Paraoxonase 3 functions as a chaperone to decrease functional expression of the epithelial sodium channel

The paraoxonase (PON) family comprises three highly conserved members: PON1, PON2, and PON3. They are orthologs of Caenorhabditis elegans MEC-6, an endoplasmic reticulum–resident chaperone that has a critical role in proper assembly and surface expression of the touch-sensing degenerin channel in nematodes. We have shown recently that MEC-6 and PON2 negatively regulate functional expression of the epithelial Na+ channel (ENaC), suggesting that the chaperone function is conserved within this family. We hypothesized that other PON family members also modulate ion channel expression. Pon3 is specifically expressed in the aldosterone-sensitive distal tubules in the mouse kidney. We found here that knocking down endogenous Pon3 in mouse cortical collecting duct cells enhanced Na+ transport, which was associated with increased γENaC abundance. We further examined Pon3 regulation of ENaC in two heterologous expression systems, Fisher rat thyroid cells and Xenopus oocytes. Pon3 coimmunoprecipitated with each of the three ENaC subunits in Fisher rat thyroid cells. As a result of this interaction, the whole-cell and surface abundance of ENaC α and γ subunits was reduced by Pon3. When expressed in oocytes, Pon3 inhibited ENaC-mediated amiloride-sensitive Na+ currents, in part by reducing the surface expression of ENaC. In contrast, Pon3 did not alter the response of ENaC to chymotrypsin-mediated proteolytic activation or [2-(trimethylammonium)ethyl]methanethiosulfonate–induced activation of αβS518Cγ, suggesting that Pon3 does not affect channel open probability. Together, our results suggest that PON3 regulates ENaC expression by inhibiting its biogenesis and/or trafficking.

The epithelial Na ϩ channel (ENaC) 2 mediates the rate-limiting step of Na ϩ uptake across the apical membrane of epithe-lia, including kidney tubules, alveoli, and distal colon (1)(2)(3). ENaC-dependent Na ϩ reabsorption in the kidney is essential for regulating extracellular fluid volume and blood pressure (BP) as well as extracellular [K ϩ ] (4 -8). ENaC-dependent Na ϩ absorption in the airway has important roles in airway surface liquid volume maintenance and mucociliary clearance (9 -12). ENaC gain-of-function mutations, as found in patients with Liddle syndrome, result in Na ϩ retention, extracellular volume expansion, hypertension, and hypokalemia (13)(14)(15)(16)(17). On the other hand, ENaC loss-of-function mutations result in renal Na ϩ wasting, hypotension, and hyperkalemia (18 -20). These changes in extracellular [K ϩ ] will likely impact the activity of the Na-Cl cotransporter (NCC), contributing to changes in BP seen in these disorders (21)(22)(23)(24). Therefore, the expression and activity of ENaC need to be tightly regulated by various endogenous and extracellular factors.
The canonical ENaC consists of three homologous subunits: ␣, ␤, and ␥ (25,26). The fourth ENaC subunit, ␦ has been identified in multiple tissues of primates and other species (27,28). ENaC is synthesized in the endoplasmic reticulum, where nascent ␣, ␤, and ␥ subunits are folded and assembled into trimeric channel complexes that are shuttled to the Golgi apparatus for additional processing and then inserted into the plasma membrane (29 -33). Misfolded subunits are subjected to endoplasmic reticulum-associated degradation via the proteasome (34 -36). When at the cell surface, ENaC can be ubiquitinated by E3 ligases, which promotes channel internalization by clathrin-mediated endocytosis (37,38). The retrieved channel complex may then be shuttled to the lysosome for degradation or deubiquitinated and recycled back to the cell surface (39). Together, this multistage machinery determines the total number of functional channels at the cell surface. Molecular chaperones have been implicated in multiple key steps during ENaC biogenesis, trafficking, and degradation (34, 40 -45), including members of the paraoxonase (PON) family.
There are three highly conserved genes in the mammalian PON family: PON1, PON2, and PON3. Although all three PONs have antioxidative and anti-atherosclerotic properties (46 -50), they differ in their tissues expression and substrate specificity. PON2 is ubiquitously expressed in various tissues and most major organs, whereas PON1 and PON3 were thought to be mainly synthesized by the liver and circulate within high-density lipoprotein particles (47,(51)(52)(53). However, recent studies have detected PON3 in epithelial cells within many tissues, including the digestive tract, respiratory system, reproductive system, and kidneys (54). Although all three PONs possess lactonase activity, only PON1 can hydrolyze organophosphates, such as paraoxon and its derivatives (48,51). PONs share evolutionary conservation with Caenorhabditis elegans MEC-6 and several other predicted proteins in nematodes. MEC-6 is a chaperone protein that is required for proper folding, assembly, and surface expression of the touch-sensitive MEC-4/MEC-10 channel in the nematode's touch receptor neurons (55,56). We have shown recently that PON2 inhibits ENaC activity by reducing channel surface expression (57), suggesting that the chaperone function is conserved between mammalian PONs and their nematode orthologs.
There is evidence suggesting that mammalian PONs have important roles in kidney function. Pon1 KO mice were hypotensive with a lower serum aldosterone level (58). Knockdown (KD) of Pon2 by renal subcapsular infusion of Pon2 siRNA led to hypertension in rats (59). Although it is unclear whether altered ENaC activity and/or expression contribute to the changes in BP in these models, they collectively support a role of PONs in maintaining normal BP. Being the most neglected member of this family, the role of PON3 in BP control has not been investigated in whole-animal studies. However, Pon3 transcripts have been detected in multiple nephron segments in the rat kidney and in principal cells and intercalated cells of the mouse distal nephron (60,61). PON3 shares high sequence homology with PON2 (Ͼ60%), which led us to hypothesize that PON3 also functions as a chaperone to regulate ENaC expression. This study examined the expression of Pon3 in the mouse kidney and its effect on ENaC functional expression by silencing endogenous Pon3 in mouse cortical collecting duct cells or overexpressing Pon3 in cultured FRT cells or Xenopus oocytes.

