606ENaC Proteolytic Regulation by Channel-activating Protease 2

8/7/2019 606ENaC Proteolytic Regulation by Channel-activating Protease 2

1/15

TheJo

urnalofGeneralPhysiolo

gy

A R T I C L E

The Rockefeller University Press $30.00J. Gen. Physiol. Vol. 132 No. 5 521535www.jgp.org/cgi/doi/10.1085/jgp.200810030 521

I N T R O D U C T I O N

Abnormal activity o the epithelial sodium channel (ENaC)is implicated in diseases o the cortical collecting ducts othe kidney (Liddle et al., 1963), in the airways (Boucher,2004; Schild, 2004), and in the middle ear (Guipponiet al., 2002). Novel therapies or these disorders may ol-low rom better understanding o ENaC regulation. Workrom several groups identiy proteases as important ENaC

regulators (Vallet et al., 1997, 2002; Vuagniaux et al., 2000,2002; Donaldson et al., 2002; Hughey et al., 2003, 2004;Caldwell et al., 2004, 2005; Harris et al., 2007). The resultso these studies suggest that ENaC is activated by its partialand selective proteolysis during channel assembly andprocessing or while channels are resident in the plasmamembrane. Despite this progress, detailed knowledgeo the molecular mechanism(s) underlying proteolyticactivation o ENaC is limited.

Vallet et al. (1997) obtained the frst evidence o an epi-thelial membrane proteaseactivating ENaC in an auto-crine ashion. They cloned a channel-activating protease(CAP), or CAP1 (prostasin), rom a Xenopuskidney epi-

thelial cell line (A6) and established that it activates ENaCwhen coexpressed in Xenopusoocytes (Vallet et al., 1997).Subsequently, two additional serine proteases, mCAP2

Correspondence to Agustn Garca-Caballero: [email protected] Abbreviations used in this paper: CAP, channel-activating protease;

ENaC, epithelial sodium channel; HA-NT, HAN-terminal; hNE, humanneutrophil elastase; MBS, modifed Barths solution; MTSET, [2-(trimethyl-ammonium)ethyl] methanethiosulonate bromide; PO, open probability;V5-CT, V5C-terminal; WT, wild-type.

(homologue o human transmembrane protease serine4 [TMPRSS4]) and mCAP3 (MT-SP1/Matriptase or epi-thin), were identifed by homology cloning and ound toincrease the activity o ENaC coexpressed in Xenopusoo-cytes (Vuagniaux et al., 2002). These studies ound no e-ects o CAPs on the number o channels at the surace,suggesting that CAPs increase ENaC activity by changing

open probability (PO) (Vallet et al., 2002; Vuagniaux et al.,2002; Andreasen et al., 2006). Direct evidence that prote-ases can increase ENaC PO was frst provided by Caldwellet al. (2004, 2005). They ound that trypsin or human neu-trophil elastase (hNE) added to the outer ace o outside-out patches o NIH-3T3 ENaC cells increased PO o nearsilent ENaCs by up to 28-old (Caldwell et al., 2004, 2005).Subsequently, a 65-kD -ENaC ragment generated at thesurace by hNE was linked to hNE-stimulated ENaC cur-rent (Harris et al., 2007). Adebamiro et al. (2007) identi-fed specifc residues in -ENaC that were required orelastase to stimulate ENaC current.

ENaC is a multimeric channel consisting o topologi-

cally similar -, -, and -subunits (Canessa et al., 1994).A key link between protease-mediated cleavage o ENaCand channel activity was discovered when Hughey et al.(2003, 2004) detected minimal consensus cleavage

ENaC Proteolytic Regulation by Channel-activating Protease 2

Agustn Garca-Caballero, Yan Dang, Hong He, and M. Jackson Stutts

Cystic Fibrosis/Pulmonary Research and Treatment Center, University of North Carolina, Chapel Hill, NC 27599

