented here do not purport to specify exact interactions between cadherins and Als proteins. Defining such interactions at the atomic level requires physicochemical methods such as ultracentrifugation and x-ray crystallography. However, our findings are strengthened by the fact that the results of our modeling of cadherin–cadherin interactions were very similar to what has been found by x-ray crystallography [45
Previously, we have reported that S. cerevisiae
expressing Als1 or Als3 binds to a variety of host constituents, including endothelial cells, epithelial cells, laminin, and fibronectin [23
]. In the affinity purification experiments, hyphae of the als3
Δ null mutant failed to bind to multiple host proteins, including N-cadherin and E-cadherin. Thus, these results are consistent with the model that Als3 binds to a broad range of host substrates. However, our data also indicate that there is some specificity in the binding of Als proteins to host constituents. For example, beads coated with rAls3-N were endocytosed much more efficiently than were beads coated with rAls1-N. These results, combined with those of the molecular modeling studies, suggest that Als3 binds with greater affinity to N- and E-cadherin than does Als1. Nevertheless, it is highly likely that additional host cell ligands, other than N- or E-cadherin, also contribute to the endocytosis of C. albicans,
and we are currently working to identify them.
For many microbial pathogens, invasion of host cells is critical for the initiation and maintenance of infection, and many of these organisms have more than one mechanism for inducing their own uptake by host cells [46
]. Thus, it is logical to speculate that C. albicans
has also evolved at least one Als3-independent mechanism to invade host cells. The Als family of proteins contains six members in addition to Als1 and Als3 [20
]. Most of these proteins are known to have adhesive function. When expressed in S. cerevisiae,
Als5 induces weak endothelial cell endocytosis similar to that induced by Als1 [23
]. Als6, Als7, and Als9 do not appear to mediate significant endothelial cell endocytosis, and Als2 and Als4 have not yet been tested in this assay. While it is possible that Als3 is the only member of its family that has significant invasive function, it is also conceivable that one or more of the other Als proteins mediates invasion of other types of host cells. Thus, a detailed analysis of these proteins will likely provide further insight into the pathogenesis of hematogenously disseminated and mucosal candidiasis, and perhaps additional therapeutic targets.