Pon3 expression in the mouse kidney
To determine whether ENaC is an endogenous substrate for PON3, we first asked whether PON3 localizes to the aldosterone-sensitive distal nephron where ENaC resides. Despite the detection of Pon3 transcripts in multiple nephron segments and cell types in rodents (60,61), the expression of PON3 protein in the kidney has not been thoroughly investigated. We approached this question with immunofluorescence staining. Kidney sections from WT C57BL/6 mice were incubated with an anti-PON3 antibody and imaged from the cortex to the medulla using a tile scan technique. As shown in the automatically merged tile scans (Fig. 1A), Pon3 localizes mainly in tubular epithelial cells within the cortex but is also present in a specific population of cells of the medulla. To further define the tubular expression of Pon3, we performed costaining experiments using specific markers of different nephron segments. As shown in the overlaid images (Fig. 1B), Pon3 was absent in tubules showing positive staining of the Na-K-Cl cotransporter or parvalbumin, suggesting that Pon3 is not expressed in the thick ascending limb or the early distal convoluted tubule (DCT1 (62,63)). In contrast, we detected Pon3 in a subpopulation of NCC-positive tubules, likely to be the late DCT (DCT2, Fig. 1B). Of a total of 148 NCC-positive tubules, 91 tubules showed positive staining for Pon3. Within these NCC/Pon3 double-positive tubules, Pon3 was ubiquitously expressed in nearly all NCC-positive cells. As shown in Fig. 1B, Pon3 also colocalized with the cytoplasmic Ca 2ϩ -binding protein calbindin D28K (CaBP 28K ), a marker for distal tubules with higher abundance in the DCT2 and connecting tubule (CNT) and lower expression in the DCT1 and collecting duct (CD) (64,65). Among 78 distal tubules with strong CaBP 28K staining, 90.6% Ϯ 5.3% cells expressed both CaBP 28K and Pon3. To further examine in which cell type Pon3 is expressed, WT mouse kidney sections were costained for Pon3 with the principal cell (PC) marker aquaporin 2 (AQP2) or the intercalated cell (IC) marker vacuolar H ϩ -ATPase (V-ATPase) (66,67). As shown in the merged images in Fig. 1B, Pon3 was expressed in both PCs and ICs but with distinctive expression patterns. Nearly all ICs within a total of 97 V-ATPase-positive tubules were also positive for Pon3. Although all AQP2-positive tubules (n ϭ 115) exhibited Pon3 staining, only 25.1% Ϯ 8.8% PCs within AQP2/Pon3 double-positive tubules expressed both AQP2 and Pon3. The specificity of the anti-PON3 antibody was validated with kidney sections of Pon3 KO mice. We did not observe significant Pon3 staining in tubules from KO animals and only weak staining within glomeruli (Fig.  1C). Together, our data suggest that Pon3 is primarily expressed in the DCT2, CNT, and CD of the aldosteronesensitive distal nephron.