Epithelial sodium channels (ENaCs) perorm diverse physiological roles by mediating Na+ absorption across epi-thelial suraces throughout the body. Excessive Na+ absorption in kidney and colon elevates blood pressure and inthe airways disrupts mucociliary clearance. Potential therapies or disorders o Na + absorption require better un-derstanding o ENaC regulation. Recent work has established partial and selective proteolysis o ENaCs as an im-portant means o channel activation. In particular, channel-activating transmembrane serine proteases (CAPs) andcognate inhibitors may be important in tissue-specifc regulation o ENaCs. Although CAP2 (TMPRSS4) requirescatalytic activity to activate ENaCs, there is not yet evidence o ENaC ragments produced by this serine proteaseand/or identifcation o the site(s) where CAP2 cleaves ENaCs. Here, we report that CAP2 cleaves at multiple sitesin all three ENaC subunits, including cleavage at a conserved basic residue located in the vicinity o the degenerinsite (-K561, -R503, and -R515). Sites in -ENaC at K149/R164/K169/R177 and urin-consensus sites in -ENaC(R205/R231) and -ENaC (R138) are responsible or ENaC ragments observed in oocytes coexpressing CAP2.However, the only one o these demonstrated cleavage events that is relevant or the channel activation by CAP2takes place in -ENaC at position R138, the previously identifed urin-consensus cleavage site. Replacement oarginine by alanine or glutamine (,,R138A/Q) completely abolished both the Na+ current (INa) and a 75-kD

-ENaC ragment at the cell surace stimulated by CAP2. Replacement o-ENaC R138 with a conserved basic resi-due, lysine, preserved both the CAP2-induced INa and the 75-kD -ENaC ragment. These data strongly support amodel where CAP2 activates ENaCs by cleaving at R138 in -ENaC.

2008 Garca-Caballero et al. This article is distributed under the terms of an AttributionNoncommercialShare AlikeNo Mirror Sites license for the first six months after the publi-cation date (see http://www.jgp.org/misc/terms.shtml). After six months it is available undera Creative Commons License (AttributionNoncommercialShare Alike 3.0 Unportedlicense, as described at http://creativecommons.org/licenses/by-nc-sa/3.0/).


2/15

522 Identification of Cleavage Sites on ENaC by CAP2

K156Q, K168Q, K170Q, R172Q, R178Q/K179Q/R180Q/K181Q,K185Q, K189Q, K200Q, K201Q, K504A, K510A, and R515A;CAP2: S387A) were generated by PCR and cloned into pCR-BluntII-TOPO (Invitrogen), linearized (HindIII), and in vitrotranscribed using T7 RNA polymerase. A PolyA tail was addedater transcription (Ambion). Mutations were perormed withthe quickchange multisite-directed mutagenesis kit (Stratagene).The WT ENaC plasmids were provided by B. Rossier (Univer-

sit de Lausanne, Lausanne, Switzerland). The sequence o allplasmids was verifed at the University o North Carolina se-quencing acility.

Functional Studies of Na+ Channels in Xenopus OocytesVVI-stage healthy oocytes were harvested as described previously(Donaldson et al., 2002) and maintained in modifed Barthssolution (MBS) at 18C. Animals were maintained and studiedunder protocols approved by the University o North CarolinaInstitutional Animal Care and Use Committee. cRNAs encoding

WT o both untagged and HA-NT and V5-CTtagged subunits ormutant HA-NT plus V5-CTtagged subunits rat-, -, and -ENaC(0.3 ng each) and CAP2 (TMPRSS4) cRNA (1 ng) were coin-

jected into oocytes. 24 h ater injection, two-electrode voltageclamping was perormed using a Genclamp amplifer (MDS Ana-lytical Technologies) in a constant perusion system. Currents

were measured in the presence and absence o 10 M amiloride,with membrane voltage clamped to100 mV. Currents were digi-tized and recorded using a Digidata 1200 A/D converter (MDS

Analytical Technologies) and Axoscope sotware. 2 g/ml trypsinwas perused or 5 min ater frst measuring the basal amiloridesensitive current (INa) o each egg. Some eggs were incubatedovernight with 50 g/ml aprotinin, as indicated. All results areexpressed as the mean SE or as old stimulation by CAP2 ortrypsin. The means o two groups were tested or signifcant di-erence using an unpaired Students t test, and dierences be-tween three or more groups were evaluated using ANOVA analysis(GraphPad Sotware, Inc.). Proteins extracted rom control andinjected oocytes were analyzed by Western blots to veriy expressiono ENaC and actin.

