The availability of two clonal A6 cell lines that stably overexpress functional α- or β-ENaC subunits as EGFP fusion proteins, provided a relevant cellular model to test the effect(s) of potential functional differences between NEDD4L-C2(+) and Neddd4l-C2(-) on the NEDD4L substrates α- or β-ENaC. Furthermore, using the A6 cell culture system allowed functional comparisons in the presence and absence of ENaC subunit overexpression. We report that NEDD4L-C2(+) and NEDD4L-C2(-) exhibit different subcellular localizations and distinct mobilization responses to a calcium stimulus. Additionally, we report that NEDD4L-C2(+) differentially affects short term trafficking of α- and β-ENaC subunits after a calcium stimulus. Moreover, the effect of NEDD4L-C2(+) on a segment of β-ENaC trafficking is ubiquitin dependent.
We observed that the presence or absence of the C2 domain affected the subcellular localization of NEDD4L. NEDD4L-C2(+) and NEDD4L-C2(-) were labeled with EGFP and transfected into A6 cells. Whereas EGFP-NEDD4L-C2(+) localized to the cytoplasm, EGFP-NEDD4L-C2(-) localized to the cytoplasm and early endosome. Although it is unclear what functional significance NEDD4L-C2(-) localization to the early endosome may have, it is plausible that the absence of the C2 domain prevents or hinders NEDD4L-C2(-) targeting to intracellular sites beyond the early endosome.
We observed the mobilization response of EGFP-NEDD4L-C2(+) and EGFP-NEDD4L-C2(-) following an increase in intracellular calcium. A6 cells that had been transfected with either fusion protein were incubated in 1.97 mM Ca2+ and imaged immediately after the addition of 10 μm ionomycin. The Ca2+ concentration is within typical ranges of Ca2+ concentrations that are used in studies that characterize intracellular C2 domain translocation [48, 49]. While EGFP-NEDD4L-C2(-) did not respond to the Ca2+ stimulus, EGFP-NEDD4L-C2(+) mobilized from the cytoplasm to the plasma membrane and then to small intracellular compartments. These compartments were not targeted by Transferrin Texas Red®, a marker of the early endosome, or other subcellular organelle markers used (see Additional file 2).
The availability of distinct A6 cell lines that stably express α- or β-ENaC, combined with the observation that like NEDD4L-C2(+), ENaC localizes in this overexpression system to unidentified intracellular compartments , provided a means and rationale to test the effect(s) of NEDD4L-C2(+) mobilization on a well characterized substrate. Because α- and β-ENaC were stably expressed as EGFP fusion proteins , NEDD4L-C2(+) was tagged with the red fluorophore, mCherry. mCherry-NEDD4L-C2(+) exhibited a similar calcium-dependent localization response in cells that expressed either α-ENaC-EGFP or EGFP-β-ENaC. This process occurred in three steps. First, mCherry-NEDD4L-C2(+) moved from the cytoplasm to the plasma membrane. Second, mCherry-NEDD4L-C2(+) transited to the periphery of some of the vesicles that contained ENaC subunits. Third, mCherry-NEDD4L-C2(+) moved back to the plasma membrane. mCherry-NEDD4L-C2(+) did not affect the localization of α-ENaC-EGFP. However, in cells that expressed EGFP-β-ENaC, mCherry-NEDD4L-C2(+) transited to the plasma membrane with EGFP-β-ENaC in the third mobilization step.
Ubiquitination is intimately involved in several aspects of the intracellular trafficking of various proteins . We therefore tested whether the co-transit of NEDD4L-C2(+) and β-ENaC to the plasma membrane in response to an increase in intracellular calcium was ubiquitination dependent. To ablate the ubiquitin ligase capability of NEDD4L-C2(+), the catalytic cysteine 943 (C943) in the HECT domain was mutated to alanine. C943 forms a thioester linkage with ubiquitin prior to substrate ligation and this association is required for substrate ubiquitination. The functional significance of C943 is supported by its conservation in NEDD4L orthologs from humans to yeast. In the context of ENaC regulation, mutation of the catalytic cysteine to alanine abolishes NEDD4L mediated β-ENaC ubiquitination . In the mobilization response to an increase in intracellular calcium, mCherry-NEDD4L-C2(+)-C943A first transited from the cytoplasm to the plasma membrane and then transited to the periphery of some vesicles that contained β-ENaC-EGFP. However, unlike the mobilization response of wild-type mCherry-NEDD4L-C2(+), neither mCherry-NEDD4L-C2(+)-C943A nor β-ENaC-EGFP subsequently transited to the plasma membrane.
The paradigm of E3 enzyme function within the ubiquitin system implies that E3s down-regulate or sequester their substrate targets. Ubiquitination results in substrate transport to the proteasome and lysosome (for degradation) or to endosomal compartments where proteins remain intact but sequestered until needed and recycled for future use . It is not clear how our observation of mCherry-NEDD4L-C2(+) and β-ENaC-EGFP mobilization from intracellular compartments to the plasma membrane would facilitate NEDD4L mediated ENaC down-regulation. Since β-ENaC is active at the cell surface this inside-out movement propels β-ENaC towards its functional locale. This trafficking step may therefore correspond to β-ENaC mobilization from an intracellular pool or recycling. Recent work using live-cell imaging and GFP-labeled ENaC subunits overexpressed in polarized kidney epithelial (MDCK) cells, has shown that cAMP can stimulate rapid ENaC trafficking, more from replenishment than recycling, to the apical surface from an intracellular pool . Furthermore, this rapid cAMP-stimulated replenishment is dependent on the presence of intact PY motifs because it is impaired by PY-motif mutants of ENaC suggesting that the PY motifs play a role in regulating exocytic trafficking of the channel. These results confirmed earlier work in rat thyroid cells demonstrating that cAMP stimulation of ENaC was dependent on the PY motif in the COOH terminus of each subunit, with the most pronounced effect seen with a truncating mutation in the β-ENaC subunit .
