Specific Palmitoyltransferases Associate with and Activate the Epithelial Sodium Channel*

The epithelial sodium channel (ENaC) has an important role in regulating extracellular fluid volume and blood pressure, as well as airway surface liquid volume and mucociliary clearance. ENaC is a trimer of three homologous subunits (α, β, and γ). We previously reported that cytoplasmic residues on the β (βCys-43 and βCys-557) and γ (γCys-33 and γCys-41) subunits are palmitoylated. Mutation of Cys that blocked ENaC palmitoylation also reduced channel open probability. Furthermore, γ subunit palmitoylation had a dominant role over β subunit palmitoylation in regulating ENaC. To determine which palmitoyltransferases (termed DHHCs) regulate the channel, mouse ENaCs were co-expressed in Xenopus oocytes with each of the 23 mouse DHHCs. ENaC activity was significantly increased by DHHCs 1, 2, 3, 7, and 14. ENaC activation by DHHCs was lost when γ subunit palmitoylation sites were mutated, whereas DHHCs 1, 2, and 14 still activated ENaC lacking β subunit palmitoylation sites. β subunit palmitoylation was increased by ENaC co-expression with DHHC 7. Both wild type ENaC and channels lacking β and γ palmitoylation sites co-immunoprecipitated with the five activating DHHCs, suggesting that ENaC forms a complex with multiple DHHCs. RT-PCR revealed that transcripts for the five activating DHHCs were present in cultured mCCDcl1 cells, and DHHC 3 was expressed in aquaporin 2-positive principal cells of mouse aldosterone-sensitive distal nephron where ENaC is localized. Treatment of polarized mCCDcl1 cells with a general inhibitor of palmitoylation reduced ENaC-mediated Na+ currents within minutes. Our results indicate that specific DHHCs have a role in regulating ENaC.

ENaCs 2 are amiloride-sensitive Na ϩ channels that are found in high resistance epithelia and other tissues. In the aldoste-rone-sensitive distal nephron (ASDN), ENaC-dependent Na ϩ transport has an important role in the maintenance of extracellular fluid volume and blood pressure, as well as extracellular K ϩ homeostasis. In the lung, ENaC has a role in regulating airway surface fluid volume, mucociliary clearance, and alveolar fluid volume (1)(2)(3). The channels are composed of three homologous subunits (termed ␣, ␤, and ␥), each with two transmembrane domains, a large extracellular loop and cytoplasmic N and C termini (3)(4)(5).
Using a fatty acid-biotin exchange assay, we found that the ␤ and ␥ subunits of mouse and human ENaC were Cys palmitoylated, whereas the ␣ subunit was not (20,21,33). Sites of cytoplasmic Cys palmitoylation included ␤Cys-43 and ␤Cys-557 in the N and C termini of the mouse ␤ subunit, respectively, as well as ␥Cys-33 and ␥Cys-41 in the N terminus of the ␥ subunit (20,21). Mutation of specific Cys palmitoylation sites to Ala reduced channel activity, whereas membrane trafficking and subunit proteolysis associated with activation were unchanged (20,21). Cell-attached patch clamp analyses of Xenopus oocytes expressing channels with subunits lacking selected palmitoylation sites (␣␤C43A,C557A␥ or ␣␤␥C33A,C41A) revealed a significantly reduced P o , when compared with wild type ENaC, indicating that palmitoylation affected channel gating (20,21). Comparison of the activities of channels lacking palmitoylation of one or both subunits (␣␤C43A,C557A␥, ␣␤␥C33A,C41A, or ␣␤C43A,C557A ␥C33A,C41A) showed that ␥ subunit palmitoylation had a dominant role in modulating ENaC activity (21).
Protein palmitoylation is catalyzed by a family of 23 mammalian palmitoyltransferases that exhibit four or more transmembrane domains and a highly conserved cysteine-rich domain adjacent to a DHHC tract (Asp-His-His-Cys) within the active site of the enzyme (34,35). These palmitoyltransferases, referred to as DHHCs, differ in their relative size and in their cytoplasmic N and C termini, which exhibit protein-interacting motifs such as PDZ-binding domains or ankyrin repeats that likely have roles in subcellular localization and conferring substrate specificity (36,37). We previously reported that DHHC 2 co-immunoprecipitates with ENaC when co-expressed in Madin-Darby canine kidney cells and that DHHC 2 enhances ENaC activity by 2.5-fold when co-expressed in Xenopus oocytes (21). Although ENaCs lacking ␤ subunit palmitoylation had reduced activation by DHHC 2, ENaCs lacking ␥ subunit palmitoylation were not activated by DHHC 2 (21), consistent with the concept that ␥ subunit palmitoylation has a dominant role controlling ENaC P o . In the current study, we examined whether ENaC is regulated by any other of the 23 known DHHCs. We identified five that activate and co-immunoprecipitate with ENaC and three that enhance ENaC activity in a ␥ subunit-specific manner.