Surface LabelingXenopusoocytes were injected with desired combinations o WTor mutant double epitopetagged , , and rat ENaC subunits(0.3 ng each) and with or without CAP2 or CAP1 cRNA (1 ng).

Ater 24 h, 70 oocytes per experimental condition were pre-chilled on ice or 30 min and labeled with 0.7 mg/ml sulo-NHS-biotin in MBS-Ca2+ (mM), 85 NaCl, 1 KCl, 2.4 NaHCO3, 0.82MgSO4, 0.41 CaCl, 0.33 Ca(NO3), 16.3 Hepes titrated to pH 8.0

with NaOH, while tumbling gently or 20 min at 4C. Oocyteswere washed twice with chilled MBS-Ca2+ buer and incubated inMBS-Ca2+ buer with 100 mM glycine or 10 min at 4C to quenchree biotin. Oocytes were washed again three times with chilledMBS-Ca2+ buer, and then lysed with lysis buer (in mM: 20Tris, 50 NaCl, 50 NaF, 10 -glycerophosphate, 5 Na4P2O7 pyro-phosphate, 1 EDTA, pH 7.5, containing protease inhibitors [com-

plete; Roche], aprotinin [Sigma-Aldrich]). Cell lysates wereprepared by passing oocytes through a 27G1/2 needle twice andby centriugation at 3,600 rpm or 10 min at 4C. Supernatants

were transerred to new tubes, and samples were spun at 14,000rpm or 20 min at 4C. Supernatants were discarded and pellets

were solubilized in solubilization buer (in mM: 50 Tris, 100 NaCl,1% Triton X-100, 1% NP-40, 0.2% SDS, 0.1% Na deoxycholate,20 NaF, 10 Na4P2O7 pyrophosphate, 10 EDTA plus protease inhibitorcocktail, pH 7.5). Total inputs were taken rom whole cell samplesrepresenting 4% o total protein. Solubilized proteins were incu-bated with 100 l o neutravidin beads (Thermo Fisher Scientifc)overnight while tumbling at 4C. Samples were washed twice with500 mM NaCl, 50 mM Tris, pH 7.5, buer and once with 150 mM

sequences or convertases o the urin amily in the- and -ENaC subunits. They demonstrated that muta-genesis o these sites eliminated specifc ragments o-and -ENaC and caused a reduced basal ENaC currentthat was recovered by application o exogenous trypsin.Although urin amily convertases are known to cleaveproteins during trafcking to the cell membrane, they

are also active at the cell surace (Thomas, 2002), leav-ing some doubt as to the subcellular location o urinaction on ENaC. In addition, whereas mutagenesis ourin consensus sequences in -ENaC had the largesteect on basal ENaC current (Hughey et al., 2004), theprecise roles o urin-mediated cleavage o the two -and single -ENaC urin sites are not known.

Based on these observations, a simple model or ENaCregulation by proteases emerged (Planes and Caughey,2007). Cleavage o- and -subunits by urin or relatedconvertases is assumed to occur during channel assem-bly and delivery to the cell membrane, but to a variableextent. This results in channels arriving at the surace

with a range o proteolytic activation. Uncleaved or per-haps mixed multimers o cleaved and uncleaved sub-units are thought to be susceptible to urther cleavageand stimulation by surace or soluble proteases (Planesand Caughey, 2007).

