**4. Discussion**

Endocytic uptake from the cell surface plays a particular role in the replication cycle of highly pathogenic HeV and NiV [11–13]. Previous work has shown the importance of endocytosis for the

maturation of F proteins to gain biological activity and thus, viral infectivity [11–13,16]. Given the fact that endocytic uptake of HeV and NiV F proteins depends on tyrosine-based motifs present in their cytoplasmic tail, the comparably high number of tyrosine residues found in the CedV F protein was of interest for this study. We first aimed to investigate whether these residues indeed mediate endocytosis of the protein. Second, we analyzed whether endocytosis is of functional relevance to the CedV F protein maturation. Taken together, our data clearly indicate that endocytic uptake of CedV F protein is signal-mediated and important for the biological activity of the protein. Furthermore, among the putative endocytosis signals, we identified a YXX Φ motif and a C-terminal di-tyrosine motif to have the strongest e ffects on endocytosis, cell surface expression, and fusion activity of CedV F protein.

The YXX Φ motif known to facilitate clathrin-mediated endocytosis of many transmembrane proteins is present in the membrane-proximal region of the cytoplasmic tail of both high and low pathogenic henipavirus F proteins. Mutation of the YXX Φ motif in the F protein of low pathogenic CedV resulted in a strongly decreased internalization rate for mutant e2 and e3, suggesting endocytic uptake of CedV F protein to be signal-mediated and largely dependent on this particular motif and most likely on Y524. For highly pathogenic HeV and NiV, several studies have similarly shown that an intact YXX Φ motif greatly contributes to the internalization of F proteins with a marked reduction in endocytosis rate upon motif disruption [11–13,43]. However, while the delayed endocytic uptake of NiV F protein led to a reduced fusogenicity of the mutant, the respective mutation in the HeV F protein rather resulted in an enhanced cell-to-cell fusion upon interaction with the viral attachment protein G [11,12]. Due to strong reduction but not the absence of endocytosis, the authors suggested an accumulation of recycled, fusogenic HeV F protein at the cell surface over time to account for the observed e ffects. For the respective CedV F mutants, we similarly found a clear delay in internalization (Figure 4b), but only moderate di fferences in total surface expression compared to the parental CedV F protein (Figure 5). However, we did indeed measure a strong 4- to 6.5-fold increase in the fusion activity of these two mutants in contrast to the parental CedV F protein, which cannot exclusively be explained by the only slightly increased cell surface expression. Although triggering of fusion is usually thought to primarily involve the ectodomain of the fusion protein, hyper- and hypofusogenicity were also observed in NiV F cytoplasmic tail mutants due to specific mutations in a membrane-proximal polybasic KKR motif [48]. The authors explained their findings by a mutation-dependent inside-out signaling mechanism resulting in conformational changes in the NiV F ectodomain followed by measurable e ffects on the fusion activity of the protein [48]. Overall, such conformational changes of the F-ectodomain due to mutations in the cytoplasmic tail can also a ffect the avidity of F and G interaction and thus, the coordinated processes required for cell–cell fusion [48–50]. Further, it is believed that mechanisms resulting in virus entry (viral—cellular membrane fusion) and cell–cell-fusion (fusion of neighboring cell membranes) are closely related [51]. Thus, cell–cell fusion levels often correlate to viral entry levels as shown for several Hendra and Nipah glycoprotein mutants [48,52–54]. In contrast, some hyperfusogenic G mutants with a modified O-glycosylation were described to display reduced entry levels but similar cell–cell fusion levels [55]. Also, a headless NiV G mutant readily triggered cell–cell fusion but pseudotyped NiV virions did not enter cells [56]. Hence, the underlying mechanism for the observed hyperfusogenicity of the CedV F mutants e2 and e3 and here the potential role of Y524 will have to be addressed in future studies.