Pon3 KD in mouse CCD (mCCD) cells enhances ENaC-mediated Na ؉ transport
As Pon3 is expressed in principal cells of the distal nephron ( Fig. 1B), we examined the role of endogenous Pon3 in regulating ENaC expression and function with siRNA-mediated KD. mCCD cells, derived from the mouse cortical CD, express all three ENaC subunits as well as mineralocorticoid and glucocorticoids receptors (68). As expected, endogenous Pon3 was detected as a single band near 40 kDa in mCCD cells (Fig. 2C). To knock down Pon3, mCCD cells were transiently transfected with Pon3-specific siRNA, and ENaC-mediated Na ϩ transport was determined with short-circuit current (I sc ) measurements. As shown in Fig. 2, Pon3 expression in mCCD cells was reduced 55% Ϯ 14% (n ϭ 11, p Ͻ 0.001) compared with negative control (NC) cells transfected with scrambled siRNAs (Fig. 2D). As a result, we observed a 1.2 Ϯ 0.2-fold (n ϭ 12, p Ͻ 0.01) increase in amiloride-sensitive I sc in Pon3 KD cells (Fig. 2B) without noticeable changes in transepithelial resistance (1.13 Ϯ 0.49 kilo-ohm for Pon3 KD cells versus 1.22 Ϯ 0.34 kilo-ohm for NC cells, n ϭ 12, p ϭ 0.62). The enhanced Na ϩ transport in Pon3 KD cells was associated with a 1.4 Ϯ 0.4-fold increase in the abundance of cleaved ␥ENaC (n ϭ 11, p Ͻ 0.05). The abundance of uncleaved ␥ENaC was not affected by Pon3 KD (Fig.  2E). In addition, the ratio of cleaved versus total ␥ENaC was PON3 regulates ENaC significantly higher in Pon3 KD cells (Fig. 2F), suggesting enhanced channel proteolytic processing. Together, our data suggest that Pon3 regulates endogenous ENaC functional expression in mCCD cells.

Pon3 interacts with ENaC subunits
We then performed coimmunoprecipitation to investigate whether PON3 forms a complex with ENaC subunits. Briefly, we transfected FRT cells with mouse Pon3 that had a C-termi-

PON3 regulates ENaC
nal Myc-FLAG epitope tag, mouse ENaC, or both. In each case, only one ENaC subunit had an N-terminal HA and a C-terminal V5 epitope tag ( HA ␣ V5 ␤␥, ␣ HA ␤ V5 ␥, or ␣␤ HA ␥ V5 ). Whole-cell lysates were either precipitated with anti-V5 antibodies to pulldown ENaC and probe for Pon3 (Fig. 3A) or precipitated with anti-FLAG antibodies to pulldown Pon3 and probe for ENaC Transfected cells were then cultured on permeable filters, and amiloride (Amil)-sensitive I sc were determined with an Ussing chamber assay. A, representative traces in NC cells (black) or Pon3 KD cells (red). After reaching a steady current, 10 M amiloride (black bar) was added to the apical side of the Ussing chamber to block ENaC. B, summary of the amiloride-sensitive I sc of Pon3 KD cells normalized (Norm.) to the average I sc of NC cells from the same experiment. C, following the recordings, whole-cell lysates were collected from the filters to blot for ␥ENaC, Pon3, and actin as a loading control. The asterisk and arrowhead indicate uncleaved and cleaved ␥ENaC, respectively. D and E, whole-cell expression of Pon3 (D) and ␥ENaC (E) were normalized to actin and are shown as a percentage of NC cells. F, to access the extent of ␥ENaC cleavage, the ratio of the cleaved ␥ENaC to total ␥ENaC (cleaved and uncleaved) was estimated for each sample and is shown as a percentage of NC cells. Experiments were repeated a total of four times with three individual transfections of scrambled siRNA or Pon3 siRNA for each experiment. Data are shown in scatter-dot plots, with a horizontal bar indicating the mean. Statistical comparisons were analyzed with a nonparametric Mann-Whitney test or with one-way ANOVA followed by Sidak's multiple comparisons test. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.

PON3 regulates ENaC
subunits (Fig. 3B). As shown in Fig. 3A, Pon3-Myc-FLAG (ϳ45 kDa) was detected in V5 precipitates of FRT cells coexpressing Pon3 and ENaC but not in cells transfected with only Pon3. Vice versa, ENaC subunits could be detected only in FLAG precipitates of FRT cells coexpressing Pon3 and ENaC but not in cells transfected with ENaC alone (Fig. 3B). Expression of Pon3 or ENaC subunits was detected in whole-cell lysate as expected (input). Our data suggest that Pon3 specifically interacts with ENaC subunits when coexpressed in FRT cells.