CAPs are the main candidates identifed thus ar assurace-bound proteases that may activate ENaC at thecell surace. A polybasic stretch o residues in -ENaCwas recently identifed by mutagenesis as being requiredor CAP1 stimulation o ENaC (Bruns et al., 2007). How-ever, studies comparing wild-type (WT) and catalyticallyinactive CAP1-3 mutants have shown that catalyticallyinactive CAP1 ully stimulates ENaC, suggesting it hasan indirect role in activating ENaC (Andreasen et al.,2006). Thus, the mechanism o CAP1 regulation is notentirely clear. In contrast to CAP1, both CAP2 andCAP3 must possess catalytic activity to stimulate ENaC(Andreasen et al., 2006). Here, we have ocused on theactivation o ENaC by CAP2. Previously, the sites withinENaC where CAP2 cleaves were not known, nor hadCAP2-mediated cleavage at a specifc ENaC residue beenshown to cause an increase in ENaC PO. Here, we reportthat CAP2 cleaves all three ENaC subunits, both withand without associated stimulation.

M AT E R I A L S A N D M E T H O D S

Plasmid PreparationFor biochemical analyses o ENaC subunit proteolysis, cDNAsencoding rat-, -, and -ENaC with HAN-terminal (HA-NT)and V5C-terminal (V5-CT) epitope tags were generated. WTand mutant constructs (-ENaC: K149A/R164A/K169A/R177A,R205A/R231A, K486A, K501A, R503A, K504A, K512A, R519A,K524A, K544A, R545A, K550A, and K561A; -ENaC: K411A,R416A, R435A, K443A, K452A, R477A, R487A, K488A, K492A,R503A, and S518C; -ENaC: R138A, R138K, R138Q, R153Q,


3/15

Garca-Caballero et al. 523

and -subunits and a mutant S518C subunit. The cys-teine-reactive compound [2-(trimethylammonium)ethyl]methanethiosulonate bromide (MTSET) locks suchchannels in the open state (PO = 1.0) and has been usedas a means o estimating channel density as well as restingPO (Kellenberger et al., 2002). INa o WT ENaC was notstimulated by 1 mM MTSET (not depicted). However,

NaCl, 50 mM Tris, pH 7.5, buer. Laemmli buer was addedand samples were loaded on a 412% gradient Tris-glycine gelater incubation or 10 min at 96C. Samples were transerred to0.45 m polyvinylidene diuoride membranes (Millipore), and

Western blot analysis was perormed using an anti-V5 (Invitro-gen), anti-HA (Covance), and anti-actin (Millipore) monoclonalantibodies. Surace ENaC-biotinylated ragments were quantifedusing the metamorph imaging 4.5 program (Hooker Microscopy

Facility, University o North Carolina). Densitometry o selectedbands was perormed using uninjected oocyte samples as back-ground signal.

Western Blot AnalysisProteins were extracted rom oocytes as described above. Biotin-

ylated and total proteins were solubilized by boiling in Laemmlisample buer or 10 min beore loading onto 412% SDS-PAGEgels. Western blots were perormed with anti-V5 (Invitrogen),anti-HA (Covance), anti-express (Invitrogen), and anti-actin (Mil-lipore) antibodies.

R E S U L T S

We applied unctional and biochemical approachesin parallel to assess the eects o CAP2 on the activity,number, and cleavage o coexpressed ENaC in Xenopusoocytes. To establish that our experimental conditionswere inormative or these analyses, we frst observedthat coexpression o CAP2 with WT ENaC elevatedbasal INa by2.5-old over WT ENaC alone (Fig. 1 A).Whereas the basal INa o WT ENaC alone was stimulated3.5-old by the application o trypsin, the already ele-vated INa o oocytes coexpressing CAP2 and ENaC wasnot stimulated urther by exogenous trypsin. We rou-tinely observed a reduction in INa ater 5 min o trypsinexposure o oocytes coexpressing ENaC and CAP2 (Fig. 1,

A and B). This apparent inhibitory eect o trypsin ismainly due to rundown o the CAP2-stimulated chan-nels during the interval o trypsin application.