The observation here that neither catalytically inactive NEDD4L (mCherry-NEDD4L-C2(+)-C943A) nor β-ENaC-EGFP mobilize to the plasma membrane suggests that ubiquitination is necessary for wild-type NEDD4L (mCherry-NEDD4L-C2(+)) and β-ENaC-EGFP co-transit to the plasma membrane. This result is surprising considering that ubiquitination of ENaC as well as myriad other proteins typically leads to internalization and functional down-regulation [10, 51]. However, our observation of ubiquitin-dependent mobilization of β-ENaC by NEDD4L-C2(+) provides a novel mechanism for the PY motif-dependent ENaC mobilization from intracellular pools by cAMP. Multiple reports suggest that intracellular calcium oscillations can be mediated by cAMP in renal collecting duct cells [52–54]. A cAMP mediated increase in intracellular calcium could trigger NEDD4L-C2(+) mobilization and targeting of intracellular compartments that contain β-ENaC. Subsequent co-transiting of NEDD4L-C2(+) with β-ENaC from intracellular pools to the cell surface could occur in a PY motif-dependent manner due to the requirement for interactions with NEDD4L-C2(+) WW domains.
The observation that mCherry-NEDD4L-C2(+) mobilizes to the plasma membrane independently in A6 cells that stably express α-ENaC (Figure 5D, see Additional file 3) suggests that fluorescently labeled ENaC subunits may not be the only substrate(s) that are ubiquitinated. Future investigations that employ alternate experimental approaches should provide more detailed insight into the mechanisms of NEDD4L and ENaC trafficking. Likewise, the functional effects of the observed co-mobilization between mCherry-NEDD4L-C2(+) and β-ENaC should be correlated with functional sodium transport studies.
The differential trafficking effect of NEDD4L-C2(+) on α- and β-ENaC subunits is consistent with differential subunit trafficking observed in other contexts. In A6 cells an increase in the density of β-ENaC subunits, but not α- or γ-ENaC subunits, at the cell membrane in response to aldosterone and vasopressin stimuli has been observed . Aldosterone-mediated-regulation of ENaC occurs at least in part by activating the serum glucocorticoid-inducible kinase 1 (Sgk1) . Subsequent phosphorylation of NEDD4L hinders the interaction between NEDD4L and ENaC thereby reducing the negative regulatory effect of NEDD4L on ENaC [57, 58]. In vivo vasopressin induces a significant increase of the β-ENaC subunit compared to the α-ENaC subunit in rat kidney  as well. Furthermore, a model that describes intracellular ENaC trafficking proposes that its subunits are transported individually, independent of each other, in a non-coordinate manner .
The differential regulation of intracellular trafficking as a function of the presence or absence of a C2 domain may extend to other substrates of NEDD4L, including the dopamine transporter (dopamine active transporter, DAT) , and to other Nedd4-like E3 ligases, including the NEDD4L paralog Nedd4-1 (Nedd4). Moreover, Nedd4-like proteins have been implicated in the process of viral budding due to the potential interaction between Nedd4-like WW domains and viral late domains that contain proline-rich motifs . Viral pathogens could therefore exploit the differential trafficking behavior and intracellular localization between NEDD4L-C2(+) and NEDD4L-C2(-) to suit aspects of their pathogenesis. Two recent publications indicate that NEDD4L-C2(-) can more potently correct human immunodeficiency virus type 1 (HIV-1) release defects in 293T cells compared to NEDD4L-C2(+) [63, 64]. We observed that NEDD4L-C2(-) localizes predominately to transferrin-labeled early endosomal compartments instead of the cytoplasm like NEDD4L-C2(+) (Figure 2). A recent study demonstrated that HIV-1 localizes to transferrin-labeled early endosomal compartments . Perhaps the ability of NEDD4L-C2(-) to correct more readily HIV-1 release compared to NEDD4L-C2(+) is the result of the preferential localization of NEDD4L-C2(-) to early endosomal compartments where HIV-1 is present, thereby facilitating interactions that promote viral egress.
Our data support the hypothesis that the G/A variant impacts ENaC-dependent sodium handling in the ASDN by affecting the relative synthesis of NEDD4L-C2(+) and NEDD4L-C2(-). Individual differences in the regulation of sodium balance mediated by this genotype may contribute modest but definite individual differences in the liability to develop essential hypertension through a sodium dependent mechanism. Our cellular observations suggest a mechanism accounting for recently reported clinical correlations between the genotype of the G/A variant and responses to acute sodium loading as well as changes in blood pressure . Greater synthesis of the NEDD4L-C2(+) isoform relative to the NEDD4L-C2(-) isoform may favor apical recycling or mobilization from intracellular pools, thereby generating a greater apical density of ENaC channels in the ASDN through regulation of the β-ENaC subunit, without concurrent effects on the α-ENaC subunit.
The cellular observations provide a framework for future investigations of the potential relationship between genetic variation in NEDD4L affecting relative expression of C2-defined isoforms, subcellular trafficking and regulation of ENaC subunits, and whether and how such genetic variation can contribute to individual susceptibility to develop essential hypertension. It is of particular interest whether our observations relate to ENaC activity. Electrophysiological experiments that would investigate amiloride sensitive sodium current in a similar tissue culture system would be a logical approach to resolve this issue. While we exploited published evidence regarding interaction between ENaC and NEDD4L as well as the availability of a cellular model system , the functional implications of C2-containing and C2-lacking NEDD4L isoforms may extend beyond blood pressure control to other NEDD4L mediated processes including viral pathogenesis [62, 63] and regulation of dopamine transport .