Pon3 reduces ENaC expression in FRT cells
As ENaC functional expression was enhanced by Pon3 KD in mCCD cells, we examined whether overexpressing Pon3 affects ENaC expression in FRT cells. We cotransfected FRT cells with ENaC ( HA ␣ V5 ␤␥) with either the Pon3 plasmid or an equal amount of the empty vector. Surface proteins were biotinylated on ice and recovered with NeutrAvidin beads. The expression of ENaC ␣ and ␥ subunits was assessed in surface precipitates and whole-cell lysates. GAPDH was absent in surface precipitates, indicating that the surface fractions were free of contamination from intracellular proteins (Fig. 4A, bottom). In FRT cells coexpressing Pon3, whole-cell ENaC expression was significantly lower for full-length ␣ENaC (48% Ϯ 22%, n ϭ 5-6, p Ͻ 0.01), the cleaved ␣ENaC (64% Ϯ 14%, n ϭ 5-6, p Ͻ 0.001) and ␥ENaC (41% Ϯ 13%, n ϭ 5-6, p Ͻ 0.01, Fig. 4B). In contrast, Pon3 did not alter the ratio of surface abundance versus whole-cell abundance for ␣ENaC or ␥ENaC (Fig. 4C), suggesting that Pon3 not only reduced ENaC subunit whole-cell expression but also decreased channel density at the cell surface in FRT cells. This inhibitory effect of Pon3 is consistent with our findings of the effects of endogenous Pon3 on ENaC in mCCD cells (Fig. 2). As a control, we tested whether ␥-glutamyl transferase (␥GT), an unrelated type II protein affects ENaC expression in FRT cells. Our data suggest that coexpressing mouse ␥GT has no effect on ENaC abundance at the whole-cell level in FRT cells (Fig. 4, D and E).

Pon3 inhibits ENaC activity by reducing channel surface expression in Xenopus oocytes
To further investigate the role of Pon3 in regulating ENaC functional expression, we injected mature oocytes with cRNA mixtures encoding WT ENaC alone or with an equal amount of mouse Pon3 cRNA. Whole-cell Na ϩ currents were measured 24 -30 h after injection by clamping oocytes at Ϫ100 mV (Fig.  5A). The average amiloride-sensitive Na ϩ current in oocytes coexpressing ENaC and Pon3 (Ϫ1.4 Ϯ 1.4 A, n ϭ 60) is ϳ40% lower than in oocytes expressing ENaC alone (Ϫ2.4 Ϯ 1.8 A, n ϭ 62, p Ͻ 0.01). This is similar to the extent of inhibition of PON2 on ENaC activity (57). The inhibitory effect of Pon3 is specific to ENaC, as Pon3 did not alter renal outer medullary K ϩ (Romk) channel activity (Ϫ2.5 Ϯ 1.8 A in oocytes expressing Romk versus Ϫ2.2 Ϯ 1.9 A in oocytes coexpressing Pon3 and Romk, Fig. 5C).
The inhibitory effect of Pon3 on ENaC activity suggests a reduction in channel open probability (P O ) and/or in the number of functional channels at the cell surface. As Pon3 reduced ENaC surface expression in FRT cells (Fig. 4), we examined the effect of Pon3 on channel surface expression in oocytes using a FLAG epitope-tagged ENaC ␤ subunit, as described previously (69). As shown in Fig. 6A, the surface expression of the nontagged channel was 0.2 Ϯ 0.5 ϫ 10 6 (shown as relative light units, n ϭ 21) but rose to 1.6 Ϯ 1.8 ϫ 10 7 in oocytes expressing FLAG-tagged ENaC (n ϭ 33, p Ͻ 0.0001). Notably, this value was significantly reduced in oocytes coexpressing Pon3 (3.3 Ϯ 3.3 ϫ 10 6 , n ϭ 36, p Ͻ 0.01). In contrast, whole-cell expression of ␤ENaC in oocytes was not altered by Pon3 (Fig. 6C).