This result indicates that ENaC at the surace o oo-cytes expressing CAP2 had undergone a proteolytic acti-vation process that was not shared by all channels in thecontrol (ENaC alone) oocytes. To biochemically detectENaC cleavage concomitant with CAP2 coexpression, weused modifed -, -, and -ENaC subunits with HA-NTand V5-CTtagged subunits (Hughey et al., 2003). Withthis approach, one injected subunit was the double epi-topetagged version o-, -, or -ENaC, combined withthe appropriate WT versions. We frst confrmed that

CAP2 stimulated ENaC containing the HA-NT/V5-CTtagged subunits similarly to WT ENaC (e.g., Fig. 1, B).The eects o CAP2 coexpression and trypsin on INa me-diated by these tagged channels were virtually identicalto its actions on WT ENaC. Thus, epitope-tagged ENaCsubunits can be used to ollow the proteolytic actions oCAP2 on ENaC (compare A and B in Fig. 1).

The stimulation o ENaC by active trypsin or elastaseis due to an increase in ENaC PO (Caldwell et al., 2004,2005). To test i CAP2 coexpression increased ENaC PO,we coexpressed CAP2 with ENaC consisting o WT -

Figure 1. CAP2 and trypsin eects on amiloride-sensitive INa

o WT or HA-NT and V5-CTtagged or -S518C mutant chan-nels in oocytes. WT -, -, -, or double tagged (HA-NT/V5-CT)and mutant-, -S518C, and -cRNA (0.3 ng each) ENaC sub-units and 1 ng CAP2 cRNA were injected into oocytes. 24 h aterinjection, two-electrode voltage clamp assays were conducted.Currents measured in the presence and absence o 10 M amil-oride, while clamping the membrane voltage to 100 mV, weredigitized and recorded. (A) Stimulation o ENaC WT channelsby coexpression o CAP2 and 2 g/ml o exogenous trypsin(n= 24). (B) CAP2 and trypsin (2 g/ml) stimulation o ENaC-, -, and -HA-NT/V5-CTtagged subunits (n= 18). (C) 1 mMMTSET activation o-, -S518C, and mutant untagged chan-nels with or without CAP2 (n= 24). (AC) Batches o oocytes

were extracted rom our to fve dierent rogs. Results are ex-pressed as the means SE. * and **, P < 0.0001, signifcant di-erence when CAP2- or trypsin-stimulated INa is compared withcontrol basal INa. Statistical signifcance was determined usingan unpaired Students ttest.


4/15


in Fig. 2 B, mutation o the catalytic triad o CAP2 abol-ished the proteases ability to stimulate basal INa.

The observation that trypsin and MTSET do not stim-ulate the INa o ENaC coexpressed with active CAP2indicates that ENaC at the surace o these oocytes aremaximally proteolytically stimulated to a high PO. Toidentiy ENaC cleavage(s) in the presence o CAP2 that

could increase PO, we examined the ragmentation pat-tern o ENaC when coexpressed with CAP2 using HA-NT and V5-CTtagged subunits (Hughey et al., 2003)(Fig. 3). With ENaC alone, we typically observed a pat-tern o ragments similar to that reported (Hughey et al.,2003). Western blots stained or epitope-tagged -ENaCor -ENaC each variably demonstrated ragments (-ENaC HA-NT,32 kD; -ENaC HA-NT,18 kD; -ENaCV5-CT,66 kD; -ENaC V5-CT,75 kD), consistent withcleavage at identifed consensus cleavage sites or mem-bers o the protein convertase amily, whereas Westernblots o-ENaC, as reported previously, did not indicatecleavage under basal conditions (Hughey et al., 2003).