Results
Five DHHCs Activate ENaC and Co-immunoprecipitate with the Channel-To identify ENaC-activating DHHCs, the 23 mouse DHHCs with N-terminal HA epitope tags were individually co-expressed with wild type mouse ␣␤␥ ENaC in Xenopus oocytes. Amiloride-sensitive currents were measured after 36 -48 h. Five DHHCs (DHHCs 1, 2, 3, 7, and 14) significantly increased ENaC activity ϳ2-fold (ranging from 1.7 Ϯ 0.9 to 2.5 Ϯ 1.5-fold (means Ϯ S.D.), p Ͻ 0.01, by one-way ANOVA), whereas the remaining 18 did not significantly alter ENaC activity (Fig. 1A). Because the expression levels of the 23 DHHCs have been variable when expressed in mammalian cells (38,39), we also assessed their expression in Xenopus oocytes by immunoblotting extracts of oocytes injected with cRNA encoding individual HA-DHHCs. Extracts of oocytes were subjected to immunoblotting with anti-HA antibodies and revealed bands of the expected size for each DHHC, except DHHC 21, although the ␤-actin signal by immunoblotting was relatively uniform for all the samples (Fig. 1B). The strongest signals were observed for DHHCs 7, 9, 14, and 25, but of these, only DHHC 7 and 14 were found to activate ENaC when co-expressed. Increased amounts of cRNA were required to obtain signals for DHHCs 4,8,11,and 13. Also, longer exposure of the immunoblot to film was needed to obtain signals for DHHCs 4, 6, 11, 20, and 23 (Fig. 1C).
To determine whether ENaC palmitoylation was increased when the channel was co-expressed with an activating DHHC, we transfected HEK293 cells with cDNA for ENaC including a ␤ subunit with a C-terminal V5 epitope tag, in the presence or absence of DHHC 7. Palmitoylation of the ␤ subunit was assessed with fatty acid exchange chemistry where palmitate is removed from Cys with hydroxylamine treatment, using Tris treatment as a negative control, and replaced with biotin as previously described (20,21). We observed a significant 2.4-fold increase in ␤ subunit palmitoylation when the channel was coexpressed with DHHC 7 (7.7% Ϯ 2.6 (Ϫ DHHC 7) versus 18.7% Ϯ 7.4 (ϩ DHHC 7), means Ϯ S.D., p Ͻ 0.05) (Fig. 2).
To determine whether the five activating DHHCs associate with ENaC within a protein complex, we co-expressed ␣␤␥ (all subunits with C-terminal V5 tags) and individual DHHCs (DHHC 1, 2, 3, 7, or 14) bearing N-terminal GFP in Fischer rat thyroid (FRT) cells. Cell extracts were immunoprecipitated with anti-V5 antibodies and immunoblotted with anti-GFP antibodies. We found that each of the five DHHCs co-immunoprecipitated with ENaC (Fig. 3, A and B). V5-tagged channels bearing mutations of the four palmitoylation sites (␣␤Cys43, 557A␥Cys33,41A) also co-immunoprecipitated with the five palmitoyltransferases (Fig. 3, A and B). Based on the intensity of the bands, the most robust co-immunoprecipitating DHHCs were 1 and 7. These results also confirmed our previous findings that DHHC 2 activated and co-immunoprecipitated with ENaC (21). The blots were also reprobed with anti-V5 antibodies and showed similar band intensities of the three V5-tagged ENaC subunits (Fig. 3C), indicating that variations in DHHC co-immunoprecipitation were not due to differential expression or immunoprecipitation of ENaC. Anti-V5 immunoprecipitates from FRT cells expressing the five GFP-DHHCs in the absence of ␣␤␥ENaC revealed no signal on an immunoblot for DHHCs 2, 3, 7, and 14 but a faint signal for DHHC 1 (Fig. 3, D and E). However, the signal for DHHC 1 in the anti-V5 immunoprecipitates was consistently enhanced by co-expression with V5-tagged ␣␤␥ENaC (Fig. 3E). Interestingly, we also found that non-activating DHHC 11 and 23 were also present in anti-V5 immunoprecipitates when co-expressed with V5tagged ␣␤␥ENaC (Fig. 3, F and G), suggesting that the nonactivating DHHCs may also be in complex with the activating DHHCs. Complexes of multiple DHHCs have been previously reported (40).