Pon3 does not alter ENaC P O in Xenopus oocytes
ENaC is gated in response to external cues, such as proteolytic cleavage, extracellular Na ϩ , and laminar shear stress, that primarily modulate channel P O (70,71). Specific proteases activate ENaC by releasing the embedded inhibitory tracts from the ␣ or ␥ subunit, increasing channel P O (72)(73)(74)(75)(76). We examined whether Pon3 affects channel activation by a protease, ␣-chymotrypsin. Oocytes were treated with ␣-chymotrypsin (2 g/ml) for 2 min until currents reached a steady state (Fig.  7A). If Pon3 reduces channel P O , we predicted that the fold increase in ENaC current following ␣-chymotrypsin treatment would be altered. However, we found that chymotrypsin elicited a 1.9 Ϯ 1.0-fold increase in whole-cell Na ϩ currents in oocytes expressing ENaC and a similar 1.9 Ϯ 0.5-fold
To examine the effect of Pon3 on channel P O in a more direct manner, we measured the response of ENaC bearing a Cys mutation at the degenerin site of the ␤ subunit (␣␤ S518C ␥) to [2-(trimethylammonium)ethyl]methanethiosulfonate MTSET. The activity of this channel rises when a disulfide bond is formed between MTSET and the introduced Cys within the channel pore, reflecting an increase in channel P O (77,78). As shown in Fig. 8, the whole-cell Na ϩ currents in oocytes expressing ␣␤ S518C ␥ was increased 2.1 Ϯ 0.7-fold by MTSET (n ϭ 16). The presence of Pon3 did not alter the effect of MTSET on ENaC activity, as a 2.0 Ϯ 0.4-fold increase in whole-cell Na ϩ currents was observed in oocytes coexpressing ␣␤ S518C ␥ and Pon3 (n ϭ 16, Fig. 8B). Together, our data suggest that Pon3 inhibits ENaC activity in oocytes by reducing channel surface expression and not P O . ENaC was labeled with biotin and recovered with NeutrAvidin beads at 4°C. Blots were probed for the ␣ subunit, ␥ subunit, Pon3, or GAPDH in the biotinylated surface fraction and in 5% of the total whole-cell lysates. B, the abundance of the ␣ subunit (full-length 95 kDa and cleaved 30 kDa, indicated by arrowheads) or the ␥ subunit (ϳ75 kDa) in whole-cell lysate was normalized to the loading control (GAPDH) and expressed as the percentage in cells transfected with ENaC alone (ϪPon3). C, the abundance of the ␣ subunit (full-length 95 kDa and cleaved 30 kDa) or the ␥ subunit (ϳ75 kDa) at the cell surface was normalized to whole-cell abundance of the ␣ or ␥ subunit, respectively, and then expressed as the percentage in cells transfected with ENaC alone (ϪPon3). D, as a control, FRT cells were cotransfected with HA ␣ V5 ␤␥ and with either ␥-glutamyl transferase (ϩ␥GT) or an equal amount of the pCDNA3 vector (Ϫ␥GT). Whole-cell lysates were collected and probed for both ␣ENaC and ␥GT (indicated by the arrow). E, the abundance of the ␣ subunit (95 kDa and 30 kDa) was normalized to the loading control (GAPDH) and expressed as the percentage in cells transfected with ENaC alone (Ϫ␥GT). Experiments were repeated a total of three times with FRT cells of different passages. The summarized data are shown in a scatter-dot plot, with a horizontal bar indicating the mean. Statistical comparisons were analyzed with one-way ANOVA followed by Sidak's multiple comparisons test (**, p Ͻ 0.01; ***, p Ͻ 0.001).

Figure 5. Pon3 inhibits ENaC activity in oocytes.
A, representative traces of whole-cell Na ϩ currents measured in oocytes expressing mouse ENaC alone (black) or coexpressing Pon3 (ϩPon3, red), with the holding potential set at Ϫ100 mV. After reaching a steady current, 10 M amiloride (Amil, black bar) was added to the bath to block ENaC. B, summary of the amiloride-sensitive Na ϩ currents in oocytes expressing ENaC or ENaC and Pon3. C, summary of the BaCl 2 -sensitive K ϩ currents of oocytes injected with Romk with or without Pon3. Experiments were repeated with four batches of oocytes, and the pooled data are shown in scatter-dot plots, with a horizontal bar indicating the mean. The number of oocytes assessed for each condition is listed. Statistical comparisons were analyzed with nonparametric Mann-Whitney tests (**, p Ͻ 0.01).