CAP2 coexpression decreased the amount o ull-lengthprotein corresponding to all three ENaC subunits inwhole cell lysates (Fig. 3). The loss o ull-length ENaCcoincided with the somewhat variable appearance osmaller ragments. CAP2 coexpression consistently en-hanced 32-kD HA-NT and complementary66-kDV5-CT staining bands o-ENaC, suggesting increasedcleavage at the urin-consensus sites (Fig. 3, A and B).In some experiments, CAP2 coexpression with -ENaCgenerated an HA-NT band o82 kD (Fig. 3 A), andmore consistently yielded an 17-kD V5-CT band (Fig.3 B). These complementary-ENaC ragments predict

MTSET typically stimulated INa o eggs coexpressing , S518C, and -ENaC by three- to sixold (Fig. 1 C). Ineggs injected with , S518C, -ENaC plus CAP2, basalINa was stimulated 2.7-old as beore. The addition oMTSET to these eggs did not change INa (Fig. 1 C). Theacute eect o MTSET on the INa o, S518C, -ENaCwas previously attributed to the net o MTSETs actions

o simultaneously increasing PO to near 1.0 and reduc-ing single channel conductance by 30% (Kellenbergeret al., 2002). Thereore, the absence o change in INa in-duced by MTSET in eggs coexpressing, S518C, -ENaCand CAP2 indicates that the PO o ENaC at the suraceo oocytes expressing CAP2 is in the range o 0.7, ingood agreement with Caldwells estimation o the PO otrypsin-treated ENaC (Caldwell et al., 2004).

To establish cleavage o ENaC in the presence o CAP2as a potential mechanism or stimulation o INa, we nextasked i CAP2 catalytic activity was required or stimula-tion o coexpressed ENaC. We addressed this question intwo ways. Because CAP2-mediated proteolysis is inhib-ited by aprotinin, we asked i aprotinin blocked CAP2regulation o ENaC. Overnight incubation o oocytes with 50 g/ml aprotinin prevented the stimulation ocurrents typically seen rom coexpressing ENaC and CAP2(Fig. 2 A). This result does not rule out indirect action oCAP2 on ENaC through another aprotinin-sensitive pro-tease, as has been proposed or CAP1 (Andreasen et al.,2006). Thereore, we constructed an inactive CAP2control by mutating serine 387 in its catalytic triad(CAP2s387a) and tested its eect on ENaC. This muta-tion alters the HDS triad at the catalytic site o CAP2 re-sulting in inactivation (Andreasen et al., 2006). As shown

Figure 2. Catalytic activity o CAP2 is required tostimulate ENaC. (A) INa stimulated by CAP2 is sen-sitive to aprotinin. Eggs coexpressing ENaC plus

CAP2 were preincubated or not with 50 g/ml apro-tinin overnight (n= 18). ENaC WT plus CAP2 versusENaC WT plus CAP2 plus aprotinin. *, P < 0.0001.(B) Inactive CAP2 s387a (CAP2s387a) does not acti-

vate ENaC. Amiloride-sensitive INa stimulated by WTbut not by inactive CAP2 (n= 30). (A and B) Batcheso oocytes were extracted rom our to fve dierentrogs. Results are expressed as the means SE. *,P < 0.0001. Statistical signifcance was determined us-ing an unpaired Students ttest. (C) Inactive CAP2is expressed as a ull-length precursor. Western blotso WT and inactive express-tagged CAP2. Represen-tative experiment is shown.


5/15

Garca-Caballero et al. 525

CAP2 coexpression induced the ormation o HA-NT-ENaC ragments o47 and 18 kD (Fig. 3 A). Wedid not pursue the 47-kD ragment due to a lack ocomplementarity with a corresponding V5-CT ragmento the expected size. The HA-NT band at 18 kD matchesa ragment reported to be generated by a convertaseamily member, urin (Hughey et al., 2003). CAP2 induced

a complementary V5-CT ragment o-ENaC at75kD, also consistent with cleavage at the -ENaC urin site.There was also an 15-kD V5-CT -ENaC band in ly-sates o oocytes expressing active CAP2 (Fig. 3 B). Thisband suggests cleavage at a site near the pre-M2 regiono-ENaC (Fig. 3 B).