Palmitoylation of the ␤ and ␥ Subunits Are Necessary for Full DHHC-mediated ENaC Activation-We previously reported that channels with mutations that block ␥ subunit palmitoylation (or both ␤ and ␥ subunit palmitoylation) are not activated by DHHC 2, whereas channels that lack ␤ subunit palmitoylation sites are activated by DHHC 2 (21). These results suggested that there was a degree of subunit specificity regarding channel activation by DHHC 2. We examined whether the other ENaCactivating DHHCs exhibited subunit specificity regarding channel activation. ENaCs lacking palmitoylation sites on either the ␤ subunit, the ␥ subunit, or both subunits were expressed in oocytes with or without DHHC 1, 2, 3, 7, or 14. As expected, ENaCs lacking ␤ and ␥ subunit palmitoylation sites (␣␤Cys43,557A␥Cys33,41A) were not activated by these five DHHCs (Fig. 4A). Furthermore, ENaCs lacking ␥ subunit palmitoylation sites (␣␤␥C33A,C41A) were not activated by the five DHHCs (Fig. 4B). However, channels lacking ␤ subunit palmitoylation sites (␣␤C43A,C557A␥) were significantly activated by DHHCs 1, 2, and 14 but not by DHHCs 3 and 7 (Fig.  4C). These data suggest that ENaC activation by DHHCs 1, 2, and 14 requires Cys palmitoylation of the ␥ subunit, whereas channel activation by DHHCs 3 and 7 requires Cys palmitoylation of both subunits. These data are also consistent with our earlier results suggesting that ␥ subunit palmitoylation has a dominant role in activating ENaC (21).
ENaC-activating DHHC 3 Is Expressed in the ASDN of the Kidney-ENaC is expressed in the latter aspects of the ASDN. Deep sequencing of dissected rat renal tubules indicated that 18 DHHCs are expressed in the ASDN, including the five ENaC-FIGURE 1. ENaC is activated by specific DHHCs when co-expressed in Xenopus oocytes. A, oocytes were injected with cRNAs for wild type ␣␤␥ alone (NA, no addition) or with one of the 23 DHHCs with an N-terminal HA epitope tag as indicated (numbered 1-25). Amiloride-sensitive currents were measured 36 -48 h after cRNA injection (n ϭ 11-50) and normalized to currents of wild type ␣␤␥ each day. The data are presented as box and whisker plots, with wild type ␣␤␥ set as 1 (n ϭ 378, dashed line). Significant increases were found for DHHCs 1, 2, 3, 7, and 14 (gray boxes) (n ϭ 17-24) when compared with ␣␤␥ expressed alone (p Ͻ 0.01, determined with one-way ANOVA followed by a Tukey test). Whiskers indicate the 10th and 90th percentiles. Median is indicated by a horizontal line, and the mean is indicated with a dot symbol within each box. B and C, oocytes were injected with cRNAs for each of the 23 HA-DHHCs or no DHHC (Ϫ), and detergent extracts of oocytes were subjected to SDS-PAGE and immunoblotting with anti-HA antibodies conjugated to HRP or anti-␤-actin antibodies (as loading control). A band of the expected size (formula weight in kDa beneath each lane) was observed for all the DHHCs except DHHC 21. A representative immunoblot is shown in B, and a longer exposure is shown in C to enhance the faint signals for DHHCs 4, 6, 11, 20, and 23. The mobility of the Bio-Rad Precision Plus protein standards is shown to the left of each blot.
activating DHHCs (41). We examined whether an ENaC-activating DHHC (DHHC 3) is expressed in principal cells of the mouse ASDN, where ENaC and aquaporin 2 are localized. We carried out immunofluorescence confocal microscopy with a rabbit anti-DHHC 3 antibody and a goat anti-aquaporin 2 antibody in paraformaldehyde (PFA)-fixed mouse kidney sections (Fig. 5, A-C). We observed staining with the two antibodies in the same cells, indicating that intracellular DHHC 3 is expressed in principal cells in mouse ASDN. DHHC 3 was also observed in adjacent aquaporin 2-negative cells, consistent with DHHC 3 expression in the intercalated cells of the ASDN. Notably, anti-DHHC 3 antibody staining was blocked by preincubation with the antigenic peptide (Fig. 5B). DHHC3 intracellular expression in principal cells was distinctly subapical, consistent with localization in intracellular vesicles (Fig. 5C). The anti-DHHC 3 antibody was also validated by immunoblot of extracts from FRT cells after transfection with the cDNA for either GFP or GFP-DHHC 3. We observed only a single band in the GFP transfected cells corresponding to the expected size of the endogenous DHHC 3 (M r ϭ 34 kDa) and a second band in the GFP-DHHC 3 transfected cells corresponding to the expected size of the GFP-tagged protein (62 kDa) (Fig. 5D). We also observed a single band in an immunoblot of kidney homogenates from three individual mice at a size expected for the mouse DHHC 3 (M r ϭ 34 kDa) that was blocked by antibody preincubation with the antigenic peptide (Fig. 5E).