PON3 regulates ENaC
Discussion PON family members are encoded by highly conserved genes, although the three members have distinct substrate specificities and tissue expression (46,47,51). A recent study by Clark and co-workers (79) identified convergent functional loss of PON1 but not PON2 or PON3 in marine mammals. PON1 contained lesions within the coding region, rendering marine mammals susceptible to neurotoxicity induced by organophosphorus compounds. In contrast, PON1 was intact in all 53 surveyed terrestrial mammal genomes (79). Like PON1, PON3 is highly expressed in the gallbladder and liver (80), and Pon3 KO mice exhibit altered metabolism of lipid and bile acids. They are also susceptible to developing obesity, atherosclerosis, and large gallstones when fed a diet containing cholic acid and cholesterol (81). ENaC and the bile acid-sensitive ion channel, a related member of the ENaC/degenerin family, are expressed in cholangiocytes lining bile ducts, and their channel activities can be regulated by unconjugated and conjugated bile acids (82)(83)(84)(85).
As the least-studied member in the PON family, little is known about the expression or function of PON3 in other organs, such as the kidneys. We show, for the first time, that Pon3 is specifically expressed in distal nephron segments of mouse kidneys, including the DCT2, CNT, and CD. These are the same nephron segments where ENaC resides. In the CNT and CD, finetuning of Na ϩ reabsorption takes place under the control of aldosterone (86 -93). Interestingly, we observed differential expression of Pon3 in distal nephron segments. It appears to have homogeneous expression within the NCC-positive DCT2 and CaBP 28K -postive CNT, where Pon3 is nearly present in all cells lining these tubules. In contrast, Pon3 has higher expression in V-ATPasepositive ICs than in AQP2-positive PCs of CDs. This is consistent with a previous study showing that higher levels of Pon3 transcripts were found in ICs of mouse CDs (60). In addition, the expression of Pon3 is quite variable in PCs of CDs, with only ϳ25% cells coexpressing both Pon3 and AQP2. This is particularly intriguing because the magnitude of ENaC-mediated Na ϩ reabsorption differs in these segments. It has been shown that CD-specific ␣ENaC KO mice were able to maintain normal Na ϩ /K ϩ homeostasis when challenged with dietary salt/water restriction (94). When ␣ENaC was deleted in the CNT and CD, KO mice developed a severe salt-wasting phenotype and a continuous reduction in body weight (95). In  Oocytes were perfused with ␣-chymotrypsin (gray bar, 2 g/ml) for 2 min until the current reached a steady state. Amiloride (Amil) was added to the bath at the beginning and the end of each recording (black bars). B, amiloride-sensitive whole-cell Na ϩ currents (I_Amil) measured prior to (I basal ) and following chymotrypsin treatment (I chymo ) are shown for individual oocytes. Statistical comparisons were analyzed with two-way ANOVA followed by Sidak's multiple comparisons test (**, p Ͻ 0.01; ****, p Ͻ 0.0001). C, the -fold increase of ENaC currents in response to ␣-chymotrypsin (I chymo /I basal ) was estimated for individual oocytes. Data pooled from three experiments are shown in a scatter-dot plot, with the horizontal bar indicating the mean. The number of oocytes assessed for each group is listed. Statistical comparisons were analyzed with a nonparametric Mann-Whitney test.

PON3 regulates ENaC
addition, ENaC currents are significantly greater in the CNT compared with the cortical CD (96,97). Together, these studies suggest that ENaC expression in the CNT has a key role in Na ϩ homeostasis in mice. As PON3 is mainly expressed in principal cells of the CNT, we predict that PON3 has an important role in regulating renal Na ϩ /K ϩ homeostasis and BP control.
This study examined the role of PON3 in regulating ENaC functional expression in different expression systems. We found that Pon3 KD in mCCD cells led to enhanced amiloridesensitive Na ϩ transport, a higher abundance of a cleaved ␥ENaC product (ϳ75 kDa), and higher ratios of cleaved ␥ENaC versus total ␥ENaC, suggesting up-regulation of ENaC proteolytic processing and surface expression in mCCD cells when Pon3 expression was reduced. Supporting this notion, surface expression of ENaC subunits was reduced by Pon3 coexpression in FRT cells and Xenopus oocytes. It is interesting to note that the anti-␥ENaC antibody recognized two forms (cleaved and uncleaved) of endogenous ␥ENaC in mCCD cells but favored cleaved ␥ENaC when ENaC was overexpressed in FRT cells. Nonetheless, our results using cultured epithelial cells or Xenopus oocytes support that PON3 decreases ENaC functional expression. Other members of this family, including PON2 and MEC-6, also inhibit ENaC activity in oocytes (57). This negative regulation of channel activity is of great interest to us, as ENaC hyperactivity has physiological consequences (11,13,(98)(99)(100)(101). There is growing evidence suggesting that members of the PON family, including MEC-6, PON2, and PON3, function as chaperones to modulate the expression of ion channels within the ENaC/degenerin family (56,57). MEC-6 is not only required for proper folding and assembly of MEC-4 channels but also to facilitate channel surface expression (56). Little is known regarding the chaperone function of PONs. We found that coexpression of Pon3 reduced ENaC whole-cell expression in FRT cells, suggesting that this trimeric channel is less stable in the presence of Pon3. Molecular chaperones have important roles in ENaC quality control within the early secretory pathway (34,42,102). It is possible that PON3 facilitates proteasomal degradation of misfolded ENaC subunits. Other chaperones have been identified that affect ENaC trafficking at later sites in the secretory pathway (40,41,43,44,103). Additional studies are needed to investigate mechanisms by which PONs regulate ENaC expression and the physiological relevance of this regulation in the kidneys and other organs.