To identiy those ENaC ragments at the cell suracethat could have derived rom the catalytic activity o CAP2,we compared ragments detected when ENaC was coex-pressed with CAP2 or a catalytically inactive CAP2 mu-tant. With coexpression o inactive CAP2 (CAP2 S387A),we observed a migration pattern o ENaC on gels, cor-responding to ull-length subunits and ENaC urinragments indicative o endogenous urin-like activity,indistinguishable rom WT ENaC alone (Fig. 3, C and D).In addition, INa due to ENaC coexpressed with CAP2S387A was similar to that o ENaC alone (Fig. 2 B).

cleavage close to the second transmembrane, or pre-M2, region. The relationship between CAP2-stimulatedcleavage at the -ENaC urin and pre-M2 regions toCAP2-stimulated INa is addressed in detail below. In ad-dition, unique HA-NT ragments o-ENaC o17 and19 kD were induced by CAP2, consistent with cleavagejust distal o the frst transmembrane segment (Fig. 3 A).

Mutagenesis o our conserved basic residues in thisregion to alanine (K149A/R164A/K169A/R177A) pre-vented 17 and 19 kD N-terminal ragments withCAP2 coexpression, but it did not aect stimulation oINa (Fig. 4). Thus, we conclude that CAP2-mediatedcleavage in the proximal region o the extracellularloop o the -ENaC subunit is not required or CAP2 tostimulate ENaC.

Previous work with other proteases revealed no cleav-age o-ENaC (Hughey et al., 2003; Harris et al., 2007).To our surprise, CAP2 induced 90-kD HA-NT and20-kD V5-CT ragments (Fig. 3, A and B) o doubleepitopetagged -ENaC, which complement each other,as the expected size o ull-length -ENaC protein is110 kD. These -ENaC ragments indicate that, simi-lar to its action on -ENaC, CAP2 promotes cleavage o-ENaC at the pre-M2 region.

Figure 3. Cleavage o-,-, and-ENaCtagged sub-units by WT but not inactive CAP2 S387A in oocytes.

WT -, -, and -ENaCdouble tagged (HA-NT/V5-CT) and -, -, or -ENaCuntagged subunits (0.3ng each) and 1 ng CAP2 cRNA were injected intooocytes. Total protein lysates were prepared romoocytes, and Western blots analysis were conductedusing anti-HA (A) and anti-V5 (B) monoclonal anti-

bodies. Western blots o surace biotinylated and to-tal pools o-ENaC HA-NT (C) and -ENaC V5-CT(D) ragments caused by WT but not inactive CAP2(CAP2 S387A). Actin expression was detected withan anti-actin monoclonal antibody as control (D).

A-ENaC 90-kD HA-NT ragment was detected atthe surace pool (not depicted). NS, nonspecifc.Batches o oocytes were extracted rom three di-erent rogs. Representative experiments are shown(n= 3). (E) Linear diagram o ENaC residues cleavedby CAP2. NH2, amino terminus; COOH, carboxy ter-minus; TM1, transmembrane domain 1; TM2, trans-membrane domain 2.


6/15


66-kD V5-CT band; or -ENaC, an HA-NT band o18kD complemented by a V5-CT band o75 kD (Fig. 3).The surace pool also contained bands (HA-NT o90kD and V5-CT o20 kD) o-ENaC consistent withcleavage in the pre-M2 region (not depicted).

Thus, active CAP2 induces ragments in all three sub-units at the oocyte surace. Because these ragments rep-

resent novel cleavage events at the pre-M2 region in -, -,and -ENaC, and apparent increased cleavage at theconvertase consensus site in - and -ENaC, we directedadditional studies to these regions. We used site-directedmutagenesis to locate the specifc residues cleaved andto link specifc cleavages to stimulation o INa.