Discussion
Results from numerous studies indicate that ϳ400 proteins in mammalian cells and tissues are palmitoylated. The identification of palmitoylated proteins has been a slow process. The first acylated proteins were described in 1979, whereas the first global analysis of palmitoylated yeast proteins and palmitoylated rat brain synaptosomal proteins were performed in 2006 and 2008, respectively (45)(46)(47)(48). Palmitoylation was established by a variety of increasingly complex methodologies starting FIGURE 2. Palmitoylation of ENaC is increased by co-expression with the activating DHHC 7. HEK293 cells were transfected with ␣␤␥ ENaC with a C-terminal V5 epitope tag on the ␤ subunit, with or without the activating DHHC 7 as indicated. Cys palmitoylation of the ␤ subunit in anti-V5 IPs was assessed with fatty acid exchange chemistry where palmitate is removed from Cys with hydroxylamine treatment, using Tris treatment as a negative control, and replaced with biotin. A fraction of the IP (10%) was reserved to assess total ␤ subunit in the initial IP, and biotinylated ␤ subunit was recovered with avidin-conjugated beads (90%) for immunoblotting (IB) with anti-V5 antibodies. The percentage of palmitoylation was calculated from the difference in ␤ subunit biotinylation after treatment with hydroxylamine (ϩ) or Tris (Ϫ), relative to the ␤ subunit recovered in the total IP. A, the percentage of ␤ palmitoylation (means and S.D.) in the presence (n ϭ 3, white bar and open circles) or absence (n ϭ 4, black bar and closed circles) of DHHC 7 was statistically significantly different by Student's t test (p Ͻ 0.05). B, a representative immunoblot from a single experiment is shown, with the mobility of Bio-Rad Precision Plus protein standards to the right of each panel. . ENaC co-immunoprecipitates with the five activating DHHCs. FRT cells were transfected with WT ␣␤␥ ENaC or ENaC with mutant subunits lacking sites for palmitoylation (␣␤Cys43,557A␥Cys33,41A) either alone (Ϫ) or with DHHC 1, 2, 3, 7, or 14 as indicated. All DHHCs had an N-terminal GFP epitope-tag and all three ENaC subunits had a C-terminal V5 epitope tag. A, an aliquot of the cell extract (10% total extract) was retained for immunoblotting (IB) with anti-GFP antibodies. B, the remainder was incubated with anti-V5 antibodies. Total extract and IPs were subjected to immunoblotting with anti-GFP antibodies. with radiolabeling with [ 3 H]palmitate and moving on to acylbiotin exchange protocols, acyl-resin-assisted capture with thiopropyl-Sepharose and click chemistry based on metabolic labeling with bioorthogonal palmitate analogs combined with quantitative mass spectroscopy (45,46,49,50).
ENaCs are among the ϳ50 ion channels known to be palmitoylated (for reviews, see Refs. 22, 26, and 51). We treated mCCD cl1 cells that express endogenous ENaC with an inhibitor of protein acylation (2-BP) and found a rapid reduction in ENaC currents consistent with the reduced activity of ENaC that we observed in heterologous systems when we expressed mutant ENaC lacking sites for Cys palmitoylation (20,21). However, the reduced currents could also reflect changes in palmitoylation of additional proteins that either alter ENaC trafficking or gating. The palmitate analog 2-BP has been regularly used in cultured cells to inhibit palmitoylation of proteins, thereby providing a complementary in vivo approach to study the function of palmitate addition to specified proteins (44,52). Although 2-BP clearly inhibits the palmitoyltransferases, there is evidence that it inhibits the thioesterases as well (44,52,53). 2-BP also inhibits metabolic enzymes, including fatty acid CoA ligase, which reduces the cellular content of the palmitoylation substrate palmitoyl-CoA (44,52,53). We also observed a decrease in TER after treatment of mCCD cl1 cells with 2-BP, which may reflect reduced palmitoylation of claudins normally found within the tight junctions (54 -56). Transcripts for 13 claudins were identified in dissected rat kidney CCDs by deep sequencing (41). A role of claudin palmitoylation in modifying TER was noted in studies of polarized Madin-Darby canine kidney cells, where TER was increased 5-fold after transfection with wild type claudin-14. However, TER was unchanged after transfection with claudin-14 lacking sites for palmitoylation because of reduced targeting to the tight junctions (55).