Plasmids
Mouse Pon3 was cloned in the pCMV6 vector (MR220409, OriGene) and has a C-terminal Myc-FLAG tag. The mouse ENaC ␣, ␤, and ␥ subunits with an N-terminal HA epitope tag and a C-terminal V5 epitope tag were cloned in the pcDNA3 vector and used to transfect FRT cells. WT or mutant mouse ENaC ␣, ␤, and ␥ subunits cloned in the pBlueScript vector were used for expressing ENaC in oocytes. Mouse ␥GT was cloned in the pcDNA3 vector (104).

PON3 regulates ENaC
Retiga 4000R Fast 1394 camera. To quantify Pon3 expression in specific nephron segments and cell types, the staining was repeated with kidney tissues from four WT mice, and four or five fields were chosen randomly for each kidney sample. The fluorescence intensity of Pon3 staining in glomeruli was used as the threshold to identify positive Pon3 staining. Numbers of total cells as well as cells with positive staining of Pon3 and/or the chosen markers were manually counted for each individual PON3-positive tubule in the field. The specificity of the anti-PON3 antibody was examined in kidney sections from Pon3 KO mice.

FRT cell coimmunoprecipitation and surface biotinylation
FRT cells were cultured in DMEM/F12 medium supplemented with 8% FBS. For the coimmunoprecipitation assay, FRT cells were seeded on 6-well dishes and transfected with plasmids encoding mouse Pon3 and mouse ENaC subunits (0.5 g/construct). In each case, only one ENaC subunit had an N-terminal HA tag and a C-terminal V5 epitope tag ( HA ␣ V5 ␤␥, ␣ HA ␤ V5 ␥ or ␣␤ HA ␥ V5 ). The next day, cells were extracted with detergent solution supplemented with protease inhibitor mixture. Five percent of the cell lysate was saved as input. The remainder was incubated overnight with 50 l of agarose-immobilized anti-V5 antibodies (S190-119, Bethyl) with endover-end mixing at 4°C to pull down ENaC subunits. To pull down Pon3, the lysates were incubated with anti-FLAG antibodies (2 g, F3165, Sigma) in the presence of rec-protein G-Sepharose (10-1241, Invitrogen). Proteins were eluted into Laemmli sample buffer (1610737, Bio-Rad) by heating the isolated beads at 95°C. The immunoprecipitate and the input were subjected to SDS-PAGE and blotted for Pon3 (0.1 g/ml, Sigma, HPA014848) or for ENaC subunits with V5 antibody (0.2 g/ml, Invitrogen, R96025). Assays were repeated three times for each condition.
To measure surface expression of ENaC, FRT cells were transfected with plasmids encoding mouse HA ␣ V5 ␤␥ with or without the Pon3 plasmid (0.5 g/construct) directly on 24-mm Transwell filters and cultured to confluency. In the control group, an equal amount of the pCMV6 vector was included. FRT cells grown on the filters were washed four times with cold Dulbecco's PBS with 1.0 mM CaCl 2 and 0.5 mM MgCl 2 (PBS, Corning Life Sciences). To label apical surface proteins, 1 mg/ml EZ-Link TM Sulfo-NHS-SS-Biotin (21331, Thermo Fisher) in a buffer containing 137 mM NaCl and 15 mM sodium borate (pH 9.0) was added to the apical side, whereas the basolateral side was incubated with just the buffer. Excess biotin was then quenched with 10% FBS in DMEM/F12. After washing with PBS, cells were lysed in detergent solution and incubated with NeutrAvidin agarose (29200, Thermo Fisher) overnight at 4°C to isolate biotinylated proteins. The recovered surface proteins and 5% of the total lysate were separated by SDS-PAGE and blotted for the ␣ subunit with HA-HRP antibodies (0.05 g/ml, 3F10, Sigma), the ␥ subunit (0.3 g/ml, StressMarq), PON3 (0.03 g/ml, Sigma), or GAPDH (0.3 g/ml, Proteintech, HRP-60004), as described previously (29,106).
In some experiments, FRT cells seeded on 6-well size plates were transfected with plasmids encoding mouse ENaC ( HA ␣ V5 ␤␥) or mouse ␥GT (0.5 g/construct). An equal amount of the pCDNA3 vector was used in the control (Ϫ␥GT) group. The next day, whole-cell lysates were extracted with detergent solution supplemented with protease inhibitors and subjected to SDS-PAGE. ␣ENaC was detected with HA-HRP antibodies (0.05 g/ml, 3F10, Sigma). ␥GT was detected using a rabbit polyclonal antibody described previously (107).