We frst sought to identiy specifc residues necessaryor CAP2-stimulated ragments in the pre-M2 regionso-, -, and -ENaC. We analyzed -, -, and -ENaCsequences corresponding to the extracellular loop at thepre-M2 region with a cleavage prediction model or CAP3/matriptase available at http://pops.csse.monash.edu.au/pops.html. Although there is no established model to

predict cleavage sites or CAP2, we have observed thatCAP2 and CAP3/matriptase generate an identical C ter-minalragmentation pattern in the -ENaC subunit(not depicted). From this observation, we reasoned thatthese CAPs could share some substrate afnity. Based onthe apparent molecular mass o the 1517-kD C termi-nal ragments generated by CAP2 in all three subunits,we surveyed a stretch o 40 amino acids in the pre-M2region o-, -, and -ENaC. Whereas multiple basic res-idues in this region had high predictive scores or mat-riptase cleavage, we noted an aligned basic residue atposition K561 in -ENaC, R503 in -ENaC, and R515 in

-ENaC that was conserved across all three subunits ohuman, mouse, and rat ENaCs (Fig. 5 B). When wereplaced basic residues at this residue with alaninesby site-directed mutagenesis, the appearance o the pre-M2 ragments with CAP2 coexpression was eliminated(Fig. 5 A). Interestingly, in the mutants there was a trendtoward lighter intensity o bands associated with cleav-age at the urin sites in - and -ENaC. Mutagenesis oother basic residues (-ENaC: K486A, K501A, R503A,K504A, K512A, R519A, K524A, K544A, R545A, K550A,and K556A; -ENaC: K411A, R416A, R435A, K443A,K452A, R477A, R487A, K488A, and K492A; -ENaC:K504A and K510A) in the 80-residue stretch had no e-

ects on CAP2-mediated ragments (not depicted).The highly conserved residue we ound to be required

or pre-M2 CAP2 ragments is 15 residues upstream o thedegenerin site present in each subunit (Fig. 5 B). To assessthe unctional eect o this cleavage event in the pre-M2region o the channel, we coexpressed a triple mutantchannel (-K561A,-R503A, and-R515A) in oocytes withor without CAP2 and measured INa. We used trypsin toevaluate the proteolytic state o the WT or mutant chan-nels, given previous reports by Caldwell et al. (2004) thattrypsin activates near-silent channels in excised patches.

To be sure the lack o any eect was not due to poorexpression o the mutant CAP2, we perormed Westernblot analyses o WT and mutant CAP2 expression. Theseresults indicated that the mutant was efciently expressed

(Fig. 2 C). As previously observed (Andreasen et al.,2006), WT CAP2 appears as the glycosylated precursor(55 kD) and cleaved N-terminus glycosylated orms(30 and 20 kD), whereas mutant CAP2 S387A ap-pears as a single precursor band o55 kD (Fig. 2 C). Inpaired experiments with active CAP2, the surace poolcontained the ragments o - and -ENaC describedabove, which indicated cleavage in the pre-M2 region.In addition, we routinely observed increased intensityo the bands believed to represent urin ragments: In -ENaC, an HA-NT band o32 kD, complemented by an

Figure 4. Eect o-ENaC K149A/R164A/K169A/R177A mu-tant on CAP2-induced INa and -ENaC N-terminal ragments.

(A) Western blots o-ENaC N-terminal 17- and 19-kD novel rag-ments at the surace (top) and total pools (bottom). Lane 1, unin-

jected eggs; lane 2, ENaC alone; lane 3, ENaC plus CAP2; lane 4,-ENaC mutant alone; lane 5, -ENaC mutant plus CAP2. FL, ull-length. A representative experiment is shown (n= 3). (B) INa o

WT and mutant-ENaC stimulated by CAP2. Batches o oocyteswere extracted rom three dierent rogs (n= 18). * and **, P

Documents

606ENaC Proteolytic Regulation by Channel-activating Protease 2