Of the 23 mouse palmitoyltransferases, DHHCs 1, 2, 3, 7, and 14 activated ENaCs expressed in Xenopus oocytes. Analysis of gene expression data in the online Xenbase indicates that DHHCs 4, 6, 8, 14, and 16 are expressed in Xenopus laevis oocytes, whereas DHHCs 1, 2, 3, 7, 9, 15, 20, and 21 are not expressed. There are no data available for DHHCs 5,11,12,13,17,23,24,and 25. Although these data indicate that oocytes do express one ENaC-activating transferase (DHHC 14), overexpression of DHHC 14 further enhanced ENaC activity. These five activating DHHCs are expressed in the ASDN (41). ENaCs are expressed in principal cells in the latter aspect of the ASDN, and we found transcripts for all five ENaC-activating DHHCs in a cultured principal cell line (mCCD cl1 ) using RT-PCR. Furthermore, in mouse kidney sections DHHC 3 was localized in cells expressing aquaporin 2, a principal cell marker. At least three of the channel activating DHHCs (DHHCs 1, 2, and 3) are expressed in mouse airway epithelia (57), whereas analyses of DHHC expression in both human kidney and lung by RT-PCR revealed expression of 17 DHHCs (including DHHCs 1, 3, 7, and 14) (58). Taken together, these data are consistent with a role for Cys palmitoylation in regulation of ENaC in both kidney and lung, because the relevant transferases are expressed at the appropriate sites in these tissues.
Although the list of palmitoylated proteins continues to expand, it has been a challenge to identify physiologically relevant DHHC-substrate pairs as (i) there are numerous reports indicating that substrates are shared by multiple DHHCs, (ii) tissue and cell type expression of the DHHCs are highly variable, and (iii) the definitive subcellular distribution of each DHHC has been described in only a few cases (for review, see Refs. 26 and 27). Expression of epitope-tagged DHHCs in HEK293 cells coupled with immunofluorescence microscopy placed DHHCs 1, 2, 3, 6, 10, 11, 12, 13, 14, 16, 19, and 22 in the Xenopus oocytes were injected with cRNAs for wild type ␣␤␥ or mutant ENaCs lacking sites for palmitoylation on both subunits (␣␤Cys43,557A␥Cys33,41A) (A), the ␥ subunit (␣␤␥C33A,C41A) (B), or the ␤ subunit (␣␤C43A,C557A␥) (C). Mutant ENaCs were expressed alone (NA, no addition, n ϭ 77-98) or co-expressed with DHHCs 1, 2, 3, 7, or 14 as indicated (n ϭ 11-33). Amiloridesensitive currents were measured 48 h after cRNA injection and normalized to wild type ␣␤␥ currents each day. The data are presented as box and whisker plots, with wild type ␣␤␥ set as 1 (dashed line, n ϭ 73-115). Co-expression of DHHCs with ENaC lacking palmitoylation sites on both the ␤ and ␥ subunit (A) or just the ␥ subunit (B) did not affect channel activity (p Ͼ 0.05 versus NA). C, co-expression of DHHCs 1, 2, or 14 with ENaC lacking ␤ subunit palmitoylation sites significantly activated the channel (gray boxes) (p Ͻ 0.01 for DHHC 1 or 2 versus NA, p Ͻ 0.05 for DHHC14 versus NA, determined with one-way ANOVA followed by a Tukey test). Whiskers indicate the 10th and 90th percentiles. The median is indicated by a horizontal line, and the mean is indicated with a cross symbol within each bar. In summary, we have identified five ENaC-activating DHHCs and found that ENaC was present in a stable complex (or complexes) with DHHCs 1, 2, 3, 7, and 14. Additional nonactivating DHHCs such as DHHC 11 and 23 are likely present in the complex (or complexes) because they were also found in ENaC immunoprecipitates when co-expressed in FRT cells. Complexes of heterogeneous DHHCs have been previously reported (40). It is also possible that ENaC could be activated by  . Transcripts for ENaC-activating DHHCs are expressed in mCCD cl1 cells. Single-stranded cDNA was generated from total RNA isolated from cultures of mCCD cl1 cells using reverse transcriptase (ϩ RT) and amplified using PCR with primer pairs specific for mouse DHHC 1, 2, 3, 7, or 14 that span at least one intron. The reactions without RT (Ϫ) were analyzed with PCR as a negative control. An aliquot of each PCR was analyzed on a 1% agarose gel as indicated and corresponded to the expected size of the amplified fragments (474 bp for DHHC1, 475 bp for DHHC2, 453 bp for DHHC3, 541 bp for DHHC7, and 420 bp for DHHC 14).  additional DHHCs in other cell types where unknown factors stabilize DHHC complexes. Because channel activating DHHCs are likely expressed at different sites in the biosynthetic pathway, including ER (for DHHCs 1, 2, 3, and 14), Golgi (DHHCs 2, 3, and 7) and endosomes (DHHCs 1 and 2), it is likely that ENaC is palmitoylated during transit through the biosynthetic pathway and during endocytic recycling at the apical cell surface.