Xenopus oocyte expression, two-electrode voltage clamp, and surface expression
cRNAs of mouse ENaC ␣, ␤, and ␥ subunits; mouse Pon3; and rat Romk were synthesized using mMESSAGE mMA-CHINE TM transcription kits (Invitrogen) following the manufacturer's protocol. Oocytes were harvested from Xenopus laevis following a protocol approved by the University of Pittsburgh Institutional Animal Care and Use Committee. Stage V-VI oocytes were injected with 2 ng of cRNA per ENaC subunit or 1 ng of rat Romk cRNA. An equal amount of mouse Pon3 cRNA was injected for the coexpression assay. The injected oocytes were incubated in modified Barth's saline (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 15 mM HEPES, 0.3 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , 0.82 mM MgSO 4 , 10 g/ml sodium penicillin, 10 g/ml streptomycin sulfate, and 100 g/ml gen-PON3 regulates ENaC tamycin sulfate (pH 7.4)) at 18°C for an additional 24 -48 h for optimalchannelexpression.Whole-cellNa ϩ currentsweremeasured by clamping oocytes at Ϫ100 mV, whereas oocytes were continuously perfused with a solution containing 110 mM NaCl, 2 mM KCl, 1.6 mM CaCl 2 , and 10 mM HEPES (pH adjusted to 7.4). To examine the effect of Pon3 on ENaC gating in response to MTSET, oocytes were injected with cRNAs encoding the WT mouse ENaC ␣ and ␥ subunits and the degenerin mutation ␤ S518C with or without an equal amount of mouse Pon3 cRNA. MTSET was freshly prepared and delivered to oocytes through perfusion. To measure the Pon3 effect on Romk activity, we replaced the NaCl in the perfusion solution with 100 mM KCl. At the end of each recording, BaCl 2 was applied at a final concentration of 5 mM to block Romk activity. Voltage clamping was performed using a GeneClamp 500B amplifier and Digi-Data 1440A interface (Molecular Devices). Electrophysiological data were analyzed with Clampfit 10.5 and plotted with Origin 2015 (OriginLab).
ENaC surface expression in oocytes was measured using a chemiluminescence assay as described previously (69). The mouse ENaC ␤ subunit with an extracellular FLAG epitope tag was included in the cRNA mixture. A nontagged WT ␤ subunit was used as a negative control. Surface expression was assessed 2 days after cRNA injection by incubating oocytes with anti-FLAG M2 antibodies (1 g/ml, Sigma F3165) and then with anti-mouse HRP antibodies (1 g/ml, Jackson ImmunoResearch Laboratories, 115036072). Chemiluminescence of each oocyte was developed with Super Signal ELISA Femto Maximum Sensitivity Substrate (37075, Thermo Scientific) and quantified as relative light units with a GloMax-Multi detection system (Promega). To investigate the effect of Pon3 on ENaC whole-cell expression, oocytes injected with ENaC or ENaC and Pon3 were homogenized, and whole-cell lysates were subjected to SDS-PAGE to probe for ␤ENaC (1 g/ml, StressMarq, SPC-404D) or GAPDH (0.3 g/ml). All oocytes experiments were repeated at least three times using different frogs.

Statistical analyses
Data were expressed as the mean Ϯ S.D. in the main text and are shown as scatter-dot plots, with a horizontal bar indicating the mean. Data distribution was examined with a D'Agostino-Pearson normality test. Statistical comparisons between two groups were determined with a nonparametric Mann-Whitney test. Statistical comparisons between three or more groups were performed with one-way or two-way ANOVA, followed by Dunn's or Sidak's multiple comparisons test, using Prism 8 (GraphPad Software, San Diego, CA). p Ͻ 0.05 was considered statistically significant