Apart from a functional YXX Φ motif, the cytoplasmic tail of CedV F protein contains two additional di-tyrosine based motifs. Mutation of the di-tyrosine motif at amino acid position Y533/534A (mutant e4) showed no e ffect in any of the assays performed. However, mutation of the C-terminal di-tyrosine motif (Y553/554A; mutant e5 and e6) led to a detectable increase of cell surface expression with slightly enhanced fusion activity. On the one hand, this phenotype could result from a marginally reduced internalization rate. Alternatively, this di-tyrosine motif may a ffect the dynamics of endosomal tra fficking and recycling of CedV F proteins, which could subsequently alter the availability of fusion-active F protein on the cell surface, thus having an impact on fusion activity. In a previous study, the lack of the di-tyrosine motif has been considered to delay the recycling of fusogenic NiV F proteins to the cell surface [43]. A tail-truncated NiV F variant lacking the C-terminal di-tyrosine motif was markedly downregulated in constitutive surface expression while F protein endocytosis and endosomal cleavage remained una ffected. Consequently, the intact di-tyrosine motif was hypothesized to act as a potential cytoplasmic recycling motif a ffecting intracellular tra fficking [43]. Interestingly, both the YXX Φ and the C-terminal di-tyrosine motif have been described to be of importance for NiV F protein tra fficking and sorting in polarized microvascular endothelial cells [57], in polarized epithelial cells [58], and in polarized neuronal cells [59].

The presence of intact tyrosine-based endocytosis and/or sorting motif in the cytoplasmic tail of transmembrane proteins has been described to be important for the replication cycle, infectivity, and virulence of many viruses such as herpes-, corona- and retroviruses [24,25,29,30,44,60,61]. Apart from signal-mediated intracellular trafficking of viral proteins to the sites of viral assembly, these motifs were shown to affect cell surface expression of different viral transmembrane proteins [24,25,29,30,44,60,61]. For instance, mutating the endocytosis signal of the envelope protein of simian immunodeficiency virus led to more e fficient incorporation of envelope proteins into budding virions with enhanced infectivity due to increased levels of envelope protein expressed on the cell surface [61]. In the case of the CedV F protein, mutation of the di-tyrosine motif had stronger e ffects on the cell surface expression than mutation of the classical YXX Φ motif. However, at this point, it is not clear whether the di-tyrosine mutants will be incorporated more e fficiently into virions due to an increase in cell surface expression. Mutation of the di-tyrosine motif in the NiV F cytoplasmic tail almost completely abrogated NiV VLP budding despite the wt-like cell surface expression of the protein [62]. Additionally, the presence of an intact and functional YXX Φ endocytosis motif in the cytoplasmic tail of NiV and HeV F protein has been considered to play a critical role in endocytic tra fficking and recycling, and thus, in e fficient viral assembly and particle release [31,62]. Similar findings have been reported for the envelope protein of HIV-1 where recent evidence suggests that tra fficking through the recycling endosome is required for efficient incorporation of viral envelope proteins into virus particles [28]. Considering the lack of an intact YXX Φ motif, the reduced endocytosis rate, and the level of cell surface expression similar to the parental F protein, it will be interesting to see whether assembly and budding of mutant CedV F e2 and e3 is impaired or rather enhanced.

The most significant e ffects on cell surface expression, internalization rate, and fusion activity were noted for CedV mutant e7, in which all putative tyrosine-based endocytosis motifs were mutated. Importantly, the combination of all motif mutations led to an almost abrogated fusion activity of the protein with the strongest delay in internalization, which is in accordance with previous findings for a NiV F protein mutant disrupted of all tyrosine-based motifs [11]. Though CedV F mutant e7 displayed the highest cell surface expression, the combination of all mutations led to a rather reversed cleavage ratio in contrast to the other mutants. These findings point towards a reduced proteolytic F activation, which might result from the marked decrease in the endocytosis rate. Considering what is known for the F proteins of highly pathogenic HeV and NiV that are cleaved within the endosomal compartment [12–15,17], it seems likely that the proteolytic activation of CedV F protein is quite similar in terms of subcellular localization. However, future studies are needed to unravel shared commonalities and potential di fferences in this process.