Experimental Procedures
Plasmids-cDNAs for wild type mouse ␣, ␤, and ␥ ENaC subunits and mutant subunits (␤C43A,C557A and ␥C33A, C41A), with and without N-terminal HA and C-terminal V5 epitope tags, or an ␣ subunit with only a C-terminal V5 tag were described previously (12,20,21). The 23 cDNAs encoding individual mouse palmitoyltransferases with an N-terminal HA epitope tag (in pEF-Bos-HA) and with an N-terminal GFP tag (in pEGFP-C1) were a gift from Masaki Fukata (National Institute for Physiological Sciences, Okazaki, Japan) and were previously described (35). Each HA-tagged DHHC cDNA was subcloned into pCDNA 3.1(Ϫ)neo, and corresponding cRNAs for oocyte injections were prepared using T7 mMESSAGE mMACHINE kit (Ambion Invitrogen).
FRT cells were provided by P. Snyder (University of Iowa) and cultured in DMEM/F-12 medium supplemented with 7.5% fetal bovine serum as previously described (64). The cells were transfected with cDNAs using Lipofectamine 2000 (Invitrogen Thermo Fisher) according to the manufacturer. FRT cells were transfected with 0.5 g/subunit for ENaC and 0.5 g for each GFP-tagged DHHC. For co-immunoprecipitation experiments, FRT cells were transfected with N-terminal HA and C-terminal V5 tagged WT ␣␤␥ or ENaC with mutant subunits lacking sites for palmitoylation (␣␤Cys43,557A␥Cys33,41A) either alone or with N-terminal GFP tagged DHHC 1, 2, 3, 7, or 14 (cDNAs described below). The following day, detergent extracts of the cells were incubated with agarose-immobilized goat anti-V5 antibodies (Bethyl Laboratories, Inc., Montgomery, TX), and immunoprecipitates were subjected to SDS-PAGE and immunoblotting with either rabbit anti-GFP antibodies (at 2 g/ml; Thermo Fisher Molecular Probes) or mouse monoclonal anti-V5 antibodies (at 1 g/ml; Bio-Rad) as previously described (21). 10% of the detergent cell extracts was subjected to SDS-PAGE and immunoblotting with a rabbit anti-GFP antibody (at 2 g/ml). Alternatively, co-immuno-precipitation experiments were carried out with HA-DHHCs 7, 11, or 23 and ENaC with non-tagged ␤ and ␥ subunits and a C-terminal V5 epitope tagged ␣ subunit. Anti-V5 immunoprecipitates or cell extracts were immunoblotted with rat monoclonal anti-HA antibodies conjugated to HRP at 0.05 g/ml prepared as directed by the manufacturer (Roche Diagnostics).
Assay for Cys Palmitoylation in HEK293 Cells-HEK293T cells (ATCC Cell Biology Collection) were cultured in DMEM/ F-12 medium supplemented with 8% fetal bovine serum. The cells were plated in a 12-well size dish and transfected the next day with ␣␤␥ ENaC (␤ subunit with a C-terminal V5 epitope tag and non-tagged ␣ and ␥ subunits), with or without DHHC 7 (with an N-terminal HA tag) using Lipofectamine 2000 according to the manufacturer. Cys palmitoylation of the ␤ subunit in anti-V5 immunoprecipitates (IPs) was assessed with fatty acid exchange chemistry where palmitate is removed from Cys with hydroxylamine treatment, using Tris treatment as a negative control, and replaced with biotin as previously described (21). A fraction of the IP (10%) was reserved to assess total ␤ subunit in the initial immunoprecipitate, and biotinylated ␤ subunit was recovered with avidin-conjugated beads (90%) for immunoblotting with mouse anti-V5 antibodies at 4 g/ml (Invitrogen; R96025). The percentage of palmitoylation was calculated from the difference in ␤ subunit biotinylation after treatment with hydroxylamine or Tris, relative to the ␤ subunit recovered in the total immunoprecipitate.
Functional ENaC Expression in Xenopus Oocytes-Two-electrode voltage clamp was performed as previously described (65,66). Oocytes were injected with wild type or mutant ␣, ␤, and ␥ mouse ENaC subunit cRNAs (0.5-1 ng/subunit) and co-injected where noted with cRNAs for one of the 23 mouse DHHCs (3 ng). The oocytes were incubated at 18°C in modified Barth's saline (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 0.3 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , 0.82 mM MgSO 4 , 15 mM HEPES, 10 g/ml streptomycin sulfate, 100 g/ml gentamycin sulfate, pH 7.4) for 36 -48 h before electrophysiological recordings were performed at room temperature. Oocytes were placed in a recording chamber and perfused with a solution containing 110 mM NaCl, 2 mM KCl, 2 mM CaCl 2 , 10 mM HEPES, pH 7.4. Inward Na ϩ currents were measured at Ϫ100 mV in the absence and presence of amiloride (10 M). The protocol for harvesting oocytes from X. laevis was approved by the University of Pittsburgh Institutional Animal Care and Use Committee.
Immunoblotting of DHHCs Expressed in Xenopus Oocytes-20 oocytes were injected with cRNAs (12 ng) for 1 of the 23 mouse DHHCs containing an N-terminal HA epitope tag. After 48 h, the oocytes were extracted in detergent for immunoblotting as described previously (20,21). Briefly, surviving oocytes (n ϭ 5-17) were disrupted in homogenization buffer (100 mM NaCl, 50 mM Tris HCl, pH 7.4) including 1% protease inhibitor mixture III (EMD Millipore, Billerica, MA) using an allergy syringe (1 cc/ml, 27-gauge ϫ 0.5-inch needle; Terumo Medical Corporation, Elkton, MD) in a 1.5-ml capped tube (15 l buffer/oocyte). The homogenates were centrifuged twice at 200 ϫ g for 10 min, and the supernatant was removed from the pellet of yolk proteins, nuclei, and cell debris to a clean tube. Each supernatant was adjusted to 1% Triton X-100 by addition of 1/3 vol-ume of 4% Triton X-100 in homogenization buffer including protease inhibitors and incubated overnight on a rotating wheel at 4°C before centrifugation at 15,300 ϫ g at 4°C to remove any insoluble material. An aliquot equivalent to 1.2 oocytes was subjected to immunoblotting with mouse anti-␤-actin ascites (A5441 clone AC-15, diluted 1:1000; Sigma-Aldrich) or rat anti-HA antibodies conjugated to HRP (0.05 g/ml; Roche Diagnostics). Signal for some DHHCs on immunoblots was observed only after increasing the amount of cRNA injected into oocytes (388 ng for DHHC 4, 46 ng for DHHC 8, 67 ng for DHHC 11, 79 ng for DHHC 13, and 67 ng for DHHC 23). No signal for DHHC21 was observed after injection of 20 ng.
Immunostaining and Imaging of Mouse Kidneys-The protocol for harvesting kidneys from mice was approved by the University of Pittsburgh Institutional Animal Care and Use Committee. Kidneys from C57BL/6 male mice were fixed in 4% PFA and embedded in Tissue-Tek optimal cutting temperature compound before staining 4 -5-m cryosections with goat anti-aquaporin 2 (C-17) (0.2 g/ml Santa Cruz Biotechnology, sc-9882) and rabbit anti-DHHC3 (at 3.3 g/ml; ab31837, Abcam, Cambridge, MA) antibodies, as previously described (67,68). Validation of the anti-DHHC 3 antibody is described below. Secondary antibodies (FITC-tagged donkey anti-goat IgG (at 7.5 g/ml, catalog no. 705-095-147) and Cy3-tagged donkey anti-rabbit IgG (at 1.9 g/ml, catalog no. 711-165-152)) were purchased from Jackson ImmunoResearch (West Grove, PA). The nuclei were counterstained with TO-PRO 3 (Thermo Fisher). Sections were imaged on a Leica TCS SP5 CW-STED confocal imaging system (without STED settings) with a 63ϫ glycerol immersion lens with a 1.25 numerical aperture. The images were acquired with Leica proprietary software, and scale bars were generated with ImageJ. For the peptide competition immunofluorescence, anti-DHHC antibody (6.6 g/ml) was preincubated with and without antigen peptide from Abcam (Human GODZ peptide ab31882 at 0.1 g/ml) for 90 min at 4°C with agitation and then combined with an equal volume of anti-aquaporin 2 antibody (0.4 g/ml) before primary antibodies were applied to tissue.
The anti-DHHC 3 antibody was further validated by immunoblotting. FRT cells were transfected with the cDNA for either GFP or DHHC 3 with an N-terminal GFP epitope tag using Lipofectamine 2000 (Invitrogen Thermo Fisher) according to the manufacturer. Cells were extracted in detergent, and an aliquot was subjected to immunoblotting with either mouse anti-␤-actin ascites (A5441 clone AC-15, diluted 1:1000; Sigma-Aldrich) or rabbit anti-DHHC 3 antibody (1 g/ml, ab31837; Abcam). Kidneys from three C57BL/6 mice were homogenized separately, and detergent extracts (30 g) of the post-nuclear supernatant were subjected to immunoblotting on two identical blots with rabbit anti-DHHC 3 antibody (1 g/ml, ab31837; Abcam) preincubated with and without the immunizing peptide in 10-fold excess (0.75 g of antibody with 7.5 g of peptide, ab31882;Abcam). Both blots were also immunoblotted with mouse anti-␤-actin ascites (A5441 clone AC-15, diluted 1:1000; Sigma-Aldrich) as a loading control. Signal for DHHC 3 on immunoblots was detected using either BioMax MR film (Carestream Health, Inc., Rochester, NY) or a Bio-Rad ChemiDoc Touch imaging system. I sc Measurements-mCCD cl1 cells were grown on Snapwell inserts (Corning Costar) until a confluent, high resistance monolayer was obtained. Monolayer resistances were greater than 1 k⍀ ϫ cm 2 for all cell monolayers tested. Snapwells were mounted in Ussing sliders (P2302; Physiological Instruments, San Diego, CA) and inserted into the chambers of an EM-CSYS Ussing system (Physiologic Instruments) equipped with a heat block for temperature control. The apical and basolateral hemichambers contained 4 ml of Krebs buffer solution (110 mM NaCl, 25 mM NaHCO 3 , 5.8 mM KCl, 2 mM MgSO 4, 1.2 mM K 2 HPO 4 , 2 mM CaCl 2 , and 11 mM glucose). The chamber temperature was maintained at 37°C. The hemichambers were continuously bubbled with 95% O 2 , 5% CO 2 , which maintained the pH at 7.4. The apical and basolateral hemichambers were connected to Ag/AgCl electrodes via 5 M NaCl agar bridges for voltage sensing and current passing. The electrodes were connected to a VCC MC6 multichannel voltage/current clamp (Physiologic Instruments). The asymmetry of voltage-sensing Ag/AgCl electrodes and the liquid junction potentials were compensated using an offset removal circuit before tissue mounting. I sc and TER were measured under voltage clamp conditions. To calculate TER, a bipolar pulse of Ϯ 10 mV with a duration of 0.5 s was applied every 60 s. The data were digitalized at 1 KHz using DigiData 1440A (Molecular Devices) and acquired with pClamp 10.3 software (Molecular Devices). After an equilibration period to achieve a stable I sc (ϳ30 min), the cells were treated with either 25 M 2-BP (2-bromohexadecanoic acid (97% pure); Aldrich) or vehicle only (DMSO (1:2,000 dilution) in the apical hemi-chamber for 30 min. The amiloride-sensitive component of the I sc was then determined by adding 10 M amiloride to the apical hemi-chamber. Typically, Ն90% of the total I sc was inhibited after amiloride addition. Average TER was calculated from five bipolar pulses of Ϯ 10 mV prior to addition of DMSO or 2-BP and five pulses prior to addition of amiloride. After 5 min, amiloride was washed out, and the I sc was allowed to stabilize again. The apical membrane ENaC Activation by Specific Palmitoyltransferases MARCH 10, 2017 • VOLUME 292 • NUMBER 10 was then permeabilized by the addition of amphotericin B (120 g/ml; Sigma).
Data and Statistical Analyses-The data are expressed as the means Ϯ S.D. Box and whisker plots are used in figures to show the distribution of data: median (middle line); mean (cross or dot); 25th to 75th percentile (box); and 10th and 90th percentiles (whisker). Experiments were repeated with a minimum of two batches of oocytes obtained from different frogs. When appropriate, statistical comparisons were obtained from unpaired Student's t test or one-way ANOVA followed by a Tukey post hoc test. A p value of less than 0.05 was considered statistically different.